ESMO 2026 GU Cancer Slides

[Slide 1] VOLGA: Study Design Patients • Adults aged ≥18 years • MIBC (cT2-T4aN0/1M0 or cT1N1M0) by local assessment • Ineligible for or refused cisplatin • UC or UC with divergent differentiation or histological subtypes • Evaluated and confirmed for cystectomy R (1:1:1) N=695 Neoadjuvant D+EV N=233 | Durvalumab Q3W × 3 cycles | Enfortumab vedotin Q3W × 3 cycles D+T+EV N=233 | Durvalumab Q3W × 3 cycles | Tremelimumab Q3W × 2 cycles | Enfortumab vedotin Q3W × 3 cycles Control N=229 | No treatment | Straight to RC RC Adjuvant D+EV: Durvalumab Q4W × 9 cycles D+T+EV: Durvalumab Q4W × 9 cycles; Tremelimumab Q4W × 1 cycle Control: Observation/locally approved treatment Endpoints Dual primary – EFSa • D+EV vs control • D+T+EV vs control Key secondary – OS • D+EV vs control • D+T+EV vs control Other secondary • pCR • EORTC QLQ-C30 Exploratory • ctDNA analysis Safety Stratification • PD-L1 status (high vs low/negative) • Clinical tumour stage (T2N0M0 vs T3-4aN0M0 vs T1-4aN1M0) • Cisplatin status (ineligible vs refusal) EFS was defined as time from randomisation to the first: • Recurrence after RC • Progression in patients who did not receive RC • Failure to undergo RC in patients with residual disease • Death due to any cause Durvalumab dose was 1500 mg, tremelimumab dose was 75 mg, and EV dose was 1.25 mg/kg body weight. aEvaluated by BICR or central pathology review (if a biopsy was required for a suspected new lesion). ClinicalTrials.gov, NCT04960709. EudraCT number, 2020-005452-38. BICR, blinded independent central review; ctDNA, circulating tumour DNA; D, durvalumab; EFS, event-free survival; EORTC QLQ-C30, European Organisation for Research and Treatment of Cancer 30-item Core Quality of Life Questionnaire; EV, enfortumab vedotin; MIBC, muscle-invasive bladder cancer; OS, overall survival; pCR, pathological complete response; PD-L1, programmed cell death ligand-1; Q3W, every 3 weeks; Q4W, every 4 weeks; R, randomisation; RC, radical cystectomy; T, tremelimumab; UC, urothelial carcinoma. [Slide 2] VOLGA: EFS by BICR for Durvalumab+EV vs Control – ITT Durvalumab+EV demonstrated a statistically significant improvement in EFS vs control | D+EV N=233 | Control N=229 Events, n (%) | 84 (36) | 123 (54) Median EFS, months (95% CI) | NR (33.2–NC) | 12.1 (7.6–27.0) HR (95% CI) | 0.51 (0.38–0.67) Stratified log-rank P valuea | <0.0001 Curve labels: 12 mo: 73% (D+EV), 50% (Control); 24 mo: 64% (D+EV), 45% (Control) Durvalumab+EV arm; Control arm Median follow-upb: 23 months Y axis: Probability of EFS; X axis: Time from randomisation (months) 0–45 No. of patients at risk D+EV | 233 | 197 | 180 | 170 | 158 | 136 | 111 | 93 | 80 | 51 | 38 | 21 | 11 | 4 | 1 | 0 Control | 229 | 157 | 131 | 106 | 98 | 80 | 66 | 59 | 43 | 35 | 22 | 16 | 8 | 5 | 0 | 0 EFS is the time from randomisation to the first: 1) recurrence after RC, 2) progression in patients who did not receive RC, 3) failure to undergo RC in patients with residual disease, or 4) death due to any cause. aThe threshold to declare statistical significance was based on a based on a Lan-DeMets alpha spending function with O'Brien-Fleming boundary – with the observed number of events, the boundary for declaring statistical significance was 0.0218 for a 2.7% overall 2-sided alpha. bIn censored patients across all study arms. Data cutoff 02 March 2026. BICR, blinded independent central review; CI, confidence interval; D, durvalumab; EFS, event-free survival; EV, enfortumab vedotin; HR, hazard ratio; ITT, intention-to-treat population; NC, not calculated; NR, not reached; RC, radical cystectomy. Survival table Endpoint: EFS by BICR (ITT) Arm | N | Median | Median CI D+EV | 233 | NR | 33.2–NC Control | 229 | 12.1 | 7.6–27.0 HR: 0.51 HR CI: 0.38–0.67 P: <0.0001 Landmark rates: 12 mo: D+EV 73% | Control 50% 24 mo: D+EV 64% | Control 45% [Slide 3] VOLGA: OS for Durvalumab+EV vs Control – ITT Durvalumab+EV demonstrated a statistically significant improvement in OS vs control | D+EV N=233 | Control N=229 Deaths, n (%) | 49 (21) | 74 (32) Median OS, months (95% CI) | NR (NC–NC) | NR (37.7–NC) HR (95% CI) | 0.58 (0.40–0.83) Stratified log-rank P valuea | 0.0027 Curve labels: 12 mo: 87% (D+EV), 81% (Control); 24 mo: 78% (D+EV), 69% (Control) Durvalumab+EV arm; Control arm Median follow-upb: 26 months Y axis: Probability of OS; X axis: Time from randomisation (months) 0–48 No. of patients at risk D+EV | 233 | 224 | 212 | 207 | 198 | 185 | 160 | 128 | 110 | 82 | 66 | 45 | 27 | 15 | 4 | 0 | 0 Control | 229 | 208 | 190 | 185 | 176 | 157 | 128 | 110 | 96 | 72 | 53 | 34 | 22 | 13 | 5 | 2 | 0 OS is the time from randomisation until death due to any cause, regardless of whether the patient withdraws from randomised therapy or receives another anticancer therapy. aThe threshold to declare statistical significance was based on a Gamma family (-3) spending function – with the observed number of events, the boundary for declaring statistical significance was 0.0104 for a 5% overall 2-sided alpha. bIn censored patients across all study arms. Data cutoff 02 March 2026. CI, confidence interval; D, durvalumab; EV, enfortumab vedotin; HR, hazard ratio; ITT, intention-to-treat population; NC, not calculated; NR, not reached; OS, overall survival. Survival table Endpoint: OS (ITT) Arm | N | Median | Median CI D+EV | 233 | NR | NC–NC Control | 229 | NR | 37.7–NC HR: 0.58 HR CI: 0.40–0.83 P: 0.0027 Landmark rates: 12 mo: D+EV 87% | Control 81% 24 mo: D+EV 78% | Control 69% [Slide 4] VOLGA: AE Summary – Safety Population (Overall Trial Period) Events, n (%) | Durvalumab+EV n=228 | Control n=229 AEs of any cause | 227 (>99) | 170 (74) Maximum grade 3 or 4 AEs | 145 (64) | 101 (44) Serious AEs | 143 (63) | 101 (44) Leading to death | 17 (7) | 14 (6) Leading to discontinuation of any study treatment | 48 (21) | NA Leading to discontinuation of durvalumab | 37 (16) | NA Leading to discontinuation of tremelimumab | NA | NA Leading to discontinuation of enfortumab vedotin | 26 (11) | NA Leading to cancellation of surgery | 2 (0.9) | 0 Leading to delay in surgerya | 8 (4) | 1 (0.4) AEs possibly related to any treatment | 207 (91) | NA Maximum grade 3 or 4 AEs | 69 (30) | NA Serious AEs | 45 (20) | NA Leading to death | 1 (0.4) | NA Any-grade immune-mediated AEs | 75 (33) | 8 (3) Maximum grade 3 or 4 | 24 (11) | 2 (0.9) Leading to death | 0 | 0 The safety population includes all patients who received study treatment (D+EV arm) or who were randomised (control arm). Investigator-assessed causality. aRecommended timeframe for surgery was within 56 days after the last dose of neoadjuvant treatment (D+EV arm), and within 12 weeks of diagnosis (control arm). Data cutoff 02 March 2026. AE, adverse event; D, durvalumab; EV, enfortumab vedotin; NA, not applicable.

[Slide 1] Proffered Paper session LBA2 - Perioperative durvalumab with or without tremelimumab plus neoadjuvant enfortumab vedotin in muscle-invasive bladder cancer: Results from the phase 3 VOLGA trial Speakers Thomas B. Powles (London, United Kingdom) Lecture Time 16:54 - 17:06 [Abstract button] Background The standard of care continues to evolve for patients (pts) with muscle-invasive bladder cancer (MIBC). We evaluated if neoadjuvant durvalumab (D) ± tremelimumab (T) plus enfortumab vedotin (EV) followed by radical cystectomy (RC) and adjuvant D±T would be an effective option. [Slide 2] Results Overall, 233, 233, and 229 pts were randomised to the D+EV, D+T+EV, and control arms, respectively. RC was completed in 85% (D+EV), 79% (D+T+EV), and 78% (control) of pts; and 55% (D+EV) and 42% (D+T+EV) completed adjuvant tx. Median EFS follow-up was 22.9 mo. D+EV showed significant improvements in EFS (HR 0.51 [95% CI, 0.38-0.67], P<0.0001; 24-mo rate 64% vs 45%) and OS (HR 0.58 [95% CI, 0.40-0.83], P=0.0027; 24-mo rate 78% vs 69%) vs control. D+T+EV also showed a significant improvement in EFS vs control (HR 0.61 [95% CI, 0.47-0.80], P=0.0004; 24-mo rate 60% vs 45%) but did not meet significance for OS; follow-up is ongoing. Pathological complete response per local review was observed in 45% (95% CI, 38.6-51.7) of pts in the D+EV arm and 10% (95% CI, 6.1-14.2) in the control arm. In the D+EV and D+T+EV arms, max. Grade 3/4 tx-related adverse events occurred in 30% and 42% of pts, respectively. ctDNA results will also be presented. Conclusions Dual primary endpoints of EFS were met. Perioperative D plus 3 cycles of neoadjuvant EV also demonstrated a significant improvement in OS vs control. Safety profiles were consistent with those of the individual agents. [Slide 3] Methods VOLGA (NCT04960709), a phase 3, randomised, open-label trial, enrolled pts with MIBC (cT2-T4aN0/1M0 or cT1N1M0) ineligible for or who refused treatment (tx) with cisplatin and were planned for RC. Pts were randomised 1:1:1 to D+EV, D+T+EV, or control. D+EV and D+T+EV arms: perioperative D±T (neoadjuvant: 3 cycles of D, 2 cycles of T; adjuvant: 9 cycles of D, 1 cycle of T) and 3 cycles of neoadjuvant EV; RC was completed after neoadjuvant tx. Control arm: RC and observation or locally approved adjuvant tx. Dual primary endpoints were event-free survival (EFS): D+EV vs control and D+T+EV vs control. Overall survival (OS) was the key secondary endpoint. Safety was also assessed. Here we present results from a pre-planned interim analysis (data cutoff: 02 Mar 2026).

VOLGA: Pre-RC ctDNA Dynamics and Association With EFS for D+EV 60% of baseline ctDNA+ patients achieved ctDNA clearance in this exploratory analysis, as assessed using Natera's ultrasensitive Signatera(TM) Genome with Phased and Structural Variants Assay ctDNA status in matched evaluable samples at baseline and pre-RC = 112 patients(a) Baseline → Neoadjuvant → Pre-RC Baseline: 22% ctDNA– (25/112); 78% ctDNA+ (87/112) 60% ctDNA clearance (52/87) Pre-RC: 68% ctDNA– (76/112); 32% ctDNA+ (36/112) EFS for durvalumab+EV [KM curve]: Probability of EFS vs Time to event or censor (months) 0-42 Curve labels at 12 / 24 months: Persistent negative 91% / 91%; Clearance 88% / 81%; Persistent positive 70% / 55% No. of patients at risk: Months | 0 | 6 | 12 | 18 | 24 | 30 | 36 | 42 Persistent negative | 24 | 21 | 20 | 15 | 11 | 4 | 1 | 0 Clearance | 52 | 49 | 45 | 33 | 23 | 14 | 3 | 0 Persistent positive | 35 | 30 | 23 | 13 | 10 | 7 | 2 | 0 (a)Analysis based on currently available ctDNA biomarker evaluable population as of July 2026; additional patient data pending. One patient with ctDNA– to ctDNA+ status is not shown in the Sankey plot and was not included in the EFS Kaplan-Meier analysis. ctDNA, circulating tumour DNA; D, durvalumab; EFS, event-free survival; EV, enfortumab vedotin; RC, radical cystectomy. Survival table Endpoint: EFS (durvalumab+EV arm) by ctDNA dynamics Persistent negative | 24 Clearance | 52 Persistent positive | 35 Landmark rates: 12 mo: Persistent negative 91% | Clearance 88% | Persistent positive 70% 24 mo: Persistent negative 91% | Clearance 81% | Persistent positive 55%

[Slide 1] The NEW ENGLAND JOURNAL of MEDICINE ORIGINAL ARTICLE Durvalumab, Tremelimumab, and Enfortumab Vedotin in Muscle-Invasive Bladder Cancer T. Powles,1 E. Grande,2 G. Simone,3 A. Gómez de Liaño,4 M. Maruzzo,5 P.I. Velho,6 E. Kikuchi,7 T.Q. Vu,8 L. Popovic,9 T. Çil,10 G. Gravis,11 R. Radkowski,12 L.R.N. Hernández,13 H.J. Cutuli,14 A. Bamias,15 T. Bauernhofer,16 M. Pavic,17 Y.J. Choi,18 C.-H. Chang,19 H. Kitamura,20 J.E. Rosenberg,21 J. Rowbottom,22 S. Ho,23 A. Gupta,24 and A. Drakaki,25 for the VOLGA Investigators* ABSTRACT BACKGROUND The standard of care for patients with muscle-invasive bladder cancer continues to evolve. Whether perioperative durvalumab, with or without tremelimumab, plus neoadjuvant enfortumab vedotin is effective in such patients is unknown. METHODS In this phase 3 open-label trial, we randomly assigned, in a 1:1:1 ratio, patients with muscle-invasive bladder cancer who were ineligible for or declined cisplatin-based chemotherapy to receive neoadjuvant durvalumab–enfortumab vedotin every 3 weeks for three cycles, followed by radical cystectomy and then adjuvant durvalumab every 4 weeks for nine cycles; neoadjuvant durvalumab–tremelimumab–enfortumab vedotin (tremelimumab for two cycles only), radical cystectomy, and adjuvant durvalumab–tremelimumab (tremelimumab for one cycle only); or radical cystectomy alone (control). The dual primary end points were event-free survival in each of the two durvalumab groups as compared with that in the control group. The two durvalumab groups were not compared with each other. The key secondary end point was overall survival. RESULTS In total, 233 patients were assigned to the durvalumab–enfortumab vedotin group, 233 to the durvalumab–tremelimumab–enfortumab vedotin group, and 229 to the control group. Estimated event-free survival at 24 months was 63.5% (95% confidence interval [CI], 56.5 to 69.7) with durvalumab–enfortumab vedotin and 45.1% (95% CI, 38.2 to 51.8) with control (hazard ratio for an event or death, 0.51; 95% CI, 0.38 to 0.67; P<0.001); estimated overall survival at 24 months was 78.4% and 69.4%, respectively (hazard ratio for death, 0.58; 95% CI, 0.40 to 0.83; P=0.003). Estimated event-free survival at 24 months was 60.1% (95% CI, 53.2 to 66.2) with durvalumab–tremelimumab–enfortumab vedotin (hazard ratio for an event or death vs. control, 0.61; 95% CI, 0.47 to 0.80; P<0.001). Grade 3 or higher adverse events occurred in 71.1% of patients in the durvalumab–enfortumab vedotin group (treatment-related, 30.7%), in 79.5% of those in the durvalumab–tremelimumab–enfortumab vedotin group (treatment-related, 44.5%), and in 50.2% of those in the control group. CONCLUSIONS Perioperative durvalumab with or without tremelimumab plus three cycles of neoadjuvant enfortumab vedotin resulted in significantly better event-free survival outcomes than surgery alone. Perioperative durvalumab plus enfortumab vedotin led to significantly better overall survival outcomes than surgery alone. Grade 3 or higher adverse events of any cause were noted in more than half the patients in the durvalumab groups. (Funded by AstraZeneca; VOLGA ClinicalTrials.gov number, NCT04960709; EudraCT number, 2020-005452-38.) The authors' full names, academic degrees, and affiliations are listed at the end of the article. Thomas Powles can be contacted at [email omitted] or at Barts Cancer Institute, Queen Mary University of London, Barts Health NHS Trust Biomedical Research Centre, Old Anatomy Bldg., Charterhouse Sq., London EC1M 6BQ, United Kingdom. *A complete list of the VOLGA Investigators is provided in the Supplementary Appendix, available at NEJM.org. This article was published on October 24, 2026, at NEJM.org. DOI: 10.1056/NEJMoa2608323 Copyright © 2026 Massachusetts Medical Society. N ENGL J MED NEJM.ORG 1 [Slide 2] Percentage of Patients 73.1 | 63.5 | Durvalumab–EV 50.3 | 45.1 | Control Median Event-free Survival (95% CI), mo Durvalumab–EV | NE (33.2–NC) Control | 12.1 (7.6–27.0) Hazard ratio for an event or death, 0.51 (95% CI, 0.38–0.67) P<0.001 Months since Randomization: 0 | 3 | 6 | 9 | 12 | 15 | 18 | 21 | 24 | 27 | 30 | 33 | 36 | 39 | 42 | 45 No. at Risk Durvalumab–EV | 233 | 197 | 180 | 170 | 158 | 136 | 111 | 93 | 80 | 51 | 38 | 21 | 11 | 4 | 1 | 0 Control | 229 | 157 | 131 | 106 | 98 | 80 | 66 | 59 | 43 | 35 | 22 | 16 | 8 | 5 | 0 | 0 Figure 1. Event-free Survival for the Durvalumab–Enfortumab Vedotin (EV) and Control Groups (Intention-to-Treat Population). Shown are Kaplan–Meier estimates of event-free survival (as assessed by blinded independent central review or by central pathological review if a biopsy was required for a suspected new lesion) as of the data-cutoff date (March 2, 2026). Event-free survival was defined as the time from randomization to the first recurrence of disease after radical cystectomy, first documented disease progression in patients who did not undergo radical cystectomy, the expected date of surgery (for patients who had residual disease and did not undergo radical cystectomy), or death from any cause, whichever occurred first. Patients in the durvalumab–EV group received perioperative durvalumab and neoadjuvant enfortumab vedotin with radical cystectomy, and patients in the control group underwent radical cystectomy without neoadjuvant treatment and then entered an observation-only period with no protocol-indicated adjuvant trial treatment. The two-sided P value was based on a stratified log-rank test (adjusted for tumor stage and cisplatin status as recorded with the use of interactive response technology at randomization). A stratified Cox proportional-hazards model was used to estimate the hazard ratio and confidence interval. A hazard ratio of less than 1 favored durvalumab–EV. Tick marks indicate censored data. Dashed lines indicate event-free survival at 12 and 24 months. NC denotes not calculated, and NE could not be estimated. Survival table Endpoint: EFS (ITT), durvalumab-EV vs control Arm | N | Median | Median CI Durvalumab–EV | 233 | NE | 33.2–NC Control | 229 | 12.1 | 7.6–27.0 HR: 0.51 HR CI: 0.38–0.67 P: P<0.001 Landmark rates: 12 months: Durvalumab–EV 73.1 | Control 50.3 24 months: Durvalumab–EV 63.5 | Control 45.1 [Slide 3] Percentage of Patients 66.8 | 60.1 | Durvalumab–tremelimumab–EV 50.3 | 45.1 | Control Median Event-free Survival (95% CI), mo Durvalumab–Tremelimumab–EV | NE (27.9–NC) Control | 12.1 (7.6–27.0) Hazard ratio for an event or death, 0.61 (95% CI, 0.47–0.80) P<0.001 Months since Randomization: 0 | 3 | 6 | 9 | 12 | 15 | 18 | 21 | 24 | 27 | 30 | 33 | 36 | 39 | 42 | 45 No. at Risk Durvalumab–tremelimumab–EV | 233 | 185 | 167 | 155 | 145 | 132 | 112 | 97 | 76 | 47 | 41 | 17 | 11 | 4 | 4 | 0 Control | 229 | 157 | 131 | 106 | 98 | 80 | 66 | 59 | 43 | 35 | 22 | 16 | 8 | 5 | 0 | 0 Figure 2. Event-free Survival for the Durvalumab–Tremelimumab–EV and Control Groups (Intention-to-Treat Population). Shown are Kaplan–Meier estimates of event-free survival (as assessed by blinded independent central review or by central pathological review if a biopsy was required for a suspected new lesion) as of the data-cutoff date (March 2, 2026). The two-sided P value was based on a stratified log-rank test (adjusted for tumor stage and cisplatin status as recorded with the use of interactive response technology at randomization). A stratified Cox proportional-hazards model was used to estimate the hazard ratio and confidence interval. A hazard ratio of less than 1 favored durvalumab–tremelimumab–EV. Tick marks indicate censored data. Dashed lines indicate event-free survival at 12 and 24 months. Survival table Endpoint: EFS (ITT), durvalumab-tremelimumab-EV vs control Arm | N | Median | Median CI Durvalumab–tremelimumab–EV | 233 | NE | 27.9–NC Control | 229 | 12.1 | 7.6–27.0 HR: 0.61 HR CI: 0.47–0.80 P: P<0.001 Landmark rates: 12 months: Durvalumab–tremelimumab–EV 66.8 | Control 50.3 24 months: Durvalumab–tremelimumab–EV 60.1 | Control 45.1 [Slide 4] Percentage of Patients 87.0 | 78.4 | Durvalumab–EV 80.9 | 69.4 | Control Median Overall Survival (95% CI), mo Durvalumab–EV | NE (NC–NC) Control | NE (37.7–NC) Hazard ratio for death, 0.58 (95% CI, 0.40–0.83) P=0.003 Months since Randomization: 0 | 3 | 6 | 9 | 12 | 15 | 18 | 21 | 24 | 27 | 30 | 33 | 36 | 39 | 42 | 45 | 48 No. at Risk Durvalumab–EV | 233 | 224 | 212 | 207 | 198 | 185 | 160 | 128 | 110 | 82 | 66 | 45 | 27 | 15 | 4 | 0 | 0 Control | 229 | 208 | 190 | 185 | 176 | 157 | 128 | 110 | 96 | 72 | 53 | 34 | 22 | 13 | 5 | 2 | 0 Figure 3. Overall Survival for the Durvalumab–EV and Control Groups (Intention-to-Treat Population). Shown are Kaplan–Meier estimates of overall survival as of the data-cutoff date (March 2, 2026). The two-sided P value was based on a stratified log-rank test (adjusted for tumor stage as recorded with the use of interactive response technology at randomization). A stratified Cox proportional-hazards model was used to estimate the hazard ratio and confidence interval. A hazard ratio of less than 1 favored durvalumab–EV. Tick marks indicate censored data. Dashed lines indicate overall survival at 12 and 24 months. Survival table Endpoint: OS (ITT), durvalumab-EV vs control Arm | N | Median | Median CI Durvalumab–EV | 233 | NE | NC–NC Control | 229 | NE | 37.7–NC HR: 0.58 HR CI: 0.40–0.83 P: P=0.003 Landmark rates: 12 months: Durvalumab–EV 87.0 | Control 80.9 24 months: Durvalumab–EV 78.4 | Control 69.4

[Slide 3] VOLGA: OS Subgroup Analyses for Durvalumab+EV vs Control OS benefit was consistent across subgroups for durvalumab+EV vs control Subgroup | Category | Events/patients Durvalumab+EV | Events/patients Control | HR (95% CI) All patients | | 49/233 | 74/229 | 0.58 (0.40–0.83) Age at randomisation | <65 years | 6/49 | 11/40 | NC (NC–NC) | ≥65 years | 43/184 | 63/189 | 0.64 (0.43–0.95) Sex | Male | 38/191 | 53/187 | 0.64 (0.42–0.96) | Female | 11/42 | 21/42 | 0.47 (0.22–0.96) Race | White | 34/153 | 45/158 | 0.71 (0.45–1.10) | Non-White | 11/56 | 23/54 | 0.42 (0.20–0.84) Geographic region | Asia | 10/44 | 19/46 | 0.49 (0.22–1.03) | Europe | 28/139 | 37/129 | 0.61 (0.37–1.00) | Americas | 11/50 | 18/54 | 0.66 (0.30–1.39) Tumour PD-L1 expression(a,b) | High | 27/120 | 38/120 | 0.69 (0.42–1.13) | Low/negative | 22/113 | 36/109 | 0.50 (0.29–0.84) Tumour stage(a) | T2N0M0 | 19/82 | 20/83 | 0.92 (0.49–1.74) | T3-4aN0M0 | 23/113 | 41/109 | 0.47 (0.28–0.78) | T1-4aN1M0 | 7/37 | 12/35 | NC (NC–NC) Cisplatin status(a) | Ineligible | 43/175 | 63/176 | 0.61 (0.41–0.89) | Refusal | 6/58 | 11/53 | NC (NC–NC) Axis: 0.1 | 0.2 | 0.4 | 0.8 | 1.0 | 2.0; Favours durvalumab+EV | Favours control [left edge cropped] HR and CI were not calculated for subgroups with <20 events across both treatment arms. (a)As reported by electronic case report form. (b)Assessed with the clinical trial assay for PD-L1 (SP263) (Roche Diagnostics); [cropped]h PD-L1 expression was defined as any of the following: ≥25% of tumour cells exhibited membrane staining; or ICP >1% and IC+ ≥25%; or ICP=1% and IC+=100%. Data cutoff 02 March 2026. CI, confidence interval; [cropped], enfortumab vedotin; HR, hazard ratio; IC+, immune cells with PD-L1 staining; ICP, tumour-associated immune cells present; NC, not calculated; OS, overall survival; PD-L1, programmed cell death ligand-1. [Slide 4] VOLGA: AEs Known to be Associated With Durvalumab or EV Durvalumab+EV had a safety profile consistent with exposure to perioperative D plus 3 cycles of neoadjuvant EV Immune-mediated AEs per preferred term(a) (any grade | max grade 3 or 4, %) Hypothyroidism | 12 | - Rash | 5 | 2 Pruritus | 4 | 0.4 Diarrhoea | 2 | 0.[?] Adrenal insufficiency | 2 | - Rash maculo-papular | 2 | 0.[?] Pneumonitis | 2 | - Patients with AEs (%) axis 0-50 Treatment-related AEs known to be associated with EV per grouped term(b) (any grade | max grade 3 or 4, %) Skin reactions | 49 | 5 Peripheral neuropathy | 18 | 0.4 | Grade 1: 12%; Grade 2: 5%; Grade 3: 0.4% Hyperglycaemia | 11 | 3 Infusion-related reactions | 3 | - Patients with AEs (%) axis 0-50 Durvalumab+EV arm, any grade (n=228) Durvalumab+EV arm, max. grade 3 or 4 (n=228) (a)Shows immune-mediated AEs in ≥1% of patients in the safety population of the D+EV arm. (b)Grouped terms are based on multiple, similar preferred terms for EV: Rosenberg JE, et al. J Clin Oncol. 2019;37:2592–2600 and The Astellas Group. European Union Risk Management Plan for Enfortumab Vedotin, Version 5.3, May 2026. Available at: https://www.ema.europa.eu/en/documents/rmp/padcev-epar-risk-management-plan_en.pdf. Access date: 23 September 2026. Investigator-assessed causality. Data cutoff 02 March 2026. AE, adverse event; D, durvalumab; EV, enfortumab vedotin.

VOLGA: pCR for Durvalumab+EV vs Control – ITT Durvalumab+EV showed improvements in pCR rate vs control Local pathology review Risk ratio, 4.7 (95% CI, 3.1–7.3) Durvalumab+EV (N=233) | 45% (95% CI, 38.6–51.7) | (105/233) Control (N=229) | 10% (95% CI, 6.1–14.2) | (22/229) Central independent review Risk ratio, 4.8 (95% CI, 3.1–7.9) Durvalumab+EV (N=233) | 40% (95% CI, 34.0–47.0) | (94/233) Control (N=229) | 8% (95% CI, 5.1–12.7) | (19/229) Analysis performed using log-binomial models adjusted using stratification factors: PD-L1 status, tumour stage and cisplatin status per interactive-response technology. Patients who did not undergo surgery or those with a missing or non-evaluable assessment were considered non-responders. Data cutoff 02 March 2026. CI, confidence interval; EV, enfortumab vedotin; ITT, intention-to-treat population; pCR, pathological complete response; PD-L1, programmed cell death ligand-1.

MADRID 2026 ESMO congress | LBA2 Perioperative Durvalumab With or Without Tremelimumab Plus Neoadjuvant Enfortumab Vedotin in Muscle-Invasive Bladder Cancer: Results From the Phase 3 VOLGA Trial Thomas Powles,1 Enrique Grande,2 Giuseppe Simone,3 Alfonso Gómez de Liaño,4 Pedro Isaacsson Velho,5 Eiji Kikuchi,6 Toan Quang Vu,7 Lazar Popovic,8 Timuçin Çil,9 Gwenaelle Gravis,10 Robert Radkowski,11 Luis Raúl Novelo Hernández,12 Yoon Ji Choi,13 Hiroshi Kitamura,14 Jonathan E. Rosenberg,15 Helen K. Angell,16 Jon Armstrong,16 Swati Sharma,17 Suliman Boulos,16 Alexandra Drakaki18 1Barts Cancer Institute, Queen Mary University of London, Barts Health NHS Trust Biomedical Research Centre, London, UK; 2Servicio de Oncología Médica, MD Anderson Cancer Center, Madrid, Spain; 3Uro-Oncology Program, IRCCS Regina Elena National Cancer Institute, Rome, Italy; 4Medical Oncology Department, Complejo Hospitalario Universitario Insular-Materno Infantil de Gran Canaria, Universidad de Las Palmas de Gran Canaria, Las Palmas, Spain; 5Hospital Moinhos de Vento, Porto Alegre, Brazil; 6St. Marianna University School of Medicine, Kawasaki, Japan; 7National Cancer Hospital, Hanoi, Vietnam; 8Oncology Institute of Vojvodina, Faculty of Medicine, University of Novi Sad, Novi Sad, Serbia; 9University of Health Sciences, Adana City Training and Research Hospital, Adana, Turkey; 10Institut Paoli-Calmettes, Aix-Marseille University, INSERM, CNRS, CRCM, Marseille, France; 11Augusta-Kranken-Anstalt gGmbH - Klinik fuer Haematologie, Onkologie & Palliativmedizin, Bochum, Germany; 12Luis Raul Novelo Hernandez, Consultorio de Medicina Especializada del Sector Privado, Colima, Mexico; 13Korea University Anam Hospital, Korea University College of Medicine, Seoul, Republic of Korea; 14Faculty of Medicine, University of Toyama, Toyama, Japan; 15Genitourinary Oncology Service, Memorial Sloan Kettering Cancer Center, New York, NY, USA; 16AstraZeneca, Cambridge, UK; 17AstraZeneca, Cambridge, MA, USA; 18Division of Hematology/Oncology, UCLA Health, Los Angeles, CA, USA

VOLGA: Patient Characteristics – ITT Baseline demographic and disease characteristics were generally balanced between the study arms Characteristics | | Durvalumab+EV N=233 | Durvalumab+T+EV N=233 | Control N=229 Age, median (range), years | | 72 (31–87) | 71 (30–87) | 72 (38–87) Sex | Male | 82% | 79% | 82% Region | Europe | 60% | 52% | 56% | Asia | 19% | 20% | 20% | Americas | 21% | 28% | 24% ECOG PS(a) | 0 | 67% | 62% | 66% | 1 | 24% | 29% | 25% | 2 | 9% | 9% | 8% Nicotine use | Current | 24% | 17% | 14% | Former | 49% | 58% | 57% Tumour PD-L1 expression(b,c) | High | 52% | 52% | 52% | Low/negative | 48% | 48% | 48% Tumour stage(b,d) | T2N0M0 | 35% | 37% | 36% | T3-4aN0M0 | 48% | 48% | 48% | T1-4aN1M0 | 16% | 14% | 15% Cisplatin status(b) | Ineligible | 75% | 76% | 77% | Refusal | 25% | 24% | 23% (a)ECOG PS missing for 1 patient in the control arm. (b)As reported by electronic case report form. (c)Assessed with the clinical trial assay for PD-L1 (SP263) (Roche Diagnostics); high PD-L1 expression was defined as any of the following: ≥25% of tumour cells exhibited membrane staining; or ICP >1% and IC+ ≥25%; or ICP=1% and IC+=100%. (d)Five patients (D+EV, n=1; D+T+EV, n=2; control, n=2) with T1-4aN0-1M1 tumour stage were enrolled and randomised in error and did not receive study treatment. Data cutoff 02 March 2026. D, durvalumab; ECOG PS, Eastern Cooperative Oncology Group performance status; EV, enfortumab vedotin; IC+, immune cells with PD-L1 staining; ICP, tumour-associated immune cells present; ITT, intention-to-treat population; PD-L1, programmed cell death ligand-1; T, tremelimumab

[Slide 1] MADRID 2026 ESMO congress | LBA2 Perioperative Durvalumab With or Without Tremelimumab Plus Neoadjuvant Enfortumab Vedotin in Muscle-Invasive Bladder Cancer: Results From the Phase 3 VOLGA Trial Thomas Powles,1 Enrique Grande,2 Giuseppe Simone,3 Alfonso Gómez de Liaño,4 Pedro Isaacsson Velho,5 Eiji Kikuchi,6 Toan Quang Vu,7 Lazar Popovic,8 Timuçin Çil,9 Gwenaelle Gravis,10 Robert Radkowski,11 Luis Raúl Novelo Hernández,12 Yoon Ji Choi,13 Hiroshi Kitamura,14 Jonathan E. Rosenberg,15 Helen K. Angell,16 Jon Armstrong,16 Swati Sharma,17 Suliman Boulos,16 Alexandra Drakaki18 [Slide 4] VOLGA: Most Frequently Reported AEs for Durvalumab+EV vs Control The safety profile was consistent with those of the individual agents Patients with AEs (%) AE | D+EV any grade (n=228) | D+EV max. grade 3 or 4 (n=228) | Control max. grade 3 or 4 (n=229) | Control any grade (n=229) Pruritus | 44 | 0.9 | – | 4 Urinary tract infection | 36 | 14 | 12 | 24 Anaemia | 32 | 8 | 7 | 21 Diarrhoea | 26 | 2 | 1 | 10 Constipation | 25 | 1 | – | 14 Decreased appetite | 24 | 2 | 0.9 | 6 Rash | 22 | 2 | 0.4 | 3 Nausea | 21 | 0.9 | – | 5 Asthenia | 21 | 0.9 | 0.4 | 3 Alopecia | 21 | – | – | – Dysgeusia | 20 | – | – | 0.4 Fatigue | 19 | 3 | – | 8 Hypothyroidism | 15 | – | – | 0.9 Legend: Durvalumab+EV arm, any grade (n=228); Durvalumab+EV arm, max. grade 3 or 4 (n=228); Control arm, any grade (n=229); Control arm, max. grade 3 or 4 (n=229) Shows all-causality AEs reported for ≥15% of patients in the safety population for either arm. Data cutoff 02 March 2026. AE, adverse event; EV, enfortumab vedotin.

A) OS (%) Months Tarlatamab (n = 254) | Chemotherapy (n = 255) Median OS, months (95% CI)a 13.6 (11.9–15.8) | 8.4 (7.2–10.2) HR (95% CI)b 0.62 (0.51–0.76) Median FU, months 23.3 | 22.8 55.9% 39.0% | 30.3% 14.9% Number of patients at risk: Tarlatamab 254 220 192 171 142 116 105 73 33 14 0 Chemotherapy 255 210 156 124 99 81 60 40 13 4 1 0 B) PFS (%) Median PFS, months (95% CI)a 4.2 (3.0–4.4) | 3.2 (2.9–4.2) HR (95% CI)b 0.69 (0.57–0.84) Median FU, months 22.2 | 21.9 20.6% 5.9% | 14.2% 2.3% Number of patients at risk: Tarlatamab 254 142 74 59 50 43 38 27 15 4 0 Chemotherapy 255 119 48 21 12 11 8 3 2 1 0 C) Responders without progression or death (%) Median DOR (95% CI), monthsa 7.1 (4.5–12.4) | 5.5 (4.2–5.7) 41.6% 10.1% Survival table Endpoint: OS Arm | N | Median | Median CI Tarlatamab | 254 | 13.6 months | 11.9–15.8 Chemotherapy | 255 | 8.4 months | 7.2–10.2 HR: 0.62 HR CI: 0.51–0.76 OS 12-mo: 55.9% vs 39.0% OS 24-mo: 30.3% vs 14.9%

VOLGA: EFS by BICR for Durvalumab+EV vs Control – ITT VOLGA: OS for Durvalumab+EV vs Control – ITT Key differences with EVP: • No adjuvant EV in VOLGA • Anti-PD-L1 vs anti-PD1 Potential differences with EVP: • Protocol therapy adherence • Event definition • Study population

[Slide 1] [Composite of 4 slides] VOLGA: Study Design Patients • Adults aged ≥18 years • MIBC (cT2-T4aN0/1M0 or cT1N1M0) by local assessment • Ineligible for or refused cisplatin • UC or UC with divergent differentiation or histological subtypes • Evaluated and confirmed for cystectomy R (1:1:1) N=695 Neoadjuvant | Adjuvant D+EV N=233: Durvalumab Q3W × 3 cycles; Enfortumab vedotin Q3W × 3 cycles | RC | Durvalumab Q4W × 9 cycles D+T+EV N=233: Durvalumab Q3W × 3 cycles; Tremelimumab Q3W × 2 cycles; Enfortumab vedotin Q3W × 3 cycles | RC | Durvalumab Q4W × 9 cycles; Tremelimumab Q4W × 1 cycle Control N=229: No treatment; Straight to RC | Observation/locally approved treatment Endpoints Dual primary – EFS • D+EV vs control • D+T+EV vs control Key secondary – OS • D+EV vs control • D+T+EV vs control Other secondary • pCR • EORTC QLQ-C30 Exploratory • ctDNA analysis Safety Stratification • PD-L1 status (high vs low/negative) • Clinical tumour stage (T2N0M0 vs T3-4aN0M0 vs T1-4aN1M0) • Cisplatin status (ineligible vs refusal) EFS was defined as time from randomisation to the first: • Recurrence after RC • Progression in patients who did not receive RC • Failure to undergo RC in patients with residual disease • Death due to any cause VOLGA: pCR for Durvalumab+EV vs Control – ITT Durvalumab+EV showed improvements in pCR rate vs control Local pathology review: Risk ratio, 4.7 (95% CI, 3.1–7.3) Durvalumab+EV (N=233): 45% (95% CI, 38.6–51.7) (105/233) Control (N=229): 10% (95% CI, 6.1–14.2) (22/229) Central independent review: Risk ratio, 4.8 (95% CI, 3.1–7.9) Durvalumab+EV (N=233): 40% (95% CI, 34.0–47.0) (94/233) Control (N=229): 8% (95% CI, 5.1–12.7) (19/229) Data cutoff 02 March 2026 VOLGA: EFS by BICR for Durvalumab+EV vs Control – ITT Durvalumab+EV demonstrated a statistically significant improvement in EFS vs control | D+EV N=233 | Control N=229 Events, n (%) | 84 (36) | 123 (54) Median EFS, months (95% CI) | NR (33.2–NC) | 12.1 (7.6–27.0) HR (95% CI) | 0.51 (0.38–0.67) Stratified log-rank P value | <0.0001 Landmarks shown: D+EV 73% (12 mo), 64% (24 mo); Control 50% (12 mo), 45% (24 mo) Median follow-up: 23 months No. of patients at risk D+EV 233 197 180 170 158 136 111 93 80 51 [illegible] 21 11 4 1 0 Control 229 157 131 106 98 80 66 59 43 35 22 16 8 5 0 0 VOLGA: OS for Durvalumab+EV vs Control – ITT Durvalumab+EV demonstrated a statistically significant improvement in OS vs control | D+EV N=233 | Control N=229 Deaths, n (%) | 49 (21) | 74 (32) Median OS, months (95% CI) | NR (NC–NC) | NR (37.7–NC) HR (95% CI) | 0.58 (0.40–0.83) Stratified log-rank P value | 0.0027 Landmarks shown: D+EV 87% (12 mo), 78% (24 mo); Control 81% (12 mo), 69% (24 mo) Median follow-up: 26 months No. of patients at risk: [illegible at this resolution] Survival table Endpoint: EFS by BICR (ITT), durvalumab+EV vs control Arm | N | Median | Median CI Durvalumab+EV | 233 | NR | 33.2–NC Control | 229 | 12.1 mo | 7.6–27.0 HR: 0.51 HR CI: 0.38–0.67 P: <0.0001 (stratified log-rank) Landmark rates: 12 mo: Durvalumab+EV 73% | Control 50% 24 mo: Durvalumab+EV 64% | Control 45% [Slide 2] Cystectomy adherence, event definition and baseline characteristics Neoadjuvant 233 assigned to durvalumab+EV (ITT) 5 did not receive treatment 228 received treatment (safety population) 192 completed neoadjuvant treatment 36 discontinued neoadjuvant treatment 229 assigned to control (ITT and safety population) Surgery 198 underwent cystectomy 198 underwent RC 35 did not undergo cystectomy(a) 87% 180 underwent cystectomy 179 underwent RC 1 had partial cystectomy 49 did not undergo cystectomy(a) 79% Adjuvant 157 received treatment(b) 128 completed adjuvant treatment 28 discontinued adjuvant treatment [Inset EFS KM curve: D+EV N=233 vs Control N=229; Events, n (%) 84 (36) | 123 (54); Median EFS, months (95% CI) NR (33.2–NC) | 12.1 (7.6–27.0); HR (95% CI) 0.51 (0.38–0.67); Stratified log-rank P value <0.0001; 73%/64% vs 50%/45%; Median follow-up: 23 months; early-curve region circled] The most common reason for not undergoing cystectomy was patient decision (D+EV, n=15; control, n=30) Event if: "Failure to undergo RC in patients with residual disease" KN905/EV303: cystectomy in 88% (EVP) and 90% (control) Michiel S. van der Heijden [Slide 4] Anti-CTLA4 in UC: wrong target or wrong design? [EFS KM inset] | D+T+EV N=233 | Control N=229 Events, n (%) | 92 (39) | 123 (54) Median EFS, months (95% CI) | NR (27.9–NC) | 12.1 (7.6–27.0) HR (95% CI) | 0.61 (0.47–0.80) Stratified log-rank P value | 0.0004 Landmarks: Durvalumab+T+EV arm 67% (12 mo), 60% (24 mo); Control arm 50% (12 mo), 45% (24 mo) Median follow-up: 23 months [OS KM inset] | D+T+EV N=233 | Control N=229 Deaths, n (%) | 60 (26) | 74 (32) Median OS, months (95% CI) | NR (NC–NC) | NR (37.7–NC) HR (95% CI) | 0.77 (0.54–1.07) Stratified log-rank P value | 0.1230 Landmarks: Durvalumab+T+EV arm 82% (12 mo), 75% (24 mo); Control arm 81% (12 mo), 69% (24 mo) Median follow-up: 26 months Events, n (%) | Durvalumab+EV n=228 | Durvalumab+T+EV n=229 | Control n=229 AEs possibly related to any treatment | 207 (91) | 216 (94) | NA Maximum grade 3 or 4 AEs | 69 (30) | 97 (42) | NA Serious AEs | 45 (20) | 70 (31) | NA Leading to death | 1 (0.4) | 5 (2) | NA Any-grade immune-mediated AEs | 75 (33) | 109 (48) | 8 (3) Maximum grade 3 or 4 | 24 (11) | 46 (20) | 2 (0.9) Leading to death | 0 | 3 (1) | 0 Neoadjuvant completed (safety population) | 84% | 69% | - Surgery (safety population) | 87% | 80% | 79% Adjuvant started (safety population) | 69% | 60% | - D+T+EV vs D+EV: • G3-4 irAE ↑ • Neoadjuvant therapy completed ↓ • Surgery ↓ OS signal still somewhat immature Unlikely to justify tremi in the future Michiel S. van der Heijden

[Slide 1] Personalising perioperative treatment today? YES Hystology/variant differentiation May influence treatment choice Renal Function (CrCl) May influence treatment choice PD-L1 ≥1% Adjuvant Nivolumab after NAC or cisplatin-ineligible/declining adjuvant cisplatin NO T stage Prognostic but not validated for treatment –selection Nectin-4 Target not validate predictive biomarker PD-L1 Not consistent in peri-operative strategy NOT YET HER-2 Promising but not validated [Slide 2] De-escalation strategies in MIBC today? YES ctDNA-guided adjuvant treatment Evidence supports treatment selection (IMvigor011) Omit adjuvant EV Within VOLGA regimen NOT YET ctDNA-guided bladder preservation: Not yet validated utDNA -guided bladder preservation Not yet validated cCR and local tretament Promising, but insufficient evidence NO Shorter neoadjuvant therapy Insufficient evidence to support fewer cycles

VOLGA: Conclusions • Perioperative durvalumab plus 3 cycles of neoadjuvant enfortumab vedotin demonstrated statistically significant improvements in EFS and OS vs control in patients with MIBC ineligible for or who refused cisplatin-based chemotherapy – EFS HR 0.51 (95% CI, 0.38–0.67), P<0.0001 – OS HR 0.58 (95% CI, 0.40–0.83), P=0.0027 – EFS and OS benefits were consistent across subgroups • Durvalumab plus enfortumab vedotin had a tolerable and manageable safety profile that was consistent with the known safety profiles of the individual agents, with quality of life maintained after surgery • Durvalumab and tremelimumab plus enfortumab vedotin also demonstrated a statistically significant improvement in EFS vs control – OS analysis was not statistically significant at this interim analysis; follow-up is ongoing VOLGA supports perioperative durvalumab plus 3 cycles of neoadjuvant enfortumab vedotin as a potential new standard of care for patients with MIBC ineligible for or who refused cisplatin-based chemotherapy Scan the QR code to access: • Slide presentation • Plain language summary Copies of this presentation obtained through the QR code are for personal use only and may not be reproduced without written permission of the authors. Data cutoff 02 March 2026. CI, confidence interval; EFS, event-free survival; HR, hazard ratio; MIBC, muscle-invasive bladder cancer; OS, overall survival.

VOLGA: Quality of Life as Assessed by EORTC QLQ-C30 Patient-reported quality of life was maintained after surgery in the durvalumab+EV arm GHS/QoL Difference vs control: 3.53 (95% CI, −0.28 to 7.33) Adjusted mean change of score from baseline (95% CI) Timepoint: Adjuvant Cycle 1 | Adjuvant Cycle 2 | Adjuvant Cycle 3 | Adjuvant Cycle 4 | Adjuvant Cycle 5 | Adjuvant Cycle 6 | Adjuvant Cycle 7 | Adjuvant Cycle 8 | Adjuvant Cycle 9 | Follow-up Day 30 No. of patients in analysis Durvalumab+EV | 98 | 93 | 92 | 97 | 92 | 85 | 86 | 81 | 84 | 62 Control | 61 | 58 | 57 | 55 | 55 | 44 | 45 | 39 | 47 | 37 Better / Worse Clinically relevant change from baseline Questionnaire compliance in the D+EV and control arms was 84% and 60% at baseline, 71% and 54% at adjuvant cycle 1, 66% and 64% at adjuvant cycle 9, and 34% and 29% at follow-up day 30, respectively. A positive difference favours D+EV vs control. MMRM analysis of change from baseline in EORTC QLQ-C30 scores with fixed, categorical effects of treatment, visit, and treatment-by-visit interaction, PD-L1 status, tumour stage, and cisplatin status and the continuous fixed covariate of baseline score-by-visit interaction. A clinically relevant change was defined as a ≥10-point change in score compared with baseline.1 1. Osoba D, et al. J Clin Oncol. 1998;16:139–144. Data cutoff 02 March 2026. CI, confidence interval; D, durvalumab; EORTC QLQ-C30, European Organisation for Research and Treatment of Cancer 30-item Core Quality of Life Questionnaire; EV, enfortumab vedotin; GHS/QoL, global health status/quality of life; MMRM, mixed model for repeated measures; PD-L1, programmed cell death ligand-1.

[KM curve] Percentage of Patients vs Months since Randomization (0-45) Durvalumab–EV: 73.1 (12 months), 63.5 (24 months) Control: 50.3 (12 months), 45.1 (24 months) ORIGINAL ARTICLE MEETING OF THE EUROPEAN SOCIETY FOR MEDICAL ONCOLOGY Durvalumab Regimens in Invasive Bladder Cancer T. Powles and Others In patients with muscle-invasive bladder cancer, event-free survival at 24 months was 63.5% with perioperative durvalumab plus neoadjuvant enfortumab vedotin as compared with 45.1% with surgery alone. Survival table Endpoint: EFS Durvalumab–EV Control Landmark rates: 12 months: Durvalumab–EV 73.1 | Control 50.3 24 months: Durvalumab–EV 63.5 | Control 45.1

[Slide 1] VOLGA: AEs Known to be Associated With Durvalumab or EV Durvalumab+EV had a safety profile consistent with exposure to perioperative D plus 3 cycles of neoadjuvant EV Immune-mediated AEs per preferred term(a) [Patients with AEs (%)] Hypothyroidism | 12 Rash | 5 (grade 3/4: 2) Pruritus | 4 (grade 3/4: 0.4) Diarrhoea | 2 (grade 3/4: [illegible]) Adrenal insufficiency | 2 Rash maculo-papular | 2 (grade 3/4: [illegible]) Pneumonitis | 2 Treatment-related AEs known to be associated with EV per grouped term(b) [Patients with AEs (%)] Skin reactions | 49 (grade 3/4: 5) Peripheral neuropathy | 18 (grade 3/4: 0.4) Grade 1: 12%; Grade 2: 5%; Grade 3: 0.4% Hyperglycaemia | 11 (grade 3/4: 3) Infusion-related reactions | 3 Durvalumab+EV arm, any grade (n=228) Durvalumab+EV arm, max. grade 3 or 4 (n=228) [footnotes small/partially illegible; appear to cite Rosenberg JE et al. J Clin Oncol 2019;37:2592-2600, EV EU risk management plan, investigator-assessed causality, data cutoff [illegible] March 2026] [Slide 2] VOLGA: Quality of Life as Assessed by EORTC QLQ-C30 Patient-reported quality of life was maintained after surgery in the durvalumab+EV arm GHS/QoL Difference vs control: 3.53 (95% CI, −0.28 to 7.33) Adjusted mean change of score from baseline (95% CI) Timepoint: Adjuvant Cycle 1 | Adjuvant Cycle 2 | Adjuvant Cycle 3 | Adjuvant Cycle 4 | Adjuvant Cycle 5 | Adjuvant Cycle 6 | Adjuvant Cycle 7 | Adjuvant Cycle 8 | Adjuvant Cycle 9 | Follow-up Day 30 Better / Worse Clinically relevant change from baseline No. of patients in analysis Durvalumab+EV | [illegible across timepoints] Control | 61 | 58 | 57 | 55 | 55 | 44 | 45 | 39 | 47 | 37 Questionnaire compliance in the D+EV and control arms was 84% and 60% at baseline, 71% and 54% at adjuvant cycle 1, 66% and 64% at adjuvant cycle 9, and 34% and 29% at follow-up day 30, respectively. [remaining footnote small; partially illegible - MMRM analysis; clinically relevant change defined as a ≥10-point change; data cutoff [illegible] March 2026]

[Slide 1] Median Event-free Survival (95% CI) mo Durvalumab–EV NE (33.2–NC) Control 12.1 (7.6–27.0) Hazard ratio for an event or death, 0.51 (95% CI, 0.38–0.67) P<0.001 Landmark labels: 12 mo Durvalumab–EV 73.1, Control 50.3; 24 mo Durvalumab–EV 63.5, Control 45.1 Months since Randomization No. at Risk (0,3,6,...,45) Durvalumab–EV 233 197 180 170 158 136 111 93 80 51 38 21 11 4 1 0 Control 229 157 131 106 98 80 66 59 43 35 22 16 8 5 0 0 Survival table Endpoint: Event-free survival Arm | N | Median | Median CI Durvalumab–EV | 233 | NE | 33.2–NC Control | 229 | 12.1 mo | 7.6–27.0 HR: 0.51 HR CI: 0.38–0.67 P: <0.001 Landmark rates: 12 mo: Durvalumab–EV 73.1% | Control 50.3% 24 mo: Durvalumab–EV 63.5% | Control 45.1% [Slide 2] Percentage of Patients Months since Randomization Durvalumab–tremelimumab–EV Control 66.8 | 60.1 | 50.3 | 45.1 Median Event-free Survival (95% CI) mo Durvalumab–Tremelimumab–EV NE (27.9–NC) Control 12.1 (7.6–27.0) Hazard ratio for an event or death, 0.61 (95% CI, 0.47–0.80) P<0.001 No. at Risk Durvalumab–tremelimumab–EV 233 185 167 155 145 132 112 97 76 47 41 17 11 4 4 0 Control 229 157 131 106 98 80 66 59 43 35 22 16 8 5 0 0 Survival table Endpoint: Event-free survival Arm | N | Median | Median CI Durvalumab–tremelimumab–EV | 233 | NE | 27.9–NC Control | 229 | 12.1 | 7.6–27.0 HR: 0.61 HR CI: 0.47–0.80 P: <0.001 Landmark rates: 12 mo: Durvalumab–tremelimumab–EV 66.8 | Control 50.3 24 mo: Durvalumab–tremelimumab–EV 60.1 | Control 45.1

What should be our priority in MIBC? PERSONALISING TREATMENT Which treatment for which patient? 67% 31/46 DE-ESCALATING TREATMENT Who can safely receive less treatment? 33% 15/46 Informal poll among 46 Italian medical oncologists, October 2026 Patrizia Giannatempo, MD

[Slide 2] 5% alpha (2-sided) EFS (2.7%) Durvalumab+EV vs control ↔ EFS (2.7%) Durvalumab+T+EV vs control ↓ OS (5%) Durvalumab+EV vs control ↓ OS (5%) Durvalumab+T+EV vs control [Slide 3] VOLGA: Patient Disposition First patient randomised: Apr 2022; Last patient randomised: Jan 2025 1009 patients enrolled → 695 patients randomised 233 assigned to durvalumab+EV (ITT); 5 did not receive treatment | 233 assigned to durvalumab+T+EV (ITT); 4 did not receive treatment | 229 assigned to control (ITT and safety population) Neoadjuvant: 228 received treatment (safety population); 192 completed neoadjuvant treatment; 36 discontinued neoadjuvant treatment | 229 received treatment (safety population); 159 completed neoadjuvant treatment; 70 discontinued neoadjuvant treatment Surgery: 198 underwent cystectomy; 198 underwent RC; 35 did not undergo cystectomy(a) | 183 underwent cystectomy; 183 underwent RC; 50 did not undergo cystectomy(a) | 180 underwent cystectomy; 179 underwent RC; 1 had partial cystectomy; 49 did not undergo cystectomy(a) Adjuvant: 157 received treatment(b); 128 completed adjuvant treatment; 28 discontinued adjuvant treatment | 137 received treatment(b); 98 completed adjuvant treatment; 38 discontinued adjuvant treatment Follow-up: 177 ongoing in study | 164 ongoing in study | 142 ongoing in study (a)The most common reasons for not undergoing cystectomy were patient decision (D+EV, n=15; D+T+EV, n=14; control, n=30), death (D+EV, n=7; D+T+EV, n=12; control, n=1), and disease progression (D+EV, n=4; D+T+EV, n=9; control, n=6). (b)One patient in the D+EV arm and one in the D+T+EV arm were ongoing adjuvant treatment at data cutoff. Data cutoff 02 March 2026.

[Slide 2] DASL-HiCaP Study Design International, randomized, placebo-controlled, phase 3 trial Patients with DASL very high-risk features • Radical radiation therapy: • ISUP Grade Group 5 • ISUP Grade Group 4 with • cT2b-T4 or rT3a-T4 or PSA> 20 ng/mL • cN1 • Salvage radiation therapy with PSA persistence or rise ≥0.1 ng/mL w/in 12 months: • pN1 or cN1 • ISUP Grade Group 5 • ISUP Grade Group 4 with pT3a-pT4 1,106 Participants (R 1:1) Radiation therapy ADT | Prostate / bed + pelvic LNs + 96 weeks of ADT Darolutamide: 600 mg | Twice daily x 96 weeks Radiation therapy ADT | Prostate / bed + pelvic LNs + 96 weeks of ADT Placebo | Twice daily x 96 weeks Primary endpoint • Metastasis-free Survival (MFS-CI) Secondary endpoints • Overall survival • Prostate cancer specific survival • PSA-progression free survival • Time to subsequent hormonal therapy • Time to castration resistance • Frequency and severity of adverse events • Health-related Quality of life Exploratory • Incremental cost-effectiveness • Prognostic/Predictive biomarkers • Fear of cancer recurrence Stratification factors: • Previous radical prostatectomy • Planned docetaxel use • Clinical or pathological Pelvic N1 Study timeline Enrolment period: April 2020 – August 2023 Current analysis — Interim analysis: Pre-specified event-based at 87 MFS events (67%) Final analysis: 130 MFS events (100%) NCT04136353. Full eligibility criteria available in the online article. ADT: androgen deprivation therapy. MFS based on conventional imaging (MFS-CI); International Society of Urological Pathology (ISUP) Tamim Niazi @Niazi_Dr7 #DASL-HiCaP [Slide 3] Primary Endpoint MFS-CI % Event-Free | 100% 75% 50% 25% 0% Months | 0 12 24 36 48 60 | RT + ADT + Darolutamide N=553 | RT + ADT + Placebo N=553 Events, n (%) | 27 (5%) | 63 (11%) MFS-CI 4 years | 95% | 88% Metastases | 9 (33%) | 46 (73%) Death (w/o prior metastases) | 18 (67%) | 17 (27%) HR | HR 0.42 (95% CI: 0.27–0.66) P-value | p=0.0001 • The planned interim analysis was triggered after the 87th MFS-CI event was centrally corroborated. • The observed P-value of 0.0001 met the rejection-boundary of 0.012 for the null hypothesis. Number at risk (number censored) Blue | 553 (0) | 535 (15) | 526 (18) | 460 (78) | 219 (313) | 58 (468) Red | 553 (0) | 536 (11) | 516 (14) | 448 (69) | 213 (288) | 63 (429) Tamim Niazi @Niazi_Dr7 #DASL-HiCaP Survival table Endpoint: Metastasis-free survival by conventional imaging (MFS-CI) RT + ADT + Darolutamide | 553 RT + ADT + Placebo | 553 HR: 0.42 HR CI: 0.27–0.66 P: p=0.0001 Landmark rates: 4 years: RT + ADT + Darolutamide 95% | RT + ADT + Placebo 88% [Slide 4] Interim Analysis of Secondary Endpoints | Events (EF 4y) RT+ADT+Darolutamide N=553 | Events (EF 4y) RT+ADT+Placebo N=553 | HR (95% CI) PSA-PFS | 46 (91%) | 95 (83%) | 0.47 (0.33 - 0.67) CRPC | 6 (99%) | 28 (95%) | 0.22 (0.09 - 0.52) Subsequent Hormonal Therapy | 19 (96%) | 52 (90%) | 0.36 (0.21 - 0.60) PCSS | 4 (99%) | 7 (99%) | 0.58 (0.17 - 1.97) OS | 21 (97%) | 24 (95%) | 0.88 (0.49 - 1.59) EF 4y: Event Free at 4 years; PFS: Progression-Free Survival; PCSS: Prostate Cancer-Specific Survival; OS: Overall Survival; CRPC: Castration-Resistant Prostate Cancer Tamim Niazi @Niazi_Dr7 #DASL-HiCaP

PROTEUS: A Randomized, Double-Blind, Placebo-Controlled Phase 3 Study in HR-LPC N=2109 Patients With HR-LPC RANDOMIZED 1:1 Neoadjuvant (C1 through C6): APA (240 mg QD) + ADT | PBO + ADT 2 wks -> RPa + ADT -> 4 wks Adjuvant (C7 through C12): APA (240 mg QD) + ADT | PBO + ADT Post treatment: • PSMA-PET (3 months after EOT) • PSA every 3 months • Conventional and PSMA-PET imaging at BCF (PSA ≥0.2 ng/mL); every 6 months thereafter Stratification: • cN0 vs N1 • Gleason score (7 vs 8-10) • Region (North America, Europe, ROW) Adjuvant or salvage RT post-RP at investigator's discretion • To complement these data, a substudy is ongoing to further inform on the role of APA + ADT + RP versus RP alone Patients randomized from July 15, 2019, to June 30, 2022. Protocol Amendment 7 for including PSMA-PET into MFS assessment (April 14, 2022). APA, apalutamide; BCF, biochemical failure; C, cycle; EOT, end of treatment; PBO, placebo; PSA, prostate-specific antigen; PSMA-PET, prostate-specific membrane antigen positron emission tomography; QD, daily; ROW, rest of the world; RT, radiation therapy. aAPA/PBO is stopped 2 weeks prior to planned RP and then resumed 4 weeks after RP. Cardiovascular risk prophylaxis and venous thromboembolism prophylaxis given based on risk.

Darolutamide N=553 events (event-free at 4y) | Placebo N=553 events (event-free at 4y) | Hazard ratio (95% CI) MFS | 27 (95%) | 63 (88%) | 0.42 (0.27 to 0.66) PSA-PFS | 46 (91%) | 95 (83%) | 0.47 (0.33 to 0.67) PCSS | 4 (99%) | 7 (99%) | 0.58 (0.17 to 1.97) OS | 21 (97%) | 24 (95%) | 0.88 (0.49 to 1.59)

DASL-HiCaP Conclusions: • The addition of darolutamide to 2 years of ADT and primary or Salvage RT including pelvic nodal radiation significantly improved MFS-CI for patients with DASL very-high risk prostate cancer. • MFS-CI is a proven surrogate (early intermediate clinical endpoint) for OS in this setting. • This benefit was consistent across all subgroups. • PSA-PFS, time to subsequent therapy, and time to CRPC, all favoured darolutamide. • The treatment burden was similar to placebo with limited additional burden. • Darolutamide reduces the risk of relapse to a clinically meaningful degree, as such it defines a new paradigm for the care of patients with prostate cancer at very high risk of recurrence and being treated with curative intent radiotherapy. Tamim Niazi @Niazi_Dr7 #DASL-HiCaP

[Slide 1] | STAMPEDE* „very high“ 11/2011-03/2016 Positive for MFS and OS | ENZARAD “high“ 03/2014-06/2018 Negative | DASL-HiCaP „very high“ 04/2020-08/2023 Positive for MFS N1 (conv imaging) | 38% | 11% | 24% Gleason <8 and N0 | 11% | 9% | 0% Gleason 9/10 | 54%* | 54% | 73% cT3-cT4 | 93% | 47% | 52% Median PSA ng/ml | 34* | 14 | 11 RT dose (prostate) | 74 Gy | 78 Gy (or 46 Gy + BT boost) | 78 Gy (or 46 Gy + BT boost) RT to prostate | 85% | 100% | 97% RT to pelvis | Recommended for N1; optional for N0 | Mandatory for N1; optional for N0 | Mandatory for N1 and N0 [Slide 2] | STAMPEDE „very high“ 2011-2016 Positive for MFS and OS | ENZARAD “high“ 2014-2018 Negative | DASL-HiCaP „very high“ 2020-2023 Positive for MFS Control arm | ADT alone | ADT + 6 mo NSAA | ADT + placebo N1 (conv imaging) | 38% | 11% | 24% Gleason <8 and N0 | 11% | 9% | 0% Gleason 9/10 | 54%* | 54% | 73% cT3-cT4 | 93% | 47% | 52% PSA ≧ 20 ng/ml | NR | 35% | 29% Median PSA ng/ml | 34* | 14 | 11 RT to pelvis | Recommended for N1; optional for N0 | Mandatory for N1; optional for N0 | Mandatory for N1 and N0 * Including patients in relapsed setting (3%) [Slide 3] | STAMPEDE M0 11/2011-03/2016 Positive for MFS and OS | ENZARAD 03/2014-06/2018 Negative | DASL-HiCaP 04/2020-08/2023 Positive for MFS Number of patients* | 774 | 88 | 316 ARPI | Abiraterone ± Enzalutamide | Enzalutamide | Darolutamide ADT duration | 3 yrs | 2 yrs | 2 yrs HR for MFS in N1 | 0.49 (95% CI 0.38, 0.64) | 0.43 (95% CI 0.20, 0.92) | 0.39 (95% CI 0.19, 0.81) RT to pelvis | NR | Mandated, almost 100% | Mandated, almost 100%

[Slide 1] LBA4: DASL-HiCaP About 2 years of systemic therapy Patients with DASL very high-risk features • Radical radiation therapy: • ISUP Grade Group 5 • ISUP Grade Group 4 with • cT2b-T4 or rT3a+ or PSA> 20 ng/mL • cN1 85% n=935 • BCR post RP + PSA persistence/rise w/in 12 months and at least: • pN1 or cN1 • ISUP Grade Group 5 • ISUP Grade Group 4 with pT3a+ 15% n=171 1,106 patients (R 1:1) Radiation therapy ADT | Prostate / bed + pelvic LNs + 96 weeks of ADT Darolutamide: 600 mg | Twice daily x 96 weeks Radiation therapy ADT | Prostate / bed + pelvic LNs + 96 weeks of ADT Placebo | Twice daily x 96 weeks Primary endpoint • Metastasis-free Survival (MFS)-CI Secondary endpoints • Overall survival • Prostate cancer specific survival • PSA-progression free survival • Time to subsequent hormonal therapy • Time to castration resistance • Frequency and severity of adverse events • Health-related Quality of life • Fear of Cancer recurrence Exploratory • Incremental cost-effectiveness • Prognostic/Predictive biomarkers Stratification factors: • Previous radical prostatectomy • Planned docetaxel use • Clinical or pathological Pelvic N1 613 pts (55%) had a PSMA-PET but patients with mi*N1 were considered as N0, patients with mi*M1 were considered M0 * Molecular imaging Adapted from Niazi T et al. ESMO 2026 Two definitions of very high-risk non-metastatic prostate cancer STAMPEDE M0 (Abiraterone ± Enzalutamide) For N0: 2 risk factors needed Localised • cN1 (conventional imaging) or • cN0 with ≥2 of: T3/4, Gleason 8–10, PSA ≥40ng/mL Relapse after local therapy with high-risk features: N1 or PSA ≥4ng/mL, rising + PSA doubling time <6 mo or PSA ≥20ng/mL. If prior ADT: ≤12 mo of total ADT with an interval of ≥12 mo without treatment DASL-HiCaP (Darolutamide) For N0: Gleason 9/10 or Gleason 8 with at least 1 risk factor Localised • cN1 (conventional imaging) or • cN0 with Gleason 9/10 (alone sufficient) • cN0 with Gleason 8 with cT2b-T4 or seminal vesicle invasion/extracapsular extension in MRI or PSA >20ng/mL Relapse after radical prostatectomy (≤1 year prior to randomisation) and planned for salvage RT for persisting or rising PSA (≥0.1ng/mL): pN1/cN1 or Gleason 9/10 or Gleason 8 with ≥ pT3a+ Attard G et al. Lancet 2022; Niazi T et al. ESMO 2026 Primary Endpoint MFS-CI % Event-Free RT + ADT + Darolutamide N=553 | RT + ADT + Placebo N=553 Events, n (%) | 27 (5%) | 63 (11%) MFS-CI 4 years | 95% | 88% Metastases | 9 (33%) | 46 (73%) Death (w/o prior metastases) | 18 (67%) | 17 (27%) HR | HR 0.42 (95% CI: 0.27-0.66) P-value | p=0.0001 • The planned interim analysis was triggered after the 87th MFS-CI event was centrally corroborated. • The observed P-value of 0.0001 met the rejection-boundary of 0.012 for the null hypothesis. Months Number at risk (number censored) Tamim Niazi Conventional imaging MFS at 4 years Interim Analysis of Secondary Endpoints Events (EF 4y) RT+ADT+Darolutamide N=553 | Events (EF 4y) RT+ADT+Placebo N=553 | HR (95% CI) PSA-PFS | 46 (91%) | 95 (83%) | 0.47 (0.33 - 0.67) CRPC | 6 (99%) | 28 (95%) | 0.22 (0.09 - 0.52) Subsequent Hormonal Therapy | 19 (96%) | 52 (90%) | 0.36 (0.21 - 0.60) PCSS | 4 (99%) | 7 (99%) | 0.58 (0.17 - 1.97) OS | 21 (97%) | 24 (95%) | 0.88 (0.49 - 1.59) EF 4y: Event Free at 4 years; PFS: Progression-Free Survival; PCSS: Prostate Cancer-Specific Survival; OS: Overall Survival; CRPC: Castration-Resistant Prostate Cancer Survival table Endpoint: MFS-CI (metastasis-free survival, conventional imaging) RT + ADT + Darolutamide | 553 RT + ADT + Placebo | 553 HR: 0.42 HR CI: 0.27-0.66 P: 0.0001 MFS-CI 4 years: 95% vs 88% [Slide 2] Differences in the two more modern trials (treatment naïve patients) | STAMPEDE „very high" 2011-2016 Positive for MFS and OS | ENZARAD "high" 2014-2018 Negative | DASL-HiCaP „very high" 2020-2023 Positive for MFS Control arm | ADT alone | ADT + 6 mo NSAA | ADT + placebo N1 (conv imaging) | 38% | 11% | 24% Gleason <8 and N0 | 11% | 9% | 0% Gleason 9/10 | 54%* | 54% | 73% cT3-cT4 | 93% | 47% | 52% PSA ≧ 20 ng/ml | NR | 35% | 29% Median PSA ng/ml | 34* | 14 | 11 RT to pelvis | Recommended for N1; optional for N0 | Mandatory for N1; optional for N0 | Mandatory for N1 and N0 * Including patients in relapsed setting (3%) S. Gillessen Attard G et al. Lancet 2022; Nguyen P et al. ESMO 2025; Williams S et al. J Clin Onc 2026, in press; Niazi T et al. ESMO 2026 [Slide 3] Potential Biomarkers are emerging 1. Multimodal AI (MMAI) model Analysis of digitized H&E tissue slides and clinical factors (approved by FDA) 2. 22-gene Genomic Classifier (Decipher) RNA-based genomic test for prostate cancer prognosis (approved by FDA) Benefit seen in MMAI very high-risk MMAI Standard high-risk | MMAI Very high-risk Benefit increases with higher Decipher scores First quartile (GC ≤0.69) | Fourth quartile (GC ≥0.95) Treatment arm ADT alone ADT + abiraterone Years since randomisation Number at risk (% distant metastasis) Tests seem to identify patients that are less likely to develop metastases or die from prostate cancer and seem to have limited benefit from treatment intensification Silke Gillessen Parker CTA et al. Ann Oncol 2026; Courtesy of Sarah Howlett, presented ESMO 2026

DASL-HiCaP Rationale Unmet need: • At study inception in 2019, long course ADT + radiation was the standard of care for patients treated with primary radiation therapy for localized/locally advanced prostate cancer at very high risk of relapse. • STAMPEDE in 2022 and ENZARAD in 2025 reported some patients do benefit from addition of abiraterone or enzalutamide (pelvic RT) to standard long course ADT plus radiation therapy. • In 2026, there is still uncertainty as to which patients having standard of care Radiation and ADT will most likely benefit from systemic treatment intensification. [References 1-3 too small to read reliably: Attard G, ... STAMPEDE ... Lancet 2022; Nguyen PL, et al. Phase III ENZARAD Trial ... Abstract ...; Xie W, et al. ICECaP Working Group, Metastasis-Free Survival is a Strong Surrogate of Overall Survival in Localized Prostate Cancer. J Clin Oncol ...] Tamim Niazi @Niazi_Dr7 #DASL-HiCaP

[Slide 1] MADRID 2026 ESMO congress THE UNIVERSITY OF SYDNEY — Sydney Clinical Trials Centre ANZUP Cancer Trials Group Limited Darolutamide with androgen deprivation therapy (ADT) and definitive or salvage locoregional radiation therapy (RT) for localised, very high-risk prostate cancer: interim analysis of an international, placebo-controlled, randomised phase 3 trial (DASL-HiCaP; ANZUP1801) Tamim Niazi, Hayley Thomas, Alison Y. Zhang, Scott Williams, Ian D. Davis, Sean M. McBride, Hans T. Chung, Annie Ebacher, Eric Khoo, Jarad Martin, Tee Sin Lim, Simon Hughes, David Pryor, Paul J Kelly, Leily Gholamrezaei, Scott C. Morgan, Wendy Parulekar, Grace Balicki, Martin R. Stockler, Christopher Sweeney, Australian and New Zealand Urogenital and Prostate Cancer Trials Group (ANZUP) Tamim Niazi MDCM Department of Oncology, McGill University Division of Radiation Oncology, Jewish General Hospital University Health Centre, Montreal, Qc. Canada Madrid Oct 2026 Tamim Niazi [Slide 2] DASL-HiCaP: Primary Endpoint Metastases Free Survival by conventional imaging (MFS-CI) Follow-Up: 46 months Interim analysis at 67% of the required events for final analysis % Event-Free; Months (0, 12, 24, 36, 48, 60) | RT + ADT + Darolutamide N=553 | RT + ADT + Placebo N=553 Events, n (%) | 27 (5%) | 63 (11%) MFS-CI 4 years | 95% | 88% Metastases | 9 (33%) | 46 (73%) Death (w/o prior metastases) | 18 (67%) | 17 (27%) HR | HR 0.42 (95% CI: 0.27−0.66) P-value | p=0.0001 Number at risk (number censored) 553 (0) | 535 (15) | 526 (18) | 460 (78) | 219 (313) | 58 (468) 553 (0) | 536 (11) | 516 (14) | 448 (69) | 213 (288) | 63 (429) Statistically significant benefit also in important secondary endpoints: PSA-Progression Free Survival Time to Subsequent Hormonal Therapy Time to Castration Resistance OS data immature Only 4/21 (Daro) and 7/24 (Plac) deaths from prostate cancer Survival table Endpoint: MFS-CI (metastases-free survival by conventional imaging) RT + ADT + Darolutamide | 553 RT + ADT + Placebo | 553 HR: 0.42 HR CI: 0.27-0.66 P: 0.0001 Landmark rates: 4 years: RT + ADT + Darolutamide 95% | RT + ADT + Placebo 88% [Slide 3] DASL-HiCaP Strengths • Academic, large, blinded, international trial with > 1000 patients • Addresses an important question in the era of modern radiotherapy • Inclusion of patients with high-risk relapsed disease • PSMA PET scans and sample collection done and will support further research Considerations • Heterogenous patient population including treatment naïve and relapsed patients • Early results from the interim analysis, final analysis awaited; OS data immature • The question which patients with N0 disease benefit from addition of RT to the pelvis cannot be answered, since all patients received it • Introducing a new definition of «very high-risk» Silke Gillessen Invited Discussant LBA4 [Slide 4] Clinical implications • DASL-HiCaP is statistically positive: 96 wks of ADT plus Darolutamide and prostate and pelvic RT improve MFS-CI for patients with very high-risk prostate cancer. Final analysis and OS awaited • Added toxicity seems limited and this should be considered in the risk to benefit ratio. Also life expectancy is relevant in this context • The clearest indication is in patients with N1 disease. This is supported by subset analysis of the trials with Abiraterone and Enzalutamide that also look positive in this subgroup • For N0 disease, the indication is less clear. Gleason seems to be important and suggests biology is relevant: more work on biomarkers is needed • Patients with very high-risk relapsed disease seem to have the same treatment effect, however few such patients were included in the trial Silke Gillessen Invited Discussant LBA4

[Slide 3] Interim Analysis of Secondary Endpoints | Events (EF 4y) RT+ADT+Darolutamide N=553 | Events (EF 4y) RT+ADT+Placebo N=553 | HR (95% CI) PSA-PFS | 46 (91%) | 95 (83%) | 0.47 (0.33 - 0.67) CRPC | 6 (99%) | 28 (95%) | 0.22 (0.09 - 0.52) Subsequent Hormonal Therapy | 19 (96%) | 52 (90%) | 0.36 (0.21 - 0.60) PCSS | 4 (99%) | 7 (99%) | 0.58 (0.17 - 1.97) OS | 21 (97%) | 24 (95%) | 0.88 (0.49 - 1.59) [footnote abbreviation line cut off at bottom - illegible] [Slide 4] Adverse Events of Special Interest Event | Darolutamide n (%) | Control n (%) Vasodilatation and flushing | 364 (66%) | 351 (64%) Fatigue/asthenic conditions | 350 (64%) | 313 (57%) Hypertension | 42 (8%) | 56 (10%) Diabetes mellitus and hyperglycemia | 41 (7%) | 46 (8%) Mental impairment disorders | 46 (8%) | 38 (7%) Rash | 42 (8%) | 32 (6%) Depressed mood disorders | 34 (6%) | 34 (6%) Fall | 26 (5%) | 31 (6%) Breast disorders/gynecomastia | 33 (6%) | 24 (4%) Cardiac disorders | 21* (4%) | 19* (3%) Weight decreased | 23 (4%) | 12 (2%) Bone fractures | 15 (3%) | 17 (3%) Cerebral ischaemia | 4 (1%) | 6 (1%) Interstitial lung disease | 0 (0%) | 4 (1%) Cerebral and intracranial hemorrhage | 3* (1%) | 1 (0%) Seizure | 0 (0%) | 2 (0%) Treatment / Grade legend: Darolutamide - Grade 1-2; Darolutamide - Grade 3-4; Darolutamide - Grade 5; Control - Grade 1-2; Control - Grade 3-4; Control - Grade 5 Percentage of Participants (x-axis 75% - 0% - 75%)

[Slide 2] Key questions for patients with high-risk localised or relapsed prostate cancer treated with radiotherapy • Does the addition of an ARPI* to RT + ADT** improve outcomes? • Who benefits from the addition of RT to the pelvis? • In which patients can we de-escalate treatment? 2 settings: 1. High-risk localised disease (treatment-naïve) 2. Post radical prostatectomy: High-risk salvage RT setting *Androgen receptor pathway inhibitor e.g. Abiraterone, Apalutamide, Darolutamide, Enzalutamide **Androgen deprivation therapy S. Gillessen Silke Gillessen Invited Discussant LBA4 [Slide 4] Two definitions of very high-risk non-metastatic prostate cancer STAMPEDE M0 (Abiraterone ± Enzalutamide) | DASL-HiCaP (Darolutamide) For N0: 2 risk factors needed | For N0: Gleason 9/10 or Gleason 8 with at least 1 risk factor Localised | Localised • cN1 (conventional imaging) or | • cN1 (conventional imaging) or • cN0 with ≥2 of: T3/4, Gleason 8–10, PSA ≥40ng/mL | • cN0 with Gleason 9/10 (alone sufficient) | • cN0 with Gleason 8 with cT2b-T4 or seminal vesicle invasion/extracapsular extension in MRI or PSA >20ng/mL Relapse after local therapy with high-risk features: N1 or PSA ≥4ng/mL, rising + PSA doubling time <6 mo or PSA ≥20ng/mL. | Relapse after radical prostatectomy (≤1 year prior to randomisation) and planned for salvage RT for persisting or rising PSA (≥0.1ng/mL): pN1/cN1 or Gleason 9/10 or Gleason 8 with ≥ pT3a+ If prior ADT: ≤12 mo of total ADT with an interval of ≥12 mo without treatment | Attard G et al. Lancet 2022; Niazi T et al. ESMO 2026 Silke Gillessen Invited Discussant LBA4

From clinical risk factors to biology Today — Clinical selection • cN1 M0 in conventional imaging • Gleason 9/10 (and other DASL-HiCaP criteria) • STAMPEDE: ≥2 of T3–4, GS 8–10, PSA ≥40 Next step — Retrospective validation • Application of MMAI and Decipher to tissue from DASL-HiCaP, ENZARAD and other trials in this setting • Meta-analysis of individual patient data across trials • Analysis of treatment effects in patients with PSMA PET positivity Future — Prospective selection Biomarker-driven trials: • PREDICT-RT (NCT04513717), using Decipher • GUIDANCE (NCT05050084), using Decipher • THUNDER (NCT06282588), using Decipher and PSMA PET These trials incorporate modern biology for treatment selection and De-escalation strategies to minimise toxicity Silke Gillessen Silke Gillessen — Invited Discussant LBA4

16:30 - 18:15 Presidential Symposium I CHAIRS: FABRICE ANDRÉ, JAMES LARKIN Conclusions • This data does help to resolve the difference of the results of the other two trials, which is based around the proportion of patients at risk. They are not contradictory, the ARPI benefit seems to appear when the risk of metastases is high and disappears in lower-risk N0 patients (ENZARAD) • The role of pelvic radiotherapy in very high-risk N0 patients has yet to be resolved • We need to re-define "very high-risk" incorporating modern imaging and biomarkers • In the future we should evaluate not only intensification, but biomarker driven de-escalation to avoid overtreatment in this group of patients, ideally with non-testosterone suppressing strategies Silke Gillessen Invited Discussant LBA4 ALICANTE AUDITORIUM - HALL 6

One question, three trials and two answers Same strategy of 2 years of ARPI intensification: positive in STAMPEDE and DASL-HiCaP, negative in ENZARAD STAMPEDE M0 Abiraterone ± Enza n=1974 2011-2016 | ENZARAD Enzalutamide n=802 2014-2018 | DASL-HiCaP Darolutamide n=1106 2020-2023 HR for MFS-CI in all patients (Prim. Endpoint): 0.53 (95% CI 0.44, 0.64) p<0.0001 | 0.88 (95% CI 0.67, 1.15) p=0.34 | 0.42 (95% CI 0.27, 0.66) p=0.0001 HR for OS in all patients: 0.6 (95% CI 0.48, 0.73) p<0.0001 | 0.87 (95% CI 0.63, 1.2) p=0.4 | immature Median FU: 72 mo | 92 mo | 46 mo (interim analysis) Silke Gillessen Invited Discussant LBA4

Baseline Characteristics for All Participants | RT + ADT + Darolutamide (N=553) | RT + ADT + Placebo (N=553) Median age, years (IQR) | 70 (65,75) | 71 (65,75) Race, n (%) White | 491 (89%) | 484 (88%) Race, n (%) Asian | 16 (3%) | 31 (6%) Race, n (%) Black or African American | 9 (2%) | 8 (1%) Race, n (%) Other | 37 (7%) | 30 (5%) ECOG PS, n (%) 0-1 | 553 (100%) | 553 (100%) Gleason, n(%) 9-10 (GG†:5) | 411 (74%) | 395 (71%) Gleason, n(%) 8 (GG:4) | 124 (22%) | 130 (24%) Gleason, n(%) Other* | 18 (4%) | 28 (5%) Median PSA ng/mL (IQR) | 9 (4,20) | 9 (3,20) Clinical/pathological pelvic nodal involvement: n (%) | 158 (29%) | 158 (29%) Radical prostatectomy, n (%) No | 468 (85%) | 467 (84%) Radical prostatectomy, n (%) Yes | 85 (15%) | 86 (16%) *Gleason 7 (3+4 and 4+3) and Unknown (only one pt in Darolutamide arm); † GG: International Society of Urological Pathology (ISUP) Grade Group (GG) Tamim Niazi @Niazi_Dr7 #DASL-HiCaP

LBA4: DASL-HiCaP Patients with DASL very high-risk features 85% n=935 • Radical radiation therapy: • ISUP Grade Group 5 • ISUP Grade Group 4 with • cT2b-T4 or rT3a+ or PSA> 20 ng/mL • cN1 15% n=171 • BCR post RP + PSA persistence/rise w/in 12 months and at least: • pN1 or cN1 • ISUP Grade Group 5 • ISUP Grade Group 4 with pT3a+ 1,106 patients (R 1:1) About 2 years of systemic therapy Radiation therapy ADT | Prostate / bed + pelvic LNs + 96 weeks of ADT Darolutamide: 600 mg | Twice daily x 96 weeks Radiation therapy ADT | Prostate / bed + pelvic LNs + 96 weeks of ADT Placebo | Twice daily x 96 weeks Primary endpoint • Metastasis-free Survival (MFS)-CI Secondary endpoints • Overall survival • Prostate cancer specific survival • PSA-progression free survival • Time to subsequent hormonal therapy • Time to castration resistance • Frequency and severity of adverse events • Health-related Quality of life • Fear of Cancer recurrence Exploratory • Incremental cost-effectiveness • Prognostic/Predictive biomarkers Stratification factors: • Previous radical prostatectomy • Planned docetaxel use • Clinical or pathological Pelvic N1 613 pts (55%) had a PSMA-PET but patients with mi*N1 were considered as N0, patients with mi*M1 were considered M0 * Molecular imaging Silke Gillessen Adapted from Niazi T et al. ESMO 2026

Primary Endpoint MFS-CI % Event-Free | 100% 75% 50% 25% 0% Months | 0 12 24 36 48 60 | RT + ADT + Darolutamide N=553 | RT + ADT + Placebo N=553 Events, n (%) | 27 (5%) | 63 (11%) MFS-CI 4 years | 95% | 88% Metastases | 9 (33%) | 46 (73%) Death (w/o prior metastases) | 18 (67%) | 17 (27%) HR | HR 0.42 (95% CI: 0.27−0.66) P-value | p=0.0001 Number at risk (number censored) Blue | 553 (0) | 535 (15) | 526 (18) | 460 (78) | 219 (313) | 58 (468) Magenta | 553 (0) | 536 (11) | 516 (14) | 448 (69) | 213 (288) | 63 (429) Tamim Niazi @Niazi_Dr7 #DASL-HiCaP Survival table Endpoint: Metastasis-free survival by conventional imaging (MFS-CI) RT + ADT + Darolutamide | 553 RT + ADT + Placebo | 553 HR: 0.42 HR CI: 0.27–0.66 P: p=0.0001 Landmark rates: 4 years: RT + ADT + Darolutamide 95% | RT + ADT + Placebo 88%

[Slide 2] OS: Use Predicted HR From Regression (H0, HR[OS] ≥ 1; H1, HR(OS) < 1) Top axis (predicted OS HR): 0.59 | 0.63 | 0.67 | 0.71 | 0.75 | 0.79 Legend: n = 700, OS / n = 700, MFS / n = 1,000, OS / n = 1,000, MFS / n = 1,500, OS / n = 1,500, MFS / n = 2,000, OS / n = 2,000, MFS Study Duration (years) vs MFS HR (0.5 - 0.75) USE MFS <- | -> USE OS MFS HR (test STE: H0, HR[MFS] ≥ 0.88; H1, HR[MFS] < 0.88) [Slide 3] Relative attribute importance (%) by Attribute MFS at 5 years | 28.9 Risk of hot flashes | 18.6 Duration of keeping interest in sex while on tx | 17.4 Risk of breast swelling | 11.7 Tx suspension | 8.2 Duration of keeping erectile function while on tx | 7.6 Risk of fatigue | 7.5 [Slide 4] % Event-Free vs Months | RT + ADT + Darolutamide N=553 | RT + ADT + Placebo N=553 Events, n (%) | 27 (5%) | 63 (11%) MFS-CI 4 years | 95% | 88% Metastases | 9 (33%) | 46 (73%) Death (w/o prior metastases) | 18 (67%) | 17 (27%) HR | HR 0.42 (95% CI: 0.27–0.66) P-value | p=0.0001 Number at risk (number censored) (0,12,24,36,48,60 mo) Darolutamide arm: 553 (0) | 535 (15) | 526 (18) | 460 (78) | 219 (313) | 58 (468) Placebo arm: 553 (0) | 536 (11) | 516 (14) | 448 (69) | 213 (288) | 63 (429) Survival table Endpoint: MFS-CI (metastasis-free survival, centrally corroborated) RT + ADT + Darolutamide | 553 RT + ADT + Placebo | 553 HR: 0.42 HR CI: 0.27–0.66 P: p=0.0001 Landmark rates: 4 years: RT + ADT + Darolutamide 95% | RT + ADT + Placebo 88%

[Slide 3] Effects on MFS-CI in Prespecified Subgroups Subgroup | Level | DARO n/N | CTRL n/N | HR (95% CI) | P-Value Interaction | Adj P-Value Interaction Pelvic nodal involvement | Yes | 10/158 | 25/158 | 0.39 (0.19 to 0.81) | | | No | 17/395 | 38/395 | 0.44 (0.25 to 0.78) | 0.80 | 0.80 Previous radical prostatectomy | Yes | 1/85 | 10/86 | 0.1 (0.01 to 0.79) | | | No | 26/468 | 53/467 | 0.48 (0.3 to 0.77) | 0.08 | 0.50 Age | 70+ | 17/298 | 37/316 | 0.5 (0.28 to 0.88) | | | <70 | 10/255 | 26/237 | 0.34 (0.16 to 0.7) | 0.42 | 0.80 Gleason score | 9 or 10 | 18/411 | 50/395 | 0.35 (0.2 to 0.6) | | | Others | 9/142 | 13/158 | 0.71 (0.3 to 1.66) | 0.17 | 0.80 Brachytherapy used | Yes | 3/51 | 5/39 | 0.54 (0.13 to 2.26) | | | No | 24/502 | 58/514 | 0.41 (0.25 to 0.66) | 0.72 | 0.80 PET imaging at baseline | Yes | 14/309 | 30/304 | 0.45 (0.24 to 0.85) | | | No | 13/244 | 33/249 | 0.39 (0.21 to 0.75) | 0.77 | 0.80 Overall | All Patients | 27/553 | 63/553 | 0.42 (0.27 to 0.66) | | <- Darolutamide benefit | Placebo benefit -> [Slide 4] PSA-Progression-Free Survival % Event-Free vs Months | RT+ADT+ Darolutamide N=553 | RT+ADT+ Placebo N= 553 Any PFS, n (%) | 46 (8%) | 95 (17%) PSA progression | 18 (39%) | 62 (65%) Death w/o progression | 16 (35%) | 17 (18%) Non-protocol Tx | 8 (17%) | 6 (6%) Distant met w/o PSA progression | 3 (7%) | 10 (11%) Loco-regional w/o PSA progression | 1 (2%) | 0 (0%) 4yr PSA-PFS | 91% | 83% HR 0.47 (95% CI: 0.33-0.67) Number at risk (number censored) (0,12,24,36,48,60 mo) Darolutamide arm: 553 (0) | 532 (14) | 519 (18) | 452 (76) | 205 (311) | 55 (452) Placebo arm: 553 (0) | 532 (10) | 504 (14) | 427 (69) | 202 (272) | 58 (404) Survival table Endpoint: PSA-progression-free survival RT+ADT+Darolutamide | 553 RT+ADT+Placebo | 553 HR: 0.47 HR CI: 0.33-0.67 Landmark rates: 4 years: RT+ADT+Darolutamide 91% | RT+ADT+Placebo 83%

Primary Endpoint MFS-CI | RT + ADT + Darolutamide N=553 | RT + ADT + Placebo N=553 Events, n (%) | 27 (5%) | 63 (11%) MFS-CI 4 years | 95% | 88% Metastases | 9 (33%) | 46 (73%) Death (w/o prior metastases) | 18 (67%) | 17 (27%) HR | HR 0.42 (95% CI: 0.27–0.66) P-value | p=0.0001 Number at risk (number censored) — Months 0 / 12 / 24 / 36 / 48 / 60 Darolutamide | 553 (0) | 535 (15) | 526 (18) | 460 (78) | 219 (313) | 58 (468) Placebo | 553 (0) | 536 (11) | 516 (14) | 448 (69) | 213 (288) | 63 (429) • The planned interim analysis was triggered after the 87th MFS-CI event was centrally corroborated. • The observed P-value of 0.0001 met the rejection boundary of 0.012 for the null hypothesis. Tamim Niazi @Niazi_Dr7 #DASL-HiCaP THE UNIVERSITY OF SYDNEY NHMRC Clinical Trials Centre; ANZUP Cancer Trials Group Limited Survival table Endpoint: Metastasis-free survival (MFS-CI), primary endpoint RT + ADT + Darolutamide | 553 RT + ADT + Placebo | 553 HR: 0.42 HR CI: 0.27–0.66 P: 0.0001 Landmark rates: 4 years: RT + ADT + Darolutamide 95% | RT + ADT + Placebo 88%

Primary Endpoint MFS-CI: | RT + ADT + Darolutamide N=553 | RT + ADT + Placebo N=553 Events, n (%) | 27 (5%) | 63 (11%) MFS-CI 4 years | 95% | 88% Metastases | 9 (33%) | 46 (73%) Death (w/o prior metastases) | 18 (67%) | 17 (27%) HR | HR 0.42 (95% CI: 0.27-0.66) P-value | p=0.0001 % Event-Free (y-axis 0%-100%); Months (x-axis 0, 12, 24, 36, 48, 60) Number at risk (number censored) Darolutamide: 553 (0) | 535 (15) | 526 (18) | 460 (78) | 219 (313) | 58 (468) Placebo: 553 (0) | 536 (11) | 516 (14) | 448 (69) | 213 (288) | 63 (429) Survival table Endpoint: MFS-CI (metastasis-free survival, primary endpoint) RT + ADT + Darolutamide | 553 RT + ADT + Placebo | 553 HR: 0.42 HR CI: 0.27-0.66 P: 0.0001 Landmark rates: 4 years: RT + ADT + Darolutamide 95% | RT + ADT + Placebo 88%

[Slide 1] Pre- and On-Treatment ctDNA Statusa Baseline — Neoadjuvant — Presurgery — RC + PLND — Postsurgery — Adjuvant EV + pembro: ctDNA+ / ctDNA− Baseline: 52.4% (133/254) Presurgery: 13.0% (27/208) Postsurgery: 4.8% (10/209) Adjuvant: 5.8% (4/69) Cis + gem: ctDNA+ / ctDNA− Baseline: 56.8% (134/236) Presurgery: 19.7% (45/228) Postsurgery: 13.5% (30/223) Adjuvant: 4.2% (5/119) The pie chart at the adjuvant time point is scaled relative to the number of participants in the corresponding subgroup at baseline. aYielded a positive or negative ctDNA result at a given time point, irrespective of other time points. Data cutoff date: October 27, 2025. [Slide 2] Kaplan-Meier Estimates of EFS per BICR by ctDNA Status at Baseline Y-axis: Event-Free Survival, %; X-axis: Months (0–54); dashed vertical line at 24 months Curve labels at dashed line: 86.1%, 82.7%, 71.2%, 54.2% | Events, n (%) | HR (95% CI)a ctDNA+ EV + pembro (n = 133) | 41 (30.8) | 0.54 (0.37-0.80) Cis + gem (n = 134) | 64 (47.8) ctDNA− EV + pembro (n = 121) | 18 (14.9) | 0.68 (0.37-1.26) Cis + gem (n = 102) | 23 (22.5) No. at risk (months 0, 6, 12, 18, 24, 30, 36, 42, 48, 54) EV + pembro: ctDNA+ 133, 119, 101, 95, 83, 51, 30, 20, 8, 0 EV + pembro: ctDNA− 121, 103, 96, 93, 80, 46, 29, 21, 10, 0 Cis + gem: ctDNA+ 134, 111, 83, 73, 62, 36, 19, 14, 5, 0 Cis + gem: ctDNA− 102, 95, 84, 82, 68, 40, 22, 13, 8, 0 aUnstratified Cox proportional hazard model with treatment as a covariate. Data cutoff date: October 27, 2025. Survival table Endpoint: EFS per BICR by baseline ctDNA status EV + pembro, ctDNA+ | 133 Cis + gem, ctDNA+ | 134 EV + pembro, ctDNA− | 121 Cis + gem, ctDNA− | 102 HR: ctDNA+: 0.54; ctDNA−: 0.68 HR CI: ctDNA+: 0.37-0.80; ctDNA−: 0.37-1.26 Landmark rates: 24 months (dashed line): EV + pembro, ctDNA− 86.1% | Cis + gem, ctDNA− 82.7% | EV + pembro, ctDNA+ 71.2% | Cis + gem, ctDNA+ 54.2% [Slide 3] ctDNA Clearance and EFS by ctDNA Status Change From Baseline at Presurgery ctDNA clearance [bar chart, ctDNA Clearance Rate, %] EV + pembro 82.1% (69/84); Cis + gem 66.7% (68/102) EV + pembro: Positive-Positive (n = 15); Positive-Negative (n = 69); Negative-Negative (n = 79) [KM: Event-Free Survival, %; Months 0–54; dashed line at 24 months] labels: 94.6%, 81.1%, 33.3% No. at risk (months 0, 6, 12, 18, 24, 30, 36, 42, 48, 54) Positive-Positive: 15, 12, 8, 6, 5, 3, 2, 1, 1, 0 Positive-Negative: 69, 69, 61, 57, 52, 33, 21, 14, 7, 0 Negative-Negative: 79, 73, 71, 69, 62, 32, 21, 16, 8, 0 Cis + gem: Positive-Positive (n = 34); Positive-Negative (n = 68); Negative-Negative (n = 77) [KM: Event-Free Survival, %; Months 0–54; dashed line at 24 months] labels: 87.8%, 63.2%, 35.3% No. at risk (months 0, 6, 12, 18, 24, 30, 36, 42, 48, 54) Positive-Positive: 34, 26, 16, 14, 12, 10, 5, 4, 2, 0 Positive-Negative: 68, 62, 49, 43, 37, 14, 8, 5, 1, 0 Negative-Negative: 77, 74, 67, 65, 53, 31, 19, 12, 7, 0 ctDNA clearance rate was defined as the percentage of participants who were ctDNA-negative at presurgery among those with detectable ctDNA at baseline and evaluable ctDNA at presurgery. Kaplan-Meier curves for subgroups with <10 EFS events are not shown. Four participants in the EV + pembro arm and 3 participants in the cis + gem arm had a change from ctDNA negative at baseline to ctDNA positive at presurgery. Data cutoff date: October 27, 2025. Survival table Endpoint: EFS by ctDNA status change baseline→presurgery (+ ctDNA clearance rate) EV + pembro Positive-Positive | 15 EV + pembro Positive-Negative | 69 EV + pembro Negative-Negative | 79 Cis + gem Positive-Positive | 34 Cis + gem Positive-Negative | 68 Cis + gem Negative-Negative | 77 Landmark rates: 24 months (dashed line): EV + pembro Negative-Negative 94.6% | EV + pembro Positive-Negative 81.1% | EV + pembro Positive-Positive 33.3% | Cis + gem Negative-Negative 87.8% | Cis + gem Positive-Negative 63.2% | Cis + gem Positive-Positive 35.3% [Slide 4] Kaplan-Meier Estimates of EFS per BICR by ctDNA Status at Postsurgery Y-axis: Event-Free Survival, %; X-axis: Months (0–54); dashed vertical line at 24 months Curve labels at dashed line: 89.9%, 82.2%, 10.0%, 10.0% | Events, n (%) | HR (95% CI)a ctDNA+ EV + pembro (n = 10) | 9 (90.0) | 0.63 (0.30-1.35) Cis + gem (n = 30) | 27 (90.0) ctDNA− EV + pembro (n = 199) | 22 (11.1) | 0.48 (0.29-0.81) Cis + gem (n = 193) | 41 (21.2) No. at risk (months 0, 6, 12, 18, 24, 30, 36, 42, 48, 54) EV + pembro: ctDNA+ 10, 9, 3, 2, 1, 0, 0, 0, 0, 0 EV + pembro: ctDNA− 199, 197, 186, 180, 162, 91, 58, 38, 16, 0 Cis + gem: ctDNA+ 30, 22, 5, 3, 3, 2, 2, 1, 1, 0 Cis + gem: ctDNA− 193, 191, 173, 162, 131, 75, 43, 26, 12, 0 aUnstratified Cox proportional hazards model with treatment as a covariate. Data cutoff date: October 27, 2025. Survival table Endpoint: EFS per BICR by postsurgery ctDNA status EV + pembro, ctDNA+ | 10 Cis + gem, ctDNA+ | 30 EV + pembro, ctDNA− | 199 Cis + gem, ctDNA− | 193 HR: ctDNA+: 0.63; ctDNA−: 0.48 HR CI: ctDNA+: 0.30-1.35; ctDNA−: 0.29-0.81 Landmark rates: 24 months (dashed line): EV + pembro, ctDNA− 89.9% | Cis + gem, ctDNA− 82.2% | EV + pembro, ctDNA+ 10.0% | Cis + gem, ctDNA+ 10.0%

[Slide 1] Health-Related Quality of Life With Neoadjuvant and Adjuvant Enfortumab Vedotin Plus Pembrolizumab in Participants With Muscle-Invasive Bladder Cancer Eligible for Cisplatin in the Phase 3 KEYNOTE-B15 Study Alfonso Gómez De Liaño1; Begona Perez Valderrama2; Marco Maruzzo3; Tudor Ciuleanu4; Thomas Deja5; Consuelo Buttigliero6; Seung Il Jung7; Alison Y. Zhang8; Karim Chamie9; Daniel Sloth Hauberg10; Eleni Theodorou11; Qinlei Huang12; Aljosja Rogiers12; Allison Martin Nguyen12; Matthew D. Galsky13 1Complejo Hospitalario Universitario Insular-Materno Infantil de Gran Canaria, Universidad de Las Palmas de Gran Canaria, Las Palmas, Spain; 2Hospital Universitario Virgen del Rocio, Seville, Spain; 3Istituto Oncologico Veneto IRCCS-Oncologia Medica 1, Padova, Italy; 4Institutul Oncologic Prof. Dr. Ion Chiricuta Cluj-Napoca, Cluj-Napoca, Romania; 5Clinical Research Center Sp. z o. o. Medic-R Sp.k, Poznań, Poland; 6Department of Oncology, University of Turin, San Luigi Gonzaga Hospital, Orbassano, Italy; 7Chonnam National University Hwasun Hospital, Hwasun Gun, Republic of Korea; 8Macquarie University Hospital, Sydney, NSW, Australia; 9UCLA, Los Angeles, CA, USA; 10Pfizer, Ballerup, Denmark; 11Astellas Pharma Inc., Addlestone, United Kingdom; 12Merck & Co., Inc., Rahway, NJ, USA; 13Icahn School of Medicine at Mount Sinai, The Tisch Cancer Center at Mount Sinai, New York, NY, USA [Slide 2] Conclusions • The use of perioperative EV plus pembrolizumab was not associated with a clinically meaningful detriment in general HRQoL or cystectomy-specific HRQoL • No clinically meaningful between-group differences were observed in the prespecified HRQoL domains assessed • The largest deterioration in most HRQoL domains was observed after surgery, followed by partial recovery in both treatment groups • Sexual function showed clinically meaningful and persistent worsening in both treatment groups, consistent with the known impact of radical cystectomy • Together with the efficacy and safety results observed in the primary analysis, these PRO findings further support the overall benefit-risk profile of perioperative EV plus pembrolizumab in participants with cisplatin-eligible MIBC1 1. Galsky MD et al. N Engl J Med. 2026;395:338-348.

[Slide 1] KEYNOTE-B15/EV-304 Study (NCT04700124) Key Eligibility Criteria • Adults with MIBC • Clinical stage T2-T4aN0M0 or T1-T4aN1M0 by central assessment • Urothelial histology ≥50% • Eligible for RC + PLND • Did not meet any Galsky criteria for cisplatin ineligibility¹ • ECOG PS score of 0 or 1 N = 808 R 1:1 N = 405 Neoadjuvant: Enfortumab vedotin + Pembrolizumab ctDNA testing time pointsᵃ (via Natera Signatera™ assay) Baseline (W1D1) n = 490ᵇ Presurgery (W13 ± 7 d) n = 436ᵇ RC + PLND Postsurgery (W6 ± 14 d) n = 432ᵇ Adjuvant (W30-34) n = 188ᵇ Adjuvant: Enfortumab vedotin + Pembrolizumab N = 403 Cisplatin + Gemcitabine → Observationᶜ • Primary end point: Event-free survival (EFS) by blinded independent central review Exploratory: Evaluation of EFS and pCR by pre- and on-treatment ctDNA status ᵃAnalysis was performed in the ctDNA-evaluable population, defined as all randomly assigned participants with available ctDNA status at ≥1 of 4 time points during the study. ᵇBoth treatment arms combined. ᶜAs of February 2023, adjuvant nivolumab was permitted when clinically indicated and regionally available. 1. Galsky MD et al. Lancet Oncol. 2011;12:211-214. Data cutoff date: October 27, 2025. [Slide 2] ctDNA-Evaluable Population at the Adjuvant Time Point ctDNA-evaluable population EV + pembro, n (%)ᵃ: Baseline 254 (62.7) | Presurgery 208 (51.4) | Postsurgery 209 (51.6) | Adjuvant 69 (17.0) Cis + gem, n (%)ᵃ: Baseline 236 (58.6) | Presurgery 228 (56.6) | Postsurgery 223 (55.3) | Adjuvant 119 (29.5) Baseline (W1D1) n = 490ᵇ; Presurgery (W13 ± 7 d) n = 436ᵇ; RC + PLND; Postsurgery (W6 ± 14 d) n = 432ᵇ; Adjuvant (W30-34) n = 188ᵇ Nonevaluable included samples that were not received or collected, had insufficient cell-free DNA for testing, or failed the assay Reasons for reduced sample availability at the adjuvant time point • The reduced sample availability at the adjuvant timepoint is likely multifactorial, reflecting: – Expected patient attrition – An elevated rate of test failures at adjuvant W30-34 associated with insufficient cell-free DNA and lower plasma volumes – Timelines for procedures in the two arms were different • Although these effects were observed across both arms, the greater frequency of assay failures and/or missed collections in the EV + pembro arm may have contributed to its higher attrition rate ᵃPercentages are based on the intention-to-treat population (EV + pembro, n = 405; cis + gem, n = 403). ᵇBoth treatment arms combined. Data cutoff date: October 27, 2025. [Slide 4] Conclusions • In this exploratory ctDNA analysis from KEYNOTE-B15, a ctDNA-positive status was associated with poor prognosis in both treatment arms – ctDNA-negative status was associated with improved EFS (baseline and postsurgery) and pCR (presurgery) compared with ctDNA-positive status in both treatment arms • Perioperative EV + pembro showed favorable EFS outcomes versus neoadjuvant cis + gem regardless of ctDNA status at baseline or postsurgery • Participants with ctDNA clearance at presurgery had more favorable EFS outcomes compared with those without ctDNA clearance • ctDNA clearance at presurgery was higher with EV + pembro versus cis + gem – Participants in the EV + pembro arm with ctDNA clearance had improved EFS versus those in the cis + gem arm with ctDNA clearance • These findings show that ctDNA is strongly prognostic, regardless of therapy in this study • ctDNA clearance was higher with EV + pembro than with cis + gem

Empirical Mean Change From Baseline Over Time in EORTC QLQ-C30 Scores Legend: EV + Pembro | Cis + Gem y-axis: Mean Change From Baseline (95% CI); Improvement / Decline arrows; SURGERY marker Panels: GHS/QoL (y 20 to -20); Physical Functioning (y 20 to -20); Role Functioning (y 20 to -40) Presurgery Week: 0, 2, 4, 5, 7, 8, 10, 11, 13 | Postsurgery Week: 9, 18, 30, 42, 54, 66, 78, 90, 102, 126, 150, 174, 198 No. (EV + Pembro / Cis + Gem), GHS/QoL panel: 383/368, 354/331, 300/289, 314/304, 289/291, 299/319, 285/293, 290/283, 308/316 | 289/295, 257/234, 242/224, 241/213, 232/202, 218/184, 208/183, 195/167, 153/135, 114/90, 74/51, 46/29, 17/12 No., Physical Functioning panel: 383/368, 354/331, 300/289, 314/305, 289/291, 299/319, 285/293, 290/283, 308/316 | 289/295, 257/234, 242/224, 241/213, 232/202, 218/184, 208/183, 195/167, 153/135, 114/90, 74/51, 46/29, 17/12 No., Role Functioning panel: 383/368, 354/331, 300/289, 314/305, 289/291, 299/319, 285/293, 290/283, 308/316 | 289/295, 257/234, 242/224, 241/213, 232/202, 218/184, 208/183, 195/167, 153/135, 114/90, 74/51, 46/29, 17/12 Database cutoff: October 27, 2025.

[Slide 2] Presurgery ctDNA+ Predicts Low pCR, Any Regimen pCR by Presurgery ctDNA Status, KEYNOTE-B15 NIAGARA7 ctDNA−: 78% (315/404) ctDNA+: 22% (89/404) [Similar pCR if ctDNA−] ctDNA−: pCR 51%; Non-pCR 49% (n = 315) ctDNA+: pCR 3%; Non-pCR 97% (n = 89) Combined arms ctDNA−: 83.5% (364/436) ctDNA+: 16.5% (72/436) ctDNA−: pCR 58.0% (211/364); Non-pCR 42.0% (153/364) ctDNA+: pCR 2.8% (2/72); Non-pCR 97.2% (70/72) Cis + gem ctDNA−: 80.3% (183/228) ctDNA+: 19.7% (45/228) [Similar pCR if ctDNA−] ctDNA−: pCR 43.7% (80/183); Non-pCR 56.3% (103/183) ctDNA+: pCR 2.2% (1/45); Non-pCR 97.8% (44/45) EV + pembro ctDNA−: 87.0% (181/208) ctDNA+: 13.0% (27/208) ctDNA−: pCR 72.4% (131/181); Non-pCR 27.6% (50/181) ctDNA+: pCR 3.7% (1/27); Non-pCR 96.3% (26/27) [pCR ↓↓↓ if ctDNA+] Pie charts are scaled according to the relative number of participants in each subgroup. Data cutoff date: October 27, 2025. NIAGARA: pre-RC ctDNA BEP who had RC; percentages as reported (Powles T, et al. J Clin Oncol. 2025;43(16 Suppl):4503). [Slide 3] ctDNA Clearers Fare Better With Perioperative EV + Pembro Higher pCR and Longer EFS vs Cis + Gem Among ctDNA− Patients EFS by change in ctDNA status, baseline → presurgery KM per BICR EV + pembro Positive-Positive (n = 15) Positive-Negative (n = 69) Negative-Negative (n = 79) Curve labels at 24 months: 94.6%; 81.1%; 33.3% Cis + gem Positive-Positive (n = 34) Positive-Negative (n = 68) Negative-Negative (n = 77) Curve labels at 24 months: 87.8%; 63.2%; 35.3% Y-axis: Event-Free Survival, % X-axis: Months No. at risk (months 0 / 6 / 12 / 18 / 24 / 30 / 36 / 42 / 48 / 54): EV + pembro Positive-Positive: 15 / 12 / 8 / 6 / 5 / 3 / 2 / 1 / 1 / 0 EV + pembro Positive-Negative: 69 / 69 / 61 / 57 / 52 / 33 / 21 / 14 / 7 / 0 EV + pembro Negative-Negative: 79 / 73 / 71 / 69 / 62 / 32 / 21 / 16 / 8 / 0 Cis + gem Positive-Positive: 34 / 26 / 16 / 14 / 12 / 10 / 5 / 4 / 2 / 0 Cis + gem Positive-Negative: 68 / 62 / 49 / 43 / 37 / 14 / 8 / 5 / 1 / 0 Cis + gem Negative-Negative: 77 / 74 / 67 / 65 / 53 / 31 / 19 / 12 / 7 / 0 pCR among ctDNA− patients (presurgery ctDNA) Cis + gem: 43.7% (80/183) EV + pembro: 72.4% (131/181) aUnstratified Cox proportional hazards model with treatment as a covariate. Data cutoff date: October 27, 2025. pCR by presurgery ctDNA status (pCR assessed at RC); post hoc χ² 1 df, p < 0.001. Source: Chatzkel J, et al. Presented at: ESMO Congress 2026; 2026 Oct 23-27; Madrid, Spain. Abstract 4734O. EV + pembro arm received adjuvant EV + pembro; cis + gem arm was observed after RC. EFS differences cannot be attributed to the neoadjuvant phase alone. Survival table Endpoint: EFS (KM per BICR) by change in ctDNA status, baseline → presurgery EV + pembro Positive-Positive | 15 EV + pembro Positive-Negative | 69 EV + pembro Negative-Negative | 79 Cis + gem Positive-Positive | 34 Cis + gem Positive-Negative | 68 Cis + gem Negative-Negative | 77 P: p < 0.001 (post hoc χ² 1 df, for pCR by presurgery ctDNA status — not an EFS comparison) Landmark rates: 24 mo: EV + pembro Negative-Negative 94.6% | EV + pembro Positive-Negative 81.1% | EV + pembro Positive-Positive 33.3% | Cis + gem Negative-Negative 87.8% | Cis + gem Positive-Negative 63.2% | Cis + gem Positive-Positive 35.3% [Slide 4] Perioperative EV + pembro with serial ctDNA assessment Proposed ctDNA-informed treatment flow, modelled on KEYNOTE-B15 / EV-304 Muscle-invasive bladder cancer [Cisplatin eligibility — struck through] → ctDNA (Baseline) → Neoadjuvant: Enfortumab vedotin + pembrolizumab → ctDNA (Presurgery) → + → Surgery: RC + PLND → ctDNA (Postsurgery) → + → Adjuvant: Enfortumab vedotin + pembrolizumab Presurgery − → ?Bladder sparing Postsurgery − → ?Surveillance Ongoing bladder-preservation trials (not ctDNA-selected) EV + P: EV-209 cystectomy-eligible; cCR-guided surveillance EV-309 cystectomy-ineligible/declining; vs chemoradiotherapy CRT-based: Indi-Blade ipilimumab + nivolumab → chemoradiotherapy ARCHER (NRG-GU015) 5-fraction SBRT vs 20-fraction CRT + ctDNA-positive − ctDNA-negative Dashed boxes: open questions, not established practice cCR, clinical complete response; CRT, chemoradiotherapy; EV, enfortumab vedotin; MIBC, muscle-invasive bladder cancer; PLND, pelvic lymph node dissection; RC, radical cystectomy; SBRT, stereotactic body radiotherapy.

[Slide 1] ctDNA-Evaluable Population at the Adjuvant Time Point ctDNA-evaluable population EV + pembro, n (%)a: Baseline 254 (62.7) | Presurgery 208 (51.4) | Postsurgery 209 (51.6) | Adjuvant 69 (17.0) Cis + gem, n (%)a: Baseline 236 (58.6) | Presurgery 228 (56.6) | Postsurgery 223 (55.3) | Adjuvant 119 (29.5) Baseline (W1D1) n = 490b | Presurgery (W13 ± 7 d) n = 436b | [RC + PLND] | Postsurgery (W6 ± 14 d) n = 432b | Adjuvant (W30-34) n = 188b Nonevaluable included samples that were not received or collected, had insufficient cell-free DNA for testing, or failed the assay Reasons for reduced sample availability at the adjuvant time point - The reduced sample availability at the adjuvant timepoint is likely multifactorial, reflecting: - Expected patient attrition - An elevated rate of test failures at adjuvant W30-34 associated with insufficient cell-free DNA and lower plasma volumes - Timelines for procedures in the two arms were different - Although these effects were observed across both arms, the greater frequency of assay failures and/or missed collections in the EV + pembro arm may have contributed to its higher attrition rate aPercentages are based on the intention-to-treat population (EV + pembro, n = 405; cis + gem, n = 403). bBoth treatment arms combined. Data cutoff date: October 27, 2025 [Slide 3] Kaplan-Meier Estimates of EFS per BICR by ctDNA Status at Baseline [KM plot: Event-Free Survival, % vs Months 0-54; dashed line at 24 months] 24-month labels: 86.1% | 82.7% | 71.2% | 54.2% Events, n (%) | HR (95% CI)a ctDNA- EV + pembro (n = 121) 18 (14.9) | 0.68 Cis + gem (n = 102) 23 (22.5) | (0.37-1.26) ctDNA+ EV + pembro (n = 133) 41 (30.8) | 0.54 Cis + gem (n = 134) 64 (47.8) | (0.37-0.80) No. at risk (0, 6, 12, 18, 24, 30, 36, 42, 48, 54 months) EV + pembro: ctDNA- 121 103 96 93 80 46 29 21 10 0 Cis + gem: ctDNA- 102 95 84 82 68 40 22 13 8 0 EV + pembro: ctDNA+ 133 119 101 95 83 51 30 20 8 0 Cis + gem: ctDNA+ 134 111 83 73 62 36 19 14 5 0 aUnstratified Cox proportional hazard model with treatment as a covariate. Data cutoff date: October 27, 2025. [small low-resolution image; faint rows read on upscale, confirmed against sharper photo 2113613933680492620_s3] Survival table Endpoint: EFS per BICR by baseline ctDNA status (exploratory) ctDNA- EV + pembro | 121 ctDNA- Cis + gem | 102 ctDNA+ EV + pembro | 133 ctDNA+ Cis + gem | 134 HR: ctDNA-: 0.68; ctDNA+: 0.54 HR CI: ctDNA-: 0.37-1.26; ctDNA+: 0.37-0.80 Landmark rates: 24 months: ctDNA- EV + pembro 86.1% | ctDNA- Cis + gem 82.7% | ctDNA+ EV + pembro 71.2% | ctDNA+ Cis + gem 54.2% [Slide 4] ctDNA Clearance and EFS by ctDNA Status Change From Baseline to Presurgery ctDNA clearance [bar chart, ctDNA Clearance Rate, %] EV + pembro 82.1% (69/84) Cis + gem 66.7% (68/102) EV + pembro [KM plot, Event-Free Survival, % vs Months 0-54; dashed line at 24 months] Negative-Negative (n = 79) — 94.6% Positive-Negative (n = 69) — 81.1% Positive-Positive (n = 15) — 33.3% No. at risk (0-54 months, 6-mo steps) Negative-Negative 79 73 71 69 62 32 21 16 8 0 Positive-Negative 69 69 61 57 52 33 21 14 7 0 Positive-Positive 15 12 8 6 5 3 2 1 1 0 Cis + gem [KM plot] Negative-Negative (n = 77) — 87.8% Positive-Negative (n = 68) — 63.2% Positive-Positive (n = 34) — 35.3% No. at risk Negative-Negative 77 74 67 65 53 31 19 12 7 0 Positive-Negative 68 62 49 43 37 14 8 5 1 0 Positive-Positive 34 26 16 14 12 10 5 4 2 0 ctDNA clearance rate was defined as the percentage of participants who were ctDNA-negative at presurgery among those with detectable ctDNA at baseline and evaluable ctDNA at presurgery. Kaplan-Meier curves for subgroups with <10 EFS events are not shown. Four participants in the EV + pembro arm and 3 participants in the cis + gem arm had a change from ctDNA negative at baseline to ctDNA positive at presurgery. Data cutoff date: October 27, 2025. [small low-resolution image; faint rows read on upscale, confirmed against sharper photo 2113614045274390767_s3] Survival table Endpoint: EFS by ctDNA status change baseline to presurgery (exploratory); ctDNA clearance EV + pembro Negative-Negative | 79 EV + pembro Positive-Negative | 69 EV + pembro Positive-Positive | 15 Cis + gem Negative-Negative | 77 Cis + gem Positive-Negative | 68 Cis + gem Positive-Positive | 34 Landmark rates: 24 months: EV + pembro Negative-Negative 94.6% | EV + pembro Positive-Negative 81.1% | EV + pembro Positive-Positive 33.3% | Cis + gem Negative-Negative 87.8% | Cis + gem Positive-Negative 63.2% | Cis + gem Positive-Positive 35.3%

[Slide 1] Change From Baseline to Postsurgery Week 42 in EORTC QLQ-C30 Scores Legend: EV + Pembro | Cis + Gem Y-axis: LS Mean Change From Baseline (95% CI) GHS/QoL: LS mean difference 0.85 (95% CI, −1.74 to 3.44) EV + Pembro −1.58 (−3.54 to 0.38); Cis + Gem −2.43 (−4.45 to −0.41) Physical Functioning: LS mean difference 0.86 (95% CI, −1.73 to 3.45) EV + Pembro −5.42 (−7.24 to −3.61); Cis + Gem −6.28 (−8.16 to −4.40) Role Functioning: LS mean difference −0.21 (95% CI, −4.31 to 3.88) EV + Pembro −9.56 (−12.50 to −6.63); Cis + Gem −9.35 (−12.40 to −6.31) IMPROVEMENT / DECLINE; dotted line at −10 marked a Baseline n = 383 / n = 368 (each domain) Week 42 n = 250 / n = 227 (each domain) Total PRO N = 403 / N = 394 (each domain) aMinimal important difference, ≥10 point worsening from baseline. 'n' is the number of participants in each treatment group with nonmissing assessments at the specific time point; for change from baseline, 'N' is the number of participants in the analysis population in each treatment group. Database cutoff: October 27, 2025. [Slide 2] Change From Baseline to Postsurgery Week 42 in BCI Scores Legend: EV + Pembro | Cis + Gem Y-axis: LS Mean Change From Baseline (95% CI) BCI Urinary: LS mean difference 2.74 (95% CI, −0.15 to 5.63) EV + Pembro −0.77 (−3.02 to 1.48); Cis + Gem −3.51 (−5.81 to −1.21) Baseline n = 379 / n = 362; Week 42 n = 245 / n = 227; Total PRO N = 402 / N = 394 BCI Bowel: LS mean difference 0.20 (95% CI, −2.16 to 2.57) EV + Pembro −1.86 (−3.63 to −0.10); Cis + Gem −2.07 (−3.88 to −0.25) Baseline n = 379 / n = 362; Week 42 n = 245 / n = 227; Total PRO N = 402 / N = 394 BCI Sexual: LS mean difference −0.31 (95% CI, −3.77 to 3.15) EV + Pembro −15.13 (−17.79 to −12.46); Cis + Gem −14.82 (−17.57 to −12.06) Baseline n = 378 / n = 356; Week 42 n = 241 / n = 220; Total PRO N = 401 / N = 390 IMPROVEMENT / DECLINE; dotted MID lines marked a and b aMinimal important difference for BCI urinary, ≥6 point worsening from baseline. bMinimal important difference for BCI bowel and sexual, ≥7 point worsening from baseline. 'n' is the number of participants in each treatment group with non-missing assessments at the specific time point; for change from baseline, 'N' is the number of participants in the analysis population in each treatment group. Database cutoff: October 27, 2025. [Slide 3] Empirical Mean Change From Baseline Over Time in EORTC QLQ-C30 Scores Legend: EV + Pembro | Cis + Gem Panels: GHS/QoL; Physical Functioning; Role Functioning (Improvement / Decline arrows); shaded SURGERY band Y-axis: Mean Change From Baseline (95% CI) X-axis weeks: 0 2 4 5 7 8 10 11 13 (Presurgery Week) | 9 18 30 42 54 66 78 90 102 126 150 174 198 (Postsurgery Week) No. (EV + Pembro): 383 354 300 314 289 299 285 290 308 | 289 257 242 241 232 218 208 195 153 114 74 46 17 No. (Cis + Gem): 368 331 289 304 291 319 293 283 316 | 295 234 224 213 202 184 183 167 135 90 51 29 12 (Physical Functioning row: Cis + Gem week 5 = 305; Role Functioning rows: EV + Pembro week 4 = 300, Cis + Gem week 4 = 289, week 5 = 305) Database cutoff: October 27, 2025. [Slide 4] Kaplan-Meier Estimates of Time to Definitive Deterioration in EORTC QLQ-C30 and BCI Scores Columns: Events, n/N (%) EV + Pembro | Cis + Gem; TTD, median (95% CI), months EV + Pembro | Cis + Gem; HR (95% CI) EORTC QLQ-C30 GHS/QoL | 79/383 (20.6) | 84/368 (22.8) | NR (NR-NR) | NR (NR-NR) | 0.84 (0.61-1.14) Physical functioning | 92/383 (24.0) | 94/368 (25.5) | NR (NR-NR) | NR (NR-NR) | 0.88 (0.66-1.18) Role functioning | 108/383 (28.2) | 121/368 (32.9) | NR (NR-NR) | NR (37.1-NR) | 0.80 (0.61-1.04) BCI Urinary | 94/379 (24.8) | 91/362 (25.1) | NR (NR-NR) | NR (NR-NR) | 0.93 (0.70-1.25) Bowel | 101/379 (26.6) | 82/362 (22.7) | NR (NR-NR) | NR (43.1-NR) | 1.11 (0.83-1.49) Sexual | 173/378 (45.8) | 139/356 (39.0) | 25.5 (18.9-33.5) | 44.3 (24.7-NR) | 1.17 (0.94-1.47) HR (95% CI) axis 0.0 0.5 1.0 1.5 2.0; Favors EV + Pembro / Favors Cis + Gem Participants with no definitive deterioration were censored at last PRO assessment regardless of PD or recurrence. Participants who did not have postbaseline PRO assessment were censored at baseline. Database cutoff: October 27, 2025.

[Slide 1] Change From Baseline to Postsurgery Week 42 in EORTC QLQ-C30 Scores EV + Pembro | Cis + Gem LS Mean Change From Baseline (95% CI) GHS/QoL: LS mean difference 0.85 (95% CI, −1.74 to 3.44) EV + Pembro −1.58 (−3.54 to 0.38) | Cis + Gem −2.43 (−4.45 to −0.41) Physical Functioning: LS mean difference 0.86 (95% CI, −1.73 to 3.45) EV + Pembro −5.42 (−7.24 to −3.61) | Cis + Gem −6.28 (−8.16 to −4.40) Role Functioning: LS mean difference −0.21 (95% CI, −4.31 to 3.88) EV + Pembro −9.56 (−12.50 to −6.63) | Cis + Gem −9.35 (−12.40 to −6.31) Baseline | n = 383 | n = 368 (each domain) Week 42 | n = 250 | n = 227 Total PRO | N = 403 | N = 394 IMPROVEMENT / DECLINE [Slide 2] Change From Baseline to Postsurgery Week 42 in BCI Scores EV + Pembro | Cis + Gem LS Mean Change From Baseline (95% CI) BCI Urinary: LS mean difference 2.74 (95% CI, −0.15 to 5.63) EV + Pembro −0.77 (−3.02 to 1.48) | Cis + Gem −3.51 (−5.81 to −1.21) Baseline n = 379 | n = 362; Week 42 n = 245 | n = 227; Total PRO N = 402 | N = 394 BCI Bowel: LS mean difference 0.20 (95% CI, −2.16 to 2.57) EV + Pembro −1.86 (−3.63 to −0.10) | Cis + Gem −2.07 (−3.88 to −0.25) Baseline n = 379 | n = 362; Week 42 n = 245 | n = 227; Total PRO N = 402 | N = 394 BCI Sexual: LS mean difference −0.31 (95% CI, −3.77 to 3.15) EV + Pembro −15.13 (−17.79 to −12.46) | Cis + Gem −14.82 (−17.57 to −12.06) Baseline n = 378 | n = 356; Week 42 n = 241 | n = 220; Total PRO N = 401 | N = 390 IMPROVEMENT / DECLINE [Slide 3] Empirical Mean Change From Baseline Over Time in EORTC QLQ-C30 Scores EV + Pembro | Cis + Gem Mean Change From Baseline (95% CI) Panels: GHS/QoL; Physical Functioning; Role Functioning (SURGERY marked between presurgery week 13 and postsurgery week 9) Weeks: 0 2 4 5 7 8 10 11 13 (Presurgery Week) | 9 18 30 42 54 66 78 90 102 126 150 174 198 (Postsurgery Week) No. EV + Pembro: 383 354 300 314 289 299 285 290 308 | 289 257 242 241 232 218 208 195 153 114 74 46 17 No. Cis + Gem: 368 331 289 304 291 319 293 283 316 | 295 234 224 213 202 184 183 167 135 90 51 29 12 (Physical Functioning row: Cis + Gem week 5 = 305 per panel; other panels show 304) Database cutoff: October 27, 2025.

[Slide 1] KEYNOTE-B15/EV-304 Study (NCT04700124) Key Eligibility Criteria • Adults with MIBC • Clinical stage T2-T4aN0M0 or T1-T4aN1M0 by central assessment • Urothelial histology ≥50% • Eligible for RC + PLND • Did not meet any Galsky criteria for cisplatin ineligibility¹ • ECOG PS score of 0 or 1 N = 808, R 1:1 N = 405 → Neoadjuvant: Enfortumab vedotin + Pembrolizumab → RC + PLND → Adjuvant: Enfortumab vedotin + Pembrolizumab N = 403 → Neoadjuvant: Cisplatin + Gemcitabine → RC + PLND → Observation^c ctDNA testing time points^a (via Natera Signatera™ assay) Baseline (W1D1) n = 490^b Presurgery (W13 ± 7 d) n = 436^b Postsurgery (W6 ± 14 d) n = 432^b Adjuvant (W30-34) n = 188^b • Primary end point: Event-free survival (EFS) by blinded independent central review Exploratory: Evaluation of EFS and pCR by pre- and on-treatment ctDNA status [Slide 3] Kaplan-Meier Estimates of EFS per BICR by ctDNA Status at Baseline [KM curves; 24-month landmark labels: 86.1%, 82.7%, 71.2%, 54.2%] | | Events, n (%) | HR (95% CI)^a | ctDNA− EV + pembro (n = 121) | 18 (14.9) | 0.68 (0.37-1.26) Cis + gem (n = 102) | 23 (22.5) | ctDNA+ EV + pembro (n = 133) | 41 (30.8) | 0.54 (0.37-0.80) Cis + gem (n = 134) | 64 (47.8) | No. at risk (Months 0 / 6 / 12 / 18 / 24 / 30 / 36 / 42 / 48 / 54) EV + pembro: ctDNA− 121 / 103 / 96 / 93 / 80 / 46 / 29 / 21 / 10 / 0 Cis + gem: ctDNA− 102 / 95 / 84 / 82 / 68 / 40 / 22 / 13 / 8 / 0 EV + pembro: ctDNA+ 133 / 119 / 101 / 95 / 83 / 51 / 30 / 20 / 8 / 0 Cis + gem: ctDNA+ 134 / 111 / 83 / 73 / 62 / 36 / 19 / 14 / 5 / 0 ^aUnstratified Cox proportional hazard model with treatment as a covariate. Data cutoff date: October 27, 2025. Survival table Endpoint: EFS per BICR by baseline ctDNA status EV + pembro: ctDNA− | 121 Cis + gem: ctDNA− | 102 EV + pembro: ctDNA+ | 133 Cis + gem: ctDNA+ | 134 Landmark rates: 24 months: EV + pembro: ctDNA− 86.1% | Cis + gem: ctDNA− 82.7% | EV + pembro: ctDNA+ 71.2% | Cis + gem: ctDNA+ 54.2% [Slide 4] Kaplan-Meier Estimates of EFS per BICR by ctDNA Status at Postsurgery [KM curves; 24-month landmark labels: 89.9%, 82.2%, 10.0%, 10.0%] | | Events, n (%) | HR (95% CI)^a | ctDNA− EV + pembro (n = 199) | 22 (11.1) | 0.48 (0.29-0.81) Cis + gem (n = 193) | 41 (21.2) | ctDNA+ EV + pembro (n = 10) | 9 (90.0) | 0.63 (0.30-1.35) Cis + gem (n = 30) | 27 (90.0) | No. at risk (Months 0 / 6 / 12 / 18 / 24 / 30 / 36 / 42 / 48 / 54) EV + pembro: ctDNA− 199 / 197 / 186 / 180 / 162 / 91 / 58 / 38 / 16 / 0 Cis + gem: ctDNA− 193 / 191 / 173 / 162 / 131 / 75 / 43 / 26 / 12 / 0 EV + pembro: ctDNA+ 10 / 9 / 3 / 2 / 1 / 0 / 0 / 0 / 0 / 0 Cis + gem: ctDNA+ 30 / 22 / 5 / 3 / 3 / 2 / 2 / 1 / 1 / 0 ^aUnstratified Cox proportional hazards model with treatment as a covariate. Data cutoff date: October 27, 2025. Survival table Endpoint: EFS per BICR by postsurgery ctDNA status EV + pembro: ctDNA− | 199 Cis + gem: ctDNA− | 193 EV + pembro: ctDNA+ | 10 Cis + gem: ctDNA+ | 30 Landmark rates: 24 months: EV + pembro: ctDNA− 89.9% | Cis + gem: ctDNA− 82.2% | EV + pembro: ctDNA+ 10.0% | Cis + gem: ctDNA+ 10.0%

Kaplan-Meier Estimates of EFS per BICR by ctDNA Status at Baseline [KM curves; y-axis: Event-Free Survival, %; x-axis: Months 0-54; dashed vertical line at 24 months] Curve labels at dashed line: 86.1% (EV + pembro ctDNA−), 82.7% (Cis + gem ctDNA−), 71.2% (EV + pembro ctDNA+), 54.2% (Cis + gem ctDNA+) | Events, n (%) | HR (95% CI)ᵃ ctDNA− EV + pembro (n = 121) | 18 (14.9) | 0.68 (0.37-1.26) Cis + gem (n = 102) | 23 (22.5) | ctDNA+ EV + pembro (n = 133) | 41 (30.8) | 0.54 (0.37-0.80) Cis + gem (n = 134) | 64 (47.8) | No. at risk (0, 6, 12, 18, 24, 30, 36, 42, 48, 54 months) EV + pembro: ctDNA− 121 103 96 93 80 46 29 21 10 0 Cis + gem: ctDNA− 102 95 84 82 68 40 22 13 8 0 EV + pembro: ctDNA+ 133 119 101 95 83 51 30 20 8 0 Cis + gem: ctDNA+ 134 111 83 73 62 36 19 14 5 0 ᵃUnstratified Cox proportional hazard model with treatment as a covariate. Data cutoff date: October 27, 2025. Survival table Endpoint: EFS per BICR by baseline ctDNA status ctDNA− EV + pembro | 121 ctDNA− Cis + gem | 102 ctDNA+ EV + pembro | 133 ctDNA+ Cis + gem | 134 HR: ctDNA− 0.68; ctDNA+ 0.54 HR CI: ctDNA− 0.37-1.26; ctDNA+ 0.37-0.80 Landmark rates: 24 mo (dashed line): ctDNA− EV + pembro 86.1% | ctDNA− Cis + gem 82.7% | ctDNA+ EV + pembro 71.2% | ctDNA+ Cis + gem 54.2%

Kaplan-Meier Estimates of EFS per BICR by ctDNA Status at Postsurgery [KM plot: Event-Free Survival, % vs Months 0-54; dashed line at 24 months] 24-month labels: 89.9% | 82.2% | 10.0% | 10.0% Events, n (%) | HR (95% CI)a ctDNA- EV + pembro (n = 199) 22 (11.1) | 0.48 Cis + gem (n = 193) 41 (21.2) | (0.29-0.81) ctDNA+ EV + pembro (n = 10) 9 (90.0) | 0.63 Cis + gem (n = 30) 27 (90.0) | (0.30-1.35) No. at risk (0, 6, 12, 18, 24, 30, 36, 42, 48, 54 months) EV + pembro: ctDNA- 199 197 186 180 162 91 58 38 16 0 Cis + gem: ctDNA- 193 191 173 162 131 75 43 26 12 0 EV + pembro: ctDNA+ 10 9 3 2 1 0 0 0 0 0 Cis + gem: ctDNA+ 30 22 5 3 3 2 2 1 1 0 aUnstratified Cox proportional hazards model with treatment as a covariate. Data cutoff date: October 27, 2025. [ESMO congress slide frame] Survival table Endpoint: EFS per BICR by postsurgery ctDNA status (exploratory) ctDNA- EV + pembro | 199 ctDNA- Cis + gem | 193 ctDNA+ EV + pembro | 10 ctDNA+ Cis + gem | 30 HR: ctDNA-: 0.48; ctDNA+: 0.63 HR CI: ctDNA-: 0.29-0.81; ctDNA+: 0.30-1.35 Landmark rates: 24 months: ctDNA- EV + pembro 89.9% | ctDNA- Cis + gem 82.2% | ctDNA+ EV + pembro 10.0% | ctDNA+ Cis + gem 10.0%

Kaplan-Meier Estimates of EFS per BICR by ctDNA Status at Postsurgery Y-axis: Event-Free Survival, % Curve labels at 24 months (dashed line): [faint green label, illegible] (EV + pembro ctDNA−), 82.2% [partially obscured] (Cis + gem ctDNA−), 10.0% (EV + pembro ctDNA+), 10.0% (Cis + gem ctDNA+) [Table] Events, n (%) | HR (95% CI)^a ctDNA− EV + pembro (n = 199) [faint] | 22 (11.1) [faint] | 0.48 Cis + gem (n = 193) | 41 (21.2) | (0.29-0.81) ctDNA+ EV + pembro (n = 10) | 9 (90.0) | 0.63 Cis + gem (n = 30) | 27 (90.0) | (0.30-1.35) X-axis: Months No. at risk (0, 6, 12, 18, 24, 30, 36, 42, 48, 54) EV + pembro: ctDNA− 199 197 186 180 162 91 58 38 16 0 [faint] Cis + gem: ctDNA− 193 191 173 162 131 75 43 26 12 0 EV + pembro: ctDNA+ 10 9 3 2 1 0 0 0 0 0 Cis + gem: ctDNA+ 30 22 5 3 3 2 2 1 1 0 ^aUnstratified Cox proportional hazards model with treatment as a covariate. Data cutoff date: October 27, 2025. Survival table Endpoint: EFS per BICR by postsurgery ctDNA status (exploratory) EV + pembro, ctDNA− | 199 Cis + gem, ctDNA− | 193 EV + pembro, ctDNA+ | 10 Cis + gem, ctDNA+ | 30 HR: ctDNA−: 0.48; ctDNA+: 0.63 HR CI: ctDNA−: 0.29-0.81; ctDNA+: 0.30-1.35 Landmark rates: 24 mo: Cis + gem, ctDNA− 82.2% | EV + pembro, ctDNA+ 10.0% | Cis + gem, ctDNA+ 10.0%

[Slide 1] ctDNA Analysis of the Phase 3 KEYNOTE-B15 Trial: Neoadjuvant and Adjuvant Enfortumab Vedotin Plus Pembrolizumab for Participants With Muscle-Invasive Bladder Cancer Who Are Eligible for Cisplatin Jonathan Chatzkel1; Matthew D. Galsky2; Alfonso Gómez de Liaño3; Begoña Pérez-Valderrama4; Marco Maruzzo5; Marco J. Schnabel6; Avivit Peer7; Albert Font Pous8; Tudor Ciuleanu9; Christopher Hoimes10; Takuya Koie11; Yousef Zakharia12; Flora Zagouri13; Leah Hogdal14; Michele Wozniak15; Julia Markensohn16; Julie Kobie16; Carol Peña16; Ja Hyeon Ku17; Thomas Powles18 1University of Florida, Gainesville, FL, USA; [current affiliation] Moffitt Cancer Center, Tampa, FL, USA; 2Icahn School of Medicine at Mount Sinai, Tisch Cancer Institute, New York, NY, USA; 3Complejo Hospitalario Universitario Insular-Materno Infantil de Gran Canaria, Las Palmas, Spain; 4Hospital Universitario Virgen del Rocío, Seville, Spain; 5Istituto Oncologico Veneto IRCCS-Oncologia Medica 1, Padova, Italy; 6Caritas - St. Josef Medical Center, University of Regensburg, Regensburg, Germany; 7Institute of Oncology, Rambam Health Care Campus, Haifa, Israel; 8Catalan Institute of Oncology (ICO) Badalona, Barcelona, Spain; 9Institutul Oncologic Prof. Dr. Ion Chiricuta Cluj-Napoca, Cluj-Napoca, Romania; 10Duke University Medical Center, Durham, NC, USA; 11Gifu University Hospital, Gifu, Japan; 12Mayo Clinic, Phoenix, AZ, USA; 13Alexandra Hospital, Athens, Greece; 14Pfizer, Bothell, WA, USA; 15Astellas Pharma Inc., Northbrook, IL, USA; 16Merck & Co., Inc., Rahway, NJ, USA; 17Seoul National University Hospital, Seoul, Republic of Korea; 18Barts Health NHS Trust and the Royal Free NHS Foundation Trust, Barts Cancer Institute, and Queen Mary University of London, London, United Kingdom [Slide 2] Pre-Treatment and Presurgery ctDNA Predict Outcome KEYNOTE-B15 vs NIAGARA and Christensen 2019 KEYNOTE-B15 / EV-304 EFS Pre-chemotherapy ctDNA EV + pembro: curve labels at 24 months: [illegible, upper/ctDNA− curve]; 71.2% [ctDNA+] No. at risk (months 0 / 6 / 12 / 18 / 24 / 30 / 36 / 42 / 48 / 54): ctDNA+: 133 / 119 / 101 / 95 / 83 / 51 / 30 / 20 / 8 / 0 ctDNA−: 121 / 103 / 96 / 93 / 89 / 46 / 29 / 21 / 10 / 0 [faint] Cis + gem: curve labels at 24 months: 82.7% [ctDNA−]; 54.2% [ctDNA+] No. at risk: ctDNA+: 134 / 111 / 83 / 73 / 62 / 36 / 19 / 14 / 5 / 0 ctDNA−: 102 / 95 / 84 / 82 / 68 / 40 / 22 / 13 / 8 / 0 [faint] Pre-surgery ctDNA EV + pembro: Positive-Positive (n = 15); Negative-Negative (n = 79); curve labels at 24 months: 94.6%; 33.3% No. at risk: 15 / 12 / 8 / 6 / 5 / 3 / 2 / 1 / 1 / 0 79 / 73 / 71 / 69 / 62 / 32 / 21 / 16 / 8 / 0 Cis + gem: Positive-Positive (n = 34); Negative-Negative (n = 77); curve labels at 24 months: 87.8%; 35.3% No. at risk: 34 / 26 / 16 / 14 / 12 / 10 / 5 / 4 / 2 / 0 77 / 74 / 67 / 65 / 53 / 31 / 19 / 12 / 7 / 0 NIAGARA EFS, combined arms [top] EFS (combined arms): ctDNA− vs ctDNA+ HR, 0.42 (95% CI, 0.30-0.60) [bottom] EFS in the baseline ctDNA+ population that did or did not clear ctDNA pre-cystectomy (combined arms): ctDNA clearance vs persistent ctDNA+ HR, [illegible] (95% CI [illegible]) Christensen et al. 2019 RFS A ctDNA Before Chemotherapy — RFS (probability) vs Time Since Cystectomy (days); P < .001 No. at risk (0 / 300 / 600 / 900 days): Negative 35 / 32 / 23 / 1; Positive 24 / 16 / 4 / 0 B ctDNA Before Cystectomy — P < .001 No. at risk: Negative 55 / 49 / 29 / 1; Positive [illegible] Survival table Endpoint: EFS (KEYNOTE-B15/EV-304) by pre-chemotherapy and pre-surgery ctDNA status; NIAGARA EFS; Christensen 2019 RFS KEYNOTE-B15 EV + pembro pre-surgery Positive-Positive | 15 KEYNOTE-B15 EV + pembro pre-surgery Negative-Negative | 79 KEYNOTE-B15 Cis + gem pre-surgery Positive-Positive | 34 KEYNOTE-B15 Cis + gem pre-surgery Negative-Negative | 77 KEYNOTE-B15 EV + pembro pre-chemo ctDNA+ | 133 KEYNOTE-B15 EV + pembro pre-chemo ctDNA- | 121 KEYNOTE-B15 Cis + gem pre-chemo ctDNA+ | 134 KEYNOTE-B15 Cis + gem pre-chemo ctDNA- | 102 HR: NIAGARA EFS ctDNA- vs ctDNA+ 0.42 HR CI: 0.30-0.60 P: Christensen 2019: P < .001 (both panels) Landmark rates: 24 mo: KEYNOTE-B15 EV + pembro pre-chemo ctDNA+ 71.2% | KEYNOTE-B15 Cis + gem pre-chemo ctDNA- 82.7% | KEYNOTE-B15 Cis + gem pre-chemo ctDNA+ 54.2% | KEYNOTE-B15 EV + pembro pre-surgery Negative-Negative 94.6% | KEYNOTE-B15 EV + pembro pre-surgery Positive-Positive 33.3% | KEYNOTE-B15 Cis + gem pre-surgery Negative-Negative 87.8% | KEYNOTE-B15 Cis + gem pre-surgery Positive-Positive 35.3% [Slide 3] [Title partially cropped at left] [Hi]gher pCR With EV + Pembro Even Among ctDNA− Patients [pC]R by Presurgery ctDNA Status, KEYNOTE-B15 NIAGARA7 [greyed] ctDNA−: 78% (315/404) ctDNA+: 22% (89/404) pCR 51%; Non-pCR 49% (n = 315) pCR 3%; Non-pCR 97% (n = 89) Combined arms [greyed] ctDNA−: 83.5% (364/436) ctDNA+: 16.5% (72/436) pCR 58.0% (211/364); Non-pCR 42.0% (153/364) pCR 2.8% (2/72); Non-pCR 97.2% (70/72) Cis + gem ctDNA−: 80.3% (183/228) ctDNA+: 19.7% (45/228) pCR 43.7% (80/183); Non-pCR 56.3% (103/183) pCR 2.2% (1/45); Non-pCR 97.8% (44/45) EV + pembro ctDNA−: 87.0% (181/208) ctDNA+: 13.0% (27/208) pCR 72.4% (131/181); Non-pCR 27.6% (50/181) pCR 3.7% (1/27); Non-pCR 96.3% (26/27) pCR 72.4% vs 43.7% +28.7 pts, p < 0.001 · post hoc, exploratory [Footnote, left edge cropped:] [Pie charts ar]e scaled according to the relative number of participants in each subgroup. Data cutoff date: October 27, 2025. Post hoc χ² (1 df) for pCR in presurgery ctDNA− patients, EV + pembro 131/181 vs cis + gem 80/183: χ² = 30.7, p [cropped] ... NIAGARA: pre-RC ctDNA BEP who had RC; percentages as reported (Powles T, et al. J Clin Oncol. 2025;43(16 Suppl):4503). Presurgery ctDNA [...] [diffe]rence 28.7 pts (95% CI 19.0–38.4). Exploratory, not reported by the authors. [...] ctDNA− groups combine baseline-negative patients and those who cleared ctDNA, so between-arm comparisons within ctDNA strata are not randomised and are hypothesis-generating. [... po]st-randomisation outcome: [...] [Slide 4] Postsurgery ctDNA Findings Replicate NIAGARA KEYNOTE-B15 vs NIAGARA KEYNOTE-B15 / EV-304 study scheme [small schematic: Key eligibility: Adults with MIBC, cisplatin eligible, RC + PLND candidate; R 1:1; Neoadjuvant Enfortumab vedotin + Pembrolizumab N=405 → RC + PLND → Adjuvant Enfortumab vedotin + Pembrolizumab [highlighted]; Cisplatin + Gemcitabine N=403 → Observation; ctDNA assessed at baseline, presurgery, postsurgery and adjuvant] NIAGARA study scheme [small schematic: Durvalumab arm — Neoadjuvant Durvalumab + Gemcitabine + cisplatin → Radical cystectomy → Adjuvant Durvalumab [highlighted]; Comparator arm — Gemcitabine + cisplatin → No treatment; plasma samples analysed at baseline, pre-RC, C1D1; study population text too small to read] KEYNOTE-B15 / EV-304 EFS per BICR by postsurgery ctDNA status | Events, n (%) | HR (95% CI)a ctDNA+ EV + pembro (n = 10) | 9 (90.0) | 0.63 (0.30-1.35) Cis + gem (n = 30) | 27 (90.0) ctDNA− EV + pembro (n = 199) | 22 (11.1) | 0.48 (0.29-0.81) Cis + gem (n = 193) | 41 (21.2) Curve labels at 24 months: 89.9% [EV + pembro ctDNA−]; 82.2% [Cis + gem ctDNA−]; 10.0%; 10.0% [ctDNA+ arms] Y-axis: Event-Free Survival, % X-axis: Months No. at risk (months 0 / 6 / 12 / 18 / 24 / 30 / 36 / 42 / 48 / 54): EV + pembro: ctDNA+: 10 / 9 / 3 / 2 / 1 / 0 / 0 / 0 / 0 / 0 EV + pembro: ctDNA−: 199 / 197 / 186 / 180 / 162 / 91 / 58 / 38 / 16 / 0 Cis + gem: ctDNA+: 30 / 22 / 5 / 3 / 3 / 2 / 2 / 1 / 1 / 0 Cis + gem: ctDNA−: 193 / 191 / 173 / 162 / 131 / 75 / 43 / 26 / 12 / 0 NIAGARA7 DFS by post-RC ctDNA status, combined arms Curves: ctDNA−; ctDNA+ Y-axis: Probability of DFS X-axis: Time from radical cystectomy (months) [No. of patients at risk row too small to read] ctDNA− vs ctDNA+ HR, 0.09 (95% CI, 0.05–0.18) aUnstratified Cox proportional hazards model with treatment as a covariate. Data cutoff date: October 27, 2025. NIAGARA: DFS from radical cystectomy, ctDNA− vs ctDNA+ HR 0.09 (95% CI 0.05–0.18); Powles T, et al. J Clin Oncol. 2025;43(16 Suppl):4503. Survival table Endpoint: KEYNOTE-B15/EV-304 EFS per BICR by postsurgery ctDNA status; NIAGARA DFS by post-RC ctDNA status (combined arms) EV + pembro ctDNA+ | 10 Cis + gem ctDNA+ | 30 EV + pembro ctDNA- | 199 Cis + gem ctDNA- | 193 HR: ctDNA+: 0.63; ctDNA-: 0.48 (EV + pembro vs cis + gem); NIAGARA ctDNA- vs ctDNA+ DFS 0.09 HR CI: ctDNA+: 0.30-1.35; ctDNA-: 0.29-0.81; NIAGARA: 0.05-0.18 Landmark rates: 24 mo: EV + pembro ctDNA- 89.9% | Cis + gem ctDNA- 82.2% | ctDNA+ (both arms) 10.0% / 10.0%

[Slide 2] KEYNOTE-B15 ctDNA-Evaluable Population^a 808 participants randomly assigned • Median study follow-up: 33.6 months (range, 22.5-53.6) EV + pembro • 405 assigned (ITT population) Cis + gem • 403 assigned (ITT population) Exploratory ctDNA analysis ctDNA-evaluable population at each time point (% of ITT) EV + pembro: Baseline n = 254 (62.7%); Presurgery n = 208 (51.4%); Postsurgery n = 209 (51.6%); Adjuvant n = 69 (17.0%) Cis + gem: Baseline n = 236 (58.6%); Presurgery n = 228 (56.6%); Postsurgery n = 223 (55.3%); Adjuvant n = 119 (29.5%) ctDNA status change-evaluable population (% of ITT) EV + pembro: Baseline→Presurgery n = 167 (41.2%); Presurgery→Postsurgery n = 154 (38.0%); Postsurgery→Adjuvant n = 64 (15.8%) Cis + gem: Baseline→Presurgery n = 182 (45.2%); Presurgery→Postsurgery n = 170 (42.2%); Postsurgery→Adjuvant n = 102 (25.3%) [Slide 4] Pre- and On-Treatment ctDNA Status^a Timeline: Baseline — Neoadjuvant — Presurgery — RC + PLND — Postsurgery — Adjuvant EV + pembro (ctDNA+ / ctDNA−) Baseline: 52.4% (133/254) Presurgery: 13.0% (27/208) Postsurgery: 4.8% (10/209) Adjuvant: 5.8% (4/69) Cis + gem (ctDNA+ / ctDNA−) Baseline: 56.8% (134/236) Presurgery: 19.7% (45/228) Postsurgery: 13.5% (30/223) Adjuvant: 4.2% (5/119)

KEYNOTE-B15/EV-304 Study (NCT04700124) Key Eligibility Criteria • Adults with MIBC • Clinical stage T2-T4aN0M0 or T1-T4aN1M0 by central assessment • Urothelial histology ≥50% • Eligible for RC + PLND • Did not meet any Galsky criteria for cisplatin ineligibility • ECOG PS score of 0 or 1 R 1:1 N = 405: Enfortumab vedotin 1.25 mg/kg d1 and d8 IV Q3W 4 cycles + Pembrolizumab 200 mg d1 IV Q3W 4 cycles → RC + PLND → Enfortumab vedotin 1.25 mg/kg d1 and d8 IV Q3W 5 cycles + Pembrolizumab 200 mg d1 IV Q3W 13 cycles N = 403: Cisplatin 70 mg/m² d1 IV Q3W 4 cycles + Gemcitabine 1000 mg/m² d1 and d8 IV Q3W 4 cycles → RC + PLND → Observation^b Stratification Factors • PD-L1 status (CPS ≥10 vs <10)^a • Clinical stage (T2N0 vs T3/T4aN0 vs T1-T4aN1) • Geographic region (US vs EU vs most of world) • Primary results from KEYNOTE-B15 showed neoadjuvant + adjuvant enfortumab vedotin + pembrolizumab and RC + PLND significantly improved EFS (HR, 0.53 [95% CI, 0.41-0.70]; P < 0.001), OS (HR, 0.65 [95% CI, 0.48-0.89]; 2-sided P = 0.006), and pCR (P < 0.001) vs neoadjuvant chemotherapy followed by RC + PLND in cisplatin-eligible participants with MIBC¹ • The safety profile of enfortumab vedotin + pembrolizumab and RC + PLND was consistent with prior reports¹ • Results from the prespecified analysis of HRQoL are presented

[Slide 1] Patient-Reported Outcome (PRO) Analyses • HRQoL was assessed by validated cancer-specific PRO questionnaires • Prespecified secondary PRO end points (not multiplicity controlled): – Change from baseline to postsurgery week 42^a and time to definitive deterioration^b in: ▪ European Organisation for Research & Treatment of Cancer (EORTC) Quality of Life Questionnaire-Core 30 (QLQ-C30): global health status/QoL (GHS/QoL), physical, and role functioning scores ▪ Bladder Cancer Index (BCI) urinary, bowel, and sexual scores • PRO analysis population: all randomly allocated participants with ≥1 PRO assessment available for the specific end point and who received ≥1 dose of the study treatment PRO assessment schedule Neoadjuvant treatment (days 1 and 8 of each 21-day cycle) | Presurgery visit | RC + PLND | Postsurgery phase - adjuvant EV + pembro or observation (on day 1 of each 21-day cycle) Week: 0, 2 | 4, 5 | 7, 8 | 10, 11 | 13 | 0 | 9^c | 12 | 18 | 30 | 42 | 54 | Y2 Q12W | Y3 Q24W EV + pembro: X, X | X, X | X, X | X, X | X | | X | X | X | X | X | X | X | X Cis + gem: X, X | X, X | X, X | X, X | X | | X | | X | X | X | X | X | X ^a Time point where PRO completion and compliance rates were ~60% and ~80%, respectively, based on blinded data. ^b Defined as the first onset of deterioration confirmed at a subsequent assessment and persistent at all subsequent visits without recovery. ^c The 6-week postsurgery visit could be combined with the first adjuvant treatment cycle visit (week 9 post-surgery). Median time from randomization to database cutoff: 33.6 months (range, 22.5-53.6). Database cutoff: October 27, 2025. [Slide 3] Empirical Mean Change From Baseline Over Time in EORTC QLQ-C30 Scores [Three line-plot panels: GHS/QoL, Physical Functioning, Role Functioning; Mean Change From Baseline (95% CI); EV + Pembro (teal) vs Cis + Gem (maroon); shaded SURGERY band; Improvement / Decline arrows] Presurgery Week: 0 2 4 5 7 8 10 11 13 | Postsurgery Week: 9 18 30 42 54 66 78 90 102 126 150 174 198 No. (EV + Pembro): 383 354 300 314 289 299 285 290 308 | 289 257 242 241 232 218 208 195 153 114 74 46 17 No. (Cis + Gem), GHS/QoL panel: 368 331 289 304 291 319 293 283 316 | 295 234 224 213 202 184 183 167 135 90 51 29 12 No. (Cis + Gem), Physical and Role Functioning panels: 368 331 289 305 291 319 293 283 316 | 295 234 224 213 202 184 183 167 135 90 51 29 12 Database cutoff: October 27, 2025.

What did XmAb819 achieve? A strong efficacy signal in general and in particular for a strategy that is immunotherapy alone Among the highest single-agent response rates (cross-trial comparison) How well it worked Confirmed or RECIST ORR, % Single agents XmAb819 · 400 µg/kg | 45.5 XmAb819 · 200 µg/kg | 44.0 LITESPARK-011 · Cabozantinib * | 40.2 Casdatifan 100 mg | 35.0 XmAb819 · 100 µg/kg | 30.0 Belzutifan | 22.7 ALLO-316 | 17.4 Combinations LITESPARK-011 · Belz + len * | 52.6 KM-U03 · Belz + lenvatinib | 50.0 Pembro + lenvatinib | 43.8 Pembro + quavonlimab | 30.0 Pembro + belzutifan | 22.6 Pembro + favezelimab | 0.0 Pembro + MK-4830 | 0.0 Responses are frequent 30–45%, three quarters still ongoing Depth takes time one CR so far – it took 42 weeks Pal SK et al., ESMO 2026, abstract [illegible]. * LITESPARK-011 (Motzer RJ et al., Lancet 2025; second interim analysis, by blinded independent central review): belzutifan + lenvatinib 14.8 vs cabozantinib 10.7 months, HR 0.70 [illegible]; ORR 52.6% vs 40.2%; CR [illegible]. [illegible] Cross-trial comparison. Manuela Schmidinger, Medical University of Vienna, Austria

[Slide 1] LITESPARK-011 Study (NCT04586231) Key Eligibility Criteria • Unresectable, locally advanced or metastatic ccRCC • KPS score ≥70% • ≤2 prior systemic regimens • Progression on/after anti–PD-(L)1 mAb as 1L or 2L therapy or progression ≤6 months of the last dose of adjuvant anti–PD-(L)1 mAb • Prior VEGFR-TKI was permitted R 1:1 N = 371 → Belzutifan 120 mg orally QD + Lenvatinib 20 mg orally QD N = 376 → Cabozantinib 60 mg orally QD Stratification Factors • IMDC prognostic score:a 0 vs. 1-2 vs. 3-6 • Line of treatment for prior anti–PD-(L)1: 1L, adjuvant, neoadjuvant-adjuvant vs. 2L • Geographic region: North America vs. western Europe vs. rest of world Dual Primary Endpoints: • PFS per RECIST 1.1 by BICR • OS Key Secondary Endpoint: • ORR per RECIST 1.1 by BICR Other Secondary Endpoints Include: • DOR per RECIST 1.1 by BICR • Safety Exploratory Analyses Include: • Subgroup efficacy analysis by type of IO combination in 1L (IO-IO vs. IO-TKI) • Progression-free survival 2 BICR, blinded independent central review; DOR, duration of response; IO, immunotherapy; KPS, Karnofsky Performance Status; OS, overall survival; QD, daily. aBased on the number of present risk factors (KPS score <80%; time from diagnosis to 1L treatment <1 year; low hemoglobin; high corrected serum calcium; high neutrophils; high levels of platelets) according to the International Metastatic Renal Cell Carcinoma Database Consortium (IMDC). [Slide 2] Primary Endpoint: PFS per RECIST 1.1 by BICR Belzutifan + Lenvatinib (N = 371) | Cabozantinib (N = 376) Events, n (%) | 262 (70.6) | 285 (75.8) Median, mo (95% CI) | 14.8 (11.2–16.6) | 10.8 (9.2–11.1) HR (95% CI) | 0.74 (0.62–0.88) Nominal P value (1-sided)a | 0.00035 Landmarks: 24 months: 35.2% (Belzutifan + lenvatinib), 20.0% (Cabozantinib); 36 months: 20.5% (Belzutifan + lenvatinib), 12.6% (Cabozantinib) No. at Risk (Months 0, 5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60) Belzutifan + lenvatinib | 371 | 292 | 204 | 151 | 131 | 104 | 70 | 44 | 28 | 13 | 3 | 0 | 0 Cabozantinib | 376 | 280 | 167 | 103 | 85 | 55 | 30 | 19 | 7 | 2 | 0 | 0 | 0 aPrimary PFS endpoint was met at previous interim analysis and was not formally statistically tested at final analysis. Data cutoff date: 2 Feb 2026 Survival table Endpoint: PFS per RECIST 1.1 by BICR Arm | N | Median | Median CI Belzutifan + lenvatinib | 371 | 14.8 | 11.2–16.6 Cabozantinib | 376 | 10.8 | 9.2–11.1 HR: 0.74 HR CI: 0.62–0.88 P: 0.00035 (nominal, 1-sided) Landmark rates: 24 mo: Belzutifan + lenvatinib 35.2% | Cabozantinib 20.0% 36 mo: Belzutifan + lenvatinib 20.5% | Cabozantinib 12.6% [Slide 3] Primary Endpoint: OS Belzutifan + Lenvatinib (N = 371) | Cabozantinib (N = 376) Events, n (%) | 195 (52.6) | 224 (59.6) Median, mo (95% CI) | 33.7 (29.0–46.9) | 28.6 (24.1–31.4) HR (95% CI) | 0.85 (0.70–1.03) P=0.04689 (1-sided boundary 0.0223) Landmarks: 36 mo: 48.9% (Belzutifan + lenvatinib), 40.6% (Cabozantinib); 42 mo: 45.8% (Belzutifan + lenvatinib), 37.0% (Cabozantinib) No. at Risk (Months 0, 5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60) Belzutifan + lenvatinib | 371 | 345 | 311 | 273 | 252 | 228 | 194 | 117 | 74 | 44 | 11 | 1 | 0 Cabozantinib | 376 | 355 | 307 | 271 | 234 | 200 | 168 | 105 | 56 | 32 | 8 | 4 | 0 Data cutoff date: 2 Feb 2026 Survival table Endpoint: OS Arm | N | Median | Median CI Belzutifan + lenvatinib | 371 | 33.7 | 29.0–46.9 Cabozantinib | 376 | 28.6 | 24.1–31.4 HR: 0.85 HR CI: 0.70–1.03 P: 0.04689 (1-sided boundary 0.0223) Landmark rates: 36 mo: Belzutifan + lenvatinib 48.9% | Cabozantinib 40.6% 42 mo: Belzutifan + lenvatinib 45.8% | Cabozantinib 37.0% [Slide 4] ORR (Key Secondary) and DOR (Secondary Endpoint) per RECIST 1.1 by BICR Estimated difference 12.9% (95% CI 5.8–19.8) Nominal P value = 0.0002a Belzutifan + Lenvatinib: ORR 53.1% (95% CI 47.9–58.3); CR=6.7%; PR=46.4% Cabozantinib: ORR 40.2% (95% CI 35.2–45.3); CR=1.3%; PR=38.8% Belzutifan + Lenvatinib (N = 371) | Cabozantinib (N = 376) Pts with response, n | 197 | 151 Median DOR, mo (range) | 22.0 (2.0 to 52.2+) | 11.9 (1.8+ to 44.1+) DOR landmark 30 mo: 37.5% (Belzutifan + lenvatinib), 22.6% (Cabozantinib) No. at Risk (Months 0, 5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60) Belzutifan + lenvatinib | 197 | 181 | 137 | 110 | 93 | 74 | 50 | 26 | 17 | 5 | 1 | 0 | 0 Cabozantinib | 151 | 137 | 80 | 59 | 48 | 31 | 19 | 10 | 3 | 0 | 0 | 0 | 0 [left edge cropped]y secondary ORR endpoint was met at previous interim analysis and was not formally statistically tested at final analysis. Data cutoff date: 2 Feb 2026 Survival table Endpoint: DOR per RECIST 1.1 by BICR (responders) Belzutifan + lenvatinib | 197 | 22.0 Cabozantinib | 151 | 11.9 Landmark rates: 30 mo: Belzutifan + lenvatinib 37.5% | Cabozantinib 22.6%

[Slide 1] 1- Does HIF-2α inhibition improve overall survival? In combination – LS-011 Primary Endpoint: OS | Belzutifan + Lenvatinib (N = 371) | Cabozantinib (N = 376) Events, n (%) | 195 (52.6) | 224 (59.6) Median, mo (95% CI) | 33.7 (29.0–46.9) | 28.6 (24.1–31.4) HR (95% CI) | 0.85 (0.70–1.03) 36 mo: 48.9% (Belzutifan + Lenvatinib) | 40.6% (Cabozantinib) 42 mo: 45.8% (Belzutifan + Lenvatinib) | 37.0% (Cabozantinib) P=0.04689 (1-sided boundary 0.0223) No. at Risk Belzutifan + lenvatinib 371 345 311 273 252 228 194 117 74 44 11 1 0 Cabozantinib 376 355 307 271 234 200 168 105 56 32 8 4 0 Combination of belzutifan with lenvatinib increased ORR and PFS vs single-agent cabozantinib but did not demonstrate a statistically significant OS benefit Data cutoff date: 2 Feb 2026 Presented by D. Heng, ESMO 2026 Heng et al. ESMO 2026 Survival table Endpoint: Overall survival (primary endpoint) Arm | N | Median | Median CI Belzutifan + Lenvatinib | 371 | 33.7 mo | 29.0–46.9 Cabozantinib | 376 | 28.6 mo | 24.1–31.4 HR: 0.85 HR CI: 0.70–1.03 P: 0.04689 (1-sided boundary 0.0223) Landmark rates: 36 mo: Belzutifan + Lenvatinib 48.9% | Cabozantinib 40.6% 42 mo: Belzutifan + Lenvatinib 45.8% | Cabozantinib 37.0% [Slide 2] 2- Should HIF-2α inhibitors be used alone or in combination? Is there a rationale to combine with VEGFR-TKI? [Figure: Preclinical Discovery (VEGFR TKI/IO, KVpR ccRCC GEMM, Hypoxyprobe IF, scRNA-seq); Human Validation (ccRCC patients, + TKI/IO or IO/IO treatment, scRNA-seq, Spatial RNA-seq, IMC); VEGFR TKI/IO responders (ccRCC, ↑VEGF, Vhl-/- pseudohypoxia, VEGFR TKI, chronic VEGFR TKI, SPP1 TAM, metastasis, True TME hypoxia); Functional Validation (VEGFR TKI/IO, Metastatic ccRCC syngeneic models, Metastatic phenotype, IF staining, Flow cytometry, Depletion antibodies); Biomarker Development (Bulk RNA-seq, Pretreatment, KVpR-derived gene signatures, Hypoxia-skewed low, Hypoxia-skewed high, TKI/IO response)] • VEGFR-TKI promotes severe intratumor hypoxia Could the pseudohypoxia related to ccRCC biology and the VEGFR-TKI-induced hypoxia have an impact on the response to HIF2α inhibition? • Pseudohypoxia and true hypoxia have distinct immune microenvironments Laurence Albiges, MD PhD Vuong et al. Cancer Cell. 2026 [Slide 4] Does casdatifan have a distinct activity? b: Pharmacodynamic response (% EPO change from the baseline) vs Total daily casdatifan dose (mg) c: Maximal sEPO change (%) by Patient order; cBOR: PD, SD, CR + PR e: Best sEPO reduction (%) CR + PR P = 0.001 (Odds ratio; CR/PR less likely <- -> CR/PR more likely) PD P = 0.003 (Odds ratio; PD less likely <- -> PD more likely) PFS P = 0.006 (HR; Progression less likely <- -> Progression more likely) Can sEPO be used as a pharmacodynamic biomarker? b: tEPO expression (log2[CPM]) by Disease control (No / Yes), P = 0.0075 Laurence Albiges, MD PhD Choueiri et al. Nature. 2026

[Slide 1] MADRID 2026 ESMO congress Relationship Between Health-Related Quality of Life and Efficacy in LITESPARK-011: Belzutifan Plus Lenvatinib Versus Cabozantinib in Advanced Clear Cell Renal Cell Carcinoma After Anti–PD-(L)1 Therapy Daniel Y. C. Heng1; Se Hoon Park2; Hernan J. Cutuli3; Arun Azad4; Sylvie Rottey5; Manuela Schmidinger6; Mauricio Burotto7; Michel Pavic8; Masatoshi Eto9; Camillo Porta10; Ray Manneh Kopp11; Guillermo de Velasco12; Elena Verzoni13; Robert A. Figlin14; Todd L. Saretsky15; Erin H. Jensen15; Rodolfo F. Perini15; Donna Vickery15; Robert J. Motzer16 1Arthur J.E. Child Comprehensive Cancer Centre, Calgary, AB, Canada; 2Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Republic of Korea; 3Uro-Oncology Research Unit, Hospital Sirio Libanês, Buenos Aires, Argentina; 4Peter MacCallum Cancer Centre, Melbourne VIC, Australia; 5Ghent University Hospital, Ghent, Belgium; 6Medical University of Vienna, Vienna, Austria; 7Bradford Hill Clinical Research Center, Santiago, Chile; 8Centre Hospitalier Universitaire de Sherbrooke, Sherbrooke, QC, Canada; 9Kyushu University Hospital, Fukuoka, Japan; 10University of Bari "A. Moro" and A.O.U. Consorziale Policlinico Di Bari, Bari, Italy; 11Sociedad de Oncología y Hematología del Cesar S.A.S., Valledupar, Colombia; 12University Hospital 12 de Octubre, Madrid, Spain; 13Istituto Nazionale dei Tumori, Milano, Italy; 14Cedars Sinai Cancer, Cedars Sinai Medical Center, Los Angeles, CA, USA; 15Merck & Co., Inc., Rahway, NJ, USA; 16Memorial Sloan Kettering Cancer Center, New York, NY, USA [Slide 2] Association Between Baseline and Longitudinal PRO Scores and PFS and OS PFS | Events/Participants | HR (95% CI)a Baseline modelb FKSI-DRS: Pooled arms 434/633 0.86 (0.81-0.91); Belz + len 197/318 0.84 (0.77-0.92); Cabo 237/315 0.87 (0.79-0.95) QLQ-C30 GHS/QoL: Pooled arms 426/621 0.90 (0.86-0.94); Belz + len 192/311 0.88 (0.83-0.95); Cabo 234/310 0.90 (0.84-0.97) Longitudinal modelc FKSI-DRS: Pooled arms 400/633 0.82 (0.76-0.89); Belz + len 175/318 0.81 (0.70-0.93); Cabo 225/315 0.83 (0.75-0.92) QLQ-C30 GHS/QoL: Pooled arms 393/621 0.92 (0.86-0.99); Belz + len 171/311 0.89 (0.80-0.99); Cabo 222/310 0.93 (0.84-1.04) HR for progression or death reduces as PRO score increases OS | Events/Participants | HR (95% CI)a Baseline modelb FKSI-DRS: Pooled arms 355/640 0.81 (0.77-0.86); Belz + len 164/319 0.83 (0.76-0.91); Cabo 191/321 0.74 (0.68-0.82) QLQ-C30 GHS/QoL: Pooled arms 347/628 0.88 (0.83-0.92); Belz + len 159/312 0.91 (0.85-0.99); Cabo 188/316 0.83 (0.78-0.90) Longitudinal modelc FKSI-DRS: Pooled arms 130/640 0.57 (0.52-0.62); Belz + len 62/319 0.51 (0.44-0.59); Cabo 68/321 0.58 (0.52-0.66) QLQ-C30 GHS/QoL: Pooled arms 127/628 0.68 (0.62-0.74); Belz + len 60/312 0.65 (0.57-0.73); Cabo 67/316 0.69 (0.60-0.80) HR for death reduces as PRO score increases aPer Cox proportional hazards model. HR was calculated as risk of event per 3-point (FKSI-DRS) or 10-point (QLQ-C30 GHS/QoL) increase from the baseline score. bIncluded randomization stratification factors as categorical terms and baseline HRQoL scores as a continuous covariate. cIncluded randomization stratification factors as categorical terms and baseline HRQoL scores and changes from baseline HRQoL scores assessed after randomization visits as continuous covariates. Data cutoff date: PFS, April 9, 2025 (IA2); OS, February 2, 2026.

PFS by BICR in Key Subgroups Events/Participants: Belzutifan + lenvatinib | Cabozantinib | HR (95% CI) Overall | 262/371 | 285/376 | 0.74 (0.62-0.88) Age <65 years | 150/209 | 171/224 | 0.75 (0.60-0.94) Age ≥65 years | 112/162 | 114/152 | 0.64 (0.49-0.84) Sex Male | 187/273 | 224/292 | 0.69 (0.56-0.84) Sex Female | 75/98 | 61/84 | 0.78 (0.56-1.09) Race White | 236/330 | 239/321 | 0.73 (0.61-0.87) Race All others | 24/39 | 44/51 | 0.58 (0.35-0.96) Region North America | 44/64 | 39/63 | 0.79 (0.51-1.23) Region Western Europe | 157/221 | 162/216 | 0.70 (0.56-0.87) Region Rest of the world | 61/86 | 84/97 | 0.70 (0.50-0.98) ECOG PS 0 | 147/216 | 166/219 | 0.71 (0.56-0.88) ECOG PS 1 | 110/150 | 116/153 | 0.66 (0.51-0.86) IMDC risk category Favorable | 56/84 | 71/96 | 0.57 (0.40-0.82) IMDC risk category Intermediate | 153/227 | 160/214 | 0.70 (0.56-0.88) IMDC risk category Poor | 53/60 | 54/66 | 1.03 (0.70-1.52) Prior lines of therapy Adjuvant or one | 194/266 | 204/269 | 0.68 (0.55-0.83) Prior lines of therapy Two | 65/101 | 81/107 | 0.75 (0.54-1.04) Axis: 0.1 0.25 0.5 1 2 — Favors belzutifan + lenvatinib | Favors cabozantinib Data cutoff date: 2 Feb 2026

(A) PFS | Belzutifan + Lenvatinib (N = 371) | Cabozantinib (N = 376) Events, n (%) | 262 (70.6) | 285 (75.8) Censored, n (%) | 109 (29.4) | 91 (24.2) Median, mo (95% CI) | 14.8 (11.2–16.6) | 10.8 (9.2–11.1) HR (95% CI) | 0.74 (0.62–0.88) Landmarks printed: 24 mo: 35.2% vs 20.0%; 36 mo: 20.5% vs 12.6% No. at Risk (0-60 by 5): Belzutifan + lenvatinib 371 | 292 | 204 | 151 | 131 | 104 | 70 | 44 | 28 | 13 | 3 | 0 | 0 Cabozantinib 376 | 280 | 167 | 103 | 85 | 55 | 30 | 19 | 7 | 2 | 0 | 0 | 0 (B) OS | Events, n (%) | 195 (52.6) | 224 (59.6) Censored, n (%) | 176 (47.4) | 152 (40.4) Median, mo (95% CI) | 33.7 (29.0–46.9) | 28.6 (24.1–31.4) HR (95% CI) | 0.85 (0.70–1.03) P=0.04689 (1-sided boundary 0.0223) Landmarks printed: 36 mo: 48.9% vs 40.6%; 42 mo: 45.8% vs 37.0% No. at Risk: Belzutifan + lenvatinib 371 | 345 | 311 | 273 | 252 | 228 | 194 | 117 | 74 | 44 | 11 | 1 | 0 Cabozantinib 376 | 355 | 307 | 271 | 234 | 200 | 168 | 105 | 56 | 32 | 8 | 4 | 0 (C) DOR | Pts with response (censored), n | 197 (67) | 151 (43) Median DOR, mo (range) | 22.0 (2.0 to 52.2+) | 11.9 (1.8+ to 44.1+) Landmarks printed: 30 mo: 37.5% vs 22.6% No. at Risk: Belzutifan + lenvatinib 197 | 181 | 137 | 110 | 93 | 74 | 50 | 26 | 17 | 5 | 1 | 0 | 0 Cabozantinib 151 | 137 | 80 | 59 | 48 | 31 | 19 | 10 | 3 | 0 | 0 | 0 | 0 Figure 1 Kaplan–Meier estimates of (A) PFS, (B) OS, and (C) DOR in the intention-to-treat population CI, confidence interval; DOR, duration of response; HR, hazard ratio; OS, overall survival; PFS, progression-free survival. Survival table Endpoint: PFS (ITT) Arm | N | Median | Median CI Belzutifan + Lenvatinib | 371 | 14.8 | 11.2–16.6 Cabozantinib | 376 | 10.8 | 9.2–11.1 HR: 0.74 HR CI: 0.62–0.88 Landmark rates: 24 mo: Belzutifan + Lenvatinib 35.2% | Cabozantinib 20.0% 36 mo: Belzutifan + Lenvatinib 20.5% | Cabozantinib 12.6%

[Slide 1] Final Analysis of the Phase 3 LITESPARK-011 Study: Belzutifan Plus Lenvatinib Versus Cabozantinib for Advanced Renal Cell Carcinoma After Anti–PD-(L)1 Therapy Daniel Y.C. Heng1, Ray McDermott2, Se Hoon Park3, Manuela Schmidinger4, Mauricio Burotto5, Roberto Iacovelli6, Elena Verzoni7, Cristina Suárez8, Masatoshi Eto9, Hernan J. Cutuli10, Camillo Porta11, Ray Manneh Kopp12, Sylvie Rottey13, Guillermo de Velasco14, Begoña P. Valderrama15, Ugo De Giorgi16, Alessandro Vagheggini17, Rodolfo F. Perini18, Donna Vickery18, Robert J. Motzer19 1Arthur J. E. Child Comprehensive Cancer Centre, Calgary, AB, Canada; 2St. Vincent's University Hospital, Cancer Trials Ireland, Dublin, Ireland; 3Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Republic of Korea; 4Medical University of Vienna, Vienna, Austria; 5Bradford Hill Clinical Research Center, Universidad Finis Terrae Santiago, Chile; 6Fondazione Policlinico Universitario Agostino Gemelli, IRCCS, Rome, Italy; 7Fondazione IRCCS Istituto Nazionale dei Tumori di Milano, Genitourinary Medical Oncology, Milan, Italy; 8Vall d'Hebron Institute of Oncology (VHIO), Vall d'Hebron University Hospital, Vall d'Hebron Barcelona Hospital Campus, Barcelona, Spain; 9Kyushu University, Graduate School of Medical Sciences, Fukuoka, Japan; 10Hospital Sirio Libanés, Buenos Aires, Argentina; 11Azienda Ospedaliero-Universitaria Consorziale Policlinico di Bari and University of Bari "A. Moro," Bari, Italy; 12Sociedad de Oncología y Hematología del Cesar, Valledupar, Colombia; 13Ghent University Hospital, Ghent, Belgium; 14Department of Medical Oncology, University Hospital 12 de Octubre, Instituto de Investigación (imas12), Madrid, Spain; 15Hospital Universitario Virgen del Rocío, Seville, Spain; 16IRCCS Istituto Romagnolo per lo Studio dei Tumori (IRST) Dino Amadori, Meldola, Italy; 17MSD Innovation & Development GmbH, Zurich, Switzerland; 18Merck & Co., Inc., Rahway, NJ, USA; 19Memorial Sloan Kettering Cancer Center, New York, NY, USA [Slide 4] Summary and Conclusions • Belzutifan + lenvatinib demonstrated continued PFS and ORR benefit versus cabozantinib at the final analysis of the phase 3 LITESPARK-011 study • OS numerically favored belzutifan + lenvatinib at final analysis but was not statistically significant; PFS2 also favored the combination arm • Belzutifan + lenvatinib improved PFS, OS, ORR, and DOR versus cabozantinib across prior 1L IO + IO or IO + TKI subgroups in exploratory analyses • Safety was generally consistent with previous analyses • These results continue to support the use of belzutifan + lenvatinib in participants with advanced clear cell RCC following anti–PD-(L)1 therapy

ORR (Key Secondary) and DOR (Secondary Endpoint) per RECIST 1.1 by BICR Belzutifan + Lenvatinib: ORR 53.1% (95% CI 47.9–58.3); CR=6.7%; PR=46.4% Cabozantinib: ORR 40.2% (95% CI 35.2–45.3); CR=1.3%; PR=38.8% Estimated difference 12.9% (95% CI 5.8–19.8) Nominal P=0.0002a | Belzutifan + Lenvatinib (N = 371) | Cabozantinib (N = 376) Pts with response, n | 197 | 151 Median DOR, mo (range) | 22.0 (2.0 to 52.2+) | 11.9 (1.8+ to 44.1+) DOR curve annotations at 30 mo: 37.5% / 22.6% No. at Risk (0, 5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60 months) Belzutifan + lenvatinib 197 181 137 110 93 74 50 26 17 5 1 0 0 Cabozantinib 151 137 80 59 48 31 19 10 3 0 0 0 0 aKey secondary ORR endpoint was met at previous interim analysis and was not formally statistically tested at final analysis. Data cutoff date: 2 Feb 2026

[Slide 1] HIF2α inhibition and RCC in VHL disease: the proof of concept A Maximum Change in Target Renal Tumors; Best objective response (N=61): Confirmed partial response; Stable disease C Duration of Treatment: Confirmed partial response; Stable disease; Progressive disease; Death; Treatment ongoing D Tumor-Reduction Procedures: Procedure after trial enrollment; Procedure before trial enrollment; Patient had at least one procedure >10 years before trial enrollment No. of Procedures per Respective Year: 142 18 7 28 15 19 13 15 18 22 24 2 [partially legible] LITESPARK-004: 61 patients treated with belzutifan 120 mg QD 50 months update: RCC ORR 67%, including 11% CR [Slide 2] Belzutifan single-agent approval: LITESPARK-005 Primary Endpoint: PFS per RECIST 1.1 by BICR Kaplan-Meier Estimate of PFS at IA2 | IA1 Belzutifan | IA1 Everolimus | IA2 Belzutifan | IA2 Everolimus Events | 257 (68.7%) | 262 (70.4%) | 289 (77.3%) | 276 (74.2%) Median, mo (95% CI) | 5.6 (3.9–7.0) | 5.6 (4.8–5.8) | 5.6 (3.8–6.5) | 5.6 (4.8–5.8) HR (95% CI) | 0.75 (0.63–0.90); P <.001* | 0.74 (0.63–0.88) Landmarks printed: 12 mo: 33.7% vs 17.6%; 18 mo: 22.5% vs 9.0% No. at Risk (0-36 by 3): Belzutifan | 374 | 218 | 156 | 135 | 113 | 93 | 66 | 45 | 35 | 21 | 14 | 4 | 0 Everolimus | 372 | 226 | 113 | 70 | 41 | 26 | 19 | 10 | 5 | 2 | 2 | 0 | 0 ORR at IA2: Belzutifan 22.7% (CR=3.5%, PR=19.3%); Everolimus 3.5% (PR=3.5%) * denotes statistical significance. Primary PFS endpoint was met at IA1 and was not formally statistically tested at IA2. Data cutoff date for IA1: November 1, 2022. Data cutoff date for IA2: June 13, 2023. Survival table Endpoint: PFS per RECIST 1.1 by BICR (IA2) Arm | N | Median | Median CI Belzutifan | 374 | 5.6 | 3.8–6.5 Everolimus | 372 | 5.6 | 4.8–5.8 HR: 0.74 HR CI: 0.63–0.88 Landmark rates: 12 mo: Belzutifan 33.7% | Everolimus 17.6% 18 mo: Belzutifan 22.5% | Everolimus 9.0% [Slide 3] Belzutifan in combination ADJ (LS-022) | L2 | L3 (LS-011) DFS by Investigator, ITT Population HR 0.72 (95% CI, 0.59–0.87) P = 0.0003* Landmarks printed: 12 months 91.9% vs 86.2%; 24 months 80.7% vs 73.7%; 30 months 75.8% vs 68.6% Pembrolizumab + belzutifan (N = 921) | Pembrolizumab + placebo (N = 920) Events, n (%) | [illegible] | [illegible] Median, mo (95% CI) | NR ([illegible]) | NR ([illegible]) No. at Risk (0, 6, 12, 18, 24, 30, 36, 42): Pembrolizumab + belzutifan 921 | 860 | 800 | 669 | 401 | 153 | 33 | 0 Pembrolizumab + placebo 920 | 823 | 769 | 619 | 417 | 140 | 38 | 0 Median follow-up: 28.4 months (range, [illegible]). Data cutoff: [illegible] Primary Endpoint: PFS per RECIST 1.1 by BICR Belzutifan + Lenvatinib (N = 371) | Cabozantinib (N = 376) Events, n (%) | 232 (62.5) | 273 (72.6) Median, mo (95% CI) | 14.8 (11.2–16.6) | 10.7 (9.2–11.1) HR 0.70, 95% CI 0.59–0.84, 1-sided P =.00007* Landmarks printed: 12 months 55.0% vs 41.0%; 24 months 35.6% vs 19.1% No. at Risk (0-50 by 5): Belzutifan + Lenvatinib 371 | 292 | 204 | 151 | 117 | 61 | 39 | 19 | 8 | 1 | 0 Cabozantinib 376 | 279 | 166 | 102 | 78 | 33 | 10 | 2 | 0 | 0 | 0 * denotes statistical significance (1-sided boundary 0.0047). Data cutoff date IA2: [illegible] April 202[illegible] Survival table Endpoint: LITESPARK-011 PFS per RECIST 1.1 by BICR (IA2) Arm | N | Median | Median CI Belzutifan + Lenvatinib | 371 | 14.8 | 11.2–16.6 Cabozantinib | 376 | 10.7 | 9.2–11.1 HR: 0.70 HR CI: 0.59–0.84 P: 1-sided P=.00007 Landmark rates: 12 mo: Belzutifan + Lenvatinib 55.0% | Cabozantinib 41.0% 24 mo: Belzutifan + Lenvatinib 35.6% | Cabozantinib 19.1%

Final Analysis of the Phase 3 LITESPARK-011 Study: Belzutifan Plus Lenvatinib Versus Cabozantinib for Advanced Renal Cell Carcinoma After Anti–PD-(L)1 Therapy Daniel Y.C. Heng1, Ray McDermott2, Se Hoon Park3, Manuela Schmidinger4, Mauricio Burotto5, Roberto Iacovelli6, Elena Verzoni7, Cristina Suárez8, Masatoshi Eto9, Hernan J. Cutuli10, Camillo Porta11, Ray Manneh Kopp12, Sylvie Rottey13, Guillermo de Velasco14, Begoña P. Valderrama15, Ugo De Giorgi16, Alessandro Vagheggini17, Rodolfo F. Perini18, Donna Vickery18, Robert J. Motzer19

Final Analysis of the Phase 3 LITESPARK-011 Study: Belzutifan Plus Lenvatinib Versus Cabozantinib for Advanced Renal Cell Carcinoma After Anti–PD-(L)1 Therapy Daniel Y.C. Heng1, Ray McDermott2, Se Hoon Park3, Manuela Schmidinger4, Mauricio Burotto5, Roberto Iacovelli6, Elena Verzoni7, Cristina Suárez8, Masatoshi Eto9, Hernan J. Cutuli10, Camillo Porta11, Ray Manneh Kopp12, Sylvie Rottey13, Guillermo de Velasco14, Begoña P. Valderrama15, Ugo De Giorgi16, Alessandro Vagheggini17, Rodolfo F. Perini18, Donna Vickery18, Robert J. Motzer19 [Affiliation list in small print, partially legible: 1Arthur J. E. Child Comprehensive Cancer Centre, Calgary, AB, Canada; 2St. Vincent's University Hospital, Cancer Trials Ireland, Dublin, Ireland; 3Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Republic of Korea; 4Medical University of Vienna, Vienna, Austria; 5Bradford Hill Clinical Research Center, Universidad Finis Terrae Santiago, Chile; ... 18Merck & Co., Inc., Rahway, NJ, USA; 19Memorial Sloan Kettering Cancer Center, New York, NY, USA]

[Slide 1] LITESPARK-011 Study (NCT04586231) Key Eligibility Criteria • Unresectable, locally advanced or metastatic ccRCC • KPS score ≥70% • ≤2 prior systemic regimens • Progression on/after anti–PD-(L)1 mAb as 1L or 2L therapy or progression ≤6 months of the last dose of adjuvant anti–PD-(L)1 mAb • Prior VEGFR-TKI therapy was permitted R 1:1 N = 371: Belzutifan 120 mg orally QD + Lenvatinib 20 mg orally QD N = 376: Cabozantinib 60 mg orally QD Stratification Factors • IMDC prognostic score:a 0 vs. 1-2 vs. 3-6 • Line of treatment for prior anti–PD-(L)1: 1L, adjuvant, neoadjuvant-adjuvant vs. 2L • Geographic region: North America vs. western Europe vs. rest of world Dual Primary Endpoints: • PFS per RECIST 1.1 by BICR • OS Key Secondary Endpoint: • ORR per RECIST 1.1 by BICR Other Secondary Endpoints Include: • DOR per RECIST 1.1 by BICR • Safety Exploratory Analyses Include: • Subgroup efficacy analysis by type of IO combination in 1L (IO + IO vs. IO + TKI) • PFS2 [Slide 3] Most Common Treatment-Emergent AEs,a As-Treated Population Incidence ≥15% in Any Arm Any grade % | Belzutifan + Lenvatinib (N = 370) | Cabozantinib (N = 371) Diarrhea | 54.3% | 71.2% Hypertension | 59.5% | 57.1% Anemia | 70.8% | 27.0% Fatigue | 46.2% | 41.0% Decreased appetite | 40.5% | 40.7% Nausea | 41.9% | 38.5% PPES | 20.5% | 51.8% Hypothyroidism | 36.2% | 29.1% Weight decreased | 29.5% | 30.5% Proteinuria | 38.4% | 20.8% AST increased | 21.9% | 34.2% Constipation | 27.3% | 23.2% ALT increased | 19.2% | 31.3% Vomiting | 32.2% | 17.0% Asthenia | 28.1% | 20.2% Arthralgia | 30.8% | 17.3% Stomatitis | 14.3% | 28.3% Dysgeusia | 12.4% | 27.8% Headache | 23.0% | 12.4% Cough | 18.1% | 17.5% Back pain | 18.9% | 14.6% Blood creatinine increased | 15.4% | 15.6% Dysphonia | 14.1% | 17.3% Abdominal pain | 15.1% | 14.8% Dyspnea | 15.9% | 10.2% Pain in extremity | 15.9% | 8.9% AEs of Clinical Interest Hypoxia | 15.7% | 0% Cardiac dysfunctionb | 13.0% | 8.1% Any grade AEs / Grade ≥3 AEs (grade ≥3 shown as bars only, no values printed)

Primary Endpoint: PFS per RECIST 1.1 by BICR 24 mo 36 mo Belzutifan + Lenvatinib (N = 371) Cabozantinib (N = 376) Events, n (%) 262 (70.6) 285 (75.8) Median, mo (95% CI) 14.8 (11.2–16.6) 10.8 (9.2–11.1) HR (95% CI) 0.74 (0.62–0.88) Nominal P value (1-sided)a 0.00035 35.2% 20.0% 20.5% 12.6% PFS, % Months No. at Risk Belzutifan + lenvatinib Cabozantinib Survival table Endpoint: PFS per RECIST 1.1 by BICR Arm | N | Median | Median CI Belzutifan + lenvatinib | 371 | 14.8 mo | 11.2–16.6 Cabozantinib | 376 | 10.8 mo | 9.2–11.1 HR: 0.74 HR CI: 0.62–0.88 P: 0.00035 (nominal, 1-sided) 24 mo: 35.2% vs 20.0% 36 mo: 20.5% vs 12.6%

[Slide 3] OS in Key Subgroups Events/Participants: Belzutifan + lenvatinib | Cabozantinib | HR (95% CI) Overall | 195/371 | 224/376 | 0.85 (0.70-1.03) Age <65 years | 107/209 | 124/224 | 0.91 (0.70-1.17) Age ≥65 years | 88/162 | 100/152 | 0.73 (0.54-0.97) Male | 146/273 | 175/292 | 0.83 (0.67-1.04) Female | 49/98 | 49/84 | 0.81 (0.55-1.20) Race White | 180/330 | 193/321 | 0.85 (0.69-1.04) Race All others | 14/39 | 29/51 | 0.52 (0.27-1.01) North America | 41/64 | 32/63 | 1.39 (0.87-2.21) Western Europe | 114/221 | 136/216 | 0.70 (0.55-0.90) Rest of the world | 40/86 | 56/97 | 0.80 (0.53-1.20) ECOG PS 0 | 102/216 | 107/219 | 0.94 (0.72-1.23) ECOG PS 1 | 88/150 | 113/153 | 0.68 (0.52-0.90) IMDC Favorable | 23/84 | 30/96 | 0.82 (0.48-1.41) IMDC Intermediate | 120/227 | 134/214 | 0.78 (0.61-0.99) IMDC Poor | 52/60 | 60/66 | 0.93 (0.64-1.35) Prior lines: Adjuvant or one | 141/266 | 161/269 | 0.82 (0.66-1.03) Prior lines: Two | 52/101 | 63/107 | 0.82 (0.57-1.18) Favors belzutifan + lenvatinib | Favors cabozantinib Objective Response and Duration of Response [Slide 4] ORR (Key Secondary) and DOR (Secondary Endpoint) per RECIST 1.1 by BICR Belzutifan + Lenvatinib: ORR 53.1% (95% CI 47.9–58.3); CR=6.7%; PR=46.4% Cabozantinib: ORR 40.2% (95% CI 35.2–45.3); CR=1.3%; PR=38.8% Estimated difference 12.9% (95% CI 5.8–19.8); Nominal P=0.0002a | Belzutifan + Lenvatinib (N = 371) | Cabozantinib (N = 376) Pts with response, n | 197 | 151 Median DOR, mo (range) | 22.0 (2.0 to 52.2+) | 11.9 (1.8+ to 44.1+) DOR at 30 mo: 37.5% vs 22.6% No. at Risk (0–60 mo by 5) Belzutifan + lenvatinib | 197 181 137 110 93 74 50 26 17 5 1 0 0 Cabozantinib | 151 137 80 59 48 31 19 10 3 0 0 0 0

[Slide 3] OS in Key Subgroups (Left-panel row labels partially cropped in photo; visible fragments: "verall", "e", "65 years", "65 years", "le", "nale", "ers", "merica", "Europe", "e world") Row (visible label fragment) | Events/Participants Belzutifan + lenvatinib | Cabozantinib | HR (95% CI) [O]verall | 195/371 | 224/376 | 0.85 (0.70-1.03) [<]65 years | 107/209 | 124/224 | 0.91 (0.70-1.17) [≥]65 years | 88/162 | 100/152 | 0.73 (0.54-0.97) [Ma]le | 146/273 | 175/292 | 0.83 (0.67-1.04) [Fe]male | 49/98 | 49/84 | 0.81 (0.55-1.20) [label cropped] | 180/330 | 193/321 | 0.85 (0.69-1.04) [All oth]ers | 14/39 | 29/51 | 0.52 (0.27-1.01) [North A]merica | 41/64 | 32/63 | 1.39 (0.87-2.21) [Western] Europe | 114/221 | 136/216 | 0.70 (0.55-0.90) [Rest of th]e world | 40/86 | 56/97 | 0.80 (0.53-1.20) ECOG PS 0 | 102/216 | 107/219 | 0.94 (0.72-1.23) ECOG PS 1 | 88/150 | 113/153 | 0.68 (0.52-0.90) IMDC risk category Favorable | 23/84 | 30/96 | 0.82 (0.48-1.41) IMDC risk category Intermediate | 120/227 | 134/214 | 0.78 (0.61-0.99) IMDC risk category Poor | 52/60 | 60/66 | 0.93 (0.64-1.35) Prior lines of therapy Adjuvant or one | 141/266 | 161/269 | 0.82 (0.66-1.03) Prior lines of therapy Two | 52/101 | 63/107 | 0.82 (0.57-1.18) Favors belzutifan + lenvatinib | Favors cabozantinib Data cutoff date: 2 Feb 2026

[Slide 1] LITESPARK-012 Study Key Eligibility Criteria • Adults with advanced ccRCC (stage IV per AJCC) • KPS score ≥70% • Measurable disease per RECIST 1.1 • No prior systemic therapy for advanced RCC Stratification Factors • IMDC prognostic score (0 [favorable] vs. 1-2 [intermediate] vs. 3-6 [poor]) • Geographic region (North America vs. western Europe vs. rest of world) • Sarcomatoid features (yes vs. no) N = 1688 R N = 560: Pembrolizumab 400 mg IV Q6W up to 18 cycles + Lenvatinib 20 mg orally QD + Belzutifan 120 mg orally QD N = 562: Quavonlimab/pembrolizumab 25 mg/400 mg (MK-1308A) IV Q6W up to 18 cycles + Lenvatinib 20 mg orally QD N = 566: CONTROL Pembrolizumab 400 mg IV Q6W up to 18 cycles + Lenvatinib 20 mg orally QD Dual primary endpoints: Progression-free survival (PFS) by blinded independent central review (BICR) per RECIST 1.1 and overall survival (OS) Secondary endpoints: Objective response rate (ORR) and duration of response (DOR) by BICR per RECIST 1.1, safety IMDC, International Metastatic Renal-Cell Carcinoma Database Consortium; IV, intravenous; KPS, Karnofsky Performance Status; Q6W, every 6 weeks; QD, daily. [Slide 2] PFS by BICR, Pembro + Len + Belz vs. Pembro + Len ITT Population | Pembro + len + belz N = 560 | Pembro + len N = 566 Events, n (%) | 337 (60.2) | 345 (61.0) Median, mo (95% CI) | 20.8 (17.9–23.7) | 23.7 (20.7–26.6) HR 1.08 (95% CI 0.93–1.26); 1-sided P=0.8408 Landmarks (12/24/36 mo): Pembro + len 69.1% / 49.6% / 36.7%; Pembro + len + belz 66.3% / 45.3% / 34.7% No. at Risk Pembro + len + belz 560 | 424 | 329 | 257 | 210 | 176 | 122 | 78 | 23 | 2 | 0 Pembro + len 566 | 455 | 365 | 298 | 243 | 199 | 129 | 75 | 24 | 1 | 0 Data cutoff date: 4 Feb 2026 Survival table Endpoint: PFS by BICR (ITT) Arm | N | Median | Median CI Pembro + len + belz | 560 | 20.8 mo | 17.9–23.7 Pembro + len | 566 | 23.7 mo | 20.7–26.6 HR: 1.08 HR CI: 0.93–1.26 P: 1-sided P=0.8408 Landmark rates: 12 mo: Pembro + len 69.1% | Pembro + len + belz 66.3% 24 mo: Pembro + len 49.6% | Pembro + len + belz 45.3% 36 mo: Pembro + len 36.7% | Pembro + len + belz 34.7% [Slide 3] PFS by BICR, Qmab/Pembro + Len vs. Pembro + Len IMDC Intermediate/Poor Risk Population (Per Sequential Testing Strategy) | Qmab/pembro + len N = 429 | Pembro + len N = 429 Events, n (%) | 259 (60.4) | 261 (60.8) Median, mo (95% CI) | 20.5 (16.6–23.5) | 23.5 (20.6–26.3) HR 1.12 (95% CI 0.94–1.33); 1-sided P=0.8954 Landmarks (12/24/36 mo): Pembro + len 67.5% / 47.6% / 34.8%; Qmab/pembro + len 62.3% / 43.7% / 31.2% No. at Risk Qmab/pembro + len 429 | 310 | 230 | 185 | 147 | 124 | 79 | 49 | 17 | 3 | 0 Pembro + len 429 | 337 | 267 | 219 | 176 | 148 | 84 | 48 | 17 | 1 | 0 Data cutoff date: 4 Feb 2026 Survival table Endpoint: PFS by BICR (IMDC intermediate/poor risk, per sequential testing strategy) Arm | N | Median | Median CI Qmab/pembro + len | 429 | 20.5 mo | 16.6–23.5 Pembro + len | 429 | 23.5 mo | 20.6–26.3 HR: 1.12 HR CI: 0.94–1.33 P: 1-sided P=0.8954 Landmark rates: 12 mo: Pembro + len 67.5% | Qmab/pembro + len 62.3% 24 mo: Pembro + len 47.6% | Qmab/pembro + len 43.7% 36 mo: Pembro + len 34.8% | Qmab/pembro + len 31.2% [Slide 4] Summary of All-Cause AEs As-Treated Population Pts with ≥1 AE, n (%) | Pembro + len + belz N = 556 | Qmab/pembro + len N = 558 | Pembro + len N = 564 Any grade | 554 (99.6) | 557 (99.8) | 562 (99.6) Grade ≥3 | 485 (87.2) | 495 (88.7) | 478 (84.8) Led to discontinuation of any drug | 287 (51.6) | 287 (51.4)(a) | 204 (36.2) Of belz | 166 (29.9) | – | – Of pembro | 181 (32.6) | – | 111 (19.7) Of qmab/pembro | – | 198 (35.5) | – Of len | 176 (31.7) | 208 (37.3) | 158 (28.0) Led to death | 42 (7.6) | 53 (9.5) | 44 (7.8) Any grade treatment-related | 545 (98.0) | 549 (98.4) | 557 (98.8) Grade ≥3 treatment-related | 431 (77.5) | 437 (78.3) | 390 (69.1) Grade 3-5 All-Cause AEs in ≥10.0% of Pts in Any Arm (Incidence, %; Pembro + len + belz / Qmab/pembro + len / Pembro + len) Hypertension | 25.0 | 28.7 | 30.7 Anemia | 23.4 | 4.1 | 3.7 Hypoxia | 13.5 | 0.2 | 0.4 ALT increased | 11.2 | 8.6 | 5.5 Lipase increased | 8.8 | 13.6 | 6.9 Proteinuria | 8.6 | 9.9 | 11.9 Diarrhea | 6.8 | 10.0 | 8.3 Weight decreased | 6.5 | 9.1 | 10.3 ALT, alanine aminotransferase. a6 pts who received pembro after qmab/pembro discontinued pembro due to an all-cause AE. Data cutoff date: 4 Feb 2026

[Slide 3] OS, Pembro + Len + Belz vs. Pembro + Len ITT Population | Pembro + len + belz N = 560 | Pembro + len N = 566 Events, n (%) | 229 (40.9) | 227 (40.1) Median, mo (95% CI) | 52.0 (48.0–NR) | NR (48.0–NR) HR 1.04 (95% CI 0.87–1.25); 1-sided P=0.6622 Landmarks: 89.9% / 88.7%; 78.2% / 76.4%; 66.3% / 64.5% No. at Risk Pembro + len + belz 560 | 514 | 491 | 456 | 420 | 386 | 298 | 198 | 73 | 10 | 0 Pembro + len 566 | 536 | 508 | 476 | 439 | 399 | 307 | 185 | 73 | 13 | 0 NR, not reached. Data cutoff date: 4 Feb 2026 Survival table Endpoint: OS (ITT) Arm | N | Median | Median CI Pembro + len + belz | 560 | 52.0 mo | 48.0–NR Pembro + len | 566 | NR | 48.0–NR HR: 1.04 HR CI: 0.87–1.25 P: 1-sided P=0.6622 Landmark rates: 12 mo (dashed line): Pembro + len 89.9% | Pembro + len + belz 88.7% 24 mo (dashed line): Pembro + len 78.2% | Pembro + len + belz 76.4% 36 mo (dashed line): Pembro + len 66.3% | Pembro + len + belz 64.5% [Slide 4] Drug Exposure As-Treated Population(a) Data are median (range) except where specified otherwise. | Pembro + len + belz N = 556 | Qmab/pembro + len N = 558 | Pembro + len N = 564 Pembrolizumab, n | 556 | – | 564 Duration of therapy, months | 14.1 (0.03–31.8) | – | 21.2 (0.03–29.5) Cycles, n | 11.0 (1.0–18.0) | – | 16.0 (1.0–18.0) Quavonlimab/pembrolizumab, n | – | 558(c) | – Duration of therapy, months | – | 11.8 (0.03–37.2) | – Cycles, n | – | 9.0 (1.0–18.0) | – Lenvatinib, n | 556 | 557 | 564 Duration of therapy, months | 18.1 (0.03–56.6) | 15.2 (0.1–56.8) | 22.2 (0.1–57.3) Total (range) dose administered, mg | 6423.0 (20.0–32620.0) | 5470.0 (40.0–32840.0) | 7676.0 (40.0–30492.0) Dose intensity, mg/day(b) | 12.6 (2.1–20.0) | 12.0 (1.6–20.0) | 12.8 (2.4–20.0) Belzutifan, n | 556 | – | – Duration of therapy, months | 18.6 (0.03–56.6) | – | – Total (range) dose administered, mg | 50100.0 (120.0–197280.0) | – | – Dose intensity, mg/day(b) | 106.8 (16.6–120.0) | – | – aDefined as all pts who received ≥1 dose of study treatment. bCalculated as total dose administered divided by duration on therapy. cPts who discontinued qmab/pembro (MK-1308A) before completing 2 years or 18 cycles due to an AE could continue with pembro alone. 21 pts in the qmab/pembro + len arm received pembro alone after qmab/pembro for a median duration of 4.4 months (range, 0.03–25.6) and median 4.0 cycles (range, 1.0–16.0). Data cutoff date: 4 Feb 2026

[Slide 1] ORIGINAL ARTICLE · Articles in Press, October 25, 2026 Download Full Issue Pembrolizumab plus lenvatinib with belzutifan or quavonlimab as first-line therapy for advanced renal-cell carcinoma: randomized open-label phase 3 LITESPARK-012 study Brian I. Rini 1 · Toni K. Choueiri 2 · Martin H. Voss 3 · ... · Joseph E. Burgents 19 · Karla G. Rodriguez-Lopez 19 · Thomas B. Powles 20 ... Show more Affiliations & Notes Article Info [Slide 2] (A) Arm A | Control Events, n (%) | 337 (60.2) | 345 (61.0) Median, months (95% CI) | 20.8 (17.9–23.7) | 23.7 (20.7–26.6) HR 1.08 (95% CI 0.93–1.26); P=0.8408 Landmarks printed: Control 69.1%, 49.6%, 36.7%; Arm A 66.3%, 45.3%, 34.7% (dashed lines at 12, 24, 36 months) No. at Risk (No. Censored), months 0-60 by 6: Arm A | 560 (0) | 424 (39) | 329 (55) | 257 (66) | 210 (73) | 176 (81) | 122 (115) | 78 (152) | 23 (200) | 2 (221) | 0 (223) Control | 566 (0) | 455 (21) | 365 (33) | 298 (46) | 243 (56) | 199 (64) | 129 (112) | 75 (153) | 24 (199) | 1 (220) | 0 (221) (B) Arm B | Control Events, n (%) | 259 (60.4) | 261 (60.8) Median, months (95% CI) | 20.5 (16.6–23.5) | 23.5 (20.6–26.3) HR 1.12 (95% CI 0.94–1.33); P=0.8954 Landmarks printed: Control 67.5%, 47.6%, 34.8%; Arm B 62.3%, 43.7%, 31.2% Arm B | 429 (0) | 310 (36) | 230 (50) | 185 (59) | 147 (67) | 124 (72) | 79 (97) | 49 (122) | 17 (153) | 3 (167) | 0 (170) Control | 429 (0) | 337 (19) | 267 (29) | 219 (37) | 176 (44) | 148 (47) | 84 (94) | 48 (121) | 17 (151) | 1 (167) | 0 (168) (C) Arm B | Control Events, n (%) | 334 (59.4) | 345 (61.0) Median, months (95% CI) | 20.7 (18.1–26.0) | 23.7 (20.7–26.6) HR 1.07 (95% CI 0.92–1.25); nominal P=0.8174 Landmarks printed: Control 69.1%, 49.6%, 36.7%; Arm B 66.2%, 46.0%, 33.1% Arm B | 562 (0) | 425 (46) | 325 (62) | 262 (73) | 209 (82) | 174 (89) | 119 (118) | 77 (154) | 26 (202) | 4 (224) | 0 (228) Control | 566 (0) | 455 (21) | 365 (33) | 298 (46) | 243 (56) | 199 (64) | 129 (112) | 75 (153) | 24 (199) | 1 (220) | 0 (221) Figure 2 Dual primary endpoint of progression-free survival. (A) Kaplan-Meier analysis of progression-free survival in Arm A vs Control in all randomized [caption cut off] Survival table Endpoint: PFS, Arm A vs Control (panel A, all randomized) Arm | N | Median | Median CI Arm A | 560 | 20.8 | 17.9–23.7 Control | 566 | 23.7 | 20.7–26.6 HR: 1.08 HR CI: 0.93–1.26 P: P=0.8408 Landmark rates: 12 mo: Arm A 66.3% | Control 69.1% 24 mo: Arm A 45.3% | Control 49.6% 36 mo: Arm A 34.7% | Control 36.7% [Slide 3] (A) Arm A | Control Events, n (%) | 229 (40.9) | 227 (40.1) Median, months (95% CI) | 52.0 (48.0–NR) | NR (48.0–NR) HR 1.04 (95% CI 0.87–1.25); P=0.6622 Landmarks printed: Control 89.9%, 78.2%, 66.3%; Arm A 88.7%, 76.4%, 64.5% No. at Risk (No. Censored), months 0-60 by 6: Arm A | 560 (0) | 514 (5) | 491 (6) | 456 (8) | 420 (9) | 386 (12) | 298 (68) | 198 (150) | 73 (264) | 10 (321) | 0 (331) Control | 566 (0) | 536 (0) | 508 (1) | 476 (3) | 439 (4) | 399 (5) | 307 (71) | 185 (168) | 73 (268) | 13 (326) | 0 (339) (B) Arm B | Control Events, n (%) | 214 (49.9) | 191 (44.5) Median, months (95% CI) | 39.1 (36.3–NR) | 46.2 (40.7–NR) HR 1.17 (95% CI 0.96–1.42); P=0.9419 Landmarks printed: Control 87.4%, 73.1%, 60.8%; Arm B 84.1%, 67.3%, 55.3% Arm B | 429 (0) | 391 (1) | 360 (1) | 320 (2) | 287 (2) | 268 (3) | 183 (58) | 115 (106) | 37 (179) | 11 (204) | 0 (215) Control | 429 (0) | 400 (0) | 374 (1) | 345 (3) | 310 (4) | 278 (4) | 194 (70) | 109 (136) | 50 (189) | 9 (229) | 0 (238) (C) Arm B | Control Events, n (%) | 245 (43.6) | 227 (40.1) Median, months (95% CI) | NR (45.2–NR) | NR (48.0–NR) HR 1.12 (95% CI 0.94–1.35); nominal P=0.8955 Landmarks printed: Control 89.9%, 78.2%, 66.3%; Arm B 87.5%, 72.7%, 61.7% Arm B | 562 (0) | 524 (1) | 491 (1) | 448 (2) | 407 (2) | 381 (3) | 292 (58) | 197 (130) | 71 (248) | 17 (300) | 0 (317) Control | 566 (0) | 536 (0) | 508 (1) | 476 (3) | 439 (4) | 399 (5) | 307 (71) | 185 (168) | 73 (268) | 13 (326) | 0 (339) Figure 3 Dual primary endpoint of overall survival. (A) Kaplan-Meier analysis of overall survival in Arm A vs Control in all randomized participants. (B) Kaplan-Meier analysis of overall survival in Arm B vs Control in all randomized [caption cut off] Survival table Endpoint: OS, Arm A vs Control (panel A, all randomized) Arm | N | Median | Median CI Arm A | 560 | 52.0 | 48.0–NR Control | 566 | NR | 48.0–NR HR: 1.04 HR CI: 0.87–1.25 P: P=0.6622 Landmark rates: 12 mo: Arm A 88.7% | Control 89.9% 24 mo: Arm A 76.4% | Control 78.2% 36 mo: Arm A 64.5% | Control 66.3%

[Slide 3] OS, Qmab/Pembro + Len vs Pembro + Len IMDC Intermediate/Poor Risk Population (Per Sequential Testing Strategy) | Qmab/pembro + len N = 429 | Pembro + len N = 429 Events, n (%) | 214 (49.9) | 191 (44.5) Median, mo (95% CI) | 39.1 (36.3–NR) | 46.2 (40.7–NR) HR 1.17 (95% CI 0.96–1.42); 1-sided P=0.9419 Landmarks: 12 mo: 87.4% (pembro + len), 84.1% (Qmab/pembro + len); 24 mo: 73.1% (pembro + len), 67.3% (Qmab/pembro + len); 36 mo: 60.8% (pembro + len), 55.3% (Qmab/pembro + len) No. at Risk (Months 0, 6, 12, 18, 24, 30, 36, 42, 48, 54, 60) Qmab/pembro + len | 429 | 391 | 360 | 320 | 287 | 268 | 183 | 115 | 37 | 11 | 0 Pembro + len | 429 | 400 | 374 | 345 | 310 | 278 | 194 | 109 | 50 | 9 | 0 Data cutoff date: 4 Feb 2026 Survival table Endpoint: OS, IMDC intermediate/poor risk Arm | N | Median | Median CI Qmab/pembro + len | 429 | 39.1 mo | 36.3–NR Pembro + len | 429 | 46.2 mo | 40.7–NR HR: 1.17 HR CI: 0.96–1.42 P: 1-sided P=0.9419 Landmark rates: 12 mo: Pembro + len 87.4% | Qmab/pembro + len 84.1% 24 mo: Pembro + len 73.1% | Qmab/pembro + len 67.3% 36 mo: Pembro + len 60.8% | Qmab/pembro + len 55.3% [Slide 4] ORR by BICR per RECIST 1.1 ITT Population | All participants: Pembro + len + belz N = 560 | Qmab/pembro + len N = 562 | Pembro + len N = 566 | IMDC intermediate/poor risk: Qmab/pembro + len n = 429 | Pembro + len n = 429 ORR, % (95% CI) | 69.1 (65.1–72.9) | 69.0 (65.0–72.8) | 76.3 (72.6–79.8) | 66.9 (62.2–71.3) | 75.5 (71.2–79.5) Confirmed best objective response, n (%) Complete response | 107 (19.1) | 83 (14.8) | 84 (14.8) | 59 (13.8) | 56 (13.1) Partial response | 280 (50.0) | 305 (54.3) | 348 (61.5) | 228 (53.1) | 268 (62.5) Stable disease^a | 109 (19.5) | 107 (19.0) | 86 (15.2) | 83 (19.3) | 62 (14.5) Progressive disease | 28 (5.0) | 27 (4.8) | 19 (3.4) | 24 (5.6) | 18 (4.2) NED,^b NE,^c or no assessment^d | 36 (6.4) | 40 (7.1) | 29 (5.1) | 35 (8.2) | 25 (5.8) Median DOR, months (range) | 26.4 (1.2+–52.5+) | 26.3 (1.4+–51.9+) | 27.7 (1.1+–51.3+) | 25.1 (1.4+–51.9+) | 27.2 (1.1+–51.3+) Bar chart ORR, % (95% CI): Pembro + len + belz ITT N=560: 69.1 (CR 19.1); Qmab/pembro + len ITT N=562: 69.0 (CR 14.8); Pembro + len ITT N=566: 76.3 (CR 14.8); Qmab/pembro + len IMDC int/p n=429: 66.9 (CR 13.8); Pembro + len IMDC int/p n=429: 75.5 (CR 13.1) ^a Includes both stable disease and non-complete response/non-progressive disease. ^b No evidence of disease. ^c Not evaluable: includes pts with insufficient data for assessment of response per RECIST 1.1. ^d No post-baseline assessment on the data cutoff date. CR, complete response; int/p, intermediate/poor. Data cutoff date: 4 Feb 2026

Annals of Oncology Available online 25 October 2026 In Press, Journal Pre-proof Original Article Pembrolizumab plus lenvatinib with belzutifan or quavonlimab as first-line therapy for advanced renal-cell carcinoma: randomized open-label phase 3 LITESPARK-012 study Brian I. Rini 1 *, Toni K. Choueiri 2, Martin H. Voss 3, Mauricio Burotto 4, Jan Oldenburg 5, Howard P. Gurney 6, Thomas R. Ferguson 7, Hans Hammers 8, Laurence Albiges 9, Yüksel Ürün 10, Roberto I. Romero Diaz 11, Mustafa Özgüroğlu 12, Patricio Yañez Weber 13, Elena Verzoni 14, Camillo Porta 15, Alexandra Drakaki 16, Jun Guo 17, Nobuaki Matsubara 18, Aiying Chen 19, Joseph E. Burgents 19...Thomas B. Powles 20

First-line Pembrolizumab Plus Lenvatinib Plus Belzutifan or MK-1308A Plus Lenvatinib in Advanced RCC: Randomized Open-label Phase 3 LITESPARK-012 Study Toni K. Choueiri1; Thomas B. Powles2; Martin H. Voss3; Mauricio Burotto4; Jan Oldenburg5; Howard P. Gurney6; Thomas R. Ferguson7; Hans Hammers8; Laurence Albiges9; Yüksel Ürün10; Roberto I. Romero Diaz11; Mustafa Özgüroğlu12; Patricio Yañez Weber13; Elena Verzoni14; Camillo Porta15; Alexandra Drakaki16; Aiying Chen17; Joseph E. Burgents17; Karla G. Rodriguez-Lopez17; Brian I. Rini18 1Dana-Farber Cancer Institute and Harvard Medical School, Boston, MA, USA; 2Barts Health NHS Trust, Barts Cancer Institute, and Queen Mary University of London, London, United Kingdom; 3Memorial Sloan Kettering Cancer Center, New York, NY, USA; 4Bradford Hill Clinical Research Center, Santiago, Chile; 5Akershus universitetssykehus, Lorenskog, Akershus, Norway; 6Maquarie University, Sydney, NSW, Australia; 7Fiona Stanley Hospital, Perth, WA, Australia; 8University of Texas Southwestern, Dallas, TX, USA; 9Département de Médecine Oncologique, Gustave Roussy, Université Paris Saclay, Villejuif, France; 10Ankara Universitesi Tip Fakultesi, Ankara, Türkiye; 11Centro de Investigacion Clinica de Oaxaca, Oaxaca City, Mexico; 12Istanbul University-Cerrahpaşa, Cerrahpaşa Faculty of Medicine, Istanbul, Türkiye; 13James Lind Cancer Research Center, Temuco, Chile; 14Genitourinary Medical Oncology, Fondazione IRCCS Istituto Nazionale dei Tumori di Milano, Milan, Italy; 15Azienda Ospedaliero-Universitaria Consorziale Policlinico di Bari and University of Bari "A. Moro," Bari, Italy; 16UCLA Hematology Oncology Santa Monica, Santa Monica, CA, USA; 17Merck & Co., Inc., Rahway, NJ, USA; 18Vanderbilt Ingram Cancer Center, Nashville, TN, USA

PFS by BICR, Pembro + Len + Belz vs. Pembro + Len ITT Population | Pembro + len + belz N = 560 | Pembro + len N = 566 Events, n (%) | 337 (60.2) | 345 (61.0) Median, mo (95% CI) | 20.8 (17.9–23.7) | 23.7 (20.7–26.6) HR 1.08 (95% CI 0.93–1.26); 1-sided P=0.8408 Landmarks printed: 12 mo 69.1% / 66.3%; 24 mo 49.6% / 45.3%; 36 mo 36.7% / 34.7% No. at Risk (months 0/6/12/18/24/30/36/42/48/54/60): Pembro + len + belz: 560 | 424 | 329 | 257 | 210 | 176 | 122 | 78 | 23 | 2 | 0 Pembro + len: 566 | 455 | 365 | 298 | 243 | 199 | 129 | 75 | 24 | 1 | 0 Data cutoff date: 4 Feb 2026 Toni Choueiri Survival table Endpoint: PFS by BICR (dual primary), ITT Arm | N | Median | Median CI Pembro + len + belz | 560 | 20.8 mo | 17.9–23.7 Pembro + len | 566 | 23.7 mo | 20.7–26.6 HR: 1.08 HR CI: 0.93–1.26 P: 1-sided P=0.8408 Landmark rates: 12 mo: Pembro + len (teal) 69.1% | Pembro + len + belz (blue) 66.3% 24 mo: Pembro + len (teal) 49.6% | Pembro + len + belz (blue) 45.3% 36 mo: Pembro + len (teal) 36.7% | Pembro + len + belz (blue) 34.7%

Summary and Conclusions In the randomized phase 3 LITESPARK-012 study: • Neither addition of the HIF-2α inhibitor belzutifan nor the CTLA-4 inhibitor quavonlimab to pembrolizumab + lenvatinib improved efficacy in participants with previously untreated advanced ccRCC • Adverse events in the investigational arms led to higher drug discontinuation rates • Addition of belzutifan to pembrolizumab + lenvatinib was generally associated with anticipated adverse events consistent with belzutifan's known safety profile • No new safety signals were observed for coformulated quavonlimab/pembrolizumab • Efficacy and safety of standard-of-care pembrolizumab + lenvatinib were consistent with prior reports, and this regimen was reinforced as a standard of care in front-line advanced ccRCC

[Slide 2] Baseline Characteristics ITT Population Characteristic, n (%) | Pembro + len + belz N = 560 | Qmab/pembro + len N = 562 | Pembro + len N = 566 Median age (range), years | 61.0 (27–90) | 61.0 (22–86) | 61.0 (22–84) Male | 423 (75.5) | 436 (77.6) | 393 (69.4) KPS score 90 or 100 | 439 (78.4) | 435 (77.4) | 458 (80.9) 70 or 80 | 121 (21.6) | 127 (22.6) | 108 (19.1) Region North America | 47 (8.4) | 46 (8.2) | 46 (8.1) Western Europe | 172 (30.7) | 169 (30.1) | 184 (32.5) Rest of world | 341 (60.9) | 347 (61.7) | 336 (59.4) IMDC risk categories Favorable | 145 (25.9) | 133 (23.7) | 137 (24.2) Intermediate | 329 (58.8) | 332 (59.1) | 344 (60.8) Poor | 86 (15.4) | 97 (17.3) | 85 (15.0) Prior nephrectomy | 363 (64.8) | 375 (66.7) | 352 (62.2) Sarcomatoid features Yes | 57 (10.2) | 67 (11.9) | 54 (9.5) No | 407 (72.7) | 412 (73.3) | 438 (77.4) Unknown | 96 (17.1) | 83 (14.8) | 74 (13.1) Data cutoff date: 4 Feb 2026 [Slide 4] OS, Pembro + Len + Belz vs. Pembro + Len Key Subgroups Subgroup | Events/Participants Pembro + len + belz | Events/Participants Pembro + len | HR (95% CI) Overall | 229/560 | 227/566 | 1.04 (0.87-1.25) Age <65 years | 130/344 | 126/348 | 1.04 (0.81-1.33) Age ≥65 years | 99/216 | 101/218 | 1.04 (0.79-1.38) Sex Male | 171/423 | 148/393 | 1.12 (0.90-1.39) Sex Female | 58/137 | 79/173 | 0.90 (0.64-1.27) Race White | 165/397 | 162/402 | 1.03 (0.83-1.29) Race All others | 64/163 | 65/162 | 1.06 (0.75-1.49) Region North America | 13/47 | 17/46 | 0.72 (0.35-1.49) Region Western Europe | 63/172 | 71/184 | 0.92 (0.65-1.29) Region Rest of world | 153/341 | 139/336 | 1.15 (0.91-1.44) Sarcomatoid features Yes | 31/57 | 23/54 | 1.29 (0.75-2.22) Sarcomatoid features No | 198/503 | 204/512 | 1.01 (0.83-1.23) IMDC prognostic risk Favorable | 36/145 | 36/137 | 0.95 (0.60-1.52) IMDC prognostic risk Intermediate | 139/329 | 140/344 | 1.07 (0.85-1.36) IMDC prognostic risk Poor | 54/86 | 51/85 | 1.03 (0.70-1.51) Axis: 0.50 0.75 1.00 1.50 2.00 3.00 Favors pembro + len + belz | Favors pembro + len Data cutoff date: 4 Feb 2026

PFS by BICR, Pembro + Len + Belz vs. Pembro + Len Key Subgroups Events/Participants: Pembro + len + belz | Pembro + len | HR (95% CI) Overall | 337/560 | 345/566 | 1.08 (0.93-1.26) Age <65 years | 210/344 | 201/348 | 1.11 (0.92-1.35) Age ≥65 years | 127/216 | 144/218 | 1.03 (0.81-1.30) Sex Male | 249/423 | 229/393 | 1.14 (0.95-1.36) Sex Female | 88/137 | 116/173 | 0.97 (0.74-1.29) Race White | 254/397 | 245/402 | 1.15 (0.97-1.37) Race All others | 83/163 | 99/162 | 0.88 (0.66-1.18) Region North America | 24/47 | 26/46 | 1.12 (0.64-1.96) Region Western Europe | 110/172 | 110/184 | 1.14 (0.88-1.49) Region Rest of world | 203/341 | 209/336 | 1.04 (0.86-1.26) Sarcomatoid features Yes | 37/57 | 35/54 | 1.04 (0.66-1.66) Sarcomatoid features No | 300/503 | 310/512 | 1.07 (0.92-1.26) IMDC prognostic risk Favorable | 76/145 | 84/137 | 0.86 (0.63-1.18) IMDC prognostic risk Intermediate | 199/329 | 201/344 | 1.19 (0.98-1.44) IMDC prognostic risk Poor | 62/86 | 60/85 | 1.04 (0.73-1.49) Favors pembro + len + belz <- -> Favors pembro + len Data cutoff date: 4 Feb 2026 Pembrolizumab + Lenvatinib + Belzutifan: Overall Survival

Arm A | Control Events, n (%) | 337 (60.2) | 345 (61.0) Median, months (95% CI) | 20.8 (17.9-23.7) | 23.7 (20.7-26.6) HR 1.08 (95% CI 0.93-1.26); P=0.8408 Landmarks: 12 mo Control 69.1%, Arm A 66.3%; 24 mo Control 49.6%, Arm A 45.3%; 36 mo Control 36.7%, Arm A 34.7% X-axis: Months; Y-axis: PFS, % No. at Risk (No. Censored) (months 0 | 6 | 12 | 18 | 24 | 30 | 36 | 42 | 48 | 54 | 60) Arm A: 560 (0) | 424 (39) | 329 (55) | 257 (66) | 210 (73) | 176 (81) | 122 (115) | 78 (152) | 23 (200) | 2 (221) | 0 (223) Control: 566 (0) | 455 (21) | 365 (33) | 298 (46) | 243 (56) | 199 (64) | 129 (112) | 75 (153) | 24 (199) | 1 (220) | 0 (221) Survival table Endpoint: Progression-free survival Arm | N | Median | Median CI Arm A | 560 | 20.8 | 17.9-23.7 Control | 566 | 23.7 | 20.7-26.6 HR: 1.08 HR CI: 0.93-1.26 P: 0.8408 Landmark rates: 12 months: Arm A 66.3% | Control 69.1% 24 months: Arm A 45.3% | Control 49.6% 36 months: Arm A 34.7% | Control 36.7%

Statistical Considerations • Multiplicity control – Overall Type I error rate was strongly controlled at 2.5% (1-sided) within each comparison – Prespecified hierarchical testing strategy across PFS and OS endpoints – Alpha allocation shown on the right • Efficacy analyses – PFS and OS: HRs and 95% CIs estimated by stratified Cox regression model; between-arm differences assessed with stratified log-rank test – ORR: estimated with exact Clopper and Pearson method – Efficacy population: ITT population • Safety analyses – Safety population: all treated participants with ≥1 dose of study treatment Pembro + len + belz vs. pembro + len PFS (dual primary) ITT population α = 0.005 ⇄ (1.0 / 1.0) OS (dual primary) ITT population α = 0.02 Qmab/pembro + len vs. pembro + len PFS (dual primary) Int/poor population α = 0.0035 → (1.0) PFS (dual primary) ITT population α = 0 PFS (dual primary) ITT population α = 0 → (1.0) OS (dual primary) Int/poor population α = 0.0215 OS (dual primary) Int/poor population α = 0.0215 → (1.0) OS (dual primary) ITT population α = 0 OS (dual primary) ITT population α = 0 → (1.0) PFS (dual primary) Int/poor population α = 0.0035 Data cutoff date: 4 Feb 2026

[Slide 2] 3586O First-line Pembrolizumab Plus Lenvatinib Plus Belzutifan or MK-1308A in Advanced RCC: Randomized Open-label Phase 3 LITESPARK-012 Study. Choueiri et. al. LITESPARK-012 Study Key Eligibility Criteria - Adults with advanced ccRCC (stage IV per AJCC) - KPS score >=70% - Measurable disease per RECIST 1.1 - No prior systemic therapy for advanced RCC Stratification Factors - IMDC prognostic score (0 [favorable] vs. 1-2 [intermediate] vs. 3-6 [poor]) - Geographic region (North America vs. western Europe vs. rest of world) - Sarcomatoid features (yes vs. no) N = 1688 R N = 560: Pembrolizumab 400 mg IV Q6W up to 18 cycles + Lenvatinib 20 mg orally QD + Belzutifan 120 mg orally QD N = 562: Quavonlimab/pembrolizumab 25 mg/400 mg (MK-1308A) IV Q6W up to 18 cycles + Lenvatinib 20 mg orally QD N = 566: CONTROL Pembrolizumab 400 mg IV Q6W up to 18 cycles + Lenvatinib 20 mg orally QD Dual primary endpoints: Progression-free survival (PFS) by blinded independent central review (BICR) per RECIST 1.1 and overall survival (OS) Secondary endpoints: Objective response rate (ORR) and duration of response (DOR) by BICR per RECIST 1.1, safety PFS by BICR, Pembro + Len + Belz vs. Pembro + Len ITT Population | Pembro + len + belz N = 560 | Pembro + len N = 566 Events, n (%) | 337 (60.2) | 345 (61.0) Median, mo (95% CI) | 20.8 (17.9-23.7) | 23.7 (20.7-26.6) HR 1.08 (95% CI 0.93-1.26); 1-sided P=0.8408 Landmark labels on curve: 69.1% | 66.3% | 49.6% | 45.3% | 36.7% | 34.7% No. at Risk (0-60 mo) Pembro + len + belz 560 | 424 | 329 | 257 | 210 | 176 | 122 | 78 | 23 | 2 | 0 Pembro + len 566 | 455 | 365 | 298 | 243 | 199 | 129 | 75 | 24 | 1 | 0 Data cutoff date: 4 Feb 2026 PFS by BICR, Qmab/Pembro + Len vs. Pembro + Len IMDC Intermediate/Poor Risk Population (Per Sequential Testing Strategy) | Qmab/pembro + len N = 429 | Pembro + len N = 429 Events, n (%) | 259 (60.4) | 261 (60.8) Median, mo (95% CI) | 20.5 (16.6-23.5) | 23.5 (20.6-26.3) HR 1.12 (95% CI 0.94-1.33); 1-sided P=0.8954 Landmark labels on curve: 67.5% | 62.3% | 47.6% | 43.7% | 34.8% | 31.2% No. at Risk (0-60 mo) Qmab/pembro + len 429 | 310 | 230 | 185 | 147 | 124 | 79 | 49 | 17 | 3 | 0 Pembro + len 429 | 337 | 267 | 219 | 176 | 148 | 84 | 48 | 17 | 1 | 0 Data cutoff date: 4 Feb 2026 Survival table Endpoint: PFS by BICR, Pembro + Len + Belz vs. Pembro + Len (ITT Population) Arm | N | Median | Median CI Pembro + len + belz | 560 | 20.8 mo | 17.9-23.7 Pembro + len | 566 | 23.7 mo | 20.7-26.6 HR: 1.08 HR CI: 0.93-1.26 P: 1-sided P=0.8408 [Slide 3] Cosmic 313 Motzer et. al. Annals of Oncol. 2026 PFS benefit was maintained with longer follow-up Median PFS (95% CI), months Cabo + nivo + ipi (n=428) | 16.6 (14.0-22.6) Placebo + nivo + ipi (n=427) | 11.2 (9.3-14.0) HR, 0.82 (95% CI, 0.69-0.98) Landmark labels on curve: 57% | 48% | 44% (cabo) ; 48% | 41% | 37% (placebo) No. at risk (0-54 mo, 3-mo steps) Cabo + nivo + ipi 428 | 359 | 269 | 232 | 192 | 162 | 148 | 144 | 126 | 112 | 100 | 84 | 71 | 54 | 37 | 20 | 11 | 2 | 0 Placebo + nivo + ipi 427 | 295 | 221 | 192 | 154 | 133 | 119 | 103 | 94 | 85 | 78 | 71 | 67 | 58 | 35 | 11 | 7 | 2 | 0 OS in the ITT Population OS was comparable between arms Median OS (95% CI), months Cabo + nivo + ipi (n=42[illegible]) | 41.9 (34.8-47.9) Placebo + nivo + ipi (n=427) | 42.0 (34.9-53.1) HR, 1.02 (95% CI, 0.85-1.23); P=0.8366 Landmark labels on curve: 67% | 65% | 54% | 54% No. at risk (0-57 mo) Cabo + nivo + ipi 428 | 404 | 386 | 363 | 341 | 328 | 313 | 298 | 279 | 264 | 250 | 237 | 223 | 187 | 160 | 101 | 54 | 21 | 4 | 0 Placebo + nivo + ipi 427 | 402 | 376 | 350 | 331 | 312 | 297 | 278 | 265 | 253 | 240 | 230 | 216 | 189 | 168 | 106 | 59 | 24 | 4 | 0 follow-up of 45.0 months Treatment Exposure and Summary of AEs (Safety Population) Treatment exposure and dose modifications | Cabo + nivo + ipi (n=426) | Placebo + nivo + ipi (n=423) Median duration of exposure of study treatment, months (range) | 13.7 (0.2-53.4) | 10.3 (0.1-55.3) Median average daily dose of cabo/placebo, mg (range) | 22.4 (3.6-40.0) | 35.2 (0.8-40.0) Median nivo infusions per patient, n (range) | 11.0 (1-52) | 9.0 (1-48) 1 / 2 / 3 / 4 cycles of ipi administered, % | 7 / 22 / 13 / 58 | 6 / 7 / 13 / 74 Dose modification (any treatment component) due to AE, % | 92 | 76 Treatment-related AE leading to discontinuation of >=1 component, % | 49 | 26 Most common treatment-related AEs (any grade in >=25% of patients) | Cabo + nivo + ipi Any Grade | Grade 3/4 | Placebo + nivo + ipi Any Grade | Grade 3/4 Treatment-related AE, % | 99 | 75 | 91 | 43 Alanine aminotransferase increased | 46 | 26 | 19 | 6 Aspartate aminotransferase increased | 44 | 19 | 17 | 5 Diarrhea | 43 | 5 | 21 | 3 Palmar-plantar erythrodysesthesia syndrome | 29 | 4 | 5 | 0 Hypothyroidism | 26 | <1 | 17 | 0 Hypertension | 25 | 9 | 7 | 2 Pruritus | 21 | 0 | 26 | <1 - Use of systemic immune-modulating medications* for AE was 73.0% in the cabo + nivo + ipi arm and 56.7% in the control arm *Includes glucocorticoids and other immunosuppressants. AE, adverse event. Advanced RCC (N~840) - No prior systemic therapy* - Clear cell component - Intermediate or poor risk per IMDC criteria - Measurable disease per RECIST v1.1 - Karnofsky Performance Status >=70% R1:1 | Stratification: IMDC risk, Region Cabo + nivo + ipi: Cabo 40 mg PO QD + nivo 3 mg/kg IV Q3W x 4 + ipi 1 mg/kg IV Q3W x 4 -> Cabo 40 mg PO QD + nivo 480 mg IV Q4W Placebo + nivo + ipi: Placebo PO QD + nivo 3 mg/kg IV Q3W x 4 + ipi 1 mg/kg IV Q3W x 4 -> Placebo PO QD + nivo 480 mg IV Q4W Tumor assessment every 8 weeks per RECIST v1.1 Treatment until loss of clinical benefit or intolerable toxicity No crossover allowed Primary endpoint: PFS per RECIST v1.1 by BIRC after the 249th event in the first 550 randomized patients (PITT population) Secondary endpoint: OS after 433 events in all randomized patients (ITT population) Additional endpoints: ORR, DOR, and safety Survival table Endpoint: PFS (COSMIC-313, longer follow-up) Arm | N | Median | Median CI Cabo + nivo + ipi | 428 | 16.6 months | 14.0-22.6 Placebo + nivo + ipi | 427 | 11.2 months | 9.3-14.0 HR: 0.82 HR CI: 0.69-0.98 [Slide 4] Characteristics | MDT+ICI N = 30 | MDT N = 120 | P Age, years, median (IQR) | 62 (52 to 71) | 66 (61 to 72) | 0.06 Male sex | 23 (77) | 88 (73) | 0.7 ECOG Performance Status | | | 0.8 0 | 20 (67) | 81 (68) 1 | 10 (33) | 35 (29) 2 | 0 (0) | 4 (3) IMDC risk score | | | 0.7 Favorable | 17 (57) | 62 (52) Intermediate | 13 (43) | 57 (47) Poor | 0 (0) | 1 (1) Prior systemic therapy* | 7 (23) | 36 (30) | 0.7 TKI | 6 (86) | 19 (53) ICI | 0 (0) | 13 (36) Dual TKI+ICI | 0 (0) | 4 (11) Interleukin-2 | 1 (14) | 0 (0) Prior nephrectomy | 27 (90) | 117 (97) | 0.09 Number of metastases | | | < 0.001 1 | 5 (17) | 71 (59) 2 | 8 (27) | 37 (31) 3 | 9 (30) | 7 (6) 4 | 5 (17) | 3 (2) 5 | 3 (10) | 2 (2) Organs involved Lung | 20 (67) | 76 (63) | 0.7 Lymph node | 9 (30) | 16 (13) | 0.03 Bone | 7 (23) | 11 (9) | 0.03 Other | 10 (33) | 29 (24) | 0.3 30 pts from RAPPORT (NCT02855203) with radiation to oligometastases and ICI and 120 from MDACC MDT alone (NCT03575611) MDT +ICI Cohort had more unfavorable features Longer PFS with SBRT+ICI vs MDT OS not significant PFS: HR 0.57; 95% CI: 0.32 to 1.02; p = 0.058 PSM: HR 0.44; 95% CI: 0.21 to 0.93; p = 0.03 Median f/u: 34 months OS: HR 1.9; 95% CI: 0.67 to 5.38; p = 0.2 PSM: HR 1.7; 95% CI: 0.48 to 5.78; p = 0.4 Legend: MDT | MDT+ICI SHERRY AD et al, ASCO, J Clin Oncol 44, 2026 (suppl 7; abstr 498) Naomi B. Haas, MD

[Slide 1] CA017-078 Overall survival by treatment arm [table] NIVO+GC (N = 325) | GC (N = 328) # of events/# of patients (%): 71/325 (21.8) | 81/328 (24.7) Median OS (95% CI), mo: NR (NE-NE) | NR (NE-NE) HR (95% CI): 0.88 (0.64-1.21) Curve labels: GC 92.8%, NIVO+GC 91.6% (12 mo); GC 85.2%, NIVO+GC 84.8% (24 mo); GC 80.5%, NIVO+GC 79.8% (36 mo); NIVO+GC 79.8%, GC 78.4% (~39 mo) Y-axis: Probability of overall survival (%); X-axis: Months No. at risk (0,6,12,18,24,30,36,42,48,54,60,66,72,78 mo) NIVO+GC: 325 305 281 268 257 248 238 204 148 113 72 36 16 0 GC: 328 299 282 269 255 249 238 205 149 105 66 32 17 0 - Median OS was not reached in either treatment arm - OS was analyzed descriptively because the EFS testing condition was not met Survival table Endpoint: OS Arm | N | Median | Median CI NIVO+GC | 325 | NR | NE-NE GC | 328 | NR | NE-NE HR: 0.88 HR CI: 0.64-1.21 Landmark rates: 12 mo: NIVO+GC 91.6% | GC 92.8% 24 mo: NIVO+GC 84.8% | GC 85.2% 36 mo: NIVO+GC 79.8% | GC 80.5% [Slide 2] Key Secondary Endpoint: OS ITT Population [table] Pembro + cis + gem n = 453 | Placebo + cis + gem n = 454 Events, n (%): 135 (29.8) | 135 (29.7) Median (95% CI), mo: NR (NR-NR) | NR (NR-NR) HR 1.00 (95% CI, 0.79-1.27)* Curve labels: 24 mo 82.8% (pembro) / 81.7% (placebo); 36 mo 76.8% / 76.6%; 48 mo 72.9% / 73.4%; 60 mo 70.8% / 70.9% Y-axis: Overall Survival, %; X-axis: Months No. at risk (0,6,12,18,24,30,36,42,48,54,60,66,72,78,84 mo) Pembro + cis + gem: 453 440 409 386 375 363 348 332 278 231 161 94 43 1 0 Placebo + cis + gem: 454 437 410 390 370 356 344 332 278 229 159 97 40 2 0 NR, not reached. *Not formally statistically tested according to the statistical design. Data cutoff: January 26, 2026. Survival table Endpoint: OS (key secondary), ITT Arm | N | Median | Median CI Pembro + cis + gem | 453 | NR | NR-NR Placebo + cis + gem | 454 | NR | NR-NR HR: 1.00 HR CI: 0.79-1.27 Landmark rates: 24 mo: Pembro + cis + gem 82.8% | Placebo + cis + gem 81.7% 36 mo: Pembro + cis + gem 76.8% | Placebo + cis + gem 76.6% 48 mo: Pembro + cis + gem 72.9% | Placebo + cis + gem 73.4% 60 mo: Pembro + cis + gem 70.8% | Placebo + cis + gem 70.9%

[Slide 1] Study design: CA017-078 Phase 3 randomized trial of neoadjuvant NIVO plus GC, followed by adjuvant NIVO, vs neoadjuvant GC alone for previously untreated MIBC (NCT03661320) Screening Key inclusion criteria - MIBC (predominant urothelial component) - cT2-T4a, N0, M0 - GFR >= 50 mL/min - ECOG PS 0-1 Stratification factors - Clinical stage (T2 vs T3/T4a) - GFR (50 to <60 mL/min vs >= 60 mL/min) - PD-L1 status (>= 1% vs < 1%/indeterminate/NE) by central reviewa R 1:1 Neoadjuvant presurgical treatmentb,c | Surgery | Adjuvant postsurgical treatment GC Q3W x 4 cycles + NIVO 360 mg Q3W x 4 cycles, N = 325 -> Radical cystectomy -> NIVO 480 mg Q4W x 9 cycles GC Q3W x 4 cycles, N = 328 -> Radical cystectomy -> No further treatment TERMINATED (Dec 2021) Excluded from formal testing of NIVO+GC vs GC alone: GC Q3W x 4 cycles + NIVO 360 mg Q3W x 4 cycles + BMS986205 (blinded) 100 mg QD -> Radical cystectomy -> NIVO 480 mg Q4W + BMS986205 (blinded) 100 mg QD x 9 cycles Co-primary endpointsd: pCR and EFS per BICR Secondary endpointsd: OS, safety Exploratory endpoints: ctDNA analyses Median follow-up (range): 54.5 months (39.4-77.0) GFR based on the Cockcroft-Gault formula. aAssessed by the Dako PD-L1 IHC 28-8 pharmDx assay. bAll patients receive GC chemotherapy on a Q3W schedule. Dose and schedule depend on the patient's calculated GFR. GFR >= 60 mL/min: gemcitabine 1000 mg/m2 D1, D8; cisplatin 70 mg/m2 D1; GFR 50 to < 60 mL/min: gemcitabine 1000 mg/m2 D1, D8; cisplatin 35 mg/m2 D1, D8. cA third arm (NIVO + linrodostat + GC/NIVO + linrodostat) was terminated. dpCR rate was defined by the proportion of randomized patients with absence of any cancer (T0 and N0) in pathology specimens after RC, based on blinded pathology review. EFS was defined as the time from randomization to any of the following events: progression of disease that precludes RC surgery, local or distant recurrence based on BICR assessments, or death due to any cause. OS was defined as the time between the date of randomization and the date of death from any cause. [Slide 2] CA017-078 Event-free survival by treatment arm (co-primary endpoint) [table] NIVO+GC (N = 325) | GC (N = 328) # of events/# of patients (%): 104/325 (32.0) | 104/328 (31.7) Median EFS (95% CI), mo: NR (NE-NE) | 67.6 (63.0-NE) HR (95% CI): 0.94 (0.72-1.24) HR (95.74% CI): 0.94 (0.71-1.25) P valuea: 0.6697 Curve labels: 12 mo NIVO+GC 82.6% / GC 78.6%; 24 mo 72.3% / 71.6%; 36 mo 68.1% / 68.5%; ~39 mo 67.3% / 65.8% Y-axis: Probability of event-free survival per BICR (%); X-axis: Months No. at risk (0,6,12,18,24,30,36,42,48,54,60,66,72,78 mo) NIVO+GC: 325 283 243 227 209 197 189 146 100 60 25 5 0 0 GC: 328 271 219 199 187 176 168 131 90 61 23 4 1 0 - Median EFS was not reached in the NIVO+GC arm and was 67.6 months (95% CI 63.0-NE) in the GC arm - The EFS co-primary endpoint did not meet the significance criterion determined by the prespecified testing procedure aTwo-sided P value from log-rank test stratified by PD-L1 status, GFR category, and T stage at screening as entered into the IRT. P value threshold for statistical significance, 0.0426. Survival table Endpoint: EFS per BICR (co-primary) Arm | N | Median | Median CI NIVO+GC | 325 | NR | NE-NE GC | 328 | 67.6 | 63.0-NE HR: 0.94 HR CI: 0.72-1.24 (95% CI); 0.71-1.25 (95.74% CI) P: 0.6697 (two-sided; threshold 0.0426) Landmark rates: 12 mo: NIVO+GC 82.6% | GC 78.6% 24 mo: NIVO+GC 72.3% | GC 71.6% 36 mo: NIVO+GC 68.1% | GC 68.5%

[Slide 1] CHAIRS: AXEL BEX, ANJA LORCH Results from a phase 3 trial of neoadjuvant chemotherapy alone vs in combination with nivolumab followed by postsurgical adjuvant nivolumab in patients with muscle-invasive bladder cancer Andrea Necchi,1* Shilpa Gupta,2 Michiel S. van der Heijden,3 Matthew D. Galsky,4 Gary Steinberg,5† Jürgen E. Gschwend,6 Juan Pablo Sade,7 Davide Bimbatti,8 Ugo De Giorgi,9‡ Sergio Vazquez Estevez,10 Sergio Panay,11 Giovanni Schinzari,12 Athanasios Kotsakis,13 Hideaki Miyake,14 JongJin Oh,15 Joshua Zhang,16‡ Chandrika Jeyamohan,16 Miye Jacques,16 Quyen Duong,16 Guru Sonpavde17§ 1Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy; 2Taussig Cancer Institute, Cleveland Clinic Foundation, Cleveland, OH, USA; 3Netherlands Cancer Institute, Amsterdam, the Netherlands; 4Tisch Cancer Institute, Icahn School of Medicine at Mount Sinai, New York, NY, USA; 5NYU Langone Health, New York, NY, USA; 6Technical University of Munich, Munich, Germany; 7University of Buenos Aires, Buenos Aires, Argentina; 8Veneto Institute of Oncology IOV-IRCCS, Padua, Italy; 9IRCCS Istituto Romagnolo per lo Studio dei Tumori (IRST) "Dino Amadori," Meldola, Italy; 10Hospital Universitario Lucus Augusti (HULA), EOXI de Lugo, Cervo e Monforte, Lugo, Spain; 11Fundación Arturo López Pérez, Santiago, Chile; 12Catholic University of the Sacred Heart, Comprehensive Cancer Center, Fondazione Policlinico Universitario Agostino Gemelli IRCCS, Rome, Italy; 13General Hospital of Larissa and Faculty of Medicine, School of Health Sciences, University of Thessaly, Larissa, Thessaly, Greece; 14Kobe University Hospital School of Medicine, Kobe, Japan; 15Seoul National University Bundang Hospital, Seoul National University College of Medicine, Seoul, South Korea; 16Bristol Myers Squibb, Princeton, NJ, USA; 17Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA, USA *Dr Necchi is now with the Department of Medical Oncology, IRCCS San Raffaele Hospital, Comprehensive Cancer Center, Milan, Italy. †Dr Steinberg is now with Rush University Medical Center, Chicago, IL, USA. ‡Dr De Giorgi is now with the University Oncology Unit, "Vito Fazzi" Hospital, Lecce, Italy. §Dr Sonpavde is now with AdventHealth Cancer Institute and Loma Linda University, Orlando, FL, USA. Presentation number 4736O [Speaker panel] Andrea Necchi - Results from a phase 3 trial of neoadjuvant (neoadj) chemotherapy alone vs in combination with nivolumab followed by postsurgical adjuvant nivolumab in patients (pts) with muscle-invasive bladder cancer (MIBC) [Slide 3] 13:30 - 15:00 Proffered paper 1: Genitourinary tumour, renal and urothelial CHAIRS: AXEL BEX, ANJA LORCH CA017-078 EFS and OS by pre-cystectomy ctDNA status (pooled)^a EFS by pre-cystectomy ctDNA status | OS by pre-cystectomy ctDNA status Legend: Negative C1D1+C4D1; Cleared by C4D1; Positive at C4D1 No. at risk - EFS (Months 0-80) Negative C1D1+C4D1: 147 127 115 103 83 45 13 2 0 Cleared by C4D1: 153 129 105 93 79 37 9 1 0 Positive at C4D1: 46 19 14 9 8 4 3 1 0 No. at risk - OS (Months 0-80) Negative C1D1+C4D1: 147 139 134 130 103 57 14 1 0 Cleared by C4D1: 153 146 131 118 98 48 17 2 0 Positive at C4D1: 46 33 24 20 13 6 3 1 0 - In a pooled analyses across treatment arms, ctDNA absence (at both C1D1 and C4D1) and ctDNA clearance was associated with longer EFS and OS ^aBased on pooled analysis of all evaluable patients in the NIVO+GC and GC arms. ctDNA testing was performed using the Signatera tumor-informed molecular residual disease assay (Natera, Austin, TX, USA). Pre-cystectomy ctDNA status was categorized as negative at C1D1 and C4D1, positive at C1D1, but negative by C4D1, or positive at C4D1. BARCELONA AUDITORIUM - HALL 9 Survival table Endpoint: EFS and OS by pre-cystectomy ctDNA status (pooled across arms) Negative C1D1+C4D1 | 147 Cleared by C4D1 | 153 Positive at C4D1 | 46 [Slide 4] 13:30 - 15:00 Proffered paper 1: Genitourinary tumour, renal and urothelial CHAIRS: AXEL BEX, ANJA LORCH CA017-078 Event-free survival by treatment arm (co-primary endpoint) | | NIVO+GC (N = 325) | GC (N = 328) | # of events/# of patients (%) | 104/325 (32.0) | 104/328 (31.7) Median EFS (95% CI), mo | NR (NE-NE) | 67.6 (63.0-NE) HR (95% CI) | 0.94 (0.72-1.24) HR (95.74% CI) | 0.94 (0.71-1.25) P value^a | 0.6697 Curve landmark labels (NIVO+GC / GC): 82.6% / 78.6% (12 mo line); 72.3% / 71.6% (24 mo line); 68.1% / 68.5% (36 mo line); 67.3% / 65.8% (fourth dashed line, ~38-39 mo, timepoint not labelled) No. at risk (Months 0, 6, 12, 18, 24, 30, 36, 42, 48, 54, 60, 66, 72, 78) NIVO+GC: 325 283 243 227 209 197 189 146 100 60 25 5 0 0 GC: 328 271 219 199 187 176 168 131 90 61 23 4 1 0 - Median EFS was not reached in the NIVO+GC arm and was 67.6 months (95% CI 63.0-NE) in the GC arm - The EFS co-primary endpoint did not meet the significance criterion determined by the prespecified testing procedure ^aTwo-sided P value from log-rank test stratified by PD-L1 status, GFR category, and T stage at screening as entered into the IRT. P value threshold for statistical significance, 0.0426. BARCELONA AUDITORIUM - HALL 9 Survival table Endpoint: EFS per BICR (co-primary) Arm | N | Median | Median CI NIVO+GC | 325 | NR | NE-NE GC | 328 | 67.6 mo | 63.0-NE HR: 0.94 HR CI: 0.72-1.24 (95% CI); 0.71-1.25 (95.74% CI) P: 0.6697 Landmark rates: 12 mo: NIVO+GC 82.6% | GC 78.6% 24 mo: NIVO+GC 72.3% | GC 71.6% 36 mo: NIVO+GC 68.1% | GC 68.5% unlabelled 4th dashed line (~38-39 mo): NIVO+GC 67.3% | GC 65.8%

[Slide 1] 13:30 - 15:00 Proffered paper 1: Genitourinary tumour, renal and urothelial CHAIRS: AXEL BEX, ANJA LORCH CA017-078 Neoadjuvant ctDNA clearance and pCR ctDNA clearance (neoadjuvant C1D1 to C4D1) Fisher exact P = 0.130 Cleared ctDNA at C4D1 (%) GC: 71.7% (66/92) NIVO+GC: 81.3% (87/107) Bars show % of C1D1 ctDNA-positive patients negative at C4D1; whiskers are Wilson 95% CIs. pCR by ctDNA clearance status (neoadjuvant C1D1 to C4D1) pCR rate by treatment arm, faceted by C1D1 to C4D1 ctDNA clearance status Pathologic complete response rate (%) ctDNA cleared by C4D1 n = 153 Fisher exact P = 0.242 GC: 33.3% (22/66) NIVO+GC: 43.7% (38/87) ctDNA not cleared by C4D1 n = 46 Fisher exact P = 1.000 GC: 7.7% (2/26) NIVO+GC: 5.0% (1/20) • ctDNA clearance was higher with NIVO+GC vs GC alone • In both treatment arms, ctDNA clearance was associated with a higher pCR rate • Among patients with ctDNA clearance, pCR rates were higher with NIVO+GC vs GC alone ctDNA testing was performed using the Signatera tumor-informed molecular residual disease assay (Natera, Austin, TX, USA). ctDNA clearance status was categorized as positive at C1D1 but negative by C4D1 (cleared), or positive at both C1D1 and C4D1 (not cleared). The ctDNA-evaluable population (patients with ctDNA results for both C1D1 and C4D1) comprised 53.0% (346/653) of all randomly assigned patients. The ctDNA-evaluable population was representative of the total study population across all demographic features and baseline demographic characteristics were well balanced between the treatment arms. [Speaker panel] Andrea Necchi Results from a phase 3 trial of neoadjuvant (neoadj) chemotherapy alone vs in combination with nivolumab followed by postsurgical adjuvant nivolumab in patients (pts) with muscle-invasive bladder cancer (MIBC) BARCELONA AUDITORIUM - HALL 9 [Slide 2] 13:30 - 15:00 Proffered paper 1: Genitourinary tumour, renal and urothelial CHAIRS: AXEL BEX, ANJA LORCH CA017-078 EFS and OS by pre-treatment ctDNA status (pooled)a EFS by pretreatment ctDNA status | OS by pretreatment ctDNA status Legend: C1D1 positive; C1D1 negative Y-axis: Survival probability; X-axis: Months No. at risk (EFS), months 0/10/20/30/40/50/60/70/80: C1D1 positive: 229 164 136 118 99 45 10 1 0 C1D1 negative: 187 149 132 118 94 53 14 1 0 No. at risk (OS), months 0/10/20/30/40/50/60/70/80: C1D1 positive: 229 202 175 158 128 62 24 2 0 C1D1 negative: 187 168 160 153 120 68 17 1 0 • In a pooled analyses across treatment arms, pretreatment ctDNA detection was associated with shorter EFS and OS aBased on pooled analysis of all evaluable patients in the NIVO+GC and GC arms. ctDNA testing was performed using the Signatera tumor-informed molecular residual disease assay (Natera, Austin, TX, USA). Pretreatment ctDNA status was categorized as positive or negative based on detectable ctDNA at C1D1. [Speaker panel] Andrea Necchi - Results from a phase 3 trial of neoadjuvant (neoadj) chemotherapy alone vs in combination with nivolumab followed by postsurgical adjuvant nivolumab in patients (pts) with muscle-invasive bladder cancer (MIBC) BARCELONA AUDITORIUM - HALL 9 Survival table Endpoint: EFS and OS by pretreatment (C1D1) ctDNA status, pooled across NIVO+GC and GC arms C1D1 ctDNA positive | 229 C1D1 ctDNA negative | 187 [Slide 3] 13:30 - 15:00 Proffered paper 1: Genitourinary tumour, renal and urothelial CHAIRS: AXEL BEX, ANJA LORCH CA017-078 EFS and OS by pre-cystectomy ctDNA status (pooled)a EFS by pre-cystectomy ctDNA status | OS by pre-cystectomy ctDNA status Legend: Negative C1D1+C4D1; Cleared by C4D1; Positive at C4D1 Y-axis: Survival probability; X-axis: Months No. at risk (EFS), months 0/10/20/30/40/50/60/70/80: Negative C1D1+C4D1: 147 127 115 103 83 45 13 2 0 Cleared by C4D1: 153 129 105 93 79 37 9 1 0 Positive at C4D1: 46 19 14 9 8 4 3 1 0 No. at risk (OS), months 0/10/20/30/40/50/60/70/80: Negative C1D1+C4D1: 147 139 134 130 103 57 14 1 0 Cleared by C4D1: 153 146 131 118 98 48 17 2 0 Positive at C4D1: 46 33 24 20 13 6 3 1 0 • In a pooled analyses across treatment arms, ctDNA absence (at both C1D1 and C4D1) and ctDNA clearance was associated with longer EFS and OS aBased on pooled analysis of all evaluable patients in the NIVO+GC and GC arms. ctDNA testing was performed using the Signatera tumor-informed molecular residual disease assay (Natera, Austin, TX, USA). Pre-cystectomy ctDNA status was categorized as negative at C1D1 and C4D1, positive at C1D1, but negative by C4D1, or positive at C4D1. [Speaker panel] Andrea Necchi - Results from a phase 3 trial of neoadjuvant (neoadj) chemotherapy alone vs in combination with nivolumab followed by postsurgical adjuvant nivolumab in patients (pts) with muscle-invasive bladder cancer (MIBC) BARCELONA AUDITORIUM - HALL 9 Survival table Endpoint: EFS and OS by pre-cystectomy (C1D1/C4D1) ctDNA status, pooled across NIVO+GC and GC arms Negative C1D1+C4D1 | 147 Cleared by C4D1 | 153 Positive at C4D1 | 46

[Slide 1] Study design: CA017-078 Phase 3 randomized trial of neoadjuvant NIVO plus GC, followed by adjuvant NIVO, vs neoadjuvant GC alone for previously untreated MIBC (NCT03661320) Screening - Key inclusion criteria: MIBC (predominant urothelial component); cT2-T4a, N0, M0; GFR >= 50 mL/min; ECOG PS 0-1 Stratification factors: Clinical stage (T2 vs T3/T4a); GFR (50 to <60 mL/min vs >= 60 mL/min); PD-L1 status (>= 1% vs < 1%/indeterminate/NE) by central review R 1:1 -> GC Q3W x 4 cycles + NIVO 360 mg Q3W x 4 cycles N = 325 -> Radical cystectomy -> NIVO 480 mg Q4W x 9 cycles GC Q3W x 4 cycles N = 328 -> Radical cystectomy -> No further treatment TERMINATED (Dec 2021) Excluded from formal testing of NIVO+GC vs GC alone: GC Q3W x 4 cycles + NIVO 360 mg Q3W x 4 cycles + BMS986205 (blinded) 100 mg QD -> Radical cystectomy -> NIVO 480 mg Q4W + BMS986205 (blinded) 100 mg QD x 9 cycles Co-primary endpoints: pCR and EFS per BICR; Secondary endpoints: OS, safety; Exploratory endpoints: ctDNA analyses Median follow-up (range): 54.5 months (39.4-77.0) [footnotes illegible at this resolution] [Slide 4] Summary - In CA017-078, adding perioperative NIVO to neoadjuvant GC increased the pCR rate vs GC alone (39.1% vs 29.6%); however, the prespecified statistical threshold for the co-primary endpoint of pCR was not met - The co-primary endpoint of EFS was not improved with NIVO+GC vs GC (HR, 0.94; P value, 0.6697), and OS data were immature with no clear treatment difference (HR, 0.88) - The safety profile of NIVO+GC was consistent with that of the individual agents1,2 - Exploratory analyses suggested that pretreatment ctDNA status and ctDNA clearance during neoadjuvant treatment were prognostic (patients who were ctDNA-negative before treatment or achieved ctDNA clearance had better EFS and OS in both treatment arms) - Although the co-primary endpoints were not met, these mature randomized phase 3 data - including exploratory ctDNA analyses - are practice-informing 1. Bajorin DF, et al. N Engl J Med. 2021;384:2102-2114. 2. van der Heijden MS, et al. N Engl J Med. 2023;389:1778-1789

[Slide 2] CA017-078 Pathological complete response by treatment arm pCR overall (co-primary endpoint)a - Overall ITT: NIVO+GC 39.1 | GC 29.6; difference 9.5% (95% CI [illegible, approx. 2.3-16.7]) (99.9x% CI [illegible, approx. -2.x to 21.6]a) pCR by subgroups (Patients %): PD-L1 ≥1%: 46.4 | 35.5; difference 10.9% (95% CI [illegible]) PD-L1 <1%: 33.3 | 24.3; difference 9.1% (95% CI [illegible]) cT2: 40.6 | 31.5; difference 9.1% (95% CI [illegible]) cT3/cT4a: 35.1 | 23.9; difference 11.2% (95% CI [illegible]) Combined response (≤ ypT1N0) - Overall ITT: 51.7 | 41.8; difference 9.9% (95% CI [illegible]) Legend: NIVO+GC; GC • Overall pCR rate was higher with NIVO+GC, but the multiplicity-adjusted analysis did not establish a significant difference • Descriptive pCR rates were higher with NIVO+GC in both PD-L1 strata and in the cT2 and cT3/cT4a subgroups • Overall combined response rate (≤ ypT1N0) was higher with NIVO+GC aResults shown are from a descriptive pCR update at the EFS final analysis cutoff. The prespecified pCR final analysis had already been completed, and formal hypothesis testing was not repeated. Therefore, no P value is reported. bThe pCR hypothesis was evaluated at the multiplicity-adjusted 2-sided nominal significance level of 0.0009[illegible] at the final analysis, corresponding to a 99.9x% CI. Conventional 95% CIs were also reported descriptively. [Slide 3] CA017-078 Overall survival by treatment arm Table: NIVO+GC (N = 325) | GC (N = 328) # of events/# of patients (%): 71/325 (21.8)[blurry] | 81/328 (24.7)[blurry] Median OS (95% CI), mo: NR (NE-NE) | NR (NE-NE) HR (95% CI): 0.88 (0.64-1.21)[blurry] Y-axis: Probability of overall survival (%); X-axis: Months 0-78 Landmark labels on curve: 91.6% (NIVO+GC) / 92.8% (GC) at 12 mo; 84.8% / 85.2% at 24 mo; 79.8% / 80.5% at 36 mo; 79.8% / 78.4% at ~39 mo dashed line No. at risk (months 0,6,12,...,78): NIVO+GC: 325 305 281 268 257 248 238 204 148 113 72 36 16 0 GC: 328 299 282 269 255 249 238 205 149 105 66 32 17 0 • Median OS was not reached in either treatment arm • OS was analyzed descriptively because the EFS testing condition was not met Survival table Endpoint: Overall survival (secondary; descriptive) Arm | N | Median | Median CI NIVO+GC | 325 | NR | NE-NE GC | 328 | NR | NE-NE HR: 0.88 [blurry] HR CI: 0.64-1.21 [blurry] Landmark rates: 12 mo: NIVO+GC 91.6% | GC 92.8% 24 mo: NIVO+GC 84.8% | GC 85.2% 36 mo: NIVO+GC 79.8% | GC 80.5% ~39 mo (dashed line, exact month not labeled): NIVO+GC 79.8% | GC 78.4% [Slide 4] CA017-078 EFS and OS by pre-treatment ctDNA status (by treatment)a EFS by pretreatment ctDNA status | OS by pretreatment ctDNA status Legend: GC C1D1 positive; GC C1D1 negative; NIVO+GC C1D1 positive; NIVO+GC C1D1 negative Y-axis: Survival probability; X-axis: Months 0-80 No. at risk (EFS), months 0-80 by 10: GC C1D1 positive: 108 71 55 48 43 21 7 1 0 [faint, low confidence] GC C1D1 negative: 103 [rest illegible] NIVO+GC C1D1 positive: 121 94 82 71 57 25 4 1 0 NIVO+GC C1D1 negative: 84 64 59 52 44 27 11 1 0 [low confidence] No. at risk (OS), months 0-80 by 10: GC C1D1 positive: 108 92 79 70 59 28 11 2 0 [faint, low confidence] GC C1D1 negative: [illegible] NIVO+GC C1D1 positive: 121 110 97 89 70 35 14 1 0 NIVO+GC C1D1 negative: 84 74 71 67 55 35 13 1 0 [low confidence] • No clear treatment difference was observed between NIVO+GC and GC alone, regardless of pretreatment ctDNA status aPretreatment ctDNA status was categorized as positive or negative based on detectable ctDNA at C1D1. ctDNA testing was performed using the Signatera tumor-informed molecular residual disease assay (Natera, Austin, TX, USA). Survival table Endpoint: EFS and OS by pretreatment (C1D1) ctDNA status and treatment arm GC C1D1 positive | 108 GC C1D1 negative NIVO+GC C1D1 positive | 121 NIVO+GC C1D1 negative | 84

[Slide 2] CA017-078 Demographics and baseline characteristics | | NIVO+GC (N = 325) | GC (N = 328) | Median age, years | 66.0 | 67.0 < 65 years, n (%) | 137 (42.2) | 136 (41.5) ≥ 65, n (%) | 188 (57.8) | 192 (58.5) Sex, n (%) Male | 264 (81.2) | 259 (79.0) Race, n (%) White | 251 (77.2) | 257 (78.4) Asian | 55 (16.9) | 55 (16.8) Other^a | 12 (3.7) | 10 (3.0) Calculated GFR, n (%)^b < 50 mL/min | 4 (1.2) | 5 (1.5) 50 to < 60 mL/min | 34 (10.5) | 30 (9.1) ≥ 60 mL/min | 284 (87.4) | 291 (88.7) Clinical T stage at study entry, n (%)^c cT2 | 249 (76.6) | 254 (77.4) cT3 | 56 (17.2) | 49 (14.9) cT4a | 18 (5.5) | 22 (6.7) Minor histological variants present, n (%) | 180 (55.4) | 180 (54.9) PD-L1 ≥ 1%, n (%)^d | 153 (47.1) | 155 (47.3) • Baseline demographic and disease characteristics were generally balanced between arms • Most patients had cT2 disease at study entry and around a half of patients had tumor PD-L1 expression ≥ 1% Randomized from September 17, 2019, to November 7, 2022. ^a"Other" category includes Black or African American, American Indian/Alaska native, and other. Race was not reported in 13 patients (2.0%). ^bCalculated GFR was not reported in 5 patients (0.8%). Nine patients with baseline calculated GFR <50 mL/min were classified as having a relevant protocol deviation at study entry. ^cClinical T stage at study entry was not reported in 5 patients (0.8%). ^dFour patients (1.2%) in the NIVO+GC arm were not evaluable for PD-L1 status. [Slide 4] CA017-078 Overall survival by treatment arm | | NIVO+GC (N = 325) | GC (N = 328) | # of events/# of patients (%) | 71/325 (21.8) | 81/328 (24.7) Median OS (95% CI), mo | NR (NE-NE) | NR (NE-NE) HR (95% CI) | 0.88 (0.64-1.21) Y-axis: Probability of overall survival (%); X-axis: Months Landmark labels on curve — NIVO+GC: 91.6% (12 mo), 84.8% (24 mo), 79.8% (36 mo), 79.8% (dashed line between 36 and 42 mo); GC: 92.8% (12 mo), 85.2% (24 mo), 80.5% (36 mo), 78.4% (dashed line between 36 and 42 mo) No. at risk (months 0, 6, 12, 18, 24, 30, 36, 42, 48, 54, 60, 66, 72, 78) NIVO+GC: 325 305 281 268 257 248 238 204 148 113 72 36 16 0 GC: 328 299 282 269 255 249 238 205 149 105 66 32 17 0 • Median OS was not reached in either treatment arm • OS was analyzed descriptively because the EFS testing condition was not met Survival table Endpoint: OS (secondary, descriptive) Arm | N | Median | Median CI NIVO+GC | 325 | NR | NE-NE GC | 328 | NR | NE-NE HR: 0.88 HR CI: 95% CI 0.64-1.21 Landmark rates: 12 mo: NIVO+GC 91.6% | GC 92.8% 24 mo: NIVO+GC 84.8% | GC 85.2% 36 mo: NIVO+GC 79.8% | GC 80.5% unlabeled dashed line between 36 and 42 mo: NIVO+GC 79.8% | GC 78.4%

Neoadjuvant ctDNA clearance and pCR CA017-078 ctDNA clearance (neoadjuvant C1D1 to C4D1) Fisher exact P = 0.130 GC: 71.7% (66/92) NIVO+GC: 81.3% (87/107) [y-axis: Cleared ctDNA at C4D1 (%)] Bars show % of C1D1 ctDNA-positive patients negative at C4D1; whiskers are Wilson 95% CIs. pCR by ctDNA clearance status (neoadjuvant C1D1 to C4D1) pCR rate by treatment arm, faceted by C1D1 to C4D1 ctDNA clearance status ctDNA cleared by C4D1, n = 153: Fisher exact P = 0.242; GC 33.3% (22/66); NIVO+GC 43.7% (38/87) ctDNA not cleared by C4D1, n = 46: Fisher exact P = 1.000; GC 7.7% (2/26); NIVO+GC 5.0% (1/20) [y-axis: Pathologic complete response rate (%)] • ctDNA clearance was higher with NIVO+GC vs GC alone • In both treatment arms, ctDNA clearance was associated with a higher pCR rate • Among patients with ctDNA clearance, pCR rates were higher with NIVO+GC vs GC alone ctDNA testing was performed using the Signatera tumor-informed molecular residual disease assay (Natera, Austin, TX, USA). ctDNA clearance status was categorized as positive at C1D1 but negative by C4D1 (cleared), or positive at both C1D1 and C4D1 (not cleared). The ctDNA-evaluable population (patients with ctDNA results for both C1D1 and C4D1) comprised 53.0% (346/653) of all randomly assigned patients. The ctDNA-evaluable population was representative of the total study population across all demographic features and baseline demographic characteristics were well balanced between the treatment arms.

[Slide 3] CA017-078 Event-free survival by treatment arm (co-primary endpoint) [Kaplan-Meier plot; y-axis: Probability of event-free survival per BICR (%); x-axis: Months 0-78] Landmark labels on curves: NIVO+GC 82.6%, 72.3%, 68.1%, 67.3%; GC 78.6%, 71.6%, 68.5%, 65.8% Table: | NIVO+GC (N = 325) | GC (N = 328) # of events/# of patients (%): 104/325 (32.0) | 104/328 (31.7) Median EFS (95% CI), mo: NR (NE-NE) | 67.6 (63.0-NE) HR (95% CI): 0.94 (0.72-1.24) HR (95.74% CI): 0.94 (0.71-1.25) P value^a: 0.6697 No. at risk NIVO+GC: 325 283 243 227 209 197 189 146 100 60 25 5 0 0 GC: 328 271 219 199 187 176 168 131 90 61 23 4 1 0 • Median EFS was not reached in the NIVO+GC arm and was 67.6 months (95% CI 63.0-NE) in the GC arm • The EFS co-primary endpoint did not meet the significance criterion determined by the prespecified testing procedure ^aTwo-sided P value from log-rank test stratified by PD-L1 status, GFR category, and T stage at screening as entered into the IRT. P value threshold for statistical significance, 0.0426. Survival table Endpoint: EFS per BICR (co-primary endpoint) Arm | N | Median | Median CI NIVO+GC | 325 | NR | NE-NE GC | 328 | 67.6 months | 63.0-NE HR: 0.94 HR CI: 0.72-1.24 (95% CI); 0.71-1.25 (95.74% CI) P: 0.6697 (threshold 0.0426; not significant) Landmark rates: unlabeled dashed line (~12 mo): NIVO+GC 82.6% | GC 78.6% unlabeled dashed line (~24 mo): NIVO+GC 72.3% | GC 71.6% unlabeled dashed line (~36 mo): NIVO+GC 68.1% | GC 68.5% unlabeled dashed line (~38-39 mo): NIVO+GC 67.3% | GC 65.8% [Slide 4] CA017-078 EFS and OS by pre-treatment ctDNA status (pooled)^a EFS by pretreatment ctDNA status [Kaplan-Meier plot; y-axis: Survival probability 0.5-1.0; x-axis: Months 0-80] Legend: C1D1 positive; C1D1 negative No. at risk C1D1 positive: 229 164 136 118 99 45 10 1 0 C1D1 negative: 187 149 132 118 94 53 14 1 0 OS by pretreatment ctDNA status [Kaplan-Meier plot; y-axis 0.6-1.0; x-axis: Months 0-80] No. at risk C1D1 positive: 229 202 175 158 128 62 24 2 0 C1D1 negative: 187 168 160 153 120 68 17 1 0 • In a pooled analyses across treatment arms, pretreatment ctDNA detection was associated with shorter EFS and OS ^aBased on pooled analysis of all evaluable patients in the NIVO+GC and GC arms. ctDNA testing was performed using the Signatera tumor-informed molecular residual disease assay (Natera, Austin, TX, USA). Pretreatment ctDNA status was categorized as positive or negative based on detectable ctDNA at C1D1. Survival table Endpoint: EFS and OS by pretreatment (C1D1) ctDNA status, pooled across arms C1D1 ctDNA positive | 229 C1D1 ctDNA negative | 187

[Slide 2] Pathological complete response by treatment arm pCR overall (co-primary endpoint)^a: Overall ITT NIVO+GC 39.1 vs GC 29.6; difference 9.5% (95% CI 2.3-16.7) (99.91% CI -2.6 to 21.6)^b pCR by subgroups: PD-L1 ≥1%: 46.4 vs 35.5; 10.9% (95% CI -0.1 to 21.5) PD-L1 <1%: 33.3 vs 24.3; 9.1% (95% CI -0.6 to 18.5) cT2: 40.6 vs 31.5; 9.1% (95% CI 0.7-17.3) cT3/cT4a: 35.1 vs 23.9; 11.2% (95% CI -3.7 to 25.3) Combined response (≤ ypT1N0): Overall ITT 51.7 vs 41.8; 9.9% (95% CI 2.3-17.5) Patients (%) Legend: NIVO+GC; GC [Slide 3] Event-free survival by treatment arm (co-primary endpoint) | NIVO+GC (N = 325) | GC (N = 328) # of events/# of patients (%) | 104/325 (32.0) | 104/328 (31.7) Median EFS (95% CI), mo | NR (NE-NE) | 67.6 (63.0-NE) HR (95% CI) | 0.94 (0.72-1.24) HR (95.74% CI) | 0.94 (0.71-1.25) P value^a | 0.6697 Landmarks: NIVO+GC 82.6%, 72.3%, 68.1%, 67.3%; GC 78.6%, 71.6%, 68.5%, 65.8% Probability of event-free survival per BICR (%) vs Months No. at risk (0,6,...,78) NIVO+GC: 325 283 243 227 209 197 189 146 100 60 25 5 0 0 GC: 328 271 219 199 187 176 168 131 90 61 23 4 1 0 - Median EFS was not reached in the NIVO+GC arm and was 67.6 months (95% CI 63.0-NE) in the GC arm - The EFS co-primary endpoint did not meet the significance criterion determined by the prespecified testing procedure Survival table Endpoint: EFS per BICR (co-primary) Arm | N | Median | Median CI NIVO+GC | 325 | NR | NE-NE GC | 328 | 67.6 | 63.0-NE HR: 0.94 HR CI: 95% CI 0.72-1.24; 95.74% CI 0.71-1.25 P: 0.6697 Landmark rates: 12 mo: NIVO+GC 82.6% | GC 78.6% 24 mo: NIVO+GC 72.3% | GC 71.6% 36 mo: NIVO+GC 68.1% | GC 68.5% 4th dashed line (timepoint not printed): NIVO+GC 67.3% | GC 65.8% [Slide 4] CA017-078 Neoadjuvant ctDNA clearance and pCR ctDNA clearance (neoadjuvant C1D1 to C4D1) Fisher exact P = 0.130 Cleared ctDNA at C4D1 (%): GC 71.7% (66/92); NIVO+GC 81.3% (87/107) Bars show % of C1D1 ctDNA-positive patients negative at C4D1; whiskers are Wilson 95% CIs. pCR by ctDNA clearance status (neoadjuvant C1D1 to C4D1) pCR rate by treatment arm, faceted by C1D1 to C4D1 ctDNA clearance status ctDNA cleared by C4D1 n = 153: Fisher exact P = 0.242; GC 33.3% (22/66); NIVO+GC 43.7% (38/87) ctDNA not cleared by C4D1 n = 46: Fisher exact P = 1.000; GC 7.7% (2/26); NIVO+GC 5.0% (1/20) Pathologic complete response rate (%)

Perioperative nivolumab plus neoadjuvant GC in resectable MIBC: final results and ctDNA analysis CA017-078 (ENERGIZE; NCT03661320): randomized phase III trial of nivolumab + GC → radical cystectomy → adjuvant nivolumab (n=325) versus GC → radical cystectomy (n=328) Key inclusion criteria • Adults ≥18 years with previously untreated cT2-T4aN0/1M0 MIBC • Fit for radical cystectomy and eligible for cisplatin (creatinine clearance ≥50 mL/min) • Eastern Cooperative Oncology Group performance status 0-1 Study highlights • Perioperative nivolumab did not significantly improve pCR or EFS; neither co-primary endpoint was met • Grade 3-4 treatment-related AEs: 50.0% vs 49.5% • There was one treatment-related death with nivolumab + GC • ctDNA presence was prognostic in exploratory analyses Key clinical outcomes Outcome | Nivolumab + GC (n=325) | GC (n=328) | Comparison pCR | 39.1% | 29.6% | Adjusted difference: 9.5% (99.91% CI: −2.6 to 21.6) Median EFS (mo) | Not reached | 67.6 | HR 0.94 (95.74% CI: 0.71-1.25); P=0.6697 Median OS (mo) | Not reached | Not reached | HR 0.88 (95% CI: 0.64-1.21); descriptive only Exploratory ctDNA analysis Pretreatment ctDNA detection EFS by C1D1 ctDNA detection (pooled) [KM curve: C1D1 positive vs C1D1 negative] No. at risk [rows present; small print, see notes] Median survival Group | Median (95% CI) | N | Events C1D1 positive | NR (44.7-NR) | 229 | 95 C1D1 negative | NR (NR-NR) | 187 | 37 OS by C1D1 ctDNA detection (pooled) [KM curve] Median survival Group | Median (95% CI) | N | Events C1D1 positive | NR (NR-NR) | 229 | 69 C1D1 negative | NR (NR-NR) | 187 | 28 C1D1/C4D1 ctDNA trajectory EFS by C1D1/C4D1 trajectory (pooled) [KM curve: Negative C1D1+C4D1, Cleared by C4D1, Positive at C4D1] Median survival Group | Median (95% CI) | N | Events Negative C1D1+C4D1 | NR (NR-NR) | 147 | 25 Cleared by C4D1 | NR (NR-NR) | 153 | 51 Positive at C4D1 | 8.5 (6.7-11.2) | 46 | 35 OS by C1D1/C4D1 trajectory (pooled) [KM curve] Median survival Group | Median (95% CI) | N | Events Negative C1D1+C4D1 | NR (NR-NR) | 147 | 19 Cleared by C4D1 | NR (NR-NR) | 153 | 35 Positive at C4D1 | 28.1 (16.4-NR) | 46 | 26 AE, adverse event; C1D1/C4D1, cycle 1/4 day 1; CI, confidence interval; ctDNA, circulating tumor DNA; EFS, event-free survival; GC, gemcitabine-cisplatin; HR, hazard ratio; MIBC, muscle-invasive bladder cancer; OS, overall survival; pCR, pathological complete response.

Primary Endpoint: EFS by Investigator Assessment ITT Population [Table] | | Pembro + cis + gem n = 453 | Placebo + cis + gem n = 454 | Events, n (%) | 157 (34.7) | 168 (37.0) Median (95% CI), mo | 74.1 (72.4-NR) | NR (67.7-NR) HR 0.88 (95% CI, 0.71-1.10), 1-sided P = 0.1366* Y-axis: Event-Free Survival, % Curve labels: Pembro 76.7%, 69.2%, 66.0%, 63.5%; Placebo 70.0%, 66.3%, 62.3%, 61.9% X-axis: Months No. at risk (0, 6, 12, 18, 24, 30, 36, 42, 48, 54, 60, 66, 72, 78) Pembro + cis + gem: 453 401 366 338 329 304 296 273 224 175 114 51 11 0 Placebo + cis + gem: 454 400 345 310 300 291 283 257 208 168 106 51 12 0 NR, not reached. *Not statistically significant (1-sided boundary 0.0225). Data cutoff: January 26, 2026. Survival table Endpoint: EFS by investigator assessment (primary), ITT Arm | N | Median | Median CI Pembro + cis + gem | 453 | 74.1 mo | 72.4-NR Placebo + cis + gem | 454 | NR | 67.7-NR HR: 0.88 HR CI: 0.71-1.10 P: 1-sided P = 0.1366 (not significant; boundary 0.0225) Landmark rates: 24 mo (dashed line at axis tick): Pembro + cis + gem 76.7% | Placebo + cis + gem 70.0% 36 mo (dashed line at axis tick): Pembro + cis + gem 69.2% | Placebo + cis + gem 66.3% 48 mo (dashed line at axis tick): Pembro + cis + gem 66.0% | Placebo + cis + gem 62.3% 60 mo (dashed line at axis tick): Pembro + cis + gem 63.5% | Placebo + cis + gem 61.9%

[Slide 1] An official website of the United States government Here's how you know FDA Search Menu Resources for Information | Approved Drugs IN THIS SECTION FDA approves talazoparib with enzalutamide for metastatic androgen pathway modulation-naïve or -sensitive prostate cancer with HRR gene mutations [Slide 3] 64 By Gene Analysis of rPFS In post hoc analyses, rPFS generally favored TALA + ENZA across HRR gene subgroups TALA + ENZA | PBO + ENZA No. of Events/No. of Patients | HR (95% CI) All HRR gene-altered patients based on all HRR tests 67/300 126/297 0.482 (0.358–0.648) ATM 16/80 33/77 0.433 (0.238–0.786) ATR 3/19 2/14 0.786 (0.131–4.711) BRCA1 5/15 6/12 0.706 (0.215–2.321) BRCA2 17/87 45/97 0.353 (0.202–0.617) CDK12 8/51 29/55 0.275 (0.125–0.603) CHEK2 12/44 19/44 0.626 (0.304–1.290) FANCA 6/10 1/6 4.277 (0.512–35.752) MLH1 0/3 1/1 NE MRE11A 0/0 2/2 NE NBN 1/5 1/4 NE PALB2 0/11 4/6 NE RAD51C 0/0 1/4 NE 0.1 1 3 6 9 Favors TALA + ENZA | Favors PBO + ENZA Data cutoff: February 18, 2026 (ITT population). Analyses included all available tumor and baseline ctDNA records. For non-BRCAm, those with co-occurring BRCA1/2m were excluded. For BRCA1m, those with co-occurring [illegible] excluded. Unstratified HRs are based on a Cox proportional hazards model, with treatment as the only covariate. ctDNA=circulating tumor DNA; NE=not evaluable. 2026 ASCO ANNUAL MEETING #ASCO26 PRESENTED BY: Neeraj Agarwal

[Slide 1] TALAPRO-3: Trial Design Key eligibility criteria • HRR gene-altered mCSPC • Metastatic disease confirmed by positive bone scan or metastatic lesion by CT or MRI scan • ECOG PS 0 or 1 • Ongoing ADT • ≤3 months of prior ADT with/without ARPI for mCSPC • Prior docetaxel for mCSPC not permitted† Stratification factors • De novo vs relapsed mCSPC • High- vs low-volume disease • BRCAm vs non-BRCAm 1:1 N=599 TALA 0.5 mg* + ENZA 160 mg QD PBO + ENZA 160 mg QD Data cutoff: February 18, 2026 Primary endpoint • rPFS by investigator assessment Key secondary endpoint • OS (alpha-protected) Other secondary endpoints • ORR per RECIST v1.1 • Duration of soft tissue response • PSA response • Time to PSA progression • Time to subsequent antineoplastic therapy • Time to first SSE • Safety • PROs Exploratory endpoint • PFS2 HRRm: ATM, ATR, BRCA1, BRCA2, CDK12, CHEK2, FANCA, MLH1, MRE11A, NBN, PALB2, RAD51C‡ *0.35 mg QD if the patient has moderate renal impairment; †As per the protocol amendment (September 2021). ClinicalTrials.gov identifier: NCT04821622. ‡HRR testing using FoundationOne®/FoundationOne®CDx and FoundationOne®Liquid CDx. BRCAm=BRCA alterations; CT=computed tomography; ECOG PS=Eastern Cooperative Oncology Group performance status; HRRm=HRR gene alterations; MRI=magnetic resonance imaging; ORR=objective response rate; [illegible] PRESENTED BY: Neeraj Agarwal 2026 ASCO ANNUAL MEETING #ASCO26 [Slide 2] Primary Endpoint: Investigator-Assessed rPFS TALA + ENZA led to a 52% reduction in the risk of progression or death Probability of rPFS (95% CI) Months No. at Risk TALA + ENZA PBO + ENZA 76.6% (70.8–81.4) 56.2% (50.0–62.0) HR, 0.481 (95% CI, 0.357–0.647); p<0.0001 TALA + ENZA (n=300) | PBO + ENZA (n=299) No. of events | 67 | 126 Median (95% CI), mo | NR (NR–NR) | 45.8 (37.7–NR) Median follow-up for rPFS was 37.6 and 37.7 months, respectively Data cutoff: February 18, 2026 (ITT population). Stratified HRs and 2-sided p values are reported throughout this presentation unless otherwise stated. NR=not reached. PRESENTED BY: Neeraj Agarwal Survival table Endpoint: Investigator-assessed rPFS (ITT) Arm | N | Median | Median CI TALA + ENZA | 300 | NR | NR–NR PBO + ENZA | 299 | 45.8 mo | 37.7–NR HR: 0.481 HR CI: 0.357–0.647 P: <0.0001 36 mo: TALA + ENZA 76.6% (70.8–81.4) vs PBO + ENZA 56.2% (50.0–62.0) [Slide 3] rPFS in the BRCAm vs non-BRCAm Subgroups A 63% risk reduction was observed in the BRCAm subgroup and 43% in the non-BRCAm subgroup BRCA1/2m 77.2% (66.9–84.7) 48.8% (38.0–58.7) TALA + ENZA (n=104) | PBO + ENZA (n=103) No. of events | 22 | 49 Median (95% CI), mo | NR (NR–NR) | 35.1 (18.6–NR) HR (95% CI) | 0.368 (0.222–0.609); p<0.0001 Non-BRCA1/2m 76.2% (68.8–82.1) 60.2% (52.5–67.0) TALA + ENZA (n=196) | PBO + ENZA (n=196) No. of events | 45 | 77 Median (95% CI), mo | NR (NR–NR) | NR (40.5–NR) HR (95% CI) | 0.567 (0.392–0.819); p=0.0022 Probability of rPFS (95% CI) Months No. at Risk Data cutoff: February 18, 2026 (ITT population). The BRCAm and non-BRCAm subgroups by randomization stratification. HRs are unstratified. PRESENTED BY: Neeraj Agarwal Survival table Endpoint: Investigator-assessed rPFS by BRCA subgroup Arm | N | Median | Median CI BRCA1/2m: TALA + ENZA | 104 | NR | NR–NR BRCA1/2m: PBO + ENZA | 103 | 35.1 mo | 18.6–NR Non-BRCA1/2m: TALA + ENZA | 196 | NR | NR–NR Non-BRCA1/2m: PBO + ENZA | 196 | NR | 40.5–NR HR: BRCA1/2m 0.368; non-BRCA1/2m 0.567 HR CI: BRCA1/2m 0.222–0.609; non-BRCA1/2m 0.392–0.819 P: BRCA1/2m <0.0001; non-BRCA1/2m 0.0022 36 mo BRCA1/2m: 77.2% (66.9–84.7) vs 48.8% (38.0–58.7) 36 mo non-BRCA1/2m: 76.2% (68.8–82.1) vs 60.2% (52.5–67.0) [Slide 4] Overall Survival (Interim Analysis) Overall survival at this interim analysis did not cross the efficacy boundary Probability of OS Months No. at Risk TALA + ENZA PBO + ENZA 77.8% (72.5–82.1) 71.6% (66.0–76.4) HR=0.767 (95% CI, 0.564–1.044); P=0.0905 TALA + ENZA (N=300) | PBO + ENZA (N=299) Events, n | 74 | 91 Median (95% CI), months | NR (NR–NR) | NR (NR–NR) Median follow-up for OS was 40.5 and 41.1 months, respectively Data cutoff: February 18, 2026. Intention-to-Treat Population. PRESENTED BY: Neeraj Agarwal Survival table Endpoint: Overall survival (interim, ITT) Arm | N | Median | Median CI TALA + ENZA | 300 | NR | NR–NR PBO + ENZA | 299 | NR | NR–NR HR: 0.767 HR CI: 0.564–1.044 P: 0.0905 36 mo: TALA + ENZA 77.8% (72.5–82.1) vs PBO + ENZA 71.6% (66.0–76.4)

FDA approves talazoparib with enzalutamide for metastatic androgen pathway modulation-naïve or -sensitive prostate cancer with HRR gene mutations On October 26, 2026, the Food and Drug Administration (FDA) approved talazoparib (Talzenna, Pfizer Inc.) in combination with enzalutamide for the treatment of adults with metastatic androgen pathway modulation-naïve or -sensitive (mAPMN/S) prostate cancer with homologous recombination repair (HRR) gene mutations, as detected by an FDA-authorized test. The FDA previously approved FoundationOne CDx and FoundationOne Liquid CDx assays as companion diagnostic devices to identify patients with prostate cancer with HRR gene alterations for treatment with talazoparib in combination with enzalutamide. Feedback

[Slide 1] Baseline demographics and treatment duration (PRO population)* TALA + ENZA (n=240) | PBO + ENZA (n=234) Median age (IQR), years: 70 (63–76) | 68.5 (63–75) Race, n (%) White: 123 (51) | 128 (55) Black or African American: 4 (2) | 2 (1) Asian: 103 (43) | 98 (42) American Indian or Alaska Native†: 4 (2) | 2 (1) Not reported: 6 (2) | 4 (2) Geographical region, n (%) Asia: 101 (42) | 98 (42) Europe: 91 (38) | 100 (43) North America: 34 (14) | 24 (10) Rest of the world: 14 (6) | 12 (5) Median treatment duration, months Talazoparib/placebo: 35.0 | 34.1 Enzalutamide: 35.9 | 34.0 Baseline PRO Completion Rates‡ (TALA + ENZA | PBO + ENZA) EORTC QLQ-C30: 68% | 66% EORTC QLQ-PR25: 66% | 66% BPI-SF: 80% | 79% EQ-5D-5L: 65% | 65% [Slide 2] Comparable mean change from baseline in GHS/QoL per EORTC QLQ-C30 EORTC QLQ-C30 Score Range: 0–100 Estimated overall mean change from baseline difference (95% CI): -3.7 (-6.5 to -1.0)* Observed mean change from baseline, 95% CI (Improved ↑ / Worsened ↓) Assessment visit (week): 5 9 13 17 21 25 33 41 49 57 69 81 93 105 117 129 141 153 165 177 189 201 213 Number evaluable TALA + ENZA: 183 173 171 173 164 172 166 161 154 148 144 136 131 129 120 117 113 103 75 50 34 18 9 PBO + ENZA: 169 164 170 167 162 168 156 153 151 146 139 134 126 117 109 106 99 87 60 39 29 20 10 [Slide 3] [Slide title cropped at top - illegible] EORTC QLQ-C30 Score Range: 0–100 Estimated overall mean change from baseline difference (95% CI): 5.2 (2.5–8.0)* Observed mean change from baseline, 95% CI (Worsened ↑ / Improved ↓) Assessment visit (week): 5 9 13 17 21 25 33 41 49 57 69 81 93 105 117 129 141 153 165 177 189 201 213 Number evaluable TALA + ENZA: 183 173 171 173 164 172 166 161 154 148 144 136 131 129 120 117 113 103 75 50 34 18 9 PBO + ENZA: 169 164 170 167 162 168 156 153 151 146 139 134 126 117 109 106 99 87 60 39 29 20 10 [Slide 4] GHS/QoL per EORTC QLQ-C30* - Time to definitive deterioration (months) TALA + ENZA: events/patients 92/240; median 38.0 (35.1–46.0) mo PBO + ENZA: 78/234; median 46.0 (37.8–49.9) mo Stratified HR (95% CI): 1.16 (0.85–1.58) Urinary symptoms per EORTC QLQ-PR25† - Time to definitive deterioration (months) TALA + ENZA: 42/240; median NR (48.8–NR) PBO + ENZA: 46/234; median 51.4 (48.8–NR) Stratified HR (95% CI): 0.94 (0.61–1.44) Worst pain per BPI-SF‡ - Time to deterioration (months) TALA + ENZA: 81/240; median NR (32.9–NR) PBO + ENZA: 93/234; median NR (23.0–NR) Stratified HR (95% CI): 0.82 (0.61–1.11) Event-free probability Survival table Endpoint: Time to definitive deterioration in GHS/QoL (EORTC QLQ-C30) Arm | N | Median | Median CI TALA + ENZA | 240 | 38.0 mo | 35.1–46.0 PBO + ENZA | 234 | 46.0 mo | 37.8–49.9 HR: 1.16 HR CI: 0.85–1.58

[Slide 1] Figure 2. PSA Progression for Alterations in Select* HRR12 Genes (All Randomized Patients) Subgroup | TALA + ENZA Events/n | PBO + ENZA Events/n | HR^† (95% CI) | 2-sided p-value All HRRm | 58/300 | 96/299 | 0.51 (0.37-0.71) | <0.0001 ATM | 18/80 | 25/77 | 0.61 (0.34-1.13) | 0.11 ATR | 2/19 | 5/14 | 0.25 (0.05-1.30) | 0.075 BRCA1 | 2/15 | 4/12 | 0.33 (0.06-1.82) | 0.19 BRCA2 | 15/85 | 30/96 | 0.45 (0.24-0.83) | 0.0092 CDK12 | 11/50 | 26/55 | 0.42 (0.20-0.84) | 0.012 CHEK2 | 6/43 | 12/44 | 0.49 (0.18-1.30) | 0.14 [Forest plot axis 0.1, 1.0, 10.0; <- Favors TALA + ENZA | Favors PBO + ENZA ->] [Slide 2] Figure 3. Patients With PSA <0.2 µg/L (A) at 6 Months and (B) at 12 Months and With Alterations in Select* HRR12 Genes (Randomized Patients With Baseline PSA) A. 6 months Subgroup | TALA + ENZA | PBO + ENZA All patients | 55.0% n/N=165/300 | 53.4% n/N=159/298 ATM | 62.5% n/N=50/80 | 62.3% n/N=48/77 ATR | 78.9% n/N=15/19 | 50.0% n/N=7/14 BRCA1 | 60.0% n/N=9/15 | 33.3% n/N=4/12 BRCA2 | 49.4% n/N=43/87 | 49.5% n/N=48/97 CDK12 | 49.0% n/N=25/51 | 40.0% n/N=22/55 CHEK2 | 50.0% n/N=22/44 | 54.5% n/N=24/44 B. 12 months Subgroup | TALA + ENZA | PBO + ENZA All patients | 58.3% n/N=175/300 | 50.3% n/N=150/298 ATM | 65.0% n/N=52/80 | 58.4% n/N=45/77 ATR | 84.2% n/N=16/19 | 64.3% n/N=9/14 BRCA1 | 53.3% n/N=8/15 | 25.0% n/N=3/12 BRCA2 | 57.5% n/N=50/87 | 42.3% n/N=41/97 CDK12 | 51.0% n/N=26/51 | 41.8% n/N=23/55 CHEK2 | 54.5% n/N=24/44 | 59.1% n/N=26/44 [Slide 3] Figure 4. Kaplan-Meier Plot of rPFS Based on Investigator Assessment According to PSA <0.2 µg/L or PSA <0.02 µg/L at 6 Months in Patients Treated With (A) TALA + ENZA and (B) PBO + ENZA (Randomized Patients With Baseline PSA) A. TALA + ENZA | n | Events | Median, mo | 95% CI Achieved PSA <0.2 µg/L at 6 months | 165 | 27 | NR | NR-NR Achieved PSA <0.02 µg/L at 6 months | 74 | 15 | NR | NR-NR Not achieved PSA <0.2 µg/L at 6 months | 135 | 40 | NR | 40.4-NR Not achieved PSA <0.02 µg/L at 6 months | 226 | 52 | NR | NR-NR Unstratified (achieved vs not achieved at 6 months): PSA <0.2 µg/L HR=0.37; 95% CI, 0.22-0.60; p<0.0001 PSA <0.02 µg/L HR=0.69; 95% CI, 0.39-1.23; p=0.21 No. at risk (0-52 months, every 2): Achieved PSA <0.2 µg/L: 165 162 162 160 158 157 155 151 145 144 142 137 133 132 129 123 121 109 84 66 65 42 23 14 10 1 0 Achieved PSA <0.02 µg/L: 74 74 74 74 73 73 71 70 67 66 65 63 61 61 59 55 54 47 37 31 31 15 10 4 3 0 0 Not achieved PSA <0.2 µg/L: 135 120 109 102 102 91 86 81 77 75 71 69 63 63 60 59 57 54 42 33 31 21 9 6 4 2 0 Not achieved PSA <0.02 µg/L: 226 208 197 188 187 175 170 162 155 153 148 143 135 134 130 127 124 116 89 68 65 48 22 11 11 3 0 B. PBO + ENZA | n | Events | Median, mo | 95% CI Achieved PSA <0.2 µg/L at 6 months | 159 | 43 | NR | NR-NR Achieved PSA <0.02 µg/L at 6 months | 70 | 10 | NR | NR-NR Not achieved PSA <0.2 µg/L at 6 months | 139 | 83 | 18.3 | 13.0-26.7 Not achieved PSA <0.02 µg/L at 6 months | 228 | 116 | 29.5 | 23.7-40.5 Unstratified (achieved vs not achieved at 6 months): PSA <0.2 µg/L HR=0.23; 95% CI, 0.16-0.34; p<0.0001 PSA <0.02 µg/L HR=0.17; 95% CI, 0.09-0.33; p<0.0001 No. at risk (0-52 months, every 2): Achieved PSA <0.2 µg/L: 159 159 159 156 153 152 149 146 144 142 138 133 130 130 123 115 115 103 75 56 56 40 26 17 10 4 0 Achieved PSA <0.02 µg/L: 70 70 70 70 68 68 68 68 68 68 68 67 67 67 64 62 62 55 40 29 29 23 14 9 4 2 0 Not achieved PSA <0.2 µg/L: 139 125 117 99 90 80 73 66 59 58 50 46 45 44 40 36 36 29 23 18 18 9 6 4 2 2 0 Not achieved PSA <0.02 µg/L: 228 214 206 185 175 164 154 144 135 132 120 112 108 107 99 89 89 77 58 45 45 26 18 12 8 4 0 Survival table Endpoint: rPFS (investigator) by PSA response at 6 months Arm | N | Median | Median CI TALA+ENZA: Achieved PSA <0.2 | 165 | NR | NR-NR TALA+ENZA: Achieved PSA <0.02 | 74 | NR | NR-NR TALA+ENZA: Not achieved PSA <0.2 | 135 | NR | 40.4-NR TALA+ENZA: Not achieved PSA <0.02 | 226 | NR | NR-NR PBO+ENZA: Achieved PSA <0.2 | 159 | NR | NR-NR PBO+ENZA: Achieved PSA <0.02 | 70 | NR | NR-NR PBO+ENZA: Not achieved PSA <0.2 | 139 | 18.3 | 13.0-26.7 PBO+ENZA: Not achieved PSA <0.02 | 228 | 29.5 | 23.7-40.5 HR: TALA+ENZA: <0.2 HR 0.37, <0.02 HR 0.69; PBO+ENZA: <0.2 HR 0.23, <0.02 HR 0.17 (achieved vs not achieved, unstratified) HR CI: 0.22-0.60; 0.39-1.23; 0.16-0.34; 0.09-0.33 P: <0.0001; 0.21; <0.0001; <0.0001 [Slide 4] Table 2. rPFS and Time to PSA Progression According to PSA <0.2 µg/L or PSA <0.02 µg/L at 12 Months by Treatment Arm (Randomized Patients With Baseline PSA) Endpoint | TALA + ENZA (n=300) | PBO + ENZA (n=298) HR (95% CI)*; p-value^† rPFS Patients achieving vs those not achieving PSA <0.2 µg/L | 0.22 (0.14-0.37); p<0.0001 | 0.16 (0.11-0.24); p<0.0001 Patients achieving vs those not achieving PSA <0.02 µg/L | 0.37 (0.21-0.67); p=0.0006 | 0.15 (0.08-0.29); p<0.0001 Time to PSA progression Patients achieving vs those not achieving PSA <0.2 µg/L | 0.12 (0.07-0.23); p<0.0001 | 0.09 (0.06-0.15); p<0.0001 Patients achieving vs those not achieving PSA <0.02 µg/L | 0.08 (0.02-0.26); p<0.0001 | 0.04 (0.01-0.16); p<0.0001 Patients without PSA assessments on Week 49 were considered as not having achieved PSA <0.2 µg/L or <0.02 µg/L at 12 months. *HR based on Cox proportional hazards model; HR <1 indicates reduction in HR in favor of patients achieving vs those not achieving (1) PSA <0.2 µg/L, or (2) PSA <0.02 µg/L, at 12 months; ^†p-value calculated using an unstratified log-rank test. CI=confidence interval; ENZA=enzalutamide; HR=hazard ratio; PBO=placebo; PSA=prostate-specific antigen; rPFS=radiographic progression-free survival; TALA=talazoparib.

[Slide 2] Baseline ctDNA Burden Was Prognostic of rPFS in Both Arms | n | Events | Median, mo | 95% CI A. TALA + ENZA ctDNA+ at BL | 83 | 24 | NR | NR–NR B. TALA + ENZA ctDNA− at BL | 147 | 32 | NR | NR–NR C. PBO + ENZA ctDNA+ at BL | 75 | 40 | 26.7 | 18.2–NR D. PBO + ENZA ctDNA− at BL | 149 | 50 | NR | 45.8–NR [KM curve: Probability of PFS vs Progression-Free Survival Time (Months), 0-52; curve labels: TALA + ENZA ctDNA− at BL; TALA + ENZA ctDNA+ at BL; PBO + ENZA ctDNA− at BL; PBO + ENZA ctDNA+ at BL] Unstratified HR (A vs B)=1.59 (95% CI, 0.94–2.70); p=0.08 HR (A vs C)=0.47 (95% CI, 0.28–0.78); p=0.003 HR (B vs D)=0.62 (95% CI, 0.39–0.96); p=0.03 HR (C vs D)=2.15 (95% CI, 1.42–3.26); p=0.0002 No. at risk (months 0,2,4,...,52) A: 83 80 74 69 69 60 57 54 52 51 49 48 47 46 44 42 42 39 30 24 23 17 8 4 3 1 0 B: 147 139 136 133 132 131 127 123 117 115 112 108 100 100 98 95 93 84 66 51 50 33 18 13 9 2 0 C: 75 69 65 59 52 49 46 44 42 42 38 36 33 32 29 26 26 21 15 13 13 9 4 3 2 1 0 D: 149 144 143 131 128 125 121 116 114 113 107 102 102 102 95 88 88 76 55 38 38 25 19 14 8 4 0 Data cutoff: February 18, 2026 Hazard ratio based on Cox proportional hazards model. P-value from an unstratified log-rank test. *Intention-to-treat population (ITT) excluding patients in China; rPFS by investigator assessment. NR=not reached. Survival table Endpoint: rPFS by baseline ctDNA status (investigator-assessed) Arm | N | Median | Median CI A. TALA + ENZA ctDNA+ at BL | 83 | NR | NR–NR B. TALA + ENZA ctDNA− at BL | 147 | NR | NR–NR C. PBO + ENZA ctDNA+ at BL | 75 | 26.7 | 18.2–NR D. PBO + ENZA ctDNA− at BL | 149 | NR | 45.8–NR HR: A vs B 1.59; A vs C 0.47; B vs D 0.62; C vs D 2.15 HR CI: A vs B 0.94–2.70; A vs C 0.28–0.78; B vs D 0.39–0.96; C vs D 1.42–3.26 P: A vs B 0.08; A vs C 0.003; B vs D 0.03; C vs D 0.0002 [Slide 3] Conversion from ctDNA+ to ctDNA- and persistent ctDNA- at Week 5 was associated with longer rPFS in the TALA + ENZA Arm | n | Events | Median, mo | 95% CI A. TALA + ENZA ctDNA+ at BL and Wk 5 | 12 | 6 | 17.3 | 9.1–NR B. TALA + ENZA ctDNA+ at BL and ctDNA− at Wk 5 | 55 | 14 | NR | NR–NR C. TALA + ENZA ctDNA− at BL and Wk 5 | 114 | 25 | NR | NR–NR D. TALA + ENZA ctDNA− at BL and ctDNA+ at Wk 5 | 4 | 3 | 14.3 | 5.6–NR [KM curve: Probability of PFS vs Progression-Free Survival Time (Months), 0-52] Unstratified (ctDNA burden at BL/Wk 5): HR (B vs A)=0.36 (95% CI, 0.14–0.93); p=0.03 HR (C vs A)=0.23 (95% CI, 0.09–0.56); p=0.0004 HR (C vs B)=0.72 (95% CI, 0.38–1.39); p=0.33 No. at risk (months 0,2,4,...,52) A: 12 11 9 8 8 6 6 6 5 4 4 3 3 3 3 3 3 3 2 2 2 2 0 0 0 0 0 B: 55 54 51 47 47 41 38 36 35 35 33 33 32 31 30 29 29 27 21 15 14 10 6 3 2 1 0 C: 114 108 106 104 104 103 100 98 94 92 89 85 79 79 77 75 73 66 51 37 37 23 13 10 7 1 0 D: 4 4 4 3 3 3 3 2 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 Data cutoff: February 18, 2026 Hazard ratio based on Cox proportional hazards model. P-value from an unstratified log-rank test. *ITT population excluding patients in China; rPFS by investigator assessment. Wk=Week. Survival table Endpoint: rPFS by ctDNA status at baseline and Week 5, TALA + ENZA arm (investigator-assessed) Arm | N | Median | Median CI A. ctDNA+ at BL and Wk 5 | 12 | 17.3 | 9.1–NR B. ctDNA+ at BL and ctDNA− at Wk 5 | 55 | NR | NR–NR C. ctDNA− at BL and Wk 5 | 114 | NR | NR–NR D. ctDNA− at BL and ctDNA+ at Wk 5 | 4 | 14.3 | 5.6–NR HR: B vs A 0.36; C vs A 0.23; C vs B 0.72 HR CI: B vs A 0.14–0.93; C vs A 0.09–0.56; C vs B 0.38–1.39 P: B vs A 0.03; C vs A 0.0004; C vs B 0.33 [Slide 4] Conclusions • In TALAPRO-3, baseline and on-treatment Week 5 ctDNA tumor fraction showed potential prognostic utility for rPFS in mCSPC • Conversion of ctDNA tumor fraction from positive at baseline to negative at Week 5, which was higher with TALA + ENZA compared with PBO + ENZA, correlated with improved rPFS • Further investigation will look at whether ctDNA provides independent prognostic value beyond established markers (e.g., PSA at 6 months) through multivariate analysis ctDNA tumor fraction has potential as a prognostic biomarker and may improve risk stratification and therapeutic monitoring in patients with mCSPC PSA=prostate-specific antigen

[Slide 1] Figure 1. Incidence of All-Cause (A) Hematologic, (B) Most Common (≥15%) Non-Hematologic TEAEs, and (C) AESIs (Safety Population) A. (Grade 1 | Grade 2 | Grade 3 | Grade 4) Anemia | 8.4 | 13.7 | 48.5 | 2.3 Neutropenia | 7.7 | 20.7 | 16.7 | 3.0 Leukopenia | 9.7 | 15.1 | 6.0 Thrombocytopenia | 9.7 | 9.0 | 6.7 | 2.7 Lymphopenia | 2.3 | 4.0 | [illegible] | 0.3 B. (Grade 1 | Grade 2 | Grade 3) Fatigue | 15.7 | [illegible] | 3.0 Asthenia | 6.0 | [illegible] | [illegible] Hot flush | 18.1 | 2.0 Decreased appetite | 13.0 | 4.7 Back pain | 10.0 | [illegible] | 1.3 Constipation | 12.0 | [illegible] Nausea | 13.0 | 4.0 Arthralgia | 10.4 | 4.7 C. (Grade 1/2 | Grade ≥3) Any AESI | 2.3 | 5.0 Secondary primary malignancies | 1.0 | 2.3 Embolic and thrombotic events | 1.0 | 1.3 MDS | 0.3 | 0.7 AML | 0.7 Patients, % The treatment-emergent period is from the first dose through 28 days after the last dose of study treatment, or before initiation of new systemic (ie, not including surgery or radiotherapy) antineoplastic therapy excluding androgen deprivation therapy, whichever occurs first. Cluster terms used for anemia, neutropenia, leukopenia, thrombocytopenia, and lymphopenia. AESI=adverse events of special interest; AML=acute myeloid leukemia; MDS=myelodysplastic syndrome; TEAE=treatment-emergent adverse event. [Slide 2] Figure 4. rPFS in Patients With and Without a Talazoparib Dose Reduction (Safety Population) | n | Events | Median, mo | 95% CI With dose reductions | 195 | 44 | NR | NR–NR Without dose reductions | 104 | 23 | NR | 40.7–NR Unstratified: HR=0.79 (95% CI, 0.48–1.31); 2-sided p=0.36 No. at risk (months 0-52, every 2) 195 191 186 180 179 173 170 164 157 156 152 148 142 141 135 129 126 115 90 70 69 47 23 13 8 2 0 104 91 85 82 81 75 71 68 65 63 61 58 54 54 54 53 52 48 36 29 27 16 9 7 6 1 0 Unstratified hazard ratio based on Cox proportional hazards model. CI=confidence interval; HR=hazard ratio; NR=not reached; rPFS=radiographic progression-free survival Survival table Endpoint: rPFS by talazoparib dose reduction Arm | N | Median | Median CI With dose reductions | 195 | NR | NR–NR Without dose reductions | 104 | NR | 40.7–NR HR: 0.79 (unstratified) HR CI: 0.48–1.31 P: 0.36 (2-sided)

[Slide 1] 13:30 - 15:00 Rapid oral: Genitourinary tumours, prostate, penile and testis CHAIRS: THOMAS HWANG, EDMOND KWAN, ALISON REID Patient-reported outcomes in men with metastatic castration-sensitive prostate cancer and homologous recombination repair gene alterations receiving talazoparib + enzalutamide vs placebo + enzalutamide: Results from the Phase 3 TALAPRO-3 study Arun A. Azad,1 Fred Saad,2 Joaquin Mateo,3 Nobuaki Matsubara,4 Shusuan Jiang,5 Eric Voog,6 Dingwei Ye,7 Neal D. Shore,8 Timuçin Çil,9 Christof Vulsteke,10 Hsiao-Jen Chung,11 Jane Chang,12 Paul Cislo,12 Melissa Kirker,12 Prachi Nandoskar,13 Karim Fizazi,14 Neeraj Agarwal15 1Peter MacCallum Cancer Centre, Melbourne, Victoria, Australia; 2Centre Hospitalier de l'Université de Montréal, University of Montreal, Montreal, Canada; 3Vall d'Hebron Institute of Oncology (VHIO), Vall d'Hebron University Hospital, Barcelona, Spain; 4National Cancer Center Hospital East, Chiba, Japan; 5Hunan Cancer Hospital, Changsha, China; 6Clinique Victor Hugo, Centre Jean Bernard, Le Mans, France; 7Fudan University Shanghai Cancer Center, Shanghai, China; 8START Carolinas/Carolina Urologic Research Center, Myrtle Beach, SC, USA; 9Adana City Education and Research Hospital, Adana Faculty of Medicine, University of Health Sciences, Yuregir/Adana, Turkey; 10Integrated Cancer Center Ghent, AZ Maria Middelares, Ghent, Belgium and Center for Oncological Research (CORE), University of Antwerp, Antwerp, Belgium; 11Taipei Veterans General Hospital, Taipei, Taiwan, ROC; 12Pfizer Inc., New York, NY, USA; 13Pfizer Inc., South San Francisco, CA, USA; 14Centre Oscar Lambret, University of Paris Saclay, Lille, France; 15Huntsman Cancer Institute (NCI-CCC), University of Utah, Salt Lake City, UT, USA October 23-27, 2026 ClinicalTrials.gov identifier: NCT04821622. This study was sponsored by Pfizer Inc. Astellas Pharma Inc. provided enzalutamide. [Speaker card] Arun Azad — Patient-reported outcomes (PROs) in men with metastatic castration-sensitive prostate cancer (mCSPC) and homologous recombination repair (HRR) gene alterations receiving talazoparib (TALA) + enzalutamide (ENZA) vs placebo (PBO) + ENZA: Results from the... BILBAO AUDITORIUM - HALL 3 [Slide 3] 13:30 - 15:00 Rapid oral: Genitourinary tumours, prostate, penile and testis CHAIRS: THOMAS HWANG, EDMOND KWAN, ALISON REID Time to (definitive) deterioration in GHS/QoL, urinary symptoms, and worst pain were comparable in the two arms GHS/QoL per EORTC QLQ-C30* [KM curve; y-axis: Event-free probability; x-axis: Time to definitive deterioration (months), 0-52] | | No. of events/patients | Median (95% CI), mo | Stratified HR (95% CI) | TALA + ENZA | 92/240 | 38.0 (35.1-46.0) | 1.15 (0.85-1.58) PBO + ENZA | 78/234 | 46.0 (37.8-49.9) | Urinary symptoms per EORTC QLQ-PR25† [KM curve; y-axis: Event-free probability; x-axis: Time to definitive deterioration (months), 0-52] | | No. of events/patients | Median (95% CI), mo | Stratified HR (95% CI) | TALA + ENZA | 42/240 | NR (48.8-NR) | 0.94 (0.61-1.44) PBO + ENZA | 46/234 | 51.4 (46.8-NR) | Worst pain per BPI-SF‡ [KM curve; y-axis: Event-free probability; x-axis: Time to deterioration (months), 0-52] | | No. of events/patients | Median (95% CI), mo | Stratified HR (95% CI) | TALA + ENZA | 81/240 | NR (32.9-NR) | 0.82 (0.61-1.11) PBO + ENZA | 83/234 | NR (23.[illegible]-NR) | *Definitive deterioration defined as a ≥10-point decrease from baseline and no subsequent observations with a <10-point decrease from baseline assessed by the EORTC QLQ-C30. †Definitive deterioration defined as a ≥10-point increase from baseline and no subsequent observations with a <10-point increase from baseline assessed by the EORTC QLQ-PR25. ‡Deterioration defined as the first of either a ≥2-point increase from baseline for item 3 of the BPI-SF for two consecutive visits at least 4 weeks apart or an increase in WHO analgesic usage score on two consecutive visits. Confidential - Not for Public Consumption or Distribution [Speaker card] Arun Azad BILBAO AUDITORIUM - HALL 3 Survival table Endpoint: Time to (definitive) deterioration — GHS/QoL (EORTC QLQ-C30), urinary symptoms (EORTC QLQ-PR25), worst pain (BPI-SF) Arm | N | Median | Median CI GHS/QoL — TALA + ENZA | 240 | 38.0 mo | 35.1-46.0 GHS/QoL — PBO + ENZA | 234 | 46.0 mo | 37.8-49.9 Urinary symptoms — TALA + ENZA | 240 | NR | 48.8-NR Urinary symptoms — PBO + ENZA | 234 | 51.4 mo | 46.8-NR Worst pain — TALA + ENZA | 240 | NR | 32.9-NR Worst pain — PBO + ENZA | 234 | NR | 23.[illegible]-NR HR: GHS/QoL 1.15; Urinary 0.94; Worst pain 0.82 HR CI: GHS/QoL 0.85-1.58; Urinary 0.61-1.44; Worst pain 0.61-1.11 [Slide 4] 13:30 - 15:00 Rapid oral: Genitourinary tumours, prostate, penile and testis CHAIRS: THOMAS HWANG, EDMOND KWAN, ALISON REID Conclusions • Treatment with talazoparib plus enzalutamide did not result in clinically meaningful differences between arms (≥5-point threshold in overall mean change for deterioration) in HRQoL based on EORTC QLQ-C30 and -PR25 scales, except for appetite loss • While numerical worsening was noted for several HRQoL parameters in the talazoparib arm compared with the control arm, these differences did not reach the threshold for clinically meaningful difference despite the addition of a third therapy to standard-of-care treatment • Reduction of HRQoL around Week 17, including fatigue and appetite loss, is consistent with the onset of anemia, which was managed with dose modifications and supportive care • Time to definitive deterioration in GHS/QoL and urinary symptoms, and time to deterioration in worst pain were comparable between arms These results can help inform the risk-benefit assessment of talazoparib plus enzalutamide as a potential treatment option for patients with mCSPC harboring HRR gene alterations [Speaker card] Arun Azad BILBAO AUDITORIUM - HALL 3

[Slide 1] Baseline ctDNA Burden Was Prognostic of rPFS in Both Arms Legend | n | Events | Median, mo | 95% CI A. TALA + ENZA ctDNA+ at BL | 83 | 24 | NR | NR-NR B. TALA + ENZA ctDNA- at BL | 147 | 32 | NR | NR-NR C. PBO + ENZA ctDNA+ at BL | 75 | 40 | 26.7 | 18.2-NR D. PBO + ENZA ctDNA- at BL | 149 | 50 | NR | 45.8-NR Curve labels: TALA + ENZA ctDNA- at BL; TALA + ENZA ctDNA+ at BL; PBO + ENZA ctDNA- at BL; PBO + ENZA ctDNA+ at BL Unstratified HR (A vs B)=1.59 (95% CI, 0.94-2.70); p=0.08 HR (A vs C)=0.47 (95% CI, 0.28-0.78); p=0.003 HR (B vs D)=0.62 (95% CI, 0.39-0.96); p=0.03 HR (C vs D)=2.15 (95% CI, 1.42-3.26); p=0.0002 Probability of PFS Progression-Free Survival Time (Months) No. at risk (months 0,2,4,...,52) A (solid dark): 83 80 74 69 69 60 57 54 52 51 49 48 47 46 44 42 42 39 30 24 23 17 8 4 3 1 0 B (dotted dark): 147 139 136 133 132 131 127 123 117 115 112 108 100 100 98 95 93 84 66 51 50 33 18 13 9 2 0 C (solid light): 75 69 65 59 52 49 46 44 42 42 38 36 33 32 29 26 26 21 15 13 13 9 4 3 2 1 0 D (dotted light): 149 144 143 131 128 125 121 116 114 113 107 102 102 102 95 88 88 76 55 38 38 25 19 14 8 4 0 Survival table Endpoint: rPFS by baseline ctDNA status Arm | N | Median | Median CI A. TALA + ENZA ctDNA+ at BL | 83 | NR | NR-NR B. TALA + ENZA ctDNA- at BL | 147 | NR | NR-NR C. PBO + ENZA ctDNA+ at BL | 75 | 26.7 mo | 18.2-NR D. PBO + ENZA ctDNA- at BL | 149 | NR | 45.8-NR HR: A vs C 0.47 HR CI: 0.28-0.78 P: 0.003 [Slide 2] ctDNA+ to ctDNA- Conversion Rates Were Higher With TALA + ENZA ctDNA+ at baseline*: TALA + ENZA 67/269 (24.9%) PBO + ENZA 66/262 (25.2%) Conversion to ctDNA- at Week 5, n (%)†,‡ TALA + ENZA: Yes 55 (82.1%) [95% CI, 70.4-90.0]; No 12 (17.9%) PBO + ENZA: Yes 45 (68.2%) [95% CI, 55.4-78.8]; No 21 (31.8%) ctDNA- at baseline*: TALA + ENZA 118/269 (43.9%) PBO + ENZA 123/262 (46.9%) Conversion to ctDNA+ at Week 5, n (%)†,‡ TALA + ENZA: No 114 (96.6%) [95% CI, 91.0-98.9]; Yes 4 (3.4%) PBO + ENZA: No 121 (98.4%) [95% CI, 93.7-99.7]; Yes 2 (1.6%) Patients, % [x-axis: TALA + ENZA | PBO + ENZA | TALA + ENZA | PBO + ENZA] [Slide 3] Conversion from ctDNA+ to ctDNA- and persistent ctDNA- at Week 5 was associated with longer rPFS in the TALA + ENZA Arm Legend | n | Events | Median, mo | 95% CI A. TALA + ENZA ctDNA+ at BL and Wk 5 | 12 | 6 | 17.3 | 9.1-NR B. TALA + ENZA ctDNA+ at BL and ctDNA- at Wk 5 | 55 | 14 | NR | NR-NR C. TALA + ENZA ctDNA- at BL and Wk 5 | 114 | 25 | NR | NR-NR D. TALA + ENZA ctDNA- at BL and ctDNA+ at Wk 5 | 4 | 3 | 14.3 | 5.6-NR Unstratified (ctDNA burden at BL/Wk 5): HR (B vs A)=0.36 (95% CI, 0.14-0.93); p=0.03 HR (C vs A)=0.23 (95% CI, 0.09-0.56); p=0.0004 HR (C vs B)=0.72 (95% CI, 0.38-1.39); p=0.33 Probability of PFS Progression-Free Survival Time (Months) No. at risk (months 0,2,4,...,52) A: 12 11 9 8 8 6 6 6 5 4 4 3 3 3 3 3 3 2 2 2 2 2 0 0 0 0 0 B: 55 54 51 47 47 41 38 36 35 35 33 33 32 31 30 29 29 27 21 15 14 10 6 3 2 1 0 C: 114 108 106 104 104 103 100 98 94 92 89 85 79 79 77 75 73 66 51 37 37 23 13 10 7 1 0 D: 4 4 4 3 3 3 3 2 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 Survival table Endpoint: rPFS by ctDNA status at BL/Week 5 (TALA + ENZA arm) Arm | N | Median | Median CI A. ctDNA+ at BL and Wk 5 | 12 | 17.3 mo | 9.1-NR B. ctDNA+ at BL and ctDNA- at Wk 5 | 55 | NR | NR-NR C. ctDNA- at BL and Wk 5 | 114 | NR | NR-NR D. ctDNA- at BL and ctDNA+ at Wk 5 | 4 | 14.3 mo | 5.6-NR HR: B vs A 0.36 HR CI: 0.14-0.93 P: 0.03 [Slide 4] Conversion from ctDNA+ to ctDNA- and persistent ctDNA- at Week 5 was associated with longer rPFS in the PBO + ENZA Arm Legend | n | Events | Median, mo | 95% CI A. PBO + ENZA ctDNA+ at BL and Week 5 | 21 | 14 | 11.5 | 3.6-29.3 B. PBO + ENZA ctDNA+ at BL and ctDNA- at Week 5 | 45 | 23 | 35.1 | 18.4-NR C. PBO + ENZA ctDNA- at BL and Week 5 | 121 | 39 | NR | 46.0-NR D. PBO + ENZA ctDNA- at BL and ctDNA+ at Week 5 | 2 | 2 | 4.7 | 3.9-NR Unstratified (ctDNA burden at BL/Wk 5): HR (B vs A)=0.43 (95% CI, 0.22-0.83); p=0.01 HR (C vs A)=0.23 (95% CI, 0.12-0.43); p<0.0001 HR (C vs B)=0.55 (95% CI, 0.33-0.92); p=0.02 Probability of PFS Progression-Free Survival Time (Months) No. at risk (months 0,2,4,...,52) A: 21 17 14 12 10 9 8 8 7 7 6 5 5 3 3 2 2 2 2 2 2 0 0 0 0 0 0 B: 45 45 45 42 37 35 33 31 31 31 28 27 27 26 23 22 22 18 13 11 11 7 4 3 2 1 0 C: 121 120 120 111 108 105 102 98 96 95 90 86 86 86 81 75 75 65 45 31 31 18 13 12 6 3 0 D: 2 2 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Survival table Endpoint: rPFS by ctDNA status at BL/Week 5 (PBO + ENZA arm) Arm | N | Median | Median CI A. ctDNA+ at BL and Week 5 | 21 | 11.5 mo | 3.6-29.3 B. ctDNA+ at BL and ctDNA- at Week 5 | 45 | 35.1 mo | 18.4-NR C. ctDNA- at BL and Week 5 | 121 | NR | 46.0-NR D. ctDNA- at BL and ctDNA+ at Week 5 | 2 | 4.7 mo | 3.9-NR HR: B vs A 0.43 HR CI: 0.22-0.83 P: 0.01

[Slide 1] Investigator-Assessed rPFS in TALAPRO-3† 77% (95% CI, 71–81) 56% (95% CI, 50–62) TALA + ENZA (n=300) | PBO + ENZA (n=299) No. of events | 67 | 126 Median (95% CI), mo | NR (NR–NR) | 45.8 (37.7–NR) HR=0.48 (95% CI, 0.36–0.65); p<0.001 Y-axis: Patients, % X-axis: Months 0 4 8 12 16 20 24 28 32 36 40 44 48 52 Survival table Endpoint: Investigator-assessed rPFS Arm | N | Median | Median CI TALA + ENZA | 300 | NR | NR–NR PBO + ENZA | 299 | 45.8 mo | 37.7–NR HR: 0.48 HR CI: 0.36–0.65 P: <0.001 Landmark rates: 36 months (dashed line): TALA + ENZA 77% (CI 71–81) | PBO + ENZA 56% (CI 50–62) [Slide 2] TALAPRO-3 trial design: PRO analysis Key eligibility criteria • HRRm mCSPC • Metastatic disease confirmed by positive bone scan or metastatic lesion by CT or MRI scan • ECOG PS 0 or 1 • Ongoing ADT • ≤3 months of prior ADT with/without ARPI for mCSPC • Prior docetaxel not permitted* Stratification factors • De novo vs relapsed mCSPC • High- vs low-volume disease • BRCAm vs non-BRCAm 1:1 N=599 Talazoparib 0.5 mg† + enzalutamide 160 mg QD Placebo + enzalutamide 160 mg QD Data cutoff: February 18, 2026 Primary endpoint • rPFS by investigator assessment Key secondary endpoint • OS (alpha-protected) PRO endpoints‡ • Change from BL in GHS/QoL and symptoms • TTDD in GHS/QoL • TTDD in EORTC QLQ-PR25 • TTD in BPI-SF PRO instruments • EORTC QLQ-C30 (cancer specific) • EORTC QLQ-PR25 (prostate cancer–specific) • BPI-SF (pain) • EQ-5D-5L (general health status) Assessed at baseline, every 4W to W25, then every 8W to W57, then every 12W up to the safety follow-up visit A ≥5-point difference between groups for deterioration in EORTC-QLQ measures was considered clinically meaningful§ TALAPRO-3 HRRm gene panel: ATM, ATR, BRCA1, BRCA2, CDK12, CHEK2, FANCA, MLH1, MRE11A, NBN, PALB2, RAD51C§

[Slide 1] Conversion from ctDNA+ to ctDNA- and persistent ctDNA- at Week 5 was associated with longer rPFS in the PBO + ENZA Arm Legend: n | Events | Median, mo | 95% CI A. PBO + ENZA ctDNA+ at BL and Week 5 | 21 | 14 | 11.5 | 3.6–29.3 B. PBO + ENZA ctDNA+ at BL and ctDNA− at Week 5 | 45 | 23 | 35.1 | 18.4–NR C. PBO + ENZA ctDNA− at BL and Week 5 | 121 | 39 | NR | 46.0–NR D. PBO + ENZA ctDNA− at BL and ctDNA+ at Week 5 | 2 | 2 | 4.7 | 3.9–NR Curve labels: PBO + ENZA ctDNA− at BL and Wk 5; PBO + ENZA ctDNA+ at BL and ctDNA− at Wk 5; PBO + ENZA ctDNA+ at BL and Wk 5; PBO + ENZA ctDNA− at BL and ctDNA+ at Wk 5 Unstratified (ctDNA burden at BL/Wk 5): HR (B vs A)=0.43 (95% CI, 0.22–0.83); p=0.01 HR (C vs A)=0.23 (95% CI, 0.12–0.43); p<0.0001 HR (C vs B)=0.55 (95% CI, 0.33–0.92); p=0.02 Y axis: Probability of PFS; X axis: Progression-Free Survival Time (Months) 0–52 No. at risk (months 0,2,4,...,52): A: 21 17 14 12 10 9 8 7 7 7 6 5 5 3 3 2 2 2 2 2 0 0 0 0 0 0 0 B: 45 45 45 42 37 35 33 31 31 31 28 27 27 26 23 22 22 18 13 11 11 7 4 3 2 1 0 C: 121 120 120 111 108 105 102 98 96 95 90 86 86 81 75 75 65 45 31 31 18 13 12 6 3 0 [illegible] D: 2 2 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Data cutoff: February 18, 2026 Hazard ratio based on Cox proportional hazards model. P-value from an unstratified log-rank test. *ITT population excluding patients in China; rPFS by investigator assessment. Wk=Week. Survival table Endpoint: rPFS (investigator) by ctDNA status at baseline and Week 5, PBO + ENZA arm Arm | N | Median | Median CI A. PBO + ENZA ctDNA+ at BL and Week 5 | 21 | 11.5 mo | 3.6–29.3 B. PBO + ENZA ctDNA+ at BL and ctDNA− at Week 5 | 45 | 35.1 mo | 18.4–NR C. PBO + ENZA ctDNA− at BL and Week 5 | 121 | NR | 46.0–NR D. PBO + ENZA ctDNA− at BL and ctDNA+ at Week 5 | 2 | 4.7 mo | 3.9–NR HR: 0.43 (B vs A) HR CI: 0.22–0.83 P: 0.01 [Slide 3] ctDNA+ to ctDNA- Conversion Rates Were Higher With TALA + ENZA ctDNA+ at baseline*: TALA + ENZA 67/269 (24.9%) | PBO + ENZA 66/262 (25.2%) Conversion to ctDNA− at Week 5, n (%)†,‡ TALA + ENZA: Yes 55 (82.1%) [95% CI, 70.4–90.0]; No 12 (17.9%) PBO + ENZA: Yes 45 (68.2%) [95% CI, 55.4–78.8]; No 21 (31.8%) ctDNA− at baseline*: TALA + ENZA 118/269 (43.9%) | PBO + ENZA 123/262 (46.9%) Conversion to ctDNA+ at Week 5, n (%)†,‡ TALA + ENZA: No 114 (96.6%) [95% CI, 91.0–98.9]; Yes 4 (3.4%) PBO + ENZA: No 121 (98.4%) [95% CI, 93.7–99.7]; Yes 2 (1.6%) Y axis: Patients, % Data cutoff: February 18, 2026 *Denominator for participants with evaluable ctDNA at both baseline and Week 5 is intention-to-treat population excluding China patients. †Counts reported for participants with evaluable ctDNA at both baseline and Week 5. ‡Denominator for conversion calculations is patients with evaluable ctDNA at both baseline and Week 5. Two-sided 95% CI calculated using the Wilson score interval with continuity correction. [Slide 4] 13:30 - 15:00 Rapid oral: Genitourinary tumours, prostate, penile and testis CHAIRS: THOMAS HWANG, EDMOND KWAN, ALISON REID Background • In the Phase 3 TALAPRO-3 trial, adding TALA (PARPi) to ENZA (ARPI) significantly improved rPFS versus PBO + ENZA in patients with HRR-deficient metastatic castration-sensitive prostate cancer (mCSPC; also known as metastatic APMN/S prostate cancer) (HR=0.48; 95% CI 0.36–0.65; p<0.001)1,2 • Circulating tumor DNA (ctDNA) tumor fraction is an emerging prognostic biomarker with broad potential utility across multiple treatment approaches and tumor types*,3,4 • We assessed the potential prognostic utility of baseline ctDNA tumor fraction and its early on-treatment dynamics, captured as the change at Week 5, in patients from TALAPRO-3 *ctDNA tumor fraction is the proportion of total cell-free DNA in a sample which is tumor derived.4 APMN/S=androgen pathway modulation-naive/sensitive; ARPI=androgen receptor pathway inhibitor; CI=confidence interval; ENZA=enzalutamide; HR=hazard ratio; HRR=homologous recombination repair; PARPi=poly(ADP-ribose) polymerase inhibitor; PBO=placebo; rPFS=radiographic progression-free survival; TALA=talazoparib. 1. Agarwal N, et al. N Engl J Med. 2026;395:427-439 2. Armstrong AJ, et al. J Clin Oncol. 2026;44:1249-1265 3. Reichert ZR, et al. Clin Cancer Res. 2023;34:111-120 4. Rolfo CD, et al. Clin Cancer Res. 2024;30:2452-2460 [Speaker panel] Arun Azad — Utility of ctDNA tumor fraction (TF) as a prognostic biomarker in TALAPRO-3: phase 3 study of talazoparib (TALA) + enzalutamide (ENZA) vs placebo (PBO) + ENZA in patients (pts) with homologous recombination repair (HRR) gene–mutated metastatic castra... BILBAO AUDITORIUM - HALL 3

[Slide 1] TALAPRO-3 trial design: PRO analysis Key eligibility criteria • HRRm mCSPC • Metastatic disease confirmed by positive bone scan or metastatic lesion by CT or MRI scan • ECOG PS 0 or 1 • Ongoing ADT • ≤3 months of prior ADT with/without ARPI for mCSPC • Prior docetaxel not permitted* Stratification factors • De novo vs relapsed mCSPC • High- vs low-volume disease • BRCAm vs non-BRCAm 1:1 N=599 Talazoparib 0.5 mg† + enzalutamide 160 mg QD Placebo + enzalutamide 160 mg QD Data cutoff: February 18, 2026 Primary endpoint • rPFS by investigator assessment Key secondary endpoint • OS (alpha-protected) PRO endpoints‡ • Change from BL in GHS/QoL and symptoms • TTDD in GHS/QoL • TTDD in EORTC QLQ-PR25 • TTD in BPI-SF PRO instruments • EORTC QLQ-C30 (cancer specific) • EORTC QLQ-PR25 (prostate cancer–specific) • BPI-SF (pain) • EQ-5D-5L (general health status) Assessed at baseline, every 4W to W25, then every 8W to W57, then every 12W up to the safety follow-up visit A ≥5-point difference between groups for deterioration in EORTC-QLQ measures was considered clinically meaningful1 TALAPRO-3 HRRm gene panel: ATM, ATR, BRCA1, BRCA2, CDK12, CHEK2, FANCA, MLH1, MRE11A, NBN, PALB2, RAD51C§ *For mCSPC, as per the protocol amendment (September 2021). ClinicalTrials.gov identifier: NCT04821622; †0.35 mg QD if the patient has moderate renal impairment; ‡A stratified log-rank test and Cox proportional hazards model were used to compare time-to-events between arms, with HR values being nominal (not alpha controlled); §HRR testing using FoundationOne®/FoundationOne®CDx and FoundationOne®Liquid CDx. ADT=androgen deprivation therapy; ARPI=androgen receptor pathway inhibitor; BL=baseline; BPI-SF=Brief Pain Inventory Short Form; EORTC QLQ-C30=European Organisation for Research and Treatment of Cancer cancer-specific global health questionnaire; EORTC QLQ-PR25=European Organisation for Research and Treatment of Cancer disease-specific urinary symptoms questionnaire; EQ-5D-5L=EuroQol 5-Dimension 5-Level; GHS/QoL=general health status/quality of life; HRRm=HRR gene alterations; QD=once daily; TT(D)D=time to (definitive) deterioration; W=week. 1. Gamper EM, et al. BMC Cancer. 2021;21:1083. Confidential - Not for Public Consumption or Distribution [Slide 2] Comparable mean change from baseline in physical functioning per EORTC QLQ-C30 EORTC QLQ-C30 Score Range: 0–100 Estimated overall mean change from baseline difference (95% CI) -2.5 (-5.1 to 0.1)* [line chart: Observed mean change from baseline, 95% CI (5 to -25) vs Assessment visit (week) 5, 9, 13, 17, 21, 25, 33, 41, 49, 57, 69, 81, 93, 105, 117, 129, 141, 153, 165, 177, 189, 201, 213; Improved (up) / Worsened (down)] Number evaluable TALA + ENZA: 183 173 171 173 164 172 166 161 154 148 144 136 131 129 120 117 113 103 75 50 34 18 9 PBO + ENZA: 169 164 170 167 162 168 156 153 151 146 139 134 126 117 109 106 99 87 60 39 29 20 10 *Differences in estimated overall mean scores between arms were assessed by longitudinal mixed-effects model for repeated measures. A ≥5-point difference between groups for deterioration was considered clinically meaningful.1 1. Gamper EM, et al. BMC Cancer. 2021;21:1083. Confidential - Not for Public Consumption or Distribution [Slide 3] Similar mean change from baseline between arms in fatigue per EORTC QLQ-C30 EORTC QLQ-C30 Score Range: 0–100 Estimated overall mean change from baseline difference (95% CI) 3.2 (0.2–6.2)* [line chart: Observed mean change from baseline, 95% CI (40 to -20) vs Assessment visit (week) 5, 9, 13, 17, 21, 25, 33, 41, 49, 57, 69, 81, 93, 105, 117, 129, 141, 153, 165, 177, 189, 201, 213; Worsened (up) / Improved (down)] Number evaluable TALA + ENZA: 183 173 171 173 164 172 166 161 154 148 144 136 131 129 120 117 113 103 75 50 34 18 9 PBO + ENZA: 169 164 170 167 162 168 156 153 151 146 139 134 126 117 109 106 99 87 60 39 29 20 10 *Differences in estimated overall mean scores between arms were assessed by longitudinal mixed-effects model for repeated measures. A ≥5-point difference between groups for deterioration was considered clinically meaningful.1 1. Gamper EM, et al. BMC Cancer. 2021;21:1083. Confidential - Not for Public Consumption or Distribution

[Slide 2] Baseline Characteristics | Pembro + cis + gem n = 453 | Placebo + cis + gem n = 454 Median age (range), years | 65.0 (31-85) | 65.0 (41-82) ≥65 years | 243 (53.6) | 232 (51.1) Male, n (%) | 365 (80.6) | 375 (82.6) ECOG PS, n (%) 0 | 346 (76.4) | 344 (75.8) 1 | 107 (23.6) | 110 (24.2) Geographic region, n (%) United States | 40 (8.8) | 41 (9.0) Western Europe | 223 (49.2) | 230 (50.7) Most of world | 190 (41.9) | 183 (40.3) PD-L1 CPS,a n (%) ≥10 | 257 (56.7) | 255 (56.2) <10 | 193 (42.6) | 196 (43.2) Tumor stage at baseline (central assessment),b n (%) cT2N0 | 67 (14.8) | 66 (14.5) cT3/cT4aN0 | 383 (84.5) | 385 (84.8) cT1-4aN1c | 3 (0.7) | 3 (0.7) Pure urothelial carcinoma histology, n (%) | 422 (93.2) | 433 (95.4) Creatinine clearance, n (%) >90 mL/min | 242 (53.4) | 271 (59.7) 60-90 mL/min | 208 (45.9) | 182 (40.1) <60 mL/mind | 3 (0.7) | 1 (0.2) aPD-L1 status was nonevaluable for 3 participant in each treatment arm. bUsing central pathology and/or imaging per protocol. cOne participant with cT4aN1M0 disease was classified as cT4aN1M1 by the central vendor because the original protocol, the vendor system was designed to capture N1-3 disease as M1. dProtocol deviation. Data cutoff: January 26, 2026. [Slide 3] Primary Endpoint: EFS by Investigator Assessment ITT Population | Pembro + cis + gem n = 453 | Placebo + cis + gem n = 454 Events, n (%) | 157 (34.7) | 168 (37.0) Median (95% CI), mo | 74.1 (72.4-NR) | NR (67.7-NR) HR 0.88 (95% CI, 0.71-1.10), 1-sided P = 0.1366* Landmarks: 24 mo 76.7% vs 70.0%; 36 mo 69.2% vs 66.3%; 48 mo 66.0% vs 62.3%; 60 mo 63.5% vs 61.9% Months: 0 6 12 18 24 30 36 42 48 54 60 66 72 78 No. at risk Pembro + cis + gem: 453 401 366 338 329 304 296 273 224 175 114 51 11 0 Placebo + cis + gem: 454 400 345 310 300 291 283 257 208 168 106 51 12 0 NR, not reached. *Not statistically significant (1-sided boundary 0.0225). Data cutoff: January 26, 2026. Survival table Endpoint: EFS by investigator assessment (ITT) Arm | N | Median | Median CI Pembro + cis + gem | 453 | 74.1 mo | 72.4-NR Placebo + cis + gem | 454 | NR | 67.7-NR HR: 0.88 HR CI: 0.71-1.10 P: 1-sided P = 0.1366 (not statistically significant; 1-sided boundary 0.0225) Landmark rates: 24 mo: Pembro + cis + gem 76.7% | Placebo + cis + gem 70.0% 36 mo: Pembro + cis + gem 69.2% | Placebo + cis + gem 66.3% 48 mo: Pembro + cis + gem 66.0% | Placebo + cis + gem 62.3% 60 mo: Pembro + cis + gem 63.5% | Placebo + cis + gem 61.9% [Slide 4] Subsequent Therapy ITT Population n (%) | Pembro + cis + gem n = 453 | Placebo + cis + gem n = 454 Any | 85 (18.8) | 108 (23.8) Participants who did not undergo cystectomy | 16 (3.5) | 25 (5.5) Chemotherapy | 9 (2.0) | 13 (2.9) Radiotherapy | 9 (2.0) | 18 (4.0) Anti-PD-(L)1 | 3 (0.7) | 7 (1.5) Enfortumab vedotin | 1 (0.2) | 2 (0.4) Enfortumab vedotin plus pembrolizumab | 1 (0.2) | 0 Other | 3 (0.7) | 2 (0.4) Participants who underwent cystectomy | 69 (15.2) | 83 (18.3) Adjuvanta | 5 (1.1) | 10 (2.2) Anti-PD-(L)1 | 5 (1.1) | 7 (1.5) Chemotherapy | 1 (0.2) | 3 (0.7) Recurrent/metastaticb | 61 (13.5) | 76 (16.7) Chemotherapy | 37 (8.2) | 44 (9.7) Anti-PD-(L)1 | 22 (4.9) | 51 (11.2) Enfortumab vedotin | 19 (4.2) | 16 (3.5) Radiotherapy | 19 (4.2) | 27 (5.9) Enfortumab vedotin plus pembrolizumab | 5 (1.1) | 1 (0.2) Anti-PD-(L)1 plus chemotherapy | 1 (0.2) | 2 (0.4) Other | 5 (1.1) | 8 (1.8) Otherc | 6 (1.3) | 4 (0.9) aSubsequent oncologic therapies received after surgery for participants who underwent cystectomy to reduce the risk of disease recurrence. bSubsequent oncologic therapies received after cystectomy to treat recurrent or metastatic disease. cTherapies given off study for participants who had off-study surgery. Data cutoff: January 26, 2026.

[Slide 1] Phase 3 KEYNOTE-866 Study Design (NCT03924856) Key Eligibility Criteria - Adults with MIBC - Clinical stage T2-T4aN0M0 or T1-T4aN1M0 by central assessment - Urothelial histology >=50% - Eligible for RC + PLND - Did not meet any Galsky criteria for cisplatin ineligibility,1 including GFR >=60 mL/min - ECOG PS 0 or 1 Stratification Factors - PD-L1 status (CPS >=10 vs <10)a - Clinical stage (T2N0 vs T3/T4aN0 vs T1-4aN1) - Geographic region (US vs Western Europe vs most of world) R 1:1 n = 453: Pembrolizumab 200 mg d1 IV Q3W 4 cycles + Cisplatin 70 mg/m2 d1 IV Q3W 4 cycles + Gemcitabine 1000 mg/m2 d1 and d8 IV Q3W 4 cycles -> RC + PLND -> Pembrolizumab 200 mg d1 IV Q3W 13 cycles n = 454: Placebo d1 IV Q3W 4 cycles + Cisplatin 70 mg/m2 d1 IV Q3W 4 cycles + Gemcitabine 1000 mg/m2 d1 and d8 IV Q3W 4 cycles -> RC + PLND -> Placebo d1 IV Q3W 13 cycles Primary endpoint: EFS by investigator assessment Key secondary endpoints: OS and pathological complete response (pCR; pT0N0, ie, absence of viable tumor in examined tissue from cystectomy) by central pathologist review Other secondary endpoints include: Safety aAssessed by PD-L1 IHC 22C3 pharmDx (Agilent, Carpinteria, CA); CPS = #PD-L1-staining cells (tumor cells, lymphocytes, and macrophages) / total # viable tumor cells x 100. 1. Galsky MD et al. Lancet Oncol. 2011;12:211-214. Data cutoff: January 26, 2026. [Slide 3] Common TEAEsa (Incidence >=15% in Either Arm) Safety Analysis Population, All Phases of Treatmentb Pembro + cis + gem (n = 450) | Placebo + cis + gem (n = 449) Terms (top to bottom): Nausea, Anemia, Neutropenia, Constipation, Fatigue, Urinary tract infection, Diarrhea, Decreased appetite, Thrombocytopenia, Blood creatinine increased, Asthenia, Pruritus, Weight decreased, Vomiting, Leukopenia, Pyrexia Pembro + cis + gem values (top to bottom): 54.9%, 54.4%, 53.1%, 35.8%, 34.9%, 26.0%, 24.7%, 24.2%, 23.3%, 22.9%, 22.4%, 20.4%, 17.8%, 17.1%, 16.2%, 15.3% Placebo + cis + gem values (top to bottom): 52.3%, 55.2%, 50.6%, 39.0%, 29.8%, 28.1%, 19.2%, 20.7%, 23.8%, 20.5%, 24.3%, 10.9%, 12.0%, 13.8%, 13.6%, 14.9% Legend: Grade 1 or 2 | Grade 3-5 X-axis: Incidence, % (100 ... 0 ... 100) aCollected up to 30 days after cessation of study treatment (90 days for serious AEs). bReporting window included neoadjuvant, surgical, and adjuvant phases. Data cutoff: January 26, 2026. [Slide 4] Conclusions - KEYNOTE-866 is the first placebo-controlled trial to add perioperative PD-1(L)1-based immunotherapy to neoadjuvant cisplatin plus gemcitabine and RC + PLND for MIBC - The addition of pembrolizumab to cisplatin plus gemcitabine did not confer a statistically significant EFS benefit in participants with MIBC who were eligible for cisplatin - EFS outcomes were generally consistent across key subgroups - The key secondary endpoints of OS and pCR rate were not formally tested for statistical significance per the statistical design - OS was similar between the treatment groups - pCR rate was numerically higher in the pembrolizumab group compared with the placebo group (44.4% vs 35.8%) - Efficacy outcomes in the placebo group surpassed conventional benchmarks for neoadjuvant cisplatin plus gemcitabine and RC + PLND in MIBC1 - The safety profile of perioperative pembrolizumab added to standard-of-care neoadjuvant cisplatin plus gemcitabine and RC + PLND was manageable and consistent with prior reports with no new safety signals2,3 - While perioperative pembrolizumab plus neoadjuvant cisplatin-based chemotherapy is not supported for participants with MIBC, perioperative enfortumab vedotin plus pembrolizumab has become a standard of care in this disease setting4,5 1. Powles T et al. N Engl J Med. 2024;391:1773-1786. 2. Bellmunt J et al. N Engl J Med. 2017;376:1015-1026. 3. Balar AV et al. Lancet Oncol. 2017;18:1483-1492. 4. Galsky MD et al. N Engl J Med. 2026;395:338-348. 5. Vulsteke C et al. N Engl J Med. 2026;394:1257-1269.

[Slide 2] Key Secondary Endpoint: OS ITT Population | | Pembro + cis + gem n = 453 | Placebo + cis + gem n = 454 | | Events, n (%) | 135 (29.8) | 135 (29.7) | | Median (95% CI), mo | NR (NR-NR) | NR (NR-NR) | HR 1.00 (95% CI, 0.79-1.27)ª [KM curve: Overall Survival, % vs Months 0-84] Landmarks: 24 mo 82.8% vs 81.7%; 36 mo 76.8% vs 76.6%; 48 mo 72.9% vs 73.4%; 60 mo 70.8% vs 70.9% No. at risk (months 0,6,12,...,84) Pembro + cis + gem: 453 440 409 386 375 363 348 332 278 231 161 94 43 1 0 Placebo + cis + gem: 454 437 410 390 370 358 344 332 278 229 159 97 40 2 0 NR, not reached. ªNot formally statistically tested according to the statistical design. Data cutoff: January 2[illegible], 2026. Survival table Endpoint: OS (ITT) Arm | N | Median | Median CI Pembro + cis + gem | 453 | NR | NR-NR Placebo + cis + gem | 454 | NR | NR-NR HR: 1.00 HR CI: 0.79-1.27 Landmark rates: 24 mo: Pembro + cis + gem 82.8% | Placebo + cis + gem 81.7% 36 mo: Pembro + cis + gem 76.8% | Placebo + cis + gem 76.6% 48 mo: Pembro + cis + gem 72.9% | Placebo + cis + gem 73.4% 60 mo: Pembro + cis + gem 70.8% | Placebo + cis + gem 70.9% [Slide 3] Key Secondary Endpoint: pCR by Central Pathology Review ITT Populationª [Bar chart: pCR Rate, % (95% CI)] Estimated difference:ᵇ 8.9% (95% CI, 2.5-15.1)* | | Pembro + cis + gem n = 450 | Placebo + cis + gem n = 453 | | pCR, n | 200 | 162 | | pCR rate, % (95% CI) | 44.4 (39.8-49.2) | 35.8 (31.3-40.4) | • pCR: defined as absence of viable tumor in examined tissue from RC + PLND (pT0N0) • Participants who did not undergo RC + PLND were considered nonresponders ªExcluding those who did not undergo RC + PLND and had a clinical CR. ᵇBased on the Miettinen-Nurminen method stratified by the randomization factors. *Not formally statistically tested according to the statistical design. Data cutoff: January 2[illegible], 2026 [Slide 4] Immune-Mediated AEs and Infusion Reactionsª Safety Analysis Population Pembro + cis + gem (n = 450) | Placebo + cis + gem (n = 449) Hypothyroidism 10.4% | 1.3% Hyperthyroidism 5.1% | 1.8% Pneumonitis 4.0% | 0.7% Severe skin reactions 3.1% | 0.4% Colitis 2.7% | 0.7% Adrenal insufficiency 2.4% | 0.4% Gastritis 2.2% | 1.8% Infusion reactions 1.1% | 0.7% Hypophysitis 0.9% | 0% Thyroiditis 0.9% | 0.2% Nephritis 0.7% | 0.9% Hepatitis 0.4% | 0% Vasculitis 0.4% | 0% Myocarditis 0.2% | 0% Myositis 0.2% | 0% Optic neuritis 0.2% | 0% Type 1 diabetes mellitus 0.2% | 0% Uveitis 0.2% | 0.2% HLH 0% | 0.2% Pancreatitis 0% | 0.2% Legend: Grade 1 or 2, Grade 3-5 (each arm) Summary n (%) | Pembro + cis + gem n = 450 | Placebo + cis + gem n = 449 Any grade | 130 (28.9) | 42 (9.4) Grade 3-5 | 40 (8.9) | 8 (1.8) Serious | 31 (6.9) | 8 (1.8) Led to death | 0 | 1 (0.2)ᵇ HLH, hemophagocytic lymphohistiocytosis. ªBased on a list of preferred terms intended to capture known risks of pembrolizumab and were considered regardless of attribution to study treatment by the investigator. ᵇDue to hemophagocytic lymphohistiocytosis. Data cutoff: January 2[illegible], 2026.

[Slide 1] 13:30 - 15:00 Proffered paper 1: Genitourinary tumour, renal and urothelial CHAIRS: AXEL BEX, ANJA LORCH Subsequent Therapy ITT Population n (%) | Pembro + cis + gem n = 453 | Placebo + cis + gem n = 454 Any | 85 (18.8) | 108 (23.8) Participants who did not undergo cystectomy | 16 (3.5) | 25 (5.5) Chemotherapy | 9 (2.0) | 13 (2.9) Radiotherapy | 9 (2.0) | 18 (4.0) Anti–PD-(L)1 | 3 (0.7) | 7 (1.5) Enfortumab vedotin | 1 (0.2) | 2 (0.4) Enfortumab vedotin plus pembrolizumab | 1 (0.2) | 0 Other | 3 (0.7) | 2 (0.4) Participants who underwent cystectomy | 69 (15.2) | 83 (18.3) Adjuvant(a) | 5 (1.1) | 10 (2.2) Anti–PD-(L)1 | 5 (1.1) | 7 (1.5) Chemotherapy | 1 (0.2) | 3 (0.7) Recurrent/metastatic(b) | 61 (13.5) | 76 (16.7) Chemotherapy | 37 (8.2) | 44 (9.7) Anti–PD-(L)1 | 22 (4.9) | 51 (11.2) Enfortumab vedotin | 19 (4.2) | 16 (3.5) Radiotherapy | 19 (4.2) | 27 (5.9) Enfortumab vedotin plus pembrolizumab | 5 (1.1) | 1 (0.2) Anti–PD-(L)1 plus chemotherapy | 1 (0.2) | 2 (0.4) Other | 5 (1.1) | 8 (1.8) Other(c) | 6 (1.3) | 4 (0.9) [left edge cut off]ubsequent oncologic therapies received after surgery for participants who underwent cystectomy to reduce the risk of disease recurrence. (b)Subsequent oncologic therapies received after cystectomy to treat recurrent or metastatic [cut off]ease. (c)Therapies given off study for participants who had off-study surgery. Data cutoff: January 26, 2026. Jens Bedke Perioperative pembrolizumab (pembro) or placebo (pbo) + neoadjuvant cisplatin (cis) and gemcitabine (gem) in participants (pts) with muscle-invasive bladder cancer (MIBC): the double-blind, phase 3 KEYNOTE-866 study BARCELONA AUDITORIUM - HALL 9 [Slide 2] 13:30 - 15:00 Proffered paper 1: Genitourinary tumour, renal and urothelial CHAIRS: AXEL BEX, ANJA LORCH Common TEAEs(a) (Incidence ≥15% in Either Arm) Safety Analysis Population, All Phases of Treatment(b) Pembro + cis + gem (n = 450) | Placebo + cis + gem (n = 449) Nausea | 54.9% | 52.3% Anemia | 54.4% | 55.2% Neutropenia | 53.1% | 50.6% Constipation | 35.8% | 39.0% Fatigue | 34.9% | 29.8% Urinary tract infection | 26.0% | 28.1% Diarrhea | 24.7% | 19.2% Decreased appetite | 24.2% | 20.7% Thrombocytopenia | 23.3% | 23.8% Blood creatinine increased | 22.9% | 20.5% Asthenia | 22.4% | 24.3% Pruritus | 20.4% | 10.9% Weight decreased | 17.8% | 12.0% Vomiting | 17.1% | 13.8% Leukopenia | 16.2% | 13.6% Pyrexia | 15.3% | 14.9% Legend: Grade 1 or 2 | Grade 3-5 (each arm) Incidence, % (axis 100 to 0 to 100) (a)Collected up to 30 days after cessation of study treatment (90 days for serious AEs). (b)Reporting window included neoadjuvant, surgical, and adjuvant phases. Data cutoff: January 26, 2026. Jens Bedke Perioperative pembrolizumab (pembro) or placebo (pbo) + neoadjuvant cisplatin (cis) and gemcitabine (gem) in participants (pts) with muscle-invasive bladder cancer (MIBC): the double-blind, phase 3 KEYNOTE-866 study BARCELONA AUDITORIUM - HALL 9 [Slide 4] 13:30 - 15:00 Proffered paper 1: Genitourinary tumour, renal and urothelial CHAIRS: AXEL BEX, ANJA LORCH Perioperative pembrolizumab or placebo plus neoadjuvant cisplatin and gemcitabine in participants with muscle-invasive bladder cancer who are eligible for cisplatin: the double-blind, phase III KEYNOTE-866 study J. Bedke, P. Arkosy, H. Nishiyama, H. Mahammedi, C. Vulsteke, M.J. Juan-Fita, I. Antonyan, O. Didas, E. Verzoni, V. Atduev, A. Rodriguez-Vida, J.J. Oh, S.-H. Kang, A. Peer, R. Zukov, B. Zurawksi, M. Gümüş, G.D. Steinberg, J. Olah, S. Abadie-Lacourtoisie, N. Jain, S. Franco, B. Homet Moreno, C.J. Hoimes, A. Siefker-Radtke PII: S0923-7534(26)06888-2 DOI: https://doi.org/10.1016/j.annonc.2026.10.106 Reference: ANNONC 2262 URL: https://www.annalsofoncology.org/article/S0923-7534(26)06888-2/fulltext [Annals of Oncology journal cover: ESMO logo, "ANNALS OF ONCOLOGY DRIVING INNOVATION IN ONCOLOGY", Elsevier logo] Jens Bedke Perioperative pembrolizumab (pembro) or placebo (pbo) + neoadjuvant cisplatin (cis) and gemcitabine (gem) in participants (pts) with muscle-invasive bladder cancer (MIBC): the double-blind, phase 3 KEYNOTE-866 study BARCELONA AUDITORIUM - HALL 9

[Journal figure with 'Journal Pre-proof' watermark] | Events, n (%) | EFS, median (95% CI), months Pembrolizumab group | 157 (34.7) | 74.1 (72.4-NR) Placebo group | 168 (37.0) | NR (67.7-NR) HR 0.88 (95% CI 0.71-1.10); P = 0.1366 Landmarks: 24 mo 76.7% vs 70.0%; 36 mo 69.2% vs 66.3%; 48 mo 66.0% vs 62.3%; 60 mo 63.5% vs 61.9% Event-free survival (%) vs Months 0-78 No. at risk Pembrolizumab group (no. censored): 453 (0) 401 (20) 366 (22) 338 (22) 329 (23) 304 (23) 296 (24) 273 (39) 224 (83) 175 (130) 114 (187) 51 (248) 11 (287) 0 (296) Placebo group (no. censored): 454 (0) 400 (20) 345 (22) 310 (23) 300 (24) 291 (25) 283 (25) 257 (43) 208 (84) 168 (123) 106 (185) 51 (238) 12 (274) 0 (286) Survival table Endpoint: EFS Arm | N | Median | Median CI Pembrolizumab group | 453 | 74.1 mo | 72.4-NR Placebo group | 454 | NR | 67.7-NR HR: 0.88 HR CI: 0.71-1.10 P: P = 0.1366 (sidedness not stated on this figure) Landmark rates: 24 mo: Pembrolizumab group 76.7% | Placebo group 70.0% 36 mo: Pembrolizumab group 69.2% | Placebo group 66.3% 48 mo: Pembrolizumab group 66.0% | Placebo group 62.3% 60 mo: Pembrolizumab group 63.5% | Placebo group 61.9%

Primary Endpoint: EFS by Investigator Assessment ITT Population Pembro + cis + gem n = 453 | Placebo + cis + gem n = 454 Events, n (%): 157 (34.7) | 168 (37.0) Median (95% CI), mo: 74.1 (72.4-NR) | NR (67.7-NR) HR 0.88 (95% CI, 0.71-1.10), 1-sided P = 0.1366* Curve labels at 24/36/48/60 months - Pembro + cis + gem: 76.7%, 69.2%, 66.0%, 63.5%; Placebo + cis + gem: 70.0%, 66.3%, 62.3%, 61.9% Y-axis: Event-Free Survival, %; X-axis: Months 0-78 No. at risk (0,6,12,...,78) Pembro + cis + gem: 453, 401, 366, 338, 329, 304, 296, 273, 224, 175, 114, 51, 11, 0 Placebo + cis + gem: 454, 400, 345, 310, 300, 291, 283, 257, 208, 168, 106, 51, 12, 0 Survival table Endpoint: EFS by investigator assessment (ITT) Arm | N | Median | Median CI Pembro + cis + gem | 453 | 74.1 mo | 72.4-NR Placebo + cis + gem | 454 | NR | 67.7-NR HR: 0.88 HR CI: 0.71-1.10 P: 1-sided P = 0.1366* Landmark rates: 24 mo: Pembro + cis + gem 76.7% | Placebo + cis + gem 70.0% 36 mo: Pembro + cis + gem 69.2% | Placebo + cis + gem 66.3% 48 mo: Pembro + cis + gem 66.0% | Placebo + cis + gem 62.3% 60 mo: Pembro + cis + gem 63.5% | Placebo + cis + gem 61.9%

Key Secondary Endpoint: pCR by Central Pathology Re[view - cut off] ITT Population(a) [Bar chart: pCR Rate, % (95% CI), y-axis 0-100] Estimated difference:(b) 8.9% (95% CI, 2.5-15.1)* Pembro + cis + gem | Placebo + cis + gem Table (right side cropped; placebo column not visible): | Pembro + cis + gem n = 450 | [cropped] pCR, n | 200 | [cropped] pCR rate, % (95% CI) | 44.4 (39.8-49.2) | [cropped] - pCR: defined as absence of viable tumo[r in] examined tissue from RC + PLND (pT0N[0]) [cropped] - Participants who did not undergo RC + [PLND were] considered nonresponders [cropped] (a)Excluding those who did not undergo RC + PLND and had a clinical CR. (b)Based on the Miettinen-Nurminen method stratified by the randomization factors. *Not formally statistically tested according to the statistical design. Data cutoff: January 26, 2026.

Perioperative Pembrolizumab or Placebo Plus Neoadjuvant Cisplatin and Gemcitabine in Parti[cipants] With Muscle-Invasive Bladder Cancer: The Dou[ble-] Blind, Phase 3 KEYNOTE-866 Study [right edge cropped] Jens Bedke(1); Peter Arkosy(2); Hiroyuki Nishiyama(3); Hakim Mahammedi(4); Christof Vulsteke(5); María José Juan-Fita(6); Igor Antonyan(7); Ossama Didas(8); Elena Verzoni(9); Vagif Atduev(10); Alejo Rodrigu[ez-Vida(11) - cropped]; Jong Jin Oh(12); Seok-Ho Kang(13); Avivit Peer(14); Ruslan Zukov(15); Nainy Jain(16); Sonia Franco(16); Blanca Homet Moreno(16); Christopher Hoimes(17); Arlene Siefker-Radtke(18) (1)Department of Urology & Transplantation Surgery, Eva Mayr-Stihl Cancer Center, Klinikum Stuttgart, Stuttgart, Germany; (2)Debreceni Egyetem Klinikai K[...], Debrecen, Hungary; (3)University of Tsukuba Hospital, Tsukuba, Japan; (4)Centre Jean Perrin, Clermont-Ferrand, France; (5)Integrated Cancer Center Ghent, [AZ] Middelares, Ghent, Belgium, and Center for Oncological Research, Antwerp University, Antwerp, Belgium; (6)Instituto Valenciano de Oncologia, Valencia, S[pain]; (7)V.I. Shapoval Regional Medical Clinical Urology and Nephrology Center of Kharkiv Reg Council, Kharkiv, Ukraine; (8)Clinique Victor Hugo, Le Mans, Franc[e]; (9)IRCCS Istituto Nazionale dei Tumori di Milano, Genitourinary Medical Oncology, Milan, Italy; (10)Volga District Medical Center, Federal Medical-Biological A[gency], Novgorod, Russia; (11)Hospital del Mar, Barcelona, Spain; (12)Seoul National University Bundang Hospital, Seongnam-si, South Korea; (13)Korea University A[...], Seoul, South Korea; (14)Fishman Oncology Center, Rambam Health Care Campus, Haifa, Israel; (15)V.F. Voyno-Yasenetsky Krasnoyarsk State Medical Univ[ersity], Krasnoyarsk, Russia; (16)Merck & Co., Inc., Rahway, NJ, USA; (17)Duke University Medical Center, Durham, NC, USA; (18)MD Anderson Cancer Center, Hous[ton] [Stage photo: presenter at podium; podium sign 'ESMO congress MADRID SPAIN 23-27 OCTOBER 2026'; chairs seated]

A | | Events, n (%) | EFS, median (95% CI), months | | Pembrolizumab group | 157 (34.7) | 74.1 (72.4-NR) | | Placebo group | 168 (37.0) | NR (67.7-NR) | HR 0.88 (95% CI 0.71-1.10); P = 0.1366 [KM curve: Event-free survival (%) vs Months 0-78] Landmarks: 24 mo 76.7% vs 70.0%; 36 mo 69.2% vs 66.3%; 48 mo 66.0% vs 62.3%; 60 mo 63.5% vs 61.9% No. at risk (months 0,6,12,18,24,30,36,42,48,54,60,66,72,78) Pembrolizumab group: 453 401 366 338 329 304 296 273 224 175 114 51 11 0 Placebo group: 454 400 345 310 300 291 283 257 208 168 106 51 12 0 Survival table Endpoint: EFS Arm | N | Median | Median CI Pembrolizumab group | 453 | 74.1 mo | 72.4-NR Placebo group | 454 | NR | 67.7-NR HR: 0.88 HR CI: 0.71-1.10 P: 0.1366 Landmark rates: 24 mo: Pembrolizumab group 76.7% | Placebo group 70.0% 36 mo: Pembrolizumab group 69.2% | Placebo group 66.3% 48 mo: Pembrolizumab group 66.0% | Placebo group 62.3% 60 mo: Pembrolizumab group 63.5% | Placebo group 61.9%

Statistical Considerations Overall type I error rate controlled at α = 2.5% (1-sided) with graphical method of Maurer and Bretz EFS (primary) 1-sided α = 0.025 → 0.36 → pCR (key secondary) 1-sided α = 0 → 0.64 → OS (key secondary) 1-sided α = 0 pCR ↔ OS: 1.0 / 1.0 IA2 was planned to occur when ~331 EFS events were observed (~60 months after first participant was randomly assigned) OS and pCR rate were to be formally statistically tested only if the EFS null hypothesis was rejected

[Slide 1] Sacituzumab Govitecan Combined With Enfortumab Vedotin and Zimberelimab in Treatment-Naive Metastatic Urothelial Cancer: Interim Results From Phase 1b TROPHY-U-01 Cohort 7 Adults with histologically documented, previously untreated, locally advanced unresectable or metastatic urothelial cancer: • Measurable disease by CT/MRI per RECIST 1.1 • Archival tumor tissue or fresh biopsy • No prior systemic therapy for locally advanced or metastatic UC (curative therapy allowed if recurrence occurred > 12 months since last systemic treatment) • ECOG performance status 0-1 Spider Plot of Tumor Response per Investigator Assessment [y-axis: Percent Change From Baseline – Target Lesions (−100 to 100); x-axis: Weeks (0-90)] Best overall response: Complete response; Partial response; Stable disease; Progressive disease [Slide 2] Sacituzumab Govitecan Combined With Enfortumab Vedotin and Zimberelimab in Treatment-Naive Metastatic Urothelial Cancer: Interim Results From Phase 1b TROPHY-U-01 Cohort 7 Participants, n (%) | Cohort 7 (n = 21) Median treatment duration (range), months SG 8.0 (0-18) EV 7.6 (0-16) ZIM 9.7 (0-20) Any TEAE/grade ≥ 3 TEAE 21 (100)/14 (67) Any TRAE/grade ≥ 3 TRAE 21 (100)/11 (52) TRAEs leading to any drug discontinuation 7 (33) Discontinuation of SG 4 (19) Discontinuation of EV 6 (29) Discontinuation of ZIM 1 (5) TEAE leading to death 0 DLTa 1/12 (8) [Butterfly chart: TEAEs, % (left) | TRAEs, % (right); light bar = all grade, dark bar = grade ≥3; values as printed: TEAE all / TEAE G≥3 | TRAE G≥3 / TRAE all] Fatigue 86 / 19 | 10 / 67 Decreased appetite 71 / 5 | 5 / 62 Diarrhea 71 / 10 | 10 / 67 Alopecia 67 / 0 | 0 / 62 Nausea 62 / 10 | 5 / 62 Constipation 57 / 10 | 5 / 29 Neuropathy peripheral 52 / 10 | 5 / 48 ALT increase 43 / 5 | 5 / 43 Dysgeusia 38 / 0 | 0 / 33 Peripheral sensory neuropathy 38 / 0 | 0 / 38 Pruritus 38 / 0 | 0 / 38 Rash 38 / 0 | 0 / 33 Anemia 33 / 10 | 0 / 19 Arthralgia 33 / 0 | 0 / 14 AST increase 33 / 0 | 0 / 33 Dry eye 33 / 0 | 0 / 24 Pyrexia 33 / 5 | 5 / 19 Vomiting 33 / 5 | 5 / 29 Abdominal pain 29 / 5 | 0 / 14 Dizziness 29 / 0 | 0 / 19 IRR 29 / 5 | 5 / 29 Lacrimation increased 29 / 0 | 0 / 24 Lipase increase 29 / 0 | 0 / 29 Amylase increase 24 / 5 | 0 / 19 Rhinorrhea 24 / 0 | 0 / 10 Stomatitis 24 / 0 | 0 / 24 Febrile neutropenia – / 10 | 10 / – Rash maculopapular 14 / 10 | 10 / 14 Spinal fracture – / 10 | – / – [Slide 3] Sacituzumab Govitecan Combined With Enfortumab Vedotin and Zimberelimab in Treatment-Naive Metastatic Urothelial Cancer: Interim Results From Phase 1b TROPHY-U-01 Cohort 7 Variable | Cohort 7 (n = 21) Median follow-up (range), months | 12.2 (1.3-20.8) Objective response rate,a n (%) (95% CI) | 17 (81) (58-95) Best overall response, n (%) Complete response | 3 (14) Partial response | 14 (67) Stable disease | 1 (5) Stable disease ≥ 6 mo | 1 (5) Progressive disease | 1 (5) Not evaluable/not assessed | 2 (10) Clinical benefit rate,b n (%) (95% CI) | 18 (86) (64-97) Median duration of responsec (95% CI), months | 16.7 (4.9-NE) Median time to responsed (range), months | 2.5 (1.2-4.2) Median PFS (95% CI), months | 19.3 (10.6-NE) Median OS (95% CI), months | NR (12.2-NE) Best Percent Change from Baseline in Total Sum of Target Lesion Diameters [Waterfall plot; y-axis: Best Percent Change From Baseline, -120 to 120; x-axis: Participant] Best overall response Complete response (n = 3) Partial response (n = 14) Stable disease (n = 14) Progressive disease (n = 1) + sign indicates treatment is still ongoing for the participant. Dotted line indicates partial response with a 30% reduction per RECIST.

[Slide 1] Sacituzumab Govitecan Combined With Enfortumab Vedotin and Zimberelimab in Treatment-Naive Metastatic Urothelial Cancer: Interim Results From Phase 1b TROPHY-U-01 Cohort 7 Randy F Sweis(1), Petros Grivas(2), Thomas Powles(3), Yohann Loriot(4), Ignacio Duran(5), Philippe Barthélémy(6), Jae Lyun Lee(7), Sang-Joon Shin(8), Andrea Necchi(9,10), Alejandro Recio-Boiles(11), Manoti Marya(12), Jana Waldes(12), Eliza Harting(12), Freda Boateng(12), Bradley A McGregor(13) (1)University of Chicago, Chicago, IL, USA; (2)University of Washington, Fred Hutch Cancer Center, Seattle, WA, USA; (3)Barts Cancer Centre, Queen Mary University of London, London, UK; (4)Gustave Roussy, Université Paris-Saclay, Villejuif, France; (5)Hospital Universitario Marqués de Valdecilla IDIVAL, Santander, Cantabria, Spain; (6)University Hospital Strasbourg, Strasbourg, France; (7)Asan Medical Center, University of Ulsan College of Medicine, Seoul, South Korea; (8)Severance Hospital, Yonsei University Health System, Seoul, South Korea; (9)Vita-Salute San Raffaele University, Milan, Italy; (10)IRCCS San Raffaele Hospital, Milan, Italy; (11)University of Arizona Cancer Center, Tucson, AZ, USA; (12)Gilead Sciences, Inc., Foster City, CA, USA; (13)Dana-Farber Cancer Institute, Boston, MA, USA Friday, October 23, 2026; 13:50-14:00 FPN-4742O [Slide 2] TROPHY-U-01 Cohort 7: Study Design Adults with histologically documented, previously untreated, locally advanced unresectable or metastatic urothelial cancer: • Measurable disease by CT/MRI per RECIST 1.1 • Archival tumor tissue or fresh biopsy • No prior systemic therapy for locally advanced or metastatic UC (curative therapy allowed if recurrence occurred > 12 months since last systemic treatment) • ECOG performance status 0-1 SG(a) + EV(b) + ZIM(c) Safety lead-in Planned N = 12 to 36 SG (7.5 mg/kg) + EV(b) + ZIM(c) Stage 1 expansion Planned N = up to 20 End points Primary • Safety and tolerability • Confirmed ORR (INV, RECIST 1.1) Secondary(d) • DOR, CBR, and PFS (INV, RECIST 1.1) • OS • DLTs were SG-related toxicities in safety lead-in participants who completed all cycle 1 dosing and 21-day evaluation period • Primary granulocyte colony stimulating factor prophylaxis from cycle 1 was mandatory during each cycle unless clinically contraindicated per investigator • At data snapshot (July 24, 2026), 21 participants (safety lead-in n = 13, expansion n = 8) were enrolled; median follow-up was 12.2 months • SG 7.5 mg/kg was used for expansion (a)SG 7.5 mg/kg administered on days 1 and 8 of a 21-day cycle (SG 7.5 mg/kg was planned for safety lead-in; based on the incidence of DLTs, 5 mg/kg and 10 mg/kg doses could be evaluated). (b)EV doses of 1.25 mg/kg are administered on days 1 and 8 of a 21-day cycle. (c)ZIM doses of 360 mg are administered on day 1 of a 21-day cycle. (d)Efficacy results in this presentation are based on INV; BICR- and iRECIST-related analyses are not included. BICR, blinded independent central review; CBR, clinical benefit rate; CT, computed tomography; DLT, dose-limiting toxicity; DOR, duration of response; ECOG, Eastern Cooperative Oncology Group; EV, enfortumab vedotin; INV, investigator assessment; iRECIST, modified RECIST 1.1 for immune-based therapeutics; MRI, magnetic resonance imaging; ORR, objective response rate; OS, overall survival; PFS, progression-free survival; RECIST, Response Evaluation Criteria in Solid Tumors; SG, sacituzumab govitecan; UC, urothelial cancer; ZIM, zimberelimab. Randy F Sweis, MD [Slide 3] Efficacy outcomes Variable | Cohort 7 (n = 21) Median follow-up (range), months | 12.2 (1.3-20.8) Objective response rate,(a) n (%) (95% CI) | 17 (81) (58-95) Best overall response, n (%) Complete response | 3 (14) Partial response | 14 (67) Stable disease | 1 (5) Stable disease ≥ 6 months | 1 (5) Progressive disease | 1 (5) Not evaluable/not assessed | 2 (10) Variable | Cohort 7 (n = 21) Clinical benefit rate,(b) n (%) (95% CI) | 18 (86) (64-97) Median DOR(c) (95% CI), months | 16.7 (4.9-NE) 9-month DOR, % (95% CI) | 89 (43-98) 12-month DOR, % (95% CI) | 76 (33-94) Median time to response(d) (range), months | 2.5 (1.2-4.2) Achieving response ≤3 months, n (%) | 14/17 (82) Median PFS (95% CI), months | 19.3 (10.6-NE) 12-month PFS, % (95% CI) | 73 (41-89) Median OS (95% CI), months | NR (12.2-NE) The combination of SG, EV, and ZIM demonstrated promising objective responses and survival outcomes Data snapshot date: July 24, 2026. Response rates and PFS were by investigator assessment. (a)Objective response rate is defined as the proportion of participants who achieved a best overall response of CR/PR. (b)Clinical benefit rate is defined as proportion of participants with CR + PR + SD ≥ 6 months. (c)Duration of response is calculated based on the time between the first date showing a documented response of CR or PR and the date of progression or death. (d)Time to response is defined as time between the first dose date to the first documented PR/CR. CR, complete response; DOR, duration of response; NE, not estimable; NR, not reached; OS, overall survival; PFS, progression-free survival; PR, partial response; SD, stable disease. Randy F Sweis, MD [Slide 4] Best Percent Change in Target Lesion Diameters and Tumor Response by Investigator Assessment Best Percent Change from Baseline in Total Sum of Target Lesion Diameters [Waterfall plot; y-axis Best Percent Change From Baseline (100 to -100); x-axis Participant] Best overall response: Complete response (n = 3); Partial response (n = 14); Stable disease (n = 1); Progressive disease (n = 1) Spider Plot of Tumor Response per Investigator Assessment [Spider plot; y-axis Percent Change From Baseline – Target Lesions (100 to -100); x-axis Weeks (0-90)] Best overall response: Complete response; Partial response; Stable disease; Progressive disease + sign indicates treatment is still ongoing for the participant. Dotted line indicates partial response with a 30% reduction per RECIST. Tumor shrinkage occurred in all evaluable participants and the responses were durable Data snapshot date: July 24, 2026. Randy F Sweis, MD

[Slide 2] 13:30 - 15:00 Proffered paper 1: Genitourinary tumour, renal and urothelial CHAIRS: AXEL BEX, ANJA LORCH Efficacy outcomes | Variable | Cohort 7 (n = 21) | Median follow-up (range), months | 12.2 (1.3-20.8) Objective response rate,^a n (%) (95% CI) | 17 (81) (58-95) Best overall response, n (%) Complete response | 3 (14) Partial response | 14 (67) Stable disease | 1 (5) Stable disease ≥ 6 months | 1 (5) Progressive disease | 1 (5) Not evaluable/not assessed | 2 (10) | Variable | Cohort 7 (n = 21) | Clinical benefit rate,^b n (%) (95% CI) | 18 (86) (64-97) Median DOR^c (95% CI), months | 16.7 (4.9-NE) 9-month DOR, % (95% CI) | 89 (43-98) 12-month DOR, % (95% CI) | 76 (33-94) Median time to response^d (range), months | 2.5 (1.2-4.2) Achieving response ≤3 months, n (%) | 14/17 (82) Median PFS (95% CI), months | 19.3 (10.6-NE) 12-month PFS, % (95% CI) | 73 (41-89) Median OS (95% CI), months | NR (12.2-NE) The combination of SG, EV, and ZIM demonstrated promising objective responses and survival outcomes Data snapshot date: July 24, 2026 Response rates and PFS were by investigator assessment. ^aObjective response rate is defined as the proportion of participants who achieved a best overall response of CR/PR. ^bClinical benefit rate is defined as proportion of participants with CR + PR + SD ≥ 6 months. ^cDuration of response is calculated based on the time between the first date showing a documented response of CR or PR and the date of progression or death. ^dTime to response is defined as time between the first dose date to the first documented PR/CR. CR, complete response; DOR, duration of response; NE, not estimable; NR, not reached; OS, overall survival; PFS, progression-free survival; PR, partial response; SD, stable disease. Randy F Sweis, MD [Speaker panel] Randy Sweis — Sacituzumab Govitecan (SG) Combined With Enfortumab Vedotin (EV) and Zimberelimab (ZIM) in Treatment-Naive Metastatic Urothelial Cancer (mUC): Interim Results From Phase 1b TROPHY-U-01 Cohort 7 BARCELONA AUDITORIUM - HALL 9 [Slide 4] 13:30 - 15:00 Proffered paper 1: Genitourinary tumour, renal and urothelial CHAIRS: AXEL BEX, ANJA LORCH Conclusions • SG + EV + ZIM showed manageable toxicity and tolerability over 12 months of median follow-up in participants with treatment-naive mUC – No treatment-related deaths were observed – Discontinuation rate due to TRAEs for any agent was 33% (highest rate was reported for EV) – The overall safety findings were consistent with the established toxicity profiles of SG, EV, and ZIM • SG + EV + ZIM demonstrated promising antitumor activity, with early and durable responses – Objective response was achieved in 81% of the participants; median DOR was 16.7 months – Median PFS was 19.3 months, and median OS was not reached • These findings support further investigation of immunotherapy combinations that include ADCs with different mechanisms of action in patients with UC – Ongoing Gilead-supported studies of SG in combination with immunotherapy include the phase 2 DAD/DAD-IO (NCT04724018) study sponsored by the DFCI, and the phase 3 ECOG-ACRIN EA8231 (NCT06524544) study DFCI, Dana-Farber Cancer Institute; EV, enfortumab vedotin; mUC, metastatic urothelial cancer; SG, sacituzumab govitecan; ZIM, zimberelimab Randy F Sweis, MD BARCELONA AUDITORIUM - HALL 9

Best Percent Change in Target Lesion Diameters and Tumor Response by Investigator Assessment Best Percent Change from Baseline in Total Sum of Target Lesion Diameters Best overall response ■ Complete response (n = 3) ■ Partial response (n = 14) ■ Stable disease (n = 1) ■ Progressive disease (n = 1) [Waterfall plot; y-axis: Best Percent Change From Baseline (100, 60, 0, -60, -100); x-axis: Participant] Spider Plot of Tumor Response per Investigator Assessment Best overall response: ■ Complete response ■ Partial response ■ Stable disease ■ Progressive disease [Spider plot; y-axis: Percent Change From Baseline – Target Lesions (100 to -100); x-axis: Weeks (0-90)] + sign indicates treatment is still ongoing for the participant. Dotted line indicates partial response with a 30% reduction per RECIST. Tumor shrinkage occurred in all evaluable participants and the responses were durable Data snapshot date: July 24, 2026. Randy F Sweis, MD [Webcast overlay:] 13:30 - 15:00 Proffered paper 1: Genitourinary tumour, renal and urothelial CHAIRS: AXEL BEX, ANJA LORCH Randy Sweis — Sacituzumab Govitecan (SG) Combined With Enfortumab Vedotin (EV) and Zimberelimab (ZIM) in Treatment-Naive Metastatic Urothelial Cancer (mUC): Interim Results From Phase 1b TROPHY-U-01 Cohort 7 BARCELONA AUDITORIUM - HALL 9

Most Common Treatment-Emergent vs Treatment-Related AEs Legend: Any grade | Grade ≥ 3 [Butterfly bar chart, TEAEs % (left) vs TRAEs % (right); AE row labels and bar values illegible at this resolution — first rows appear to be Fatigue, Decreased appetite, Diarrhea, Alopecia, Nausea [partially illegible]] • The most common treatment-related AEs were fatigue, decreased appetite, diarrhea, alopecia, and nausea • 20/21 (95%) participants received primary G-CSF prophylaxis from cycle 1. Grade 3 neutropenia(a) was reported in 3/21 participants, including 1 participant who did not receive primary G-CSF prophylaxis. All three events occurred during cycle 1 Overall safety data were consistent with the known toxicity profile of SG, EV, and ZIM [footnotes illegible] Data snapshot date: July 24, 2026 [partially legible] Randy F Sweis, MD

Most Common Treatment-Emergent vs Treatment-Related AEs ■ Any grade ■ Grade ≥ 3 [Butterfly bar chart: TEAEs, % (left) vs TRAEs, % (right); axis 100 80 60 40 20 0 20 40 60 80 100. Row labels top to bottom: Fatigue, Decreased appetite, Diarrhea, Alopecia, Nausea, Constipation, Neuropathy peripheral, ALT increase, Dysgeusia, Peripheral sensory neuropathy, Pruritus, Rash, Anemia, Arthralgia, AST increase, Dry eye, Pyrexia, Vomiting, Abdominal pain, Dizziness, IRR, Lacrimation increased, Lipase increase, Neutropenia^a, Amylase increase, Rhinorrhea, Stomatitis, Rash maculopapular, Spinal fracture. Per-bar numeric labels are printed inside bars but too small to transcribe reliably — not transcribed.] TEAEs, % TRAEs, % • The most common treatment-related AEs were fatigue, decreased appetite, diarrhea, alopecia, and nausea • 20/21 (95%) participants received primary G-CSF prophylaxis from cycle 1. Grade 3 neutropenia^a was reported in 3/21 participants, including 1 participant who did not receive primary G-CSF prophylaxis. All three events occurred during cycle 1 Overall safety data were consistent with the known toxicity profile of SG, EV, and ZIM Data snapshot date: July 24, 2026. TEAEs were included if they occurred in ≥ 20% (any-grade) or ≥ 10% (grade ≥ 3) of participants. ^aNeutropenia includes neutropenia, neutrophil count decreased, and febrile neutropenia. AE, adverse event; ALT, alanine aminotransferase; AST, aspartate aminotransferase; EV, enfortumab vedotin; G-CSF, granulocyte colony-stimulating factor; IRR, infusion-related reaction; SLI, safety lead-in; SG, sacituzumab govitecan; TEAE, treatment-emergent adverse event; TRAE, treatment-related adverse event; ZIM, zimberelimab. Randy F Sweis, MD

Demographics and Baseline Characteristics Characteristic | Cohort 7 (N = 21)ᵃ Male, n (%) | 16 (76) Median age (range), yr | 67 (39-83) ≥ 65 yr, n (%) | 12 (57) Race or ethnic group, n (%) White | 10 (48) Asian | 7 (33) Black or African-American | 1 (5) Not reported | 3 (14) ECOG PS, n (%) 0 | 14 (67) 1 | 7 (33) Tumor stage at screening, n (%) Metastatic | 16 (76) Locoregional | 5 (24) Baseline renal function, n (%) Creatinine clearance < 30 mL/min | 0 Creatinine clearance 30 to < 60 mL/min | 10 (48) Creatinine clearance ≥ 60 mL/min | 11 (52) Characteristic | Cohort 7 (N = 21)ᵃ Baseline hepatic function, n (%) Bilirubin ≤ upper limit of normal | 21 (100) Bellmunt risk factor, n (%) 0 | 12 (57) 1 | 7 (33) 2 | 2 (10) Site of disease at baseline, n (%) Visceral | 13 (62) Lung/pleura | 8 (38) Liver | 4 (19) Other visceral | 4 (19) Non-visceral | 8 (38) Lymph node only, n (%) | 6 (29) Median time since initial UC diagnosis, months (range) | 5.9 (0.3-143.1) Data snapshot date: July 24, 2026. ᵃIncludes 13 participants from safety lead in and 8 from Phase 1 expansion. ECOG PS, Eastern Cooperative Oncology Group performance status; UC, urothelial cancer. Randy F Sweis, MD

[Slide 1] TROPHY-U-01 Cohort 7: Study Design Adults with histologically documented, previously untreated, locally advanced unresectable or metastatic urothelial cancer: • Measurable disease by CT/MRI per RECIST 1.1 • Archival tumor tissue or fresh biopsy • No prior systemic therapy for locally advanced or metastatic UC (curative therapy allowed if recurrence occurred > 12 months since last systemic treatment) • ECOG performance status 0-1 SG(a) + EV(b) + ZIM(c) Safety lead-in Planned N = 12 to 36 SG (7.5 mg/kg) + EV(b) + ZIM(c) Stage 1 expansion Planned N = up to 20 End points Primary • Safety and tolerability • Confirmed ORR (INV, RECIST 1.1) Secondary(d) • DOR, CBR, and PFS (INV, RECIST 1.1) • OS • DLTs were SG-related toxicities in safety lead-in participants who completed all cycle 1 dosing and 21-day evaluation period • Primary granulocyte colony stimulating factor prophylaxis from cycle 1 was mandatory during each cycle unless clinically contraindicated per investigator • At data snapshot (July 24, 2026), 21 participants (safety lead-in n = 13, expansion n = 8) were enrolled; median follow-up was 12.2 months • SG 7.5 mg/kg was used for expansion [Slide 2] Safety Summary Participants, n (%) | Cohort 7 (n = 21) Median treatment duration (range), months SG 8.0 (0-18) EV 7.6 (0-16) ZIM 9.7 (0-20) Any TEAE/grade ≥ 3 TEAE 21 (100)/14 (67) Any TRAE/grade ≥ 3 TRAE 21 (100)/11 (52) Participants, n (%) | Cohort 7 (n = 21) TRAEs leading to any drug discontinuation 7 (33) Discontinuation of SG 4 (19) Discontinuation of EV 6 (29) Discontinuation of ZIM 1 (5) TEAE leading to death 0 DLT(a) 1/12 (8) • 3 participants died of disease progression; there were no treatment-related deaths [Slide 3] Efficacy outcomes Variable | Cohort 7 (n = 21) Median follow-up (range), months 12.2 (1.3-20.8) Objective response rate,(a) n (%) (95% CI) 17 (81) (58-95) Best overall response, n (%) Complete response 3 (14) Partial response 14 (67) Stable disease 1 (5) Stable disease ≥ 6 months 1 (5) Progressive disease 1 (5) Not evaluable/not assessed 2 (10) Variable | Cohort 7 (n = 21) Clinical benefit rate,(b) n (%) (95% CI) 18 (86) (64-97) Median DOR(c) (95% CI), months 16.7 (4.9-NE) 9-month DOR, % (95% CI) 89 (43-98) 12-month DOR, % (95% CI) 76 (33-94) Median time to response(d) (range), months 2.5 (1.2-4.2) Achieving response ≤3 months, n (%) 14/17 (82) Median PFS (95% CI), months 19.3 (10.6-NE) 12-month PFS, % (95% CI) 73 (41-89) Median OS (95% CI), months NR (12.2-NE) The combination of SG, EV, and ZIM demonstrated promising objective responses and survival outcomes Data snapshot date: July 24, 2026 [partially cut off] [Slide 4] Swimmer Plot of Tumor Response by Investigator Assessment [swimmer plot, x-axis Months 0-24] Legend: Time to first assessment; Progressive disease; Discontinued treatment; Complete response; Treatment ongoing; Not estimable; Partial response; Response assessment; Stable disease; Death date Responses were sustained or deepened over time and most participants were still on treatment at the time of data snapshot Data snapshot date: July 24, 2026.

Safety Summary Participants, n (%) | Cohort 7 (n = 21) Median treatment duration (range), months SG 8.0 (0-18) EV 7.6 (0-16) ZIM 9.7 (0-20) Any TEAE/grade ≥ 3 TEAE 21 (100)/14 (67) Any TRAE/grade ≥ 3 TRAE 21 (100)/11 (52) Participants, n (%) | Cohort 7 (n = 21) TRAEs leading to any drug discontinuation 7 (33) Discontinuation of SG 4 (19) Discontinuation of EV 6 (29) Discontinuation of ZIM 1 (5) TEAE leading to death 0 DLT^a 1/12 (8) • 3 participants died of disease progression; there were no treatment-related deaths Data snapshot date: July 24, 2026. ^a1/13 SLI participants was not DLT evaluable; the 1 DLT event was febrile neutropenia lasting < 24 hours. AE, adverse event; DLT, dose-limiting toxicity; EV, enfortumab vedotin; SLI, safety lead-in; SG, sacituzumab govitecan; TEAE, treatment-emergent adverse event; TRAE, treatment-related adverse event; ZIM, zimberelimab. Randy F Sweis, MD

[Slide 1] KEYNOTE-905/EV-303 Study (NCT03924895) Key Eligibility Criteria • Adults with muscle-invasive bladder cancer (MIBC) • Clinical stage T2-T4aN0M0 or T1-T4aN1M0 by central assessment • ≥50% Urothelial histology • Cisplatin-ineligible per Galsky criteria(a) or cisplatin-declining • ECOG PS 0-2 Stratification Factors • Cisplatin eligibility (ineligible vs. eligible but declining) • Clinical stage (T2N0 vs. T3/T4aN0 vs. T1-4aN1) • Region (US vs. EU vs. most of world) [Schema: R →] Randomization stopped in November 2022 Pembrolizumab 200 mg IV Q3W 3 cycles → RC + PLND → Pembrolizumab 200 mg IV Q3W 14 cycles → RC + PLND → Observation(b) Opened in August 2020 Enfortumab vedotin 1.25 mg/kg IV day 1 and day 8 Q3W 3 cycles + Pembrolizumab 200 mg IV Q3W 3 cycles → RC + PLND → Enfortumab vedotin 1.25 mg/kg IV day 1 and day 8 Q3W 6 cycles + Pembrolizumab 200 mg IV Q3W 14 cycles Primary endpoint: EFS by BICR for EV + pembro and RC + PLND vs. RC + PLND alone Key secondary endpoints: OS and pCR (pT0N0, i.e. absence of viable tumor in examined tissue from surgery) by central pathologist review for EV + pembro and RC + PLND vs. RC + PLND alone Other secondary endpoints include: Pathological downstaging (pDS; <pT2 and N0) by central pathologist review, safety BICR, blinded independent central review; IV, intravenous; Q3W, every 3 weeks. (a)Per protocol, defined as having ≥1 of: impaired renal function (creatinine clearance 30–59 mL/min), ECOG PS 2, gr ≥2 [remainder cut off at image edge] [Slide 2] Event-Free Survival Ad Hoc Analysis ITT Population [KM curves; y-axis: Event-free survival, %; x-axis: Months] Landmarks: 12 mo 77.2% / 66.4% / 55.0%; 24 mo 75.8% / 55.8% / 41.0%; 36 mo 69.5% / 49.6% / 37.4% | EV + pembro (n = 85) | Pembro (n = 83) | Control (n = 87) Events, n (%) | 27 (31.8) | 41 (49.4) | 52 (59.8) Median (95% CI) | NR (41.2–NR) | 28.7 mo (13.9–NR) | 15.7 mo (9.5–31.6) HR(a) vs. Control (95% CI) | 0.40 (0.25–0.65) | 0.71 (0.47–1.07) | No. at Risk (0, 6, 12, 18, 24, 30, 36, 42, 48, 54, 60 mo) EV + pembro: 85 70 59 57 55 42 35 16 3 0 0 Pembro: 83 62 52 44 41 27 22 13 3 0 0 Control: 87 56 44 37 32 22 14 9 1 0 0 mo, months; NR, not reached. (a)Comparison analysis was descriptive only (not a prespecified hypothesis test). Data cutoff date: 6 June 2025 Survival table Endpoint: Event-free survival (ad hoc analysis ITT population) Arm | N | Median | Median CI EV + pembro | 85 | NR | 41.2–NR Pembro | 83 | 28.7 mo | 13.9–NR Control (RC + PLND alone) | 87 | 15.7 mo | 9.5–31.6 HR: EV + pembro vs control 0.40; Pembro vs control 0.71 HR CI: 0.25–0.65; 0.47–1.07 Landmark rates: 12 mo: EV + pembro 77.2% | Pembro 66.4% | Control 55.0% 24 mo: EV + pembro 75.8% | Pembro 55.8% | Control 41.0% 36 mo: EV + pembro 69.5% | Pembro 49.6% | Control 37.4% [Slide 3] Overall Survival Ad Hoc Analysis ITT Population [KM curves; y-axis: Overall survival, %; x-axis: Months] Landmarks: 12 mo 83.3% / 78.2% / 74.6%; 24 mo 78.4% / 66.9% / 62.6%; 36 mo 74.5% / 61.4% / 55.7% | EV + pembro (n = 85) | Pembro (n = 83) | Control (n = 87) Events, n (%) | 23 (27.1) | 34 (41.0) | 42 (48.3) Median (95% CI) | NR (NR–NR) | 51.0 mo (32.0–NR) | 41.7 mo (26.4–NR) HR(a) vs. Control (95% CI) | 0.48 (0.29–0.80) | 0.76 (0.48–1.20) | No. at Risk (0, 6, 12, 18, 24, 30, 36, 42, 48, 54, 60 mo) EV + pembro: 85 79 70 64 63 56 45 28 10 0 0 Pembro: 83 76 64 57 53 45 31 22 7 0 0 Control: 87 72 64 54 51 45 30 18 6 0 0 mo, months; NR, not reached. (a)Comparison analysis was descriptive only (not a prespecified hypothesis test). Data cutoff date: 6 June 2025 Survival table Endpoint: Overall survival (ad hoc analysis ITT population) Arm | N | Median | Median CI EV + pembro | 85 | NR | NR–NR Pembro | 83 | 51.0 mo | 32.0–NR Control (RC + PLND alone) | 87 | 41.7 mo | 26.4–NR HR: EV + pembro vs control 0.48; Pembro vs control 0.76 HR CI: 0.29–0.80; 0.48–1.20 Landmark rates: 12 mo: EV + pembro 83.3% | Pembro 78.2% | Control 74.6% 24 mo: EV + pembro 78.4% | Pembro 66.9% | Control 62.6% 36 mo: EV + pembro 74.5% | Pembro 61.4% | Control 55.7% [Slide 4] Pathological Downstaging Ad Hoc Analysis ITT Population [Bar chart; y-axis: pDS, % (95% CI)] EV + pembro 65.9 | Pembro 33.7 | Control 14.9 | EV + pembro (n = 85) | Pembro (n = 83) | Control (n = 87) pDS, n | 56 | 28 | 13 pDS rate, % (95% CI) | 65.9 (54.8–75.8) | 33.7 (23.7–44.9) | 14.9 (8.2–24.2) Estimated(a) difference vs. Control, % (95% CI) | 51.0 (37.5–62.5) | 19.6 (6.9–32.1) | • pDS: <pT2 and N0 in examined tissue from RC + PLND by central pathologist review • (a)Determined based on stratified Miettinen & Nurminen method and may vary slightly from the numerical difference. Comparison analysis was descriptive only (not a prespecified hypothesis test). Data cutoff date: 6 June 2025

[Slide 1] Summary of Analysis Populations Present ad hoc analysis ITT population (n = 255) An ad hoc analysis was performed in participants concurrently randomized to all 3 arms between Aug 2020 and Nov 2022 Primary analysis (n = 344) All concurrently randomized participants to the EV + pembro vs. control arms, presented previously^1 [Timeline bars] Pembro monotherapy (n = 166) Control (RC + PLND alone; n = 346) EV + pembro (n = 170) 2019: Study start Aug 2020: EV + pembro arm opened Nov 2022: Enrollment to pembro arm stopped 2024: Last participant enrolled 1. Vulsteke C et al. N Engl J Med 2026. [Slide 3] Summary and Conclusions - In this ad hoc analysis of participants with MIBC who were cisplatin-ineligible or cisplatin-declining and were concurrently randomized to all 3 KEYNOTE-905 study arms: - EFS and OS favored perioperative EV + pembro compared with RC + PLND alone - Perioperative pembro was associated with numerically better EFS and OS compared with RC + PLND alone - pCR and pDS rates favored perioperative EV + pembro or pembro versus RC + PLND alone - Safety results with perioperative EV + pembro or pembro were generally consistent with the known safety profiles of the individual drugs, the underlying disease and/or surgery, or the preexisting comorbidities within the study population - The pattern of efficacy outcomes across all 3 KEYNOTE-905 study arms was consistent with a contribution of both EV and pembro to the observed activity of the perioperative EV + pembro regimen in participants with MIBC who are cisplatin-ineligible - However, these observations should be interpreted with caution due to the ad hoc nature of this analysis - Perioperative EV + pembro is approved for patients with MIBC irrespective of eligibility for cisplatin-based neoadjuvant therapy, representing a new standard of care in this setting [Slide 4] Pathological Complete Response Ad Hoc Analysis ITT Population [Bar chart] pCR, % (95% CI): EV + pembro 58.8; Pembro 25.3; Control 10.3 | | EV + pembro (n = 85) | Pembro (n = 83) | Control (n = 87) | pCR, n | 50 | 21 | 9 pCR rate, % (95% CI) | 58.8 (47.6-69.4) | 25.3 (16.4-36.0) | 10.3 (4.8-18.7) Estimated^a difference vs. Control, % (95% CI) | 48.6 (35.6-60.1) | 15.7 (4.4-27.4) | - pCR: absence of viable tumor (pT0N0) in examined tissue from RC + PLND by central pathologist review - Participants who did not undergo surgery, including those with clinical complete response after neoadjuvant therapy, were considered nonresponders ^aDetermined based on stratified Miettinen & Nurminen method and may vary slightly from the numerical difference. Comparison analysis was descriptive only (not a prespecified hypothesis test). Data cutoff date: 6 June 2025

[Slide 1] Response-adapted approach increasingly becoming clinical reality cT2–4a N0–1 MIBC Enfortumab vedotin Checkpoint inhibitor → Cystectomy Clinical CR (ctDNA neg?) → Bladder preservation? pCR (ctDNA undetectable?) → De-escalate in case of poor tolerance? Incomplete response? → Enfortumab vedotin / Checkpoint inhibitor Nonresponse (ctDNA positive?) → Trials to switch therapy? Michiel S. van der Heijden [Slide 2] Response-adapted approach increasingly becoming clinical reality cT2–4a N0–1 MIBC Enfortumab vedotin Checkpoint inhibitor → Cystectomy Clinical CR (ctDNA neg?) → Bladder preservation? pCR (ctDNA undetectable?) → De-escalate in case of poor tolerance? Incomplete response? → Enfortumab vedotin / Checkpoint inhibitor Nonresponse (ctDNA positive?) → Trials to switch therapy? Michiel S. van der Heijden [Slide 3] Response-adapted approach to the adjuvant component in perioperative therapy? cT2–4a N0–1 MIBC Enfortumab vedotin Checkpoint inhibitor → Cystectomy → pCR (ctDNA undetectable?) → Contribution of more (cytotoxic) therapy if major response achieved? Response primary tumor = micrometastases?? [Inset 1] EFS by pCR Status ITT Population KN905/EV303 Landmark labels (small, partly legible): 88.4%, 87.2%, 85.7%, 67.3%, 61.5%, 55.0%, [illegible], [illegible] No. at Risk | 0 | 6 | 12 | 18 | 24 | 30 | 36 | 42 | 48 | 54 | 60 EV + pembro (pCR) | 97 | 93 | 80 | 54 | 41 | 30 | 25 | 13 | 2 | 0 | 0 EV + pembro (no pCR) | [illegible] Control (pCR) | 15 | 12 | 12 | 6 | 6 | 3 | 2 | 1 | 0 | 0 | 0 Control (no pCR) | [illegible] [Inset 2] EFS and ctDNA in VOLGA Persistent negative | Clearance | Persistent positive Landmark labels: 91%, 91%, 88%, 81%, 70%, 55% Time to event or censor (months) Michiel S. van der Heijden

[Slide 3] Pathological Complete Response Ad Hoc Analysis ITT Population Bar chart pCR, % (95% CI): EV + pembro 58.8; Pembro 25.3; Control 10.3 | EV + pembro (n = 85) | Pembro (n = 83) | Control (n = 87) pCR, n | 50 | 21 | 9 pCR rate, % (95% CI) | 58.8 (47.6–69.4) | 25.3 (16.4–36.0) | 10.3 (4.8–18.7) Estimated(a) difference vs. Control, % (95% CI) | 48.6 (35.6–60.1) | 15.7 (4.4–27.4) | • pCR: absence of viable tumor (pT0N0) in examined tissue from RC + PLND by central pathologist review • Participants who did not undergo surgery, including those with clinical complete response after neoadjuvant therapy, were considered nonresponders [Slide 4] Most Common All-Cause AEs, All Phases AEs Shown Are Among the Five Most Frequent AEs in ≥1 Arm(a) EV + pembro (n = 84): Anemia 29.8%; Constipation 26.2%; Decreased appetite 29.8%; Diarrhea 33.3%; Fatigue 29.8%; Hypothyroidism 14.3%; Nausea 25.0%; Procedural pain 3.6%; Prostate cancer 8.3%; Pruritus 47.6%; Urinary tract infection 23.8% Pembro (n = 82): Anemia 34.1%; Constipation 13.4%; Decreased appetite 13.4%; Diarrhea 11.0%; Fatigue 14.6%; Hypothyroidism 17.1%; Nausea 14.6%; Procedural pain 8.5%; Prostate cancer 12.2%; Pruritus 18.3%; Urinary tract infection 34.1% Control (n = 79): Anemia 11.4%; Constipation 6.3%; Decreased appetite 3.8%; Diarrhea 0%; Fatigue 2.5%; Hypothyroidism 0%; Nausea 7.6%; Procedural pain 6.3%; Prostate cancer 11.4%; Pruritus 0%; Urinary tract infection 10.1% X-axis: Incidence, % Legend: Grade 1 to 2 AEs; Grade ≥3 AEs

[Slide 1] Summary of Analysis Populations Future pembro vs. control analysis (n = 338) The study continues to evaluate prespecified hypotheses under multiplicity control for concurrently randomized participants to pembro vs. control Present ad hoc analysis ITT population (n = 255) An ad hoc analysis was performed in participants concurrently randomized to all 3 arms between Aug 2020 and Nov 2022 Primary analysis (n = 344) All concurrently randomized participants to the EV + pembro vs. control arms, presented previously^1 [Timeline bars] Pembro monotherapy (n = 166) Control (RC + PLND alone; n = 346) EV + pembro (n = 170) 2019: Study start Aug 2020: EV + pembro arm opened Nov 2022: Enrollment to pembro arm stopped 2024: Last participant enrolled 1. Vulsteke C et al. N Engl J Med 2026 [Speaker panel] Anders Ullén - Perioperative pembrolizumab (pembro) or enfortumab vedotin (EV) plus pembro in pts with muscle-invasive bladder cancer (MIBC) who are cisplatin ineligible: ad hoc analysis from KEYNOTE-905 [Slide 2] Event-Free Survival Ad Hoc Analysis ITT Population | | EV + pembro (n = 85) | Pembro (n = 83) | Control (n = 87) | Events, n (%) | 27 (31.8) | 41 (49.4) | 52 (59.8) Median (95% CI) | NR (41.2-NR) | 28.7 mo (13.9-NR) | 15.7 mo (9.5-31.6) HR^a vs. Control (95% CI) | 0.40 (0.25-0.65) | 0.71 (0.47-1.07) | Landmark labels (EV + pembro / Pembro / Control): 12 mo: 77.2% / 66.4% / 55.0% 24 mo: 75.8% / 55.8% / 41.0% 36 mo: 69.5% / 49.6% / 37.4% No. at Risk (Months 0, 6, 12, 18, 24, 30, 36, 42, 48, 54, 60) EV + pembro: 85 70 59 57 55 42 35 16 3 0 0 Pembro: 83 62 52 44 41 27 22 13 3 0 0 Control: 87 56 44 37 32 22 14 9 1 0 0 mo, months; NR, not reached. ^aComparison analysis was descriptive only (not a prespecified hypothesis test). Data cutoff date: 6 June 2025 [low-res, best reading] Survival table Endpoint: EFS (ad hoc ITT, 3 concurrently randomized arms) Arm | N | Median | Median CI EV + pembro | 85 | NR | 41.2-NR Pembro | 83 | 28.7 mo | 13.9-NR Control (RC + PLND alone) | 87 | 15.7 mo | 9.5-31.6 HR: EV + pembro vs control 0.40; Pembro vs control 0.71 HR CI: 0.25-0.65; 0.47-1.07 Landmark rates: 12 mo: EV + pembro 77.2% | Pembro 66.4% | Control 55.0% 24 mo: EV + pembro 75.8% | Pembro 55.8% | Control 41.0% 36 mo: EV + pembro 69.5% | Pembro 49.6% | Control 37.4% [Slide 3] Overall Survival Ad Hoc Analysis ITT Population | | EV + pembro (n = 85) | Pembro (n = 83) | Control (n = 87) | Events, n (%) | 23 (27.1) | 34 (41.0) | 42 (48.3) Median (95% CI) | NR (NR-NR) | 51.0 mo (32.0-NR) | 41.7 mo (26.4-NR) HR^a vs. Control (95% CI) | 0.48 (0.29-0.80) | 0.76 (0.48-1.20) | Landmark labels (EV + pembro / Pembro / Control): 12 mo: 83.3% / 78.2% / 74.6% 24 mo: 78.4% / 66.9% / 62.6% 36 mo: 74.5% / 61.4% / 55.7% No. at Risk (Months 0, 6, 12, 18, 24, 30, 36, 42, 48, 54, 60) EV + pembro: 85 79 70 64 63 56 45 28 10 0 0 Pembro: 83 76 64 57 53 45 31 22 7 0 0 Control: 87 72 64 54 51 45 30 18 6 0 0 mo, months; NR, not reached. ^aComparison analysis was descriptive only (not a prespecified hypothesis test). Data cutoff date: 6 June 2025 [low-res, best reading] Survival table Endpoint: OS (ad hoc ITT, 3 concurrently randomized arms) Arm | N | Median | Median CI EV + pembro | 85 | NR | NR-NR Pembro | 83 | 51.0 mo | 32.0-NR Control (RC + PLND alone) | 87 | 41.7 mo | 26.4-NR HR: EV + pembro vs control 0.48; Pembro vs control 0.76 HR CI: 0.29-0.80; 0.48-1.20 Landmark rates: 12 mo: EV + pembro 83.3% | Pembro 78.2% | Control 74.6% 24 mo: EV + pembro 78.4% | Pembro 66.9% | Control 62.6% 36 mo: EV + pembro 74.5% | Pembro 61.4% | Control 55.7%

Pathological Complete Response Ad Hoc Analysis ITT Population [Bar chart] y-axis: pCR, % (95% CI); 0-100 EV + pembro: 58.8 Pembro: 25.3 Control: 10.3 [Table, right side cropped] | | EV + pembro (n = 85) | [cropped] | pCR, n | 50 pCR rate, % (95% CI) | 58.8 (47.6-69.4) Estimated^a difference vs. Control, % (95% CI) | 48.6 (35.6-60.1) - pCR: absence of viable tumor ([cropped]) from RC + PLND by central path[cropped] - Participants who did not underg[cropped] with clinical complete response[cropped] were considered nonresponde[cropped] ^aDetermined based on stratified Miettinen & Nurminen method and may vary slightly from the numerical difference. Comparison analysis was descriptive only (not a prespecified hypothesis test).

Summary of Analysis Populations Future pembro vs. control analysis (n = 338) The study continues to evaluate prespecified hypotheses under multiplicity control for concurrently randomized participants to pembro vs. control Present ad hoc analysis ITT population (n = 255) An ad hoc analysis was performed in participants concurrently randomized to all 3 arms between Aug 2020 and Nov 2022 Primary analysis (n = 344) All concurrently randomized participants to the EV + pembro vs. control arms, presented previously(1) Pembro monotherapy (n = 166) Control (RC + PLND alone; n = 346) EV + pembro (n = 170) [Timeline] 2019: Study start | Aug 2020: EV + pembro arm opened | Nov 2022: Enrollment to pembro arm stopped | 2024: Last participant enrolled 1. Vulsteke C et al. N Engl J Med 2026.

PD-1 Blockade Shows Similar Benefit Across Perioperative Timings Perioperative Pembro (EV-303) vs Adjuvant Nivo and Pembro [Panel 1] EV-303 · perioperative pembro - EFS · blue = pembro, red = RC + PLND alone [small KM thumbnail; landmark labels illegible at this resolution] [Panel 2] CheckMate 274 · adjuvant nivo - DFS vs placebo (ITT) [KM thumbnail: Nivolumab, Placebo] [Panel 3] AMBASSADOR · adjuvant pembro - DFS vs observation (ITT) [KM thumbnail: Pembrolizumab, Observation] Hazard ratio vs no active IO (95% CI) EFS for EV-303; DFS for adjuvant trials EV-303 Perioperative pembro 0.71 (0.47-1.07) CheckMate 274 Adjuvant nivo 0.70 (0.55-0.90) AMBASSADOR Adjuvant pembro 0.73 (0.59-0.90) Axis 0.4 0.6 0.8 1.0 1.2: <- Favours PD-1 blockade | Favours control -> Footnote (very small, best-effort): HRs as reported; not directly comparable: EV-303 EFS counted from randomisation before surgery in cisplatin-ineligible patients; CheckMate 274 and AMBASSADOR DFS counted from randomisation after surgery in high-risk pathology. CheckMate 274 HR 98.22% CI. EV-303 pembro: 3 neoadjuvant + 14 adjuvant cycles; adjuvant arms ~1 year. Cross-trial comparison of different populations, endpoints and time zero; not a formal test of treatment timing.

[Slide 1] Timeline of Study Design Development 2019 — Study initiated with 2 arms: • Perioperative pembro with RC + PLND vs. RC + PLND alone • Randomization was 1:1 • Participants had to be cisplatin-ineligible 2020 — Third arm added: • Perioperative EV + pembro with RC + PLND • Randomization was updated to 1:1:1 2022 — Inclusion criteria updated to include participants who were cisplatin-eligible but declined cisplatin-based chemotherapy 2022 — • Randomization to pembro with RC + PLND was stopped • Randomization updated to 1:1 to perioperative EV + pembro with RC + PLND vs. RC + PLND alone • Adjuvant nivolumab permitted in the RC + PLND alone arm when clinically indicated and regionally available Vulsteke C et al. N Engl J Med 2026. [Slide 3] Overall Survival Ad Hoc Analysis ITT Population | | EV + pembro (n = 85) | Pembro (n = 83) | Control (n = 87) | | Events, n (%) | 23 (27.1) | 34 (41.0) | 42 (48.3) | | Median (95% CI) | NR (NR–NR) | 51.0 mo (32.0–NR) | 41.7 mo (26.4–NR) | | HR^a vs. Control (95% CI) | 0.48 (0.29–0.80) | 0.76 (0.48–1.20) | | Landmarks: 12 mo: 83.3% / 78.2% / 74.6%; 24 mo: 78.4% / 66.9% / 62.6%; 36 mo: 74.5% / 61.4% / 55.7% (EV + pembro / Pembro / Control) Y-axis: Overall survival, %; X-axis: Months No. at Risk (months 0, 6, 12, 18, 24, 30, 36, 42, 48, 54, 60) EV + pembro: 85, 79, 70, 64, 63, 56, 45, 28, 10, 0, 0 Pembro: 83, 76, 64, 57, 53, 45, 31, 22, 7, 0, 0 Control: 87, 72, 64, 54, 51, 45, 30, 18, 6, 0, 0 Survival table Endpoint: Overall survival (ad hoc, ITT) Arm | N | Median | Median CI EV + pembro | 85 | NR | NR–NR Pembro | 83 | 51.0 mo | 32.0–NR Control | 87 | 41.7 mo | 26.4–NR HR: EV + pembro vs control 0.48; Pembro vs control 0.76 HR CI: 0.29–0.80; 0.48–1.20 Landmark rates: 12 mo: EV + pembro 83.3% | Pembro 78.2% | Control 74.6% 24 mo: EV + pembro 78.4% | Pembro 66.9% | Control 62.6% 36 mo: EV + pembro 74.5% | Pembro 61.4% | Control 55.7%

OS: ADT+Docetaxel population (n=710) | SOC+Abi (n = 355) | SOC (n = 355) Median, y (95.1% CI) | 6.2 (5.2-7.3) | 4.4 (4.0-4.9) 7-year OS, % (95.1% CI) | 45.7 (40.3-51.2) | 30.7 (25.8-36.1) 8-year OS, % (95.1% CI) | 40.3 (34.3-46.[illegible]) | 27.1 (22.0-32.9) Events | 192 | 240 Adjusted HR* (95.1% CI) | 0.68 (0.56-0.82) P value | <0.0001 [Curve annotations: Triplet; Doublet; 1.8 y; 46%; 31%] Overall survival (%) / Time from randomization (in years) 0-8 Number at risk (censored): No abiraterone: 355(0) 329(4) 281(6) 235(7) 189(8) 153(9) 111(24) 55(66) 20(96) Abiraterone: 355(0) 329(3) 288(5) 257(6) 221(7) 194(10) 156(32) 91(81) 36(129) *Cox model adjusted for prostate RT, ECOG, type of castration and disease burden Karim Fizazi Survival table Endpoint: OS (ADT+docetaxel population, n=710) Arm | N | Median | Median CI SOC+Abi (triplet) | 355 | 6.2 y | 5.2-7.3 (95.1% CI) SOC (doublet) | 355 | 4.4 y | 4.0-4.9 (95.1% CI) HR: 0.68 (adjusted) HR CI: 0.56-0.82 (95.1% CI) P: <0.0001 Landmark rates: 7 years: SOC+Abi 45.7% (CI 40.3-51.2) | SOC 30.7% (CI 25.8-36.1) 8 years: SOC+Abi 40.3% (CI 34.3-46.[illegible]) | SOC 27.1% (CI 22.0-32.9)

[Slide 1] Overall Survival (OS): Overall population (n=1172) | SOC (+/- RXT)+Abi (n = 583) | SOC (+/- RXT) (n = 589) Median, y (95.1% CI) | 6.1 (5.3-6.6) | 4.6 (4.2-5.2) 7-year OS, % (95.1% CI) | 44.9 (40.8-49.2) | 34.2 (30.3-38.3) 8-year OS, % (95.1% CI) | 40.3 (36.0-44.8) | 29.5 (25.6-33.8) Events | 330 | 399 Adjusted HR* (95.1% CI) | 0.75 (0.65-0.87) P value | 0.0001 1.5 y Restricted Mean Survival Time difference at 8 years = 0.48y (0.17-0.79) P value RMST test of equality = 0.0021 Overall survival (%) vs Time from randomization (in years) 0-8 Number at risk (censored) No abiraterone 589(0) 556(5) 481(7) 396(10) 327(12) 273(14) 213(30) 140(76) 77(123) Abiraterone 583(0) 542(7) 470(12) 414(16) 360(18) 314(21) 261(43) 184(93) 104(158) Survival table Endpoint: Overall survival, overall population (n=1172) Arm | N | Median | Median CI SOC (+/- RXT)+Abi | 583 | 6.1 y | 5.3-6.6 SOC (+/- RXT) | 589 | 4.6 y | 4.2-5.2 HR: 0.75 (adjusted) HR CI: 95.1% CI 0.65-0.87 P: 0.0001 Landmark rates: 7-year: SOC (+/- RXT)+Abi 44.9% (CI 40.8-49.2) | SOC (+/- RXT) 34.2% (CI 30.3-38.3) 8-year: SOC (+/- RXT)+Abi 40.3% (CI 36.0-44.8) | SOC (+/- RXT) 29.5% (CI 25.6-33.8) [Slide 2] | SOC+Abi (n = 355) | SOC (n = 355) Median, y (95.1% CI) | 6.2 (5.2-7.3) | 4.4 (4.0-4.9) 7-year OS, % (95.1% CI) | 45.7 (40.3-51.2) | 30.7 (25.8-36.1) 8-year OS, % (95.1% CI) | 40.3 (34.3-46.) | 27.1 (22.0-32.9) Events | 192 | 240 Adjusted HR* (95.1% CI) | 0.68 (0.56-0.82) P value | <0.0001 1.8 y Triplet | Doublet 46% | 31% Overall survival (%) vs Time from randomization (in years) 0-8 Number at risk (censored) No abiraterone 355(0) 329(4) 281(6) 235(7) 189(8) 153(9) 111(24) 55(66) 20(96) Abiraterone 355(0) 329(3) 288(5) 257(6) 221(7) 194(10) 156(32) 91(81) 36(129) Survival table Endpoint: Overall survival, ADT + docetaxel population (triplet vs doublet) Arm | N | Median | Median CI SOC+Abi (Triplet) | 355 | 6.2 y | 5.2-7.3 SOC (Doublet) | 355 | 4.4 y | 4.0-4.9 HR: 0.68 (adjusted) HR CI: 95.1% CI 0.56-0.82 P: <0.0001 Landmark rates: 7-year: SOC+Abi 45.7% (CI 40.3-51.2) | SOC 30.7% (CI 25.8-36.1) 8-year: SOC+Abi 40.3% (CI 34.3-46. (as printed)) | SOC 27.1% (CI 22.0-32.9) [Slide 3] High volume | SOC+Abi (n = 224) | SOC (n = 231) Median, y (95.1% CI) | 4.6 (3.9-6.1) | 3.5 (3.2-4.0) 7-year OS, % (95.1% CI) | 40.9 (34.4-47.7) | 22.4 (17.2-28.7) Events | 132 | 173 Adjusted HR* (95.1% CI) | 0.67 (0.54-0.85) P value | 0.0007 1.1 y | Triplet | Doublet | 41% | 22% Number at risk (censored) No abiraterone 231(0) 210(3) 170(4) 137(5) 97(6) 75(7) 54(14) 28(32) 12(46) Abiraterone 224(0) 202(3) 172(3) 146(4) 119(4) 106(5) 83(18) 53(43) 23(70) Low volume | SOC+Abi (n = 131) | SOC (n = 124) Median, y (95.1% CI) | 7.8 (6.4-NE) | 6.2 (5.4-8.1) 7-year OS, % (95.1% CI) | 54.2 (45.0-63.1) | 46.0 (36.9-55.4) Events | 67 | 60 Adjusted HR* (95.1% CI) | 0.74 (0.51-1.05) P value | 0.091 1.6 y | Triplet | Doublet Number at risk (censored) No abiraterone 124(0) 119(1) 111(2) 98(2) 92(2) 78(2) 57(10) 27(34) Abiraterone 131(0) 127(0) 116(2) 111(2) 102(3) 88(5) 73(14) 38(38) Survival table Endpoint: Overall survival by disease volume (triplet vs doublet) Arm | N | Median | Median CI High volume SOC+Abi | 224 | 4.6 y | 3.9-6.1 High volume SOC | 231 | 3.5 y | 3.2-4.0 Low volume SOC+Abi | 131 | 7.8 y | 6.4-NE Low volume SOC | 124 | 6.2 y | 5.4-8.1 HR: High volume 0.67; Low volume 0.74 (adjusted) HR CI: High 0.54-0.85; Low 0.51-1.05 (95.1% CI) P: High 0.0007; Low 0.091 Landmark rates: 7-year: High volume SOC+Abi 40.9% (CI 34.4-47.7) | High volume SOC 22.4% (CI 17.2-28.7) | Low volume SOC+Abi 54.2% (CI 45.0-63.1) | Low volume SOC 46.0% (CI 36.9-55.4)

[Slide 1] Design of PEACE-1 (2x2) Key Eligibility Criteria De novo mCSPC ECOG PS 0-2 On-Study Requirement Continuous ADT Permitted ADT <= 3 months Stratification ECOG PS (0 vs 1-2) Metastatic sites (LN vs bone vs visceral) Type of castration (orchidectomy vs LHRH agonist vs LHRH antagonist) Docetaxel (yes vs no) Nov 2013 - Dec 2018 RANDOMIZATION 1:1:1:1 n = 1173 SOC (n = 296) SOC+Abiraterone (n = 292) SOC+Radiotherapy (n = 293) SOC+Abiraterone+Radiotherapy (n = 292) ECOG PS, Eastern Cooperative Oncology Group performance status [Map of Europe highlighting participating countries: Ireland, France, Belgium, Switzerland, Spain, Italy, Romania] Karim Fizazi [Slide 4] rPFS by Metastatic extent (« Volume »): ADT+Docetaxel population (n=710) High volume: SOC (+/- RXT)+Abi (n =224) | SOC (+/- RXT) (n =231) Median, y (99.9% CI) 3.2 (2.1-5.1) | 1.5 (1.2-2.0) 7-year rPFS, % (99.9% CI) 33.4 (23.7-44.7) | 10.9 (5.7-19.8) Events 148 | 201 Adjusted HR* (99.9% CI) 0.50 (0.35-0.73) P value <0.0001 Number at risk (censored) No abiraterone 231(0) 162(3) 92(3) 66(4) 45(5) 32(6) 23(8) 13(17) Abiraterone 224(0) 181(3) 139(3) 116(4) 99(4) 88(4) 68(15) 43(36) Low volume: SOC (+/- RXT)+Abi (n =131) | SOC (+/- RXT) (n =124) Median, y (99.9% CI) 5.9 (3.1-NE) | 2.7 (2.0-4.8) 7-year rPFS, % (99.9% CI) 43.4 (29.6-58.2) | 20.9 (10.9-36.4) Events 74 | 95 Adjusted HR* (99.9% CI) 0.54 (0.32-0.91) P value <0.0001 Number at risk (censored) No abiraterone 124(0) 110(1) 80(2) 57(2) 50(2) 43(2) 29(5) 13(17) Abiraterone 131(0) 123(0) 111(2) 89(2) 75(3) 66(4) 53(12) 30(30) [KM curves labeled Triplet / Doublet; panels "High volume" and "Low volume"] Radiographic progression-free survival (%); Time from randomization (in years) *Cox models adjusted for prostate RT, ECOG, and type of castration Karim Fizazi Survival table Endpoint: rPFS by metastatic volume (ADT+Docetaxel population, n=710) Arm | N | Median | Median CI High volume: SOC (+/- RXT)+Abi | 224 | 3.2 y | 2.1-5.1 (99.9% CI) High volume: SOC (+/- RXT) | 231 | 1.5 y | 1.2-2.0 (99.9% CI) Low volume: SOC (+/- RXT)+Abi | 131 | 5.9 y | 3.1-NE (99.9% CI) Low volume: SOC (+/- RXT) | 124 | 2.7 y | 2.0-4.8 (99.9% CI) HR: High volume 0.50; Low volume 0.54 (adjusted) HR CI: High 0.35-0.73; Low 0.32-0.91 (99.9% CI) P: <0.0001 (both) Landmark rates: 7 years: High volume SOC+Abi 33.4% (CI 23.7-44.7) | High volume SOC 10.9% (CI 5.7-19.8) | Low volume SOC+Abi 43.4% (CI 29.6-58.2) | Low volume SOC 20.9% (CI 10.9-36.4)

Conclusion: PEACE-1 • Survival is clearly improved by abiraterone in the long-term • Almost twice more long-term survivors with triplet systemic therapy in subgroups ! • Triplet systemic therapy active regardless of metastatic « volume » • Early is truly better in mCSPC • Almost systematic use of an ARPI beyond progression in the control arm of PEACE-1 • No long-term safety signal Karim Fizazi

OS by Metastatic extent (« Volume ») ADT+Docetaxel population (n=710) High volume: | SOC+Abi (n = 224) | SOC (n = 231) Median, y (95.1% CI) | 4.6 (3.9-6.1) | 3.5 (3.2-4.0) 7-year OS, % (95.1% CI) | 40.9 (34.4-47.7) | 22.4 (17.2-28.7) Low volume: | SOC+Abi (n = 131) | SOC (n = 124) Median, y (95.1% CI) | 7.8 (6.4-NE) | 6.2 (5.4-8.1) 7-year OS, % (95.1% CI) | 54.2 (45.0-63.1) | 46.0 (36.9-55.4) No Interaction between effect of Triplet treatment and « Volume » (Interaction test: p=0.5552) [High volume panel: Triplet vs Doublet curves; annotations 1.1 y (median gap), 41% vs 22%] [Low volume panel: Triplet vs Doublet curves; annotation 1.6 y] Overall survival (%) vs Time from randomization (in years) Number at risk (censored): [illegible] Karim Fizazi *Cox models adjusted for prostate RT, ECOG and type of castration Survival table Endpoint: Overall survival by metastatic volume (ADT+docetaxel population, n=710) Arm | N | Median | Median CI High volume SOC+Abi (triplet) | 224 | 4.6 y | 3.9-6.1 High volume SOC (doublet) | 231 | 3.5 y | 3.2-4.0 Low volume SOC+Abi (triplet) | 131 | 7.8 y | 6.4-NE Low volume SOC (doublet) | 124 | 6.2 y | 5.4-8.1 P: Interaction test p=0.5552 Landmark rates: 7 years: High volume SOC+Abi 40.9% (CI 34.4-47.7) | High volume SOC 22.4% (CI 17.2-28.7) | Low volume SOC+Abi 54.2% (CI 45.0-63.1) | Low volume SOC 46.0% (CI 36.9-55.4)

But.....there is still room for improvement rPFS: PEACE-1 de novo high volume | rPFS: PEACE-1 de novo low volume [KM curves: Radiographic progression-free survival vs Time from randomization (in years); Triplet vs Doublet; High volume annotations 3[?]% (Triplet) vs 11% (Doublet); Low volume annotations 43% (Triplet) vs 21% (Doublet)] Number at risk (censored): High volume Abiraterone 224(5) 181(3) 139(3) 116(4) 99(4) 88(4) 68(15) 43(36); No abiraterone row [illegible]; Low volume rows [illegible] Early progressors despite «triplet therapy» → Precision treatment / intensification • BRCA/HRD alterated → benefit PARPi in mCSPC1,2 • PTEN loss → benefit Akt inhibitor3 • Potential benefit from early 177Lutetium-PSMA-6174,5 • Poor responders PSA > 0.2 ng/mL at nadir → clinical trial (e.g. PEACE-6 Poor Responders6, TRIPLE-SWITCH7, REINFORCE8, EvoPAR-PR059) Excellent responders – e.g. PSA nadir ≤ 0.2 ng/mL → De-escalation • Clinical trial: De-escalation (EORTC 2238 DE-ESCALATE10) Fizazi K et al., ESMO 2026 References: Attard G et al., Nat Med 2025; Agarwal N et al., N Engl J Med 2026; Fizazi K et al., Ann Oncol 2026; STAMPEDE2 (177Lu-PSMA-617) NCT06320067; Tagawa ST et al., Lancet 2026; PEACE-6 Poor Responders (177Lu-PSMA-617) NCT06496581; TRIPLE-SWITCH (docetaxel) NCT06592924; REINFORCE (docetaxel) NCT07333060; EvoPAR-PR05 (saruparib) NCT07071002; EORTC 2238 DE-ESCALATE NCT05974774 [small print; NCT digits uncertain] Ursula Vogl Survival table Endpoint: Radiographic PFS, PEACE-1 de novo mCSPC (triplet vs doublet) by volume High volume Triplet (abiraterone) | 224 High volume Doublet Low volume Triplet Low volume Doublet Landmark rates: ~7 y (arrow position): High volume Triplet 3[?]% | High volume Doublet 11% | Low volume Triplet 43% | Low volume Doublet 21%

OS by Metastatic extent (« Volume ») ADT+Docetaxel population (n=710) High volume [KM: Triplet vs Doublet; annotation '1.1 y'] SOC+Abi (n = 224) | SOC (n = 231) Median, y (95.1% CI): 4.6 (3.9-6.1) | 3.5 (3.2-4.0) 7-year OS, % (95.1% CI): 40.9 (34.4-47.7) | 22.4 (17.2-28.7) Events: 132 | 173 Adjusted HR* (95.1% CI): 0.67 (0.54-0.85) P value: 0.0007 Number at risk (censored): No abiraterone 231(0) 210(3) 170(4) 137(5) 97(6) 75(7) 54(14) 28(32) 12(46); Abiraterone 224(0) 202(3) 172(3) 146(4) 119(4) 106(5) 83(18) 53(43) 23(70) Low volume [KM: Triplet vs Doublet; annotation '1.6 y'] SOC+Abi (n = 131) | SOC (n = 124) Median, y (95.1% CI): 7.8 (6.4-NE) | 6.2 (5.4-8.1) 7-year OS, % (95.1% CI): 54.2 (45.0-63.1) | 46.0 (36.9-55.4) Events: 67 | 60 Adjusted HR* (95.1% CI): 0.74 (0.51-1.05) P value: 0.091 Number at risk (censored): No abiraterone 124(0) 119(1) 111(2) 98(2) 92(2) 78(2) 57(10) 27(34); Abiraterone 131(0) 127(0) 116(2) 111(2) 102(3) 88(5) 73(14) 38(38) Karim Fizazi *Cox models adjusted for prostate RT, ECOG and type of castration Survival table Endpoint: OS by metastatic volume, ADT+docetaxel population Arm | N | Median | Median CI High volume SOC+Abi | 224 | 4.6 y | 3.9-6.1 High volume SOC | 231 | 3.5 y | 3.2-4.0 Low volume SOC+Abi | 131 | 7.8 y | 6.4-NE Low volume SOC | 124 | 6.2 y | 5.4-8.1 HR: High volume 0.67; Low volume 0.74 HR CI: 0.54-0.85; 0.51-1.05 (95.1% CI, adjusted) P: 0.0007; 0.091 Landmark rates: 7 years: High volume SOC+Abi 40.9% (CI 34.4-47.7) | High volume SOC 22.4% (CI 17.2-28.7) | Low volume SOC+Abi 54.2% (CI 45.0-63.1) | Low volume SOC 46.0% (CI 36.9-55.4)

[Slide 1] Key Eligibility Criteria De novo mCSPC ECOG PS 0 -2 On-Study Requirement Continuous ADT Permitted ADT ≤ 3 months Stratification ECOG PS (0 vs 1-2) Metastatic sites (LN vs bone vs visceral) Type of castration (orchidectomy vs LHRH agonist vs LHRH antagonist) Docetaxel (yes vs no) Nov 2013 – Dec 2018 RANDOMIZATION 1:1:1:1 n = 1173 SOC (n = 296) SOC+Abiraterone (n = 292) SOC+Radiotherapy (n = 293) SOC+Abiraterone+ Radiotherapy (n = 292) ECOG PS, Eastern Cooperative Oncology Group performance status [Slide 2] SOC+Abi (n = 583) | SOC (n = 589) Median, y (95% CI) | 4.3 (3.2-5.5) | 2.2 (1.8-2.6) Events | 377 | 489 HR (95% CI)* | 0.55 (0.44-0.70) P value | <0.0001 2.1 y Y-axis: Radiographic progression-free survival (%) Number at risk (censored) (years 0-8) No abiraterone 589(0) 451(5) 308(6) 223(8) 170(10) 137(12) 111(17) 69(41) 39(65) Abiraterone 583(0) 496(7) 406(10) 334(14) 293(15) 257(16) 212(35) 154(75) 89(126) Survival table Endpoint: Radiographic progression-free survival Arm | N | Median | Median CI SOC+Abi | 583 | 4.3 y | 3.2-5.5 SOC | 589 | 2.2 y | 1.8-2.6 HR: 0.55 HR CI: 0.44-0.70 P: <0.0001 [Slide 4] SOC+Abi (n = 355) | SOC (n = 355) Median, y (99.9% CI) | 4.1 (3.0-5.9) | 2.0 (1.6-2.5) 7-year rPFS, % (99.9% CI) | 37.0 (28.8-46.1) | 14.5 (9.2-22.2) Events | 222 | 296 Adjusted HR* (99.9% CI) | 0.50 (0.38-0.68) P value | <0.0001 2.1 y Triplet Doublet 37% 15% Y-axis: Radiographic progression-free survival (%) Number at risk (censored) (years 0-7) No abiraterone 355(0) 272(4) 172(5) 123(6) 95(7) 75(8) 52(13) 26(34) Abiraterone 355(0) 304(3) 250(5) 205(6) 174(7) 154(8) 121(27) 73(66) Survival table Endpoint: Radiographic progression-free survival (Triplet vs Doublet population, n=710) Arm | N | Median | Median CI SOC+Abi (Triplet) | 355 | 4.1 y | 3.0-5.9 SOC (Doublet) | 355 | 2.0 y | 1.6-2.5 HR: 0.50 (adjusted) HR CI: 0.38-0.68 (99.9% CI) P: <0.0001 Landmark rates: 7 years: SOC+Abi 37.0% (CI 28.8-46.1) | SOC 14.5% (CI 9.2-22.2)

[Slide 1] Median follow-up and OS of combination phase 3 trials in mCSPC reported before ESMO 2026 Strategy | Trial (n) | Experimental treatment | Median FU | Median OS, experimental | HR (95% CI) ADT + docetaxel | GETUG-AFU 15 (n=385)1 | ADT + 6x docetaxel | 7.0 y | 5.2 y | 0.88 (0.68–1.14) ADT + docetaxel | CHAARTED (n=790)2 | ADT + 6x docetaxel | 9.7 y | 5.0 y | 0.77 (0.65–0.92) ADT + docetaxel | STAMPEDE arm C, M1 (n=1,086)3 | ADT + 6x docetaxel | 6.5 y | 4.9 y | 0.81 (0.69–0.95) ADT + ARPI | LATITUDE (n=1,199)4 | ADT + abiraterone | 4.3 y | 4.4 y | 0.66 (0.56–0.78) ADT + ARPI | STAMPEDE arm G, M1 (n=1,003)5 | ADT + abiraterone | 6.1 y | 6.6 y | 0.60 (0.50–0.71) ADT + ARPI | ARCHES (n=1,150)6 | ADT + enzalutamide | 5.1 y | NR (5-y OS 66%) | 0.70 (0.58–0.85) ADT + ARPI | ENZAMET (n=1,125)7 | ADT + enzalutamide | 8.1 y | 7.9 y | 0.73 (0.63–0.86) ADT + ARPI | TITAN (n=1,052)8 | ADT + apalutamide | 3.7 y | NR (4-y OS 65%) | 0.65 (0.53–0.79) ADT + ARPI | ARANOTE (n=669)9–11 | ADT + darolutamide | 2.6 y | NR | 0.78 (0.58–1.05) ADT + docetaxel + ARPI | ENZAMET, de novo M1 (n=362)7 | ADT + 6x docetaxel + enzalutamide | 8.1 y | 8-y OS 40% | 0.81 (0.62–1.05) ADT + docetaxel + ARPI | PEACE-1, docetaxel population (n=710)12 | ADT + 6x docetaxel + abiraterone ± RXT | 4.4 y | NR | 0.75 (0.59–0.95) ADT + docetaxel + ARPI | ARASENS (n=1,306)13 | ADT + 6x docetaxel + darolutamide | 3.6 y | NR (4-y OS 63%) | 0.68 (0.57–0.80) NR= not reached; ARPI= androgen receptor pathway inhibitor; RXT= radiotherapy; mCSPC=metastatic castration-sensitive prostate cancer Pre-planned long-term analysis Median follow-up= 7.7 years [7.5 ; 7.9] Ursula Vogl [Slide 2] PEACE-1: ADT + Doce +/- Abi Overall survival (%) SOC+Abi | SOC 7-year OS, % (95.1% CI) 45.7 (40.3-51.2) | 30.7 (25.8-36.1) 8-year OS, % (95.1% CI) 40.3 (34.3-46.3) | 27.1 (22.0-32.9) Adjusted HR* (95.1% CI) 0.68 (0.56-0.82) Time from randomization (in years) Number at risk (censored) No abiraterone Abiraterone ENZAMET: ADT + Doce +/- Enza Percent Alive HR = 0.81 (95% CI: 0.62–1.05) 5-year OS: 60% vs. 52% 8-year OS: 40% vs. 34% Enzalutamide + DOC Control + DOC Years Number at risk (number of events) Survival table Endpoint: OS (PEACE-1 docetaxel population, panel 1 of 2) SOC+Abi | 355 SOC | 355 HR: 0.68 HR CI: 0.56-0.82 (95.1% CI, adjusted) 7-year OS 45.7% (40.3-51.2) vs 30.7% (25.8-36.1) 8-year OS 40.3% (34.3-46.3) vs 27.1% (22.0-32.9) [Slide 3] ARPI resistance in advanced prostate cancer: mechanisms and key alterations AR-driven resistance 1 AR reactivation • AR amplification / overexpression • AR-LBD mutations (L702H, T878A) • AR-V7 splice variant 2 Androgen bypass • Glucocorticoid receptor takeover 3 Alternative survival signaling • PTEN loss / PI3K–AKT activation 4 Epigenetic escape • EZH2 upregulation • FOXA1/GATA2 ARPI resistance in mCRPC Resistance beyond AR 5 DNA repair deficiency • BRCA2, BRCA1, ATM, other HRR genes • MSI-high / dMMR 6 Lineage plasticity (AR-indifferent) • TP53 / RB1 loss • Neuroendocrine transdifferentiation 7 Targets downstream of AR • PSMA, STEAP1, hK2, B7-H3 (surface)

[Slide 1] Safety and Efficacy of JNJ-6143, a First-in-Class Oral AR-Targeted Regulated Induced Proximity Targeting Chimera Molecule in Patients With mCRPC/mAPMR, Including High-Risk Molecular Features Benjamin Garmezy,1 Sheena Sahota,2 Judy S. Wang,3 Sreenivasa R. Chandana,4 Gerald S. Falchook,5 Neal D. Shore,6 Xin Gao,7 Matthew R. Smith,7 Joseph W. Kim,8 Dillon Cockrell,9 Charles J. Ryan,10 Daniel P. Petrylak,8 Michael J. Morris,10 Andrew J. Armstrong,9 Johann S. De Bono,11 Alex Yu,12 Lilla Di Scala,13 Malhar Khatu,14 Lena N. Lupey-Green,15 Sebastiano Battaglia,12 Beerinder Pal Karir,12 Katherine J. Kayser-Bricker,15 Eyal Attar,16 Lindsey Hoffman,17 Andrew W. Hahn18 1Sarah Cannon Research Institute, Nashville, TN, USA; 2NEXT Oncology, Austin, TX, USA; 3Florida Cancer Specialists/Sarah Cannon Research Institute, Sarasota, FL, USA; 4START Center for Cancer Research - Midwest, Grand Rapids, MI, USA; 5Sarah Cannon Research Institute at HealthONE, Denver, CO, USA; 6START Carolinas/Carolina Urologic Research Center, Myrtle Beach, SC, USA; 7Massachusetts General Hospital and Harvard Medical School, Boston, MA, USA; 8Yale School of Medicine, New Haven, CT, USA; 9Duke University, Durham, NC, USA; 10Memorial Sloan Kettering Cancer Center, New York, NY, USA; 11The Institute of Cancer Research and the Royal Marsden NHS Foundation Trust, London, UK; 12Johnson & Johnson, Spring House, PA, USA; 13Johnson & Johnson, Basel, Switzerland; 14Johnson & Johnson, Mumbai, India; 15Johnson & Johnson, New Haven, CT, USA; 16formerly of Halda Therapeutics, Inc, New Haven, CT, USA; 17Johnson & Johnson, Los Angeles, CA, USA; 18The University of Texas MD Anderson Cancer Center, Houston, TX, USA AR, androgen receptor; mAPMR, metastatic androgen pathway modulation-resistant prostate cancer; mCRPC, metastatic castration-resistant prostate cancer. Presented by B Garmezy at the European Society for Medical Oncology (ESMO) Annual Congress; October 23–27, 2026; Madrid, Spain https://www.congresshub.com/Oncology/ESMO2026/ProstateEarlyAssets/Garmezy The QR code is intended to provide scientific information for individual reference, and the information should not be altered or reproduced in any way. [Slide 2] JNJ-6143,^a First-in-Class RIPTAC™ Therapeutic: Targeted Dual-Protein Lock in an Oral Small Molecule JNJ-6143: A Regulated Induced Proximity Targeting Chimera (RIPTAC™) molecule [Diagram: Prostate cancer cell; Nucleus; AR; BRD4] Ternary Complex Formation: AR-JNJ-6143-BRD4 JNJ-6143 selectively binds AR and BRD4 Ternary complex sequesters and functionally inactivates BRD4 BRD4 cannot engage chromatin to drive transcription → Apoptosis • Selectively binds AR and BRD4, resulting in BRD4 loss of function and death in PC cells1-3 • Induces tumor regression across preclinical in vivo CRPC models with AR and non-AR alterations2,3 • Preliminary phase 1 data suggest a favorable safety profile with evidence of tumor shrinkage and PSA decline in patients with mCRPC3 We report updated phase 1 data evaluating JNJ-6143 in mCRPC ^aFormerly HLD-0915. BRD4, bromodomain-containing protein 4; CRPC, castration-resistant prostate cancer; PC, prostate cancer; PSA, prostate-specific antigen; RIPTAC™, Regulated Induced Proximity Targeting Chimera. 1. Raina K, et al. Cell Chem Biol. 2024;31(8):1490-1502. 2. Raina K, et al. J Clin Oncol. 2023;41(suppl 6):184. 3. Hahn A, et al. Presented at: AACR-NCI-EORTC Molecular Targets and Cancer Therapeutics Symposium; October 22-26, 2025; Boston, MA, USA. Presented by B Garmezy at the European Society for Medical Oncology (ESMO) Annual Congress; October 23–27, 2026; Madrid, Spain [Slide 3] JNJ-6143: High PSA Response Rate Irrespective of High-Risk Molecular Features [Waterfall: Best PSA change,^a % by dose cohort 12.5 mg, 25 mg, 50 mg, 100 mg, 150 mg; PSA50 and PSA90 reference lines] Alteration: SNV; Loss; Indel; Amplification [Oncoprint rows: ARamp, AR LBD mut, TP53, PTEN, RB1, BRCA1/2] PSA50 rate by high-risk genomic alteration status in the 50 mg cohort (n=27) Gene | Altered | Unaltered ARamp | 13/15 | 9/12 AR LBD mut | 2/3 | 20/24 TP53 | 17/21 | 5/6 PTEN | 8/10 | 14/17 RB1 | 9/12 | 13/15 BRCA1/2 | 9/11 | 13/16 X-axis: PSA50 rate,^b % (0–100) PSA50 was achieved in 5/6 (83%) ARv7+ patients with evaluable CTCs (n=7) Data cutoff date: September 21, 2026. ^aPatients with both baseline and ≥1 postbaseline PSA assessment and available Cycle 1 Day 1 ctDNA data from Guardant Reveal (Guardant Health, Palo Alto, CA). ^bError bars represent Wilson 95% CI. ARamp, androgen receptor amplification; AR LBD mut, androgen receptor ligand binding domain mutation; ARv7, androgen receptor splice variant 7; BRCA, breast cancer gene; CTC, circulating tumor cell; Indel, insertion or deletion; PSA50, 50% reduction in PSA from baseline; PSA90, 90% reduction in PSA from baseline; PTEN, phosphatase and tensin homolog; RB1, RB transcriptional corepressor 1; SNV, single nucleotide variant; TP53, tumor protein p53. Presented by B Garmezy at the European Society for Medical Oncology (ESMO) Annual Congress; October 23–27, 2026; Madrid, Spain [Slide 4] JNJ-6143: High Rate of Radiographic Responses at the RP2D in Advanced mCRPC ORR-evaluable patients in the 50 mg cohort 69% Confirmed ORR (11/16) 13% CR 56% PR^c 19% SD* 88% DCR (14/16) Y-axis: Response rate,^b % | 50 mg (n=16)^a DOR, median (95% CI), months | 11.0 (5.9, NE) *After the data cutoff, 1 patient with SD converted to confirmed PR. Data cutoff date: September 21, 2026. ^aPatients with measurable disease at baseline and either off-study or with ≥1 postbaseline assessment. ^bResponses are based on best overall response per RECIST version 1.1. CR and PR are confirmed and without bone progression by PCWG3. ^cThree patients with PR had visceral disease. CR, complete response; CI, confidence interval; DCR, disease control rate; DOR, duration of response; NE, not evaluable; ORR, objective response rate; PCWG3, Prostate Cancer Working Group 3; PR, partial response; RECIST, Response Evaluation Criteria in Solid Tumors; SD, stable disease. Presented by B Garmezy at the European Society for Medical Oncology (ESMO) Annual Congress; October 23–27, 2026; Madrid, Spain

[Slide 2] JNJ-6143: At RP2D, TRAEs Were Mostly Low Grade With Low Rates of Treatment Discontinuations Due to TRAEs (3%) n (%) | 50 mg (n=38) Any grade | 50 mg Grade ≥3 | All doses(a) (N=105) Any grade | All doses Grade ≥3 Any TRAE | 22 (58) | 2 (5)* | 71 (68) | 18 (17) Most common TRAE(b) | | | | Hypertriglyceridemia | 6 (16) | 1 (3)* | 29 (28) | 5 (5) Nausea | 6 (16) | 0 | 20 (19) | 0 Decreased appetite | 4 (11) | 0 | 19 (18) | 1 (1) Fatigue | 5 (13) | 0 | 19 (18) | 0 Hypercholesterolemia | 7 (18) | 0 | 18 (17) | 3 (3) Anemia | 5 (13) | 1 (3)* | 17 (16) | 7 (7) Increased AST | 3 (8) | 0 | 10 (10) | 1 (1) Acute kidney injury | 0 | 0 | 2 (2) | 2 (2) Syncope | 0 | 0 | 2 (2) | 2 (2) TRAEs leading to dose interruption | 4 (11) | | 17 (16) | TRAEs leading to dose reduction | 1 (3) | | 4 (4) | TRAEs leading to discontinuation | 1 (3) | | 5 (5) | DLTs | 0 | | 1 (1)(c) | Serious TRAEs | 0 | | 3 (3) | TRAEs leading to death | 0 | | 1 (1)(d) | *Both grade ≥3 TRAEs were reclassified after data cutoff as not related to study drug. Data cutoff date: September 21, 2026. (a)Doses ranged from 12.5 to 150 mg. (b)The most common TRAEs of any grade that occurred in ≥10% of patients or the most common grade ≥3 TRAEs that occurred in ≥2% of patients in all doses. (c)Hepatotoxicity in setting of extrinsic bile duct compression. (d)Grade 5 acute kidney injury: patient died after acute hemodynamic event while undergoing hemodialysis. AST, aspartate aminotransferase; DLT, dose-limiting toxicity; TRAE, treatment-related adverse event. Presented by B Garmezy at the European Society for Medical Oncology (ESMO) Annual Congress; October 23–27, 2026; Madrid, Spain [Slide 3] JNJ-6143: Most Patients (69/105 [66%]) Remain on Treatment [Swimmer plot by dose: 150 mg | 100 mg | 50 mg | 25 mg | 12.5 mg; x-axis Months on treatment 0-16; ► Ongoing] 150 mg: 8/10 (80%) ongoing 100 mg: 15/19 (79%) ongoing 50 mg: 27/38 (71%) ongoing; median follow-up of 5.6 months (range: 0.2, 14.5) 25 mg: 13/29 (45%) ongoing 12.5 mg: 6/9 (67%) ongoing Data cutoff date: September 21, 2026. Presented by B Garmezy at the European Society for Medical Oncology (ESMO) Annual Congress; October 23–27, 2026; Madrid, Spain

[Slide 1] JNJ-6143: First-in-Human, Multicenter, Phase 1 Dose-Escalation Study Dose escalation and backfills: 150 mg (n=10) | 100 mg (n=19) | 50 mg (n=18) | 25 mg (n=15) | 12.5 mg (n=9) Dose optimization: 50 mg (n=20) | 25 mg (n=14) N=105(a) JNJ-6143 administered orally once daily in 21-day cycles Study design • Phase 1 dose escalation with backfill and dose optimization in patients with mCRPC Key eligibility criteria • Progressive mCRPC with rising PSA (≥2 ng/mL for dose escalation; ≥20 ng/mL for dose optimization) • Prior treatment with ≥1 ARPI and may have had ≤2 prior taxanes and ≤1 RLT • ECOG PS score of 0 or 1 Objectives • Safety and tolerability • PK/PD and efficacy (ie, PSA, ORR, ctDNA) • Determine recommended doses for expansion 50 mg once daily was selected as the RP2D based on safety, efficacy, and PK ClinicalTrials.gov identifier: NCT06800313. (a)Patients dosed as of September 21, 2026. ARPI, androgen receptor pathway inhibitor; ctDNA, circulating tumor DNA; ECOG PS, Eastern Cooperative Oncology Group performance status; ORR, objective response rate; PD, pharmacodynamics; PK, pharmacokinetics; RLT, radioligand therapy. Presented by B Garmezy at the European Society for Medical Oncology (ESMO) Annual Congress; October 23–27, 2026; Madrid, Spain [Slide 2] JNJ-6143: Phase 1 Study Enrolled a Heavily Pretreated Population with Advanced mCRPC Characteristic | 50 mg (n=38) | All doses(a) (N=105) Age, years, median (range) | 73 (56, 90) | 72 (51, 90) PSA, µg/L, median (range) | 122.2 (17.3, 2235) | 90.0 (3.6, 5000) Lines of prior systemic therapy,(b) median (range) | 4 (1, 9) | 4 (1, 10) Prior systemic therapy,(c) n (%) | | RLT | 17 (46) | 45 (44) Taxane | 20 (54) | 68 (66) Taxane + RLT | 12 (32) | 36 (35) PARP inhibitor | 2 (5) | 10 (10) Visceral disease,(d) n (%) | 11 (29) | 26 (27) Liver | 4 (11) | 5 (5) Lung | 3 (8) | 12 (13) Other | 5 (16) | 10 (10) Data cutoff date: September 21, 2026. (a)Doses ranged from 12.5 to 150 mg. (b)Lines of prior systemic therapy counting only those used in the metastatic setting. (c)n=97 in all doses. (d)n=96 in all doses. PARP, poly (ADP-ribose) polymerase. Presented by B Garmezy at the European Society for Medical Oncology (ESMO) Annual Congress; October 23–27, 2026; Madrid, Spain

[Slide 1] n (%) | 50 mg (n=38) Any grade | 50 mg Grade ≥3 | All doses^a (N=105) Any grade | All doses Grade ≥3 Any TRAE | 22 (58) | 2 (5)* | 71 (68) | 18 (17) Most common TRAE^b Hypertriglyceridemia | 6 (16) | 1 (3)* | 29 (28) | 5 (5) Nausea | 6 (16) | 0 | 20 (19) | 0 Decreased appetite | 4 (11) | 0 | 19 (18) | 1 (1) Fatigue | 5 (13) | 0 | 19 (18) | 0 Hypercholesterolemia | 7 (18) | 0 | 18 (17) | 3 (3) Anemia | 5 (13) | 1 (3)* | 17 (16) | 7 (7) Increased AST | 3 (8) | 0 | 10 (10) | 1 (1) Acute kidney injury | 0 | 0 | 2 (2) | 2 (2) Syncope | 0 | 0 | 2 (2) | 2 (2) TRAEs leading to dose interruption | 4 (11) | 17 (16) TRAEs leading to dose reduction | 1 (3) | 4 (4) TRAEs leading to discontinuation | 1 (3) | 5 (5) DLTs | 0 | 1 (1)^c Serious TRAEs | 0 | 3 (3) TRAEs leading to death | 0 | 1 (1)^d *Both Grade ≥3 TRAEs were reclassified after data cutoff as not related to study drug [Slide 2] Best PSA change,^a % Alteration: SNV | Loss | Indel | Amplification PSA50 | PSA90 Dose cohorts: 12.5 mg | 25 mg | 50 mg | 100 mg | 150 mg Oncoprint rows: ARamp | AR LBD mut | TP53 | PTEN | RB1 | BRCA1/2 PSA50 rate by high-risk genomic alteration status in 50 mg cohort (n=27) Altered | Unaltered ARamp: 13/15 | 9/12 AR LBD mut: 2/3 | 20/24 TP53: 17/21 | 5/6 PTEN: 8/10 | 14/17 RB1: 9/12 | 13/15 BRCA1/2: 9/11 | 13/16 PSA50 rate,^b % PSA50 was achieved in 5/6 (83%) ARv7+ patients with evaluable CTCs (n=7) [Slide 3] PSA-evaluable patients in the 50 mg cohort n (%) | 50 mg (n=31)^a Confirmed PSA50^b | 24 (77) Confirmed PSA90^b | 15 (48) Change in PSA from baseline, % PSA50 | PSA90 Months on treatment [Slide 4] ORR-evaluable patients in the 50 mg cohort Response rate, % 13% CR 56% PR^b 19% SD* 69% Confirmed ORR^b (11/16) 88% DCR (14/16) | 50 mg (n=16)^a DOR, median (95% CI), months | 11.0 (5.9, NE) *After the data cutoff, 1 patient with SD converted to confirmed PR.

JNJ-6143: Conclusions • JNJ-6143 is a first-in-class RIPTAC™ molecule that forms a targeted dual-protein lock between AR (PC-selective tumor antigen) and BRD4 (essential cell survival protein) to trigger cell death • JNJ-6143 was highly active with an encouraging safety profile in this phase 1 study of patients with heavily pretreated mCRPC • At the RP2D of 50 mg once daily (oral): – <60% any grade TRAEs – 77% confirmed PSA50 rate and 48% confirmed PSA90 rate – High PSA response rate irrespective of high-risk molecular features – 69% confirmed ORR • A phase 2 study (NCT07832292) has been initiated with 50 mg once daily These early data suggest that JNJ-6143 has the potential to combine high tumor selectivity and robust efficacy with a favorable safety profile and convenient oral administration Presented by B Garmezy at the European Society for Medical Oncology (ESMO) Annual Congress; October 23–27, 2026; Madrid, Spain

[Slide 1] The AR Escape Map AR-dependent: amplification, rearrangements, mutations, splice variants AR-bypass: PI3K/AKT, glucocorticoid receptor, other alternative pathways AR-independent: lineage plasticity [Metro-map style diagram; station labels include: AR amplification; LBD mutations; Rearrangements; Feria de Madrid (AR-dependent); EZH2 / NSD2 axis; RB1 / TP53 loss; Nuevos Ministerios; PI3K/AKT feedback - PTEN loss; Glucocorticoid receptor (GR); Wnt/beta-catenin pathway; AR-bypass; ASCL1-driven; AR-independent; small sub-labels illegible] Legend: AR-dependent / AR-bypass / AR-independent Alice BERNARD-TESSIER, MD Conception with Claude Alice Bernard-Tessier - Novel AR targeted therapies [Slide 2] AR alterations are enriched and frequent in mCRPC AR alteration in 80% of mCRPC patients [Horizontal bar chart of gene alteration frequency, AR at top highlighted by arrow; legend: mutation, deep deletion, amplification, fusion; x-axis Alteration frequency; gene labels illegible] AR alteration frequency in ARPI-naive vs -exposed tumours [Bar chart: ARPI-naive (lower) vs ARPI-exposed (higher); axis values illegible] Different genomic AR alterations [Schematic of AR gene/protein: AR enhancer, NTD, DBD, LBD; Copy number gain; Androgen receptor wild type; Androgen receptor splicing variants; Androgen receptor mutations - small labels partly illegible] Alice BERNARD-TESSIER, MD Abida et al., PNAS 2019 Abida et al., JCO Precision Oncol 2017 [Slide 3] AR alterations are associated with worse prognosis AR amplification ~60% [KM curve: Overall survival (%), Months 0-42; labels include AR gain (>= copies) and AR WT - small text illegible] AR mutations ~20% [KM curve, Survival Probability vs days; small text illegible] AR splicing variants ~10% [KM curve: AR-V7 negative vs AR-V7 positive; HR and median survival text too small to read reliably] Poor prognosis is well established; but their role as predictive biomarkers of ARPI resistance remains incompletely established. Alice BERNARD-TESSIER, MD Annala et al., Clin Cancer Res 2021 Annala et al., JCO Precis Oncol 2023 [partly illegible] Antonarakis et al., JAMA 2017 [partly illegible] [Slide 4] Lessons from next generation AR inhibitors Cumelutamide (TAS3681) [Waterfall plot of PSA change; inset footnotes (low resolution): aConfirmed PSA 50 response: 3 in 600 mg QD, 1 in 300 mg BID, 1 in 400 mg BID; bConfirmed PSA 30 response: 1 in 600 mg QD, 1 in 300 mg BID, 2 in 400 mg BID - [low resolution, verify]] [AR schematic: N-terminal domain, DNA-binding domain, Ligand-binding domain (highlighted)] Masofaniten (EPI-7386) [Waterfall plot; AR schematic with N-terminal domain highlighted] Alice BERNARD-TESSIER, MD De Bono et al., ASCO 2021 Pachynski et al., ASCO GU 2023

JNJ-6143,a First-in-Class RIPTAC™ Therape[utic; cropped]: Targeted Dual-Protein Lock in an Oral Smal[l; cropped] JNJ-6143: A Regulated Induced Proximity Targeting Chimera (RIPTAC™) molecule Prostate cancer cell | Nucleus | AR | BRD4 Ternary Complex Formation: AR-JNJ-6143-BRD4 JNJ-6143 selectively binds AR and BRD4 Ternary complex sequesters and functionally inactivates BRD4 BRD4 cannot engage chromatin to drive transcription → Apoptosis [Right column bullets cropped: 'Sel... res... dea...', 'Ind... pre... AR...', 'Pre... fav... of... in...'] We report updated phase 1 data evaluating JNJ-6143 in[cropped] aFormerly HLD-0915. BRD4, bromodomain-containing protein 4; CRPC, castration-resistant prostate cancer; PC, prostate cancer; PSA, prostate-specific antigen; RIPTAC™, Regulated[cropped] 1. Raina K, et al. Cell Chem Biol. 2024;31(8):1490-1502. 2. Raina K, et al. J Clin Oncol. 2023;41(suppl 6):184. 3. Hahn A, et al. Presented at: AACR-NCI-EORTC [cropped] Therapeutics Symposium; October 22-26, 2025; Boston, MA, USA. Presented by B Garmezy at the European Society for Medical Oncology (ESMO) Annual Congress; October 23–27, 2026; [cropped]

[Slide 1] JNJ-6143: A Regulated Induced Proximity Targeting Chimera (RIPTAC™) molecule Prostate cancer cell Nucleus AR BRD4 Ternary Complex Formation: AR-JNJ-6143-BRD4 JNJ-6143 JNJ-6143 selectively binds AR and BRD4 Ternary complex sequesters and functionally inactivates BRD4 BRD4 cannot engage chromatin to drive transcription Apoptosis [Slide 2] Dose escalation and backfills | Dose optimization 150 mg (n=10) 100 mg (n=19) 50 mg (n=18) | 50 mg (n=20) 25 mg (n=15) | 25 mg (n=14) 12.5 mg (n=9) N=105^a JNJ-6143 administered orally once daily in 21-day cycles [Slide 3] [Swimmer plot by dose: 150 mg, 100 mg, 50 mg, 25 mg, 12.5 mg] Months on treatment (0-16) ► Ongoing

[Slide 1] N=408 Patients with HR LPC/LAPC -> RANDOMIZED 1:3 Perioperative open-label APA + ADT Neoadjuvant (C1 through C6): APA (240 mg QD) + ADT -> 2 wks -> RP + PLND(a) -> 4 wks -> Adjuvant (C7 through C12): APA (240 mg QD) + ADT RP + PLND -> Local standard institutional practice/observation Primary end point - EFS: Time to first of BCF (PSA progression) or LRR or DM, or death(c) Secondary and exploratory end points include: - MFS - Time to first subsequent therapy Randomization: 27 July 2022 to 16 October 2023 Stratification: - cN0 vs cN1(b) - Gleason score (7 vs 8-10) - Region (North America, Europe, ROW) Local or systemic adjuvant treatment based on physician's discretion and local standard institutional practice [Slide 2] Event-Free Survival (%) T recovery to >=200 ng/dL: Median 20.7 months (~9 months from EOT) APA + ADT Median not reached | 63.0% (95% CI, 52.2-71.9) RP + PLND alone Median 19.1 months | 41.4% (95% CI, 35.6-47.1) HR, 0.38(a) 95% CI, 0.26-0.55 p<0.0001(b) Months from Randomization: 0 6 12 18 24 30 36 42 No. at risk APA + ADT | 102 | 99 | 96 | 91 | 82 | 71 | 38 | 10 RP + PLND alone | 306 | 195 | 169 | 146 | 131 | 121 | 74 | 15 Survival table Endpoint: Event-free survival APA + ADT | 102 | Not reached RP + PLND alone | 306 | 19.1 months HR: 0.38 HR CI: 0.26-0.55 P: <0.0001 Landmark rates: 36 months: APA + ADT 63.0% (CI 52.2-71.9) | RP + PLND alone 41.4% (CI 35.6-47.1) [Slide 3] Metastasis-Free Survival (%) APA + ADT RP + PLND alone HR, 0.49(a) 95% CI, 0.26-0.90 nominal p=0.0193(b) Metastasis detection method, n (%) | APA + ADT | RP + PLND PSMA PET | 9 (8.8) | 50 (16.3) Conventional imaging | 1 (1.0) | 6 (2.0) Histopathology | 1 (1.0) | 0 Months from Randomization: 0 6 12 18 24 30 36 42 No. at risk APA + ADT | 102 | 99 | 97 | 94 | 90 | 85 | 52 | 13 RP + PLND alone | 306 | 262 | 255 | 241 | 231 | 220 | 130 | 25 Survival table Endpoint: Metastasis-free survival APA + ADT | 102 RP + PLND alone | 306 HR: 0.49 HR CI: 0.26-0.90 P: nominal p=0.0193 [Slide 4] Time to First Subsequent Therapy (%) APA + ADT Median 46.0 months 64.6% (95% CI, 54.1-73.3 ) RP + PLND alone Median 12.3 months 32.0% (95% CI, 26.7-37.5) HR, 0.29a 95% CI, 0.20-0.43 nominal p<0.0001b Months from Randomization 0 | 6 | 12 | 18 | 24 | 30 | 36 | 42 No. at risk APA + ADT | 102 | 97 | 91 | 87 | 81 | 73 | 42 | 12 RP + PLND alone | 306 | 178 | 145 | 126 | 105 | 98 | 59 | 14 Survival table Endpoint: Time to first subsequent therapy APA + ADT | 102 | 46.0 months RP + PLND alone | 306 | 12.3 months HR: 0.29 HR CI: 0.20-0.43 P: nominal p<0.0001 Landmark rates: 36 months: APA + ADT 64.6% (CI 54.1-73.3) | RP + PLND alone 32.0% (CI 26.7-37.5)

[Slide 1] PROTEUS: A Significant Advance in Surgery Outcomes in HR LPC/LAPC but No Surgery-Alone Comparator • In HR LPC/LAPC, despite refinement of prostatectomy as an SOC,1 the frequency of disease recurrence is high and patients may be subject to additional morbidity from subsequent therapy2,3 • The PROTEUS main study, with a blinded design, demonstrated superior efficacy of 1-year perioperative apalutamide plus ADT compared with an active comparator4: Unprecedented Study Design • Largest size to date • Novel end points • PSMA PET inclusion Enhanced Disease Control With Perioperative Apalutamide + ADT Significant benefit: • 9x greater pCR/MRD • 20% lower hazard for metastasis or death • 29% lower risk of recurrence or death • ~3 years longer time to next treatment Limitation Control group did not receive SOC (surgery alone) The PROTEUS Substudy was designed to compare perioperative apalutamide + ADT with RP + PLND versus SOC (surgery alone) ADT, androgen deprivation therapy; HR LPC/LAPC, high-risk localized/locally advanced prostate cancer; pCR/MRD, pathologic complete response/minimal residual disease; PLND, pelvic lymph node dissection; PSMA PET, prostate-specific membrane antigen positron emission tomography; RP, radical prostatectomy; SOC, standard of care. 1. Cooley LF, Shore ND. Transl Androl Urol. 2025;14:493-495. 2. Eiber M, et al. J Clin Oncol. 2024;42:5027. 3. Stattin P, et al. BJU Int. 2024;134:96-102. 4. Taplin ME, et al. N Engl J Med. 2026;395:546-560. Presented by A Briganti at ESMO; October 23-27, 2026; Madrid, Spain [Slide 2] Perioperative APA + ADT Improved MFS by Conventional or PSMA PET Imaging (BICR) SECONDARY END POINT [KM curve: Metastasis-Free Survival (%) vs Months from Randomization; APA + ADT vs RP + PLND alone] HR, 0.49a 95% CI, 0.26-0.90 nominal p=0.0193b Metastasis detection method, n (%) | APA + ADT | RP + PLND PSMA PET | 9 (8.8) | 50 (16.3) Conventional Imaging | 1 (1.0) | 6 (2.0) Histopathology | 1 (1.0) | 0 No. at risk (months 0, 6, 12, 18, 24, 30, 36, 42) APA + ADT | 102 99 97 94 90 85 52 13 RP + PLND alone | 306 262 255 241 231 220 130 25 • 51% reduction in risk of metastasis or death • MFS by local investigator assessment: • HR, 0.38a 95% CI, 0.22-0.66 nominal p=0.0004b MFS: Time to first occurrence of distant metastasis on PSMA PET or conventional imaging by BICR, pathologic finding of distant metastasis, or death from any cause. aHR is from stratified proportional hazards model; HR <1 favors APA + ADT. bNominal p value is from a log-rank test stratified by Gleason score (7, ≥8), nodal status (N0, N1), and geographic region (North America, Europe, and ROW). Presented by A Briganti at ESMO; October 23-27, 2026; Madrid, Spain Survival table Endpoint: Metastasis-free survival by BICR (secondary endpoint), PROTEUS substudy APA + ADT | 102 RP + PLND alone | 306 HR: 0.49 HR CI: 0.26-0.90 P: nominal p=0.0193

[Slide 1] The PROTEUS Substudy Compared APA + ADT With RP + PLND vs RP + PLND Alone • Immediate RP + PLND without neoadjuvant treatment, followed by subsequent adjuvant treatment based on physician's discretion, is used to reflect SOC based on current guideline recommendations(1) N=408 | Patients with HR LPC/LAPC → RANDOMIZED 1:3 Perioperative open-label APA + ADT: Neoadjuvant (C1 through C6) APA (240 mg QD) + ADT → 2 wks → RP + PLND(a) → 4 wks → Adjuvant (C7 through C12) APA (240 mg QD) + ADT RP + PLND → Local standard institutional practice/observation Primary end point • EFS: Time to first of BCF (PSA progression) or LRR or DM, or death(c) Secondary and exploratory end points include: • MFS • Time to first subsequent therapy Randomization: 27 July 2022 to 16 October 2023 Stratification: • cN0 vs cN1(b) • Gleason score (7 vs 8-10) • Region (North America, Europe, ROW) Local or systemic adjuvant treatment based on physician's discretion and local standard institutional practice APA, apalutamide; BCF, biochemical failure; BICR, blinded independent central review; C, cycle; DM, distant metastases; EFS, event-free survival; LRR, local or regional recurrence; PSA, prostate-specific antigen; QD, daily; ROW, rest of the world. a APA + ADT is stopped 2 weeks prior to planned RP and then resumed 4 weeks after RP. Cardiovascular risk prophylaxis and venous thromboembolism prophylaxis given based on risk. b Based on conventional imaging. c BCF defined as 2 consecutive PSA rises ≥1 week apart, with the second test resulting in an absolute PSA ≥0.2 ng/mL following RP; LRR and DM based on PSMA PET or conventional imaging, both by BICR, or histopathology. 1. Cornford P, et al. Eur Urol. 2026;S0302-2838(26)02114. Presented by A Briganti at ESMO; October 23-27, 2026; Madrid, Spain [Slide 2] Baseline Characteristics Were Well Balanced ITT population | Perioperative APA + ADT (n=102) | RP + PLND alone (n=306) Median age (range), years | 66.0 (51-82) | 66.0 (45-84) ≥75, n (%) | 9 (8.8) | 31 (10.1) Gleason score at initial diagnosis, n (%) 7 | 4 (3.9) | 8 (2.6) Gleason ≥8 | 98 (96.1) | 298 (97.4) Baseline PSA(a) Median (range), ng/mL | 14.1 (1.4-422.6)(b) | 14.2 (1.5-384.3)(c) Baseline PSA <20, n (%) | 66 (65.3) | 183 (64.0) Baseline PSA ≥20, n (%) | 35 (34.7) | 103 (36.0) Tumor stage at diagnosis, n (%)(d) ≤cT2 | 69 (67.6) | 217 (70.9) ≥cT3 | 31 (30.4) | 86 (28.1) Regional lymph node stage at diagnosis, n (%) cN0 | 94 (92.2) | 285 (93.1) cN1 | 8 (7.8) | 21 (6.9) • At clinical cutoff on 13 July 2026: – Median follow-up - 37.8 months – Patients still on the study - 94% (perioperative APA + ADT, n=96) and 89% (RP + PLND alone, n=306) ITT, intent to treat. a PSA before first dose was the closest value assessment on or prior to the start date for ADT or APA and the closest value assessment on or prior to the date of randomization for the RP + PLND alone group. b n101. c n=286. d TX n=2 (APA + ADT); n=3 (RP + PLND alone). Presented by A Briganti at ESMO; October 23-27, 2026; Madrid, Spain [Slide 3] PROTEUS Substudy Met Primary End Point: EFS Significantly Improved T recovery to ≥200 ng/dL: Median 20.7 months (~9 months from EOT) APA + ADT Median not reached | 63.0% (95% CI, 52.2-71.9) RP + PLND alone Median 19.1 months | 41.4% (95% CI, 35.6-47.1) HR, 0.38(a) 95% CI, 0.26-0.55 p<0.0001(b) Y-axis: Event-Free Survival (%); X-axis: Months from Randomization (0-42); dashed landmark line at 36 months No. at risk | 0 | 6 | 12 | 18 | 24 | 30 | 36 | 42 APA + ADT | 102 | 99 | 96 | 91 | 82 | 71 | 38 | 10 RP + PLND alone | 306 | 195 | 169 | 146 | 131 | 121 | 74 | 15 • 62% reduction in risk of oncologic event or death • Median 19.1 months EFS with SOC highlights the unmet need EFS: Time to first of BCF (PSA progression) or LRR or DM, or death by any cause. LRR and DM based on PSMA PET or conventional imaging, both by BICR, or histopathology. CI, confidence interval. a HR is from stratified proportional hazards model; HR <1 favors APA + ADT. b p Value is from a log-rank test stratified by Gleason score (7, ≥8), nodal status (N0, N1), and geographic region (North America, Europe, and ROW). Presented by A Briganti at ESMO; October 23-27, 2026; Madrid, Spain Survival table Endpoint: Event-free survival (EFS) APA + ADT (perioperative) | 102 | Not reached RP + PLND alone | 306 | 19.1 months HR: 0.38 HR CI: 0.26-0.55 P: <0.0001 Landmark rates: 36 months: APA + ADT 63.0% (CI 52.2-71.9) | RP + PLND alone 41.4% (CI 35.6-47.1) [Slide 4] Treatment Benefit for EFS Was Consistent Over Time With Perioperative APA + ADT, Even After T Recovery • T recovery to ≥200 ng/dL with perioperative APA + ADT: Median 20.7 months (~9 months from EOT) Time after randomization | EFS rate, % (95% CI) Perioperative APA + ADT (n=102) | EFS rate, % (95% CI) RP + PLND alone (n=306) | EFS rate difference, % (95% CI) | Nominal p value 20 months | 91 (85-97) | 49 (43-55) | 42 (33-51) | <0.001 24 months | 84 (75-90) | 46 (40-52) | 37 (27-47) | <0.001 36 months | 63 (52-72) | 41 (36-47) | 22 (10-33) | <0.001 EFS: Time to first of BCF (PSA progression) or LRR or DM, or death by any cause. LRR and DM based on PSMA PET or conventional imaging, both by BICR, or histopathology. EOT, end of treatment; T, testosterone. Presented by A Briganti at ESMO; October 23-27, 2026; Madrid, Spain

[Slide 1] Fewer Local/Regional Recurrence or Distant Metastasis Events With Perioperative APA + ADT | Perioperative APA + ADT (n=102) | RP + PLND alone (n=306) First event (denominator for events below), n (%) | 35/102 (34.3) | 172/306 (56.2) Biochemical failure (BCF) | 27 (77.1) | 92 (53.5) Local or regional recurrence by PSMA PET/conventional imaging | 4 (11.4) | 51 (29.7) LRR by PSMA PET | 4 (11.4) | 50 (29.1) LRR by PSMA PET without BCF | 4 (11.4) | 48 (27.9) LRR by conventional imaging | 0 | 1 (0.6) Distant metastasis by PSMA PET/conventional imaging/histopathology | 4 (11.4) | 28 (16.3) DM by PSMA PET | 2 (5.7) | 24 (14.0) DM by PSMA PET without BCF | 2 (5.7) | 23 (13.4) DM by conventional imaging/histopathology | 2 (5.7) | 4 (2.3) Death | 0 | 1 (0.6) Censored, n (%) | 67/102 (65.7) | 134/306 (43.8) • ~70% of PSMA PET–detected LRR/DM without BCF in the RP group were associated with PSA persistence (PSA ≥0.1 ng/mL) 3 months after RP • Excluding patients (both arms) with PSMA PET–detected events (LRR or DM) without prior BCF: HR, 0.50; 95% CI, 0.33-0.76; p=0.001 Presented by A Briganti at ESMO; October 23-27, 2026; Madrid, Spain [Slide 3] EXPLORATORY END POINT Time to First Subsequent Therapy Was Prolonged by ~3 Years With Perioperative APA + ADT APA + ADT Median 46.0 months 64.6% (95% CI, 54.1-73.3 ) RP + PLND alone Median 12.3 months 32.0% (95% CI, 26.7-37.5) HR, 0.29a 95% CI, 0.20-0.43 nominal p<0.0001b Y-axis: Time to First Subsequent Therapy (%) X-axis: Months from Randomization (0, 6, 12, 18, 24, 30, 36, 42) No. at risk APA + ADT | 102 | 97 | 91 | 87 | 81 | 73 | 42 | 12 RP + PLND alone | 306 | 178 | 145 | 126 | 105 | 98 | 59 | 14 • 71% reduced risk of starting subsequent therapy • ~50% fewer patients started subsequent therapy by 3 years Time to first subsequent therapy: Time from randomization to first local, regional, or systemic therapy, including reinitiation of ADT. aHR is from stratified proportional hazards model; HR <1 favors APA + ADT. bNominal p value is from a log-rank test stratified by Gleason score (7, ≥8), nodal status (N0, N1), and geographic region (North America, Europe, and ROW). Presented by A Briganti at ESMO; October 23-27, 2026; Madrid, Spain Survival table Endpoint: Time to first subsequent therapy (exploratory) APA + ADT | 46.0 months RP + PLND alone | 12.3 months HR: 0.29 HR CI: 0.20-0.43 P: nominal p<0.0001 Landmark rates: 36 months: APA + ADT 64.6% (CI 54.1-73.3) | RP + PLND alone 32.0% (CI 26.7-37.5)

PROTEUS Substudy Conclusions • Together with the PROTEUS main study, the PROTEUS Substudy supports 1-year perioperative APA + ADT as a treatment option for patients with HR LPC/LAPC fit for surgery and systemic therapy • Perioperative APA + ADT improves disease control and gives patients a longer time before starting subsequent therapy Enhanced Disease Control With Perioperative APA + ADT • 62% lower risk of recurrence or death • 51% lower risk of death or developing metastases over time • ~50% fewer patients starting subsequent therapy by ~3 years • >3× longer time to first subsequent therapy Presented by A Briganti at ESMO: October 23-27, 2026; Madrid, Spain

Perioperative Apalutamide Plus Androgen Deprivation Therapy With Radical Prostatectomy Plus Pelvic Lymph Node Dissection (RP + PLND) vs RP + PLND Alone in Patients With High-Risk Localized or Locally Advanced Prostate Cancer (HR LPC/LAPC): PROTEUS Substudy Alberto Briganti1, Thomas Steuber2, Neal D Shore3, Martin Gleave4, Alexander Kretschmer5, Lisa Wetherhold5, Eleni Efstathiou6, David J Cahn7, Ronaldo Damião8, Alexandre De La Taille9, Boris A Hadaschik10, Bernardo Herrera Imbroda11, Álvaro Juárez Soto12, Giuseppe Simone13, Zhuokai Li14, Sharon A McCarthy15, Mark Wildgust16, Suneel D Mundle16, Adam S Kibel17, Mary-Ellen Taplin18 Presented by A Briganti at ESMO; October 23-27, 2026; Madrid, Spain

The PROTEUS Substudy Compared APA + ADT With RP + PLND vs RP + PLND Alone • Immediate RP + PLND without neoadjuvant treatment, followed by subsequent adjuvant treatment based on physician's discretion, is used to reflect SOC based on current guideline recommendations N=408 Patients with HR LPC/LAPC → RANDOMIZED 1:3 Perioperative open-label APA + ADT: Neoadjuvant (C1 through C6) APA (240 mg QD) + ADT → 2 wks → RP + PLNDa → 4 wks → Adjuvant (C7 through C12) APA (240 mg QD) + ADT RP + PLND → Local standard institutional practice/observation Local or systemic adjuvant treatment based on physician's discretion and local standard institutional practice Primary end point: EFS: Time to first of BCF (PSA progression) or LRR or DM, or deathc Secondary and exploratory end points include: MFS; Time to first subsequent therapy Stratification: cN0 vs cN1b; Gleason score (7 vs 8-10); Region (North America, Europe, ROW) Randomization: 27 July 2022 to 16 October 2023 Presented by A Briganti at ESMO; October 23-27, 2026; Madrid, Spain

Treatment Effect of Perioperative (Neoadjuvant and Adjuvant) Apalutamide plus Androgen-deprivation Therapy Versus Radical Prostatectomy Alone in Patients with High-risk Localized or Locally Advanced Prostate Cancer from the PROTEUS Substudy Alberto Briganti, Thomas Steuber, Neal D. Shore, Martin Gleave, Alexander Kretschmer, Lisa Wetherhold, Eleni Efstathiou, David J. Cahn, Ronaldo Damião, Alexandre De La Taille, Seok-Soo Byun, Boris A. Hadaschik, Bernardo Herrera Imbroda, Álvaro Juárez Soto, Axel S. Merseburger, Giuseppe Simone, Brendan Rooney, Denis Mikerin, Zhuokai Li, Branko Miladinovic, Sharon A. McCarthy, Mark Wildgust, Suneel D. Mundle, Adam S. Kibel, Mary-Ellen Taplin [European Urology journal cover] Presented by A Briganti at ESMO, October 23-27, 2026, Madrid, Spain

[Slide 1] XmAb819, ENPP3 x CD3 bispecific antibody in development for ccRCC [Diagram labels: T cell releases cytotoxic granzymes & perforins; Apoptosis of tumor cell; High-expressing ENPP3+ tumor cell; ENPP3; CD3; Activated CD3+ T-cell; XmAb819 ENPP3 x CD3 Bispecific Antibody] • ENPP3 is an extracellular nucleotide metabolizing enzyme with high expression in RCC and other solid tumors, and low expression in normal tissues1 • XmAb819 targets T cells for potent lysis of ENPP3+ tumor cells • XmAb 2+1 format selectively binds and kills high density ENPP3 tumor cells, but spares low density normal tissue [Box plot: ENPP3 gene expression (TPM) (log2); Primary Tumor vs Normal Tissue; Bladder, Breast, Colon, Kidney, Liver, Lung, Ovary, Prostate, Skin, Soft tissue,Bone] ENPP3 ectonucleotide pyrophosphatase / phosphodiesterase family member 3 1Doñate, Fernando, et al. Clinical Cancer Research 22.8 (2016) Pal SK, et al. [Slide 2] Patient characteristics Baseline Characteristic | All IV Doses (n=119) | Expansion Doses (n=68) Median Age, Years (range) | 61 (34, 78) | 62 (34, 78) Male, n (%) | 97 (81.5) | 58 (85.3) Race – White, n (%) | 101 (84.9) | 56 (82.4) Time Since Disease Diagnosis, mo. (range) | 52.5 (7, 287) | 49.5 (7, 287) IMDC Risk Score at Baseline1, n (%) Favorable | 8 (6.7) | 4 (5.9) Intermediate | 81 (68.1) | 46 (67.6) Poor | 30 (25.2) | 18 (26.5) Sarcomatoid Features, n (%) | 16 (13.4) | 13 (19.1) ≥3 Metastatic Sites, n (%) | 84 (70.6) | 46 (67.6) Bone metastasis, n (%) | 58 (48.7) | 32 (47.1) Liver metastasis, n (%) | 48 (40.3) | 27 (39.7) Target Lesions SOD (mm), Median | 90.0 | 90.5 ENPP3 Positive From Cases With Tissues2 (%) | 83/89 (93.3) | 50/52 (96.2) Prior Therapy | All IV Doses (n=119) | Expansion Doses (n=68) # Prior Regimens, Median (Min, Max) | 3 (1, 10) | 3 (1, 10) 1, n (%) | 7 (5.9) | 4 (5.9) 2 | 35 (29.4) | 22 (32.4) 3 | 30 (25.2) | 18 (26.5) ≥4 | 47 (39.5) | 24 (35.3) Prior Treatments, n (%) Any Immune Checkpoint Inhibitor | 119 (100) | 68 (100) Any VEGFR-TKI | 119 (100) | 68 (100) 2 or More VEGFR-TKIs | 76 (63.9) | 43 (63.2) HIF2α Inhibitor | 42 (35.3) | 23 (33.8) Nephrectomy, n (%) | 81 (68.1) | 46 (67.6) HIF2α, hypoxia-inducible factor 2-alpha; IMDC, International Metastatic RCC Database Consortium; IHC, immunohistochemistry; SOD, sum of diameters; TKI, tyrosine kinase inhibitor; VEGFR, vascular endothelial growth factor receptor Subjects who received ≥1 protocol-specified dose of XmAb819 were included in the baseline characteristics analysis. Data cut-off: 01 September 2026. Excludes 1 participant who experienced approximately 10-fold higher exposure than intended at the first and only dose administered, and who subsequently developed Grade 4 CRS and discontinued treatment. 1IMDC risk derived at study baseline using central laboratory values; KPS <80% defined as ECOG ≥2. 2Tumor ENPP3 positivity was defined as a tumor proportion score (TPS) ≥1, regardless of expression intensity, by a clinical trial IHC assay (Roche Diagnostics). Pal SK, et al. [Slide 3] Patient case study: Deepening of response over time • Background: Male, 57 years old, RCC diagnosed 10 years ago, metastatic disease ~5 years ago • Prior therapy: 2 prior therapy lines cabozantinib + nivolumab (1 mo) → progression & lenvatinib + everolimus (3 yrs) → progression • Baseline: Two enlarged mediastinal lymph nodes both > 2 cm longest diameter multiple small lung and pancreas lesions membrane ENPP3+ <25% tumor cells • Response: PR week 6, deepened to CR week 42 normalization of lymph nodes disappearance of lung and pancreas lesions • Treatment and response ongoing at month 16 [CT images: Baseline | 15.5 Months] Pal SK, et al. [Slide 4] Conclusions • XmAb819 (ENPP3 x CD3) showed meaningful antitumor activity in patients with advanced ccRCC. • Antitumor activity was observed in a dose-dependent manner, with a cORR of 30% at 100 µg/kg, 44% at 200 µg/kg, and 45% at 400 µg/kg, and a majority of responses ongoing at data cut. • Consistent activity in subgroups with prior exposure to HIF2α inhibitors (cORR 35%) and ≥2 VEGFR-TKIs (cORR 45%). • These data support continued development of XmAb819 as a monotherapy for advanced clear cell RCC and evaluation of rational combination strategies in earlier lines of treatment. cORR, confirmed objective response rate; HIF2α, hypoxia-inducible factor 2-alpha; TKI, tyrosine kinase inhibitor; VEGF, vascular endothelial growth factor receptor Pal SK, et al.

[Slide 1] XmAb819-01: Phase 1 study design in advanced ccRCC Study Schema – IV Administration Dose Escalation | Dose Expansion 400 µg/kg | Enrolling | n = 11 200 µg/kg | Complete | n = 27 100 µg/kg | Complete | n = 30 Outcomes reported in current presentation. Eight dose levels explored below expansion (n = 51) Key Eligibility Criteria • R/R ccRCC – prior ICI & VEGFR-TKI • Measurable disease by RECIST v1.1 • No prior anti-ENPP3 therapy Endpoints • Primary: Safety & Tolerability • Secondary: • ORR, DOR, PFS per RECIST 1.1 • Pharmacokinetics NCT05433142. US & EU sites. Screening | Day 1 Priming | 8 Step-Up | 15 Step-Up | 22 TD | 29 TD | 36 TD | 43 TD | 57 TD | 71 TD | 85 TD | 99 TD Cycle 1 | Cycle 2 | Cycle 3 DOR, duration of response; ICI, immune checkpoint inhibitor; IV, intravenous; ORR, objective response rate; PFS, progression free survival; R/R, relapsed / refractory; TD, target dose; TKI, tyrosine kinase inhibitor; VEGFR, vascular endothelial growth factor receptor Pal SK, et al. [Slide 3] Cytokine-driven adverse events • Rash, an on-target effect due to ENPP3 expression on basophils and mast cells, is well managed with daily antihistamine (oral) and steroids (topical & oral). • Dose delays due to rash occurred in 4.4% of subjects, with no treatment discontinuations. • Transaminase elevations were asymptomatic, transient, and self-limited, with a median time to resolution of 7 days and no clinical sequelae. • Dose delays due to transaminase elevations occurred in 14.7% of subjects, with no treatment discontinuations. Rash by Max Grade & Dose (n=68) Rash Grade: 1 | 2 | 3 Bar totals: PD1 79.4% | PD2 52.2% | PD3 42.4% | TD1 26.2% | TD2 7.9% | TD3 9.7% | TD4 1.7% | TD5 5.7% | TD6 [illegible, cropped] Segment labels: PD1 18(26.5), 28(41.2), 8(11.8) | PD2 17(25.4), 13(19.4), 5(7.5) | PD3 14(21.2), 9(13.6), 5(7.6) | TD1 13(20.0), 4(6.2) | TD2 2(3.2), 3(4.8) | TD3 4(6.5), 2(3.2) | TD4 1(1.7) | TD5 3(5.7) Dosed, n | PD1 68 | PD2 67 | PD3 66 | TD1 65 | TD2 63 | TD3 62 | TD4 58 | TD5 53 | TD6 51 Transaminase kinetics by Dose (n=68) ★ Predose ○ 24/48 hr post dose [line plot, Median x ULN (Max of ALT/AST), Grade 1 Max reference line] Combined rash terms (10 PTs): rash (unspecified, maculo-papular, erythematous, [illegible], pustular, pruritic, anal), erythema, dermatitis acneiform. Reported TEAEs were defined per MedDRA and CTCAE. Data cut-off: 01 September 2026. PD, priming dose; TD, target dose Pal SK, et al. [Slide 4] Antitumor activity of 100-400 µg/kg doses in prior therapy subgroups Prior HIF2α Inhibitor (n=23) [waterfall plot, ▶ Ongoing Tx] Prior ≥2 VEGFR-TKIs (n=42) [waterfall plot, ▶ Ongoing Tx] Expanded Doses | Prior HIF2α (n = 23) | ≥2 VEGFR-TKIs (n=42) Confirmed ORR, % (95% CI) | 34.8 (16.4, 57.3) | 45.2 (29.9, 61.3) DCR, % (95% CI) | 78.3 (56.3, 92.5) | 85.7 (71.5, 94.6) Efficacy analysis includes subjects with ≥1 protocol-specified dose of XmAb819 and ≥1 post-baseline tumor assessment. Confirmed responses per investigator assessment, RECIST v1.1. Data cut-off: 01 September 2026. CR subject has two resolved lymph node target lesions with residual mass <10 mm. DCR, disease control rate; HIF2α, hypoxia-inducible factor 2-alpha; ORR, objective response rate; TKI, tyrosine kinase inhibitor; Tx, treatment; VEGFR, vascular endothelial growth factor receptor Pal SK, et al.

[Slide 1] MADRID 2026 ESMO congress Phase 1 results of XmAb819, a novel ENPP3 x CD3 bispecific, in advanced clear cell renal cell carcinoma (ccRCC) Sumanta K. Pal1, Benjamin Garmezy2, Randy F. Sweis3, Mehmet A. Bilen4, Scott S. Tykodi5, Yuanquan Yang6, Karie Runcie7, David A. Braun8, David J. VanderWeele9, Claud M. Grigg10, Christopher Hoimes11, Adanma Ayanambakkam12, Joseph J. Maly13, Robert A. Franklin14, Biren Saraiya15, Chet Bohac16, Jitendra Kanodia16, Nicholas Rinella16, Ritesh R. Kotecha17, Laurence Albiges18 1City of Hope Comprehensive Cancer Center, Duarte, USA; 2Sarah Cannon Research Institute, Nashville, USA, 3University of Chicago, Chicago, USA, 4Winship Cancer Institute of Emory University, Atlanta, USA, 5University of Washington and Fred Hutchinson Cancer Center, Seattle, USA, 6The Ohio State University Comprehensive Cancer Center, Columbus, USA, 7Columbia University, New York, USA, 8Yale School of Medicine, New Haven, USA, 9Northwestern University, Chicago, USA, 10Atrium Health Wake Forest Baptist Comprehensive Cancer Center, Charlotte, NC, USA, 11Duke University Medical Center, Durham, USA, 12University of Oklahoma Health Sciences Center, Oklahoma City, USA, 13Norton Cancer Institute - St Matthews Campus, Louisville, USA, 14University of Cincinnati Medical Center, Cincinnati, OH, USA, 15Rutgers Cancer Institute, New Brunswick, USA, 16Xencor, Inc., Pasadena, California, USA, 17 Memorial Sloan Kettering Cancer Center, New York, USA, 18Institut Gustave Roussy, Villejuif, France 25 October 2026 Abstract 2684 [Slide 3] Antitumor activity at target dose of 100 µg/kg [Waterfall plot: Best % Change from Baseline; Ongoing Tx] [Spider plot: % Change from Baseline vs Study Months; Ongoing Tx] Efficacy Evaluable | 100 µg/kg (n = 30) Confirmed ORR, % (95% CI) | 30.0 (14.7, 49.4) BOR, n (%) | CR | 1 (3.3) PR | 8 (26.7) SD | 17 (56.7) PD | 4 (13.3) DCR, % (95% CI) | 86.7 (69.3, 96.2) Median follow-up, months (range) | 4.1 (1.1, 28.3) Efficacy analysis includes subjects with ≥1 protocol-specified dose of XmAb819 and ≥1 post-baseline tumor assessment. Confirmed responses per investigator assessment, RECIST v1.1. Data cut-off: 01 September 2026. CR subject has two resolved lymph node target lesions with residual mass <10 mm BOR, best overall response; CR, complete response; DCR, disease control rate; ORR, objective response rate; PD, progressive disease; PR, partial response; SD, stable disease; Tx, treatment Pal SK, et al. [Slide 4] Antitumor activity at target dose of 400 µg/kg [Waterfall plot: Best % Change from Baseline; bar labels PD, SD, SD, SD, SD, PR, PR, PD, PR, PR, PR; Ongoing Tx] [Spider plot: % Change from Baseline vs Study Months; Ongoing Tx] Efficacy Evaluable | 400 µg/kg (n = 11) Confirmed ORR, % (95% CI) | 45.5 (16.8, 76.6) BOR, n (%) | CR | - PR | 5 (45.5) SD | 4 (36.4) PD | 2 (18.2) DCR, % (95% CI) | 81.8 (48.2, 97.7) Median follow-up, months (range) | 6.0 (1.4, 7.9) Efficacy analysis includes subjects with ≥1 protocol-specified dose of XmAb819 and ≥1 post-baseline tumor assessment. Confirmed responses per investigator assessment, RECIST v1.1. Data cut-off: 01 September 2026. BOR, best overall response; CR, complete response; DCR, disease control rate; ORR, objective response rate; PD, progressive disease; PR, partial response; SD, stable disease; Tx, treatment Pal SK, et al.

[Slide 1] What remains to be answered (1) How limiting is toxicity? Can we actually give it - and can every unit do it? What was reported · expansion doses, n=68 TEAE, % | All grades | Grade >=3 Rash (combined term) | 91.2 | 22.1 Cytokine release syndrome | 67.6 | 5.9 ALT increased | 52.9 | 33.8 AST increased | 48.5 | 35.3 Hypotension | 29.4 | 7.4 72% related grade >=3 AEs 54% related serious AEs 7% discontinued for AEs 0 deaths from AEs One further patient - ~10-fold overexposure at the first dose, grade 4 CRS - is excluded from every analysis. Pal SK et al, ESMO 2026, abstract 2684 (data cutoff 1 Sep 2026). CRS graded per ASTCT; TEAEs in >25% of subjects shown. Manuela Schmidinger, Medical University of Vienna, Austria What it means in practice The first month needs a centre Priming doses need a unit that recognises and treats CRS - dexamethasone before each one, prophylactic tocilizumab before the first. After that it settles New CRS after day 29 in 18%, new rash in 29% and almost none severe. Transaminase rises transient, median resolution 7 days. Defined CRS management protocol [Slide 2] In summary: XmAb819: where this leaves us Points to consider: Durability Dose: ORR rises with dose, but PD in 36% at 200 vs 13% at 100 µg/kg ENPP3 density threshold: the complete responder had ENPP3 on fewer than 25% of tumor cells: does intensity predict response? Deliverability in the first month Verdict A strong signal: enthusiasm is justified - and this is only the floor... worth a phase 3 - after IO and VEGF-TKI or earlier Pal SK et al, ESMO 2026, abstract 2684 (data cutoff 1 Sep 2026) Manuela Schmidinger, Medical University of Vienna, Austria [Slide 3] 2 What makes ivonescimab special? One molecule instead of two drugs - does it make treatment less toxic? And does it make it more effective? WHY IT COULD BE LESS TOXIC It takes the ligand, not the receptor VEGFR-TKI endothelial cell VEGFR blocked and inhibits these as well MET | AXL | KIT | FGFR | RET diarrhoea · hand-foot · fatigue · hypothyroidism IVONESCIMAB endothelial cell VEGFR untouched VEGF-A caught before it binds no other kinase is inhibited hypertension · proteinuria only Fc-silenced (L234A/L235A): it does not kill the PD-1+ T cells it binds - in vitro The toxicity argument is about what is not in the molecule WHY IT COULD BE MORE EFFECTIVE Where VEGF-A is high, ivo grips with both arms TUMOUR VEGF-A high PD-1 T cell MUTUAL COOPERATIVITY Once VEGF-A is bound -> PD-1 binding becomes ~18 times stronger Once PD-1 is bound -> VEGF-A binding becomes >4 times stronger in vitro, manufacturer's own data Design claim NORMAL TISSUE VEGF-A low - but not absent PD-1 T cell ONE ARM ONLY only the PD-1 arm holds no VEGF-A to tighten the grip -> ordinary PD-1 blockade a gradient, not an on/off switch [illegible footnote text: references incl. HARMONi-2 (Xiong A et al, Lancet 2025...) and Zhong T et al, Science 2025] Manuela Schmidinger, Medical University of Vienna, Austria [Slide 4] Take-home messages XmAb819 · REDIRECTING ENPP3 × CD3 T-cell engager · after IO and VEGF-TKI MODE OF ACTION A new mode of action WHAT IT SHOWED Enthusiasm is justified - the floor, not the ceiling NEXT Ready for phase 3 - and for earlier lines Ivonescimab · COMBINING PD-1 × VEGF-A bispecific · treatment-naive, favourable risk MODE OF ACTION A known mode of action in a new format WHAT IT SHOWED In the IO-TKI range - with a toxicity profile worth enthusiasm NEXT Ready to randomise - against what, and in whom? Pal SK et al, ESMO 2025, abstract 2684 (data cutoff 1 Sep 2026); Yan X et al., ESMO 2026, abstract 8473 (data cutoff 14 Jul 2026) Manuela Schmidinger, Medical University of Vienna, Austria

Depth and duration of responses Spider Plot of Confirmed Responders Across All IV Doses (n = 27) ■ Low doses (N=2) ■ Expanded doses (N=25) ▶ Ongoing Tx [Spider plot: % Change from Baseline vs Study Months; axis ticks not transcribed] Responders | Expansion Doses (n=25) Ongoing Responses, n (%) | 19 (76.0) Median TTR, months (range) | 2.7 (1.3, 20.9) Median time on treatment, months (range) | 7.0 (3.4, 28.3) Efficacy analysis includes subjects with ≥1 protocol-specified dose of XmAb819 and ≥1 post-baseline tumor assessment. Confirmed responses per investigator assessment, RECIST v1.1. Data cut-off: 01 September 2026. TTR, time to response; Tx, treatment Pal SK, et al.

[Slide 1] Cytokine release syndrome (CRS) CRS by Max Grade & Dose (n=68) CRS Grade 1 2 3 4 PD1: 45.6%; 13(19.1) grade 1; 16(23.5) grade 2; 2(2.9) grade 3 PD2: 35.8%; 7(10.4) grade 1; 15(22.4) grade 2; 2(3.0) grade 3 PD3: 25.8%; 4(6.1) grade 1; 12(18.2) grade 2; 1(1.5) grade 4 TD1: 6.2%; 2(3.1) grade 1; 1(1.5) grade 2; 1(1.5) grade 3 TD3: 1.6%; 1(1.6) TD4: 1.7%; 1(1.7) TD7: 4%; 1(2.0); 1(2.0) TD8: 2.1%; 1(2.1) TD9: 2.3%; 1(2.3) Dosed, n | PD1 68 | PD2 67 | PD3 66 | TD1 65 | TD2 63 | TD3 62 | TD4 58 | TD5 53 | TD6 51 | TD7 50 | TD8 48 | TD9 43 • CRS was predominately low-grade and occurred during priming • No dose delays and one treatment discontinuation due to CRS • Premedications for XmAb819: ▪ Acetaminophen, diphenhydramine, and IV normal saline ▪ Dexamethasone prior to priming doses ▪ Tocilizumab prior to priming dose 1 (PD1) • Graded per ASTCT consensus criteria; managed per institutional standards, or NCCN guidelines when local standards were absent Subjects who received ≥1 protocol-specified dose of XmAb819 were included in the safety analysis. Data cut-off: 01 September 2026. Excludes 1 participant who experienced approximately 10-fold higher exposure than intended at the first and only dose administered, and who subsequently developed Grade 4 CRS and discontinued treatment. Reported TEAEs were defined per MedDRA and CTCAE; CRS graded ASTCT1. Per-protocol PK samples were not collected for subject in plot (left) with reported Grade 4 CRS at priming dose 3; drug exposure could not be determined. Two subjects each experienced 2 events of Grade ≥3 CRS. 1Lee DW, et al. Biol Blood Marrow Transplant. 2019;25:625-638. ASTCT, American Society for Transplantation and Cellular Therapy; CRS, cytokine release syndrome; CTCAE, Common Terminology Criteria for Adverse Events; MedDRA, Medical Dictionary for Regulatory Activities; NCCN, National Comprehensive Cancer Network; PD, priming dose; TEAE, treatment-emergent adverse event; TD, target dose Pal SK, et al. [Slide 2] Antitumor activity at target dose of 200 µg/kg [Waterfall plot: Best % Change from Baseline; Ongoing Tx] [Spider plot: % Change from Baseline vs Study Months; Ongoing Tx] Efficacy Evaluable1 | 200 µg/kg (n = 25) Confirmed ORR, % (95% CI) | 44.0 (24.4, 65.1) BOR, n (%) | CR | - PR | 11 (44.0) SD | 5 (20.0) PD | 9 (36.0) DCR, % (95% CI) | 64.0 (42.5, 82.0) Median follow-up, months (range) | 4.5 (0.8, 11.6) Efficacy analysis includes subjects with ≥1 protocol-specified dose of XmAb819 and ≥1 post-baseline tumor assessment. Confirmed responses per investigator assessment, RECIST v1.1. Data cut-off: 01 September 2026. 1Two subjects at 200 µg/kg discontinued treatment without a post-baseline response assessment and without documented progression. BOR, best overall response; CR, complete response; DCR, disease control rate; ORR, objective response rate; PD, progressive disease; PR, partial response; SD, stable disease; Tx, treatment Pal SK, et al.

XmAb819-01: Phase 1 study design in advanced ccRCC Study Schema – IV Administration Dose Escalation | Dose Expansion 400 µg/kg — Enrolling — n = 11 200 µg/kg — Complete — n = 27 100 µg/kg — Complete — n = 30 Outcomes reported in current presentation. Eight dose levels explored below expansion (n = 51) Key Eligibility Criteria • R/R ccRCC – prior ICI & VEGFR-TKI • Measurable disease by RECIST v1.1 • No prior anti-ENPP3 therapy Endpoints • Primary: Safety & Tolerability • Secondary: ORR, DOR, PFS per RECIST 1.1; Pharmacokinetics NCT05433142. US & EU sites. Screening | Day 1 Priming | 8 Step-Up | 15 Step-Up | 22 TD | 29 TD | 36 TD (Cycle 1) | 43 TD | 57 TD (Cycle 2) | 71 TD | 85 TD (Cycle 3) | 99 TD

Phase 1 results of XmAb819, a novel ENPP3 x CD3 bispecific, in advanced clear cell renal cell carcinoma (ccRCC) Sumanta K. Pal(1), Benjamin Garmezy(2), Randy F. Sweis(3), Mehmet A. Bilen(4), Scott S. Tykodi(5), Yuanquan Yang(6), Karie Runcie(7), David A. Braun(8), David J. VanderWeele(9), Claud M. Grigg(10), Christopher Hoimes(11), Adanma Ayanambakkam(12), Joseph J. Maly(13), Robert A. Franklin(14), Biren Saraiya(15), Chet Bohac(16), Jitendra Kanodia(16), Nicholas Rinella(16), Ritesh R. Kotecha(17), Laurence Albiges(18) (1)City of Hope Comprehensive Cancer Center, Duarte, USA; (2)Sarah Cannon Research Institute, Nashville, USA; (3)University of Chicago, Chicago, USA; (4)Winship Cancer Institute of Emory University, Atlanta, USA; (5)University of Washington and Fred Hutchinson Cancer Center, Seattle, USA; (6)The Ohio State University Comprehensive Cancer Center, Columbus, USA; (7)Columbia University, New York, USA; (8)Yale School of Medicine, New Haven, USA; (9)Northwestern University, Chicago, USA; (10)Atrium Health Wake Forest Baptist Comprehensive Cancer Center, Charlotte, NC, USA; (11)Duke University Medical Center, Durham, USA; (12)University of Oklahoma Health Sciences Center, Oklahoma City, USA; (13)Norton Cancer Institute - St Matthews Campus, Louisville, USA; (14)University of Cincinnati Medical Center, Cincinnati, OH, USA; (15)Rutgers Cancer Institute, New Brunswick, USA; (16)Xencor, Inc., Pasadena, California, USA; (17)Memorial Sloan Kettering Cancer Center, New York, USA; (18)Institut Gustave Roussy, Villejuif, France 25 October 2026 Abstract 2684

[Slide 1] MADRID 2026 ESMO congress PErioperative PEmbrolizumab (P) plus metastasis-directed therapy (MDT) versus MDT alone in patients (pts) with oligometastatic renal cell carcinoma (omRCC): the randomized phase II PE-PE trial. Roberto Iacovelli1, Alberto Dalla Volta2, Davide Bimbatti3, Sebastiano Buti4, Emanuela Fantinel5, Carlo Arcara6, Francesco Massari7, Lucia Bonomi8, Mimma Rizzo9, Elisa Zanardi10, Franco Nolè11, Helga Lipari12, Paolo Andrea Zucali13, Claudia Piombino14, Mirko Bonalda15, Elena Verzoni16, Giuseppe Prati17, Matteo Santoni18, Claudia Caserta19, Chiara Ciccarese20. 1-Fondazione Policlinico Universitario Agostino Gemelli, IRCCS, Rome, Italy. 2-Unit of Medical Oncology, Department of Medical and Surgical Specialties, Radiological Sciences, and Public Health, ASST Spedali Civili di Brescia, University of Brescia, Brescia, Italy. 3-Oncology 1 Unit, Istituto Oncologico Veneto, IOV - IRCCS, Padua, Italy. 4-Medicine and Surgery Department, University of Parma, Viale Antonio Gramsci, 14, 43126, Parma, Italy. 5-Department of Oncology, Azienda Ospedaliera Universitaria Integrata di Verona, University of Verona, Verona, Italy. 6-Medical Oncology Unit, La Maddalena Hospital, 90146 Palermo, Italy. 7-Medical Oncology, IRCCS Azienda Ospedaliero-Universitaria di Bologna, Bologna, Italy. 8-Medical Oncology Unit, ASST Papa Giovanni XXIII, Bergamo, Italy. 9-Division of Medical Oncology, A.O.U. Consorziale Policlinico di Bari, Bari, Italy. 10-IRCCS Azienda Ospedaliera Metropolitana, Genoa, Italy. 11-IRCCS Istituto Europeo di Oncologia, Milan, Italy. 12-Medical Oncology Unit, Ospedale Cannizzaro, Catania, Italy. 13-Department of Biomedical Sciences, Humanitas University, Pieve Emanuele, Milan, Italy. 14-Division of Oncology, Department of Oncology and Hematology, University Hospital of Modena, Modena, Italy. 15-Medical Oncology Unit 2, Santa Chiara Hospital, Azienda Ospedaliero-Universitaria Pisana, Pisa, Italy. 16-Department of Medical Oncology, Fondazione IRCCS Istituto Nazionale Dei Tumori di Milano, Milan, Italy. 17-Medical Oncology. Comprehensive Cancer Centre, AUSL-IRCCS di Reggio Emilia, Reggio Emilia, Italy. 18-Medical Oncology Unit, Macerata Hospital, Macerata, Italy. 19-Oncology Unit, Santa Maria della Misericordia Hospital Perugia, Italy. [Slide 2] Study Design 85 Pts with • Clear cell histology; • Previous nephrectomy; • ECOG PS 0 or 1; • No brain mets; • Oligometastatic disease defined by PE-PE criteria; • No previous medical treatment for RCC R 2:1 Pembro 400 mg, 1 cycle IV 3 wks before MDT -> MDT (surgery and/or SABR) -> Pembro 400 mg, 8 cycles IV Q6 wks MDT (surgery and/or SABR) Primary endpoint: relapse-free survival (RFS)*. Secondary endpoints: distant RFS*; OS*; safety. Additional secondary endpoint not presented today: local progression in SABR population. MDT: metastasis directed therapy SABR: stereotactic ablative radiotherapy * Evaluated in patients who received MDT. Roberto Iacovelli M.D. Ph.D @DrIacovelli [Slide 3] Population: PE-PE criteria for oligometastatic RCC Timing Nephrectomy -> < 5 years -> metastasis (eligible) OR Nephrectomy -> Metastasectomy -> < 5 years -> metastasis (eligible) Tumor Extension Max 3 lesions allowed. If bone mets. -> Max 2 lesions allowed. AND Each lesion diameter should not be more than 3 cm; AND Sum of the lesions' diameters should not be more than 5 cm. Roberto Iacovelli M.D. Ph.D @DrIacovelli [Slide 4] Primary endpoint: relapse-free survival (RFS) Relapse-Free Survival [%] vs time [months] | Pembro + MDT N=54 | MDT N=29 Events | 19 | 16 Median (mos) | 25.4 | 11.0 HR (95%CI) | 0.47 (0.24 – 0.92); p=0.024 The median follow-up was 17.5 months (cut-off date of July 22nd, 2026). N° at risk (0,6,12,18,24,30,36,42 months): Pembro + MDT: 54 | 39 | 27 | 16 | 10 | 3 | 2 | 0 MDT: 29 | 17 | 7 | 4 | 3 | 1 | 0 | 0 Roberto Iacovelli M.D. Ph.D @DrIacovelli Survival table Endpoint: Relapse-free survival (primary) Pembro + MDT | 54 | 25.4 mo MDT | 29 | 11.0 mo HR: 0.47 HR CI: 0.24 – 0.92 P: p=0.024

Baseline characteristics Characteristic, n (%) | Pembrolizumab + MDT (N = 54) | MDT (N = 29) Median age (range), years | 64 (45-77) | 64 (40-73) Sex Male | 40 (74.1) | 22 (75.9) Female | 14 (25.9) | 7 (24.1) Nephrectomy Partial | 9 (16.7) | 4 (13.8) Radical | 45 (83.3) | 25 (86.2) Initial stage T1-T2,N0,M0 | 21 (38.9) | 13 (44.8) T3-T4 or N1 or M1 | 33 (61.1) | 16 (55.2) Sarcomatoid features | 1/35 (2.9) | 2/16 (12.5) Time from Nx to Rand >1 year | 46 (85.2) | 27 (93.1) Previous metastasectomy | 17 (31.5) | 15 (51.7) Characteristic, n (%) | Pembrolizumab + MDT (N = 54) | MDT (N = 29) Number of lesions 1 | 32 (59.3) | 13 (44.8) 2-3 | 22 (40.7) | 16 (55.2) Sum of LTDs, median (range), mm | 23 (10 - 54) | 23 (10 - 49) Site of lesions Lung | 19 (35.2) | 10 (34.5) Lymph nodes | 8 (14.8) | 7 (24.1) Adrenal gland | 8 (14.8) | 3 (10.3) Bone | 8 (14.8) | 2 (6.9) Pancreas | 1 (1.9) | 4 (13.8) Other | 10 (18.5) | 3 (10.3) IMDC prognostic risk categories Favorable | 42 (77.8) | 22 (75.9) Intermediate | 8 (14.8) | 6 (20.7) Poor | 4 (7.4) | 1 (3.4) LTD: Longest tumor diameter Nx: Nephrectomy Rand: randomization Roberto Iacovelli M.D. Ph.D | @DrIacovelli

[Slide 2] Primary endpoint: relapse-free survival (RFS) Prespecified explorative subgroup analysis: | HR (95% CI) Overall | 0.47 (0.24–0.92) Sex Male | 0.54 (0.24–1.23) Female | 0.37 (0.11–1.24) Age <65 | 0.50 (0.20–1.27) ≥65 | 0.48 (0.18–1.24) Local Procedure SABR | 0.47 (0.22–0.97) Surgery | 0.62 (0.12–3.08) N of lesions 1 | 0.34 (0.10–1.15) 2–3 | 0.62 (0.27–1.42) Axis: 0.5 | 1 | 2 ; Better Pembro + MDT | Better MDT Roberto Iacovelli M.D. Ph.D @DrIacovelli [Slide 3] Secondary endpoint: Safety Pts with AEs, n (%) | Pembrolizumab + MDT (N = 56) | MDT (N = 29) Any | 45 (80.4) | 6 (20.7) Grade ≥ 3 | 4 (7.1) | 1 (3.4) Pembro-related AEs Any | 33 (58.9) | - Grade ≥ 3 | 2 (3.6)* | - Tx discontinuation | 4 (7.1) | - SABR-related AEs Any | 3 (5.4) | 2 (6.9) Grade ≥ 3 | 0 | 0 Surgery related AEs Any | 0 | 0 * One G3 hyperglycemia and one G3 diarrhea. Median number of pembrolizumab cycles was 6.6 AEs in ≥5% of pts, n (%) | Pembrolizumab + MDT (N = 56) All Grades | High Grade | MDT (N=29) All Grades | High Grade Hypothyroidism | 13 (23.2) | 0 | 0 | 0 Fatigue | 10 (17.9) | 0 | 0 | 0 Diarrhea | 9 (16.1) | 1 (1.8) | 0 | 0 Constipation | 5 (8.9) | 0 | 0 | 0 Cough | 5 (8.9) | 0 | 1 (3.4) | 0 Pneumonitis | 5 (8.9) | 0 | 0 | 0 ALT increase | 4 (7.1) | 0 | 0 | 0 Nausea | 4 (7.1) | 0 | 0 | 0 Pruritus | 4 (7.1) | 0 | 0 | 0 Anemia | 3 (5.4) | 0 | 0 | 0 Arthralgia | 3 (5.4) | 0 | 0 | 0 Depression | 3 (5.4) | 1 (1.8) | 0 | 0 Asthenia | 3 (5.4) | 0 | 1 (3.4) | 0 Hypertension | 3 (5.4) | 0 | 0 | 0 Roberto Iacovelli M.D. Ph.D @DrIacovelli [Slide 4] Conclusions • The addition of pembrolizumab resulted in a statistically significant and clinically meaningful improvement in relapse-free survival (RFS) among patients with omRCC eligible for MDT (HR 0.47; 95% CI 0.24–0.92; p = 0.024). • Despite the low disease burden and the long interval from previous nephrectomy, patients with omRCC included in PE-PE trial and treated with MDT alone continued to have a poor prognosis, with >50% experiencing disease progression within one year. • Due to the immature follow-up, no difference in overall survival (OS) can be demonstrated. • The safety profile of pembrolizumab was consistent with previous reports of adjuvant pembrolizumab in RCC with no concerns for combination with SABR. • PE-PE trial is the first to show that perioperative/adjuvant pembrolizumab increases the RFS in patients with omRCC eligible for definitive therapy (surgery or SABR) within five years from previous nephrectomy or metastasectomy. Roberto Iacovelli M.D. Ph.D @DrIacovelli

Secondary endpoint: distant-RFS | Pembro + MDT N=54 | MDT N=29 Events | 18 | 15 Median (mos) | 25.4 | 11.5 95%CI | N.E. | 5.7 – 17.3 94% of progressions were distant events N° at risk: Pembro + MDT: 54 39 27 16 10 3 2 0 MDT: 29 17 7 4 3 1 0 0 (time [months] 0, 6, 12, 18, 24, 30, 36, 42) Survival table Endpoint: Distant relapse-free survival (secondary) Arm | N | Median | Median CI Pembro + MDT | 54 | 25.4 mo | N.E. MDT | 29 | 11.5 mo | 5.7–17.3

[Slide 1] PErioperative PEmbrolizumab (P) plus metastasis-directed therapy (MDT) versus MDT alone in patients (pts) with oligometastatic renal cell carcinoma (omRCC): the randomized phase II PE-PE trial. Roberto Iacovelli1, Alberto Dalla Volta2, Davide Bimbatti3, Sebastiano Buti4, Emanuela Fantinel5, Carlo Arcara6, Francesco Massari7, Lucia Bonomi8, Mimma Rizzo9, Elisa Zanardi10, Franco Nolè11, Helga Lipari12, Paolo Andrea Zucali13, Claudia Piombino14, Mirko Bonalda15, Elena Verzoni16, Giuseppe Prati17, Matteo Santoni18, Claudia Caserta19, Chiara Ciccarese20. [Affiliations list in small print: 1-Fondazione Policlinico Universitario Agostino Gemelli, IRCCS, Rome, Italy; ... 19-Oncology Unit, Santa Maria della Misericordia Hospital Perugia, Italy - remainder [illegible] at this resolution] [Slide 4] Secondary endpoint: distant-RFS | Pembro + MDT N=54 | MDT N=29 Events | 18 | 15 Median (mos) | 25.4 | 11.5 95%CI | N.E. | 5.7 – 17.3 94% of progressions were distant events N° at risk (months 0/6/12/18/24/30/36/42): Pembro + MDT: 54 | 39 | 27 | 16 | 10 | 3 | 2 | 0 MDT: 29 | 17 | 7 | 4 | 3 | 1 | 0 | 0 Roberto Iacovelli M.D. Ph.D @DrIacovelli Survival table Endpoint: Distant relapse-free survival (secondary) Arm | N | Median | Median CI Pembro + MDT | 54 | 25.4 mo | N.E. MDT | 29 | 11.5 mo | 5.7–17.3

Secondary endpoint: OS | Pembro + MDT N=54 | MDT N=29 Events | 1 | 1 24-mos OS | 97.8% | 96.2% N° at risk (0, 6, 12, 18, 24, 30, 36, 42 months) Pembro + MDT: 54 47 38 27 19 7 3 0 MDT: 29 24 18 11 9 5 2 0 Roberto Iacovelli M.D. Ph.D @DrIacovelli Survival table Endpoint: Overall survival Pembro + MDT | 54 MDT | 29 Landmark rates: 24 months: Pembro + MDT 97.8% | MDT 96.2%

[Slide 2] Study Design 85 Pts with • Clear cell histology; • Previous nephrectomy; • ECOG PS 0 or 1; • No brain mets; • Oligometastatic disease defined by PE-PE criteria; • No previous medical treatment for RCC R 2:1 → Pembro 400 mg, 1 cycle IV 3 wks before MDT → MDT (surgery and/or SABR) → Pembro 400 mg, 8 cycles IV Q6 wks → MDT (surgery and/or SABR) Primary endpoint: relapse-free survival (RFS)*. Secondary endpoints: distant RFS*; OS*; safety. Additional secondary endpoint not presented today: local progression in SABR population [Slide 3] Patient disposition 108 patients screened between March 2023 and February 2026 85 patients enrolled Pembrolizumab + MDT 56 enrolled | MDT 29 enrolled 2 had not MDT: 1 due to pembro-related G3 AE. 1 due to SABR technical issues 54 treated with MDT: 41 (75.9%) SABR; 13 (24.1%) surgery | 29 treated with MDT: 21 (72.4%) SABR; 8 (27.6%) surgery 25 (44.6%) Completed 9 cycles of pembro; 19 (33.9%) Progressed; 11 (19.6%) Ongoing | 16 (55.1%) Progressed Pembro discontinued before 9 cycles for 10 (17.9%) Progression of disease 6 (10.7%) Patient's decision/withdrawal 4 (7.1%) Adverse event

[Slide 1] CAPItello-281 updated DCO: OS Clinically meaningful improvement in OS for capi + abi | Capi + abi (N=507) | Pbo + abi (N=505) Events, n (%) | 195 (38.5) | 226 (44.8) Median, months (95% CI) | 57.4 (47.0, NC) | 46.5 (40.7, 53.2) HR (95% CI) | 0.82 (0.68, 0.99) P-value | 0.044* *Descriptive OS analysis: alpha controlled, did not meet statistical significance. Capivasertib + abiraterone Placebo + abiraterone Probability of OS Time from randomisation (months) Number of patients at risk Months | 0 | 3 | 6 | 9 | 12 | 15 | 18 | 21 | 24 | 27 | 30 | 33 | 36 | 39 | 42 | 45 | 48 | 51 | 54 | 57 | 60 | 63 | 66 Capi + abi | 507 | 486 | 473 | 454 | 441 | 415 | 398 | 387 | 366 | 338 | 291 | 256 | 215 | 176 | 149 | 118 | 91 | 60 | 45 | 25 | 10 | 2 | 0 Pbo + abi | 505 | 493 | 478 | 457 | 438 | 409 | 381 | 359 | 344 | 313 | 263 | 233 | 195 | 158 | 126 | 106 | 81 | 58 | 42 | 25 | 13 | 5 | 0 Survival table Endpoint: OS (updated DCO) Arm | N | Median | Median CI Capivasertib + abiraterone | 507 | 57.4 | 47.0, NC Placebo + abiraterone | 505 | 46.5 | 40.7, 53.2 HR: 0.82 HR CI: 0.68, 0.99 P: 0.044* [Slide 2] Capi + abi (N=507) | Pbo + abi (N=505) Events, n (%) | 249 (49.1) | 283 (56.0) Median, months (95% CI) | 35.5 (29.7, 40.5) | 26.0 (22.4, 31.3) HR (95% CI) | 0.83 (0.70, 0.99) Nominal P-value | 0.033* *Descriptive rPFS analysis and not formally tested. 26.0 months | 9.5 months | 35.5 months Capivasertib + abiraterone Placebo + abiraterone Probability of rPFS Time from randomisation (months) Number of patients at risk Months | 0 | 3 | 6 | 9 | 12 | 15 | 18 | 21 | 24 | 27 | 30 | 33 | 36 | 39 | 42 | 45 | 48 | 51 | 54 | 57 | 60 | 63 Capi + abi | 507 | 460 | 437 | 385 | 341 | 314 | 302 | 277 | 250 | 213 | 184 | 176 | 140 | 106 | 86 | 60 | 46 | 36 | 26 | 11 | 2 | 0 Pbo + abi | 505 | 479 | 440 | 389 | 341 | 302 | 286 | 266 | 238 | 194 | 161 | 152 | 119 | 96 | 74 | 58 | 42 | 39 | 26 | 12 | 2 | 0 Survival table Endpoint: rPFS (updated DCO) Arm | N | Median | Median CI Capivasertib + abiraterone | 507 | 35.5 | 29.7, 40.5 Placebo + abiraterone | 505 | 26.0 | 22.4, 31.3 HR: 0.83 HR CI: 0.70, 0.99 P: 0.033* (nominal) [Slide 3] Capi + abi (N=507) | Pbo + abi (N=505) Events, n (%) | 223 (44.0) | 255 (50.5) Median, months (95% CI) | 43.8 (39.1, 53.1) | 35.9 (32.1, 42.4) HR (95% CI) | 0.83 (0.69, 0.99) Nominal P-value | 0.044 35.9 months | 7.9 months | 43.8 months Capivasertib + abiraterone Placebo + abiraterone Probability of SSE-FS Time from randomisation (months) Number of patients at risk Months | 0 | 3 | 6 | 9 | 12 | 15 | 18 | 21 | 24 | 27 | 30 | 33 | 36 | 39 | 42 | 45 | 48 | 51 | 54 | 57 | 60 | 63 | 66 Capi + abi | 507 | 461 | 444 | 423 | 400 | 373 | 352 | 334 | 310 | 285 | 250 | 218 | 171 | 148 | 119 | 94 | 67 | 44 | 36 | 19 | 6 | 2 | 0 Pbo + abi | 505 | 481 | 457 | 429 | 399 | 366 | 338 | 320 | 295 | 258 | 223 | 179 | 143 | 126 | 93 | 80 | 56 | 41 | 30 | 18 | 6 | 2 | 0 Survival table Endpoint: SSE-FS (updated DCO) Arm | N | Median | Median CI Capivasertib + abiraterone | 507 | 43.8 | 39.1, 53.1 Placebo + abiraterone | 505 | 35.9 | 32.1, 42.4 HR: 0.83 HR CI: 0.69, 0.99 P: 0.044 (nominal) [Slide 4] Capi + abi (N=503) | Pbo + abi (N=503) Total (%)/Grade >=3 (%) | Grade 1 | Grade 2 | >=Grade 3 AE | Capi + abi Total/Grade >=3 | Pbo + abi Total/Grade >=3 Diarrhoea | 52.7/6.4 | 9.5/0.4 Hyperglycaemia* | 40.4/11.3 | 13.9/1.0 Rash† | 36.4/12.7 | 8.0/0.2 Anaemia | 27.0/5.6 | 14.9/1.6 Hypokalaemia | 24.3/10.1 | 15.7/6.2 Hypertension | 24.1/7.6 | 26.4/9.5 Fatigue | 16.3/0.4 | 14.3/0.8 Urinary tract infection | 15.3/4.6 | 11.9/1.6 ALT increased | 15.1/4.4 | 14.1/3.6 AST increased | 13.5/2.8 | 12.9/2.2 Nausea | 13.1/0.8 | 5.4/0.2 Asthenia | 12.5/1.0 | 5.8/0.4 COVID-19 | 12.1/0.8 | 9.3/0.6 Back pain | 11.7/0.6 | 13.1/0.6 Pyrexia | 11.3/1.2 | 3.2/0 Hot flush | 11.1/0.2 | 13.9/0 Pruritus | 11.1/0.2 | 2.6/0 Blood creatine increase | 10.7/0.6 | 4.0/0 Arthralgia | 10.1/0.4 | 11.1/0.4 Decreased appetite | 10.1/1.0 | 4.2/0 Constipation | 9.9/0.4 | 13.9/0

CAPItello-281 updated DCO: OS Clinically meaningful improvement in OS for capi + abi Events/patients, n/N (%) Capi + abi (N=507) | Pbo + abi (N=505) | HR (95% CI) All Patients 195/507 (38.5) | 226/505 (44.8) | 0.82 (0.68, 0.99) M1 volume/visceral mets: High with visceral mets 44/98 (44.9) | 50/94 (53.2) | 0.74 (0.49, 1.11) High without visceral mets 122/276 (44.2) | 141/283 (49.8) | 0.89 (0.70, 1.14) Low 28/131 (21.4) | 35/127 (27.6) | 0.70 (0.42, 1.15) Forest axis 0.5 - 1.0 - 2.0: Favours capi + abi <- HR (95% CI) -> Favours pbo + abi Legend: Capivasertib + abiraterone; Placebo + abiraterone Y axis: Probability of OS; X axis: Time from randomisation (months) Number of patients at risk (0,3,...,66 mo): Capi + abi: 507 486 473 454 441 415 398 387 366 338 291 256 215 176 149 118 91 60 45 25 10 2 0 Pbo + abi: 505 493 478 457 438 409 381 359 344 313 263 233 195 158 126 106 81 58 42 25 13 5 0 Data cut-off: 9 Mar 2026. Tick marks represent censored observations. The analysis was carried out using a stratified Cox proportional-hazards model adjusting for the variables selected in the pooling strategy: volume of disease/visceral metastases. HR (95% CI) for OS at the primary rPFS analysis was 0.90 (0.71, 1.15). Predefined OS subgroup analyses are available in the article. abi, abiraterone; capi, capivasertib; CI, confidence interval; DCO, data cut-off; HR, hazard ratio; NC, not calculable; OS, overall survival; pbo, placebo; rPFS, radiographic progression-free survival. Karim Fizazi MD, PhD Survival table Endpoint: OS (updated DCO) Capivasertib + abiraterone | 507 Placebo + abiraterone | 505 HR: 0.82 HR CI: 0.68, 0.99

[Slide 1] 13:30 - 15:00 Rapid oral: Genitourinary tumours, prostate, penile and testis CHAIRS: THOMAS HWANG, EDMOND KWAN, ALISON REID CAPItello-281: Study design A global, multicentre, Phase III, randomised, double-blind study Patients with PTEN-deficient de novo mHSPC • PTEN deficiency: (diagnostic cut-off of ≥90% of viable malignant cells with no specific cytoplasmic staining by IHC)* 1012 patients randomised (1:1) Capivasertib 400 mg BID 4 days on, 3 days off + Abiraterone/pred + ADT 1000 mg QD + ADT Placebo BID 4 days on, 3 days off + Abiraterone/pred + ADT 1000 mg QD + ADT Stratification factors:† • High-volume disease with/without visceral metastases; low-volume disease • Geography Primary endpoint • Investigator-assessed rPFS Secondary endpoints analysed at ad hoc DCO • Overall survival • Symptomatic skeletal event-free survival • Time to castration resistance • Time to PSA progression • Safety Exploratory post hoc PTEN deficiency subgroups NCT04493853. Full eligibility criteria available in Fizazi et al.¹ *Determined using investigational antibody for PTEN (SP218) (Roche Diagnostics). †High-volume disease (per CHAARTED criteria) with visceral metastases, high-volume disease without visceral metastases, low-volume disease; North America, Western Europe and Australia; Latin America and Eastern Europe; Asia. ADT, androgen deprivation therapy; BID, twice daily; DCO, data cut-off; IHC, immunohistochemistry; mHSPC, metastatic hormone-sensitive prostate cancer; pred, prednisone/prednisolone; PSA, prostate specific antigen; PTEN, phosphatase and tensin homolog; QD, once daily; rPFS, radiographic progression-free survival 1. Fizazi K et al. Ann Oncol 2026;37:53–68 Karim Fizazi MD, PhD [Side panel: presenter video] Karim Fizazi — Capivasertib (capi) + abiraterone (abi) vs placebo (pbo) + abi in patients (pts) with PTEN-deficient metastatic hormone-sensitive prostate cancer (mHSPC): Updated overall survival (OS) from CAPItello-281 BILBAO AUDITORIUM - HALL 3 [Slide 2] 13:30 - 15:00 Rapid oral: Genitourinary tumours, prostate, penile and testis CHAIRS: THOMAS HWANG, EDMOND KWAN, ALISON REID CAPItello-281 updated DCO: OS Clinically meaningful improvement in OS for capi + abi Capi + abi (N=507) | Pbo + abi (N=505) Events, n (%): 195 (38.5) | 226 (44.8) Median, months (95% CI): 57.4 (47.0, NC) | 46.5 (40.7, 53.2) HR (95% CI): 0.82 (0.68, 0.99) P-value: 0.044* *Descriptive OS analysis: alpha controlled, did not meet statistical significance Legend: Capivasertib + abiraterone; Placebo + abiraterone Y-axis: Probability of OS; X-axis: Time from randomisation (months) 0-66 Number of patients at risk Capi + abi: 507 486 473 454 441 415 398 387 366 338 291 256 215 176 149 118 91 60 45 25 10 2 0 Pbo + abi: 505 493 478 457 438 409 381 359 344 313 263 233 195 158 126 106 81 58 42 25 13 5 0 Data cut-off: 9 Mar 2026. Tick marks represent censored observations. The analysis was carried out using a stratified Cox proportional hazards model adjusting for the variables selected in the pooling strategy: volume of disease/visceral metastases. HR (95% CI) for OS at the primary rPFS analysis was 0.90 (0.71, 1.15). Predefined OS subgroup analyses are available in the article. abi, abiraterone; capi, capivasertib; CI, confidence interval; DCO, data cut-off; HR, hazard ratio; NC, not calculable; OS, overall survival; pbo, placebo; rPFS, [illegible] Karim Fizazi MD, PhD [Side panel: presenter video] Karim Fizazi — Capivasertib (capi) + abiraterone (abi) vs placebo (pbo) + abi in patients (pts) with PTEN-deficient metastatic hormone-sensitive prostate cancer (mHSPC): Updated overall survival (OS) from CAPItello-281 BILBAO AUDITORIUM - HALL 3 Survival table Endpoint: Overall survival (CAPItello-281 updated DCO, 9 Mar 2026) Arm | N | Median | Median CI Capi + abi | 507 | 57.4 mo | 47.0, NC Pbo + abi | 505 | 46.5 mo | 40.7, 53.2 HR: 0.82 HR CI: 0.68-0.99 P: 0.044 (descriptive; alpha controlled, did not meet statistical significance) [Slide 3] 13:30 - 15:00 Rapid oral: Genitourinary tumours, prostate, penile and testis CHAIRS: THOMAS HWANG, EDMOND KWAN, ALISON REID CAPItello-281 updated DCO: Conclusions • In CAPItello-281, ~25% of tested patients had PTEN-deficient tumours • A clinically meaningful improvement in OS (42% maturity) was observed for capi + abi vs pbo + abi (HR 0.82, 95% CI 0.68, 0.99; median OS: 57.4 vs 46.5 months) • rPFS and SSE-FS outcomes were consistent with the primary analysis and continued to favour capi + abi • The safety profile remained consistent with the primary analysis – Most common AEs were on-target diarrhoea, rash and hyperglycaemia These updated results support capivasertib plus abiraterone as a first-in-class treatment for this PTEN-deficient mHSPC population (high- and low-volume disease) with poor prognosis Data cut-off: 9 Mar 2026 abi, abiraterone; AE, adverse event; capi, capivasertib; CI, confidence interval; DCO, data cut-off; HR, hazard ratio; mHSPC, metastatic hormone-sensitive prostate cancer; OS, overall survival; pbo, placebo; PTEN, phosphatase and tensin homolog; rPFS, radiographic progression-free survival Karim Fizazi MD, PhD [Side panel: presenter video] Karim Fizazi — Capivasertib (capi) + abiraterone (abi) vs placebo (pbo) + abi in patients (pts) with PTEN-deficient metastatic hormone-sensitive prostate cancer (mHSPC): Updated overall survival (OS) from CAPItello-281 BILBAO AUDITORIUM - HALL 3 [Slide 4] 13:30 - 15:00 Rapid oral: Genitourinary tumours, prostate, penile and testis CHAIRS: THOMAS HWANG, EDMOND KWAN, ALISON REID CAPItello-281 updated DCO: Adverse events (≥10% of patients) With 17 months additional follow-up, the safety profile of capi + abi was consistent with the primary analysis¹ Total (%)/Grade ≥3 (%) — Capi + abi (N=503) [Grade 1 | Grade 2 | ≥Grade 3] vs Pbo + abi (N=503) [≥Grade 3 | Grade 2 | Grade 1] Capi + abi values (listed top to bottom): 52.7/6.4; 40.4/11.3; 36.4/12.7; 27.0/5.6; 24.3/10.1; 24.1/7.6; 16.3/0.4; 15.3/4.6; 15.1/4.4; 13.5/2.8; 13.1/0.8; 12.5/1.0; 12.1/0.8; 11.7/0.6; 11.3/1.2; 11.1/0.2; 11.1/0.2; 10.7/0.6; 10.1/0.4; 10.1/1.0; 9.9/0.4 Pbo + abi values (listed top to bottom): 9.5/0.4; 13.9/1.0; 8.0/0.2; 14.9/1.6; 15.7/6.2; 26.4/9.5; 14.3/0.8; 11.9/1.6; 14.1/3.6; 12.9/2.2; 5.4/0.2; 5.8/0.4; 9.3/0.6; 13.1/0.6; 3.2/0; 13.9/0; 2.6/0; 4.0/0; 11.1/0.4; 4.2/0; 13.9/0 AE rows (listed top to bottom): Diarrhoea; Hyperglycaemia*; Rash†; Anaemia; Hypokalaemia; Hypertension; Fatigue; Urinary tract infection; ALT increased; AST increased; Nausea; Asthenia; COVID-19; Back pain; Pyrexia; Hot flush; Pruritus; Blood creatine increase; Arthralgia; Decreased appetite; Constipation [Butterfly bar chart, x-axis 100-0-100 %] Data cut-off: 9 Mar 2026. DKA was reported in six patients (1.2%) in the capi + abi arm, and one patient (0.2%) in the pbo + abi arm; the pbo + abi arm case was euglycaemic DKA. No new cases of DKA were reported in the capi + abi arm since the primary analysis. *Grouped term (includes the preferred terms of blood glucose increased, hyperglycaemia). †Grouped term (includes the preferred terms of erythema, rash, rash erythematous, rash macular, rash maculo-papular, rash papular, rash pruritic). abi, abiraterone; AE, adverse event; ALT, alanine aminotransferase; AST, aspartate aminotransferase; capi, capivasertib; DKA, diabetic ketoacidosis; pbo, placebo. 1. Fizazi K et al. Ann Oncol 2026;37:53–68. Karim Fizazi MD, PhD [Side panel: presenter video] Karim Fizazi — Capivasertib (capi) + abiraterone (abi) vs placebo (pbo) + abi in patients (pts) with PTEN-deficient metastatic hormone-sensitive prostate cancer (mHSPC): Updated overall survival (OS) from CAPItello-281 BILBAO AUDITORIUM - HALL 3

CAPItello-281 updated^a descriptive analyses at 42% OS data maturity A [Kaplan-Meier plot; y-axis: Probability of OS 0.0-1.0; x-axis: Time from randomization (months) 0-66; dashed line at 0.5] Capivasertib + abiraterone; Placebo + abiraterone Callout: Median OS was reached for the first time in both arms, with ~11 months difference between arms | Capi + abi (n = 507) | Pbo + abi (n = 505) Events, n (%): 195 (38.5) | 226 (44.8) Median OS, months (95% CI): 57.4 (47.0-NC) | 46.5 (40.7-53.2) HR (95% CI): 0.82 (0.68-0.99) P-value: 0.044 Callout: Outcomes in the control arm remained similar to the primary analysis and highlight the poor prognosis of patients with PTEN-deficient mHSPC on standard-of-care treatment Number of patients at risk Capi + abi: 507 486 473 454 441 415 398 387 366 338 291 256 215 176 149 118 91 60 45 25 10 2 0 Pbo + abi: 505 493 478 457 438 409 381 359 344 313 263 233 195 158 126 106 81 58 42 25 13 5 0 With an additional 17 months' follow-up, findings from this updated descriptive analysis support the clinically meaningful OS benefit of adding capivasertib to abiraterone in the treatment of patients with PTEN-deficient mHSPC ^aData cut-off: March 9, 2026. Abi, abiraterone; Capi, capivasertib; CI, confidence interval; HR, hazard ratio; mHSPC, metastatic hormone sensitive prostate cancer; NC, not calculable; OS, overall survival; Pbo, placebo; PTEN, phosphatase and tensin homolog Survival table Endpoint: Overall survival (updated descriptive analysis, 42% maturity) Arm | N | Median | Median CI Capivasertib + abiraterone | 507 | 57.4 months | 47.0-NC Placebo + abiraterone | 505 | 46.5 months | 40.7-53.2 HR: 0.82 HR CI: 0.68-0.99 P: 0.044

[Slide 1] PTEN-deficient cell AR signalling | PI3K/AKT signalling T RTK Cell membrane Cytoplasm Abiraterone + ADT PI3Kβ PI3Kα PTEN T / AR AKT Capivasertib Nucleus T / AR Tumour proliferation | Tumour proliferation [Slide 2] Primary endpoint: investigator-assessed rPFS (primary rPFS DCO: 7 Oct 2024) Capi + abi (N=507) | Pbo + abi (N=505) Events, n (%): 183 (36.1) | 215 (42.6) Median rPFS, months (95% CI): 33.2 (25.8, 44.2) | 25.7 (22.0, 29.9) HR (95% CI): 0.81 (0.66, 0.98) P-value: 0.034 Annotation: 25.7 months — 7.5 months — 33.2 months Legend: Capivasertib + abiraterone; Placebo + abiraterone Y axis: Probability of rPFS; X axis: Time from randomisation (months) Number of patients at risk (0,3,...,45 months) Capi + abi: 507 460 435 353 282 233 217 165 123 93 69 62 41 21 6 0 Pbo + abi: 505 479 440 359 276 215 198 154 113 83 59 51 37 23 8 0 Survival table Endpoint: Investigator-assessed rPFS (primary DCO 7 Oct 2024) Arm | N | Median | Median CI Capivasertib + abiraterone | 507 | 33.2 months | 25.8, 44.2 Placebo + abiraterone | 505 | 25.7 months | 22.0, 29.9 HR: 0.81 HR CI: 0.66, 0.98 P: 0.034

[Slide 1] Worse with talazoparib Global QoL 3.7 Physical functioning 2.5 Fatigue 3.2 Appetite loss 5.2 0 | 5 points | 8 Worse with talazoparib -> [Slide 2] IPATential150 (Phase III) | Ipatasertib + abi (mCRPC, PTEN loss) | rPFS 0.77 (0.61-0.98) | OS 0.94 (0.76-1.17), final ProCAID (Phase II) | Capi + docetaxel (mCRPC, unselected) | cPFS 0.92 (80% CI 0.73-1.16) | OS 0.70 (0.47-1.05) CAPItello-280 (Phase III) | Capi + docetaxel (mCRPC, chemo-naive) | Stopped for futility, Apr 2025 | Unlikely to meet OS or rPFS CAPItello-281 (Phase III) | Capi + abi (mHSPC, PTEN-deficient) | rPFS 0.81 (0.66-0.98) | OS 0.82 (0.68-0.99), ad hoc Legend: progression-free survival; overall survival; ad hoc, not significant Axis: 0.5 | 0.75 | 1.0 | 1.25 | Hazard ratio (95% CI)

[Slide 2] CAPItello-281 updated DCO: OS Clinically meaningful improvement in OS for capi + abi Capi + abi (N=507) | Pbo + abi (N=505) Events, n (%) 195 (38.5) | 226 (44.8) Median, months (95% CI) 57.4 (47.0, NC) | 46.5 (40.7, 53.2) HR (95% CI) 0.82 (0.68, 0.99) P-value 0.044* *Descriptive OS analysis: alpha controlled, did not meet statistical significance. [KM plot: Probability of OS vs Time from randomisation (months) 0-66] Capivasertib + abiraterone / Placebo + abiraterone Number of patients at risk (0-66 months, 3-mo steps) Capi + abi 507 486 473 454 441 415 398 387 366 338 291 256 215 176 149 118 91 60 45 25 10 2 0 Pbo + abi 505 493 478 457 438 409 381 359 344 313 263 233 195 158 126 106 81 58 42 25 13 5 0 Data cut-off: 9 Mar 2026. Tick marks represent censored observations. The analysis was carried out using a stratified Cox proportional hazards model adjusting for the variables selected in the pooling strategy: volume of disease/visceral metastases. HR (95% CI) for OS at the primary rPFS analysis was 0.90 (0.71, 1.15). Predefined OS subgroup analyses are available in the article. abi, abiraterone; capi, capivasertib; CI, confidence interval; DCO, data cut-off; HR, hazard ratio; NC, not calculable; OS, overall survival; pbo, placebo; rPFS, radiographic progression-free survival. Karim Fizazi MD, PhD Survival table Endpoint: Overall survival (updated DCO) Arm | N | Median | Median CI Capi + abi | 507 | 57.4 mo | 47.0, NC Pbo + abi | 505 | 46.5 mo | 40.7, 53.2 HR: 0.82 HR CI: 0.68, 0.99 P: 0.044 (descriptive; alpha controlled, did not meet statistical significance) [Slide 3] CAPItello-281 updated DCO: OS in post hoc PTEN deficiency cut-off subgroups HRs for OS at increasing PTEN deficiency cut-offs ranged from 0.76-0.74 [KM plot: Probability of OS vs Time from randomisation (months) 0-66] Cut-off: ≥90% of cells without PTEN staining (all patients) Capivasertib + abiraterone / Placebo + abiraterone PTEN cut-off | Patients, n Capi + abi | Pbo + abi | Median OS, months Capi + abi | Pbo + abi | HR (95% CI) All patients (≥90%) | 507 | 505 | 57.4 | 46.5 | 0.82 (0.68, 0.99) ≥95% | 404 | 410 | 57.4 | 41.1 | 0.76 (0.61, 0.93) ≥99% | 205 | 196 | 57.7 | 41.1 | 0.75 (0.56, 1.02) 100% | 169 | 162 | 50.2 | 40.9 | 0.74 (0.53, 1.02) [Forest plot, HR axis 0.5, 1.0, 2.0] Favours capi + abi <- HR (95% CI) -> Favours pbo + abi 95% CI: all patients (≥90% PTEN cut-off) Data cut-off: 9 Mar 2026. abi, abiraterone; capi, capivasertib; CI, confidence interval; DCO, data cut-off; HR, hazard ratio; OS, overall survival; pbo, placebo; PTEN, phosphatase and tensin homolog. Karim Fizazi MD, PhD Survival table Endpoint: Overall survival by post hoc PTEN deficiency cut-off (updated DCO) All patients (≥90%) Capi + abi | 507 | 57.4 mo All patients (≥90%) Pbo + abi | 505 | 46.5 mo ≥95% Capi + abi | 404 | 57.4 mo ≥95% Pbo + abi | 410 | 41.1 mo ≥99% Capi + abi | 205 | 57.7 mo ≥99% Pbo + abi | 196 | 41.1 mo 100% Capi + abi | 169 | 50.2 mo 100% Pbo + abi | 162 | 40.9 mo HR: ≥90%: 0.82; ≥95%: 0.76; ≥99%: 0.75; 100%: 0.74 HR CI: ≥90%: 0.68, 0.99; ≥95%: 0.61, 0.93; ≥99%: 0.56, 1.02; 100%: 0.53, 1.02 [Slide 4] CAPItello-281 updated DCO: Symptomatic skeletal event-free survival Outcomes remained consistent with the primary analysis and continued to favour capi + abi Capi + abi (N=507) | Pbo + abi (N=505) Events, n (%) 223 (44.0) | 255 (50.5) Median, months (95% CI) 43.8 (39.1, 53.1) | 35.9 (32.1, 42.4) HR (95% CI) 0.83 (0.69, 0.99) Nominal P-value 0.044 [KM plot: Probability of SSE-FS vs Time from randomisation (months) 0-66] 35.9 months -> 7.9 months -> 43.8 months Capivasertib + abiraterone / Placebo + abiraterone Number of patients at risk (0-66 months, 3-mo steps) Capi + abi 507 461 444 423 400 373 352 334 310 285 250 218 171 148 119 94 67 44 36 19 6 2 0 Pbo + abi 505 481 457 429 399 366 338 320 295 258 223 179 143 126 93 80 56 41 30 18 6 2 0 Data cut-off: 9 Mar 2026. SSE-FS is a composite endpoint of pathological fracture, spinal cord compression, use of radiation, surgical intervention, and death. Tick marks represent censored observations. The analysis was carried out using a stratified Cox proportional hazards model adjusting for the variables selected in the pooling strategy: volume of disease/visceral metastases and geographic location. abi, abiraterone; capi, capivasertib; CI, confidence interval; DCO, data cut-off; HR, hazard ratio; pbo, placebo; SSE-FS, symptomatic skeletal event-free survival. Karim Fizazi MD, PhD Survival table Endpoint: Symptomatic skeletal event-free survival (SSE-FS, updated DCO) Arm | N | Median | Median CI Capi + abi | 507 | 43.8 mo | 39.1, 53.1 Pbo + abi | 505 | 35.9 mo | 32.1, 42.4 HR: 0.83 HR CI: 0.69, 0.99 P: 0.044 (nominal)

[Slide 1] CARE1: a pragmatic biomarker-driven phase 3 trial NCT06364631 Metastatic renal cell carcinoma with a clear-cell component Intermediate or poor risk IMDC Naive from systemic therapy Local PD-L1 TPS ≥ 1% or < 1% PD-L1 positive -> R 1:1 -> DUAL ICI: NIVOLUMAB + IPILIMUMAB / ICI + TKI: PEMBROLIZUMAB + AXITINIB, PEMBROLIZUMAB + LENVATINIB, NIVOLUMAB + CABOZANTINIB -> Overall survival; Hypothesis dual ICI > ICI + TKI PD-L1 negative -> R 1:1 -> DUAL ICI: NIVOLUMAB + IPILIMUMAB / ICI + TKI: PEMBROLIZUMAB + AXITINIB, PEMBROLIZUMAB + LENVATINIB, NIVOLUMAB + CABOZANTINIB -> Progression-free survival, Overall survival; Hypothesis ICI + TKI > dual ICI Ronan Flippot MD PhD CARE-1 has received funding from the European Union's Horizon Europe research and innovation programme under grant agreement No. 101104801. [Slide 2] Local PD-L1 reproducibility N=374 | Local PD-L1 PD-L1 - | Local PD-L1 PD-L1 + Central PD-L1 PD-L1 - | 191 (94%) | 23 (14%) Central PD-L1 PD-L1 + | 13 (6%) | 147 (86%) Concordance rate 90% Minor discordance rate in PD-L1+ patients 26% Characteristic | Concordance (%) IMDC risk: Intermediate | 90 Poor | 90 ISUP grade: 2 | 89 3 | 90 4 | 93 Sarcomatoid / rhabdoid | 91 PD-L1 tissue assessment: Biopsy | 91 Metastasis surgery | 89 Nephrectomy | 90 Local PD-L1 antibody: 22C3 (Dako) | 90 SP263 (Ventana) | 94 QR1 (Diagomics) | 90 Ronan Flippot MD PhD CARE-1 has received funding from the European Union's Horizon Europe research and innovation programme under grant agreement No. 101104801.

[Slide 1] MADRID 2026 ESMO congress CARE1 PRAGMATIC CLINICAL TRIAL Feasibility and reproducibility of PD-L1 assessment within CARE1: a phase III, randomized, biomarker-driven pragmatic trial to optimize first line treatment in patients with metastatic clear-cell renal cell carcinoma Ronan Flippot, Nathalie Rioux-Leclercq, Stéphanie Foulon, Bohuslav Melichar, Sofie Wilgenhof, Manuela Schmidinger, Elena Verzoni, Katriina Jalkanen, Viktor Gruenwald, Cristina Suarez, Guillermo de Velasco, Cristiane Bergerot, Rose Woodward, Pierre Moricheau-Beauchant, Stéphane Sockeel, Maia Claveau, Malek Kammoun, Bernard Escudier, Thomas Powles, Laurence Albigès October 25th, 2026 [Slide 4] Conclusions PD-L1 is evaluated as a pragmatic biomarker in CARE1, the 1st phase 3 trial comparing dual ICI and ICI + TKI in patients with mRCC Local PD-L1 assessment within the CARE1 trial was highly reproducible regardless of disease features or antibodies PD-L1 expression was associated with adverse clinical features including higher grade, presence of sarcomatoid or rhabdoid features, and necrosis PD-L1 staining was associated with distinct vascular and immune phenotypes which could portend distinct sensitivity to dual ICI or ICI + TKI Local PD-L1 has potential to inform patient selection for immunotherapy combinations Ronan Flippot MD PhD CARE-1 has received funding from the European Union's Horizon Europe research and innovation programme under grant agreement No. 101104801.

[Slide 3] PD-L1 positivity and baseline characteristics Characteristic | PD-L1- N=214 (%) | PD-L1+ N=160 (%) | Overall N=374 (%) Female | 48 (22%) | 43 (27%) | 91 (24%) Age (IQR) | 68.0 [60.0;74.0] | 68.5 [60.0;74.0] | 68.0 [60.0;74.0] Prior nephrectomy | 83 (40%) | 50 (32%) | 133 (37%) IMDC risk group Intermediate | 138 (64%) | 93 (58%) | 231 (62%) Poor | 76 (36%) | 67 (42%) | 143 (38%) ISUP grade 2 | 83 (39%) | 19 (12%) | 102 (27%) 3 | 67 (31%) | 48 (30%) | 115 (31%) 4 | 56 (26%) | 86 (54%) | 142 (38%) Sarcomatoid / rhabdoid | 36 (18%) | 58 (37%) | 94 (26%) Necrosis | 65 (31%) | 90 (56%) | 155 (42%) Ronan Flippot MD PhD CARE-1 has received funding from the European Union's Horizon Europe research and innovation programme under grant agreement No. 101104801. [Slide 4] PD-L1 positivity and disease tissue phenotypes Microvascular Density (MVD): PD-L1- N=115 | PD-L1+ N=100 T cell (CD3) infiltrate Tumour center (Proportion): PD-L1- N=184 | PD-L1+ N=138 T cell (CD3) infiltrate Invasive Margin*: PD-L1- N=85 | PD-L1+ N=44 Legend: Abundant | Abundant with tertiary lymphoid structures | None / Scarce * Nephrectomy specimens only Ronan Flippot MD PhD

[Slide 2] Local PD-L1 assessment 1 Sections 4-5 µm HE Staining No necrosis 100 tumour cells 2 PDL1 IHC Antibodies If available Dako 22C3 If not Routine Ab % of +PDL1 tumor cells [Slide 3] Local PD-L1 reproducibility N=374 Local PD-L1 PD-L1 - | PD-L1 + Central PD-L1 PD-L1 - | 191 (94%) | 23 (14%) PD-L1 + | 13 (6%) | 147 (86%) Concordance rate 90% Minor discordance rate in PD-L1+ patients 26% Characteristic | Concordance (%) IMDC risk Intermediate | 90 Poor | 90 ISUP grade 2 | 89 3 | 90 4 | 93 Sarcomatoid / rhabdoid | 91 PD-L1 tissue assessment Biopsy | 91 Metastasis surgery | 89 Nephrectomy | 90 Local PD-L1 antibody 22C3 (Dako) | 90 SP263 (Ventana) | 94 QR1 (Diagomics) | 90

PFS by TROP2 and PD-L1 Left panel: Median (95% CI), mo TROP2 H-score <median | 11.2 (3.7–NR) TROP2 H-score ≥median | 11.9 (5.6–21.1) PFS, % 0–100; Time, mo 0 3 6 9 12 15 18 21 24 27 30 No. at risk TROP2 H-score <median: 18 16 11 9 7 7 7 6 3 1 0 TROP2 H-score ≥median: 19 17 12 10 9 8 6 4 0 0 0 Right panel: Median (95% CI), mo PD-L1 CPS <10 | 15.6 (7.2–NR) PD-L1 CPS ≥10 | 5.7 (0.5–NR) No. at risk PD-L1 CPS <10: 32 30 22 18 15 15 13 9 4 1 0 PD-L1 CPS ≥10: 8 5 3 3 3 2 2 2 0 0 0 • In the overall cohort, median PFS was 11.2 (95% CI, 5.8–NR) months and median OS was NR (95% CI, 17.2 mo to NR) Median study follow-up duration, 22.8 (range, 19.8–30.0) months. Median PFS and OS are based on the Kaplan-Meier method for censored data. Data cutoff, Jan 26, 2026. Survival table Endpoint: PFS by TROP2 H-score (left panel) Arm | N | Median | Median CI TROP2 H-score <median | 18 | 11.2 mo | 3.7–NR TROP2 H-score ≥median | 19 | 11.9 mo | 5.6–21.1

Upcoming bladder preservation trials After cCR: can we omit local therapy? EV-209 Phase 2 Cystectomy-eligible EV + pembro -> cCR? -> cCR → continue EV+P, surveillance / No cCR → local therapy Primary endpoint cCR rate 2-yr BI-EFS FIDELIO Phase 2 cCR after EV + pembro EV + pembro -> cCR -> R -> No consolidation / Radiotherapy Primary endpoint to be confirmed Can systemic therapy replace or strengthen CRT? EV-309 Phase 3, randomised Cystectomy-ineligible or declining R -> EV + pembro (EV ≤ 9 cycles, pembro ≤ 17) / Chemoradiotherapy Primary endpoint BI-EFS OS TroFuse-055 Phase 3, randomised Elect bladder preservation R -> Sac-TMT + pembro induction -> CRT + pembro consolidation / Chemoradiotherapy Primary endpoint EFS OS

[Header row] ENERGYZE (n=646) Nivolumab | KN866(n=907) Pembrolizumab | NIAGARA (n=1063) Durvalumab | B15/EV304 (n=808) EV + pembro Study arm | GCN | GC | GCP | GC | GCD | GC | EVP | GC Medan FU | 54 months | 61 months | 46 months | 33 months pCR/ctDNA clearance | 39% 82% | 29% 58% | 44% | 36% | 37% | 27% | 56% | 33% 2 yr EFS | 72% | 72% | 76% | 70% | 68% | 60% | 79% | 66% 2 yr OS | 85% | 85% | 83% | 82% | 74% | 68% | 87% | 81% EFS HR | 0.94 | 0.88 | 0.68 | 0.53 OS HR | 0.88 | 1.00 | 0.75 | 0.6, Competed IO | 56% | 46% | 54% | 51% surgery | 90% | 88% | 91% | 87% | 88% | 85% | 87% | 91% T2 at | 77% | 15% | 40% | 20%

[Slide 1] 13:30 - 15:00 Proffered paper 1: Genitourinary tumour, renal and urothelial CHAIRS: AXEL BEX, ANJA LORCH Is the combination hypothesis incorrect? pCR outcomes were consistent across trials KEYNOTE-866 [Bar chart; y-axis: pCR Rate, % (95% CI)] Pembro + cis + gem vs Placebo + cis + gem Estimated difference:b 8.9% (95% CI, 2.5-15.1)* CA017-078 pCR overall (co-primary endpoint)a 9.5% (95% CI 2.3-16.7) (99.91% CI −2.6 to 21.6)b [Bars; y-axis: Patients (%)] 39.1 | 29.6 Overall ITT NIAGARA Re-analysis (Apr 2024) Odds ratio 1.60 (95% CI, 1.23-2.08) nominal P=0.0005 Durvalumab arm N=533: 37.3% 199/533 95% CI, 33.2-41.6 Comparator arm N=530: 27.5% 146/530 95% CI, 23.8-31.6 [y-axis: pCR rate (%)] Robert Jones Invited Discussant 47350, 47360 and 47420 BARCELONA AUDITORIUM - HALL 9 [Slide 2] 13:30 - 15:00 Proffered paper 1: Genitourinary tumour, renal and urothelial CHAIRS: AXEL BEX, ANJA LORCH Were populations different? EFS outcomes were good in control arms of these studies [Thumbnail 1] Primary End Point: EFS by Investigator Assessment, ITT Population — Pembro + cis + gem n = 453 vs Placebo + cis + gem n = 454; Events, n (%) 157 (34.7) vs 168 (37.0); Median (95% CI), mo 74.1 (72.4-NR) vs NR (67.7-NR); HR 0.88 (95% CI, 0.71-1.10), 1-sided P = 0.1366*; landmark labels 76.7%/70.0%, 69.2%/66.3%, 66.0%/62.3%, 63.5%/61.9%; No. at risk Pembro 453 401 366 338 329 304 296 273 224 175 114 51 11 0; Placebo 454 400 345 310 300 291 283 257 208 168 106 51 12 0 [Thumbnail 2] Event-free survival by treatment arm (co-primary endpoint) — NIVO+GC (N = 325) vs GC (N = 328); events 104/325 (32.0) vs 104/328 (31.7); median NR vs 67.6 (63.0-NE); HR (95% CI) 0.94 (0.72-1.24); HR (95.74% CI) 0.94 (0.71-1.25); P 0.6697; landmark labels 82.6%/78.6%, 72.3%/71.6%, 68.1%/68.5%, 67.3%/65.8%; No. at risk NIVO+GC 325 283 243 227 209 197 189 146 100 60 25 5 0 0; GC 328 271 219 199 187 176 168 131 90 61 23 4 1 0 KEYNOTE-866 | CA017-078 | NIAGARA T2 (%) 15 | 76 | 40 N1 (%) 0.7 | Excluded | 5 ECOG 0 (%) 76 | - | 78 24 month EFS in control arm (%) 70.0 | 71.6 | 59.8 Robert Jones Invited Discussant 47350, 47360 and 47420 BARCELONA AUDITORIUM - HALL 9 [Slide 3] 13:30 - 15:00 Proffered paper 1: Genitourinary tumour, renal and urothelial CHAIRS: AXEL BEX, ANJA LORCH How I explain these two studies • More similarities than differences with NIAGARA • Patient mix had better than expected prognosis - this makes it harder to demonstrate modest improvement • Small imbalances, unseen at baseline, may explain different findings between the trials Robert Jones Invited Discussant 47350, 47360 and 47420 BARCELONA AUDITORIUM - HALL 9 [Slide 4] 13:30 - 15:00 Proffered paper 1: Genitourinary tumour, renal and urothelial CHAIRS: AXEL BEX, ANJA LORCH Were studies designed differently? Callouts: Only KN-866 used placebo | Differences in cisplatin eligibility | Only CA017-078 excluded N1 | Subtle differences in primary endpoint definitions | Some differences in statistical design [Panel] Phase 3 KEYNOTE-866 Study Design (NCT03924856) — Adults with MIBC; clinical stage T2-T4aN0M0 or T1-T4aN1M0; urothelial histology ≥50%; eligible for RC + PLND; did not meet Galsky criteria for cisplatin ineligibility incl. GFR ≥60 mL/min; ECOG PS 0 or 1. R 1:1: n = 453 Pembrolizumab 200 mg + cisplatin 70 mg/m2 + gemcitabine 1000 mg/m2 x4 → RC + PLND → Pembrolizumab 200 mg Q3W 13 cycles; n = 454 Placebo + cisplatin + gemcitabine → Placebo Q3W 13 cycles. Primary end point: EFS by investigator assessment. Key secondary: OS and pCR (pT0N0) by central pathologist review. Other secondary: Safety. [Panel] Study design: CA017-078 — Phase 3 randomized trial of neoadjuvant NIVO plus GC, followed by adjuvant NIVO, vs neoadjuvant GC alone for previously untreated MIBC (NCT03661320). Key inclusion: MIBC (predominant urothelial component); cT2-T4a, N0, M0; GFR ≥ 50 mL/min; ECOG PS 0-1. R 1:1: GC Q3W x 4 cycles + NIVO 360 mg Q3W x 4 cycles (N = 325) → radical cystectomy → NIVO 480 mg Q4W x 9 cycles; GC Q3W x 4 cycles (N = 328) → radical cystectomy → no further treatment. [Greyed arm: GC + NIVO 360 mg + BMS986205 (blinded) 100 mg QD → NIVO 480 mg Q4W + BMS986205 — TERMINATED (Dec 2021), excluded from formal testing of NIVO+GC vs GC alone.] Stratification: clinical stage (T2 vs T3/T4a); GFR (50 to <60 mL/min vs ≥60 mL/min); PD-L1 status (≥1% vs <1%/indeterminate/NE) by central review. Co-primary endpoints: pCR and EFS per BICR. Secondary: OS, safety. Exploratory: ctDNA analyses. Median follow-up (range): 54.5 months (39.4-77.0). [Panel] NIAGARA: Study Design — Adults; cisplatin-eligible MIBC (cT2-T4aN0/1M0); UC or UC with divergent differentiation or histologic subtypes; evaluated and confirmed for RC; CrCl of ≥40 mL/min. R 1:1: Durvalumab arm N=533: neoadjuvant Durvalumab 1500 mg IV Q3W + Gemcitabine + cisplatin (4 cycles) → radical cystectomy → adjuvant Durvalumab 1500 mg IV Q4W (8 cycles); Comparator arm N=530: Gemcitabine + cisplatin → radical cystectomy → no treatment. Dual primary endpoints: EFS, pCR. Key secondary endpoint: OS. Safety. EFS was defined as: progressive disease that precluded RC; recurrence after RC; date of expected surgery in patients who did not undergo RC; death from any cause. Placebo, primary endpoints Robert Jones Invited Discussant 47350, 47360 and 47420 BARCELONA AUDITORIUM - HALL 9

[Slide 1] Practical Decision Framework* – High-volume1 mHSPC (fit patients) | Synchronous | Metachronous | Trials ADT | Yes | Yes | + Abiraterone/prednisone2 + Enzalutamide + Apalutamide + Darolutamide | One of the options | One of the options | LATITUDE (only synchronous) STAMPEDE, TITAN, ARCHES, ENZAMET, ARANOTE + Docetaxel + abiraterone/prednisone3 + Docetaxel + darolutamide | If fit for chemotherapy | Selected patients, fit for chemotherapy | PEACE-1 (only synchronous) ARASENS + Docetaxel | As sole addition to ADT generally not recommended | GETUG-AFU15, CHAARTED, STAMPEDE + Radiation primary tumour4 | Selected cases, goal: Reduction of local complications | Primary tumour already treated, local therapy only for complications | HORRAD, STAMPEDE, PEACE-1 | Synchronous, biomarker selected | Metachronous, biomarker selected | + Docetaxel + abiraterone followed by abiraterone + niraparib5 + Enzalutamide + Talazoparib3 | HRR: Yes, if BRCA1/2 In the future possibly also other HRR | HRR: Yes, if BRCA1/2 In the future possibly also other HRR | AMPLITUDE TALAPRO-3 + Abiraterone + Capivasertib3,6 | PTEN deficiency: Selected cases in the future, otherwise docetaxel combination | Not included in CAPItello-281 | CAPItello-281 (only synchronous) + ARPI + 177Lu-PSMA-6173 | PSMA: Selected cases in the future | PSMA: Selected cases in the future | PSMAddition [Slide 2] Practical Decision Framework* – Low-volume1 mHSPC (fit patients) | Synchronous | Metachronous | Trials ADT | Yes | Yes, possibly limited in time or MDT alone in selected cases | + Abiraterone/Prednisone2 + Enzalutamide + Apalutamide + Darolutamide | One of the options | Yes (limited duration) | LATITUDE (only synchronous), STAMPEDE, TITAN, ARCHES, ENZAMET, ARANOTE + Docetaxel + abiraterone/Prednisone3 + Docetaxel + darolutamide | Very selected cases | No | PEACE-1 (only synchronous), ARASENS + Docetaxel | As sole addition to ADT generally not recommended | GETUG-AFU15, CHAARTED, STAMPEDE + Radiation primary tumour4 | Yes | Primary tumour already treated, local therapy only for complications | HORRAD, STAMPEDE, PEACE-1 Metastasis-directed therapy (MDT) | Selected cases with the aim of limiting the duration of systemic therapy? | Yes, in combination with systemic therapy (limited duration) or without systemic therapy | Synchronous: EXTEND Metachronous: STOMP, ORIOLE, SABR-COMET, POPSTAR, RADIOSA | Synchronous, biomarker selected | Metachronous, biomarker selected | + Abiraterone + Niraparib5 + Enzalutamide + Talazoparib3 | HRR: selected cases, if BRCA1/2 In the future possibly also other HRR | Generally, no | AMPLITUDE TALAPRO-3 + Abiraterone + Capivasertib3,6 | PTEN deficiency: Very selected cases in the future, otherwise docetaxel combination | Not included in CAPItello-281 | CAPItello-281 (only synchronous) + ARPI + 177Lu-PSMA-6173 | PSMA: Selected cases in the future | No | PSMAddition

LBA2 - Perioperative durvalumab with or without tremelimumab plus neoadjuvant enfortumab vedotin in muscle-invasive bladder cancer: Results from the phase 3 VOLGA trial Speakers Thomas B. Powles (London, United Kingdom) Lecture Time 16:54 - 17:06 Abstract Background The standard of care continues to evolve for patients (pts) with muscle-invasive bladder cancer (MIBC). We evaluated if neoadjuvant durvalumab (D) ± tremelimumab (T) plus enfortumab vedotin (EV) followed by radical cystectomy (RC) and adjuvant D±T would be an effective option. Methods VOLGA (NCT04960709), a phase 3, randomised, open-label trial, enrolled pts with MIBC (cT2-T4aN0/1M0 or cT1N1M0) ineligible for or who refused treatment (tx) with cisplatin and were planned for RC. Pts were randomised 1:1:1 to D+EV, D+T+EV, or control. D+EV and D+T+EV arms: perioperative D±T (neoadjuvant: 3 cycles of D, 2 cycles of T; adjuvant: 9 cycles of D, 1 cycle of T) and 3 cycles of neoadjuvant EV; RC was completed after neoadjuvant tx. Control arm: RC and observation or locally approved adjuvant tx. Dual primary endpoints were event-free survival (EFS): D+EV vs control and D+T+EV vs control. Overall survival (OS) was the key secondary endpoint. Safety was also assessed. Here we present results from a pre-planned interim analysis (data cutoff: 02 Mar 2026). Results Overall, 233, 233, and 229 pts were randomised to the D+EV, D+T+EV, and control arms, respectively. RC was completed in 85% (D+EV), 79% (D+T+EV), and 78% (control) of pts; and 55% (D+EV) and 42% (D+T+EV) completed adjuvant tx. Median EFS follow-up was 22.9 mo. D+EV showed significant improvements in EFS (HR 0.51 [95% CI, 0.38–0.67], P<0.0001; 24-mo rate 64% vs 45%) and OS (HR 0.58 [95% CI, 0.40–0.83], P=0.0027; 24-mo rate 78% vs 69%) vs control. D+T+EV also showed a significant improvement in EFS vs control (HR 0.61 [95% CI, 0.47–0.80], P=0.0004; 24-mo rate 60% vs 45%) but did not meet significance for OS; follow-up is ongoing. Pathological complete response per local review was observed in 45% (95% CI, 38.6–51.7) of pts in the D+EV arm and 10% (95% CI, 6.1–14.2) in the control arm. In the D+EV and D+T+EV arms, max. Grade 3/4 tx-related adverse events occurred in 30% and 42% of pts, respectively. ctDNA results will also be presented. Conclusions Dual primary endpoints of EFS were met. Perioperative D plus 3 cycles of neoadjuvant EV also demonstrated a significant improvement in OS vs control. Safety profiles were consistent with those of the individual agents.

EFS hazard ratio Absolute pCR gain vs control (percentage points) * control = cystectomy alone (no neoadjuvant therapy) ENERGIZE +10 pp | HR 0.94 KEYNOTE-866 +9 pp | HR 0.88 Pembro alone* +15 pp | HR 0.71 NIAGARA +10 pp | HR 0.68 KEYNOTE-B15 +23 pp | HR 0.53 VOLGA* +32 pp | HR 0.51 KEYNOTE-905* +49 pp | HR 0.40

[Slide 1] Perioperative trials in MIBC – how much is enough? Trial | IO agent (all perioperative) | Comparator | Neoadj. cytotoxic | Adj. cytotoxic | Adjuvant cytotoxic received (% of randomised) | EFS | OS Chemotherapy backbone NIAGARA1 | Durvalumab | Neoadj. GC | GC x4 | None | Not applicable | Positive HR 0.68 | Positive HR 0.75 KEYNOTE-8662 | Pembrolizumab | Neoadj. GC | GC x4 | None | Not applicable | Negative HR 0.88 | Negative HR 1.00 ENERGIZE3 | Nivolumab | Neoadj. GC | GC x4 | None | Not applicable | Negative HR 0.94 | Negative HR 0.88 (immature) Enfortumab vedotin backbone KN-905/EV-3034 | Pembrolizumab | Surgery alone | EV x3 | EV x6 | Adjuvant started 59% (100/170); all 6 EV cycles ≥29%a | Positive HR 0.40 | Positive HR 0.50 KN-B15/EV-3045 | Pembrolizumab | Neoadj. GC | EV x4 | EV x5 | Adjuvant started 65% (262/405); adj. EV received 56% (225/405)b | Positive HR 0.53 | Positive HR 0.65 VOLGA (D+EV)6 | Durvalumab | Surgery ± adj. | EV x3 | None | Not applicable | Positive HR 0.51 | Positive HR 0.58 VOLGA (D+T+EV)6 | Durvalumab + tremelimumab | Surgery ± adj. | EV x3 | None | Not applicable | Positive HR 0.61 | Not significant HR 0.77 (interim) aData cutoff Jun 2025; median 6 adjuvant EV cycles among starters, so at least half of starters received all 6. bMedian 5 (range 1-5) adjuvant EV cycles among recipients (Suppl. Table S3); 208/405 (51%) completed the whole adjuvant phase. GC, gemcitabine-cisplatin; EV, enfortumab vedotin; D, durvalumab; T, tremelimumab. Shaded rows: adjuvant cytotoxic component. HR: EFS / OS vs comparator. Niklas Klümper, X: @niklas_kluemper Powles T, NEJM 2024; Bedke J, ESMO 2026 (4735O); Necchi A, ESMO 2026 (4736O); Vulsteke C, NEJM 2026; Galsky MD, NEJM 2026; Powles T, ESMO 2026 LBA2 [Slide 3] VOLGA: perioperative Durva, only neoadjuvant/ induction EV n = 695, cisplatin-ineligible or -declining | 3x EV before cystectomy, no adjuvant EV Arm 1: D + T + EV 1:1:1, n = 695 in total Arm 2: D + EV Arm 3: RC ± approved adjuvant therapy RC Enfortumab vedotin Durvalumab Tremelimumab Niklas Klümper, X: @niklas_kluemper Powles T et al., ESMO 2026 LBA2 (Presidential Symposium); VOLGA NCT04960709; data cutoff 02 March 2026 [Slide 4] VOLGA: durvalumab + 3 cycles of EV vs cystectomy Durvalumab + EV arm | ITT | median follow-up 23 (EFS) and 26 (OS) months Event-free survival HR 0.51 (0.38-0.67) 24 mo: 64% vs 45% Overall survival HR 0.58 (0.40-0.83) 24 mo: 78% vs 69% D+EV N=233 | Control N=229 Probability of EFS Probability of OS Time from randomisation (months) Durvalumab+EV arm Control arm No. of patients at risk 40% vs 8% pCR, central review (local: 45% vs 10%) 60% ctDNA clearance before cystectomy (52/87, exploratory) 85% vs 79% underwent cystectomy (198/233 vs 180/229) For patients: similar benefit to EV-P with only 3 EV cycles - 64% vs 45% event-free and 78% vs 69% alive at 2 years Niklas Klümper, X: @niklas_kluemper Survival table Endpoint: Event-free survival and overall survival, ITT (D+EV vs control/cystectomy) Durvalumab + EV (D+EV) | 233 Control (radical cystectomy) | 229 HR: EFS 0.51; OS 0.58 HR CI: EFS 0.38-0.67; OS 0.40-0.83 24-mo EFS 64% vs 45% 24-mo OS 78% vs 69% pCR 40% vs 8% central review (local 45% vs 10%) Underwent cystectomy 85% vs 79% (198/233 vs 180/229) ctDNA clearance before cystectomy 60% (52/87, exploratory)

Depth of response tracks with benefit Absolute pCR gain versus EFS hazard ratio across four phase 3 trials NIAGARA ΔpCR +10 pp | HR 0.68 KEYNOTE-B15 ΔpCR +23 pp | HR 0.53 VOLGA* ΔpCR +32 pp | HR 0.51 KEYNOTE-905* ΔpCR +49 pp | HR 0.40 * KEYNOTE-905 and VOLGA: control = cystectomy alone no neoadjuvant therapy (pCR 8-9%) → ΔpCR inflated; NIAGARA and B15 vs active GC EFS hazard ratio (y-axis 0,3–0,8); Absolute pCR gain vs control (percentage points) +10 pp | NIAGARA: 37% vs 28%, HR 0.68 +23 pp | B15: 56% vs 33%, HR 0.53 +49 pp* | KN-905: 57% vs 9%, HR 0.40 - control without neoadjuvant therapy +32 pp* | VOLGA: 40% vs 8%, HR 0.51 - control without neoadjuvant therapy Niklas Klümper, X: @niklas_kluemper Powles T et al., NEJM 2024; Vulsteke C et al., NEJM 2026; Galsky MD et al., NEJM 2026; Powles T et al., ESMO 2026 LBA2. Trial-level, hypothesis-generating; *control: cystectomy alone [PowerPoint artifact label 'Zeichnungsfläche' overlays x-axis]

[Slide 1] EV+P is approved across the UC continuum and is the recommended 1L treatment for unresectable/mUC*1 ESMO Clinical Practice Guidelines2 Treatment-naïve mUC EV+P†§ Disease progression: • PBCT‡ • Erdafitinib¶ If EV+P is unavailable or contraindicated Treatment-naïve mUC Cis-eligible: Nivolumab + CisGem§ Cis- or carbo-eligible: CisGem or CarboGem Disease progression: • Pembrolizumab§ • Atezolizumab No disease progression: • Maintenance avelumab§†† Disease progression: • Erdafitinib¶§ • EV§ • Sacituzumab govitecan†§ • Vinflunine or taxanes‡‡¶¶§§ Figure adapted from Figure 1 from Powles T et al. 2024.2 *Disclaimer: The ESMO Clinical Practice Guideline aligns with the EU regulatory approval for P 'for the 1L treatment of unresectable or mUC in adults'; †No longer approved in the US; not EMA approved; ‡Rechallenge with single-agent ICI is not encouraged without further evidence; ¶In tumours with selected FGFR3 DNA fusions and mutations; §ESMO-MCBS v1.1 was used to calculate scores for new therapies/indications approved by the EMA or FDA. The scores have been calculated and validated by the ESMO-MCBS Working Group and reviewed by the authors (https://www.esmo.org/guidelines/esmo-mcbs/esmo-mcbs-evaluation-forms); ††This should be assessed within 10 weeks of completion of chemotherapy; ‡‡Rechallenge with PBCT may be considered if progression occurred 12 months after the end of previous PBCT or 12 months after the end of previous PBCT and maintenance avelumab; ¶¶Platinum doublets to be considered if the treatment-free interval from the last PBCT is >1 year; §§To be considered when other therapies are not available. 1L, first-line; Carbo, carboplatin; Cis, cisplatin; EMA, European Medicines Agency; ESMO, European Society for Medical Oncology; EV, enfortumab vedotin; FDA, U.S. Food and Drug Administration; FGFR3, fibroblast growth factor receptor 3; Gem, gemcitabine; ICI, immune checkpoint inhibitor; MCBS, Magnitude of Clinical Benefit Scale; mUC, metastatic urothelial carcinoma; P, pembrolizumab; PBCT, platinum-based chemotherapy; UC, urothelial carcinoma. 1. Enfortumab vedotin. Summary of Product Characteristics; 2. Powles T et al. Ann Oncol 2024;35:485–490. [Slide 3] EV-304 study: Pivotal Phase III trial of perioperative EV+P vs. CisGem and RC+PLND in cisplatin-eligible patients with MIBC Adults with MIBC Key eligibility criteria • Clinical stage T2–T4aN0M0 or T1–T4aN1M0 by central assessment • Tumours with ≥50% urothelial histology • Did not meet any Galsky criteria for cisplatin ineligibility • ECOG PS 0–1 • Eligible for curative-intent RC+PLND N=808 | R 1:1 n=405: Four cycles neoadjuvant EV 1.25 mg/kg Day 1 and Day 8 IV Q3W + Four cycles neoadjuvant pembrolizumab 200 mg Day 1 IV Q3W → RC+PLND → Five cycles adjuvant EV 1.25 mg/kg Day 1 and Day 8 IV Q3W + 13 cycles adjuvant pembrolizumab 200 mg Day 1 IV Q3W n=403: Four cycles neoadjuvant CisGem 70 mg/m2 Day 1 IV (Cis) + 1000 mg/m2 Day 1 and Day 8 IV (Gem) → RC+PLND → Observation* Endpoints1 • Primary: EFS by BICR • Key secondary: OS and pCR by central pathologist review • Other secondary: pDS and safety Participants were randomised between 12 May 2021 and 12 December 2023. Data cutoff date: 27 October 2025. *As of February 2023, adjuvant nivolumab was permitted when clinically indicated and regionally available. BICR, blinded independent central review; Cis, cisplatin; ECOG PS, Eastern Cooperative Oncology Group performance status; EFS, event-free survival; EV, enfortumab vedotin; Gem, gemcitabine; IV, intravenous; MIBC, muscle-invasive bladder cancer; OS, overall survival; P, pembrolizumab; pCR, pathological complete response; pDS, pathological downstaging rate; Q3W, every 3 weeks; R, randomised; RC+PLND, radical cystectomy with standard pelvic lymph node dissection. Galsky MD et al. N Engl J Med 2026;395:338-348. [Slide 4] Summary EV+P is the recommended 1L SOC for LA/mUC in patients who are platinum-eligible, as included in multiple international treatment guidelines including ESMO and EAU guidelines1,2 In EV-302, ~3.5-year follow-up data confirm sustained OS benefit and higher ORR and CR rates for EV+P vs. PBCT in 1L LA/mUC, with patients continuing to achieve CR over time. No new safety findings were identified3 Real-world outcomes closely mirror those of EV-302, with consistent ORR, CR, and survival benefit observed4–7 Data from the EV-303 and EV-304 trials supports perioperative EV+P as an effective treatment option for patients with resectable MIBC, regardless of cisplatin-eligibility8,9 1L, first-line; CR, complete response; EAU, European Association of Urology; ESMO, European Society for Medical Oncology; EV, enfortumab vedotin; LA, locally advanced; MIBC, muscle-invasive bladder cancer; mUC, metastatic urothelial carcinoma; ORR, objective response rate; OS, overall survival; P, pembrolizumab; SOC, standard of care. 1. European Association of Urology. Muscle-invasive and metastatic bladder cancer. Available at: https://uroweb.org/guidelines/muscle-invasive-and-metastatic-bladder-cancer. Last accessed: October 2026; 2. Powles T et al. Ann Oncol 2024;35:485–490; 3. Powles TB. Presented at ASCO 2026. #4507; 4. Zschäbitz S et al. Cancer Immunol Immunother 2026;75:184; 5. Niedersüß-Beke D et al. Int J Cancer 2025;158:1654–1665; 6. Jindal T et al. Presented at ESMO 2024. Abstract 1988P; 7. Jain P et al. Clin Genitourin Cancer 2025;23:102453; 8. Vulsteke C et al. N Engl J Med 2026;394:1257–1269; 9. Galsky MD et al. N Engl J Med 2026;395:338-348.

[Slide 1] KIM-1 Dynamics Complement Tissue Biomarkers & Could Inform Adaptive Therapy 1 BASELINE: Tumour transcriptional programmes vary between regions, over time and between sites Baseline Tissue | Checkmate 214(1) | No baseline established transcriptional signature associated with PFS or OS on Ipi/Nivo Within Primary Tumours | TRACERx Renal(2) | 31% of same-patient sample pairs differ Over Time | Immotion 010 (Placebo Arm)(3) | Recurrent sites enrich for inflammatory, stromal and proliferative signatures Between Sites | OPTIC-RCC(4) | GES vary between metastatic & primary site. Moved to metastatic site biopsy 2 DYNAMIC measures can be tools for ADAPTIVE studies Metastatic ccRCC → Biomarker sampling (baseline) → Nivolumab + ipilimumab → Biomarker sampling (Day 21) KIM-1 ≤30% decrease (~70% of patients) → R → Zanzalintinib + nivolumab / Nivolumab + ipilimumab (mPFS <8 months) KIM-1 >30% decrease (~30% of patients) → Nivolumab + ipilimumab (mPFS >70 months) NEED: 1 Analytical Validity: Reproducible in plasma ✓ cut off thresholds pending 2 Clinical Validity: Discriminates Response & Regimen ✓ 3 Clinical Utility: Prospective/ Adaptive Trials Needed ↑ Zayd Tippu Invited Discussant 3593RO, 3594RO and 3595RO 1 Checkmate 214 Motzer JITC 2022, 2 Fernandez Sanroman Cancer Discov 2025 3 Immotion 010 Xu European Urology 2026 4 OPTIC RCC Haake, ESMO 2025 2591O 5 COSMIC 313 Xu ESMO 2026, 3593RO [Slide 2] Integrated Profiling Could Guide Upfront & Response-Adapted Therapy 1 BASELINE: static profiling Contemporaneous metastatic biopsy PD-L1 ± transcriptomics (OPTIC-RCC/BIONIKK) Baseline KIM-1 Baseline HRQoL 2 ON TREATMENT: dynamic profiling Serial blood KIM-1 ± ctDNA/ cfDNA (recapitulates TME) Serial HRQoL trajectory +/- wearables Dynamic change from baseline UPFRONT STRATIFICATION ICI + ICI or ICI + VEGF? 3594RO PROGNOSTICATION Risk beyond IMDC? 3593RO, 3595RO RESPONSE-ADAPTED THERAPY Timing of intensification? 3593RO SEQUENCING AFTER ICI Selection and optimisation 3595RO Advanced ccRCC | 1st line SACT | Dynamic monitoring | Progressive disease | 2nd line SACT Integrated Features Across Tumour, Blood and the Patient: May Optimise Upfront & Dynamic Therapy Choices Requires prospective validation Zayd Tippu Invited Discussant 3593RO, 3594RO and 3595RO Flippot 3594RO (CARE1), Xu 3593RO (COSMIC 313), Heng 3595RO (LITESPARK 011), Haake 2591O (OPTIC RCC, ESMO 2025), Vano Lancet Oncology 2022, Zhang, Nature 2026 [Slide 3] Guiding Clinical Decision-Making: Interrogating PD-L1 Advanced ccRCC | 1st line SACT | Dynamic Monitoring | Progressive Disease | 2nd line SACT 1 3594RO · tissue UPFRONT STRATIFICATION ICI + ICI or ICI + VEGF? Nil implementable biomarkers • Patient, clinical & histological features Absence of prospective studies Need: A stratifier between regimens The BIOMARKER ASK: Analytical Validity ↔ Measurable/Reproducible Scalable tissue test across local labs 1 Tissue · 3594RO PD-L1 CARE-1 CARE-1 Study Metastatic renal cell carcinoma with a clear-cell component Intermediate or poor risk IMDC Naive from systemic therapy Local PD-L1 TPS ≥ 1% or < 1% Routine antibody Main objective of the PD-L1 analysis: Concordance between local and central PD-L1 positivity PD-L1 positive → R 1:1 → DUAL ICI NIVOLUMAB + IPILIMUMAB / ICI + TKI PEMBROLIZUMAB + AXITINIB, PEMBROLIZUMAB + LENVATINIB, NIVOLUMAB + CABOZANTINIB Overall survival Hypothesis dual ICI > ICI + TKI PD-L1 negative → R 1:1 → DUAL ICI NIVOLUMAB + IPILIMUMAB / ICI + TKI PEMBROLIZUMAB + AXITINIB, PEMBROLIZUMAB + LENVATINIB, NIVOLUMAB + CABOZANTINIB Progression-free survival Overall survival Hypothesis ICI + TKI > dual ICI Zayd Tippu Invited Discussant 3593RO, 3594RO and 3595RO [Slide 4] Guiding Clinical Decision-Making: the Need for Validated Biomarkers and PROs Advanced ccRCC | 1st line SACT | Dynamic Monitoring | Progressive Disease | 2nd line SACT 1 3594RO · tissue UPFRONT STRATIFICATION ICI + ICI or ICI + VEGF? Nil implementable biomarkers • Patient, clinical & histological features Absence of prospective studies Need: A stratifier between regimens 2 exploratory: triplet responders UPFRONT INTENSIFICATION Role for Triplets? Triplets have not improved OS • LITESPARK-012: HR 1.04 (Arm A) • HR 1.12 (Arm B)(1) • COSMIC 313: HR 1.02(2) Toxicity/tolerance pay-off Need: Who benefits from more 2 3593RO · blood RESPONSE-ADAPTED THERAPY Timing of Intensification? Radiological response • 12 weeks: PDIGREE (NCT03793166) Role for blood-based signal? Need: Early signal that tracks disease response 3 3595RO · patient SEQUENCING AFTER ICI Selection and Optimisation Combination regimen has not improved OS • LITESPARK-011: HR 0.85 (95% CI 0.68-1.05)(3) Prognostic role of HRQoL measures? Need: Robust monitoring and strategies for intervention 1 Tissue · 3594RO PD-L1 CARE-1 2 Blood · 3593RO KIM-1 COSMIC-313 3 Patient · 3595RO HRQoL LITESPARK-011 Zayd Tippu Invited Discussant 3593RO, 3594RO and 3595RO 1 #3586O ESMO 2026, 2 Motzer et al. Ann Oncol 2026 3 Motzer et al Lancet 2026

[Slide 1] ESMO ANNALS OF ONCOLOGY ORIGINAL ARTICLE · Articles in Press, October 25, 2026 Final prespecified analysis of the phase 3 LITESPARK-011 study: belzutifan plus lenvatinib versus cabozantinib for advanced renal cell carcinoma after anti–PD-(L)1 therapy D.Y.C. Heng 1,* · M. Schmidinger 2 · R. McDermott 3 · ... · D. Vickery 29 · R.J. Motzer 30,* on behalf of the LITESPARK-011 Investigators ... Show more HIGHLIGHTS • Belzutifan plus lenvatinib continues to demonstrate PFS and ORR benefit versus cabozantinib at the final analysis of LITESPARK-011 [Slide 2] ESMO ANNALS OF ONCOLOGY ORIGINAL ARTICLE · Articles in Press, October 25, 2026 Pembrolizumab plus lenvatinib with belzutifan or quavonlimab as first-line therapy for advanced renal-cell carcinoma: randomized open-label phase 3 LITESPARK-012 study Brian I. Rini 1 · Toni K. Choueiri 2 · Martin H. Voss 3 · ... · Joseph E. Burgents 19 · Karla G. Rodriguez-Lopez 19 · Thomas B. Powles 20 ... Show more Highlights • LITESPARK-012 evaluated novel first-line combination therapies for advanced RCC. • Neither HIF-2α (belzutifan) nor CTLA-4 (quavonlimab) blockade added to the efficacy of first-line pembrolizumab plus lenvatinib in advanced RCC.

[Slide 1] How much systemic therapy? ADT remains essential – but who needs intensification? • Long-term ADT – backbone of treatment (typically 18–36 months) • ADT + 2 years abiraterone – standard for selected very-high-risk / cN1 M0 • STAMPEDE – major change in practice • ENZARAD – negative for MFS; 8-y MFS 74% vs 72% • ARPI intensification – not simply a class effect • Docetaxel – very limited role in M0 locally advanced disease and DASL-HiCaP... Bottom line: ADT remains the backbone. Intensify selectively — strongest evidence supports abiraterone, while enzalutamide cannot be considered interchangeable. Selected references Bolla et al. N Engl J Med. 2009 Nabid et al. Eur Urol. 2018 Attard et al. Lancet. 2022 (STAMPEDE) Nguyen et al. ESMO Congress 2025, LBA86 (ENZARAD) James et al. JNCI Cancer Spectr. 2022 [Slide 2] PROTEUS: a new piece of the puzzle 1 Neoadjuvant 6 months ADT ± apalutamide 2 Radical prostatectomy + PLND 3 Adjuvant 6 months ADT ± apalutamide pCR/MRD 8.9% vs 1.0% 5-year MFS 78.2% vs 73.5% Post-op RT ~1 in 3 patients High-risk / locally advanced prostate cancer often requires multimodal care. Taplin ME, et al. Perioperative Apalutamide in High-Risk Localized Prostate Cancer. N Engl J Med. 2026;395:546–560. [Slide 3] PROTEUS, STAMPEDE, DASL-HiCaP and ENZARAD: what can we conclude? Different clinical contexts and designs — NOT head-to-head comparisons PROTEUS Surgery pathway Perioperative apalutamide + ADT 5-year MFS 78.2% vs 73.5% STAMPEDE Definitive RT pathway ADT ± abiraterone-based intensification MFS HR 0.54; OS HR 0.63 ENZARAD Definitive RT pathway 24-month enzalutamide + ADT 8-year MFS 74% vs 72%; HR 0.88 No significant MFS benefit DASL-HiCaP Definitive or salvage RT 96-week ADT ± darolutamide 4-year MFS 95% vs 88%; HR 0.42 Systemic intensification may strengthen local treatment · but results are not uniform Sources: Taplin et al. NEJM 2026 (PROTEUS); STAMPEDE; DASL-HiCaP ESMO 2026 abstract; ENZARAD ESMO 2025 abstract. [Slide 4] The Evidence Mandates 18–36 Months for High-Risk Disease Global Consensus ESMO and EAU guidelines remain unified: true high-risk localized disease requires 18–36 months (2–3 years) of ADT. Proven Survival Benefit Meta-analyses (Kishan 2022, Zaorsky 2025) and landmark trials (RADICALS-HD, STAMPEDE) prove premature cessation sacrifices Metastasis-Free Survival (MFS) and Overall Survival (OS). The ARPI Intensification Adding next-generation androgen receptor pathway inhibitors (ARPIs) enhances efficacy but does not replace the need for long-term ADT duration.

[Slide 1] Stage Migration with PSMA PET Imaging MIGRATION BETWEEN CONVENTIONAL IMAGING AND PSMA PET Conventional imaging (CI) (CT/MRI + Bone Scan) ------> PSMA-PET High-Volume Disease (HVD-CI) n = 17 Low-Volume Disease (LVD-CI) n = 50 High-Volume Disease (HVD-PET) n = 27 Low-Volume Disease (LVD-PET) n = 24 Non-Metastatic Disease (M0-PET) n = 16 94% 22% 48% 30% Consider over- but also under-treatment depending on imaging used* Note: only 67 patients Aurelius Omlin *Unterrainer et al. J Nucl Med 2025; 66:54–60 Sundahl et al. Eur Urol. 2021 May;79(5):565-567. [Slide 2] Available Biomarkers for Treatment Personalisation Method | Biomarker Result | Potential Treatment Implication PSA / Androgen Receptor (Serum, IHC) | Approx. 99% positive | ADT + ARPI Imaging by PSMA PET/CT* | 87% positive (syn- and metachronous mHSPC/APMS) | 177Lu-PSMA-617# Immunohistochemistry on tumour tissue** | 25% PTEN-loss ≥90% (synchronous mHSPC/mAPMS) | ADT + abiraterone + capivasertib# Next-Generation Sequencing (Tumour and/or ctDNA/Plasma) AMPLITUDE Panel*** | 18% HRR-alterations — of which 56% BRCA1/2 (syn- and metachronous mHSPC/APMS) | ADT + abiraterone + niraparib (optional 6x docetaxel upfront)## TALAPRO-3 Panel**** | 19% HRR-alterations — of which 35% BRCA1/2 (syn- and metachronous mHSPC/APMS) | ADT + enzalutamide + talazoparib# * PSMAddition (68Ga-PSMA-11 PET/CT) ** CAPItello-281 (Antibody for PTEN, SP218; Roche Diagnostics, Tucson, AZ) *** AMPLITUDE: BRCA1/2, BRIP1, CHEK2, CDK12, FANCA, PALB2, RAD51B, RAD54L **** TALAPRO-3: ATM, ATR, BRCA1/2, CDK12, CHEK2, FANCA, MLH1, MRE11A, NBN, PALB2, RAD51C # Not approved yet by EMA ## approved by EMA only for BRCA1/2 Aurelius Omlin [Slide 3] STAMPEDE – Biomarker Analysis Associated with shorter survival on abiraterone Cellular plasticity / Immune activity / Tumor proliferation / Higher AR signalling / Longer survival RNA profiling of advanced prostate cancer Randomized in STAMPEDE trial (survival status after ~14 years) Abiraterone Docetaxel Tumor biopsy RNA Transcriptome classifiers PTEN inactivity High Decipher score Predict survival benefit from docetaxel in metastatic patients Transcriptome-based classification of PTEN activity PTEN inactive 53% of pts PTEN active 47% of pts A M1, PTEN inactive — Overall survival — Years since randomization B M1, PTEN active — Overall survival — Years since randomization PTEN active / PTEN inactive — Lower Decipher / High Decipher G Docetaxel treatment effect (all-cause death hazard ratio) Aurelius Omlin * Note: PTEN IHC by Cell Signaling Technology, Clone D4.3 XP antibody Grist et al. Cell, 2025 Survival table Endpoint: Overall survival, M1 patients, ADT vs ADT + docetaxel, by transcriptome PTEN status PTEN inactive: ADT (reference) vs ADT + doce PTEN active: ADT (reference) vs ADT + doce [Slide 4] ADT plus ARPI ± Docetaxel: Experts' Treatment Choice | Disease | ADT + ARPI | ADT + ARPI + docetaxel | ADT + ARPI + 177Lu-PSMA | Comment PSMA positive, no HRR, no PTEN loss | Synchronous, high-volume | 25% | 68% | 7% | | Metachronous, high-volume | 45% | 41% | 13% | 1% for ADT + ARPI (± docetaxel) + PARPi PSMA positive, no HRR, no PTEN loss | Synchronous, low-volume | 90% | 3% | 6% | 1% for ADT + ARPI (± docetaxel) + PARPi | Metachronous, low-volume | 90% | 1% | 7% | 2% for ADT + ARPI (± docetaxel) + PARPi Most uncertainty around metachronous high-volume disease Aurelius Omlin Source: APCCC 2026 consensus voting, publication pending

[Slide 1] 5- Who are the right patients for HIF-2α inhibition? Slowly growing disease? The INDOLENT population BRAVE-RCC (NCT07023432) Belzutifan's Role in Active Surveillance Versus Treatment for Indolent Metastatic Clear Cell Renal Cell Carcinoma A phase II randomized study in treatment-naïve metastatic clear-cell renal cell carcinoma Key inclusion criteria: Metastatic clear-cell RCC; IMDC favorable/intermediate risk (0–1); No prior systemic therapy* Randomization (1:1) N = 78 Belzutifan (oral, once daily) | Active surveillance (standard monitoring) Primary endpoint: Progression-free survival (PFS) (RECIST v1.1) Key secondary endpoints: Safety and tolerability (CTCAE v5.0); Time to start of new systemic therapy; Objective response rate (CR and PR, RECIST v1.1); Overall survival; Quality of life (FKSI-19) Sponsor: The University of Texas MD Anderson Cancer Center Planned sample size N = 78 * Prior adjuvant immunotherapy/targeted therapy is allowed if metastatic progression occurs ≥ 12 months after the last dose. Generated by Claude from ClinicalTrials.gov Laurence Albiges, MD PhD [Slide 2] OUTLOOK Where cellular therapy for RCC goes next Better targets: CD70 selection by TPS; ENPP3 and new antigens; monitor antigen heterogeneity Armoured and off-the-shelf: CTX131 (Regnase-1 + TGFβR2 KO); anti-rejection allogeneic designs Smarter TIL: TCR-seq-guided enrichment of tumour-reactive CD8+ clones Manufacturability biomarkers: CD8/FoxP3 ratio and PD-1+ Tregs predict expansion failure Toxicity control: Recognize HLH early; infection prophylaxis after lymphodepletion Combinations, earlier lines: After prior anti-PD1; lower tumour burden; randomised designs Pal SK et al. Cancer Discov 2024. Srour SA et al. JCO 2026. Einstein DJ et al. J Immunother 2024. Lee MH et al. Cancer Res Commun 2023. [Slide 3] BARRIER 3: Payload Does the delivered payload actually kill RCC? Pharmacokinetics | Tumor penetration | Target recognition | Intracellular trafficking | Payload | TME modulation Mosele et al. Nat Med (32)2774-2785, 2026 Potential barriers in ccRCC 2. Therapeutic window → Can we deliver enough payload to kill the tumor without unacceptable toxicity? AMG 172 | CD70 | DM1 (mertansine) Microtubule inhibitor DLT: Thrombocytopenia 59% (all grades); Liver injury / Hepatic dysfunction 30% (all grades) Clinical activity: ORR 5.4% (2/37); Best response: 2 PR, 14 SD, 21 PD Clinical consequence: DLTs: thrombocytopenia, liver injury; MTD 1.6 mg/kg; Limited further development Massard C, et al. Cancer Chemother Pharmacol. 2019;83:1057-1063. SGN-CD70A | CD70 | PBD dimer DNA cross-linker DLT: Thrombocytopenia 56% (all grades) Clinical activity: ORR 5.6% (1/18); Best response: 1 PR, 8 SD, 4 PD Clinical consequence: DLT: thrombocytopenia; MTD 30 µg/kg; Limited further development Pal S, et al. Cancer. 2019;125:1124-1132. CDX-014 | TIM-1 | MMAE Microtubule inhibitor DLT: 1 Grade 3 rash + Grade 3 liver dysfunction + Grade 4 neutropenia (at 2.0 mg/kg Q3W) → Grade 5 multiorgan failure (fatal, 1/16); 2 Grade 3 hyperglycemia (at 2.0 mg/kg Q3W) Clinical activity: ORR 6% (1/16); Best response: 1 PR (ongoing), 9 SD, 7 PD; Clinical benefit rate 31% (5/16) Clinical consequence: MTD not reached; 1 treatment-related death (Grade 5); Therapeutic window challenged; development discontinued McGregor BA, et al. Invest New Drugs. 2020;38:1807-1814.

[Slide 1] My Thoughts - Did LITESPARK-012 meet its primary endpoint? - No, dual primary endpoints of PFS and OS were not met for either drug combinations of lenvatinib/belzutifan/ pembrolizumab or lenvatinib + quavonlimab + pembrolizumab - Are these findings practice changing? No - Bispecific T cell engagers are under investigation as future first-line therapies - What this means for patients: - 4 approved SOC regimens remain as benchmark first-line therapies for clear cell renal cancer: - Ipilimumab +nivolumab - Axitinib +pembrolizumab - Cabozaninib + nivolumab - Lenvatinib + pembrolizumab [Slide 2] Positioning HIF-2α inhibition - HIF-2α pathway modulation studies are ongoing HIF-2α inhibitors Completed (NCT03634540) Phase II LITESPARK-003 belzutifan + cabozantinib (Choueiri, Lancet 2025) -active and tolerable Ongoing (NCT07011719) PEAK-1 (Phase 3): casdatifan + cabozantinib versus cabozantinib alone in patients with metastatic ccRCC post immunotherapy ( NCT05536141) ARC-20 casdatifan monotherapy post PD-(L)1 and VEGFR-TKI therapies (arms include ivonescimab VEGFTKI-IO bispecific) (NCT07489495) LITESPARK-034/LS-034/MK-6482-034 Zanzanitinib + belzutifan vs belzutifan post VEFRTKI +ICI (NCT07227402) LITESPARK-033 Belzutifan and Zanzalintinib Versus Cabozantinib post Adjuvant Therapy Glue Degraders (NCT07300241) Phase I/II NEO-811 investigational molecular glue degrader (NCT07195682) BMS-986506: cereblon E3 ligase modulator (CELMoD) / molecular glue candidate Pre-clinical Molecular Glue cc-885 Inhibits VHL-Deficient Clear Cell Renal Cell Carcinoma

[Slide 1] 16:30 - 18:15 Presidential Symposium I CHAIRS: FABRICE ANDRÉ, JAMES LARKIN EFS and pCR in VOLGA and KN905/EV303: hint at contribution of components? [Top-left inset: reproduced VOLGA EFS by BICR D+EV vs control KM curve; small text] D+EV N=233 | Control N=229 HR (95% CI) 0.51 (0.38–0.67); Stratified log-rank P value <0.0001 73% | 64% (D+EV); 50% | 45%[?] (Control) Durvalumab+EV arm | Control arm | Median follow-up: 23 months [Bottom-left: KEYNOTE-905/EV-303 EFS KM curve] | EV + pembro (N = 170) | Control (N = 174) Events, n (%) | 48 (28.2) | 95 (54.6) Median, mo (95% CI) | NR (37.3–NR) | 15.7 (10.3–20.5) HR 0.40, 95% CI 0.28–0.57, one-sided P <.0001* 77.8% | 74.7% EVP; 55.1% | 39.4% control Event-free survival, %; Months 0-60 No. at Risk EV + pembro | 170 | 140 | 116 | 73 | 56 | 42 | 35 | 16 | 3 | 0 | 0 Control | 174 | 116 | 84 | 48 | 33 | 22 | 14 | 9 | 1 | 0 | 0 Cross-trial differences limit comparison • Both trials positive for EFS • Lower HR in KN905/EV303 (0.40 vs 0.51) • 95% CIs overlap pCR: lower with D+EV compared to EVP Less neoadjuvant activity than EV+pembro? EV + durvalumab | EV + pembrolizumab VOLGA investigator-assessed 45% | VOLGA central review 40% | KEYNOTE-905/EV-303 57% | KEYNOTE-B15/EV-304 56% Michiel van der Heijden Invited Discussant LBA2 ALICANTE AUDITORIUM - HALL 6 Survival table Endpoint: EFS, KEYNOTE-905/EV-303 (EV + pembro vs control), as reproduced on discussant slide Arm | N | Median | Median CI EV + pembro | 170 | NR | 37.3–NR Control | 174 | 15.7 | 10.3–20.5 HR: 0.40 HR CI: 0.28–0.57 P: one-sided P <.0001 Landmark rates: 12 months: EV + pembro 77.8% | Control 55.1% 24 months: EV + pembro 74.7% | Control 39.4% [Slide 2] 16:30 - 18:15 Presidential Symposium I CHAIRS: FABRICE ANDRÉ, JAMES LARKIN Perioperative EV-IO trials in MIBC - Conclusions • Three positive trials with perioperative EV+IO for patients with MIBC • EFS HRs somewhat different, landmark 2yr OS similar, further follow-up needed • VOLGA shows high efficacy without adjuvant EV • Trials on the adjuvant component(s) of peri-operative regimens are needed • Whereas less may be enough, with anti-CTLA-4, more seems to be too much. • Congratulations to the study team for completing VOLGA in a difficult to study disease setting!! Trial | EV | EV cycles | Control arm | 24-mo EFS Exp | 24-mo EFS Control | EFS HR (95% CI) | 24-mo OS Exp | 24-mo OS Control | OS HR (95% CI) VOLGA (D+EV)1 | Neoadjuvant only | 3 | RC ± approved adjuvant | 64% | 45% | 0.51 (0.38-0.67)* | 78% | 69% | 0.58 (0.40-0.83)* VOLGA (D+T+EV)1 | Neoadjuvant only | 3 | RC ± approved adjuvant | 60% | 45% | 0.61 (0.47–0.80)* | 75% | 69% | 0.77 (0.54–1.07) KEYNOTE-905/EV-3032 | Perioperative | 3+6=9 | RC alone | 74.7% | 39.4% | 0.40 (0.28–0.57)* | 80% | 63.1% | 0.50 (0.33–0.74)* KEYNOTE-B15/EV-3043 Cisplatin eligible | Perioperative | 4+5=9 | Neoadj GC → RC | 79.4% | 66.2% | 0.53 (0.41–0.79)* | 87% | 81.3% | 0.65 (0.48–0.89)* 1Powles, ESMO 2026 2Vulsteke, NEJM 2026; 3Galsky, NEJM 2026; *significant. RC, radical cystectomy; GC, gemcitabine-cisplatin; D, durvalumab; T, tremelimumab; Exp, experimental arm. 24-month rates are Kaplan-Meier estimates. Michiel van der Heijden Invited Discussant LBA2 ALICANTE AUDITORIUM - HALL 6

[Slide 1] Trial | EV | EV cycles | Control arm | 24-mo EFS Exp | 24-mo EFS Control | EFS HR (95% CI) | 24-mo OS Exp | 24-mo OS Control | OS HR (95% CI) VOLGA (D+EV)1 | Neoadjuvant only | 3 | RC ± approved adjuvant | 64% | 45% | 0.51 (0.38-0.67)* | 78% | 69% | 0.58 (0.40-0.83)* VOLGA (D+T+EV)1 | Neoadjuvant only | 3 | RC ± approved adjuvant | 60% | 45% | 0.61 (0.47–0.80)* | 75% | 69% | 0.77 (0.54–1.07) KEYNOTE-905/EV-3032 | Perioperative | 3+6=9 | RC alone | 74.7% | 39.4% | 0.40 (0.28–0.57)* | 80% | 63.1% | 0.50 (0.33–0.74)* KEYNOTE-B15/EV-3043 Cisplatin eligible | Perioperative | 4+5=9 | Neoadj GC → RC | 79.4% | 66.2% | 0.53 (0.41–0.79)* | 87% | 81.3% | 0.65 (0.48–0.89)* 1Powles, ESMO 2026 2Vulsteke, NEJM 2026; 3Galsky, NEJM 2026; *significant. RC, radical cystectomy; GC, gemcitabine-cisplatin; D, durvalumab; T, tremelimumab; Exp, experimental arm. 24-month rates are Kaplan-Meier estimates. [Slide 2] KN905/EV303 & EV304/KN-B15: a paradigm shift for patients with neoadjuvant MIBC [Top] Primary Endpoint: EFSa by BICR - ITT Population | EV + pembro (N = 170) | Control (N = 174) Events, n (%) | 48 (28.2) | 95 (54.6) Median, mo (95% CI) | NR (37.3–NR) | 15.7 (10.3–20.5) HR 0.40, 95% CI 0.28–0.57, one-sided P <.0001* Curve labels: 77.8%, 74.7% (EV + pembro); 55.1%, 39.4% (control) 29/174 pts (16.7%) in the control arm received adjuvant nivolumab as of data cutoff per local guidelines/prescribing information No. at Risk EV + pembro | 170 | 140 | 116 | 73 | 56 | 42 | 35 | 16 | 3 | 0 | 0 Control | 174 | 116 | 84 | 48 | 33 | 22 | 14 | 9 | 1 | 0 | 0 KEYNOTE-905/EV-303 EVP vs resection alone MIBC, n=344 Neoadjuvant: 3 cycles EVP Adjuvant: 6 cycles (total 9) Vulsteke et al, ESMO&NEJM 2025 pCR bar chart: Estimated differencea 48.3% (95% CI 39.5–56.5) one-sided P <.000001*; EV + pembro vs Control [Bottom] Primary Endpoint: EFS by BICR - ITT Population | EV + pembro (N = 405) | Cis + gem (N = 403) Events, n (%) | 87 (21.5) | 146 (36.2) Median, mo (95% CI) | NR (NR–NR) | 48.5 (43.3–NR) HR 0.53, 95% CI 0.41–0.70 [?], 1-sided P <.0001* Curve labels: 86.0%, 79.4% (EV + pembro); 75.4%, 66.2% (Cis + gem) No. at Risk EV + pembro | 405 | 351 | 317 | 303 | 257 | 154 | 100 | 63 | 23 | 0 Cis + gem | 403 | 350 | 286 | 259 | 216 | 131 | 70 | 45 | 16 | 0 KEYNOTE-B15/EV-304 EVP vs cisplatin/gemcitabine MIBC, n= 808 Neoadjuvant: 4 cycles EVP Adjuvant: 5 cycles (total 9) Galsky et al, ASCO-GU&NEJM 2026 pCR bar chart: Estimated differencea 23.4% (95% CI 16.7–29.8) 1-sided P <.0001*; EV + pembro vs Cis + gem Data cutoff date: 27 October 2025 [?] [Slide 3] [Reproduced VOLGA study design: Patients - Adults aged ≥18 years; MIBC (cT2-T4aN0/1M0 or cT1N1M0) by local assessment; Ineligible for or refused cisplatin; UC or UC with divergent differentiation or histological subtypes; Evaluated and confirmed for cystectomy. R (1:1:1) N=695. D+EV N=233: Durvalumab Q3W × 3 cycles, Enfortumab vedotin Q3W × 3 cycles -> RC -> Durvalumab Q4W × 9 cycles. D+T+EV N=233: Durvalumab Q3W × 3 cycles, Tremelimumab Q3W × 2 cycles, Enfortumab vedotin Q3W × 3 cycles -> RC -> Durvalumab Q4W × 9 cycles, Tremelimumab Q4W × 1 cycle. Control N=229: No treatment, Straight to RC -> Observation/locally approved treatment.] • D+EV positive for EFS and OS • D+T+EV positive for EFS • Another great result for bladder cancer patients!! VOLGA: EFS by BICR for Durvalumab+EV vs Control – ITT D+EV N=233 vs Control N=229: Events 84 (36) vs 123 (54); Median EFS NR (33.2–NC) vs 12.1 (7.6–27.0); HR (95% CI) 0.51 (0.38–0.67); Stratified log-rank P value <0.0001; curve labels 73%, 64% (D+EV), 50%, 45% (control); Median follow-up: 23 months VOLGA: OS for Durvalumab+EV vs Control – ITT Deaths 49 (21) vs 74 (32); Median OS NR (NC–NC) vs NR (37.7–NC); HR (95% CI) 0.58 (0.40–0.83); Stratified log-rank P value 0.0027; curve labels 87%, 78% (D+EV), 81%, 69% (control); Median follow-up: 26 months [Slide 4] EFS and pCR in VOLGA and KN905/EV303: hint at contribution of components? [Reproduced VOLGA EFS curve: D+EV N=233 vs Control N=229; Events 84 (36) vs 123 (54); Median EFS NR (33.2–NC) vs 12.1 (7.6–27.0); HR 0.51 (0.38–0.67); Stratified log-rank P <0.0001; labels 73%, 64%, 50%, 45%; Median follow-up: 23 months] [Reproduced KN905/EV303 EFS curve] | EV + pembro (N = 170) | Control (N = 174) Events, n (%) | 48 (28.2) | 95 (54.6) Median, mo (95% CI) | NR (37.3–NR) | 15.7 (10.3–20.5) HR 0.40, 95% CI 0.28–0.57, one-sided P <.0001* Curve labels: EVP 77.8%, 74.7%; control 55.1%, 39.4% No. at Risk EV + pembro | 170 | 140 | 116 | 73 | 56 | 42 | 35 | 16 | 3 | 0 | 0 Control | 174 | 116 | 84 | 48 | 33 | 22 | 14 | 9 | 1 | 0 | 0 Cross-trial differences limit comparison • Both trials positive for EFS • Lower HR in KN905/EV303 (0.40 vs 0.51) • 95% CIs overlap pCR: lower with D+EV compared to EVP Less neoadjuvant activity than EV+pembro? pCR (%) bar chart: EV + durvalumab: VOLGA investigator-assessed 45% | VOLGA central review 40% EV + pembrolizumab: KEYNOTE-905/EV-303 57% | KEYNOTE-B15/EV-304 56%

[Slide 2] Background • Belzutifan is a first-in-class HIF-2α inhibitor approved as: – Monotherapy in patients with advanced ccRCC following anti–PD-(L)1 therapy and ≥1 (US) or ≥2 (EU) VEGFR-TKIs – Adjuvant therapy in combination with pembrolizumab for adults with ccRCC at increased risk of recurrence following surgery (US) • Addition of belzutifan to VEGFR-TKI therapy in advanced ccRCC may enhance therapeutic benefit through orthogonal inhibition of HIF-2α and VEGF-mediated signaling pathways • Belzutifan + lenvatinib significantly improved PFS and ORR versus cabozantinib in participants with advanced ccRCC following anti–PD-(L)1 therapy in the phase 3 LITESPARK-011 study^1 – PFS at IA2: 14.8 vs 10.7 mo, HR 0.70 (95% CI 0.59–0.84), P=0.00007 – ORR at IA1: 52.6% vs 39.6%, P=0.0002 – These findings led to the US approval of belzutifan in combination with lenvatinib in adults with advanced ccRCC following anti–PD-(L)1 therapy on 24 September 2026 • We present LITESPARK-011 final analysis results

[Slide 1] [Composite discussant slide: VOLGA study design schema + EFS and OS KM thumbnails; small embedded text low resolution] Patients • Adults aged ≥18 years • MIBC (cT2-T4aN0/1M0 or cT1N1M0) by local assessment • Ineligible for or refused cisplatin • UC or UC with divergent differentiation or histological subtypes • Evaluated and confirmed for cystectomy Neoadjuvant | Adjuvant D+EV N=233: Durvalumab Q3W × 3 cycles; Enfortumab vedotin Q3W × 3 cycles | Durvalumab Q4W × 9 cycles D+T+EV N=233: Durvalumab Q3W × 3 cycles; Tremelimumab Q3W × 2 cycles; Enfortumab vedotin Q3W × 3 cycles | Durvalumab Q4W × 9 cycles; Tremelimumab Q4W × 1 cycle Control N=229: No treatment, Straight to RC | Observation/locally approved treatment • D+EV positive for EFS and OS • D+T+EV positive for EFS • Another great result for bladder cancer patients!! VOLGA: EFS by BICR for Durvalumab+EV vs Control – ITT [thumbnail; numbers too small to read reliably] VOLGA: OS for Durvalumab+EV vs Control – ITT [thumbnail; numbers too small to read reliably] Michiel van der Heijden Invited Discussant LBA2 [Slide 2] Contribution of adjuvant component in perioperative therapy? Perioperative trials in MIBC – how much is enough? Trial | IO agent: all Perioperative | Comparator | Neoadj cytotoxic | Adj cytotoxic | Adjuvant cytotoxic Received in % of randomized | EFS | OS Chemotherapy backbone NIAGARA1 | Durvalumab | Neoadj GC | GC ×4 | None | Not applicable | Positive | Positive KEYNOTE-8662 | Pembrolizumab | Neoadj GC | GC ×4 | None | Not applicable | Negative | Negative ENERGIZE3 | Nivolumab | Neoadj GC | GC x4 | None | Not applicable | Negative | Negative Enfortumab vedotin backbone KN-905/EV-3034 | Pembrolizumab | Surgery alone | EV ×3 | EV ×6 | EV started 59% (100/170) all 6 cycles ≥29%(a) | Positive | Positive KN-B15/EV-3045 | Pembrolizumab | Neoadj GC | EV ×4 | EV ×5 | EV started 65% (262/405) all 5 cycles ≥32%(b) | Positive | Positive VOLGA (D+EV)6 | Durvalumab | Surgery ± adj | EV ×3 | None | Not applicable | Positive | Positive VOLGA (D+T+EV)6 | Durvalumab + tremelimumab | Surgery ± adj | EV ×3 | None | Not applicable | Positive | Not significant (interim) 1Powles, NEJM 2024; 2Bedke, ESMO 2026 (4735O); 3Necchi, ESMO 2026 (4736O); 4Vulsteke, ESMO 2025; NEJM 2026; 5Galsky, ASCO GU,NEJM 2026; 6ESMO 2026 LBA2; (a)Data cutoff Jun 2025; median 6 adjuvant EV cycles among starters (ESMO 2025), so at least half of starters received all 6. (b)Median 5 (range 1-5) adjuvant EV cycles among starters; EV-specific completion not reported. 208/405 (51%) completed the whole adjuvant phase. GC, gemcitabine-cisplatin; EV, enfortumab vedotin; D, durvalumab; T, tremelimumab; NR, not reported. Lower bounds derived from median cycle numbers. Shaded rows: adjuvant cytotoxic component. Michiel van der Heijden Invited Discussant LBA2 [Slide 3] OS in VOLGA and KN905/EV303 [Top-left inset: reproduced VOLGA OS D+EV vs control KM curve, small text] D+EV N=233 | Control N=229 Deaths, n (%) 49 (21) | 74 (32); Median OS, months (95% CI) NR (NC–NC) | NR (37.7–NC) HR (95% CI) 0.58 (0.40–0.83); Stratified log-rank P value 0.0027 87% | 78% (Durvalumab+EV arm); 81% | 69% (Control arm) Median follow-up: 26 months [Bottom-left: KEYNOTE-905/EV-303 OS KM curve] | EV + pembro (N = 170) | Control (N = 174) Events, n (%) | 38 (22.4) | 68 (39.1) Median, mo (95% CI) | NR (NR–NR) | 41.7 (31.8–NR) HR 0.50, 95% CI 0.33–0.74, one-sided P =.0002* 86.3% | 79.7% (EV + pembro); 75.7% | 63.1% (Control) Overall survival, %; Months 0-60 No. at Risk EV + pembro | 170 | 159 | 144 | 94 | 68 | 56 | 45 | 28 | 10 | 0 | 0 Control | 174 | 150 | 130 | 75 | 54 | 45 | 30 | 18 | 6 | 0 | 0 OS HRs | HR (95% CI) KN905/EV303 | 0.50 (0.33–0.74) VOLGA: 2 arms without adjuvant EV EV + Durva | 0.58 (0.40–0.83) EV + Durva + Treme | 0.77 (0.54–1.07) Hazard ratio (log scale) 0.25 0.5 1 2 ← Favours experimental | Favours control → Michiel van der Heijden Invited Discussant LBA2 Survival table Endpoint: OS, KEYNOTE-905/EV-303 (EV + pembro vs control), as reproduced on discussant slide Arm | N | Median | Median CI EV + pembro | 170 | NR | NR–NR Control | 174 | 41.7 | 31.8–NR HR: 0.50 HR CI: 0.33–0.74 P: one-sided P =.0002 Landmark rates: 12 months: EV + pembro 86.3% | Control 75.7% 24 months: EV + pembro 79.7% | Control 63.1%

[Slide 2] Is the combination hypothesis incorrect? pCR outcomes were consistent across trials KEYNOTE-866 | CA017-078 | NIAGARA [KEYNOTE-866 bar chart] Estimated difference:b 8.9% (95% CI, 2.5-15.1)* ; Pembro + cis + gem | Placebo + cis + gem [CA017-078 bar chart] pCR overall (co-primary endpoint) 39.1 | 29.6 ; Overall ITT [NIAGARA bar chart] Re-analysis Apr 2024 [small text illegible at this resolution] [Slide 4] Were populations different? EFS outcomes were good in control arms of these studies Primary End Point: EFS by Investigator Assessment, ITT Population [KEYNOTE-866 KM curve - small text illegible] Event-free survival by treatment arm (co-primary endpoint) [CA017-078 KM curve - small text illegible] | KEYNOTE-866 | CA017-078 | NIAGARA T2 (%) | 15 | 76 | 40 N1 (%) | 0.7 | Excluded | 5 ECOG 0 (%) | 76 | - | 78 24 month EFS in control arm (%) | 70.0 | 71.6 | 59.8

Imvigor011: ctDNA- post RC INV-assessed DFS | | No treatment (n=357) | Events, n (%) | 39 (10.9) DFS, median (95% CI), mo | NE (NE) Median 21.8 mo fro[m ... - cropped] [Kaplan-Meier curve; shaded region 0-12 mo labelled: q6w ctDNA monitoring; dashed markers at 12 and 24 mo] 95.4% (at 12 mo marker) 88.4% (at 24 mo marker) Y axis: DFS (%) X axis: Time from cystectomy (mo) No. at risk No treatment: 357 357 357 342 337 329 303 289 275 233 185 156 116 102 78 66 46 38 20 11 3 (at 0,2,4,...,40 mo) 8 DFS events were deaths not clearly attributed to disease recurrence Survival table Endpoint: INV-assessed DFS (ctDNA-negative post-RC, no treatment cohort) Arm | N | Median | Median CI No treatment | 357 | NE | NE Landmark rates: 12 mo: No treatment 95.4% 24 mo: No treatment 88.4%

[Slide 1] Baseline ctDNA status may be a valuable prognostic factor in MIBC EFS and OS by pre-treatment ctDNA status (pooled)a [KM panel: EFS by pretreatment ctDNA status - ctDNA positive vs ctDNA negative; numbers at risk illegible] [KM panel: OS by pretreatment ctDNA status; numbers at risk illegible] NIAGARA Baseline: ctDNA Detection Was Prognostic for EFS [KM panel; EFS (combined arms); HR/CI text illegible] [ASCO logo] ….. And may have been an unstratified source of inbalance in CA017-078 | Gem-cis | Gem-cis nivolumab ctDNA positive at baseline | 51% | 59% Survival table Endpoint: EFS and OS by pre-treatment ctDNA status (pooled); NIAGARA EFS by baseline ctDNA ctDNA positive ctDNA negative [Slide 2] Does pre-cystectomy ctDNA clearance identify a group of patients for bladder preservation? EFS by ctDNA clearance status [KM curve: Survival probability vs Months 0-80; dashed vertical line with circled point around month ~24] Legend: C1D1 to C4D1 cleared; C1D1 to C4D1 not cleared No. at risk C1D1 to C4D1 cleared: [low-res; approx. 152 129 118 86 86 57 21 1 0 - low confidence] C1D1 to C4D1 not cleared: [low-res; approx. 46 21 14 9 9 4 4 1 0 - low confidence] Survival table Endpoint: EFS by ctDNA clearance status (C1D1 to C4D1) C1D1 to C4D1 cleared C1D1 to C4D1 not cleared [Slide 3] Could triplet therapy become a new standard in first line mUC? Drug class | Exemplar | ORR | Duration of response | DoR > 12 months Chemotherapy | Platinum + gemcitabine (EV-302) | 44% | 7 months | 35% Immune checkpoint inhibitor | Pembrolizumab (KN-052) | 29% | 30 months | 67% ADC + ICPI | EV + pembrolizumab (EV-302) | 68% | 23 months | 68% Vuky et al. Journal of Clinical Oncology Volume 38, Number 23 August 2020 Powles et al. Volume 36, Issue 10 – 2025 McGregor et al. ESMO 2024 [Slide 4] Can this be delivered in practice? Safety Summary Participants, n (%) | Cohort 7 (n = 21) Median treatment duration (range), months SG | 8.0 (0-18) EV | 7.5 (0-16) ZIM | 9.7 (0-20) Any TEAE/grade ≥ 3 TEAE | 21 (100)/14 (67) Any TRAE/grade ≥ 3 TRAE | 21 (100)/11 (52) [circled] Participants, n (%) | Cohort 7 (n = 21) TRAEs leading to any drug discontinuation | 7 (33) [circled] Discontinuation of SG | 4 (19) Discontinuation of EV | 6 (29) Discontinuation of ZIM | 1 (5) TEAE leading to death | 0 DLTa | 1/12 (8)

[Photo of presenter (badge: Laura Matouskova) beside poster; section signs: GERM CELL (TESTICULAR) CANCER AND PENILE CANCER; board 1217P; BEST POSTER rosette] #6234 Novel candidate genes implicated in the pathogenesis and cisplatin resistance of testicular germ cell tumors [logos: EuTTN, CLIP, FN Motol, Thomayer University Hospital, NICR, TTIP] [M]atouskova L1,2, Bakardjieva-Mihaylova V1, Zwyrtkova M1, Soffrova K1, Krivonoskova S2, Winkowska L1, Stuchly J1, Skvarova Kramarzova K1, Boublikova L1,2 1 Dept. of Pediatric Hematology and Oncology, Second Faculty of Medicine, Charles University and University Hospital in Motol and Homolka, Prague, Czech Republic; 2 Dept. of Oncology, 1st Faculty of Medicine, Charles University and Thomayer University Hospital, Prague, Czech Republic CONCLUSION - WES revealed distinct somatic alteration profiles between cisplatin-naive (primary tumors) and cisplatin-exposed (metastatic) TGCT samples - Significant proportion of somatic variants present in cfDNA at diagnosis was detectable in metastases and not in primary tumors, highlighting the potential of liquid biopsy for early identification of metastasis-associated molecular alterations. INTRODUCTION [...]h testicular germ cell tumors (TGCTs) respond well to cisplatin (cDDP)-based chemotherapy. [...] develop cDDP resistance, resulting in treatment failure with no effective alternative treatments [...] molecular mechanisms underlying cDDP resistance and TGCT pathogenesis remain incompletely [...] this issue, we performed whole-exome sequencing (WES) of primary tumors, metastases and [...] samples from patients with advanced TGCTs to identify genetic alterations that may contribute [...] and therapy resistance. MATERIALS AND METHODS Sample characteristics: Primary tumors 25; Metastases 7; cfDNA from diagnosis 6; gDNA (background control) 25 Whole exome sequencing - Library preparation: SureSelect XT HS Human All Exon + SureSelect XT HS Reagent kits (Agilent) - Sequencing: Illumina platform (NextSeq500, 2000) - Sequencing depth = 50 million read pairs/sample - Target mean coverage = 100x Variant filtering - Minimum tumor sequencing depth: >=15x - Minimum alternative allele depth: >=10x - Variant allele frequency (VAF): >=5% - Exclusion of variants detected in matched germline - Exclusion of non protein-altering variants - Selection of genes with qualifying alterations in >=3 of 25 patients for primary tumor analysis and >=2 of 7 patients for metastasis analysis - Evaluation of candidate genes based on their biological function, recurrence of detected variants across patient samples and their predicted functional consequences Figure panels: Somatic alterations in primary TGCTs; Somatic alterations in metastatic TGCTs; Figure 1: Oncoplot of top 20 altered genes shared with at least 3 patiens (primary tumors, left) and at least 2 patiens (metastases, right); Somatic variant distribution across primary tumors and metastases; Overlap of diagnostic cfDNA variants with primary tumors and metastases; Recurrent gene variants present in all sample types in more than one patient [gene-level figure text illegible] laura.matouskova@lfmotol.cuni.cz European Testicular Tumor Network [QR code]

[Slide 1] 4741RO Exploratory Analysis of Molecular Subtypes and Association With Treatment Outcomes in Patients With Muscle-Invasive Bladder Cancer From NIAGARA Matthew D. Galsky,1 Michiel S. van der Heijden,2 Ricky Joshi,3 James W. F. Catto,4 Joshua J. Meeks,5 Hiroyuki Nishiyama,6 Hikmat Al-Ahmadie,7 Toan Quang Vu,8 Lorenzo Antonuzzo,9 Tae Hwan Kim,10 Hiroaki Kikukawa,11 Vagif Atduev,12 Bernhard Eigl,13 Kazuo Nishimura,14 Yousef Zakharia,15 Jae Lyun Lee,16 Svetlana Ho,17 Huiling Xiong,18 Agrin Moeini,3 Thomas Powles19 1Icahn School of Medicine at Mount Sinai, New York, NY, USA; 2Netherlands Cancer Institute, Amsterdam, Netherlands; 3AstraZeneca, Barcelona, Spain; 4University of Sheffield, Sheffield, UK; 5Northwestern University, Feinberg School of Medicine, Chicago, IL, USA; 6University of Tsukuba, Tsukuba, Japan; 7Memorial Sloan Kettering Cancer Center, New York, NY, USA; 8Vietnam National Cancer Hospital, Hanoi, Vietnam; 9SODC Oncologia Medica-Azienda Ospedaliera-Universitaria Careggi, Florence, Italy; 10Kyungpook National University School of Medicine, Kyungpook National University Chilgok Hospital, Daegu, Republic of Korea; 11National Hospital Organization Kumamoto Medical Center, Kumamoto, Japan; 12Volga District Medical Center, Federal Medical-Biological Agency, Nizhny Novgorod, Russia; 13British Columbia Cancer Agency, Vancouver, BC, Canada; 14Osaka International Cancer Institute, Osaka, Japan; 15University of Iowa Hospitals and Clinics, Holden Comprehensive Cancer Center, Iowa City, IA, USA; 16Asan Medical Center, University of Ulsan College of Medicine, Seoul, Republic of Korea; 17AstraZeneca, Gaithersburg, MD, USA; 18AstraZeneca, Waltham, MA, USA; 19Barts Cancer Institute ECMC/BRC, Queen Mary University of London, Barts Health NHS Trust, London, UK [Slide 3] NIAGARA: Increased T Effector Infiltration and IFNγ Signalling are Associated with Enhanced Perioperative Durvalumab + NAC Benefit MIBC transcriptional subtypes (basal-squamous & luminal-infiltrated) are enriched for T effector infiltration and IFNγ signalling Association of gene signature reflective of chemotherapy or immune response with EFS ITT BEP Gene signature enrichment: Platinum sensitivity | Gemcitabine sensitivity | T effector infiltration | IFNγ signalling | Stromal TGF-β signalling High Sig score | Low Sig score EFS HR (log scale) 0.25 0.5 1 2 Favours D arm | Favours C arm Panels: T effector infiltration | IFNγ signalling | Stromal TGF-β signalling (immunosuppressive) High Sig score / Low Sig score Basal-squamous | Luminal | Luminal-infiltrated | Luminal papillary | Neuronal Left graph depicts EFS HR (square) and 95% confidence intervals (lines), and was assessed by the Cox proportional hazards model, with treatment as the only covariate and ties handled by the Efron approach. BEP, biomarker-evaluable population; C, comparator; D, durvalumab; EFS, event-free survival; ITT, intention-to-treat population; MIBC, muscle-invasive bladder cancer. [Slide 4] NIAGARA: Increased T Effector Infiltration and IFNγ Signalling are Associated with Enhanced Perioperative Durvalumab + NAC Benefit MIBC transcriptional subtypes (basal-squamous & luminal-infiltrated), enriched for T effector infiltration and IFNγ signalling, showed up to 16% higher pCR rates in the D arm vs the C arm MIBC subtype | Gemcitabine + platinum sensitivity | Stromal TGF-β signalling | T effector infiltration | IFNγ signalling | pCR rates (D arm vs C arm) Basal-squamous | High | High | High | High | 50% vs 34% (16%) Luminal | High | Low | Low | Low | 44% vs 44% Luminal-infiltrated | Low | High | High | High | 36% vs 25% (11%) Luminal papillary | Low | Low | Low | Low | 38% vs 37% Neuronal | High | Mid | Low | Low | 25% vs 47% C, comparator; D, durvalumab; MIBC, muscle-invasive bladder cancer; pCR, pathological complete response.

NIAGARA: EFS Benefit in MIBC Transcriptional Subtypes Addition of perioperative durvalumab to NAC appeared to improve EFS across most transcriptional subtypes, including basal-squamous Subtype | Patients with UC with DDHS | EFS rate at 24 months – C arm | Number of patients D arm | C arm | HR (95% CI) BEP n=402 | 15% | 66.7% | 207 | 195 | 0.65 (0.46–0.90) Basal-squamous (23%) n=91 | 26% | 67.8% | 47 | 44 | 0.47 (0.20–1.11) Luminal (34%) n=139 | 10% | 65.9% | 70 | 69 | 0.52 (0.28–0.94) Luminal-infiltrated (19%) n=77 | 15% | 67.3% | 47 | 30 | 0.67 (0.35–1.31) Luminal papillary (17%) n=68 | 10% | 78.3% | 34 | 34 | 1.35 (0.64–2.87) Neuronal (7%) n=27 | 11% | 44.4% | 9 | 18 | Not calculated^a EFS HR (log scale) 0.25 0.5 0.75 1.0 1.5 2.0 3.0 Favours D arm | Favours C arm P value P<0.001 Graph depicts EFS HR (square) and 95% confidence intervals (lines), and was assessed by the Cox proportional hazards model, with treatment as the only covariate and ties handled by the Efron approach. A significant p-value provides evidence against the null hypothesis that observations are equally likely to fall into each category, suggesting the category probabilities are not all equal, p-value obtained from a Chi-squared test. ^aNot calculated as there are fewer than 20 patients per arm. BEP, biomarker-evaluable population; C, comparator; CI, confidence interval; D, durvalumab; DDHS, divergent differentiation or histological subtype; EFS, event-free survival; HR, hazard ratio; MIBC, muscle-invasive bladder cancer; NAC, neoadjuvant chemotherapy; UC, urothelial carcinoma.

Is the combination hypothesis incorrect? pCR outcomes were consistent across trials KEYNOTE-866 [Bar chart: pCR Rate, % (95% CI)] Estimated difference:b 8.9% (95% CI, 2.5-15.1)* Pembro + cis + gem | Placebo + cis + gem CA017-078 pCR overall (co-primary endpoint)a 9.5% (95% CI 2.3-16.7) (99.91% CI -2.6 to 21.6)* 39.1 | 29.6 Overall ITT NIAGARA Re-analysis (Apr 2024) Odds ratio 1.60 (95% CI 1.23-2.08) nominal P=0.0005 Durvalumab arm N=533: 37.3% 199/533 95% CI 33.2-41.6 Comparator arm N=530: 27.5% 146/530 95% CI 23.8-31.6

Chemo-IO: KEYNOTE-866 and ENERGIZE negative Same ~9-10 pp pCR gain in all three trials - EFS and OS benefit only in NIAGARA Hazard ratio (95% CI) NIAGARA durvalumab | n = 1,063 | EFS | 0.68 (0.56-0.82) NIAGARA | OS | 0.75 (0.59-0.93) KEYNOTE-866 pembrolizumab | n = 907 | EFS | 0.88 (0.71-1.10) KEYNOTE-866 | OS | 1.00 (0.79-1.27) ENERGIZE nivolumab | n = 653 | EFS | 0.94 (0.71-1.25) ENERGIZE | OS* | 0.88 (0.64-1.21) * OS immature (ENERGIZE) pCR, % | GC (+ placebo) | GC + PD-(L)1 NIAGARA | 28 | 37 KN-866 | 36 | 44 ENERGIZE | 30 | 39 Niklas Klümper, X: @niklas_kluemper Powles T et al., NEJM 2024; Bedke J et al., ESMO 2026 (4735O, KEYNOTE-866); Necchi A et al., ESMO 2026 (4736O, ENERGIZE/CA017-078 final analysis)

[Slide 2] Pembrolizumab Adds to EV Perioperatively EV + Pembro Outperforms Either Drug Alone (EV-103 Cohort K, EV-303) EV-103 cohort K: EV + pembro > EV alone Longer PFS (median NR vs 8.0 mo) -> justified the combo in EV-302 Progression-free survival - EV + pembro (n = 76): median NR; EV monotherapy (n = 73): median 8.0 mo Overall survival - EV + pembro (n = 76): median 22.3 mo; EV monotherapy (n = 73): median 21.7 mo EV-303: EV + pembro > pembro alone > RC alone EFS HR 0.40 vs 0.71; combo now standard for cisplatin-ineligible MIBC Event-free survival - HR vs. control (95% CI)*: EV + pembro 0.40 (0.25-0.65) - Pembro 0.71 (0.47-1.07) EFS landmarks (12 / 24 / 36 mo): EV + pembro 77.2% / 75.8% / 69.5%; Pembro 66.4% / 55.8% / 49.6%; RC + PLND 55.0% / 41.0% / 37.4% Overall survival - HR vs. control (95% CI)*: EV + pembro 0.48 (0.29-0.80) - Pembro 0.76 (0.48-1.20) OS landmarks (12 / 24 / 36 mo): EV + pembro 83.3% / 78.4% / 74.5%; Pembro 78.2% / 66.9% / 61.4%; RC + PLND 74.6% / 62.6% / 55.7% Left: EV-103 cohort K curves redrawn by digitising published single-arm KM figures (approximate; censoring not shown); arms randomised but not formally compared. NR, not reached. Source: O'Donnell PH, et al. J Clin Oncol. 2023;41(25):4107-17. Right: KEYNOTE-905/EV-303 ad hoc analysis; blue = pembro, teal = EV + pembro, red = RC + PLND alone. Survival table Endpoint: EV-303 EFS and OS (ad hoc 3-arm analysis); EV-103 cohort K PFS/OS EV-303 EV + pembro EV-303 Pembro monotherapy EV-303 RC + PLND alone EV-103 cohort K EV + pembro | 76 | PFS NR; OS 22.3 mo EV-103 cohort K EV monotherapy | 73 | PFS 8.0 mo; OS 21.7 mo HR: EFS: EV+pembro 0.40, Pembro 0.71 vs control; OS: EV+pembro 0.48, Pembro 0.76 vs control HR CI: EFS 0.25-0.65 / 0.47-1.07; OS 0.29-0.80 / 0.48-1.20 Landmark rates: 12 mo: EFS EV+pembro 77.2% | EFS Pembro 66.4% | EFS RC+PLND 55.0% | OS EV+pembro 83.3% | OS Pembro 78.2% | OS RC+PLND 74.6% 24 mo: EFS EV+pembro 75.8% | EFS Pembro 55.8% | EFS RC+PLND 41.0% | OS EV+pembro 78.4% | OS Pembro 66.9% | OS RC+PLND 62.6% 36 mo: EFS EV+pembro 69.5% | EFS Pembro 49.6% | EFS RC+PLND 37.4% | OS EV+pembro 74.5% | OS Pembro 61.4% | OS RC+PLND 55.7% [Slide 3] PD-1 Blockade Shows Similar Benefit Across Perioperative Timings Perioperative Pembro (EV-303) vs Adjuvant Nivo and Pembro EV-303 - perioperative pembro; EFS - blue = pembro, red = RC + PLND alone [KM thumbnail; landmarks 66.4% / 55.0% (12 mo), 55.8% / 41.0% (24 mo), 49.0% / 37.4% (36 mo) - small print] CheckMate 274 - adjuvant nivo; DFS vs placebo (ITT) [KM thumbnail: Nivolumab, Placebo] AMBASSADOR - adjuvant pembro; DFS vs observation (ITT) [KM thumbnail: Pembrolizumab, Observation] Hazard ratio vs no active IO (95% CI) - EFS for EV-303; DFS for adjuvant trials EV-303 Perioperative pembro 0.71 (0.47-1.07) CheckMate 274 Adjuvant nivo 0.70 (0.55-0.90) AMBASSADOR Adjuvant pembro 0.73 (0.59-0.90) <- Favours PD-1 blockade | Favours control -> [Slide 4] ORIGINAL SLIDE FROM THE PRESENTER/AUTHORS ctDNA Is Prognostic; EV + Pembro Clears It More Often KEYNOTE-B15 Exploratory ctDNA Analysis (4734O) Legend: EV + pembro ctDNA+; EV + pembro ctDNA-; Cis + gem ctDNA+; Cis + gem ctDNA- EFS by baseline ctDNA - 24-mo EFS shown: [89.1%], [82.7%], 71.2%, 54.2% [upper two values low-res] EFS by postsurgery ctDNA - 24-mo EFS shown: [89.9%], [82.7%], 10.0%, 10.0% [upper two values low-res] pCR by ctDNA status - combined arms Baseline ctDNA-: pCR 58.7% (131/223); Non-pCR 41.3% (92/223) Baseline ctDNA+: pCR 36.3% (97/267); Non-pCR 63.7% (170/267) Presurgery ctDNA-: pCR 58.0% (211/364); Non-pCR 42.0% (153/364) Presurgery ctDNA+: pCR 2.8% (2/72); Non-pCR 97.2% (70/72) pCR among presurgery ctDNA- - by arm EV + pembro: ctDNA-: 87.0% (181/208); pCR 72.4% (131/181); Non-pCR 27.6% (50/181) Cis + gem: ctDNA-: 80.3% (183/228); pCR 43.7% (80/183); Non-pCR 56.3% (103/183) Authors' conclusions - ctDNA is strongly prognostic in both arms: ctDNA- -> better EFS and pCR - EV + pembro beat cis + gem on EFS regardless of ctDNA status - Clearance by presurgery -> better EFS; more frequent with EV + pembro - Among clearers, EV + pembro still had better EFS ctDNA-negative status was linked to better outcomes in both arms Higher pCR at baseline and presurgery; longer EFS at baseline and postsurgery (exploratory analysis)

[Slide 1] Cisplatin-ineligible MIBC: an unmet need was met Primary Endpoint: EFSa by BICR ITT Population | EV + pembro (N = 170) | Control (N = 174) Events, n (%) | 48 (28.2) | 95 (54.6) Median, mo (95% CI) | NR (37.3–NR) | 15.7 (10.3–20.5) HR 0.40, 95% CI 0.28–0.57, one-sided P <.0001* 12 months: 77.8% vs 55.1%; 24 months: 74.7% vs 39.4% • 29/174 pts (16.7%) in the control arm received adjuvant nivolumab as of data cutoff per local guidelines/prescribing information No. at Risk EV + pembro 170 140 116 73 56 42 35 16 3 0 0 Control 174 116 84 48 33 22 14 9 1 0 0 KEYNOTE-905/EV-303 EVP vs resection alone MIBC, n=344 Neoadjuvant: 3 cycles EVP Adjuvant: 6 cycles (total 9) Vulsteke et al, ESMO&NEJM 2025 Survival table Endpoint: EFS by BICR (ITT) Arm | N | Median | Median CI EV + pembro | 170 | NR | 37.3–NR Control (resection alone) | 174 | 15.7 | 10.3–20.5 HR: 0.40 HR CI: 0.28–0.57 P: one-sided P<.0001 Landmark rates: 12 months: EV + pembro 77.8% | Control 55.1% 24 months: EV + pembro 74.7% | Control 39.4% [Slide 2] KN905/EV303 & EV304/KN-B15: a paradigm shift for patients with neoadjuvant MIBC Primary Endpoint: EFSa by BICR ITT Population [KEYNOTE-905 KM, same as prior slide: HR 0.40, 95% CI 0.28–0.57, one-sided P <.0001; 12-mo 77.8% vs 55.1%; 24-mo 74.7% vs 39.4%] KEYNOTE-905/EV-303 EVP vs resection alone MIBC, n=344 Neoadjuvant: 3 cycles EVP Adjuvant: 6 cycles (total 9) Vulsteke et al, ESMO&NEJM 2025 pCR: Estimated differencea 48.3% (95% CI 39.5–56.5) one-sided P <.000001* (EV + pembro vs Control) Primary Endpoint: EFS by BICR ITT Population EV + pembro N=405 | Cis + gem N=403 Events, n (%) | [illegible] | [illegible] Median, mo (95% CI) | [illegible] | [illegible] HR 0.53, 95% CI 0.41–0.70, 1-sided P <.0001* 12 months: 86.0% vs 75.4%; 24 months: 79.4% vs 66.2% No. at Risk EV + pembro 405 351 317 303 257 154 100 63 23 0 Cis + gem 403 350 286 259 216 131 70 45 16 0 KEYNOTE-B15/EV-304 EVP vs cisplatin/gemcitabine MIBC, n= 808 Neoadjuvant: 4 cycles EVP Adjuvant: 5 cycles (total 9) pCR: Estimated differencea 23.4% (95% CI 16.7–29.8) 1-sided P <.0001* (EV + pembro vs Cis + gem) Galsky et al, ASCO-GU&NEJM 2026 Survival table Endpoint: EFS by BICR (ITT) — KEYNOTE-B15/EV-304 panel EV + pembro | 405 Cis + gem | 403 HR: 0.53 HR CI: 0.41–0.70 P: 1-sided P<.0001 Landmark rates: 12 months: EV + pembro 86.0% | Cis + gem 75.4% 24 months: EV + pembro 79.4% | Cis + gem 66.2% [Slide 3] Patients • Adults aged ≥18 years • MIBC (cT2-T4aN0/1M0 or cT1N1M0) by local assessment • Ineligible for or refused cisplatin • UC or UC with divergent differentiation or histological subtypes • Evaluated and confirmed for cystectomy R (1:1:1) N=695 D+EV N=233: Neoadjuvant Durvalumab Q3W × 3 cycles + Enfortumab vedotin Q3W × 3 cycles → RC → Adjuvant Durvalumab Q4W × 9 cycles D+T+EV N=233: Durvalumab Q3W × 3 cycles + Tremelimumab Q3W × 2 cycles + Enfortumab vedotin Q3W × 3 cycles → RC → Durvalumab Q4W × 9 cycles + Tremelimumab Q4W × 1 cycle Control N=229: No treatment, Straight to RC → Observation/locally approved treatment • D+EV positive for EFS and OS • D+T+EV positive for EFS • Another great result for bladder cancer patients!! VOLGA: EFS by BICR for Durvalumab+EV vs Control – ITT D+EV N=233 | Control N=229 Events, n (%) | [illegible] | [illegible] Median EFS, months (95% CI) | NR ([illegible]) | [illegible] HR (95% CI) | 0.51 (0.38–0.67) Stratified log-rank P value | <0.0001 12 months: 73% vs 50%; 24 months: 64% vs 45% Median follow-up: 23 months VOLGA: OS for Durvalumab+EV vs Control – ITT D+EV N=233 | Control N=229 Deaths, n (%) | [illegible] | [illegible] Median OS | [illegible] | [illegible] HR (95% CI) | 0.58 (0.40–0.83) Stratified log-rank P value | 0.0027 12 months: 87% vs 81%; 24 months: 78% vs 69% Median follow-up: [illegible] months No. of patients at risk [illegible] Survival table Endpoint: VOLGA EFS by BICR and OS, D+EV vs control (ITT) Durvalumab + EV | 233 Control | 229 HR: EFS 0.51; OS 0.58 HR CI: EFS 0.38–0.67; OS 0.40–0.83 P: EFS <0.0001; OS 0.0027 (stratified log-rank) Landmark rates: 12 months (EFS): D+EV 73% | Control 50% 24 months (EFS): D+EV 64% | Control 45% 12 months (OS): D+EV 87% | Control 81% 24 months (OS): D+EV 78% | Control 69%

[Slide 1] Hazard ratio (95% CI) NIAGARA durvalumab | n = 1,063 EFS 0.68 (0.56-0.82) OS 0.75 (0.59-0.93) KEYNOTE-866 pembrolizumab | n = 907 EFS 0.88 (0.71-1.10) OS 1.00 (0.79-1.27) ENERGIZE nivolumab | n = 653 EFS 0.94 (0.71-1.25) OS* 0.88 (0.64-1.21) * OS immature (ENERGIZE) pCR, % NIAGARA 28 | 37 KN-866 36 | 44 ENERGIZE 30 | 39 GC (+ placebo) GC + PD-(L)1 [Slide 2] Event-free survival HR 0.53 (0.41-0.70) median NR vs 48.5 mo Event-free Survival (%) 86.0 79.4 75.4 66.2 EV–pembro Cis–gem Months No. at Risk Overall survival HR 0.65 (0.48-0.89) 2-yr 86.9% vs 81.3% Overall Survival (%) 93.1 86.9 89.6 81.3 EV–pembro Cis–gem Months No. at Risk Survival table Endpoint: Event-free survival EV–pembro | 405 | NR Cis–gem | 403 | 48.5 mo HR: 0.53 HR CI: 0.41-0.70 12-mo EFS 86.0% vs 75.4% 24-mo EFS 79.4% vs 66.2% [Slide 3] Event-free survival, % Months 97.3% 94.6% 93.9% 90.7% 69.4% 58.3% 65.7% 53.4% Median, Disease-free Survival (95% CI) mo EV–Pembro (pCR): NR (NR to NR) Cis–Gem (pCR): NR (NR to NR) EV–pembro (no pCR): 37.1 (22.7 to NR) Cis–gem (no pCR): 31.2 (19.1 to 43.3) pCR: HR 0.54 (95% CI 0.27 to 1.07) No pCR: HR 0.90 (95% CI 0.68 to 1.21) No. at Risk Survival table Endpoint: Event-free survival by pCR status Arm | N | Median | Median CI EV–pembro (pCR) | 226 | NR | NR to NR Cis–gem (pCR) | 131 | NR | NR to NR EV–pembro (no pCR) | 179 | 37.1 mo | 22.7 to NR Cis–gem (no pCR) | 272 | 31.2 mo | 19.1 to 43.3 HR: pCR: 0.54; No pCR: 0.90 HR CI: pCR: 0.27 to 1.07; No pCR: 0.68 to 1.21 12-mo: EV-pembro pCR 97.3%, Cis-gem pCR 93.9%, EV-pembro no pCR 69.4%, Cis-gem no pCR 65.7% 24-mo: EV-pembro pCR 94.6%, Cis-gem pCR 90.7%, EV-pembro no pCR 58.3%, Cis-gem no pCR 53.4% [Slide 4] % of patients Cystectomy alone Pembrolizumab EV + pembrolizumab 2-year EFS 41 | 56 | 76 pCR (pT0N0) 10 | 25 | 59 Downstaging (<pT2N0) 15 | 34 | 66 0.71 EFS HR pembro vs control (0.47-1.07) 0.40 EFS HR EV-P vs control (0.25-0.65) 15.7 / 28.7 / NR median EFS (months): control / pembro / EV-P

[Slide 1] 1047P Patient-Reported toxicity precedes clinician-graded DLTs and tolerability events in Phase I Trials: A prospective DITEP cohort K. Ouali^1, A. Yazigi^2, S. Jetley^2, A. Hollebecque^1, Y. Loriot^1, M. Sakkal^1, A. Gazzah^1, J.-M. Michot^1, R. Bahleda^1, T. Hueso^1, K. Amrane^1, A. Patrikidou^1, S. Postel-Vinay^1, F.-X. Danlos^1, V. Ribrag^1, A. Marabelle^1, A. Italiano^1, J. Ribeiro^1, C. Ferte^2, C. Massard^1 ^1Drug Development Department, Institut Gustave Roussy, Villejuif, France, ^2Resilience Care, Resilience, Paris, France BACKGROUND [partially obscured by presenter] RESULTS Among 608 pts included in phase 1 trials at DITEP between September 2024 and November 2025, 431 with evaluable PRO-CTCAE data were included in the analysis. Median number of completed questionnaire was 11 [95% CI, 10-30]. Population, n (%): n 431; Age, median [Q1,Q3] 62.0 [54.0,69.0]; Gender: Female 242 (56.1), Male 189 (43.9); Cancer stage: Localized 10 (2.3), Locally advanced 29 (6.7), Metastatic 392 (90); [Type] of drug: Targeted therapy (TT) 182 (42.2), Immunotherapy (IO) 156 (36.2), ADC 68 (15.8), IO+ Chemo 10 (2.3), IT IO 6 (1.4), IO + RT 3 (0.7), IO + TT 3 (0.7), IO + ADC 1 (0.2), Radioligand 1 (0.2), TT + RT 1 (0.2) Figure 1: Percentage of pts experiencing PRO-CTCAE G>=3 events, by symptom [bar chart; values largely illegible] Figure 2: Frequency of PRO-CTCAE G>=3 events by symptom [stacked bar chart; values largely illegible] Table 2: Time to first DLT, tolerability event and PRO-CTCAE G>=3 event (Kaplan-Meier) Event | Median Time-to-event | Event-free Probability at Day 30 DLT | NR | 0.96 Tolerability | NR | 0.92 PRO-CTCAE G>=3 | 23.0 days, 95%CI [17.0 - 35.0] | 0.47 Figure 3: Kaplan-Meier estimates of time to first PRO-CTCAE G>=3 event, DLT and TE - DLT rate was 4.6% (n=20) with a median time to DLT of 14 days [95% CI, 7-21] among pts with a DLT. 12.5% pts had a dose interruption, 8.4% a dose reduction and 6.5% discontinued treatment due to toxicity making a total of 90 tolerability events (20.9%). Median time to tolerability event was 42 days [95% CI, 31-63] among pts with a TE. - Grade >=3 PRO-CTCAE events occurred earlier, preceding tolerability events by a median of 11 days [95% CI, 8-21] and DLTs by 2.5 days [95% CI, -3 to 12]. - Incorporating grade >=3 PRO-CTCAE data improved prediction of safety/tolerability events (AUROC 0.65 vs. 0.61). The PRO-CTCAE G>=3 flag was an independent predictor beyond clinical variables (adjusted OR 5.9 [95% CI, 2.2-15.5], p<0.001). Figure 4: Mean cross-validated AUROC (±1 SD) of the model including the PRO-CTCAE G>=3 flag (ablative analysis) [CONC]LUSION: [PRO]-CTCAE events preceded tolerability events and added independent predictive information beyond clinical variables (adjusted OR 5.9). Continuous PRO-CTCAE monitoring [...] both earlier and complementary toxicity signals to refine dose-escalation decisions and RP2D identification. Contact: kaissa.ouali@gustaveroussy.fr [Presenter Kaissa Ouali standing beside poster] [Slide 2] 4465P 4465P - Prognostic value of the Decipher prostate genomic classifier (GC) and neuroendocrine (NE) gene expression signatures in real-world de novo metastatic (M1) prostate cancer Alice Bernard-Tessier^1, E. Walker^2, J. A. Proudfoot^2, Qi Joselove Xu^2, Yang Liu^2, Elai Davicioni^2, Himisha Beltran^3 ^1Gustave Roussy, Departement de medecine oncologique, Villejuif, France. ^2Veracyte Inc, San Diego, USA. ^3Department of Medical Oncology, Dana-Farber Cancer Institute, Boston, MA, USA INTRODUCTION The prognostic impact of molecular features in de novo metastatic (M1) hormone-sensitive prostate cancer (mHSPC) remains largely underexplored in real-world settings. The Decipher Genomic Classifier (GC) predicts benefit from docetaxel (Grist et al., Cell 2025; Sweeney et al., ASCO 2026). In addition, PEACE-1 data showed that approximately 25% of de novo mHSPC presented neuroendocrine markers on IHC (Pobel et al., ESMO 2024), suggesting that transcriptomic features beyond GC may carry prognostic information. Objective: to evaluate the prognostic value of GC and of validated gene expression signatures for overall survival in a large real-world cohort of de novo mHSPC, independently of clinical variables. METHODS Population. US patients with a diagnosis of prostate cancer (PCa) and a GC test performed on the diagnostic biopsy, with metastases diagnosed within 90 days before or after the PCa diagnosis (de novo M1). Data. Clinical outcomes were obtained by linkage to electronic health record data. Transcriptome-wide gene expression profiling was performed in a clinical laboratory (Veracyte, San Diego, CA). Signatures. GC and validated signatures: AR activity, small cell/NE-like, PTEN, TP53, PAM50, PSC; MKI67 and individual AR-dependent and NE-related genes. Statistics. Associations with overall survival (OS) from diagnostic biopsy were assessed using Cox proportional hazards models, with GC and signatures dichotomized using published thresholds, MKI67 and individual genes as top 20% vs bottom 80%. Multivariable analyses (MVA) were adjusted for age, T stage, Gleason grade group and PSA. OS was estimated using the Kaplan-Meier method and compared with the log-rank test. RESULTS 1 2,303 de novo mHSPC patients: aggressive disease, most with high-risk GC 209,649 pts with PCa and GC testing -> 2,303 pts (1.1%) with de novo M1 disease (M1 within 90 days of PCa diagnosis) Most patients had aggressive disease: 55% had Gleason grade group 4-5 and 76% had high-risk GC, including 55% very high risk (score >0.85). A small cell NE-like transcriptomic signature was found in 6% of tumors. 2 Higher GC is independently associated with worse OS On MVA, higher GC was independently associated with worse OS (HR 1.37 per 0.1-unit increase; p<0.001). Very high vs lower GC risk conferred a 3-fold increased risk of death (HR 3.31; p<0.001). 3 GC and TP53 combine to identify the highest-risk patients Combined GC risk (very high vs lower) and TP53 stratification demonstrated incremental risk, with 5-year OS rates of 93.3% for lower GC + wildtype, 93.3% for lower GC + mutant, 67.6% for very high GC + wildtype, and 58.7% for very high GC + mutant. 4 NE-like, TP53, PTEN, MKI67 and low AR activity signatures predict worse OS On MVA, the small cell NE-like signature (HR 2.57; 95% CI 1.56-4.24), the TP53 mutant signature (HR 2.31; 1.47-3.63), the PTEN inactivity signature (HR 1.58; 1.06-2.36), lower AR activity (HR 2.37; 1.61-3.48) and higher MKI67 expression (top 20%; HR 1.82, 1.24-2.67) were associated with worse OS. No deaths among the 30 patients (1%) classified the NE classified according to Beltran [partially obscured] 5 Neuroendocrine marker expression is associated with worse OS On MVA, high KLK2 expression was associated with better OS (HR 0.46; 95% CI 0.25-0.87). Among NE-related genes, high DLL3 (HR 1.54; 1.03-2.30), CEACAM5 (HR 1.63; 1.08-2.47) and EZH2 (HR 1.55; 1.04-2.32) expression were associated with worse OS. Chromogranin A, SEZ6, INSM1 and synaptophysin showed similar trends (HR 1.12-1.37) without reaching significance. [Table 1, Fig. 1-3 and forest plot: small text largely illegible] [CONC]LUSION [partially obscured by presenter] Disclosures. Study funded by Veracyte. Genomic data and assay provided by Veracyte (San Diego, CA). [Presenter Alice Bernard-Tessier standing beside poster] [Slide 3] 3274P Predicting informative liquid biopsy in [advan]ced thoracic malignancies to optimize ctDNA-only molecular profiling at diagnosis Marianna Peroni^1, Daniela Miliziano^1, Francesca Colamartini^1, Noemie Simon-Tillaux^2, Olfa Trabelsi-Grati^3, Maxime Borgeaud^1, Giulia La Cava^1, Lodovica Zullo^1, Arianna Marinello^1, Adrien Rousseau^1, Pernelle Lavaud^1, Pamela Abdayem^1, Maxime Frelaut^1, Jordi Remon^1, Fabrice Barlesi^4, Benjamin Besse^1, Antoine Italiano^5, David Planchard^1, Etienne Rouleau^3, Mihaela Aldea^4 [affiliations small] BACKGROUND In advanced thoracic malignancies, liquid biopsy (LB) accelerates molecular diagnosis and treatment initiation. Since its sensitivity remains variable and concurrent tissue NGS (tNGS) raises cost-efficiency concerns, identifying clinical predictors of LB positivity is essential to define when LB alone may suffice. Can we predict, using baseline clinical features available at first visit, which patients are likely to have an informative baseline LB? METHODS We retrospectively included patients with advanced thoracic malignancies and available baseline LB at Gustave Roussy (2021-2026). LB samples were analysed through FoundationOne Liquid CDx. Baseline clinical features (age, sex, smoking habit (pack-years), stage at diagnosis, metastatic burden/sites, ECOG PS) were analysed through multivariable logistic regression. The predictive model was developed using clinically plausible variables and internally validated using bootstrap optimism correction. RESULTS 516 baseline available LBs; 71 (14%) non-informative -> Multivariable logistic regression -> Predictive model development (variables: sex, pack-years, PS, stage at diagnosis, liver and bone metastases) -> Discrimination and internal validation (complete-case analysis on 417 patients) -> Potential clinical impact (using a >=90% predicted probability threshold) INDEPENDENT PREDICTORS [forest plot, Multivariable logistic regression; odds ratio values small/partly illegible]: Age 1.00 (0.97-1.03), p = 0.927; Female 0.48 (0.25-0.94), p = 0.031; Pack-years 1.03 (1.01-1.05), p = 0.001; PS 2-4 2.46 (0.96-6.18), p = 0.055; Stage IV at diagnosis 1.48 (0.67-3.29), p = 0.333; N. of metastatic sites 1.02 (0.75-1.39), p = 0.890; Liver metastases 3.12 (0.97-10.04), p = 0.057; Brain metastases 1.45 (0.70-2.98), p = 0.320; Bone metastases 2.46 (1.14-5.33), p = 0.022 PREDICTIVE MODEL DISCRIMINATION [ROC curve]: Apparent AUC = 0.763 (95% CI 0.700-0.825); Optimism-corrected AUC = 0.736 (95% CI 0.675-0.800) PREDICTIVE MODEL CALIBRATION [calibration plot]: Apparent; Optimism-corrected = 0.87; O/E = 1.00 CLINICAL IMPACT 417 patients included Predicted probability >=90% n = 216 (52%) -> Liquid biopsy first -> 11/216 (5.1%) Non-informative LB (missing one KRAS non-G12C mut and one ROS1 fusion); 205/216 (PPV 94.9%) True informative LB -> 49% Potential reduction in concurrent tissue NGS (205/417) Predicted probability <90% n = 201 (48%) -> Liquid biopsy + tissue NGS CONCLUSION A simple prediction model using clinical features available at first oncology visit may identify pts with high probability of informative LB, potentially sparing more than 40% of concurrent tNGS and optimizing resources. marianna.peroni@gustaveroussy.fr I have no COI to declare [Presenter Marianna Peroni standing beside poster] [Slide 4] [Poster number partially obscured] Role of metastasis directed therapy (MDT) in hormone sensitive [oligometa]static prostate cancer (omHSPC): early toxicity and efficacy data from the phase III randomized PEACE 6 Oligo-PRESTO trial Y. Ghannam, C. Bargain, S. Supiot, J. Chamois, P. Ronchin, G. Loos, P. Sargos, J. Khalifa, L. Cartier, G. Crehange, G. Gravis, J.-C. Trone, P.E. Brachet, I. Latorzeff, D. Pasquier, L. Lefebvre, G. Bera, V. Schartner, S. Foulon, P. Blanchard [affiliations small/illegible] [OBJECTIVE and participant flow panels partially obscured by presenter; visible: ARM B SOC* n = 283] MDT MODALITY: SBRT data on 400 lesions (245 patients): 71,5% in 3 fractions / 27,5% in 5 fractions STUDY POPULATION Baseline characteristics were well balanced between randomized arms. Characteristic | Overall N=565 | Arm A (SOC+MDT) N=282 | Arm B (SOC) N=283 Age, years - median (IQR) | 71 (65-75) | 71 (65-75) | 72 (66-76) ECOG performance status 0 | 486 (86.0%) | 241 (85.5%) | 245 (86.6%) Number of metastases: 1-2 | 421 (74.5%) | 211 (74.8%) | 210 (74.2%) 3-5 | 144 (25.5%) | 71 (25.2%) | 73 (25.8%) Disease presentation: De novo | 289 (51.2%) | 145 (51.4%) | 144 (50.9%) Recurrent | 276 (48.8%) | 137 (48.6%) | 139 (49.1%) Metastatic sites: Bone | 558 (98.8%) | 279 (98.9%) | 279 (98.6%) Lymph nodes | 209 (37.0%) | 104 (36.9%) | 105 (37.1%) Staging imaging: Choline PET/CT | 354 (62.7%) | 178 (63.1%) | 176 (62.2%) PSMA PET/CT | 211 (37.3%) | 104 (36.9%) | 107 (37.8%) Additional systemic treatment: None | 85 (15.0%) | 42 (14.9%) | 43 (15.2%) Docetaxel | 2 (0.4%) | 1 (0.4%) | 1 (0.4%) New-generation hormonal therapy | 478 (84.6%) | 239 (84.8%) | 239 (84.5%) Baseline PSA, ng/mL - median (IQR) | 6 (1-22) | 7 (2-22) | 6 (1-22) Data are n (%) unless otherwise stated. IQR, interquartile range. ACUTE SAFETY WITHIN 6 MONTHS Most frequent acute AEs related to hormonal therapy: hot flush, asthenia, pollakiuria, fatigue and diarrhoea. Most were grade 1-2, with no apparent excess following SBRT. Acute AEs within 6 months by randomized arm, n (%) Maximum grade | ARM A (SOC + MDT) n=279 | ARM B (SOC) n=282 Any acute AE | 253 (90.7) | 277 (98.2) Grade 1 | 92 (33.0) | 108 (38.3) Grade 2 | 115 (41.2) | 126 (44.7) Grade 3 | 39 (14.0) | 37 (13.1) Grade 4 | 3 (1.1) | 4 (1.4) Grade 5 | 4 (1.4) | 2 (0.7) Modified ITT: treated patients by randomized arm; maximum grade per patient. Quality of life analysis showed no impact of MDT on EPIC subdomains (urinary, bowel, sexual, hormonal), nor on pain scores (BPI, below) Brief Pain Inventory Scores in patients with bone pain at baseline [two line charts: Brief Pain interference score; Brief Pain severity score; SOC + MDT vs SOC; Baseline, Week 6, Month 3, Month 6; numbers of patients at visit small] EARLY PSA ASSESSMENT At 6 months, PSA remained low in both arms: SOC + MDT median 0 (mean 0.3) vs SOC alone median 0 ng/mL (mean 0.1); n=525, p=0.27 CONCLUSION No apparent excess of high-grade acute toxicity was observed with SBRT. Longer follow-up is required to determine efficacy. Acknowledgements, Funding: French Government grant (PHRC-K) and Johnson and Johnson; Disclosures [small] corresponding author email: youssef.ghannam@gustaveroussy.fr [Presenter Youssef Ghannam standing beside poster]

[Slide 1] ...e going to talk about EV-P? ...rm were alive vs 24.6% in the... Events/N | Median OS (95% CI), mo | Stratified HR (95% CI) 240/442 | 33.6 (26.6, 39.8) | 0.53 (0.45, 0.63) 329/444 | 15.9 (13.6, 18.3) Curve labels: 44.0% EV+P; 24.6% Chemotherapy Number at risk rows partially visible, [illegible] Powles T. ASCO2026 OncoDaily Grand Rounds During ESMO 2026 Survival table Endpoint: Overall survival Arm | N | Median | Median CI [row 1, arm label cropped] | 442 | 33.6 | 26.6, 39.8 [row 2, arm label cropped] | 444 | 15.9 | 13.6, 18.3 HR: 0.53 HR CI: 0.45, 0.63 Landmark rates: [not visible]: EV+P 44.0% | Chemotherapy 24.6% [Slide 3] What are we going to talk about? Cis-Eligible: Niagara (reported 2024; FDA approval March 2025; EU approval July 2025) | GC x 4 / GC + Durva x 4 | Radical Cystectomy and PLND | Observation / Durva x 8 EV-304/KN-B15 (reported: Feb 2026) | EV + Pembro x 4 / GC x 4 | Radical Cystectomy and PLND | EV x 5 + Pembro x 13 / Observation Cis-Ineligible: EV-303/KN-905 (reported 2025; FDA approval Nov 2025) | EV + Pembro x 3 / No preop therapy | Radical Cystectomy and PLND | EV x 6 + Pembro x 14 / Observation VOLGA | Durva + EV x 3 / Durva/Treme + EV x 3 / No preop therapy | Radical Cystectomy and PLND | Durva x 9 / Treme x 1 + Durva x 9 / Per physician choice ...ADDER ...NCER EDITION | OncoDaily Grand Rounds During ESMO 2026

Mechanisms of resistance to nivolumab in Older vs Younger Adults with metastatic Renal Cell Carcinoma Renée Maria Saliby, Wassim Daoud Khatoun, Mustafa Saleh, Pablo Moura Barrios, Eddy Saad, Marc Eid, Liliana Ascione, Marc Machaalani, Clara Steiner, Toni Choueiri, David Braun BACKGROUND • RCC incidence rises with age -> Older Adults (OA) are a growing share of patients receiving immune checkpoint inhibitors (ICI). • Whether age influences immune response remains unclear. RESULTS Table 1. Baseline characteristics of patients with resistance to nivolumab (PD, or SD with PFS <6 months) by age group. Figure 1. OA have higher tumor mutational burden (N = 75 OA, 73 YA). Figure 2. OA have higher ERV expression (mean log2 expression across all ERV loci). (N = 35 OA, 34 YA). Figure 3. Immune composition (CIBERSORTx) was similar between groups, except for fewer memory activated CD4+ T cells in OA (N= 35 OA, 34 YA). CONCLUSIONS The mechanisms of resistance to PD-1 inhibition in mRCC may differ by age. In older adults, resistance to nivolumab may reflect age-associated immune dysfunction despite preserved tumor immunogenicity, as suggested by higher TMB and ERV expression but reduced memory activated CD4+ T-cell infiltration. In contrast, resistance in younger adults may be driven more by tumor-intrinsic programs, including CNA-associated alterations and proliferative signaling. These data support further study of age-specific resistance biology and age-tailored strategies in mRCC. @BraunMDPh @DrChoueiri @ReneeSaliby FPN: 3628P

[Slide 1] Cohort I of KEYNOTE-365 (NCT02861573) Key Eligibility Criteria - Aged >=18 years - Pathologicallya confirmed treatment-emergentb or de novo metastatic NEPC per central histology reviewc - No prior platinum-based chemotherapy for NEPC - ECOG PS 0 or 1 N = 70, Rd 1:1 or 3:1 n = 40: Pembrolizumab 200 mg IV Q3W x <=35 cycles + Carboplatin AUC IV 5 Q3W x 4-6 cycles + Etoposide 100 mg/m2 IV d1-3 Q3W x 4-6 cycles n = 30: Carboplatin AUC 5 IV Q3W x 4-6 cycles + Etoposide 100 mg/m2 IV d1-3 Q3W x 4-6 cycles -> Treatment continued until unacceptable toxicity, PD, or participant or physician decision to withdraw Stratification factor - ECOG PS (0 vs 1) Primary endpoints included - ORR per RECIST v1.1 by blinded independent central review (BICR) - Safety Secondary endpoints included - DOR per RECIST v1.1 by BICR - rPFS per PCWG3-modified RECIST v1.1 by BICR - OS Post hoc analysis - ORR, rPFS, and OS by PD-L1 status Median study follow-up (time from randomization to data cutoff date [November 19, 2025]): 15.4 months (range, 0.8-43.5) aBy morphology and immunohistochemistry. bParticipants had progression <=6 months before screening and prior treatment with ADT for metastatic disease. Prior treatment with <=2 chemotherapy regimens for mCRPC and <=2 prior NHA for mCRPC was allowed. Docetaxel for mHSPC was allowed in addition to docetaxel for mCRPC. If docetaxel chemotherapy was used more than once, it was considered 1 therapy. cParticipants had one of the following: small cell carcinoma, large cell neuroendocrine carcinoma, or mixed (adenocarcinoma with either small or large cell neuroendocrine carcinoma). Participants with adenocarcinoma with neuroendocrine features were not eligible. Neuroendocrine histology according to the Epstein criteria (Epstein JI et al. Am J Surg Pathol. 2014;38:756-767) was confirmed by central histology review. dRandomization was 1:1 initially, then updated to 3:1 after a protocol amendment effective February 7, 2025. [Slide 2] Response Summary per RECIST v1.1 by BICR Among Participants With RECIST-Measurable Disease Columns: Total population - Pembro + chemo n = 36 | Chemo n = 27; PD-L1 subgroup CPS >=1 - Pembro + chemo n = 15 | Chemo n = 11; CPS <1 - Pembro + chemo n = 20 | Chemo n = 15 ORR, % (95% CI)a: 28 (14-45) | 15 (4-34) | 40 (16-68) | 18 (2-52) | 15 (3-38) | 13 (2-40) DCR, % (95% CI)b: 50 (33-67) | 48 (29-68) | 67 (38-88) | 55 (23-83) | 35 (15-59) | 47 (21-73) Best overall response, n (%) CR: 0 | 0 | 0 | 0 | 0 | 0 PR: 10 (28) | 4 (15) | 6 (40) | 2 (18) | 3 (15) | 2 (13) SD of any duration: 8 (22) | 9 (33) | 4 (27) | 4 (36) | 4 (20) | 5 (33) SD >=6 mo: 3 (8) | 1 (4) | 1 (7) | 0 | 2 (10) | 1 (7) PD: 13 (36) | 6 (22) | 3 (20) | 2 (18) | 10 (50) | 3 (20) No assessmentc: 5 (14) | 7 (26) | 2 (13) | 3 (27) | 3 (15) | 4 (7) Not evaluabled: 0 | 1 (4) | 0 | 0 | 0 | 1 (27) DOR, median (range), mo: 6.2 (2.0+ to 10.2+) | 4.1 (2.8+ to 6.2) | 8.3 (3.2+ to 10.2+) | 4.9 (3.5 to 6.2) | 4.1 (2.0+ to 4.1) | 4.1 (2.8 to 4.1+) "+" indicates no PD at the time of the last disease assessment. aIncludes participants with confirmed CR or PR. bIncludes participants with CR, PR, or SD of any duration. cIncludes participants without postbaseline assessments as of the data cutoff date, including missed treatment, discontinued treatment, or death before first postbaseline assessment. dIncludes participants who had a baseline assessment but no postbaseline assessment at the data cutoff date. Data cutoff date: November 19, 2025. [Slide 3] rPFS per PCWG3-Modified RECIST v1.1 by BICR Total population [table] Events, n (%) | Median (95% CI), mo | HR (95% CI) Total population Pembro + chemo (n = 40): 23 (58) | 5.1 (4.1-6.3) | 0.66 (0.36-1.21) Chemo (n = 30): 23 (77) | 4.0 (2.1-5.0) PD-L1 CPS >=1 Pembro + chemo (n = 15): 8 (53) | 6.0 (3.9-NR) | 0.40 (0.16-1.01) Chemo (n = 12): 11 (92) | 3.8 (0.4-5.0) PD-L1 CPS <1 Pembro + chemo (n = 23): 14 (61) | 5.1 (4.0-6.3) | 0.87 (0.38-1.97) Chemo (n = 16): 11 (69) | 4.1 (1.9-7.0) Curve label at 6 months: 49% (pembro + chemo) / 26% (chemo) Y-axis: rPFS, %; X-axis: Months No. at risk (0,2,4,6,8,10,12,14,16 mo) Pembro + chemo: 40 32 22 12 3 1 0 0 0 Chemo: 30 20 13 5 2 0 0 0 0 Data cutoff date: November 19, 2025. Survival table Endpoint: rPFS per PCWG3-modified RECIST v1.1 by BICR, total population (subgroups in table) Arm | N | Median | Median CI Pembro + chemo (total) | 40 | 5.1 | 4.1-6.3 Chemo (total) | 30 | 4.0 | 2.1-5.0 Pembro + chemo (CPS>=1) | 15 | 6.0 | 3.9-NR Chemo (CPS>=1) | 12 | 3.8 | 0.4-5.0 Pembro + chemo (CPS<1) | 23 | 5.1 | 4.0-6.3 Chemo (CPS<1) | 16 | 4.1 | 1.9-7.0 HR: 0.66 (total); 0.40 (CPS>=1); 0.87 (CPS<1) HR CI: 0.36-1.21 (total); 0.16-1.01 (CPS>=1); 0.38-1.97 (CPS<1) Landmark rates: 6 mo: Pembro + chemo (total) 49% | Chemo (total) 26% [Slide 4] Overall Survival Total population [table] Events, n (%) | Median (95% CI), mo | HR (95% CI) Total population Pembro + chemo (n = 40): 25 (63) | 10.2 (8.0-11.4) | 0.33 (0.17-0.63) Chemo (n = 30): 25 (83) | 7.1 (4.0-8.2) PD-L1 CPS >=1 Pembro + chemo (n = 15): 9 (60) | 11.0 (8.7-11.9) | 0.16 (0.05-0.49) Chemo (n = 12): 12 (100) | 4.8 (1.9-8.6) PD-L1 CPS <1 Pembro + chemo (n = 23): 14 (61) | 8.1 (6.6-13.0) | 0.49 (0.21-1.11) Chemo (n = 16): 12 (75) | 7.1 (3.6-9.9) Curve label at 6 months: 85% (pembro + chemo) / 59% (chemo) Y-axis: OS, %; X-axis: Months No. at risk (0,2,4,6,8,10,12,14,16,18,20,22,24 mo) Pembro + chemo: 40 34 31 26 18 13 5 3 3 3 3 2 2 Chemo: 30 26 21 17 10 2 1 0 0 0 0 0 0 Data cutoff date: November 19, 2025. Survival table Endpoint: OS, total population (subgroups in table) Arm | N | Median | Median CI Pembro + chemo (total) | 40 | 10.2 | 8.0-11.4 Chemo (total) | 30 | 7.1 | 4.0-8.2 Pembro + chemo (CPS>=1) | 15 | 11.0 | 8.7-11.9 Chemo (CPS>=1) | 12 | 4.8 | 1.9-8.6 Pembro + chemo (CPS<1) | 23 | 8.1 | 6.6-13.0 Chemo (CPS<1) | 16 | 7.1 | 3.6-9.9 HR: 0.33 (total); 0.16 (CPS>=1); 0.49 (CPS<1) HR CI: 0.17-0.63 (total); 0.05-0.49 (CPS>=1); 0.21-1.11 (CPS<1) Landmark rates: 6 mo: Pembro + chemo (total) 85% | Chemo (total) 59%

ABBV-969 achieved deep and durable PSA responses in patients with bone and soft tissue lesions who progressed on taxanes and RLTs • Confirmed PSA50 at ≥ 3 mg/kg was 67.4% (29/43) • Confirmed PSA90 at ≥ 3 mg/kg was 30.2% (13/43) • Median duration of PSA response was 10.8 months (95% CI: 6.7, NE) • Median time to PSA response was 3.0 months (95% CI: 2.0, 5.2) [Waterfall plot] PSA change from baseline, % (30 to −100); PSA50; PSA90 Dose: 1 mg/kg · 2 mg/kg · 3 mg/kg · 5 mg/kg · 8 mg/kg · 10 mg/kg · 12.5 mg/kg Prior PSMA-RLT / Prior Taxane / Target Lesion / RECISTᵇ [per-patient grid] Yes · No · CR · PR · SD · NAᵃ

[Slide 1] POTOMAC: Disease-Free Survival in Patients With Papillary Only and With T1 Tumours D+BCG (I+M) showed a consistent benefit across papillary only and T1 subgroups Papillary only (64%-65% of ITT population)a,b,1 | D+BCG (I+M) N=217 | BCG (I+M) N=220 Number of events, n (%) | 37 (17) | 65 (30) Median DFS, months (95% CI) | NR (NR-NR) | NR (74.0-NR) Median follow-up in censored patients: 60.7 months Landmark labels: D+BCG 87% / 84% / 84%; BCG 81% / 76% / 74% (at dashed lines 24, 36, 48 months) HR 0.56 (95% CI, 0.37-0.84) No. of patients at risk (0-81 months, every 3 months) D+BCG (I+M) 217 204 190[?] 184 183 178 177 174 168 167 164 160[?] 154 150 148 145 144 133 127 102 98[?] 69 57 38 19 9 1 0 BCG (I+M) 220 207 195 188 180 177 172 167 163 159 155 150 146 145 142 141 140 130 122 99 92 63 53 32 17 7 0 0 T1 subset of papillary only (40%-43% of ITT population)c,d | D+BCG (I+M) N=137 | BCG (I+M) N=145 Number of events, n (%) | 22 (16) | 45 (31) Median DFS, months (95% CI) | NR (NR-NR) | NR (69.3-NR) Median follow-up in censored patients: 60.7 months Landmark labels: D+BCG 88% / 85% / 85%; BCG 79% / 75% / 73% (at dashed lines 24, 36, 48 months) HR 0.48 (95% CI, 0.28-0.79) No. of patients at risk D+BCG (I+M) 137 130 120 117 117 114 114 112 109 108 106 103 101 99 98[?] 97 96[?] 90 87 70 68 52 45 30 17 6[?] 1 0 BCG (I+M) 145 136 128 125 117 114 111 106 103 102 98 95 92 91 91 90 90 83 78 68 61 43 35 21 11 4 0 0 x-axis: Time from randomisation (months); y-axis: Probability of DFS Survival table Endpoint: DFS (papillary-only subgroup; T1 subset reported in notes) Arm | N | Median | Median CI D+BCG (I+M), papillary only | 217 | NR | NR-NR BCG (I+M), papillary only | 220 | NR | 74.0-NR D+BCG (I+M), T1 subset | 137 | NR | NR-NR BCG (I+M), T1 subset | 145 | NR | 69.3-NR HR: 0.56 (papillary only); 0.48 (T1 subset) HR CI: 0.37-0.84 (papillary only); 0.28-0.79 (T1 subset) Landmark rates: 24 mo: D+BCG papillary only 87% | BCG papillary only 81% | D+BCG T1 subset 88% | BCG T1 subset 79% 36 mo: D+BCG papillary only 84% | BCG papillary only 76% | D+BCG T1 subset 85% | BCG T1 subset 75% 48 mo: D+BCG papillary only 84% | BCG papillary only 74% | D+BCG T1 subset 85% | BCG T1 subset 73% [Slide 2] POTOMAC: Early High-Risk Disease Events in Patients With Papillary Only and With T1 Tumours The addition of D to BCG (I+M) resulted in fewer early high-risk recurrences, which are associated with worse outcomes(a,b) Papillary only D+BCG (I+M) | BCG (I+M) Patients Time from randomisation to high-risk event (months) 0 3 6 9 12 T1 D+BCG (I+M) | BCG (I+M) Early high-risk events: 16 patients in D+BCG (I+M) arm vs 29 patients in BCG (I+M) arm Early high-risk events: 10 patients in D+BCG (I+M) arm vs 20 patients in BCG (I+M) arm [Slide 3] POTOMAC: Cystectomy-Free Survival in Patients With Papillary Only and With T1 Tumours Cystectomy-free survival showed a favourable trend for D+BCG (I+M) vs BCG (I+M)(a) Papillary only | D+BCG (I+M) N=217 | BCG (I+M) N=220 Total events, n (%) | 31 (14) | 52 (24) Median CFS, months (95% CI) | NR (NR–NR) | NR (78.3–NR) HR 0.59 (95% CI, 0.38–0.92) T1 | D+BCG (I+M) N=137 | BCG (I+M) N=145 Total events, n (%) | 18 (13) | 39 (27) Median CFS, months (95% CI) | NR (NR–NR) | 78.3 (78.3–NR) HR 0.45 (95% CI, 0.25–0.78) Probability of cystectomy-free survival Time from randomisation to cystectomy-free survival (months) 0–84 No. of patients at risk (papillary only) D+BCG (I+M) 217 214 208 205 203 199 196 193 192 191 189 187 185 184 181 179 179 177 176 153 140 114 86 72 48 24 12 1 0 BCG (I+M) 220 218 215 210 207 204 202 199 196 196 191 190 188 185 181 177 174 171 168 145 135 112 82 64 41 21 8 0 0 No. of patients at risk (T1) D+BCG (I+M) 137 135 130 129 127 125 124 121 121 121 121 120 119 118 117 116 115 114 114 108 94 74 62 49 35 19 11 1 0 BCG (I+M) 145 144 141 139 136 133 132 130 129 128 126 125 123 120 117 114 113 111 109 97 89 73 54 42 28 14 4 0 0 Survival table Endpoint: Cystectomy-free survival (papillary-only subgroup; T1 subgroup reported separately) Arm | N | Median | Median CI D+BCG (I+M) — papillary only | 217 | NR | NR–NR BCG (I+M) — papillary only | 220 | NR | 78.3–NR D+BCG (I+M) — T1 | 137 | NR | NR–NR BCG (I+M) — T1 | 145 | 78.3 | 78.3–NR HR: 0.59 (papillary only); 0.45 (T1) HR CI: 0.38–0.92 (papillary only); 0.25–0.78 (T1)

[Slide 1] PURE-01 Pembrolizumab | SURE-01 Sacituzumab govitecan | SURE-02 SG + pembrolizumab | NURE-Combo Nivolumab + nab-paclitaxel MIBC cT2-4N0M0 · pure/predominant (≥50%) urothelial carcinoma Neoadjuvant or perioperative therapy · restaging with CT + multiparametric bladder MRI N = 279 · enrolled 02/2017-07/2025 Radical cystectomy n = 213 (76.4%) | Declined RC: reTURBT n = 54 (19.3%) | No surgery* n = 12 (4.3%) Adjuvant / maintenance therapy per protocol reTURBT 30/54 (56%) vs RC 43/213 (20%) Deferred RC or chemo-RT only for residual tumor / recurrence [Slide 2] Figure 1A EFS, reTURBT group (N = 53) Legend: TURBT only Number at risk (months 0 | 6 | 12 | 18 | 24 | 30 | 36): 53 | 52 | 41 | 25 | 20 | 10 | 6 Cumulative number of censoring: 0 | 1 | 4 | 12 | 19 | 28 | 31 36-mo EFS 66.7% (95% CI 54.5-81.7); MFS 65.9% (50.9-85.4) Figure 1B OS, reTURBT group (N = 53) Legend: TURBT only Number at risk (months 0 | 6 | 12 | 18 | 24 | 30 | 36): 53 | 53 | 49 | 32 | 25 | 13 | 8 Cumulative number of censoring: 0 | 0 | 2 | 14 | 22 | 33 | 37 36-mo OS 82.7% (95% CI 72.4-94.6) X-axis: Time from therapy (months) Survival table Endpoint: EFS and OS, reTURBT group reTURBT group (TURBT only) | 53 Landmark rates: 36 months: reTURBT group - EFS 66.7% (CI 54.5-81.7) | reTURBT group - MFS 65.9% (CI 50.9-85.4) | reTURBT group - OS 82.7% (CI 72.4-94.6) [Slide 3] Figure 2A EFS by pathological response Legend: TURBT no RD; TURBT RD; RC pCR; RC no pCR Number at risk (months 0 | 3 | 6 | 9 | 12 | 15 | 18 | 21 | 24) TURBT no RD (green): 35 | 35 | 35 | 35 | 32 | 27 | 23 | 23 | 19 TURBT RD (orange): 18 | 17 | 17 | 13 | 9 | 7 | 2 | 1 | 1 RC pCR (purple): 72 | 71 | 70 | 70 | 69 | 69 | 69 | 69 | 67 RC no pCR (pink): 141 | 128 | 121 | 119 | 112 | 104 | 91 | 88 | 80 Cumulative number of censoring green: 0 | 0 | 0 | 1 | 2 | 5 | 8 | 12 | 14 orange: 0 | 0 | 1 | 2 | 2 | 4 | 4 | 5 | 5 purple: 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 pink: 0 | 0 | 0 | 0 | 3 | 8 | 11 | 13 | 18 24-mo EFS, reTURBT vs RC: pCR 87.6% vs HR 1.96 (0.53-7.27); non-pCR 19.3% vs 67.8% Figure 2B OS by pathological response Number at risk (months 0 | 3 | 6 | 9 | 12 | 15 | 18 | 21 | 24) green: 35 | 35 | 35 | 35 | 33 | 29 | 26 | 26 | 22 orange: 18 | 18 | 18 | 18 | 16 | 11 | 6 | 4 | 3 purple: 72 | 72 | 72 | 72 | 71 | 71 | 71 | 71 | 70 pink: 141 | 140 | 137 | 136 | 131 | 126 | 112 | 109 | 98 Cumulative number of censoring green: [illegible - column alignment uncertain] orange: 0 | 0 | 0 | 0 | 0 | 5 | 6 | 8 | 8 purple: 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 pink: 0 | 0 | 0 | 0 | 4 | 10 | 14 | 16 | 21 24-mo OS, reTURBT vs RC: pCR 96.7%, HR 0.90 (0.10-8.40); non-pCR 50.4% vs 82.5% TURBT no RD / RD: ypT0Nx / residual disease at reTURBT; RC pCR: ypT0N0. Log-rank p<0.001 across the four groups for EFS, MFS and OS. Survival table Endpoint: EFS and OS by pathological response (reTURBT vs RC) TURBT no RD (ypT0Nx) | 35 TURBT RD (residual disease) | 18 RC pCR (ypT0N0) | 72 RC no pCR | 141 HR: pCR EFS HR 1.96; pCR OS HR 0.90 HR CI: EFS 0.53-7.27; OS 0.10-8.40 P: Log-rank p<0.001 across the four groups for EFS, MFS and OS Landmark rates: 24 months: reTURBT pCR - EFS 87.6% | reTURBT non-pCR - EFS 19.3% | RC non-pCR - EFS 67.8% | reTURBT pCR - OS 96.7% | reTURBT non-pCR - OS 50.4% | RC non-pCR - OS 82.5%

[Slide 1] ADC Development Landscape ADC development has accelerated since 2020 First-in-human studies [stacked bar chart 2010-2024; Number of FIH studies] Legend: ADC (monospecific); ADC (bispecific); ADC (unconventional) Multiple antigens are being explored as ADC targets Monospecifics [bar chart; Number of ADC; Early stage / Late stage] Targets: HER2, B7-H3, TROP-2, Claudin 18.2, Cadherin-17, FR α, Nectin-4, HER3, cMET, DLL3, ROR1, Tissue factor, B7-H4, Cadherin-6, FGFR2b, PTK7, EGFR, GPC3, 5T4, CD37, CEA, Claudin-6, Integrin β6, PSMA ADC development is advancing in various cancer types, but progress in renal cell carcinoma remains limited [Horizontal bar chart, Number of ADCs; Approved / Late stage (including regulatory review)] Tumour types: Prostate, Endometrial cancer, SCLC, Pancreatic cancer, Colorectal, Esophageal squamous cell carcinoma, Blastic plasmacytoid dendritic cell neoplasm, HER2-positive solid tumors, Nasopharyngeal carcinoma, Cervical cancer, Urothelial cancer, Gastric cancer, Breast cancer, Ovarian, peritoneal, or fallopian tube cancer, NSCLC, B cell NHL, Acute myeloid leukemia, Cutaneous T cell lymphoma, Anaplastic large cell lymphoma, Hodgkin lymphoma, Acute lymphoblastic leukemia, Multiple myeloma Conilh L. et al. Cell 189:2791,2026 [Slide 2] ADC Development in RCC: Limited Success So Far ADC | Target | Payload | Trial | Key results | Clinical trial Status SGN-75(1) | CD70 | MMAF | Phase I | Efficacy: 1 CR, 2 PRs; Safety: Ocular toxicity, thrombocytopenia | Inactive AMG 172(2) | CD70 | DM1 | Phase I | Efficacy: ORR 5.4%; Safety: Thrombocytopenia; hepatotoxicity | Inactive AGS-16M8F / AGS-16C3F(3) | ENPP3 | MMAF | Phase I / II | Efficacy: ORR 7.5%; mPFS 2.9 vs 5.7 mo (axitinib); Safety: Ocular toxicity | Inactive CDX-014(4) | TIM-1 | MMAE | Phase I | Efficacy: ORR 6%; Safety: Treatment-related death (multi-organ failure) | Inactive HKT288(5) | CDH6 | DM4 | Phase I | Efficacy: No objective responses; Safety: Neurological toxicity | Inactive SGN-CD70A(6) | CD70 | PBD dimer | Phase I | Efficacy: ORR 5.6%; Safety: Thrombocytopenia (DLT) | Inactive MDX-1203(7) | CD70 | Duocarmycin | Phase I | Efficacy: ORR not reported; Safety: Delayed edema / effusions | Inactive DS-6000a(8) | CDH6 | DXd | Phase I / II | Efficacy: Early RCC activity (1 confirmed PR); Safety: Thrombocytopenia (early Phase I) | Ongoing Inactive: No ongoing clinical development identified as of October 2026. 1 Tannir et al. Invest New Drugs. 32(6);1246, 2014. 2 Massard et al. Cancer Chemother Pharmacol. 83:1057, 2019. 3 Thompson et al. Clin Cancer Res.24(18);4399, 2018 Kollmannsberger et al. Oncologist. 26(3):182, 2021. 4 McGregor et al. J Immunother Cancer. 38(6):1807,2020. 5 Schöffski et al. Oncol Res Treat 44(10):547,2021. 6 Pal et al. Cancer. 125:1124,2019. 7 Owonikoko et al. Cancer Chemother Pharmacol. 77(1):155,2016 8 Hamilton et al. J Clin Oncol. 40(16) suppl 3002(2022); NCT04707248; NCT06660654. [Slide 3] BARRIER 1: Delivery Can the ADC reach the tumor cell? [Schematic: Pharmacokinetics; Tumor penetration (highlighted); Target recognition; Intracellular trafficking; Payload; TME modulation] Mosele et al. Nat Med (32)2774-2785, 2026 Potential barriers in ccRCC 1. Hypervascularity ü uniform ADC delivery [Images A-F: MRI, perfusion map (scale 0-300), gross specimen, IHC] Yuan Q. et al. Clin Genitourin Cancer 14(6); e585-e594, 2016 • Intratumoral perfusion is spatially heterogeneous even in hypervascular ccRCC. • Heterogenous perfusion may result in spatially uneven ADC delivery to tumor cells. [Slide 4] ADC targets are abundant in ccRCC — but is high expression enough? Expression of ADC-related target genes BeforeIO Tokyo ccRCC samples only [Violin plot, log2(TPM + 1): CD70, ENPP3, B7-H3, CDH6, CAIX, PSMA, TROP2, Nectin-4] Bulk RNAseq in 24 ccRCC (unpublished data) IHC expression of ADC target antigens in renal cancer Legend: High; Medium; Low; Not detected CAIX | CAB017107 | 5 (High) | 6 (Medium) | n=12 CDH6 | HPA007047 | 2 (High) | 10 (Medium) | n=12 B7-H3 | CAB017826 | [High, unlabelled] | 4 (Medium) | 6 (Low) | n=11 TROP2 | HPA043104 | 11 (Not detected) | n=11 Nectin-4 | HPA010775 | 4 (Low) | 8 (Not detected) | n=12 ENPP3 | HPA043772 | [Medium, unlabelled] | 6 (Low) | 5 (Not detected) | n=12 CD70 | IHC pending | No IHC data HPA renal cancer IHC | One illustrative antibody per target; not an inter-protein abundance ranking. Antibody-dependent staining; histology not restricted to ccRCC. CD70: IHC pending. https://www.proteinatlas.org/

[Slide 1] No PSMA PET at diagnosis EAU and ESMO recommendations for initial staging of prostate cancer INITIAL STAGING BY RISK CATEGORY Low / favourable intermediate: No routine metastatic imaging Unfavourable intermediate: Consider PSMA PET/CT High risk / locally advanced: PSMA PET/CT if available WHY PSMA PET? proPSMA: staging accuracy 92% PSMA PET vs 65% CT + bone scan More sensitive detection of nodal and distant disease before radical treatment PROTEUS enrolled conventionally staged M0 patients without baseline PSMA PET. Nicholas van As [Slide 2] Progression or delayed detection? Baseline conventional imaging, but PSMA PET used to identify subsequent events BASELINE: CONVENTIONAL IMAGING | FOLLOW-UP: PSMA PET CT + bone scan M0 → Occult PSMA-avid micrometastasis → PSMA PET positive "MFS event" Two biologically different possibilities can produce the same observed event: A. Truly new metastatic progression B. Pre-existing disease first detected later Without baseline PSMA PET, these cannot reliably be distinguished. ~70% of RP patients with PSMA PET-detected relapse had persistent PSA. Nicholas van As [Slide 4] Is avoiding subsequent treatment a patient benefit? 100% of these patients have already received a second + third treatment Time to First Subsequent Therapy (%) Median 46.0 months APA + ADT Median 12.3 months RP + PLND alone 64.6% (95% CI, 54.1-73.3) 32.0% (95% CI, 26.7-37.5) HR, 0.29 95% CI, 0.20-0.43 nominal p<0.001^b Months from Randomization: 0 6 12 18 24 30 36 42 No. at risk APA + ADT 102 97 91 87 81 73 42 12 RP + PLND [illegible - row cropped] • Postoperative RT is initiated by BCF • 20% vs 43% at 3 years • ....but there is an approximately 20 month lead time bias Survival table Endpoint: Time to first subsequent therapy APA + ADT | 102 | 46.0 months RP + PLND alone | 12.3 months HR: 0.29 HR CI: 0.20-0.43 P: nominal p<0.001 Landmark rates: 36 mo: APA + ADT 64.6% (CI 54.1-73.3) | RP + PLND alone 32.0% (CI 26.7-37.5)

[Slide 1] Is avoiding subsequent treatment a patient benefit? One year of ADT + Apalutamide comes at a price Safety population, n (%) | Perioperative APA + ADT (n=101): All grades | Grade ≥3 | RP + PLND alone (n=286): All grades | Grade ≥3 Hot flush | 65 (64) | 0 | 2 (0.70) | 0 Fatigue | 22 (22) | 0 | 8 (2.8) | 0 Rash | 16 (16) | 3 (3.0) | 2 (0.70) | 0 Arthralgia | 16 (16) [uncertain - may read 15 (15)] | 1 (1.0) | 10 (3.5) | 0 Pruritus | 13 (13) | 0 | 1 (0.35) | 0 Hypertension | 10 (9.9) | 5 (5.0) | 13 (4.5) | 7 (2.4) COVID-19 | 10 (9.9) | 0 | 4 (1.4) | 0 Asthenia | 10 (9.9) | 0 | 2 (0.70) | 0 [Slide 2] What happens AFTER control-arm failure? PROTEUS EFS records the first event — not the outcome after effective salvage treatment PSA persistence / recurrence or PET-detected relapse → SALVAGE / SUBSEQUENT TREATMENT RT ± ADT | ADT + ARPI Other MDT-directed care → LONG-TERM OUTCOME NOT SHOWN KEY QUESTION: How many control patients were successfully salvaged — and what were their subsequent outcomes? Nicholas van As [Slide 3] PROTEUS: a positive result, but significant limitations EFS and MFS at 3 years median follow-up is too short INITIAL STAGING: No baseline PSMA PET in a very-high-risk population PRIMARY ENDPOINT: EFS includes biochemical failure during prolonged androgen suppression MFS: PSMA PET-detected events complicate interpretation without baseline PET TOXICITY: Significant increase in toxicity COMPARATOR: RP alone versus multimodality care, or ADT +RT Nicholas van As [Slide 4] Should practice change now? I would not currently support perioperative apalutamide + ADT with radical prostatectomy in high-risk prostate cancer. Longer follow-up, mature survival data and outcomes after salvage treatment are needed before changing practice. Nicholas van As

Efficacy summary Encouraging early responses were observed across all tumor types Efficacy parameter | HNSCC (N=39) | Melanoma (N=26) | HCC (N=21) | CC (N=22) | PROC (N=31) | Total* (N=139) Prior treatment, median (range), lines | 2 (1–4) | 2 (1–5) | 2 (1–2) | 2 (1–4) | 2 (1–4) | 2 (1–5) Follow-up duration, median (range), months | 4.8 (0.5–10.0) | 7.4 (3.6–12.5) | 7.7 (1.4–9.3) | 6.1 (4.0–11.4) | 6.5 (2.6–8.6) | 6.3 (0.5–12.5) Unconfirmed BOR, n (%) CR | 0 | 0 | 0 | 1 (4.5) | 2 (6.5) | 3 (2.2) PR | 15 (38.5) | 8 (30.8) | 9 (42.9) | 10 (45.5) | 23 (74.2) | 65 (46.8) SD | 19 (48.7) | 13 (50.0) | 10 (47.6) | 9 (40.9) | 3 (9.7) | 54 (38.8) PD | 4 (10.3) | 5 (19.2) | 1 (4.8) | 2 (9.1) | 3 (9.7) | 15 (10.8) NE/missing | 1 (2.6) | 0 | 1 (4.8) | 0 | 0 | 2 (1.4) uORR, % (95% CI) | 38.5 (23.4–55.4) | 30.8 (14.3–51.8) | 42.9 (21.8–66.0) | 50.0 (28.2–71.8) | 80.6 (62.5–92.5) | 48.9 (40.4–57.5) cORR, % (95% CI) | 23.1 (11.1–39.3) | 26.9 (11.6–47.8) | 38.1 (18.1–61.6) | 40.9 (20.7–63.6) | 71.0 (52.0–85.8) | 39.6 (31.4–48.2) Pending confirmation, n (%)† | 3 (7.7)‡ | 1 (3.8) | 0 | 1 (4.5) | 2 (6.5) | 7 (5.0) DCR, % (95% CI) | 87.2 (72.6–95.7) | 80.8 (60.6–93.4) | 90.5 (69.6–98.8) | 90.9 (70.8–98.9) | 90.3 (74.2–98.0) | 87.8 (81.1–92.7) *Efficacy-evaluable participants: received ≥1 dose of study drug, baseline efficacy assessment, and ≥1 post-baseline assessment or early treatment discontinuation. †Pending confirmation includes patients with an observed CR/PR whose confirmed BOR was not yet available as of September 4, 2026, excluding those who received new anticancer therapy, died/reached EOS, or had confirmed PD before response confirmation. ‡As of Oct 09, of the 7 responses pending confirmation at the data cutoff, 3 have since been confirmed, while the remaining 4 are awaiting the next tumor assessment. [abbreviations footnote illegible] Data cutoff date: September 4, 2026

Efficacy-ORR • A total of 23 patients have received at least one tumor assessment • ORR was 65.2%(15/26), DCR was 95.7%(22/26). Tumor shrinkage were obsevered in 87% (20/23) patients • irORR was 69.6% (Only one patient had PD at the first assessment, subsequently classified as irPR per iRECIST) | Ivonescimab (n=23) n (%) Best Overall Response (BOR) | CR | 0 (0.0%) PR | 15 (65.2%) SD | 7 (30.4%) PD | 1 (4.3%) Objective Response Rate (ORR) | 65.2% Disease Control Rate (DCR) | 95.7% [Waterfall plot legend: PR, SD, PD; y-axis Best percentage of change from baseline (%)] Data cutoff date: July 14, 2026 Jun Guo, MD

Metro Madrid The mCRPC metro map Same destination. Different journey. The map ahead depends on the route already travelled. Initial treatment in mHSPC ADT only ADT + ARPI ADT + ARPI + PSMA-RLT ADT + ARPI + PARPi ADT + ARPI + AKTi ADT + ARPI + docetaxel AR-naïve Taxane-naïve AR-exposed Taxane-naïve AR and PSMA-exposed Taxane-naïve AR and PARP-exposed Taxane-naïve AR and AKT-exposed Taxane-naïve AR and taxane-exposed mCRPC/ AMPR Next treatment options (mechanisms) Multiple paths beyond mCRPC – the most appropriate depends on prior treatment, tumour biology, phenotype and patient factors. Taxane (microtubule inhibition) PARP (DNA repair) PSMA-RLT (PSMA targeting) Radium-223 + enza (bone-targeted α-emitter) AR pathway (when appropriate) Immunotherapy / novel targets (selected patients)

[Photo of poster with red 'BEST POSTER' rosette; partial view, left side cropped] [...]ctomy (RC) [...]i,1 Giovanni L. Pastorino, [...] Lacava,1 Gaia [...] [...]don Foundation Trust, London, UK. 5. Depart[ment ...] Oncology 2, Azienda Ospedaliero-[...] [...]ital, Comprehensive Cancer Center, Milan, [...] Table (partial): Site of relapse | Status Bladder pTaHG > systemic | Dea[d] Bladder pT2 | De[ad] Systemic | [obscured] Systemic | Alive Bladder pT2 | Alive Bladder pT1HG | Alive Systemic | Alive Systemic | Alive Bladder pT2 > systemic | Alive Bladder pT1HG > systemic | Alive [...]erapy (chemoradiotherapy in all cases). [...]residual tumor or intravesical [...]py, recurrence/progression occurred in [...] Full Paper Here DOI: 10.1016/j.esmoop.2026.108634 CONCLUSIONS Patients with ypT0Nx at reTURBT achieved 24-month MFS and OS >95%, comparable to those with ypT0N0 after RC, supporting biopsy-confirmed cCR (Milan criteria) as a [cr]iterion for bladder preservation.

Stronger and more consistent evidence of benefit → Weaker or uncertain evidence of benefit BRCA2 | BRCA1 | PALB2 | CDK12 | ATM | CHEK2 | Other HRR genes (FANCA, RAD51C/D, BARD1, BRIP1, etc.) | HRR-negative (non-HRRm) Prevalence in mCRPC (approx.): ~5–8% | ~1–2% | ~1–2% | ~3–5% | ~3–5% | ~1–2% | ~3–5% | ~70% (no HRR alteration) rPFS HR (95% CI) PARPi + ARPI vs ARPI: 0.31 (0.23–0.42) | 0.51 (0.23–1.10) | 0.52 (0.23–1.17) | 0.50 (0.32–0.80) | 1.05 (0.74–1.49) | 1.06 (0.67–1.66) | Variable (insufficient data) | PROpel Olaparib + Abiraterone rPFS HR 0.76 (0.60–0.97) OS HR 0.89 (0.70–1.14) OS HR (95% CI) PARPi + ARPI vs ARPI: 0.66 (0.49–0.89) | 0.74 (0.34–1.61) | 0.78 (0.34–1.80) | 0.63 (0.39–0.99) | 1.18 (0.82–1.71) | 1.53 (0.95–2.46) | Variable (insufficient data) | TALAPRO-2 Talazoparib + Enzalutamide rPFS HR 0.69 (0.49–0.98) OS HR 0.78 (0.58–1.05) Key interpretation: BRCA2: • Strongest and most consistent benefit across trials • Highest response rates and longest duration of benefit BRCA1: • Evidence of benefit but smaller magnitude than BRCA2 • Responses seen across trials PALB2: • Signals of benefit (rFS and OS) • Data limited and not always statistically significant CDK12: • Potential benefit with PARPi + ARPI (FDA pooled analysis) • Less clear benefit with PARPi monotherapy • Interpret with caution ATM: • No significant benefit with PARPi + ARPI • Consider alternative strategies or clinical trial enrolment CHEK2: • No clear benefit with PARPi + ARPI • Consider alternative strategies or clinical trial enrolment Other HRR genes: • Heterogeneous results • Limited and inconsistent data • Consider on an individual basis, preferably within clinical trials HRR-negative: • Modest rPFS benefit with PARPi + ARPI in selected trials • No established OS benefit • Routine use requires careful patient selection and consideration of toxicity and prior therapy Key trials contributing data: BRCA2: PROfound MAGNITUDE TALAPRO-2 PROpel BRCA1: PROfound MAGNITUDE TALAPRO-2 PROpel PALB2: PROfound TALAPRO-2 PROpel CDK12: PROpel MAGNITUDE TALAPRO-2 ATM: PROpel MAGNITUDE TALAPRO-2 CHEK2: PROpel MAGNITUDE TALAPRO-2 Other HRR genes: PROfound MAGNITUDE TALAPRO-2 PROpel HRR-negative: PROpel TALAPRO-2

available at www.sciencedirect.com journal homepage: euoncology.europeanurology.com eau European Association of Urology [EUROPEAN UROLOGY ONCOLOGY journal cover thumbnail] Clinical Outcomes of Bladder-sparing Management Versus Early Radical Cystectomy in cT1 Micropapillary Bladder Cancer: A Systematic Review and Meta-analysis Sagar R. Patel a, Mikael Rauf a, Yimin Geng b, David Hinojosa-Gonzalez c, Joseph B. Black a, Taeweon Lee a, Amanda A. Myers d, Charles Guo e, Roger Li f, Joshua J. Meeks g, Piyush Agarwal h, Stephen Williams i, Omar Alhalabi j, Andrea Necchi k, Laura S. Mertens l, Paolo Gontero m, Neema Navai a, Ashish M. Kamat a,*

[Slide 1] [Photo of presenter standing beside poster 4758P] 4758P [Queen Mary University of London / Barts Cancer Institute logo] bci.qmul.ac.uk 4758P: Outcomes of patients who successfully switched from chemotherapy to avelumab in the randomised DISCUS trial Sara Membribes, Francesca Jackson-Spence, Syed Hussain, Miguel Angel Climent, Iciar García Carbonero, Louis-Marie Dourthe, Pablo Borrega, Javier Molina-Cerrillo, Javier Puente, Pratik Das, Jahangeer Malik, Daniel Castellano, Rob Jones, Ignacio Duran, Yohann Loriot, Yue Qi Wang, Garima Priyadarshini, Charlotte Ackerman, Thomas Powles, Enrique Grande On behalf of the DISCUS investigators [affiliations: illegible] Background • Maintenance avelumab after platinum-based chemotherapy is standard of care in advanced urothelial cancer. • A substantial proportion of patients never start planned maintenance treatment. • Factors associated with failure to receive avelumab and the impact of the switch on quality of life (QoL) remain poorly characterised in prospective datasets. Methods • Exploratory, descriptive analysis from DISCUS (NCT06892860): investigator-initiated, international, randomised phase II trial. • Randomisation to 3 (3C) or 6 (6C) cycles of gemcitabine plus cisplatin or carboplatin. Patients without progression received avelumab. • PROs: EORTC QLQ-C30 GHS/QoL at baseline and relevant post-chemotherapy assessment (C4D1 for 3C and C7D1 for 6C). The pre- and post- immunotherapy GHS/QoL was restricted to patients who had avelumab (baseline and C7D1 for 3C and C9D1 for 6C). [Three cycles] increased transition to avelumab [bar chart partially obscured: 3C ...(74%) | 19 (15%) | 13 (10%); 6C ... | 40 (33%) | 12 (10%); legend: ...ce avelumab / ...ression / death / ...ity / patient / investigator decision] Baseline characteristics: ECOG PS distinguished patients who did not proceed to receive avelumab | Avelumab 162 (66%) | No avelumab 84 (34%) | P Median age, years | 71.5 | 71.4 | Age ≥65 years | 120 (74%) | 58 (69%) | 0.40 Male | 118 (73%) | 62 (74%) | 0.87 White | 146 (90%) | 75 (89%) | 0.84 Liver metastases | 29 (18%) | 17 (20%) | 0.66 ECOG PS ≥1 | 35 (22%) | 32 (38%) | 0.006 Gemcitabine-cisplatin | 73 (45%) | 30 (36%) | 0.30 Any dose delay | 94 (58%) | 39 (46%) | 0.08 Any dose reduction | 31 (19%) | 12 (14%) | 0.34 Grade 3/4 AEs, n* | 228 (63%) | 135 (37%) | Subsequent systemic therapy | 36 (22%) | 23 (27%) | *Multiple adverse events per patient therefore number of events used as denominator QoL deterioration during chemotherapy identifies patients who will not reach avelumab • GHS/QoL declined with 6C of chemotherapy but not with 3C • Transition to avelumab was associated with stable QoL in both arms [Bar chart, mean change in GHS/QoL from baseline: 3C + avelumab +2.0; 3C no avelumab -9.8; 6C + avelumab -4.6; 6C no avelumab -10.0; group n values illegible] Shaded area = clinically meaningful decline (beyond -9 points). Arrow = 95% CI extending beyond the axis [rest illegible] Patients who did not reach avelumab had poor outcomes: median PFS 3.9 and OS 7.5 months PFS: HR 0.13 (95% CI: 0.08-0.20); legend: 3/6 cycles chemotherapy + avelumab; 3/6 cycles chemotherapy; x axis Time from Randomisation (months) OS: HR 0.22 (95% CI: 0.13-0.35); legend: 3/6 cycles chemotherapy + avelumab; 3/6 cycles chemotherapy; x axis Time from randomisation (months) Conclusions • 34% of eligible patients never received avelumab, their outcomes were poor and only 27% received subsequent therapy. • 3C increased transition to avelumab (74% vs 57%), driven by fewer progression/death events (15% vs 33%). Once avelumab started, QoL did not deteriorate further. • Only baseline ECOG PS differed among patients who did and did not get avelumab. [Slide 2] [Photo of three authors standing in front of poster 4758P; same poster as the first image] 4758P 4758P: Outcomes of patients who successfully switched from chemotherapy to avelumab in the randomised DISCUS trial Baseline characteristics: ECOG PS distinguished patients who did not proceed to receive avelumab | Avelumab 162 (66%) | No avelumab 84 (34%) | P Median age, years | 71.5 | 71.4 | Age ≥65 years | 120 (74%) | 58 (69%) | 0.40 Male | 118 (73%) | 62 (74%) | 0.87 White | 146 (90%) | 75 (89%) | 0.84 Liver metastases | 29 (18%) | 17 (20%) | 0.66 ECOG PS ≥1 | 35 (22%) | 32 (38%) | 0.006 Gemcitabine-cisplatin | 73 (45%) | 30 (36%) | 0.30 Any dose delay | 94 (58%) | 39 (46%) | 0.08 Any dose reduction | 31 (19%) | 12 (14%) | 0.34 Grade 3/4 AEs, n* | 228 (63%) | 135 (37%) | Subsequent systemic therapy | 36 (22%) | 23 (27%) | Patients who did not reach avelumab had poor outcomes: median PFS 3.9 and OS 7.5 months PFS HR 0.13 (95% CI: 0.08-0.20); OS HR 0.22 (95% CI: 0.13-[obscured]) [Background, Methods and Conclusions panels partly obscured by people] Badges: Charlotte Ackerman; Francesca Jackson-Spence; Sara Coca Membribes

ABSTRACT #1900 / FPN ESMO CONGRESS 2026 • TRIAL IN PROGRESS • NCT07484022 A phase 1 study of GB-4362, a free MMAE-neutralizing antibody, with enfortumab vedotin and pembrolizumab in advanced urothelial cancer Generate:Biomedicines S. Funt(1), J. Rosenberg(2), O. Alhalabi(3), V. Szenes(4), A. Tripathi(5), K. Mayawala(4), D. De Alwis(4), L. Goyal(4), S. Straley(4), K. Zhang(6), B. Adesunloye(7), G. O'Sullivan Coyne(8), M. Devitt(9), J. Chahoud(10), H. Hassanein(11), B. Saraiya(12), C. Jiang(13) Background Enfortumab vedotin (EV) plus pembrolizumab is a first-line standard for locally advanced or metastatic urothelial cancer. Systemic release of monomethyl auristatin E (MMAE) from EV contributes to off-target toxicity. Peripheral neuropathy is a key treatment-limiting toxicity of EV, reported in 67%(a) of patients in EV-302, leading to dose interruption (22%), reduction (13%), or discontinuation (15%); symptoms may persist after treatment and limit the duration and intensity of an otherwise effective regimen. Therapeutic Concept GB-4362 is designed to selectively bind and neutralize circulating free MMAE without binding intact EV, with the goal of reducing neuropathy and improving the therapeutic window. Reducing treatment-related toxicities may help patients remain on EV plus pembrolizumab longer and support broader use across community and other care settings. Phase 1 data may also inform development of GB-4362 with other MMAE-based antibody-drug conjugates, potentially extending the approach to additional cancer types. GB-4362 has received Fast Track designation from the U.S. FDA for the prevention and/or reduction of EV-induced toxicity in patients with urothelial cancer. Preclinical Data In vitro, GB-4362 blocked MMAE, and not ADC cytotoxicity [curve: %viability vs Concentration mAb (nM); GB-4362 + MMAE; GB-4362 + ADC] In mouse, EV's efficacy was preserved up to 80% MMAE reduction with GB-4362 [Tumor Volume (mm3) vs Time (day); GB-4362 administrations; Vehicle (Negative control); MMAE reduction ≥ ~85% (Postive control); No efficacy impact MMAE reduction ≤ ~80%] In monkeys, single dose GB-4362 resulted in dose-dependent reduction in free MMAE which was associated with neutropenia improvement Treatment | Grade 4 Neutropenia (%) | Free MMAE % (AUC0-2wk) 4mg/kg EV | 100% | 100% 4mg/kg EV + 0.5mg/kg GB-4362 | 100% | 74% 4mg/kg EV + 2mg/kg GB-4362 | 50% | 52% 4mg/kg EV + 4mg/kg GB-4362 | 33% | 32% Grade 4 neutropenia is defined as ANC < 0.5 × 10^3 cells/µL. EV = enfortumab vedotin Dose-escalation study design Dose Level 1 GB-4362 → Dose Level 2 GB-4362 → Dose Level 3 (optional) GB-4362 DOSE EXPANSION: Preliminary recommended Phase 2 dose of GB-4362 - EV+Pembrolizumab administered per standard of care - Dose Escalation (6+3 design): GB-4362 only administered in cycles 2-3. - GB-4362 administered on same days as EV - Target ~50% MMAE reduction using intra-patient MMAE data Key eligibility - Locally advanced or metastatic urothelial carcinoma with measurable disease - Planned to receive standard of care EV plus pembrolizumab - ECOG 0-1 (escalation); ECOG 0-2 (expansion) - No uncontrolled diabetes - No active untreated CNS metastases - No grade >1 sensory or motor peripheral neuropathy Endpoints Primary: GB-4362-emergent AEs/SAEs incidence and severity Secondary: PK and ADA of GB-4362. Reduction in free MMAE exposure Study Status Recruiting. Multiple active U.S. sites No clinical outcome data are presented. Source: ClinicalTrials.gov NCT07484022 (updated 18 Jun 2026) a. PADCEV® (enfortumab vedotin-ejfv) U.S. Prescribing Information. Astellas Pharma US, Inc.; revised Nov 2025. Table 7

[Slide 1] Immune activation Releasing the brakes - Immune checkpoint inhibitors Pressing the accelerator - Novel cytokine therapies RCC tumor T cell immunity Immune navigation Clearing the traffic - Modulating T cell-extrinsic factors Adding a steering wheel - Antigen-directed therapies Yochum & Braun, Targeted Oncology, 2025 [Slide 2] A neoantigen vaccine for kidney cancer Trial overview Tumor procurement | Target selection | Personal vaccine manufacture | Vaccine administration ccRCC Stage III-IV; No evidence of disease (NED) after surgery WES, RNAseq → Neoantigen prediction (HLAthena) → Identification of up to 20 personal neoantigens Pools of synthetic long peptides (A, B, C, D) Vaccine + poly ICLC 1/2 s.c., 1/2 i.d. s.c. ipilimumab (+/-) Prime, Boost, Boost; weeks 0, 4, 8, 12, 16, 20, 24; Blood collection Successful vaccine manufacturing for all patients No. of coding mutations; No. of peptides per vaccine Pt. ID: 101 | 102 | 104 | 105 | 106 | 107 | 108 | 109 | 110 RCC driver mutations: PIK3CA, VHL, BAP1, KDM5C, BAP1, KDM5C, PBRM1 SNV mutations; Frameshift indel mutations [Bra]un et al, Nature, 2025. [Slide 3] Encouraging patient outcomes (but only n = 9) Swimmer plot: Pt. ID 101, 102, 105, 106, 104, 107, 109, 108, 110; Months from surgery (0, 12, 24, 36, 48) Stage: III; IV NED Vaccine priming; Vaccine boost; Alive & NED Death (unrelated; NED at last follow-up) [patient 110] Disease-free survival vs Months from first vaccination (0, 10, 20, 30, 40) Number at risk: 9 | 9 | 8 | 5 | 2 [Bra]un et al, Nature, 2025. Survival table Endpoint: Disease-free survival from first vaccination Vaccinated patients | 9 [Slide 4] How to move vaccines to advanced RCC? Combination therapy 1. Release of cancer cell antigens (cancer cell death) 2. Cancer antigen presentation 3. Priming and activation (APCs and T cells) 4. Trafficking of T cells to tumors (CTLs) 5. Infiltration of immune cells into tumors and stroma 5a. Accumulation of activated T cells in and through stroma 5b. Interaction with immune cells (e.g., TAMs, T-regs, TLS) 5c. Maintained effector state and function 6. Recognition of cancer cells by T cells 7. Killing of cancer cells (Immune and cancer cells) [...]shizuka, and Braun, ESMO Immuno-Oncology and Technology, 2026.

Resistance mechanism | Key alterations | Strategies to overcome resistance AR reactivation | AR amplification / overexpression; AR-LBD mutations (L702H, T878A); AR-V7 splice variant | AR degraders: BMS-986365(1); PROTAC: luxdegalutamide(2,3), bavdegalutamide(4,5); RIPTAC: HLD-0915(6) Androgen bypass | Steroids and progestins activate mutant AR; Glucocorticoid receptor upregulation | CYP11A1 inhibition: opevesostat(7,8) Alternative survival signaling | PTEN loss / PI3K–AKT activation | AKT inhibition: capivasertib(9), ipatasertib(10) Epigenetic escape | EZH2 upregulation; FOXA1/GATA2 | EZH2 inhibition: mevrometostat(11) DNA repair deficiency | BRCA2, BRCA1, ATM, other HRR genes; MSI-high / dMMR | PARP inhibitors ± ARPI(12–14); PD-1 inhibitors (MSI-high)(15) Lineage plasticity (AR-indifferent) | TP53 / RB1 loss; Neuroendocrine transdifferentiation | Platinum-based chemotherapy; EZH2 inhibition AR-independent targets | Surface antigens: PSMA, STEAP1, hK2, B7-H3 | PSMA radioligand: 177Lu-PSMA-617(16,17); T-cell engagers: xaluritamig(18) (STEAP1), pasritamig (hK2)(19); B7-H3 ADCs: ifinatamab deruxtecan, DB-1311(20,21) Robinson D et al. Cell 2015; Antonarakis ES et al. N Engl J Med 2014 [rest cropped]

[Slide 1] 15:15 - 16:00 What oncologists should know about mRNA vaccines CHAIRS: YELENA JANJIGIAN, JAMES LARKIN Which Cancers Should We Target? Bridging from peptide vaccine trials -> mRNA vaccine development Neoantigen Peptide Vaccine Development Pipeline for RCC [Hand-drawn ACIR.org summary of phase 1 clinical trial, 9 patients, stage III and IV kidney cancer (RCC); small text, see 2113632411162669127_s4 for full legible version] Randomized Phase II mRNA INTerpath-004 RCC Key Eligibility Criteria - Histologically or cytologically confirmed RCC with clear cell or papillary histology with or without sarcomatoid features - Intermediate-high-risk, high-risk, or M1 NED N ≈ 272 R (1:1) n ≈ 136 -> V940 1 mg IM Q3W for 9 doses + Pembrolizumab 400 mg IV Q6W for 9 cycles n ≈ 136 -> Placebo Q3W for 9 doses + Pembrolizumab 400 mg IV Q6W for 9 cycles Stratification Factors - Histology (clear cell vs papillary) - Disease risk (intermediate-high vs high vs M1 NED) End Points - Primary: DFS by investigator assessment - Secondary: DMFS by investigator assessment, OS, safety Choueiri TK, et al. Oncologist, 2024 ACIR.org Toni K. Choueiri, MD ALICANTE AUDITORIUM - HALL 6 [Slide 2] 15:15 - 16:00 What oncologists should know about mRNA vaccines CHAIRS: YELENA JANJIGIAN, JAMES LARKIN Which Vaccine Should We Use? Shared vs Personalized Neoantigen Strategies Shared Neoantigens Predefined antigens shared across tumors Produced in bulk and available in advance Rapidly available and scalable Potentially Lower Cost Heavily Restricted to Specific Antigens Hybrid Solutions Personalized Neoantigens Tumor Sequencing-Informed Neoantigen Selection Individual manufacturing based on neoantigen-defining algorithm Complex Production Tailored to cancer features Manufacturing Delays Toni K. Choueiri, MD Adapted from Ott, et al. Nature Biotechnology 2026. ALICANTE AUDITORIUM - HALL 6 [Slide 3] 15:15 - 16:00 What oncologists should know about mRNA vaccines CHAIRS: YELENA JANJIGIAN, JAMES LARKIN Where do vaccines fit in the disease course? Early vs Advanced Setting considerations Clinical benefit has been demonstrated most clearly in the adjuvant setting, while efficacy in advanced disease remains less established. [Illustration of immune cells and tumors of increasing size] Eliminate MRD and Prevent Recurrence | Control Disease and Prolong Survival Lower Tumor Burden and Less Heterogeneity | Higher Burden and More Heterogeneity High Ratio of Vaccine-Induced T-cells to Cancer Cells | TME induces Immune Suppression Toni K. Choueiri, MD Adapted from Baharom, et al. Nature Reviews Immunology 2025. Created with BioRender.com ALICANTE AUDITORIUM - HALL 6 [Slide 4] 15:15 - 16:00 What oncologists should know about mRNA vaccines CHAIRS: YELENA JANJIGIAN, JAMES LARKIN Braun DA...Wu, Ott, Choueiri. Nature 2025 phase 1 clinical trial 9 patients stage III and IV kidney cancer (RCC) tumor sample, peripheral blood median of 45 high-quality somatic coding mutations per sample median of 15 neoantigen peptides per sample, including -at least 1 neoORF -1 or more derived from common RCC or pan-cancer driver mutations in 7/9 pts adjuvant setting prime-boost vaccination 4 peptide pools administered half sc and half id 2.5mg sc anti-CTLA-4 close to vaccine site in 5/9 patients median follow-up of 40 months from surgery 35 months from start of vaccination clinical result - well tolerated - no cancer recurrence in 9/9 patients ex vivo ELISpot with patient PBMCs - reactivity against a median of 7 vaccine peptides per patient - mostly from polyfunctional CD4+ PD-1+ CD45RO+ T cells - response was mostly strongest after the 2nd booster 44% (50/112) of passenger mutations 86% (6/7) of common RCC mutations 65% (11/17) of other pan-cancer mutations -> elicited an immune response (all de novo) skin biopsies of the vaccine site (before and after priming phase) ↑ number of myeloid cells ↓ proportions of Langerhans and dendritic cells ↑ number of lymphoid cells ↑ proportions of cytotoxic CD8+ T cells and proliferating NK cells ↑ expression of effector cytokine and cytotoxicity genes ↑ unique T cell clonotypes peripheral T cells - identification of 98 vaccine-expanded T cell clones - formed rapidly within 3 weeks (during the priming phase) - present at high levels up to 166-fold - persisted at high levels for up to 3 years - levels of putative pre-existing tumor- and viral-specific T cell clones remained the same during treatment reactivity against autologous tumor cells - at least 1 peptide pool with reactivity in 7/9 patients - reactivity against multiple RCC driver mutations addition of anti-CTLA-4 - made no apparent difference in this study ACIR.org ALICANTE AUDITORIUM - HALL 6

Summary Small-cell GU neuroendocrine carcinoma - focus on the bladder ▪ Early systemic risk ▪ Expert pathology and systemic staging are essential, even in cT1 bladder disease. ▪ Platinum remains the backbone – analogy with small cell lung cancer (SCLC) ▪ Platinum–etoposide produces 40-60% responses, but durability is limited. ▪ Localised bladder disease needs multimodal therapy ▪ Combine platinum-etoposide with radical cystectomy or definitive chemoradiotherapy. ▪ Immunotherapy in analogy with SCLC is promising option, but requires caution ▪ Bladder-specific survival benefit remains unproven. Curative use remains investigational. ▪ DLL3 is a promising therapeutic target ▪ Early activity supports biomarker-directed trials. GU efficacy and survival benefit need confirmation. Evidence remains limited: multidisciplinary decisions and clinical-trial enrolment are central

Acronym | Patient Population | Intervention | Primary completion date EvoPAR-PR01 NCT06120491 | N=1,898, mHSPC | HRR positive cohort: ADT plus ARPI ± saruparib (AZD5305) HRR negative cohort: ADT plus ARPI ± saruparib (AZD5305) | 09/2027 MEVPRO-3 NCT07028853 | N=1,000, mHSPC | enzalutamide ± mevrometostat (EZH2 inhibitor) | 09/2028 ASPIRE NCT06931340 | N=1,260, mHSPC | ADT + apalutamide ± docetaxel | 11/2031 Triple Switch NCT06592924 | N=830, mHSPC, PSA after ADT (6-12m) + ARPI (≥4m), PSA ≥0.2 | Continue ADT + ARPI vs ADT + ARPI + docetaxel | 03/2031 PEACE-6 Poor Responders EU CT 2022-502408-57-00 | N=500, synch. mHSPC ongoing SOC with PSA ≥0.2 ng/mL at 6-8m | Continuing SOC vs SOC + 4x 177Lu-PSMA-617 | 02/2033 PEACE-6 Vulnerable NCT04916613 | N=300, mHSPC, vulnerable/frail | ADT ± darolutamide | 03/2028 Oligo-PRESTO NCT04115007 | N=550, mHSPC, oligometastatic | Standard of care ± SBRT of metastases (max 5, at least one bone or lung ± nodes) | 06/2026 SWOG-S1802 NCT03678025 | N=1,273, mHSPC | Standard systemic therapy ± definitive treatment (OR or RT) of the primary tumour | 04/2033 PLATON NCT03784755 | N=409, low-volume mHSPC | SOC + local treatment of the prostate (low volume metastatic disease burden) vs. SOC + local treatment of all sites (primary and metastases) | 12/2032 DE-Escalate NCT05974774 | N=1,600, mHSPC | mHSPC treated with ADT + ARPI for 6-12 months and presenting with a PSA ≤ 0.2 ng/mL, randomised to continued vs intermittent therapy | 05/2035 LIBERTAS NCT05884398 | N=420, mHSPC | mHSPC treated with ADT + ARPI for 6 months and presenting with a PSA ≤ 0.2 ng/mL, randomised to continued apalutamide (Stop ADT) vs continue ADT + apalutamide | 10/2026 STAMPEDE-2 NCT06320067 | N=3,360, mHSPC | • Standard of care vs. SOC plus SABR (1-5 metastatic lesions: bone and/or non-regional lymph node sites) using conventional imaging • SOC vs SOC plus 177Lu-PSMA-617 | 04/2031

3602P - Clinical and transcriptomic correlates of primary resistance to nivolumab plus ipilimumab in metastatic renal cell carcinoma (mRCC) Elio Ibrahim¹, Wanling Xie¹, Razane El Hajj Chehade¹, Neeraj Agarwal², Benoit Beuselinck³, Suzanna Lee⁴, Jae Lyun Lee⁵, Heber Reyes Garcia⁶, Connor Wells⁶, Jessica Voce⁷, Sylvan Baca¹, Rana R. McKay⁷, Andrew Simmons⁸, David Braun⁹, Daniel Y.C. Heng⁶, Toni K. Choueiri¹. [affiliations: Dana-Farber Cancer Institute, Boston; Huntsman Cancer Institute, University of Utah; University Hospitals Leuven; Queen Elizabeth II Health Sciences Centre, Dalhousie University; Asan Medical Center, Seoul; Arthur JE Child Comprehensive Cancer Centre, Calgary; University of California San Diego, Moores Cancer Center; Exelixis, Inc., Alameda, CA; Yale Cancer Center — small print partly illegible] BACKGROUND • Nivolumab + ipilimumab (NIVO+IPI) is a standard first-line treatment for metastatic renal cell carcinoma (mRCC). • Despite durable responses in many patients, >20% experience primary progressive disease (PPD) at first restaging. • This study evaluated clinical factors and transcriptomic correlates associated with PPD. METHODS Clinical analysis | IMDC 1 Study population: IMDC real-world patients receiving first-line NIVO+IPI (n = 1,601). 2 Response definition: PPD = best overall response of PD per physician-assessed RECIST 1.1. 3 Statistical analysis: Multivariable logistic regression of baseline factors (evaluable n = 1,339). Endpoint Definitions •Time to Next-Line Therapy or Death (TNTD): Time from index date to next-line therapy or death. •Mortality Without Next-Line Therapy (MWNT): Death before initiation of next-line therapy. •Overall Survival (OS): Time from index date to death from any cause Exploratory RNA-seq | COSMIC-313 1 Transcriptomic cohort: COSMIC-313 placebo + NIVO+IPI arm: 38 PPD vs 70 CR/PR (n = 108). 2 Differential expression: DESeq2 with adjustment for liver metastasis status. 3 Pathway enrichment: GSEA using MSigDB Hallmark gene sets. Two complementary cohorts • Clinical associations and exploratory biology @elioibrahim8 @DrChoueiri CLINICAL RESULTS | IMDC PPD rate: 30.4% (487/1,601 patients) Baseline factors associated with PPD Variable | OR (95% CI) | P value Prior nephrectomy | 0.69 (0.54–0.88) | 0.003 Liver metastases | 1.67 (1.24–2.26) | 0.001 KPS <80% | 1.54 (1.12–2.12) | 0.007 Hemoglobin <LLN | 1.35 (1.04–1.74) | 0.023 Elevated neutrophils | 1.50 (1.09–2.06) | 0.013 Multivariable model: n = 1,339 | KPS, Karnofsky performance status; LLN, lower limit of normal Outcomes among patients with PPD Overall survival and subsequent-treatment outcomes Median (95% CI), months: OS: 13.9 (12.2-16.7); TNTD: 3.6 (3.4-3.9); TEFL: 2.1 (2.1-2.3) At Risk (0,3,6,9,12,15,18,21,24 months): OS 487 418 344 259 215 179 142 120 104 TNTD 487 300 117 60 38 30 23 19 15 TEFL 487 130 42 18 14 11 6 6 3 Figure 1. Kaplan-Meier outcome estimates among patients with PPD. [Figure 2: cumulative incidence — Initiating next line therapy; Death without starting next line] At risk 487 300 117 60 38 30 23 19 15 Figure 2. Next-line therapy or death before initiating treatment. TRANSCRIPTOMIC RESULTS | COSMIC-313 Hallmark pathways associated with primary progressive d[isease] GSEA adjusted for liver metastasis status | All FDR q-values < 0.001 Enriched in PPD: Fatty acid metabolism; Xenobiotic metabolism; Oxidative phosphorylation; Hypoxia; Epithelial-mesenchymal transition; Glycolysis Depleted in PPD: TNFα signaling via NF-κB (-1.81); IL6-JAK-STAT3 signaling (-1.94); Inflammatory response (-2.01); Interferon alpha response (-2.10); Interferon gamma response (-2.33); Allograft rejection (-2.49) [enriched NES values small/partly obscured] Figure 3. Hallmark pathway enrichment (adjusted for liver metastasis status). Primary PD vs Responders [volcano plot; gene labels small] Figure 4. Differential expression: primary PD versus resp[onders] [obscured] CONCLUSION PPD was associated with adverse clinical features [obscured by presenter] ... enriched tumor transcriptome. Findings are explo[ratory] [obscured]

[Slide 1] TAKE HOME MESSAGES EV - P is the preferred first-line benchmark for suitable patients Level of evidence is low to recommend any specific sequence Platinum retains activity after EV-P; expected benefit is often short FGFR3 and HER2 may open important later-line treatment options Clinical trials ++++ Perioperative exposure is changing every metastatic sequence [Slide 2] Sequencing in advanced urothelial cancer: navigating treatment decisions in 2026 1. The first-line landscape has changed 2. The post EV – pembrolizumab setting is our greatest current unmet need 3. Prior perioperative treatment is becoming a major determinant of metastatic sequencing

Does pre-cystectomy ctDNA clearance identify a group of patients for bladder preservation? EFS by ctDNA clearance status Legend: C1D1 to C4D1 cleared | C1D1 to C4D1 not cleared [KM curve: Survival probability vs Months 0-80; dashed vertical line near ~25 months with circle on cleared curve] No. at risk (months 0, 10, 20, 30, 40, 50, 60, 70, 80): C1D1 to C4D1 cleared: 153, 129, 110, 96, 86, 57, 23, 1, 0 C1D1 to C4D1 not cleared: 46, 21, 14, 9, 9, 4, 4, 1, 0 Survival table Endpoint: EFS by ctDNA clearance status (C1D1 to C4D1) C1D1 to C4D1 cleared | 153 C1D1 to C4D1 not cleared | 46

[Slide 1] [Photo of presenter standing beside poster 3620P; adjacent poster partly visible] 3620P Genomic and phenotypic characterization of GPR65, HPK1, ENPP3 and CD70, novel targets of early phase therapeutics for metastatic renal cell carcinoma (mRCC) Rahul Winayak(1), Nicholas J Salgia(2), Vitor Abreu de Góes(1), Ali Moradi(1), Salvador Jaime-Casas(1), Koral Shah(1), Xiaochen Li(1), Benjamin Mercier(1), Michelle Li(1), Jadon Fann(1), George Zhang(1), Ruchi Agarwal(1), Joann Hsu(1), Hedyeh Ebrahimi(3), Alexander Chehrazi-Raffle(1), Sumanta Kumar Pal(1) (1)Department of Medical Oncology, City of Hope Comprehensive Cancer Center, Duarte, CA, USA; (2)Department of Immunology, Roswell Park Comprehensive Cancer Center, Buffalo, NY, USA; (3)Beth Israel Deaconess Medical Center, Boston, MA, USA Background • Ongoing trials for mRCC are investigating novel inhibitors of GPR65 and HPK1 (MAP4K1), ENPP3xCD3 T cell engagers (TCEs), and CAR-T cells targeting CD70. Methods • Normalized RNA-sequencing gene expression data from IMmotion151, JAVELIN Renal 101, and HCRN-GU16-260 were analyzed. • Gene co-expression was assessed using Pearson correlation coefficients. • Patients were stratified into high/low expressors by median gene expression within each cohort. • TME phenotypes were defined using NMF clusters and compared using ANOVA and chi-square tests.(1) • PFS was analyzed using KM methods. Results: gene co-expression GPR65 and HPK1 (MAP4K1) were co-expressed with PDCD1, CTLA4 and CD274; ENPP3 and CD70 with KDR, EPAS1 and VEGFA in IMmotion151. [Correlation heatmap: CTLA4, TIGIT, LAG3, PDCD1, MAP4K1, PTPN22, GPR65, CD274, MTOR, HAVCR2, CD70, ENPP3, VEGFA, EPAS1, KDR] Results: GPR65, HPK1 (MAP4K1) GPR65 and HPK1 (MAP4K1) were more highly expressed in patients bearing an immunogenic profile, defined by NMF cluster 4 in IMmotion151, p<0.05 for both. [Stacked bar charts: Relative Abundance of NMF clusters, GPR65high/low, MAP4K1high/low] Increased GPR65 and HPK1 (MAP4K1) expression were associated with prolonged PFS to IO-containing regimens (HCRN-GU16-260, JAVELIN Renal 101, respectively), p<0.05. [KM panels: PFS - Nivolumab (GPR65 High/Low); Axitinib/Avelumab (MAP4K1 High/Low); number-at-risk rows small] Results: ENPP3, CD70 ENPP3 and CD70 were more highly expressed in patients bearing an angiogenic profile, defined by either NMF clusters 2-3 in IMmotion151, p<0.05 for both. [Stacked bar charts: ENPP3high/low, CD70high/low] Increased ENPP3 and CD70 expression were associated with prolonged PFS to sunitinib (IMmotion151), (p=0.057, p=0.002, respectively). [KM panels: Sunitinib (ENPP3 High/Low); Sunitinib (CD70 High/Low)] Conclusions • Emerging therapeutic targets in mRCC appear to have distinct associations with either angiogenic or immunogenic subtypes of the disease. • Given ongoing trials with small molecules, TCEs and CAR-T approaches directed at these entities, our data inform combinatorial strategies for their use in mRCC. Correspondence rwinayak@coh.org X @RWinayak References 1. Motzer et al., Cancer Cell. 2020 14;38(6):803-817.e4. Abbreviations TCEs, T cell engagers; CAR-T cell, chimeric antigen receptor-T cell; TME, tumor microenvironment; PFS, progression-free survival; NFM, non-nuclear matrix factorization; IO, immunotherapy Disclosures The presenting author does not have any financial disclosures. [Slide 2] 3620P Genomic and phenotypic characterization of GPR65, HPK1, ENPP3 and CD70, novel targets of early phase therapeutics for metastatic renal cell carcinoma (mRCC) Rahul Winayak(1), Nicholas J Salgia(2), Vitor Abreu de Góes(1), Ali Moradi(1), Salvador Jaime-Casas(1), Koral Shah(1), Xiaochen Li(1), Benjamin Mercier(1), Michelle Li(1), Jadon Fann(1), George Zhang(1), Ruchi Agarwal(1), Joann Hsu(1), Hedyeh Ebrahimi(3), Alexander Chehrazi-Raffle(1), Sumanta Kumar Pal(1) (1)Department of Medical Oncology, City of Hope Comprehensive Cancer Center, Duarte, CA, USA; (2)Department of Immunology, Roswell Park Comprehensive Cancer Center, Buffalo, NY, USA; (3)Beth Israel Deaconess Medical Center, Boston, MA, USA Background • Ongoing trials for mRCC are investigating novel inhibitors of GPR65 and HPK1 (MAP4K1), ENPP3xCD3 T cell engagers (TCEs), and CAR-T cells targeting CD70. Methods • Normalized RNA-sequencing gene expression data from IMmotion151, JAVELIN Renal 101, and HCRN-GU16-260 were analyzed. • Gene co-expression was assessed using Pearson correlation coefficients. • Patients were stratified into high/low expressors by median gene expression within each cohort. • TME phenotypes were defined using NMF clusters and compared using ANOVA and chi-square tests.(1) • PFS was analyzed using KM methods. Results: gene co-expression GPR65 and HPK1 (MAP4K1) were co-expressed with PDCD1, CTLA4 and CD274; ENPP3 and CD70 with KDR, EPAS1 and VEGFA in IMmotion151. [Correlation heatmap: CTLA4, TIGIT, LAG3, PDCD1, MAP4K1, PTPN22, GPR65, CD274, MTOR, HAVCR2, CD70, ENPP3, VEGFA, EPAS1, KDR] Results: GPR65, HPK1 (MAP4K1) GPR65 and HPK1 (MAP4K1) were more highly expressed in patients bearing an immunogenic profile, defined by NMF cluster 4 in IMmotion151, p<0.05 for both. [Stacked bar charts: Relative Abundance of NMF clusters, GPR65high/low, MAP4K1high/low] Increased GPR65 and HPK1 (MAP4K1) expression were associated with prolonged PFS to IO-containing regimens (HCRN-GU16-260, JAVELIN Renal 101, respectively), p<0.05. [KM panels: PFS - Nivolumab (GPR65 High/Low); Axitinib/Avelumab (MAP4K1 High/Low); number-at-risk rows small] Results: ENPP3, CD70 ENPP3 and CD70 were more highly expressed in patients bearing an angiogenic profile, defined by either NMF clusters 2-3 in IMmotion151, p<0.05 for both. [Stacked bar charts: ENPP3high/low, CD70high/low] Increased ENPP3 and CD70 expression were associated with prolonged PFS to sunitinib (IMmotion151), (p=0.057, p=0.002, respectively). [KM panels: Sunitinib (ENPP3 High/Low); Sunitinib (CD70 High/Low)] Conclusions • Emerging therapeutic targets in mRCC appear to have distinct associations with either angiogenic or immunogenic subtypes of the disease. • Given ongoing trials with small molecules, TCEs and CAR-T approaches directed at these entities, our data inform combinatorial strategies for their use in mRCC. Correspondence rwinayak@coh.org X @RWinayak References 1. Motzer et al., Cancer Cell. 2020 14;38(6):803-817.e4. Abbreviations TCEs, T cell engagers; CAR-T cell, chimeric antigen receptor-T cell; TME, tumor microenvironment; PFS, progression-free survival; NFM, non-nuclear matrix factorization; IO, immunotherapy Disclosures The presenting author does not have any financial disclosures. Number at risk (as printed, small): PFS - Nivolumab: GPR65 High 36 / 22 / 11 / 8 / 1 / 0; GPR65 Low 36 / 11 / 4 / 4 / 2 / 0 (time 0/10/20/30/40/50) Axitinib/Avelumab: MAP4K1 High 173 / 138 / 64 / 30 / 1 / 0; MAP4K1 Low 179 / 112 / 47 / 17 / 4 / 0 (time 0/5/10/15/20/25) Sunitinib: ENPP3 High 205 / 148 / 91 / 49 / 13 / 0; ENPP3 Low 211 / 129 / 74 / 45 / 12 / 0 Sunitinib: CD70 High 209 / 148 / 98 / 59 / 15 / 0; CD70 Low 207 / 129 / 67 / 35 / 10 / 0

3601P Characterizing the long tail of somatic mutations in clear-cell renal cell carcinoma (ccRCC) Eddy Saad, Kevin Meli, Ro Malik, Jad El Masri, Matias Vergara, Harry Kane, Renée Maria Saliby, Chris Labaki, Ziad Bakouny, Razan Mohanna, Berkay Simsek, Sabina Signoretti, Brendan Reardon, Jihye Park, Toni K. Choueiri, Eliezer M. Van Allen, David A. Braun BACKGROUND • Despite major therapeutic advances, precision oncology has had limited impact in the treatment of ccRCC. • Common genomic drivers are well described, but the biological and clinical significance of rare recurrent somatic alterations remains unclear. • We hypothesized that a large-scale harmonized multi-cohort analysis could uncover low-frequency yet clinically relevant molecular subsets. METHODS RESULTS • Out of 118 SMGs, 72 occurred in fewer than 2% of tumors (long tail of mutations). • We identified NF2 loss-of-function (LOF) mutations as a recurrent, low-frequency (2%) alteration across cohorts (a). • NF2 LOF was associated with upregulation of proliferative pathways and downregulation of IFN-γ signaling (b). • Clinically, NF2 LOF was associated with more aggressive disease and worse survival (c), independent of stage, sarcomatoid features, age, and treatment type (d). • Other rare LOF alterations included: MLH1 (0.8%) associated with increased MSI and TMB, a trend towards better outcomes with immunotherapy (IO) and worse survival with VEGF TKI. B2M (0.6%) associated with downregulation of IFN-γ signaling. STK11 (0.4%) associated with downregulation of immune pathways and worse outcomes with IO. Comprehensive multi-omic profiling of large harmonized datasets can reveal clinically important subsets missed by smaller studies. Rare genomic alterations can refine prognostic stratification and therapeutic selection in ccRCC. Email: eddy_saad@dfci.harvard.edu

[Slide 1] Figure 2: Active (Blue) and Planned (Gray) Sites of Enrollment for EV-209 [World map with site callouts] Indianapolis, IN, USA - Indiana University Melvin and Bren Simon Comprehensive Cancer Center - Active Downers Grove, IL, USA - Accelacare of Duly; Chicago, IL, USA - University of Illinois Cancer Center - Active Farmington Hills, MI, USA - Profound Research LLC - Active Austin, TX, USA - Urology Austin PLLC - Active; San Antonio, TX, USA - Urology San Antonio - Active Little Rock, AR, USA - Arkansas Urology Research Center - Active Myrtle Beach, SC, USA - START Carolinas - Active New York, NY, USA - Laura and Isaac Perlmutter Cancer Center - NYU Langone Health - Active; Perlmutter Cancer Center - NYU Langone Hospital [illegible]; Mount Sinai Tisch Cancer Center - Planned; Mineola, NY, USA - NYU Langone Hospital - Long Island - Active Philadelphia, PA, USA - Thomas Jefferson University - Active; Centers for Advanced Urology, LLP d/b/a MidLantic Urology - Active; Willow Grove, PA, USA - Sidney Kimmel Comprehensive Cancer - Asplundh Cancer Pavilion - Active Sewell, NJ, USA - Sidney Kimmel Comprehensive Cancer Center - Washington Township - Active; Cherry Hill, NJ, USA - Sidney Kimmel Comprehensive Cancer Center - Cherry Hill - Active London, UK - St Bartholomew's Hospital - Barts Health NHS Trust - Active Herne, Germany - Marien Hospital Herne - Active Stuttgart, Germany - Klinikum Stuttgart - Planned Poznań, Poland - Clinical Research Center, Medic-R - Active Tarnów, Poland - Szpital Wojewódzki im. św. Łukasza SPZOZ - Active Rennes, France - Centre Eugene Marquis - Active Villejuif, France - Institut Gustave Roussy - Planned Milan, Italy - IRCCS San Raffaele Hospital - Planned Toulouse, France - Oncopole Claudius Regaud, IUCT-Oncopole - Active Madrid, Spain - Hospital Universitario La Paz - Active Ourense, Spain - Complejo Hospitalario Universitario de Ourense - Active Granada, Spain - Hospital Universitario Virgen de las Nieves - Active Beijing, China - Chinese Academy of Medical Sciences and Peking Union Medical College - Planned; Cancer Hospital, Chinese Academy of Medical Sciences - Planned Goyang-si, South Korea - National Cancer Center - Active Bundang-gu, South Korea - Seoul National University Bundang Hospital - Active Seoul, South Korea - Seoul National University Hospital - Active; Asan Medical Center (AMC) - Active; Ewha Womans University Mokdong Hospital - Active; Korea University Anam Hospital - Active; Seoul St. Mary's Hospital - Active; Samsung Medical Center - Active Gwangju, South Korea - Chonnam National University Hwasun Hospital - Active Daegu, South Korea - Kyungpook National University Chilgok Hospital - Active Gifu, Japan - Gifu University Hospital - Active Tsukuba, Japan - University of Tsukuba Hospital - Active Tokyo, Japan - Institute of Science Tokyo Hospital - Active; Tokyo Metropolitan Komagome Hospital - Active; Cancer Institute Hospital of JFCR - Active Kawasaki, Japan - St. Marianna University Hospital - Active Kobe, Japan - Kobe City Medical Center General Hospital - Active Kyoto, Japan - Kyoto University Hospital - Active Osaka, Japan - Osaka Metropolitan University Hospital - Active Ube, Japan - Yamaguchi University Hospital - Active Southport, Australia - Tasman Oncology Research - Active Melbourne, Australia - Austin Health - Planned [Slide 2] Figure 1: EV-209 Study Design Patient population (N≈240) ↓ EV 1.25 mg/kg IV Q3W on Days 1 and 8 P 200 mg IV Q3W on Day 1 (4 cycles; 21-day cycles) ↓ Disease assessments • MRI or CT • Cytoscopy + biopsy cCR achieved → EV 1.25 mg/kg IV Q3W on Days 1 and 8 (5 cycles; 21-day cycles) P 200 mg IV Q3W on Day 1 (≤13 cycles; 21-day cycles) → Post-treatment follow-up • Safety, efficacy, OS (≤5 years) No cCR achieved ↔ For clinical downstage only (≤T2): EV 1.25 mg/kg IV Q3W on Days 1 and 8; P 200 mg IV Q3W on Day 1 (≤5 cycles; 21-day cycles) → Repeat disease assessment (≤9 cycles; 21-day cycles) → cCR achieved → P 200 mg IV Q3W on Day 1 (≤8 cycles; 21-day cycles) → Post-treatment follow-up Optional (based on investigator's decision) No cCR achieved → RC or SoC treatment based on local standard practice → Post-treatment follow-up

[Slide 1] MADRID 2026 ESMO congress Plasma Kidney Injury Molecule-1 Dynamics During First-Line Immunotherapy With or Without Cabozantinib in Advanced Renal Cell Carcinoma Wenxin Xu,1,a Habib Hamidi,2,a Robert J. Motzer,3 Thomas Powles,4 Laurence Albiges,5 Marc Machaalani,6 Christian Scheffold,7 Saurabh Gupta,8 Venu Valmeekam,2 Toni K. Choueiri,1 Bradley A. McGregor,6 Andrew Simmons2 1Department of Medical Oncology, Dana-Farber Cancer Institute, Brigham and Women's Hospital and Harvard Medical School, Boston, MA, USA; 2Translational Medicine, Exelixis, Inc., Alameda, CA, USA; 3Department of Medical Oncology, Memorial Sloan Kettering Cancer Center, New York, NY, USA; 4Department of Medical Oncology, Barts Cancer Institute, Cancer Research UK Experimental Cancer Medicine Centre, Queen Mary University of London, Royal Free National Health Service Trust, London, UK; 5Department of Cancer Medicine, Institut Gustave Roussy, Université Paris Saclay, Villejuif, France; 6Department of Medical Oncology, Dana-Farber Cancer Institute, Boston, MA, USA; 7Clinical Development, Exelixis, Inc., Alameda, CA, USA; 8Translational Development, Solid Tumor Oncology, Translational Medicine, Bristol Myers Squibb, Princeton, NJ, USA aCo-primary authors [Slide 2] Summary How a Blood Biomarker Predicts Treatment Benefit in Kidney Cancer Scientific Insight: KIM-1 reflects biologically distinct RCC states Clinical Relevance: Early KIM-1 change may predict benefit from dual IO therapy Development Strategy: Biomarker-driven escalation with TKI in patients with IO-resistant biology Potential Impact: • Biomarker-informed trial design • Precision treatment • Increased probability of response IO, immuno-oncology; KIM-1, kidney injury molecule-1; RCC, renal cell carcinoma; TKI, tyrosine kinase inhibitor.

[Poster board number] 3624P [ESMO Congress Madrid Spain 23-27 October 2026 logo] FPN: 3624P PCSK9 inhibitor evolocumab combined with nivolumab in refractory metastatic renal cell carcinoma: Stage I analysis of BOOST-RCC Tian Zhang1, Rebecca Slack Tidwell2, Qian Qin1, Lauren Wood2, Amy Rowell1, Molly McGuire1, Navneet Kaur1, Caroline S. Greenstone1, Wadih Issa1, Shahed Abdullah1, Zhongya Wang2, Hans Hammers1, James Brugarolas1, Andrew Hahn2, Amishi Shah2, Pavlos Msaouel2, Nizar Tannir2, Andrew Wang1, Eric Jonasch2 1Harold C Simmons Comprehensive Cancer Center, Department of Internal Medicine, UT Southwestern Medical Center, Dallas, TX; 2Department of Genitourinary Oncology, MD Anderson Cancer Center, Houston, TX Background • Despite advancements in the treatment of metastatic renal cell carcinoma (mRCC), there are few treatments in the immunotherapy (IO) refractory setting. • Preclinical studies have shown that PCSK9 inhibition combined with IO stunt tumor growth in multiple solid tumor models by downregulating MHC-1 expression and tumor antigen presentation, independent on its effect on cholesterol (Figure 1).1 • PCSK9 inhibition combined with IO has been shown to improve overall survival in patients with metastatic solid tumors (Figure 2).2 • We conducted an investigator-initiated trial combining PCSK9 inhibitor evolocumab with IO nivolumab in IO-refractory mRCC, BOOST-RCC (NCT06284564). Figure 1 [PCSK9 pathway diagram] Figure 1. Depiction of PCSK9 pathway for MHC-I expression1. MHC-1 is degraded in presence of PCSK9. When PCSK9 is absent or inhibited, MHC-1 is recycled, resulting in higher tumor antigen presentation and T-cell activation. Figure 2 [KM plot: Overall survival, PCSK9 users vs Statin users; Log-rank P=.006; numbers at risk small, illegible] Figure 2. Overall survival in patients with metastatic cancer receiving IO combined with either PCSK9 inhibitors vs statins.2 Methods • BOOST-RCC is an investigator-led, phase 2 clinical trial which enrolled patients at MD Anderson and UT Southwestern between June 2024 and February 2025. • A Simon 2-stage design was used with 10 patients in the first cohort. Stage 1 would be passed if >1/10 patients had an objective response. • Patients were treated with evolocumab, a PCSK9 inhibitor, and nivolumab every 4 weeks and median follow up was 18 months. • Circulating tumor cells (CTCs) were captured on a dendrimer-based device3, and HLA detected by cell surface immunofluorescence at baseline and on-treatment timepoints. Plasma PCSK9 and cytokines were measured via Multiplex ELISA. LDL and cholesterol were measured as standard of care testing. • Paired, two-tailed t-tests with Welch corrections were performed. All timepoints were compared to baseline. Table 1. Baseline patient characteristics. N=10; Age (median, range) 67 (56.5, 69.5); [remaining rows small/partially obscured — illegible] Figure 3 [Best percentage change in target lesions, waterfall — values illegible] Figure 4 [Treatment and follow-up by patient, swimmer plot — illegible] [Center text obscured by presenter] Correlative Findings Figure 6 [HLA ratio plots A-F; illegible] Figure 7 A) LDL B) Cholesterol [C1, C2, EOT] Figure 7. Correlative LDL and cholesterol levels on treatment. As expected, PCSK9 inhibition led to decrease in LDL (A) and cholesterol (B) from baseline to on treatment. Asterisks represent significance. Figure 6. HLA expression detected on CTCs for all 10 patients (A-C) and Patient 9 with stable disease. (D-F). HLA expression was measured as ratio to PanCK, DAPI, and total expression on CTCs. Figure 8 A) PCSK-9 B) MCP-1/CCL2 C) TNFα D) CCL4 E) IL-10 [C1, C2, C3, EOT, Follow up] [Figure 8 legend partially obscured] ...Cytokine levels including PCSK9. Observed ...centrations of PCSK9 increase on treatment (A). ...icant cytokine changes include MCP-1/CCL2 ...C), CCL4 (D) and IL-10 (E). Asterisks: statistical... Conclusions • In the BOOST-RCC trial, 9 of 10 patients experienced progression of disease. • One patient (patient 9) with multiple lines of prior treatment had stable disease for 14 months, with a 16% decrease in target lesions and on-target effect of increased HLA expression on CTCs. • The most common related grade 3 adverse events were fatigue (1) and elevated AST(1). There were no trial limiting toxicities. • Evolocumab combined with nivolumab had on-target effects with significant decrease in plasma LDL and cholesterol and significant increase of plasma PCSK9 on treatment. • Given no objective response to PCSK9 and nivolumab in this first stage population, this trial was stopped for further recruitment. This highlights heterogeneity of IO resistance in IO-refractory mRCC. References [partially obscured; illegible] Disclosures TZ: [small text, illegible] Twitter/X: @tiansterzhang | Email: tian.zhang@utsouthwestern.edu Presented at ESMO Annual Congress 2026 | October 23, 2026 | Madrid, Spain

[Slide 1] Study Design • This is an open-label, single arm , phase Ib/II trial to evaluate the efficacy and safety of Ivonescimab monotherapy as first-line treatment for favourable risk advanced renal cell carcinoma (aRCC) Key eligibility criteria • Histologically or cytologically confirmed RCC with mainly clear cell carcinoma with advanced disease • IMDC favourable risk (IMDC score 0) • No prior systemic therapy • ECOG PS 0-1 Ivonescimab 20mg/kg Q3W N=3~6 Ivonescimab 30mg/kg Q3W N=3~6 → Ivonescimab 30mg/kg Q3W (N=30) No DLT was observed. RP2D of ivonescimab was 30 mg/kg Q3W. Primary endpoint: ORR Secondary endpoints: PFS, OS, DCR, DoR, TTR, safety PFS, progression-free survival; OS, overall survival; ORR, overall response rate; DCR, disease control rate; DoR, duration of response; TTR, time to response; DLT, dose limiting toxicity; RP2D: Recommended Phase 2 Dose; Q3W, Every 3 Weeks; Jun Guo, MD [Slide 2] Efficacy-PFS • Median follow-up was 13.3 months • The median PFS was not reached. The 12-month and 24-month PFS rates were 85.5% and 55.4%, respectively • No deaths have occurred to date 12m-rate 85.5% 24m-rate 55.4% Follow-up time (months) AK112 At risk | 23 | 22 | 18 | 14 | 9 | 7 | 5 | 2 | 1 | 0 Censored | 0 | 0 | 3 | 6 | 11 | 11 | 12 | 15 | 16 | 17 Events | 0 | 1 | 2 | 3 | 3 | 5 | 6 | 6 | 6 | 6 Jun Guo, MD Data cutoff date: July 14, 2026 Survival table Endpoint: PFS Ivonescimab (AK112) | 23 | not reached Landmark rates: 12 months: Ivonescimab (AK112) 85.5% 24 months: Ivonescimab (AK112) 55.4% [Slide 3] Conclusions ➢ Ivonescimab monotherapy demonstrated promising antitumor activity in patients as first line treatment with IMDC favorable-risk aRCC, and further follow-up for survival endpoints is ongoing. • ORR was 65.2%, and DCR was 95.7%. • The median PFS was not reached. • The 12-month and 24-month PFS rates were 85.5% and 55.4%, respectively. ➢ Ivonescimab monotherapy showed a manageable safety profile. • The majority of TRAE were grade 1-2, and grade ≥3 TRAEs only occurred in 7 (26.9%) patients. ➢ Ivonescimab monotherapy could be a potentially safe and effective treatment strategy for patients with favorable-risk aRCC Jun Guo, MD

[Slide 1] STAMPEDE ADT HR (95% CI): 1.72 (1.34 - 2.21) p-value: <0.001* Distant metastasis (%) vs Years since randomisation (0-8) Number at risk (% distant metastasis) GC > 0.85: 101 | 91 (10%) | 64 (27%) | 9 (39%) GC 0.61 - 0.85: 65 | 62 (5%) | 50 (12%) | 13 (20%) GC ≤ 0.60: 42 | 40 (2%) | 36 (5%) | 10 (5%) STAMPEDE ADT + abi HR (95% CI): 1.27 (0.96 - 1.67) p-value: 0.10 Number at risk (% distant metastasis) GC > 0.85: 105 | 96 (4%) | 69 (16%) | 17 (19%) GC 0.61 - 0.85: 58 | 57 (0%) | 53 (2%) | 14 (11%) GC ≤ 0.60: 30 | 30 (0%) | 27 (3%) | 6 (14%) Decipher risk group: GC ≤ 0.60 | GC 0.61 - 0.85 | GC > 0.85 Survival table Endpoint: Distant metastasis cumulative incidence by Decipher risk group (STAMPEDE ADT arm; ADT + abi arm in notes) ADT: GC > 0.85 | 101 ADT: GC 0.61-0.85 | 65 ADT: GC ≤ 0.60 | 42 ADT + abi: GC > 0.85 | 105 ADT + abi: GC 0.61-0.85 | 58 ADT + abi: GC ≤ 0.60 | 30 HR: ADT: 1.72; ADT + abi: 1.27 HR CI: ADT: 1.34-2.21; ADT + abi: 0.96-1.67 P: ADT: <0.001*; ADT + abi: 0.10 [Slide 2] STAMPEDE ADT HR (95% CI): 1.43 (1.14 - 1.81) p-value: 0.002* Prostate cancer specific mortality (%) vs Years since randomisation (0-8) Number at risk (% PCSM) GC > 0.85: 101 | 98 (2%) | 80 (15%) | 49 (26%) GC 0.61 - 0.85: 65 | 65 (0%) | 55 (8%) | 43 (12%) GC ≤ 0.60: 42 | 41 (2%) | 38 (5%) | 34 (7%) STAMPEDE ADT + abi HR (95% CI): 1.60 (1.16 - 2.21) p-value: 0.004* Number at risk (% PCSM) GC > 0.85: 105 | 97 (4%) | 78 (10%) | 60 (22%) GC 0.61 - 0.85: 58 | 57 (0%) | 53 (2%) | 47 (5%) GC ≤ 0.60: 30 | 30 (0%) | 28 (3%) | 22 (3%) Decipher risk group: GC ≤ 0.60 | GC 0.61 - 0.85 | GC > 0.85 Survival table Endpoint: Prostate cancer-specific mortality by Decipher risk group (STAMPEDE ADT arm; ADT + abi arm) ADT: GC > 0.85 | 101 ADT: GC 0.61-0.85 | 65 ADT: GC ≤ 0.60 | 42 ADT + abi: GC > 0.85 | 105 ADT + abi: GC 0.61-0.85 | 58 ADT + abi: GC ≤ 0.60 | 30 HR: ADT: 1.43; ADT + abi: 1.60 HR CI: ADT: 1.14-1.81; ADT + abi: 1.16-2.21 P: ADT: 0.002*; ADT + abi: 0.004* [Slide 3] STAMPEDE All: Treatment x Decipher Interaction for DM: p = 0.11 STAMPEDE N0: Treatment x Decipher Interaction for DM: p = 0.01 Y-axis: 10-year distant metastasis (%) (0, 25, 50) X-axis: Decipher (0.25, 0.50, 0.75, 1.00) Legend: RT + ADT (solid) | RT + ADT + Abi (dashed) Distribution strips: RT + ADT + Abi | RT + ADT [Slide 4] Endpoint / Cohort | Adj. (s)HR (95% CI) | p-value DM RWD NCCN high | 1.38 (1.27 - 1.50) | <0.001* RWD NCCN very high | 1.39 (1.17 - 1.66) | <0.001* MFS RWD NCCN high | 1.29 (1.23 - 1.36) | <0.001* RWD NCCN very high | 1.29 (1.15 - 1.44) | <0.001* OS RWD NCCN high | 1.23 (1.14 - 1.32) | <0.001* RWD NCCN very high | 1.23 (1.03 - 1.47) | 0.02* X-axis: Adjusted (subdist.) hazard ratio per 0.1 (0.75-1.75)

[Slide 1] NIAGARA: Background - Muscle-invasive bladder cancer (MIBC) is molecularly heterogeneous, with recurrent alterations in cell-cycle regulation, receptor tyrosine kinase signalling, chromatin remodelling, and DNA damage response (DDR) pathways, alongside high genomic instability. In multiple studies, DDR alterations have been associated with improved response to cisplatin-based neoadjuvant chemotherapy (NAC)1-3 - Transcriptomic-based molecular subtypes further shape prognosis and treatment response by reflecting differences in differentiation state, stromal and immune contexture, and underlying biology4 - NIAGARA is a Phase 3, open-label, randomised trial assessing perioperative durvalumab plus NAC and radical cystectomy (RC) vs NAC and RC alone in patients with MIBC5 - Results from the interim analyses demonstrated that the addition of perioperative durvalumab to NAC resulted in statistically significant improvements in event-free survival (hazard ratio [HR] 0.68) and overall survival (HR 0.75), and a 10% improvement in pathological complete response rate vs NAC5 Here, we report exploratory analyses of MIBC molecular drivers and subtypes in relation to treatment benefit from NIAGARA 1. Geynisman DM, et al. J Clin Oncol. 2025;43:1113-1122. 2. Miron et al. Eur Urol Oncol. 2020;3:544-547. 3. Liu D, et al. Jama Oncol. 2016;2:1094-1096. 4. Robertson AG, et al. Cell. 2017;171:540-556.e25. 5. Powles T, et al. N Engl J Med. 2024;391:1773-1786. [Slide 2] NIAGARA: Conclusions - Baseline WES and RNA-seq captured the landscape of key mutational drivers and transcriptome-based subtypes, characteristic of MIBC - Consistent with cisplatin sensitivity, DDR-related gene alterations (eg, ERCC2) were significantly associated with a higher pCR rate across arms - Exploratory integrated baseline genomics and transcriptomics from NIAGARA suggest a broad benefit for perioperative durvalumab plus NAC in patients with MIBC - Shared biology associated with sensitivity to immunotherapy was observed across molecular subtypes of MIBC and was enriched in the basal squamous and luminal infiltrated subtypes, which appeared to derive the most benefit with the addition of perioperative durvalumab to NAC Exploratory analyses show the addition of perioperative durvalumab to NAC consistently improves EFS across a diverse molecular landscape Scan the QR code to access: - Slide presentation - Plain language summary We are also presenting data on the UC with DDHS subgroup from NIAGARA - please see poster: 4764P Outcomes in Patients With Muscle-Invasive Bladder Cancer With Divergent Differentiation or Histologic Subtypes From NIAGARA DDHS, divergent differentiation or histological subtype; DDR, DNA damage response; EFS, event-free survival; MIBC, muscle-invasive bladder cancer; NAC, neoadjuvant chemotherapy; pCR, pathological complete response; UC, urothelial carcinoma; WES, whole-exome sequencing.

[Slide 1] Enfortumab Vedotin plus Pembrolizumab in collecting duct and renal medullary carcinoma (REPRINT – MeetURO 4/25 trial). Preliminary results from Simon's first stage (NCT06302569) Giuseppe Procopio, MD Genitourinary Medical Oncology Prostate Program Fondazione IRCCS Istituto Nazionale dei Tumori Milan, Italy 25th Oct 2026 [Slide 2] REPRINT Study Design Key Eligibility criteria • Histologically confirmed diagnosis of CDC and RMC (central pathology review required) • Metastatic or advanced disease with measurable lesions according to RECIST 1.1 • ECOG PS 0-1 N of Patients = 23 Pembrolizumab 200 mg IV q21 + Enfortumab Vedotin 1,25 mg/kg 1,8q21 → X 3 cycles → Pembrolizumab 200 mg q21 until progression, unacceptable toxicity or 24 months* *Patients who will experience PD RECIST v1.1 during Pembrolizumab monotherapy could re-start Enfortumab Vedotin Simon's two stage optimal design I stage: 2/9 responses required II stage: 6/14 responses required Primary endpoint: ORR Secondary endpoints: PFS, OS and safety profile [abbreviation footnote, too small to read: illegible] Giuseppe Procopio [Slide 3] REPRINT Results OBJECTIVE RESPONSE 22% ORR (2/9 pts) 2/2 RCCU-MP vs 0/7 CDC DISEASE CONTROL 55% DCR (5/9 pts) 2 PR + 3 SD PROGRESSION-FREE SURVIVAL 5.2 mo Median PFS 95% CI: 3.4–NR SAFETY PROFILE 0% Grade 3–4 AEs All pts received 3 cycles P+EV RCCU-MP Cohort (n=2) 100% ORR • Patient 1: Complete Response (CR) in lung disease; Stable Disease (SD) in nodal sites. • Patient 2: Partial Response (PR) in lung disease. Pie: PR 22% | SD 33% | PD 45% CDC Cohort (n=7) 0% ORR • No objective responses observed (0/7 pts). • Disease Control Rate (DCR) achieved in 3/7 pts (SD). Giuseppe Procopio [Slide 4] Conclusion • Pembrolizumab + Enfortumab Vedotin demonstrated preliminary antitumor activity in RMC and CDC, meeting the prespecified criteria to proceed beyond the first stage of Simon's two-stage design in the REPRINT trial • Patients with SMARCB1-deficient RMC not associated with sickle hemoglobinopathies derived promising benefit from the study combination • The study is ongoing and continue to recruit (NCT06302569), with biological samples being collected for traslational analyses Giuseppe Procopio

[Slide 1] 16:15 - 17:45 Emerging therapeutic strategies in advanced prostate cancer CHAIRS: ANA APARICIO, AURELIUS OMLIN KLK2 × CD3 T-cell engager: pasritamig 174 patients (phase 1) 42.4% PSA50 at RP2D Stein et al. J Clin Oncol 2025 — pasritamig (JNJ-78278343) First-in-class KLK2 × CD3 bispecific. RP2D: 3.5 mg day 1, 18 mg day 8, 300 mg day 15, then every 6 weeks. Largest T-cell-engager dataset in prostate cancer to date. [Waterfall plot of PSA change] Table: All-Treated Population (SC/IV) N=174 | RP2D Safety Population (IV Only) N=45 Participants with ≥1 TRAE, n (%): 144 (82.8) | 27 (60.0) Serious TRAEs, n (%): 12 (6.9) | 2 (4.4)a Grade ≥3 TRAEs, n (%): 17 (9.8) | 2 (4.4) TRAEs leading to treatment discontinuation, n (%): 1 (0.6) | 0 KLK2 is more prostate-restricted than PSMA, which may widen the therapeutic window. Stein MN et al. J Clin Oncol 2025;43:2515-2526 (pasritamig). Niven Mehra — Emerging immune-based therapies BILBAO AUDITORIUM - HALL 3 [Slide 2] 16:15 - 17:45 Emerging therapeutic strategies in advanced prostate cancer CHAIRS: ANA APARICIO, AURELIUS OMLIN STEAP1 × CD3 T-cell engager: xaluritamig [antibody schematic: Fab, CD3 scFv, Fc] Kelly et al. Cancer Discov 2024 — xaluritamig (AMG 509) STEAP1 × CD3, XmAb 2+1 immune therapy, mCRPC. [Waterfall plot of PSA change; legend and axis labels illegible] Safety: CRS in 72% (grade ≥3 in 2%), largely confined to cycle 1 under step-up dosing Off-tumor on-target MSK events common and well managed by NSAIDs, steroids Why STEAP1 matters STEAP1 (six-transmembrane epithelial antigen of the prostate 1) is associated with proliferation and metastatic potential Broadly expressed, including in AR-indifferent and neuroendocrine disease; retained as the AR axis is exhausted. [IHC images: PSMA, STEAP1] STEAP1 staining detected in 87.7% of evaluable matched mCRPC tissues Patient-level analysis: STEAP1 expression in 95% of patients compared to 68% for PSMA Kelly WK et al. Cancer Discov 2024;14:76-89 (xaluritamig phase 1). Niven Mehra — Emerging immune-based therapies BILBAO AUDITORIUM - HALL 3 [Slide 3] 16:15 - 17:45 Emerging therapeutic strategies in advanced prostate cancer CHAIRS: ANA APARICIO, AURELIUS OMLIN CAR-T in prostate cancer: 3 recent phase 1 trials PSCA-CAR T: Second-generation 4-1BB/CD3ζ CAR, ± lymphodepletion. Dorff, Nat Med 2024 [construct schematic - City of Hope; labels illegible] BPX-601 GoCAR-T + rimiducid: First-generation CD3ζ CAR + rimiducid-inducible MyD88/CD40 · Stein, Nat Commun 2024 [construct schematic; labels illegible] P-PSMA-101 · stem-memory CAR: Second-generation (Centyrin binder + 4-1BB), piggyBac, iCasp9 off-switch · Slovin, CCR 2026 [construct schematic; labels illegible] Recognition has been established; persistence and the therapeutic window are not. In prostate cancer, CAR-T remains at proof-of-concept stage. Dorff TB et al. Nat Med 2024;30:1636–44 (NCT03873805) · Stein MN et al. Nat Commun 2024 (NCT02744287) · Slovin SF et al. Clin Cancer Res 2026;32:3540–51 (NCT04249947). Niven Mehra — Emerging immune-based therapies BILBAO AUDITORIUM - [HALL 3, partly obscured]

CONCLUSIONS • There is a high prevalence of predicted CH among HRR gene mutations in liquid-biopsy samples (56% overall), including in BRCA1 (32%) and PALB2 (24%). • Patients with predicted CH-only HRR alterations do not appear to benefit from PARPi monotherapy, highlighting the importance of accounting for HRR mutation origin in therapy selection. • Some CH-predicted mutations are enriched in samples after PARPi exposure, including TP53, NF1, pTERT and CBL. Emmanuel S. Antonarakis, MD Emmanuel Antonarakis Distinguishing clonal hematopoiesis (CH) from tumor homologous recombination repair (HRR) alterations in liquid biopsy: Implications for PARP inhibitor (PARPi) use in metastatic castration-resistant prostate cancer (mCRPC)

Decipher score identifies RWD cases at higher distant metastasis risk Left panel: RWD:NCCN high risk | Right panel: RWD:NCCN very high risk Y-axis: Distant metastasis (%) (0-50); X-axis: Years since randomisation (0-8) Number at risk (% distant metastasis) - RWD:NCCN high risk: GC > 0.85: 2230 | 982 (9%) | 52 (14%) | 15 (16%) GC 0.61-0.85: 1565 | 745 (4%) | 43 (5%) | 21 (5%) GC ≤ 0.60: 1402 | 659 (2%) | 54 (3%) | 23 (5%) Number at risk (% distant metastasis) - RWD:NCCN very high risk: GC > 0.85: 517 | 191 (25%) | 10 (29%) | 4 (36%) GC 0.61-0.85: 157 | 76 (7%) | 6 (10%) | 2 (10%) GC ≤ 0.60: 85 | 40 (9%) | 2 (12%) | 1 (12%) Decipher risk group: GC ≤ 0.60 (blue) | GC 0.61 - 0.85 (red) | GC > 0.85 (purple) Sarah Howlett Survival table Endpoint: Distant metastasis cumulative incidence by Decipher (GC) risk group, real-world data NCCN high risk, GC > 0.85 | 2230 NCCN high risk, GC 0.61-0.85 | 1565 NCCN high risk, GC ≤ 0.60 | 1402 NCCN very high risk, GC > 0.85 | 517 NCCN very high risk, GC 0.61-0.85 | 157 NCCN very high risk, GC ≤ 0.60 | 85 Landmark rates: at-risk column 2 (year not explicitly printed): NCCN high risk, GC > 0.85 9% | NCCN high risk, GC 0.61-0.85 4% | NCCN high risk, GC ≤ 0.60 2% | NCCN very high risk, GC > 0.85 25% | NCCN very high risk, GC 0.61-0.85 7% | NCCN very high risk, GC ≤ 0.60 9% at-risk column 3 (year not explicitly printed): NCCN high risk, GC > 0.85 14% | NCCN high risk, GC 0.61-0.85 5% | NCCN high risk, GC ≤ 0.60 3% | NCCN very high risk, GC > 0.85 29% | NCCN very high risk, GC 0.61-0.85 10% | NCCN very high risk, GC ≤ 0.60 12% at-risk column 4 (year not explicitly printed): NCCN high risk, GC > 0.85 16% | NCCN high risk, GC 0.61-0.85 5% | NCCN high risk, GC ≤ 0.60 5% | NCCN very high risk, GC > 0.85 36% | NCCN very high risk, GC 0.61-0.85 10% | NCCN very high risk, GC ≤ 0.60 12%

[Slide 1] MADRID 2026 ESMO congress Plasma Kidney Injury Molecule-1 Dynamics During First-Line Immunotherapy With or Without Cabozantinib in Advanced Renal Cell Carcinoma Wenxin Xu,1,a Habib Hamidi,2,a Robert J. Motzer,3 Thomas Powles,4 Laurence Albiges,5 Marc Machaalani,6 Christian Scheffold,7 Saurabh Gupta,8 Venu Valmeekam,2 Toni K. Choueiri,1 Bradley A. McGregor,6 Andrew Simmons2 aCo-primary authors [affiliations as on title slide: Dana-Farber Cancer Institute; Exelixis, Inc.; Memorial Sloan Kettering Cancer Center; Barts Cancer Institute; Institut Gustave Roussy; Bristol Myers Squibb] [Slide 2] Higher Baseline KIM-1 Levels Are Associated With IMDC Poor-Risk Disease, Absence of Prior Nephrectomy, and Greater Tumor Burden IMDC Risk: Wilcoxon, P = 1e-10 (Intermediate vs Poor; KIM-1 (pg/mL)) Prior Nephrectomy: Wilcoxon, P < 2.2e-16 (No vs Yes; KIM-1 (pg/mL)) Radiographic Tumor Burden: R = 0.58, P < 2.5e-16 (Week 1 KIM-1 (pg/mL) vs Sum of diameters at week 1) Box plots depict median values with interquartile ranges. IMDC, International Metastatic Renal Cell Carcinoma Database Consortium; KIM-1, kidney injury molecule-1. [Slide 3] Early KIM-1 Decrease Is Predictive of Clinical Outcome in IO Arm (N+I) Key Finding: • A >30% decrease at week 3 strongly predicts PFS, OS, and ORR (59% vs. 31%) with N+I • Under triplet therapy, this threshold does not stratify outcomes, indicating distinct biology Distribution by Arm (Density vs Percent Change from W1D1 to W4D4; Treatment: C+N+I, N+I; 30% Decrease line) Placebo + Nivolumab + Ipilimumab (N+I) PFS: HR 0.46 (95% CI, 0.33–0.64); N+I >30% Decrease (n=109); N+I ≤30% Decrease (n=212); P < 0.0001 OS: HR 0.55 (95% CI, 0.38–0.78); N+I >30% Decrease (n=109); N+I ≤30% Decrease (n=212); P = 0.00081 Cabozantinib + Nivolumab + Ipilimumab (C+N+I) PFS: HR 0.83 (95% CI, 0.58–1.19); C+N+I >30% Decrease (n=246); C+N+I ≤30% Decrease (n=68); P = 0.31 OS: HR 1.01 (95% CI, 0.69–1.47); C+N+I >30% Decrease (n=246); C+N+I ≤30% Decrease (n=68); P = 0.96 Median follow up 45.0 months. CI, confidence interval; D, day; HR, hazard ratio; IO, immuno-oncology; KIM-1, kidney injury molecule-1; N+I, nivolumab + ipilimumab; ORR, objective response rate; OS, overall survival; PFS, progression-free survival; W, week. Survival table Endpoint: PFS and OS by week-3 KIM-1 decrease (>30% vs <=30%), within each arm N+I >30% decrease | 109 N+I <=30% decrease | 212 C+N+I >30% decrease | 246 C+N+I <=30% decrease | 68 HR: N+I PFS 0.46; N+I OS 0.55; C+N+I PFS 0.83; C+N+I OS 1.01 HR CI: 0.33–0.64; 0.38–0.78; 0.58–1.19; 0.69–1.47 P: N+I PFS P<0.0001; N+I OS P=0.00081; C+N+I PFS P=0.31; C+N+I OS P=0.96

[Slide 1] Patient reported outcomes in resected renal cell carcinoma: adjuvant durvalumab versus active monitoring in the RAMPART trial Sophie Merrick, Elena Frangou, Laura Murphy, Hannah L Rush, Grisma Patel, Hannah Plant, Matthew Nankivell, Paul Nathan, Grant D Stewart, Craig Gedye, Axel Bex, Tim Eisen, Ian D Davis, Balaji Venugopal, Thomas Powles, Mahesh Parmar, Martin Stockler, James Larkin, Angela Meade Speaker: Sophie Merrick, on behalf of the RAMPART Investigators 25th October 2026 [Slide 2] Conclusion • The primary QoL endpoint (OHQL at Month 15) showed no evidence of a difference between arms. • Detailed domain-specific analyses identified clinically meaningful deterioration in QoL during treatment, with some effects persisting and others emerging later. • A similar pattern was observed in the previously reported DT vs AM comparison, where the primary OHQL endpoint showed no evidence of a difference despite clinically meaningful QoL impacts across domains at different timepoints. Sophie Merrick #RAMPART

[Slide 1] First-in-Human Phase I Study of UCL70802, a LACO-Stim armored Anti-CD70 CAR-T Cell Therapy, in patients with Metastatic clear cell renal cell carcinoma (mccRCC) • GAN XINXIN • The First Affiliated Hospital of Naval Medical University [Slide 2] CONCLUSIONS UCL70802 Phase I — Patient Characteristics and Dose-Level Outcomes (mccRCC, n = 11) Data cut-off: 27 Aug 2026 · 11 participants enrolled and infused · 11 efficacy-evaluable Patient ID | Age / Sex | Prior lines of therapy | ECOG | Lymphodepletion | Infused dose (x10^6/kg) | Cmax (copies/µg DNA) | Best response | CRS grade 1 | 54 y / M | 1 | 1 | Standard | 0.6 | 12,486 | PD | G1 2 | 70 y / F | 2 | 1 | None | 1.2 | 209 | SD | G1 3 | 54 y / M | 5 | 1 | Standard | 1.2 | 1,368 | SD | G2 4 | 61 y / M | 2 | 1 | Standard | 2.0 | 37,750 | SD | G2 5 | 59 y / M | 3 | 0 | Standard | 2.0 | 61,565 | PR | G1 6 | 62 y / M | 2 | 0 | Standard | 2.0 | 19,565 | PR | G3 7 | 73 y / F | 2 | 1 | Standard | 2.0 | 107,135 | SD | G3 8 | 66 y / M | 1 | 1 | Standard | 3.0 | 115,067 | PR | G3 9 | 58 y / M | 2 | 0 | Standard | 3.0 | 32,655 | SD | G3 10 | 62 y / M | 2 | 0 | Standard | 3.0 | 25,439 | PR | G3 11 | 43 y / F | 2 | 0 | Standard | 3.0 | 109,021 | SD | G2 Standard lymphodepletion: cyclophosphamide 300 mg/m² × 3 days + fludarabine 30 mg/m² × 3 days; Pt 2 received no lymphodepletion. CONCLUSIONS — UCL70802 demonstrates manageable safety and promising preliminary antitumor activity in previously treated mccRCC. With dual mechanisms — CAR-mediated precise tumor targeting and LACO-Stim–mediated systemic immune activation — it may overcome the resistance bottleneck of conventional CAR-T therapy. Dose escalation is ongoing to further characterize clinical efficacy and safety. 3615RO

[Slide 3] Strategies for antigen-directed therapies Surface antigens | HLA-restricted antigens Personalized: CAR T cell (autologous); scFv; Surface antigen; TIL therapy; TCR; Engineered TCR therapy; Neoantigen vaccine 'Off the shelf': CAR T cell (allogeneic); Bispecific antibodies; Monoclonal antibodies; Antibody–drug conjugates; Antibody–radionuclide conjugate; ImmTAC; Shared antigen vaccine; Identify shared antigens (such as ERVs and tumour-associated antigens) Tumor cell; Tumour antigen (presented on HLA class I) [...] al, Nature Reviews Clinical Oncology, 2021. [Slide 4] Initial conclusions • Neoantigen vaccines are feasible to manufacture, even in modest mutation burden tumors like renal cell carcinoma • Neoantigen vaccines can effectively elicit T cell responses • Vaccination can induce anti-tumor immunity • Preliminary signals of clinical activity in the adjuvant setting, but more and larger studies are needed – rationale for INTerpath-004 (NCT06307431)

[Slide 1] Median PFS by INV was 10.4 months in the pooled DE/A/K EV+P cohort Pooled DE/A/K EV+P cohort Events/n | Median PFS (95% CI), mo 77/121 | 10.4 (8.2, 20.9) Landmark labels: 39.9% (30.6, 49.1); 29.1% (20.1, 38.6); 24.3% (15.1, 34.7); 20.2% (10.6, 32.1) No. at risk: 121 102 79 55 49 42 38 35 34 32 28 23 23 22 21 20 19 17 15 11 9 7 7 6 6 5 4 2 1 1 0 EV monotherapy cohort Events/n | Median PFS (95% CI), mo 50/73 | 6.4 (6.1, 8.4) Landmark labels: 13.7% (5.6, 25.5); 13.7% (5.6, 25.5); 13.7% (5.6, 25.5) No. at risk: 73 57 40 20 15 13 9 7 5 5 5 5 5 5 5 5 4 4 1 1 1 1 1 1 1 0 Data cutoff: February 20, 2026. Tumor response data evaluated by BICR have been previously published,1,2 and BICR review was no longer performed after November 2024. BICR, blinded independent central review; DE, dose escalation; EV, enfortumab vedotin; INV, investigator; P, pembrolizumab; PFS, progression-free survival. 1. Rosenberg JE, et al. ESMO Open. 2026;11:107700. 2. O'Donnell PH, et al. J Clin Oncol. 2023;41:4107-4117. Presented by: Dr. Jose Pablo Maroto Rey Survival table Endpoint: PFS by investigator (INV) Arm | N | Median | Median CI Pooled DE/A/K EV+P | 121 | 10.4 mo | 8.2, 20.9 EV monotherapy (Cohort K) | 73 | 6.4 mo | 6.1, 8.4 Landmark rates: 24 mo: Pooled DE/A/K EV+P 39.9% (CI 30.6, 49.1) | EV monotherapy (Cohort K) 13.7% (CI 5.6, 25.5) 48 mo: Pooled DE/A/K EV+P 29.1% (CI 20.1, 38.6) | EV monotherapy (Cohort K) 13.7% (CI 5.6, 25.5) 60 mo: Pooled DE/A/K EV+P 24.3% (CI 15.1, 34.7) | EV monotherapy (Cohort K) 13.7% (CI 5.6, 25.5) 84 mo: Pooled DE/A/K EV+P 20.2% (CI 10.6, 32.1) [Slide 2] Median DOR by INV was 29.6 months in the pooled DE/A/K EV+P cohort Pooled DE/A/K EV+P cohort ORR by INV, n/N (%) | Events/n | Median DOR (95% CI), mo 81/121 (66.9) | 41/81 | 29.6 (10.3, 71.9) Landmark labels: 53.7% (41.3, 64.6); 41.1% (28.5, 53.3); 37.0% (23.6, 50.4); 30.8% (16.2, 46.8) No. at risk: 81 67 59 46 40 37 36 32 29 27 22 21 21 19 19 17 16 15 12 10 8 7 7 6 5 5 2 1 1 0 EV monotherapy cohort ORR by INV, n/N (%) | Events/n | Median DOR (95% CI), mo 33/73 (45.2) | 18/33 | 10.2 (4.4, 19.2) Landmark labels: 29.4% (12.3, 48.8); 29.4% (12.3, 48.8) No. at risk: 33 29 16 13 10 8 6 5 5 5 5 5 5 5 5 4 4 2 0 Data cutoff: February 20, 2026. Tumor response data evaluated by BICR have been previously published,1,2 and BICR review was no longer performed after November 2024. BICR, blinded independent central review; DE, dose escalation; DOR, duration of response; EV, enfortumab vedotin; INV, investigator; ORR, overall response rate; P, pembrolizumab. 1. Rosenberg JE, et al. ESMO Open. 2026;11:107700. 2. O'Donnell PH, et al. J Clin Oncol. 2023;41:4107-4117. Presented by: Dr. Jose Pablo Maroto Rey Survival table Endpoint: Duration of response by investigator (INV) Arm | N | Median | Median CI Pooled DE/A/K EV+P | 81 | 29.6 mo | 10.3, 71.9 EV monotherapy (Cohort K) | 33 | 10.2 mo | 4.4, 19.2 Landmark rates: 24 mo: Pooled DE/A/K EV+P 53.7% (CI 41.3, 64.6) | EV monotherapy (Cohort K) 29.4% (CI 12.3, 48.8) 48 mo: Pooled DE/A/K EV+P 41.1% (CI 28.5, 53.3) | EV monotherapy (Cohort K) 29.4% (CI 12.3, 48.8) 60 mo: Pooled DE/A/K EV+P 37.0% (CI 23.6, 50.4) 84 mo: Pooled DE/A/K EV+P 30.8% (CI 16.2, 46.8) [Slide 3] Phase 1b/2 EV-103 (KEYNOTE-869)a Study Overview • EV-103 first demonstrated durable clinical benefit with 1L EV+P in cisplatin-ineligible la/mUC, supporting the initial indication in this population1,2 • EV-302 (KEYNOTE-A39)b subsequently confirmed these findings and established EV+P as a standard of care in la/mUC3-5 • This long-term follow-up evaluated clinical benefit durability and long-term safety of EV+P • With the longest follow-up of EV+P in la/mUC to date, these results may help guide real-world management and optimize patient benefit EV-103 dose escalation and expansion cohorts A and K (1L cisplatin-ineligible la/mUC)2,6 Dose Escalation (DE) | EV+P (n=5)c Recommended EV dose → Cohort A (expansion cohort) | EV+P (n=40) Cohort Kd (randomized cohort) 1:1 randomizatione | EV+P (n=76) | EV monotherapy (n=73) This analysis | Pooled EV+P DE/A/K cohort (N=121) | EV monotherapy Cohort K (n=73) • Baseline characteristics were previously reported and consistent with the population1,7 • At data cutoff (February 20, 2026), all patients were off study treatment • Median follow-upf: 64.2 months (pooled EV+P) and 58.6 months (EV monotherapy) aNCT03288545. bNCT04223856. cParticipants who received EV 1.25 mg/kg. dStudy was not designed to compare EV+P vs EV monotherapy. eOf 151 randomized in EV-103 Cohort K, 149 received ≥1 dose of study treatment. fMedian follow-up for DE/Cohort A (n=45) and EV+P arm in Cohort K (n=76) was 89.0 months and 58.8 months, respectively. 1L, first line; DE, dose escalation; EV, enfortumab vedotin; la/mUC, locally advanced/metastatic urothelial carcinoma; P, pembrolizumab. 1. Rosenberg JE, et al. ESMO Open. 2026;11:107700. 2. Friedlander TW, et al. ESMO 2025. Abstract 3074P (Poster). 3. Powles T, et al. N Engl J Med. 2024;390:875-888. 4. Powles T, et al. Ann Oncol. 2025;36:1212-1219. 5. Powles T, et al. Ann Oncol. 2024;35:485-490. 6. Rosenberg JE, et al. ASCO 2020. Abstract 5044 (Poster). 7. O'Donnell PH, et al. J Clin Oncol. 2023;41:4107-4117. Presented by: Dr. Jose Pablo Maroto Rey

npj | precision oncology Brief communication Published in partnership with The Hormel Institute, University of Minnesota Article in Press https://doi.org/10.1038/s41698-026-01745-y An individual sample predictive model to classify patients with metastatic renal cell carcinoma Anupama Reddy1, W. Jared Brewer1, Scott Haake2 & Brian Rini2 Using RNAseq data from the IMmotion 151 clinical trial (IM151), 7 RCC biologic clusters have been defined. Several groups have attempted to predict clusters on RCC datasets without observing the association with therapy response from IM151. We developed IRIS-RCC (Individual RNA-seq Intrinsic Subtyping for RCC; available at https://irisrcc.vindhyadatascience.com/), to predict molecular clusters using an individual-sample model. IRIS-RCC normalizes gene expression within a given sample using ratiometric gene expression values.

REPRINT Results Biomarker / Feature | Responder #1 | Responder #2 Histological Subtype | RCCU-MP | RCCU-MP SMARCB1 Status | Deficient | Deficient Membranous NECTIN-4 | Negative | Negative PD-L1 CPS | 10% | 20% Tumor Mutational Burden | 1.6 mut/Mbp | 3.8 mut/Mbp INI1 (SMARCB1) loss in tumor cells with preserved expression in control lymphocytes, a defining diagnostic feature of renal medullary carcinoma Giuseppe Procopio

Conclusions: EV-103 cohorts DE, A, and K • This analysis reports the longest follow-up to date from the phase 1b/2 EV-103 study (64.2 mo for the pooled EV+P cohort), demonstrating sustained responses and survival with 1L EV+P in patients with la/mUC who are cisplatin ineligible • Median OS was 27.7 months in the pooled EV+P cohort and 22.8 months in the EV monotherapy cohort, and an estimated 33% of patients in the pooled EV+P cohort and 23.9% in the EV monotherapy cohort were alive at 5 years – These data are notable given the historically poor outcomes in advanced UC (eg, 9.6% 5-y relative survival for metastatic bladder cancer, SEER; 8.4% of cisplatin-ineligible patients alive at 4.5-y median follow-up, EORTC 30986)(1,2) • Consistent with previous reports,(3-5) the safety profile was manageable and most treatment-related AEs of special interest for EV improved or resolved over time and no new safety signals were observed • These data are consistent with EV-302, reinforcing the long-term benefit-risk profile, sustained survival benefits, and 1L standard-of-care role of EV+P in la/mUC 1L, first line; AE, adverse event; DE, dose escalation; EV, enfortumab vedotin; la/mUC, locally advanced/metastatic urothelial carcinoma; P, pembrolizumab. [References small print, partly legible: 1. National Cancer Institute. Cancer Stat Facts: Bladder Cancer ... Accessed August 2026. 2. De Santis M, et al. J Clin Oncol. 2012;30:191-199. 3. Rosenberg JE, et al. ESMO Open. 2026[?] ... 4. Friedlander TW, et al. ESMO 2025. Abstract 3074P (Poster). 5. O'Donnell PH, et al. J Clin Oncol. 2023;41:4107-4117.] Presented by: Dr. Jose Pablo Maroto Rey

[Slide 1] CATALINA-1 Study Design Phase 1 Dose Escalation and Expansion with Ixo-V Monotherapy in Solid Tumors GCT patient population - Confirmed advanced or metastatic solid tumor malignancy - No available standard of care therapy - Not amenable to surgery - Prior HDCT and ASCT allowed - For dose expansion: CLDN6 positivitya by IHC was required Key endpoints Primary: - AEs by CTCAE 5.0 Secondary: - Best objective response by investigator per RECIST v1.1 Escalation First-in-human starting dose (0.2 mg/kg Q3W) -> Accelerated dose titration (1 patient per cohort) -> Grade ≥2 related AE -> Standard "3 + 3" dose escalation -> Recommended phase 2 dose Expansion - CLDN6-positive platinum-resistant ovarian cancer expansion - CLDN6-positive refractory NSCLC expansion - Other CLDN6-positive refractory cancers expansion (including GCT) [highlighted] - CLDN6-low/negative expansion - CLDN6-low/positive ovarian cancer Ixo-V + combinations (including bevacizumab) - CLDN6-positive NSCLC Ixo-V + pembrolizumab AEs, adverse events; ASCT, autologous stem cell transplantation; CLDN6, claudin 6; CTCAE, Common Terminology Criteria for Adverse Events; GCT, germ cell tumors; HDCT, high-dose chemotherapy; IHC, immunohistochemistry; Ixo-V, ixotatug vedotin; NSCLC, non-small cell lung cancer; Q3W, every 3 weeks; RECIST, Response Evaluation Criteria in Solid Tumors. aCLDN6 positivity was defined as >30% tumor cell membrane staining of any intensity by IHC. TORL BioTherapeutics Data on File, 2026. [Slide 2] CLDN6 as a Novel Therapeutic Target in GCT - CLDN6 is a cell surface protein and a member of the claudin family of tight junction molecules1 - CLDN6 is highly expressed in multiple human cancers, including GCT, with little or no expression in healthy adult tissues2-4 - GCT are the most common solid malignancy in adolescent and young adult men and are typically diagnosed in the third and fourth decades of life5-7 - Although most patients with newly-diagnosed GCT are successfully treated with platinum-based chemotherapy, 20-30% relapse5,8,9 - Long-term survival is achieved in <10% of patients who progress post-HDCT/ASCT5,10 - Chemotherapy regimens produce ORRs of ~6-51%9 - Clinically effective salvage therapies are limited5,9 [Figure: CLDN6 mRNA Expression Across Tumor Types2 — Log(FKPM)-RNA-seq by tumor type (AML n=173 ... Testicular cancer n=156, boxed); tumor-type labels small] Reprinted from Clinical Cancer Research, 2023;29(11):2131-2143, McDermott MSJ, et al., with permission from AACR. There is high unmet clinical need for novel therapeutic options in recurrent GCT. AML, acute myeloid leukemia; CLDN6, claudin 6; GCT, germ cell tumors; HDCT/ASCT, high-dose chemotherapy with autologous stem cell transplant; ORRs, objective response rates. 1. Qu H, et al. Int J Mol Sci. 2021;22(24):13416. 2. McDermott MSJ, et al. Clin Cancer Res. 2023;29(11):2131-2143. 3. Silverstein J, et al. Presented at: SGO Annual Meeting on Women's Cancer; April 10-13, 2026; San Juan, PR. 4. Yang Z, et al. Pharmacol Res. 2026;230:108276 5. Feldman DR, et al. Cancer Med. 2026;15:e72017. 6. Cancer stat facts: testicular cancer. Accessed October 5, 2026. https://seer.cancer.gov/statfacts/html/testis.html. 7. JAMA patient page: what are testicular germ cell tumors? Accessed October 7, 2026. https://jamanetwork.com/journals/jama/fullarticle/2834695. 8. McHugh DJ, Feldman DR. Adv Urol. 2018; 7272541. 9. Oltrecolli M, et al. Ther Adv Med Oncol. 2026; 18:1-15. 10. Feldman DR, et al. Cancer. 2012;118:981-986. [Slide 3] Baseline Demographics | | Ixo-V <2.4 mg/kg (n=3) | 2.4 mg/kg (n=3) | 3 mg/kga (n=4) | 3.4 mg/kgb (n=1) | Total N=11 | | Median age, years (range) | 30.0 (27.0, 44.0) | 45.0 (31.0, 76.0) | 39.0 (26.0, 46.0) | 23.0 (23.0, 23.0) | 37.0 (23.0, 76.0) | | Sex, n (%) Male | 3 (100.0) | 3 (100.0) | 3 (75.0) | 1 (100.0) | 10 (90.9) | | Female | 0 | 0 | 1 (25.0) | 0 | 1 (9.1) | | ECOG PS at baseline, n (%) 0 | 2 (66.7) | 1 (33.3) | 3 (75.0) | 1 (100.0) | 7 (63.6) | | 1 | 1 (33.3) | 2 (66.7) | 1 (25.0) | 0 | 4 (36.4) | | Number of prior lines of therapy, n (%) 1-2 | 0 | 1 (33.3) | 1 (25.0) | 0 | 2 (18.2) | | ≥3 | 3 (100.0) | 2 (66.7) | 3 (75.0) | 1 (100.0) | 9 (81.8) | | Median prior lines of therapy, n (range) | 4.0 (3.0, 5.0) | 4.0 (1.0, 4.0) | 3.0 (2.0, 4.0) | 3.0 (3.0, 3.0) | 3.0 (1.0, 5.0) | | Type of prior therapy, n (%) Platinum compounds | 3 (100.0) | 3 (100.0) | 4 (100.0) | 1 (100.0) | 11 (100.0) | | Taxanes | 3 (100.0) | 2 (66.7) | 3 (75.0) | 0 | 8 (72.7) | | HDCT/ASCT | 3 (100.0) | 0 | 2 (50.0) | 1 (100.0) | 6 (54.5) | - Among enrolled patients, GCT primary tumor sites were testicular (n=9), ovarian (n=1), and mediastinal (n=1) - 10 of 11 patients had nonseminoma and 1 had seminoma - CLDN6 expression was positive (>30%) in 10 of 10 patients (100%) tested by IHCc CLDN6, claudin 6; ECOG PS, Eastern Cooperative Oncology Group performance status; G-CSF, granulocyte colony stimulating factor; GCT, germ cell tumors; HDCT/ASCT, high-dose chemotherapy with autologous stem cell transplant; IHC, immunohistochemistry; Ixo-V, ixotatug vedotin. a2 of the 4 patients treated with Ixo-V at this dose level received prophylactic G-CSF. bThis patient received prophylactic G-CSF. cOut of the 11 total patients, 1 patient specimen was missing. [Slide 4] Response Rates in Patients with GCTa,b [Waterfall plot: Best change from baseline in sum of diameter, % by Patient ID (OC04, T01, T06, T04, T03, T07, T08, T02, T05, S07); dose legend 0.4 / 1.0 / 2.0 / 2.4 / 3.0 / 3.4 mg/kg; markers: Ovarian tumor, Mediastinal tumor, Received prior HDCT/ASCT, * Tx ongoing] | Parametera | All dose levels (N=11) | | BOR, n (%) PR | 5 (45.5) | | SD | 3 (27.3) | | PD | 2 (18.2) | | NE | 1 (9.1) | | ORR, % (95% CI) | 45.5 (16.7-76.6) | | DCR, % (95% CI) | 72.7 (39.0-94.0) | | mPFS, months (95% CI) | 5.8 (1.4-13.8) | aDOR was immature at data cutoff date. - Responses were observed at Ixo-V dose levels ranging from 1 mg/kg to 3.4 mg/kg, with a median follow-up of 13.8 months - Three of 6 (50.0%) patients who received prior HDCT/ASCT had confirmed responses to Ixo-V - One patient with testicular GCT has maintained a response for 47.4 months with treatment ongoing at time of data cutoff Ixo-V monotherapy exhibits preliminary antitumor clinical activity at doses ≥1 mg/kg in heavily pretreated patients with recurrent GCT, including those with disease refractory to ASCT BOR, best overall response; DCR, disease control rate; DOR, duration of response; GCT, germ cell tumors; HDCT/ASCT, high-dose chemotherapy with autologous stem cell transplant; Ixo-V, ixotatug vedotin; mPFS, median progression-free survival; NE, not evaluable; ORR, objective response rate; PD, progressive disease; PR, partial response; SD, stable disease; Tx, treatment. aData cutoff was March 16, 2026. bPrimary sites for the 11 GCT patients include testicular (n=9), ovarian (n=1), and mediastinal (n=1).

[Slide 1] Background Papillary renal cell carcinoma (PRCC) is a rare type of kidney cancer, representing approximately 10-15% of all renal cell carcinomas1 S1500 (PAPMET) established cabozantinib as an effective first-line treatment compared to other TKIs based on improved PFS (5.6 months vs 9.0 months for sunitinib and cabozantinib respectively), though no difference in OS was observed across the study arms2,3 More recent single-arm trials have shown high ORR and long PFS with combination TKI/IO therapy. • Keynote-B61 tested lenvatinib/pembrolizumab with an ORR of 45% and median PFS of 17.5 months4 in the PRCC cohort 1. Saigia NJ et al. Am Soc Clin Oncol Educ Book. 2024 Jun;44(3):e438642 2. Pal SK, et al. Lancet. 2021 Feb 20;397(10275):695-703 3. Barata P, et al. J Clin Oncol. 2024 Nov 20;42(33):3911-3916. Leading cancer research. Together. SWOG [Slide 2] Conclusions S2200 (PAPMET2) closed prematurely due to slow accrual likely reflecting an evolving standard of care with greater adoption of combination therapy for PRCC based on single-arm studies The observed prolonged PFS with cabozantinib/atezolizumab compared to cabozantinib supports the hypothesis that combination therapy is an appropriate first-line treatment for PRCC Rare cancer trials can be challenging to conduct due to multiple factors including high costs at individual sites as well as low patient volumes. Additional structural changes are warranted to better support rare cancer trials in the future. Results from STELLAR-304 (NCT05678673), SAMETA (NCT05043090) and AXIPAP (NCT02489695) testing monotherapy versus combination therapies are still pending Leading cancer research. Together. SWOG

[Slide 2] POTOMAC: Early High-Risk Disease Events in Patients With Papillary Only and With T1 Tumours The addition of D to BCG (I+M) resulted in fewer early high-risk recurrences, which are associated with worse outcomes a,b Papillary only: D+BCG (I+M) | BCG (I+M) [swimmer plots; y-axis Patients; x-axis Time from randomisation to high-risk event (months) 0-12] Early high-risk events: 16 patients in D+BCG (I+M) arm vs 29 patients in BCG (I+M) arm T1: D+BCG (I+M) | BCG (I+M) Early high-risk events: 10 patients in D+BCG (I+M) arm vs 20 patients in BCG (I+M) arm aPost-hoc exploratory analysis. bIncludes patients who experience high-risk events as per DFS, except death. Early high-risk recurrences: within 12 months from randomisation. T1 disease: papillary disease stage T1, without CIS. Data cutoff 03 April 2025. BCG, bacillus Calmette-Guérin; CIS, carcinoma in situ; D, durvalumab; DFS, disease-free survival; I, induction; M, maintenance. [Slide 3] POTOMAC: Conclusions • Durvalumab plus BCG (I+M) reduced the risk of high-risk disease recurrence or death by 44% in patients with papillary only tumours compared with BCG (I+M) alone (HR 0.56, 95% CI, 0.37–0.84)1 • In patients with papillary only tumours, the addition of durvalumab to BCG (I+M) therapy led to: • Fewer early high-risk disease recurrences • Fewer BCG unresponsive recurrences • Fewer cystectomies performed, with a favourable trend for cystectomy-free survival • Similar results were observed in patients with T1 disease • Safety of durvalumab plus BCG (I+M) in patients with papillary only tumours was consistent with the overall safety population and with the known safety profiles of the individual therapies • The majority of imAEs were low grade and most resolved or were resolving at the data cutoff; imAEs were consistent with the known safety profile of durvalumab These data further support 1 year of durvalumab in combination with BCG (I+M) as a new treatment option for BCG-naïve, high-risk NMIBC, including for patients with papillary only tumours [QR code] Scan the QR code to access: • Slide presentation • Plain language summary 1. De Santis M, et al. Lancet. [illegible small print reference]; BCG, bacillus Calmette-Guérin; CI, confidence interval; I, induction; imAE, immune-mediated adverse event; HR, hazard ratio; M, maintenance; NMIBC, non-muscle invasive bladder cancer.

In Patients Predicted to Be N+I KIM-1 Non-Decreasers, Addition of a TKI Results in Significantly Longer PFS and Higher ORR Predicting N+I KIM-1 Non-Decreasers (COSMIC-313) N+I-treated patients / C+N+I-treated patients -> Baseline tumor RNA-seq -> Pre-trained RNA-seq model -> Predicted >30% KIM-1 decrease on N+I (Yes/No) ANALYSIS POPULATION: Predicted NOT to achieve >30% KIM-1 decrease on N+I N+I (n=109) vs. C+N+I (n=130) Definition of non-decreaser: Predicted to NOT have >30% KIM-1 decrease on N+I Rationale: These patients are unlikely to benefit from N+I alone based on predicted lack of early KIM-1 decrease. Does adding a TKI improve outcomes? PFS in Predicted Non-Decreasers (COSMIC-313) Number at risk (0,10,20,30,40,50 months): C+N+I 130 | 61 | 44 | 28 | 15 | 0; N+I 109 | 41 | 22 | 9 | 5 | 0 Cox Proportional Hazard Model (PFS) Study (All ITT Patients*): HR 0.76 (95% CI 0.59–0.98); P = 0.0318 Predicted Non-Decreasers: HR 0.63 (95% CI 0.46–0.87); P = 0.0052 In patients predicted NOT to achieve a >30% KIM-1 decrease on N+I, adding cabozantinib to N+I reduces risk of progression by 37% Best Overall Response in Predicted Non-Decreasers (COSMIC-313) N+I | C+N+I CR: 2.0% | 6.5% PR: 31.7% | 42.3% SD: 37.6% | 41.5% PD: 28.7% | 9.8% ORR (CR+PR) 34% vs. 49%; P = 0.02 Disease Control Rate (CR+PR+SD) 71% vs. 90%; P = 0.0005 *RNA-seq evaluated. C, cabozantinib; CI, confidence interval; CR, complete response; HR, hazard ratio; I, ipilimumab; ITT, intention to treat; KIM-1, kidney injury molecule-1; N, nivolumab; ORR, objective response rate; PD, progressive disease; PFS, progression-free survival; PR, partial response; SD, stable disease; TKI, tyrosine kinase inhibitor.

[Slide 1] Schema mPRCC • Histologically confirmed diagnosis of mPRCC • Measurable disease • 0-1 prior therapy allowed • Zubrod PS 0-2 • No prior IO or cabo Randomize 1:1 → Cabozantinib 60mg | Cabozantinib 60mg Atezolizumab 1200mg Stratification Factors • IMDC (favorable vs intermediate/poor) • Sarcomatoid expression (present vs absent) • Prior systemic therapy for metastatic pRCC: Yes vs. no Total Accrual 200 patients Primary Endpoint Progression-Free Survival Secondary Endpoints Overall Survival Objective Response Rate Safety (per CTCAE) SWOG [Slide 2] Baseline Characteristics Characteristic | All Patients(n=33) | Cabozantinib(n=16) | Cabozantinib/Atezolizumab(n=17) Median Age (range), years | 67 (22.8-90.3) | 62.7 (22.8-90.3) | 67.3 (50.7-81.0) Sex [n (%)] Males | 23 (70%) | 11 (69%) | 12 (71%) Females | 10 (30%) | 5 (31%) | 5 (29%) Hispanic [n (%)] Yes | 3 (9%) | 2 (13%) | 1 (6%) No | 30 (91%) | 14 (88%) | 16 (94%) Race [n (%)] White | 20 (61%) | 9 (56%) | 11 (65%) Black | 6 (18%) | 3 (19%) | 3 (18%) Asian | 3 (9%) | 1 (6%) | 2 (12%) Multi-Racial | 1 (3%) | 1 (6%) | 0 (0%) Unknown | 3 (9%) | 2 (13%) | 1 (6%) Prior Systemic Therapy [n (%)] Yes | 3 (9%) | 1 (6%) | 2 (12%) No | 30 (91%) | 15 (94%) | 15 (88%) Sarcomatoid Features [n (%)] Yes | 2 (6%) | 1 (6%) | 1 (6%) No | 31 (94%) | 15 (94%) | 16 (94%) IMDC Risk Score [n (%)] Favorable Risk | 11 (33%) | 5 (31%) | 6 (35%) Intermediate or Poor Risk | 22 (67%) | 11 (69%) | 11 (65%) Histological Subtype [n (%)] Type I | 6 (18%) | 3 (19%) | 3 (18%) Type II | 8 (24%) | 3 (19%) | 5 (29%) Unknown | 19 (58%) | 10 (62%) | 9 (53%) Zubrod PS 0 | 22 (67%) | 12 (75%) | 10 (59%) 1 | 10 (30%) | 3 (19%) | 7 (41%) 2 | 1 (3%) | 1 (6%) | 0 (0%) SWOG [Slide 3] Progression Free Survival Treatment Arm: Cabo Only | Cabo + Atezo PFS Events: 14 (Cabo Only); PFS Events: 9 (Cabo + Atezo) Cabo 10 months Cabo/Atezo 25 months HR 0.46 95% CI 0.22 – 0.97 One-sided P-value 0.043 Number at risk (0, 5, 10, 15, 20, 25, 30 months) Cabo Only 16 | 13 | 8 | 2 | 0 | 0 | 0 Cabo + Atezo 17 | 12 | 9 | 5 | 5 | 3 | 2 SWOG Survival table Endpoint: Progression-free survival Cabo Only | 16 | 10 months Cabo + Atezo | 17 | 25 months HR: 0.46 HR CI: 0.22 – 0.97 P: One-sided P-value 0.043

[Slide 1] Diagnostic and Staging Performance of CAIX-Targeted [68Ga]Ga-DPI-4452 PET/CT in Clear Cell Renal Cell Carcinoma: A Prospective Trial Presenting author Liangyou Gu, MD Senior Department of Urology Chinese PLA General Hospital, Beijing, China October 25, 2026 [Slide 2] Results: Cohort A TABLE1 Patient Characteristics (n = 30) Characteristic | Value Age (y) | 57.2 (52.3–61.8) Sex Female | 14 (46.7) Male | 16 (53.3) Pathology ccRCC | 23 (76.7) Unclassified RCC | 2 (6.7) pRCC | 2 (6.7) chRCC | 1 (3.3) AML | 1 (3.3) UTUC | 1 (3.3) No. positive lesions* Primary lesions | 26 Lymph nodes | 6 Adrenal glands | 2 Lungs | 24 Pancreas | 3 Bone | 2 Other | 5 TABLE 2 Diagnostic Efficacy of [68Ga]Ga-DPI-4452 PET/CT Variable | Primary tumor | Extrarenal metastases | Regional lymph nodes | Distant metastases Sensitivity (%) | 100 | 97.7 | 100 | 97.3 Total no. lesions | 23 | 43 | 6 | 37 No. lesions identified | 23 | 42 | 6 | 36 95% CI (%) | 85.2–100.0 | 87.7–99.9 | 54.1–100.0 | 85.8–99.9 Specificity | 85.7 | 100 | NA | 100 Total no. lesions | 7 | 1 | NA | 1 No. lesions identified | 6 | 1 | NA | 1 95% CI (%) | 42.1–99.6 | 2.5–100.0 | NA | 2.5–100.0 Accuracy | 96.7 | 97.7 | 100 | 97.4 Total no. lesions | 30 | 44 | 6 | 38 No. lesions identified | 29 | 43 | 6 | 37 95% CI (%) | 82.8–99.9 | 88.0–99.9 | 54.1–100.0 | 86.2–99.9 *Patients could have more than 1 positive lesion. RCC = renal cell carcinoma; pRCC = papillary renal cell carcinoma; chRCC = chromophobe renal cell carcinoma; AML = angiomyolipoma; UTUC = upper tract urothelial carcinoma. NA = not applicable Dr. Liangyou Gu

[Slide 2] Background 01 Clinical need Conventional imaging may not reliably characterize renal masses or detect small-volume metastatic disease. Unmet need: more specific diagnosis and staging Wu Q, et al. 2025. 02 Biological target CAIX is a cell-surface protein expressed in more than 90% of ccRCC, making it an attractive imaging target. Tumor biology: VHL/HIF pathway activation Queen, A., et al. 2022. 03 Imaging approach The peptide [68Ga]Ga-DPI-4452 enables rapid, same-day, whole-body CAIX-targeted PET/CT. Potential value: primary diagnosis and metastatic detection Galetic I., et al. 2024 Can [68Ga]Ga-DPI-4452 improve the diagnosis and staging of ccRCC? Dr. Liangyou Gu [Slide 3] Study design Cohort A - Inclusion • Age ≥18 years • CT/MRI/US suggests RCC • Surgery or biopsy within 4 weeks after PET/CT • ECOG ≤ 2 Patient screening Cohort B - Inclusion • Age ≥18 years • Suspected or confirmed advanced ccRCC • Paired DPI-4452 and FDG PET/CT before treatment • ECOG ≤ 2 Diagnostic validation: Cohort A → [68Ga]Ga-DPI-4452 PET/CT → Surgery or biopsy within 4 weeks → Cohort A - Endpoint • Primary ccRCC diagnosis: Histopathology-referenced diagnostic performance • Metastatic disease: Incremental detection versus conventional imaging Head-to-head comparison: Cohort B → [68Ga]Ga-DPI-4452 PET/CT / Paired comparison / [18F]FDG PET/CT → Within 14 days → Cohort B - Endpoint • Metastatic lesion detection: DPI-4452 versus FDG • Secondary analyses: Detection by site and size; quantitative uptake Trial registration: ChiCTR2600121783 Dr. Liangyou Gu [Slide 4] Conclusion and Future perspectives From disease detection to CAIX-guided precision theranostics What We Showed CAIX-targeted molecular imaging provides complementary diagnostic and biologic information in ccRCC • Accurate identification of primary ccRCC • Superior detection over 2-[18F]FDG for lymph node, lung and bone metastases, particularly small-volume lesions Potential Clinical Pathway • Precision Diagnosis and Staging: Noninvasive ccRCC characterization and detection of small-volume disease • Personalized Patient Management: Guide treatment planning and response monitoring • CAIX-Targeted Theranostics: Help select patients for radioligand therapy Next step: Multicentre validation of clinical impact, patient outcomes, and radioligand therapy dosimetry. Dr. Liangyou Gu

[Slide 1] Fig 1: Piechart of the distribution of deep PSA response — Deep PSA reponse 65%; 35% PSA response according to treatment (Chi-square p<0.001) Deep PSA response: ADT 36.6%; ADT + D 38.3%; ADT + ARPI 76.6%; ADT + D + ARPI 67.5% Non-deep PSA response: ADT 63.4%; ADT + D 61.7%; ADT + ARPI 23.4%; ADT + D + ARPI 32.5% Legend: ADT, ADT + D, ADT + ARPI, ADT + D + ARPI Fig 2: Barplot showing the distribution of deep PSA response according to treatment [Slide 2] CRPC-FS Deep PSA response 88m; Non-deep response 13.3m; p<0.001 Y-axis: Survival probability; X-axis: Time (months) 0–192 Number at risk (0, 24, 48, 72, 96, 120, 144, 168, 192) Non-deep (orange): 211, 32, 6, 1, 0, 0, 0, 0, 0 Deep (blue): 390, 244, 77, 29, 11, 3, 1, 1, 0 CRPC-FS Multivariate analysis (Hazard Ratio (95% CI), log scale 0.1–10) PSA<0.2 p<0.001; Stage IV at diagnosis; High volume p 0.009; ADT + D vs ADT p 0.03; ADT + ARPI vs ADT p 0.02; Triplet vs ADT p 0.036 OS Deep PSA response 83.5m; Non-deep response 31.5m; p<0.001 X-axis: Time (months) 0–240 Number at risk (0, 24, 48, 72, 96, 120, 144, 168, 192, 216, 240) Non-deep (orange): 211, 116, 45, 17, 3, 0, 0, 0, 0, 0, 0 Deep (blue): 392, 279, 111, 41, 20, 6, 4, 4, 2, 1, 0 OS Multivariate analysis (Hazard Ratio (95% CI)) PSA<0.2 p<0.001; Stage IV at diagnosis; High volume p<0.001; ADT + D vs ADT; ADT + ARPI vs ADT; Triplet vs ADT Survival table Endpoint: CRPC-free survival and overall survival by deep PSA response Deep PSA response | CRPC-FS 88m; OS 83.5m Non-deep response | 211 | CRPC-FS 13.3m; OS 31.5m P: <0.001 (both endpoints) [Slide 3] Differentially expressed genes in non-deep vs deep PSA responders [Volcano plot; y-axis: -Log10 Adjusted P-value; x-axis: Log2 Fold Change; legend padj < 0.01 FALSE/TRUE; gene labels largely illegible at this resolution, e.g. PPFIA4, PAX5, CFAP54, ABCA12, CDH13, KRT4, SPINK1 [illegible remainder]] Fig 4: Volcano plot of differentially expressed genes (RNA-seq) in patients with non-deep PSA response (n=74) compared with deep PSA response (n= 16) Hallmark pathways in non-deep vs deep PSA responders HALLMARK_E2F_TARGETS CIN70 CIN25 HALLMARK_DNA_REPAIR HALLMARK_GLYCOLYSIS NEUROENDOCRINE GOBP_EPITHELIAL_CELL_DIFFERENTIATION HALLMARK_PI3K_AKT HALLMARK_WNT_BETA_CATENIN_SIGNALING HALLMARK_P53_PATHWAY POLYCOMB_TARGET_GENES HALLMARK_FATTY_ACID_METABOLISM KEGG_NOTCH_SIGNALING_PATHWAY HALLMARK_ESTROGEN_RESPONSE_EARLY HALLMARK_ESTROGEN_RESPONSE_LATE HALLMARK_NOTCH_SIGNALING IDNAV_INTERFERON_SIGNATURE_IN_CANCER HALLMARK_EPITHELIAL_MESENCHYMAL_TRANSITION HALLMARK_KRAS_SIGNALING_DN HALLMARK_CHOLESTEROL_HOMEOSTASIS HALLMARK_IL2_STAT5_SIGNALING HALLMARK_KRAS_SIGNALING_UP HALLMARK_COMPLEMENT HALLMARK_ANDROGEN_RESPONSE HALLMARK_IL6_JAK_STAT3_SIGNALING HALLMARK_INFLAMMATORY_RESPONSE Enriched in non-deep responders padj < 0.1: FALSE / TRUE; padj 0.25 / 0.50 / 0.75 Normalized Enrichment Score (NES) Fig 5: Dot plot showing gene set enrichment analysis (RNA-seq) between patients with non-deep PSA response (n=16) and deep responders (n=74) [Heatmap] Deep responders | Non-deep responders 0.36 Tumor suppresor genes 0.047 Androgen receptor 0.029 NF-kB 0.026 Cell cycle 0.042 Estrogen Receptor 0.21 Neuroendocrine 0.028 Immune Normalized expression: -1 0 1 Fig 6: Heatmap of differentially expressed nCounter genes in deep PSA responders (n=280) versus non-deep responders (n=169), grouped by gene signatures [Bar plot] Non-deep responders / Deep responders TP53 33.5% / 26.3% PTEN 17.9% / 11.1% BRCA2 5.1% / 4.0% CHEK2 5.1% / 4.0% ATM 2.6% / 5.1% TMPRSS2-ERG 5.1% / 5.1% RB1 5.1% / 3.0% SPOP 0% / 4.0% [arrow highlighting SPOP] AR-V7 2.6% / 1.0% BRAF 2.6% / 1.0% Fig 7: Bar plot showing the frequency of somatic mutations in deep responders (n=39) versus non-deep responders (n=98).

[Top-left photo, partial forest plot on screen, partially legible]: Participants with no definitive deterioration were censored at last PRO assessment regardless of PD or recurrence. Participants who did not have postbaseline PRO assessment were censored at ... [cropped] Database cutoff: October 27, 2025. HR (95% CI) Favors EV + Pembro | Favors Cis + Gem [Bottom-right photo, discussant slide, top cropped]: ... indication [cropped] BENEFIT-HARM Uncertain: +20 per 100 with grade 3-4 TRAEs, some permanent MAGNITUDE Small: ~7 fewer high-risk recurrences per 100; progression unchanged CERTAINTY Low: few events; composites partly driven by deaths; inconsistent across trials PATIENT VALUES Preference-sensitive: bladder preservation vs toxicity and time APPLICABILITY (EU) Not authorised: EU application withdrawn (June 2026); FDA-approved since May 2026 What would change the assessment Individual-patient meta-analysis of POTOMAC, CREST and ALBAN with prespecified subgroups · causes of death and competing-risk analysis · mature OS · patient-reported outcomes · a predictive biomarker Consistent with the overall trial. In my reading, not sufficient to change the current recommendation (EAU 2026: discuss benefits and harms in selected patients). MJ Ribal · Uro-Oncology Unit · Clinic Comprehensive Cancer Center Barcelona

[Slide 1] The Trial Design of NCT05723991 (GUARD-01) An Exploration of ADC plus chemotherapy [Mechanism graphic: Gemcitabine, MMAE, cell-cycle G1/S/G2/M] Eligibility Criteria ➢Clinical stage T2–4aNxM0 ➢HER2 IHC 2+ or 3+ ➢Intolerance to or refusal of cisplatin-based chemotherapy ➢ECOG PS 1–2 ➢Planned to undergo RC + PLND C1 C2 C3 Day 1 RC48 2mg/kg IV, Day 2 Gemcitabine 1000mg/m2 IV; 14 days/cycle, 3 or 4 cycles C1 C2 C3 C4 Imaging assess / pathologic assess ➢RC ➢PLND ➢Refuse to RC Safety surveillance Primary endpoint: ➢ Pathological complete response rate Secondary endpoints: ➢ Pathological downstaging rate ➢ Event-free survival ➢ Safety and tolerability A Phase II Multicenter Study of Neoadjuvant Disitamab Vedotin Combined with Gemcitabine in MIBC Patients with HER2 2-3+: Interim Prognostic Outcomes MADRID 2026 ESMO congress [Slide 2] 4 more patients (Since study completion, enrollment has been expanded) ,who were treated the same as the 26-patients sample were included into pCR, EFS/OS analysis. Patient | age | gender | ECOG PS | Tissue source | Stage | HER2 IHC | Cycle | Imaging assess | pathologic assess Case-27 | 74 | male | 2 | TURBT | T3N0M0 | 2 | 4 | CR | N Case-28 | 64 | male | 1 | BP | T4aN0M0 | 2 | 4 | PR | N Case-29 | 68 | male | 2 | TURBT | T3N0M0 | 2 | 4 | CR | ypT0N0M0 Case-30 | 59 | female | 1 | TURBT | T3N0M0 | 2 | 4 | CR | ypT0N0M0 | Overall pCR | 3-cycle | 4-cycle PP | 15/24 | 4/10 | 11/14 ITT | 15/30 | 4/13 | 11/17 Pie charts: Overall: pCR 62.5, no pCR 37.5; 3-cycle: pCR 40, no pCR 60; 4-cycle: pCR 78.6, no pCR 21.4 A Phase II Multicenter Study of Neoadjuvant Disitamab Vedotin Combined with Gemcitabine in MIBC Patients with HER2 2-3+: Interim Prognostic Outcomes MADRID 2026 ESMO congress [Slide 3] Survival analysis of Final 30 patients Panel 1: Event-free survival rate (%) vs Time from neoadjuvant therapy (months); labels 92.9%, 85.0%, 67.2% Panel 2: Overall survival rate (%) vs Time from neoadjuvant therapy (months); labels 96.7%, 92.8%, 86.6% Panel 3: Event-free survival rate (%) by RC+PLND vs Non-RC; labels RC+PLND 100%, 90.6%, 78.1%; Non-RC 62.5%, 62.5%, 31.2%; P=0.026 ➢ 30 patients; median follow-up 31.6 months (cutoff Sep 1, 2026). 7 progressed/died; 3 died (2 without RC). ➢ 12/24/36-mo EFS: 92.9% / 85% / 67.2%; OS: 96.7% / 92.8% / 86.6%. ➢ Median EFS: not reached with RC vs 36 months without RC (statistically significant). A Phase II Multicenter Study of Neoadjuvant Disitamab Vedotin Combined with Gemcitabine in MIBC Patients with HER2 2-3+: Interim Prognostic Outcomes Survival table Endpoint: EFS and OS (single-arm, N=30) All patients (disitamab vedotin + gemcitabine) | 30 RC+PLND (EFS) | not reached Non-RC (EFS) | 36 months P: P=0.026 (EFS RC+PLND vs Non-RC) Landmark rates: 12 mo: All (EFS) 92.9% | All (OS) 96.7% 24 mo: All (EFS) 85% | All (OS) 92.8% 36 mo: All (EFS) 67.2% | All (OS) 86.6% [Slide 4] A phase III RCT for RC48+Gemcitabine VS. RC48+Toripalimab (NCT07693959)—Recruiting in China GUARD-02 High suspicion of MIBC in outpatient setting. → Imaging, perform biopsy or TURBT. → informed → Open-label; Centralized randomization. samples: tumor tissue, blood. ← Whole Exome Sequencing; ctDNA examination Neoadjuvant phase RC48: 2 mg/kg, IV, Day 1; Gemcitabine: 1000 mg/m², IV, Day 2. 14-day cycle for 4-6 cycles. RC48: 2 mg/kg, IV, Day 1; Toripalimab: 3 mg/kg, IV, Day 2. 14-day cycle for 4-6 cycles. ctDNA examination; samples: blood. Imaging assessment Pathological response assessment → RC+PLND*; Refuse RC → TMT N<20 Record tumor recurrence, metastasis, and death events. → pelvic and chest imaging q6 m.; cystoscopy q3 m; pelvic and chest imaging q6 m. *note:Adjuvant therapy only for poor pathological downstaging. Inclusion Criteria: 1. T2-T4aNxM0; 2. HER2 1-3+ by immunohistochemistry; 3. Histopathologically confirmed urothelial carcinoma as the predominant component; 4. Medically eligible for and willing to undergo RC with PLND; 5. Intolerance to or refusal of cisplatin-based neoadjuvant chemotherapy. Exclusion Criteria: 1.Prior systemic chemotherapy, immune checkpoint inhibitors, or antibody-drug conjugate therapy within the past 6 months; 2.Known hypersensitivity to the investigational drugs or their components; 3.Currently lactating or pregnant; 4.Severe and uncontrolled concurrent medical conditions. I. Stratification Factors: TNM stage, HER2. II. Randomization: Central system randomization. III. Independent Data Monitoring Committee. Primary Endpoint: Pathological complete response ([illegible red text]). Secondary Endpoints: TRAEs ≥ Grade 3 in the neoadjuvant phase ([illegible red text]), event-free survival, overall survival, etc. A Phase II Multicenter Study of Neoadjuvant Disitamab Vedotin Combined with Gemcitabine in MIBC Patients with HER2 2-3+: Interim Prognostic Outcomes

[Slide 1] Enfortumab vedotin with or without pembrolizumab as first-line treatment for cisplatin-ineligible locally advanced or metastatic urothelial carcinoma: Long-term follow-up analysis from EV-103 cohorts Dose Escalation, A and K Jose Pablo Maroto Rey,1 Peter H. O'Donnell,2 Terence W. Friedlander,3 Matthew I. Milowsky,4 Daniel Petrylak,5 Christopher J. Hoimes,6 Thomas W. Flaig,7 Shilpa Gupta,8 Nataliya Mar,9 Helen H. Moon,10 Rana R. McKay,11 Sandy Srinivas,12 Saby George,13 Mehmet Asim Bilen,14 Delphine Borchiellini,15 Abhishek Bavle,16 Jasmine Lichfield,17 Yalin Zhu,18 Changting Meng,18 Jonathan E. Rosenberg19 1Hospital de la Santa Creu i Sant Pau, Barcelona, Spain; 2University of Chicago, Chicago, IL, USA; 3UCSF Medical Center, San Francisco, CA, USA; 4University of North Carolina Lineberger Comprehensive Cancer Center, Chapel Hill, NC, USA; 5Yale Cancer Center, New Haven, CT, USA; 6Duke Cancer Institute, Durham, NC, USA; 7Division of Medical Oncology, Colorado University School of Medicine, Aurora, CO, USA; 8Cleveland Clinic Taussig Cancer Institute, Cleveland, OH, USA; 9University of California Irvine, Irvine, CA, USA; 10Kaiser Permanente Southern California, Riverside, CA, USA; 11UC San Diego, San Diego, CA, USA; 12Stanford University Medical Center, Stanford, CA, USA; 13Roswell Park Comprehensive Cancer Center, Buffalo, NY, USA; 14Winship Cancer Institute of Emory University, Atlanta, GA, USA; 15Centre Antoine Lacassagne, Université Côte d'Azur, Nice, France; 16Merck & Co., Inc., Rahway, NJ, USA; 17Astellas Pharma Europe Ltd, UK; 18Pfizer Inc., Bothell, WA, USA; 19Memorial Sloan Kettering Cancer Center, New York, NY, USA [Slide 3] Median OS was 27.7 months in the pooled DE/A/K EV+P cohort Pooled DE/A/K EV+P cohort Events/n 83/121 | Median OS (95% CI), mo 27.7 (20.9, 39.6) Landmark labels: 24 mo 52.9% (43.6, 61.3); 48 mo 38.8% (30.2, 47.4); 60 mo 33.0% (24.6, 41.6); 84 mo 28.2% (19.6, 37.4) No. at risk | 0 | 5 | 10 | 15 | 20 | 25 | 30 | 35 | 40 | 45 | 50 | 55 | 60 | 65 | 70 | 75 | 80 | 85 | 90 | 95 | 100 | 121 | 112 | 103 | 87 | 75 | 62 | 59 | 55 | 50 | 47 | 45 | 38 | 26 | 18 | 16 | 14 | 14 | 12 | 3 | 1 | 0 EV monotherapy cohort Events/n 54/73 | Median OS (95% CI), mo 22.8 (16.3, 28.4) Landmark labels: 24 mo 46.4% (34.6, 57.3); 48 mo 25.5% (16.0, 36.2); 60 mo 23.9% (14.6, 34.5) No. at risk: 73 70 61 57 51 48 44 40 33 28 23 22 21 19 18 18 16 12 9 6 4 3 1 0 [small print; unverified] • 56.2% of patients in the pooled EV+P cohort and 72.6% in the EV monotherapy cohort had subsequent systemic therapya – PD-(L)1 inhibitor–containing therapy was administered in 23.1% of all patients in the pooled EV+P cohort and 42.5% of all patients in the EV monotherapy cohort Data cutoff: February 20, 2026. aIncludes systemic therapies received after discontinuation of study treatment that were not for progressive disease, maintenance use, or secondary malignancy. DE, dose escalation; EV, enfortumab vedotin; OS, overall survival; P, pembrolizumab. Presented by: Dr. Jose Pablo Maroto Rey Survival table Endpoint: Overall survival Arm | N | Median | Median CI Pooled DE/A/K EV+P | 121 | 27.7 mo | 20.9, 39.6 EV monotherapy | 73 | 22.8 mo | 16.3, 28.4 Landmark rates: 24 months [?]: Pooled DE/A/K EV+P 52.9% (CI 43.6, 61.3) | EV monotherapy 46.4% (CI 34.6, 57.3) 48 months [?]: Pooled DE/A/K EV+P 38.8% (CI 30.2, 47.4) | EV monotherapy 25.5% (CI 16.0, 36.2) 60 months: Pooled DE/A/K EV+P 33.0% (CI 24.6, 41.6) | EV monotherapy 23.9% (CI 14.6, 34.5) 84 months [?]: Pooled DE/A/K EV+P 28.2% (CI 19.6, 37.4)

[Slide 1] PFS High volume‡ Events, n (%): 37 (38.1) — Time to event, median (95% CI): NR (22.6–NR) [Darolutamide + ADT] Events, n (%): 76 (83.5) — Time to event, median (95% CI): 9.1 months (6.4–11.6) [ADT (CHAARTED)] HR 0.28, 95% CI 0.19–0.41 Legend: Darolutamide + ADT; ADT (CHAARTED) Y-axis: PFS probability; X-axis: Months Patients at risk, n (months 0,3,6,9,12,15,18,21,24,27,30,33) Darolutamide + ADT: 97, 94, 89, 80, 70, 61, 54, 51, 48, 30, 19, 10 ADT CHAARTED: 91, 79, 57, 45, 34, 30, 25, 19, 12, 11, 4, 3 Low volume‡ Events, n (%): 9 (14.3) — Time to event, median (95% CI): NR (NR–NR) [Darolutamide + ADT] Events, n (%): 35 (50.7) — Time to event, median (95% CI): 22.7 months (15.9–NR) [ADT (CHAARTED)] HR 0.21, 95% CI 0.10–0.44 Legend: Darolutamide + ADT; ADT (CHAARTED) Patients at risk, n (months 0,3,6,9,12,15,18,21,24,27,30,33) Darolutamide + ADT: 63, 62, 60, 60, 59, 58, 56, 53, 46, 24, 14, 6 ADT CHAARTED: 69, 66, 60, 56, 49, 43, 41, 32, 29, 19, 10, 5 Survival table Endpoint: PFS Arm | N | Median | Median CI High volume: Darolutamide + ADT | 97 | NR | 22.6–NR High volume: ADT (CHAARTED) | 91 | 9.1 months | 6.4–11.6 Low volume: Darolutamide + ADT | 63 | NR | NR–NR Low volume: ADT (CHAARTED) | 69 | 22.7 months | 15.9–NR HR: High volume 0.28; Low volume 0.21 HR CI: High volume 0.19–0.41; Low volume 0.10–0.44 [Slide 2] OS High volume‡ Events, n (%): 19 (19.6) — Time to event, median (95% CI): NR (34.4 months–NR) [Darolutamide + ADT] Events, n (%): 47 (51.7) — Time to event, median (95% CI): 40.7 months (29.8–NR) [ADT (CHAARTED)] HR 0.51, 95% CI 0.30–0.88 Legend: Darolutamide + ADT; ADT (CHAARTED) Y-axis: OS probability; X-axis: Months Patients at risk, n (months 0,3,6,9,12,15,18,21,24,27,30,33,36,39,42) Darolutamide + ADT: 97, 95, 92, 90, 84, 81, 75, 70, 66, 49, 30, 15, 8, 2, 1 ADT CHAARTED: 91, 89, 87, 85, 83, 75, 71, 65, 63, 59, 52, 45, 43, 41, 37 Low volume‡ Events, n (%): 3 (4.8) — Time to event, median (95% CI): NR (NR–NR) [Darolutamide + ADT] Events, n (%): 15 (21.7) — Time to event, median (95% CI): NR (NR–NR) [ADT (CHAARTED)] HR 0.31, 95% CI 0.09–1.10 Legend: Darolutamide + ADT; ADT (CHAARTED) Patients at risk, n (months 0,3,6,9,12,15,18,21,24,27,30,33,36,39,42) Darolutamide + ADT: 63, 62, 60, 60, 59, 59, 57, 56, 54, 36, 22, 12, 3, 0, 0 ADT CHAARTED: 69, 69, 68, 68, 67, 65, 63, 60, 59, 58, 55, 52, 50, 48, 46 Survival table Endpoint: OS Arm | N | Median | Median CI High volume: Darolutamide + ADT | 97 | NR | 34.4 months–NR High volume: ADT (CHAARTED) | 91 | 40.7 months | 29.8–NR Low volume: Darolutamide + ADT | 63 | NR | NR–NR Low volume: ADT (CHAARTED) | 69 | NR | NR–NR HR: High volume 0.51; Low volume 0.31 HR CI: High volume 0.30–0.88; Low volume 0.09–1.10 [Slide 3] Time to mCRPC High volume‡ Events, n (%): 36 (37.1) — Time to event, median (95% CI): NR (24.7 months–NR) [Darolutamide + ADT] Events, n (%): 75 (82.4) — Time to event, median (95% CI): 9.1 months (6.4–11.7) [ADT (CHAARTED)] HR 0.27, 95% CI 0.18–0.40 Legend: Darolutamide + ADT; ADT (CHAARTED) Y-axis: Event-free probability; X-axis: Months Patients at risk, n (months 0,3,6,9,12,15,18,21,24,27,30,33) Darolutamide + ADT: 97, 94, 89, 79, 70, 61, 54, 51, 48, 30, 19, 10 ADT CHAARTED: 91, 78, 57, 45, 34, 30, 25, 19, 12, 11, 4, 3 Low volume‡ Events, n (%): 6 (9.5) — Time to event, median (95% CI): NR (NR–NR) [Darolutamide + ADT] Events, n (%): 35 (50.7) — Time to event, median (95% CI): 22.7 months (15.9–NR) [ADT (CHAARTED)] HR 0.14, 95% CI 0.06–0.33 Legend: Darolutamide + ADT; ADT (CHAARTED) Patients at risk, n (months 0,3,6,9,12,15,18,21,24,27,30,33) Darolutamide + ADT: 63, 61, 60, 59, 59, 58, 56, 52, 46, 24, 14, 6 ADT CHAARTED: 69, 66, 60, 56, 49, 43, 41, 32, 29, 19, 10, 5 Survival table Endpoint: Time to mCRPC Arm | N | Median | Median CI High volume: Darolutamide + ADT | 97 | NR | 24.7 months–NR High volume: ADT (CHAARTED) | 91 | 9.1 months | 6.4–11.7 Low volume: Darolutamide + ADT | 63 | NR | NR–NR Low volume: ADT (CHAARTED) | 69 | 22.7 months | 15.9–NR HR: High volume 0.27; Low volume 0.14 HR CI: High volume 0.18–0.40; Low volume 0.06–0.33 [Slide 4] rPFS High volume‡ Events, n (%): 17 (17.5) — Time to event, median (95% CI): NR (NR–NR) [Darolutamide + ADT] Events, n (%): 57 (62.6) — Time to event, median (95% CI): 18.2 months (12.1–23.2) [ADT (CHAARTED)] HR 0.24, 95% CI 0.14–0.41 Legend: Darolutamide + ADT; ADT (CHAARTED) Y-axis: rPFS probability; X-axis: Months Patients at risk, n (months 0,3,6,9,12,15,18,21,24,27,30,33,36) Darolutamide + ADT: 97, 94, 88, 84, 75, 64, 59, 54, 36, 20, 14, 6, 1 ADT CHAARTED: 91, 83, 71, 58, 50, 42, 39, 32, 28, 27, 24, 22, 19 Low volume‡ Events, n (%): 8 (12.7) — Time to event, median (95% CI): NR (NR–NR) [Darolutamide + ADT] Events, n (%): 22 (31.9) — Time to event, median (95% CI): NR (NR–NR) [ADT (CHAARTED)] HR 0.44, 95% CI 0.19–0.99 Legend: Darolutamide + ADT; ADT (CHAARTED) Patients at risk, n (months 0,3,6,9,12,15,18,21,24,27,30,33,36) Darolutamide + ADT: 63, 61, 60, 60, 59, 58, 56, 44, 32, 18, 9, 3, 0 ADT CHAARTED: 69, 68, 65, 63, 60, 56, 53, 49, 47, 44, 42, 40, 38 Survival table Endpoint: rPFS Arm | N | Median | Median CI High volume: Darolutamide + ADT | 97 | NR | NR–NR High volume: ADT (CHAARTED) | 91 | 18.2 months | 12.1–23.2 Low volume: Darolutamide + ADT | 63 | NR | NR–NR Low volume: ADT (CHAARTED) | 69 | NR | NR–NR HR: High volume 0.24; Low volume 0.44 HR CI: High volume 0.14–0.41; Low volume 0.19–0.99

2274P Prevalence of pathogenic germline variants (PGVs) across age in metastatic prostate cancer (mPC) S. Brown-Araujo¹, D. Tello¹, A. Jambrina², M. Alonso¹, M. Balongo², I. Gonzalez-Ginel¹,⁴, T. Pineda¹, J. Esteban-Villarrubia²,³, M. Ovejero¹, P. Iglesias¹, P. Diaz², N. Lorente², J.M. Sanchez²,⁵, B. Hidalgo⁵, J. Funes², P. Lopez-Casas¹, A. Rodriguez-Antolin⁴, D. Castellano³, D. Olmos¹,³, E. Castro²,³ ¹Genomics and Therapeutics of prostate Cancer, Instituto de Investigación Hospital 12 de Octubre. ²Translational Cancer Genetics group, Instituto de Investigación Hospital 12 de Octubre. ³Oncology department, Hospital Universitario 12 de Octubre. ⁴Urology Department, Hospital Universitario 12 de Octubre. ⁵Hereditary Cancer Laboratory, Clinical Biochemistry Department. Hospital Universitario 12 de Octubre 1. BACKGROUND • Germline testing is recommended for all mPC patients irrespective of age, family history or tumour characteristics. • In other malignancies, early-onset and de novo disease are enriched for PGVs, and younger patients are often priorised for testing. • Whether early-onset prostate cancer (<55y) truly shows increased PGV frequency has not been established. 2. OBJECTIVES • Evaluate the prevalence of germline PGVs in mPC according to age. • Assess whether age, de novo or recurring disease are associated with PGV frequency. 3. STUDY COHORT 1,036 mPC patients Age groups (years): <55 n=99 (9.6%); 55-65 n=326 (31.5%); 65-75 n=402 (38.8%); >75 n=209 (20.2%) de novo mPC (56.8%); Recurrent mPC (43.1%) Targeted sequencing of 11 genes recommended by ASCO for germline testing in PC: BRCA1, BRCA2, EPCAM, HOXB13, MLH1, MSH2, MSH6, PMS2, ATM, CHEK2 AND PALB2 4. METHODS Germline DNA extraction, library preparation & hybridisation capture → Targeted sequencing (Illumina sequencing; ASCO recommended gene panel) → Variant calling & classification (ACMG criteria) → PGV definition: Pathogenic / Likely pathogenic = PGV carrier 5. RESULTS 5.1 Overall PGV Prevalence: 5.31% (n=55) Gene | % (n) BRCA2 | 2.70% (28) CHEK2 | 0.87% (9) ATM | 0.77% (8) PALB2 | 0.29% (3) BRCA1 | 0.29 (3) MSH2 | 0.19% (2) PMS2 | ≤0.1% (1) MSH6 | ≤0.1% (1) 5.2 PGV prevalence by presentation: p > 0.5 (Fisher's test); de novo 5.43%; Recurrent 5.15% 5.3 PGV prevalence by age group: <55 5.05% (n=5); 55-65 7.67% (n=25); 65-75 5.22% (n=21); >75 1.91% (n=4); Overall: 5.31% • No significant trend across age groups (Cochran-Armitage p=0.24) • ≤55 vs >55: 5.05% vs 5.5% (Fisher's test p>0.5) • ≤75 vs >75: lower prevalence in >75 (Fisher's test p = 0.04) 6. KEY FINDINGS Overall PGV prevalence 5.31% BRCA2 most frequent 2.70% Similar de novo vs recurrent 5.48% vs 5.15% (p > 0.5) No significant age trend p = 0.24 (Cochran-Armitage) 7. CONCLUSIONS →Age at diagnosis was not significantly as[sociated with PGV] prevalence in mPC. [partly obscured] →No enrichment in PGVs was observed [in de] novo cases. →The reduced prevalence in patients [>75 did not translate] into a significant overall trend. [partly obscured] →Overall PGV prevalence aligns with [obscured] Age and timing of metastasis [are not] predictors of PGV carrier status [and] support germline testing acros[s ...] [obscured by presenter] Presenting author: sbrown.imas12@h12o.es Funding: This work was funded by Fundación CRIS Contra el Cáncer Disclosures: Authors declare no conflicts of interest Presented at the European Society for Medical [Oncology] ... 22-27 October

[Slide 1] Figure 1. Study design Patient population (N≈390) - MIBC (cT2-T4aN0M0) - No prior systemic therapy or radiotherapy - ECOG PS 0-2 - Eligible for cCRT - Fit for systemic therapy and ineligible for or elect not to undergo cystectomy Max TURBT R 1:1 Arm 1: EV 1.25 mg/kg IV Q3W on Days 1 and 8; Pembrolizumab 200 mg IV Q3W on Day 1 (9 cycles) -> Pembrolizumab 200 mg IV Q3W on Day 1^a (8 cycles) Arm 2: cCRT^b 4 or 6.5 weeks -> Long-term follow-up - Years 1-2, Q12W; Q24W thereafter: tumor response assessment via imaging, cystoscopy ± biopsy, urine cytology, and repeated TURBT as clinically indicated - Survival follow-up Randomization stratification factors - Tumor stage (cT2 vs cT3-T4a) - Cystectomy eligibility (ineligible vs eligible but elect not to undergo) - Geographic region (North America vs Europe vs rest of world) [Slide 2] [World map showing trial site countries] Legend: Active (dark blue) - United States, Canada, South Korea Planned (light blue) - e.g. United Kingdom, France, Spain, Germany, Italy, Türkiye, China, Japan, Taiwan, Brazil, Argentina Active Planned

[Slide 1] ACCOMPLISH Study design Open-label, multicenter, single-arm, investigator-initiated phase II trial Key Eligibility Criteria - Age ≥ 19 years - Locally advanced or metastatic ACC - Previously received both cisplatin- and mitotane-based chemotherapy* - measurable disease according to RECIST v1.1 - ECOG PS 0-1 - No previous immune checkpoint inhibitor and/or VEGF inhibitor *Patients were required to discontinue mitotane before study treatment N=30 Study Treatment Lenvatinib 20 mg once daily orally Plus Pembrolizumab 200 mg iv. Every 3 weeks (maximum 35 cycles) Until disease progression, unacceptable adverse events, or patient withdrawal Endpoints Primary - ORR according to RECIST v1.1 (by investigator) Secondary - PFS (by investigator) - OS - Safety profile NCT05036434. Data cut-off: 6 Mar, 2026 Hypothesis and sample size determination - Simon's optimal two-stage design - Null ORR 10%(P0) vs. target ORR 30% (P1), α=0.05 and 80% power - Stage 1: 10 patients; ≥2 responses required to proceed to stage 2 (19 additional patients) - Planned enrollment: 30 (29 evaluable, allowing of 5% drop-out) - Primary endpoint met if ≥6 of 29 evaluable patients achieve an object response. [Slide 2] Conclusion - ACCOMPLISH met its primary endpoint, with a confirmed ORR of 27% (8 of 30 patients; 95% CI 12-46%). Lenvatinib plus pembrolizumab demonstrated meaningful antitumor activity in advanced ACC after mitotane and platinum-based chemotherapy. - The safety profile was consistent with previous reports of this combination, and no new safety signals were identified. - These results support lenvatinib plus pembrolizumab as a potential treatment option in this setting. [Slide 3] Secondary Endpoints Adverse Events Overview Adverse events summary AE | N (%) Any AE | 30 (100.0) Grade ≥3 AE | 23 (76.7) Serious AE | 11 (36.7) AEs leading to death (grade 5 AE) | 1 (3.3)* AE leading to interruption of drugs Any | 22 (73.3) Both | 8 (26.7) Lenvatinib | 22 (73.3) Pembrolizumab | 9 (30.0) AE leading to discontinuation of drugs Any | 0 (0.0) AEs leading to dose reduction for lenvatinib | 19 (63.3) *Grade 5 pneumonia (not related to study drugs by investigator's assessment) Adverse events occurring in ≥15% of patients) Adverse Event | Any grade | Grade 3-4 Mucositis oral | 17 (56.7) | 1 (3.3) Proteinuria | 17 (56.7) | 5 (16.7) Hypertension | 17 (56.7) | 8 (26.7) Hypothyroidism | 12 (40.0) | 0 (0.0) Fatigue | 11 (36.7) | 0 (0.0) Diarrhea | 9 (30.0) | 1 (3.3) Anorexia | 9 (30.0) | 3 (10.0) Platelet count decreased | 8 (26.7) | 3 (10.0) Fever | 7 (23.3) | 0 (0.0) Nausea | 7 (23.3) | 0 (0.0) Constipation | 6 (20.0) | 0 (0.0) Aspartate aminotransferase increased | 6 (20.0) | 2 (6.7) Abdominal pain | 5 (16.7) | 1 (3.3) [Slide 4] Results: Primary Endpoint ORR (Investigator Assessed) Best Overall Response | n (%) Complete Response (CR) | 0 (0) Partial Response (PR) | 8 (27)* Stable Disease (SD) | 20 (67) - of which unconfirmed PR | 1 (3) Progressive Disease (PD) | 2 (7) Objective Response Rate, % (n/N; 95% CI) | 27% (8/30; 12-46%) Disease Control Rate, % (n/N) | 93% (28/30) *Note: The submitted abstract reported 9 responses based on unconfirmed best response. The protocol-specified confirmation requirement (RECIST 1.1) was not applied at the time of abstract preparation; with confirmation applied, 8 responses were confirmed. The updated results are presented with the agreement of the ESMO Scientific Programmes Office. **Computed using the Clopper-Pearson method Abbreviations: 95% CI, 95% confidence interval Waterfall plot [Best Change in Target Lesion from Baseline (%) by Patients; Best Overall Response PR / SD / PD; * Unconfirmed PR]

[Slide 1] University of Sheffield | School of Medicine and Population Health Expression of molecular targets appears to vary in molecular subtypes of UC4 Luminal papillary | Luminal non-specified | Luminal unstable | Stroma-rich | Basal/squamous | Neuroendocrine-like Incidence | ~24% | ~8% | ~15% | ~15% | ~35% | ~4% Differentiation | Urothelial–luminal | Basal NECTIN4 / HER2 / FGFR3 (decreasing luminal→basal); PD-L1 EGFR / Immune infiltration (increasing) Discussant: Prof. Syed A. Hussain, University of Sheffield 1. Kamoun A, et al. Cell. 2020;181:540-556. 2. Lambrechts D, et al. Eur Urol. 2021;79:420-433. 3. Choi W, et al. J Clin Oncol. 2025 (ES-Sep Abstract 504). 4. Balar AV, et al. Lancet Oncol. 2021;22:512-523. 5. Coenen W, et al. J Clin Oncol. 2022;40:522-533. Expression of ADC targets appear to differ in UC molecular subtypes1 HER2 (****) | EGFR (***) | PD-L1 (****) | NECTIN4 (***) — mRNA expression (log2) by LumP, LumNS, LumU, Stroma, Basal, NE HER2 and NECTIN4 expression were highest in luminal subtypes of UC UC, urothelial carcinoma; LumP, luminal papillary; LumNS, luminal non-specified; LumU, luminal unstable; NE, neuroendocrine-like. Adapted from: Choi W, et al. J Clin Oncol. 2025 (ES-Sep Abstract 504).

PAPILLARY RCC: TRIAL LANDSCAPE TRIAL | REGIMEN / COHORT | PAPILLARY RCC (n) | ORR IN PAPILLARY RCC SAVOIR | Savolitinib | 33 | 27% (MET-Driven) PAPMET | Cabozantinib | 44 | 23% CALYPSO | Savolitinib + durvalumab | 41 | 29% (53% in MET-driven) KEYNOTE-427 (cohort B) | Pembrolizumab | 118 | 28.8% CheckMate 374 | Nivolumab | 24 | 8.3% KEYNOTE-B61 | Pembrolizumab + lenvatinib | 93 | 54% SUNNIFORECAST | Nivolumab + ipilimumab | 72 | 29.2% Cabozantinib + nivolumab | Cabozantinib + nivolumab | 32 | 47% COSMIC-021 | Cabozantinib + atezolizumab | 15 | 47% Bevacizumab + erlotinib | HLRCC-associated / sporadic papillary RCC | 83 (43 / 40) | 72% / 35% Lenvatinib + everolimus | Lenvatinib + everolimus | 20 | 15% CaNI | Cabozantinib + nivolumab + ipilimumab | 19* | 35%* Cristina Suárez MD PhD MADRID 2026 ESMO congress

# 3309eP Molecular Co-Occurrence and Mutual Exclusivity Patterns in Prostate Cancer: Insights from a Prospective Real-World Genomic Database WAYFIND-R. Authors: Martin Angel1*; Lucas E. Coradini1; Federico Losco1; Florencia Tsou1; Juan Pablo Sade 1: Medical Oncologist. Instituto Alexander Fleming, Buenos Aires, Argentina mangel@alexanderfleming.org *Principal investigator of WAYFIND-R trial. INTRODUCTION: Molecular profiling has become an integral component of precision oncology in prostate cancer, particularly in advanced disease, where the identification of actionable alterations may inform the use of targeted therapies and refine prognostic assessment. Genomic alterations involving androgen receptor signaling, tumor suppressor genes, homologous recombination repair, and mismatch repair define biologically relevant subgroups; however, these events do not occur in isolation. Their patterns of co-occurrence may reflect cooperating pathways that contribute to tumor progression, whereas mutual exclusivity may identify distinct molecular lineages, redundant oncogenic mechanisms, or biologically incompatible genomic states. Characterizing these relationships in real-world populations may therefore improve the interpretation of complex molecular reports. The aim of this analysis was to identify patterns of co-occurrence and mutual exclusivity among predefined molecular alteration groups in patients with prostate cancer included in the prospective WAYFIND-R real-world genomic database. Methods: Data were prospectively collected from the WAYFIND-R database. Patients were stratified into 7 groups according to the predominant molecular profile: G0_SPOP, G1_AR, G2_TumorSuppressorGenes, G3_HRR, G4_MMR, G5_CHIP, and G6_Other. Results: 97 patients profiled ; 533 of 904 screened genes (59%) harbored a detectable mutation across the cohort; 41% showed no mutation detected in the tested panel. G2_Tumor Suppressors was the most frequent clinically relevant subgroup (48.4%), followed by G3_HRR (35.0%). CONCLUSION TumorSuppressorGenes appear to act as a common molecular denominator, suggesting that tumor suppressor gene alterations may represent a common genomic background upon which additional molecular events accumulate. Robust patterns of mutual exclusivity were not observed across most molecular subgroups, except SPOP ↔ CHIP relationship. The presence of CHIP mutations in nearly 20% of patients represents a clinically relevant finding that should be considered to avoid false-positive interpretation of somatic tumor variants. We want to thanks all the WAYFIND-R´s patients and their families CO-OCCURRENCE MATRIX [colored heatmap; cell values illegible] OVERALL MUTATION DETECTION RATE [bar chart, Mutated patients (n)]: TP53 52; AR 24; PTEN 16; BRCA2 15; DNMT3A 14; ATM 11; TET2 10; ASXL1 9; CHEK2 8; CDK12 8; RB1 8; BRCA1 7 [References footnote illegible]

[Slide 1] Determinants of ADC Activity Pharmacokinetics | Tumor penetration | Target recognition | Intracellular trafficking | Payload | TME modulation | Resistance Antibody Linker Payload Drug-to-Antibody-Ratio (DAR) Bystander effect Mosele et al. Nature Medicine (2026); Lin et al. npj Precis Onc (2025); Tsuchikama et al. Nat Rev Clin Oncol (2024) Ana Aparicio, MD [Speaker panel: Ana Aparicio — New targets (ADCs, new biologies)] [Slide 2] Target Features Cell surface density | Intratumoral density | Intertumoral heterogeneity | Normal tissue expression / Antigen shedding Mosele et al. Nature Medicine (2026); Lin et al. npj Precis Onc (2025); Tsuchikama et al. Nat Rev Clin Oncol (2024) Ana Aparicio, MD [Slide 4] Cell State Dependence of Target Expression AR Activity/Sensitivity → Lineage Promiscuity KLK2 PSMA STEAP1 STEAP2 B7H3 TROP2 HER3 CD46 CEACAM5 PSCA NE restricted: DLL3 SEZ6 Sayar et al. Mol Cancer Res (2026); Ajkunic et al. npj Precis Onc (2024); Shen et al. Clin Cancer Res (2025); Zaratt et al. J Clin Invest (2025); Makovec et al. Clin Cancer Res (2026); Busanova et al. Nat Rev Urol (2025); Ashraf et al. Prostate Cancer and Prostatic Diseases (2026) Ana Aparicio, MD

4 Results · Radiomic signatures 0.78 external AUC for pCR (95% CI 0.68-0.87) 0.77 external AUC for pMR (95% CI 0.67-0.86) 0.75-0.76 AUC in the RC-only subset (N=59) C AUC by cohort, features and weights unchanged from PURE-01 pCR (ypT0): PURE-01 (test) 0,86; External, full (N=97) 0,78; External, RC only (N=59) 0,76 pMR (ypT<2): PURE-01 (test) 0,96; External, full (N=97) 0,77; External, RC only (N=59) 0,75 D ROC curves and calibration in the external cohort Full cohort (N=97) RC-only subset (N=59) chance / perfect calibration pCR (ypT0) · ROC (left) and calibration by deciles (right) AUC (95% CI) Full: 0.78 (0.68, 0.87) Cystectomy: 0.76 (0.62, 0.89) pMR (ypT<2) · ROC (left) and calibration by deciles (right) AUC (95% CI) Full: 0.77 (0.67, 0.86) Cystectomy: 0.75 (0.62, 0.88) Sensitivity / 1 - Specificity / Observed proportion / Predicted probability

[Photo of presenter standing beside poster; 'BEST POSTER' rosette; board sign METASTATIC PROSTATE CANCER; poster number 2240P] BioChiP: A phase II, open-label, biomarker-guided study of carboplatin efficacy in pretreated metastatic castration resistant prostate cancer TAKE HOME MESSAGES - Without biomarker selection, carboplatin showed a combined response rate of 21.5%. - Higher response rates were associated with a RAD51-defined HRD phenotype and with BRCA1/2 alterations. Primary endpoint - Response Rate: 21.5% (95% CI 12.3 - 33.5) BIOCHEMICAL RESPONSE RATE 16.9% (95% CI 8.8 - 28.3) RADIOGRAPHIC RESPONSE RATE 22.7% (95% CI 11.5 - 37.8) Key secondary endpoint - RESPONSE RATE BY RAD51 [Small text, tentative: N=65; Not evaluable 38.5% (25/65); HRP 50.8% (33/65); HRD 10.8% (7/65); RAD51 - HRD n/total 4/7, 57.1%, 95% CI 18.4-90.1; RAD51 - HRP/NE n/total 10/58, 17.2%, 95% CI 8.6-29.4] Sections visible: Background; Design; Key eligibility criteria; Procedures; Biomarker analyses; Endpoints; Statistical analysis; Conclusions; Acknowledgments; Personal information; Genomic landscape of study population; Patients characteristics; Safety; Secondary endpoints (OVERALL POPULATION / BY RAD51 STATUS: rPFS, CLINICAL PFS, OVERALL SURVIVAL); Exploratory endpoints (BY GENOMIC ALTERATION) - body text and curve values illegible

[Slide 1] Phase 2, single-arm, multicenter (8 institutions), open-label trial · N = 51 treated patients Key eligibility criteria • Histologically confirmed prostate adenocarcinoma • Metastatic castration-resistant prostate cancer • Radiographic, PSA, or clinical progression on or after abiraterone • Chemotherapy-naïve for mCRPC • ECOG performance status 0–1 • Archival or fresh tumor tissue and blood available for molecular analyses Serum testosterone (supraphysiologic -- castrate) AR blockade Testosterone cypionate (BAT) | Darolutamide | Washout Day 1 - Weeks 1–4 | Days 29–56 · Weeks 5–8 | 7 days · Week 9 One cycle = 9 weeks (63 days) Repeat cycle until progression or unacceptable toxicity Ongoing LHRH agonist/antagonist maintained throughout treatment PSA and physical examination every 4 weeks CT abdomen/pelvis and bone scan every 12 weeks Endpoints Primary Rate of radiographic progression-free survival (rPFS) at 12 months (PCWG3), compared with a historical control Secondary • Median rPFS and time to progression • PSA50 response rate • Overall survival • Quality of life • Safety and tolerability [Slide 2] Radiographic Progression-Free Survival (%) Median (95% CI) 8.3m (3.9-NE) [Non-Amplified] Median (95% CI) 2.9m (1.8-NE) [Amplified] P-value: 0.0284 HR: 0.39 AR | Events/Total | % 12m rPFS (95% CI) | HR (95% CI) Amplified | 7/8 | 0.0 (NE-NE%) | Reference Non-Amplified | 23/38 | 42.3 (26.2-57.6%) | 0.39 (0.16-0.94) Months At Risk (0, 2, 4, 6, 8, 10, 12, 14) Amplified | 8 | 6 | 2 | 2 | 2 | 0 Non-Amplified | 38 | 30 | 22 | 18 | 18 | 16 | 15 | 0 Survival table Endpoint: Radiographic progression-free survival (by AR amplification status) Arm | N | Median | Median CI AR Amplified | 8 | 2.9m | 1.8-NE AR Non-Amplified | 38 | 8.3m | 3.9-NE HR: 0.39 HR CI: 0.16-0.94 P: 0.0284 Landmark rates: 12 months: AR Amplified 0.0% (CI NE-NE%) | AR Non-Amplified 42.3% (CI 26.2-57.6%) [Slide 3] Overall Survival (%) 12m OS (95% CI) 88.9% (73.0-95.7%) [Non-amplified] 12m OS (95% CI) 50% (15.2-77.5%) [Amplified] P-value: 0.001 HR: 0.19 AR | Events/Total | Median (95% CI) | HR (95% CI) Amplified | 6/8 | 11.6 (3.7-NE) | Reference Non-Amplified | 9/38 | NR (21.6-NE) | 0.19 (0.07-0.55) Months At Risk (0, 2, 4, 6, 8, 10, 12, 14) Amplified | 8 | 8 | 7 | 6 | 6 | 5 | 3 | 2 Non-amplified | 38 | 38 | 37 | 36 | 34 | 32 | 31 | 25 Survival table Endpoint: Overall survival (by AR amplification status) Arm | N | Median | Median CI AR Amplified | 8 | 11.6 | 3.7-NE AR Non-Amplified | 38 | NR | 21.6-NE HR: 0.19 HR CI: 0.07-0.55 P: 0.001 Landmark rates: 12 months: AR Non-Amplified 88.9% (CI 73.0-95.7%) | AR Amplified 50% (CI 15.2-77.5%) [Slide 4] Genomic Profile and Individual Clinical Outcomes Genes: AR | TP53 | DNMT3A | PPM1D | TET2 | ASXL1 | TMPRSS2 | MUTYH | ATM | CDK12 | BRCA2 | CHEK2 | Other gene** | Sample Patient (1–48) Time from enrollment (months) (0, 5, 10, 15, 20, 25) Genomic alteration key Mutation (oncogenic/likely oncogenic) | AR mutation + amplification | Not evaluable AR amplification (CNV) | Not detected Swimmer plot key No PSA decline† | Death Any PSA decline† | Alive / last contact (censored) No post-baseline PSA assessment | Radiographic/clinical progression (rPFS event)

[Slide 1] Methods Study design (single-centre, single-arm phase 2) MIBC cT2–T3bN0M0* · ECOG PS 0–1 Cisplatin-ineligible or refusing · scheduled for RC N = 49 Neoadjuvant SG + Pembro × 4 cycles, Q3W SG 7.5 mg/kg D1 + D8 · Pembro 200 mg D1 Primary G-CSF prophylaxis Response assessment: CT + pelvic MRI + cystoscopy; multidisciplinary tumour board Radical cystectomy n = 16 (33%) re-TURBT n = 29 (59%) No surgery (progression) n = 4 (8%) bladder sparing** Adjuvant / maintenance Pembro 200 mg Q3W × 13 cycles Off-protocol systemic therapy *Four cT4a patients enrolled before the protocol amendment are included in the ITT analysis. **re-TURBT offered after tumour board discussion to patients refusing RC (protocol v5, June 2024). Numbers refer to the ITT population (N = 49). [Slide 2] Results: Response and survival 39% cCR (19/49) 95% CI 25–54 24 mo median follow-up (reverse KM; 20–29) 63.8% 24-mo EFS, ITT 95% CI 51.3–79.3 75.1% 24-mo OS, ITT 95% CI 63.8–88.4 Bladder sparing: in the re-TURBT group (n = 29), 5 patients underwent a delayed RC and 3 started salvage chemoradiotherapy. Three intravesical relapses occurred among the 19 patients with cCR. Survival estimates are from the updated analysis (August 2026) and differ slightly from the submitted abstract. A EFS, ITT population (N = 49) B OS, ITT population (N = 49) 24-mo EFS 63.8% (95% CI 51.3–79.3) 24-mo OS 75.1% (95% CI 63.8–88.4) C EFS by clinical response D OS by clinical response No cCR / cCR Number at risk Time from therapy (months) 24-mo EFS: cCR 83.3% (67.8–100) vs no cCR 51.7% (36.2–74.1); P = 0.019 24-mo OS: cCR 94.4% (84.4–100) vs no cCR 63.3% (48.2–83.2); P = 0.01 Discussion Survival table Endpoint: EFS and OS (24-month landmarks), ITT and by clinical response ITT (EFS) | 49 ITT (OS) | 49 cCR | 19 no cCR | 30 P: EFS by response P = 0.019; OS by response P = 0.01 24-mo EFS ITT 63.8% (95% CI 51.3–79.3) 24-mo OS ITT 75.1% (95% CI 63.8–88.4) 24-mo EFS cCR 83.3% (67.8–100) vs no cCR 51.7% (36.2–74.1) 24-mo OS cCR 94.4% (84.4–100) vs no cCR 63.3% (48.2–83.2) [Slide 3] 4750P Urothelial cancer | ESMO Congress 2026, Madrid, 23–27 October Overall survival (OS) and mature safety results of SURE-02, a phase 2 study of neoadjuvant sacituzumab govitecan (SG) plus pembrolizumab (Pembro), followed by response-adapted bladder sparing and adjuvant Pembro, in patients with muscle-invasive bladder cancer (MIBC) Debora Serafin, Giuseppe Basile, Chiara Mercinelli, Brigida A. Maiorano, Antonio Cigliola, Michela Piacentini, Gualtiero Guandalini, Giorgio Brembilla, Enrico Tomasi, Gaia Latini, Maurizio Colecchia, Francesco de Cobelli, Marco Moschini, Alberto Briganti, Jeffrey S. Ross, Francesco Montorsi, Andrea Necchi Abstract 4750P Methods Results: Response and survival Results: Mature safety 81.6% 42.9% 22.4% 65% Discussion Conclusions

[Slide 1] Final Protocol-Prespecified Analysis of the Phase 3 KEYNOTE-564 Trial of Adjuvant Pembrolizumab Versus Placebo in Clear Cell Renal Cell Carcinoma Toni K. Choueiri1; Piotr Tomczak2; Se Hoon Park3; Balaji Venugopal4; Thomas Ferguson5; Naomi B. Haas6; Yen-Hwa Chang7; Stefan N. Symeonides8; Jae Lyun Lee9; Piotr Sawrycki10; Naveed Sarwar11; Mauricio Mahave12; Howard Gurney13; Marine Gross-Goupil14; Guryjot K. Doshi15; John M. Burke16; Rui Kang17; Joseph E. Burgents17; Priyanka Chablani17; Thomas Powles18 1Dana-Farber Cancer Institute, Lank Center for Genitourinary Oncology at Dana-Farber Cancer Institute, Brigham and Women's Hospital, and Harvard Medical School, Boston, MA, USA; 2Poznan University of Medical Sciences, Poznan, Poland; 3Sungkyunkwan University Samsung Medical Center, Seoul, Republic of Korea; 4The Beatson West of Scotland Cancer Centre and University of Glasgow, Glasgow, United Kingdom; 5Fiona Stanley Hospital, Perth, WA, Australia; 6Abramson Cancer Center, Philadelphia, PA, USA; 7Taipei Veterans General Hospital, Taipei, Taiwan; 8Edinburgh Cancer Research Centre and University of Edinburgh, Edinburgh, United Kingdom; 9Asan Medical Center and University of Ulsan College of Medicine, Seoul, Republic of Korea; 10Provincial Hospital in Torun, Torun, Poland; 11Imperial College Healthcare NHS Trust, London, United Kingdom; 12Arturo López Pérez Foundation, Santiago, Chile; 13Macquarie University, Sydney, NSW, Australia; 14Centre Hospitalier Universitaire de Bordeaux - Hôpital Saint-André, Bordeaux, France; 15Texas Oncology, US Oncology Research, Woodlands, TX, USA; 16Rocky Mountain Cancer Centers, Aurora, CO, USA; 17Merck & Co., Inc., Rahway, NJ, USA; 18Barts Health NHS Trust and the Royal Free NHS Foundation Trust, Barts Cancer Institute and Queen Mary University of London, London, United Kingdom [Slide 2] OS, ITT Population | Pembro (N = 496) | Placebo (N = 498) Events, n (%) | 86 (17.3) | 118 (23.7) Median (95% CI), months | NR (NR-NR) | NR (NR-NR) HR 0.70 (95% CI, 0.53-0.92) Landmark labels on curve (Pembro): 96.3% | 93.9% | 91.2% | 87.7% | 85.9% | 82.3% Landmark labels on curve (Placebo): 93.9% | 89.5% | 85.8% | 82.2% | 79.3% | 76.6% No. at risk (0–108 months, every 6): Pembro: 496 489 485 479 472 468 460 452 445 440 428 422 417 383 227 131 49 1 0 Placebo: 498 492 484 478 464 455 442 429 423 416 404 398 390 358 204 119 45 2 0 ...OS end point was met at IA3 and was not formally statistically tested thereafter. Data cutoff date: February 4, 2026. Survival table Endpoint: Overall survival, ITT Arm | N | Median | Median CI Pembro | 496 | NR | NR-NR Placebo | 498 | NR | NR-NR HR: 0.70 HR CI: 0.53-0.92 Landmark rates: dashed line 1 (unlabelled): Pembro 96.3% | Placebo 93.9% dashed line 2 (unlabelled): Pembro 93.9% | Placebo 89.5% dashed line 3 (unlabelled): Pembro 91.2% | Placebo 85.8% dashed line 4 (unlabelled): Pembro 87.7% | Placebo 82.2% dashed line 5 (unlabelled): Pembro 85.9% | Placebo 79.3% dashed line 6 (unlabelled): Pembro 82.3% | Placebo 76.6% [Slide 3] AE Summary, As-Treated Population | IA4 (median 69.5 months follow-up)1 Pembro (n = 488) | IA4 Placebo (n = 496) | FA (median 85.9 months follow-up) Pembro (n = 488) | FA Placebo (n = 496) Duration of therapy, median (range), months | 11.1 (0.03-14.3) | 11.1 (0.03-15.4) | 11.1 (0.03-14.3) | 11.1 (0.03-15.4) All-cause AEs | 470 (96.3) | 453 (91.3) | 471 (96.5) | 454 (91.5) Grade 3-5 | 156 (32.0) | 88 (17.7) | 156 (32.0) | 88 (17.7) Led to treatment discontinuation | 103 (21.1) | 11 (2.2) | 103 (21.1) | 11 (2.2) Led to death | 2 (0.4) | 1 (0.2) | 2 (0.4) | 1 (0.2) All-cause serious AEs | 101 (20.7) | 57 (11.5) | 101 (20.7) | 57 (11.5) Led to treatment discontinuation | 49 (10.0) | 5 (1.0) | 49 (10.0) | 5 (1.0) Treatment-related AEs | 386 (79.1) | 263 (53.0) | 386 (79.1) | 263 (53.0) Grade 3-4 | 91 (18.6) | 6 (1.2) | 91 (18.6) | 6 (1.2) Led to treatment discontinuation | 89 (18.2) | 4 (0.8) | 89 (18.2) | 4 (0.8) Led to death | 0 | 0 | 0 | 0 Treatment-related serious AEs | 58 (11.9) | 1 (0.2) | 58 (11.9) | 1 (0.2) Data are n (%) unless otherwise specified. Data cutoff date: February 4, 2026. 1. Haas NB et al. J Clin Oncol. 2025;43:4514.

[Slide 1] Baseline Characteristics | Sac-TMT + pembro (N = 40) Age, median (range), y | 66.5 (42–81) Male | 26 (65) Race: Asian | 37 (93) Race: White | 3 (8) ECOG PS 0 | 3 (8) ECOG PS 1 | 32 (80) ECOG PS 2a | 5 (13) Histology: Urothelial or transitional-cell carcinoma | 38 (95) Histology: Urothelial carcinoma, mixed types | 2 (5) Location of metastasis: Lymph node only | 7 (18) Location of metastasis: Liver | 8 (20) Primary tumor location: Lower tract | 16 (40) Primary tumor location: Upper tract | 24 (60) PD-L1 CPS <10 | 32 (80) PD-L1 CPS ≥10 | 8 (20) TROP2 H-scoreb <200 | 5 (13) TROP2 H-scoreb ≥200 | 32 (80) TROP2 H-score Missingc | 3 (8) Cisplatin ineligibility criteria: ECOG PS of 2a | 5 (13) Creatinine clearanced <60 mL/min but >30 mL/min | 33 (83) Grade ≥2 audiometric hearing loss | 1 (3) Grade ≥2 peripheral neuropathy | 3 (8) CPS, combined positive score. Median study follow-up duration (time from first dose to data cutoff), 22.8 (range, 19.8–30.0) months. Data are n (%) unless otherwise specified. aBefore a protocol amendment, participants with ECOG PS of 2 were eligible for enrollment. bAs determined by Roche Tissue Diagnostics IHC assay. cParticipants who did not have TROP2 IHC data available. dCalculated or measured. Data cutoff, Jan 26, 2026. [Slide 3] Summary of Treatment-Relateda AEs | Sac-TMT 4 mg/kg + pembro (n = 26) | Sac-TMT 5 mg/kg + pembro (n = 14) | Sac-TMT + pembro (N = 40) Treatment duration, median (range), mo — Sac-TMT | 11.8 (0.03–25.1) | 10.0 (0.5–29.1) | 10.9 (0.03–29.1) Treatment duration — Pembrolizumab | 9.2 (0.03–22.8) | 9.3 (0.03–23.5) | 9.3 (0.03–23.5) Treatment-related AEs | 25 (96) | 14 (100) | 39 (98) Grade 3 or 4 | 17 (65) | 8 (57) | 25 (63) Serious | 3 (12) | 3 (21) | 6 (15) Led to any treatment discontinuation | 2 (8) | 4 (29) | 6 (15) Led to sac-TMT discontinuation | 1 (4)b | 2 (14)c | 3 (8) Led to pembrolizumab discontinuation | 1 (4)d | 2 (14)e | 3 (8) Led to death | 0 | 0 | 0 Chart: Treatment-Related AEs Occurring in ≥25% of Participants (Participants, %; Grade 1 or 2 / Grade 3 or 4) Anemia 73 | Stomatitis 73 | Alopecia 43 | Hypoalbuminemia 43 | Decreased appetite 40 | Rash 40 | Neutrophil count decreased 35 | Blood creatinine increased 33 | WBC count decreased 33 | Hyperglycemia 30 | Lymphocyte count decreased 30 | Urinary tract infection 28 | Proteinuria 25 Data in the table are n (%). aDetermined by the investigator to be related to any study treatment; in the 4 and 5 mg/kg groups respectively, 2 (8%) and 3 (21%) participants had ocular surface toxicity (all were grade ≤2), and no participants had interstitial lung disease/pneumonitis. bDue to infusion-related reaction. cDue to infusion-related reaction and platelet count decreased. dDue to glucocorticoid deficiency. eOne participant due to adrenal insufficiency and one participant due to ALT increased and AST increased. Data cutoff, Jan 26, 2026.

[Slide 1] Castration Resistant-free Survival (CRFS) Survival Probability (%) vs Time (months) CRFS12m | CRFS18m TRIDA 83.8% (95% CI, 79.5-88.1) | 71.6% (95% CI, 64.2-79.0) ARASENS ~78% | ~86 Number at risk: 0: 402 | 6: 323 | 12: 146 | 18: 33 | 24: 5 | 30: 4 | 36: 1 | 42: 1 | 48: 1 | 54: 0 | 60: 0 Survival table Endpoint: Castration resistant-free survival (CRFS) TRIDA (darolutamide + ADT + docetaxel, real-world) | 402 Landmark rates: 12 months: TRIDA 83.8% (CI 79.5-88.1) | ARASENS (reference, approximate) ~78% 18 months: TRIDA 71.6% (CI 64.2-79.0) | ARASENS (reference, approximate) ~86 [Slide 2] Overall Survival Survival Probability (%) vs Time (months) OS12m | OS18m TRIDA 90.4% (95% CI, 86.9-93.9) | 86.5% (95% CI, 81.0-92.0) ARASENS ~93% | ~88 Number at risk: 0: 402 | 6: 333 | 12: 153 | 18: 38 | 24: 8 | 30: 5 | 36: 3 | 42: 3 | 48: 3 | 54: 2 | 60: 2 Survival table Endpoint: Overall survival TRIDA (darolutamide + ADT + docetaxel, real-world) | 402 Landmark rates: 12 months: TRIDA 90.4% (CI 86.9-93.9) | ARASENS (reference, approximate) ~93% 18 months: TRIDA 86.5% (CI 81.0-92.0) | ARASENS (reference, approximate) ~88 [Slide 3] Biochemical response: PSA<0.2 48.5% | PSA 0.2-0.4 32.3% | PSA>0.4 19.2% Global response: Complete response 4.2% | PR/SD/in treatment 83.3% | Progression 12.4% Legend: PSA<0.2; PSA 0.2-0.4; PSA>0.4; Complete response; PR/SD/in treatment; Progression [Slide 4] [Horizontal stacked bar chart, Patients (%) 0-50; legend Grade 1/2, Grade 3/4, Grade 5] Adverse events listed (top to bottom): Asthenia; Alopecia; Neuropathy; Diarrhea; Neutropenia; Onychodystrophy; Artromialgy/Miositis; Anemia; Dysgeusia; Edema; Xerostomy/mucositis; Rash/dermatitis; Constipation; ALT increased; Emesis; AST increased; Anorexia; Infusion reaction; Conjunctival or corneal toxicity; Heart failure; Renal failure; Bacteremia/sepsis

Abstract FPN 357P: AlphaMissense (AM) reclassifies DNA damage repair (DDR) variants of unknown significance and identifies a subset associated with survival in metastatic urothelial carcinoma (mUC) Aditi Gupta, Karissa Whiting, Irina Ostrovnaya, Ritesh R. Kotecha, Gopa Iyer, Jonathan E. Rosenberg Memorial Sloan Kettering Cancer Center, New York, NY BACKGROUND • Alterations in DNA damage repair (DDR) genes are associated with improved survival on platinum-based chemotherapy and higher response rates to PD-1/PD-L1 blockade in metastatic urothelial carcinoma (mUC). • Most DDR alterations in mUC are missense variants of unknown significance (VUS). Clinical annotation tools such as OncoKB classify only a small fraction of them, so most patients with a DDR missense variant cannot be assigned a functional status. • In a retrospective analysis of 60 patients with mUC treated with PD-1/PD-L1 blockade, ORR was 80% with deleterious DDR mutations, 54% with DDR VUS and 19% without DDR alterations, suggesting some VUS are functional. • AlphaMissense (AM) is a deep learning model adapted from AlphaFold. It predicts pathogenicity for nearly all possible human missense variants and classifies 89% as likely benign or likely pathogenic. • We hypothesized that AM can reclassify OncoKB-Unknown DDR missense variants and identify a pathogenic subset with clinical significance in mUC METHODS • Cohort: Patients with mUC sequenced with MSK-IMPACT at or after first metastasis were identified from the MSK-MET 2021 cohort (N=590; mUC n=463, mUTUC n=127) • Gene panel: Missense variants were restricted to a previously published 34-gene DDR panel (n=511 variants) • Variant classification: Each variant was scored by OncoKB (OKB; Oncogenic, Neutral, Unknown) and by AM, which assigns a continuous pathogenicity score (0–1) categorized using thresholds calibrated against ClinVar and by AM (Pathogenic >0.564, Ambiguous 0.34–0.564, Benign <0.34) • Patient classification: Patients were assigned to a class by "worst-case wins" across all of their DDR missense variants • Cross-classification: OKB-Unknown patients were stratified by AM class to identify the pathogenic variants that OKB could not classify • Statistical analysis: Overall survival (OS) was measured from the date of sequencing. Cox regression was adjusted for age, sex and tumor mutation burden (TMB) Figure 1. Classification of missense variants (n=511) by OncoKB versus AM OncoKB: Oncogenic / Likely Oncogenic 46 (9.0%); Inconclusive 3 (0.6%); Likely Neutral 6 (1.2%); Unknown 456 (89.2%) AlphaMissense: Likely Pathogenic 139 (27.2%); Ambiguous 40 (7.8%); Likely Benign 327 (64.0%); Unscored 5 (1.0%) Fig 1. AM classifies 99.0% of missense variants, compared with 10.8% classified by OncoKB. Table 2 (Cox models vs WT): Comparison | N (events) | UV HR (95% CI) | p | MV HR (95% CI) | p AM-Pathogenic | 105 (38) | 0.49 (0.35–0.70) | <0.001 | 0.62 (0.42–0.91) | 0.016 AM-Ambiguous | 23 (12) | 0.94 (0.52–1.68) | 0.83 | 1.24 (0.67–2.31) | 0.49 AM-Benign | 113 (51) | 0.71 (0.52–0.96) | 0.027 | 0.76 (0.57–1.07) | 0.12 OKB-Oncogenic | 41 (14) | 0.51 (0.30–0.88) | 0.016 | 0.69 (0.38–1.24) | 0.22 OKB-Unknown | 198 (88) | 0.64 (0.50–0.82) | <0.001 | 0.74 (0.56–0.97) | 0.032 OKB-Unknown / AM-Pathogenic | 66 (26) | 0.51 (0.34–0.78) | 0.001 | 0.57 (0.36–0.91) | 0.019 OKB-Oncogenic / AM-Pathogenic | 38 (12) | 0.45 (0.25–0.80) | 0.007 | 0.58 (0.29–1.18) | 0.13 OKB-Unknown / AM-Ambiguous | 22 (11) | 0.87 (0.47–1.59) | 0.65 | 1.27 (0.64–2.52) | 0.49 OKB-Unknown / AM-Benign | 110 (49) | 0.68 (0.50–0.93) | 0.017 | 0.76 (0.54–1.06) | 0.11 Table 2. UV, univariable; MV, multivariable Cox model adjusted for age, sex, and TMB. HRs are relative to WT unless otherwise specified. Fig 3. OncoKB-Unknown patients reclassified as AM-Pathogenic had the strongest association with longer OS vs WT (p=0.001). Those reclassified as AM-Benign were also associated with longer OS vs WT (p=0.016). CONCLUSIONS • AM classified 99% of DDR missense variants vs 11% by OncoKB, and reclassified most OncoKB-Unknown variants in mUC. • AM-Pathogenic status was associated with improved OS vs WT (adjusted HR 0.62, p = 0.016). • AM identified a pathogenic subset that OncoKB could not classify. OncoKB-Unknown patients reclassified as AM-Pathogenic had improved OS (adjusted HR 0.57, p = 0.019), similar to OncoKB-Oncogenic patients. • AM offers a complementary classification system for identifying patients with pathogenic DDR mutations, which warrants validation in a larger mUC cohort. This study was sponsored by Memorial Sloan Kettering Cancer Center. Abstract #357P; Presented at the ESMO 2026 Madrid, Spain. [Left lower poster (Table 1, Figure 2) partially obscured by presenter]

[Slide 1] Fig 1: Cumulative incidence of SCNEPC, by treatment group Y-axis: Cumulative Incidence of SCNEPC in ENZAMET (%) Legend: Enzalutamide; Conventional NSAA X-axis: Years Number at risk (event) Years: 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 Enzalutamide: 563 (0) | 540 (2) | 498 (3) | 449 (6) | 408 (8) | 367 (10) | 336 (10) | 309 (11) | 214 (13) | 78 (15) Conventional NSAA: 562 (0) | 531 (0) | 467 (2) | 407 (3) | 347 (3) | 312 (3) | 263 (4) | 237 (4) | 159 (4) | 49 (4) Survival table Endpoint: Cumulative incidence of treatment-emergent SCNEPC Enzalutamide | 563 Conventional NSAA | 562 [Slide 2] Median OS (IQR) from SCNEPC diagnosis in ALL 19 pts 9 months (2-18), ENZA vs NSAA 9 vs 7, p=0.3 (see Fig 2) Fig 2: OS from diagnosis of SCNEPC, by treatment group Y-axis: Proportion Alive; X-axis: Months since SCNEPC diagnosis (0, 6, 12, 18, 24) Legend: NSAA; ENZA Number at risk (number of events) NSAA: 4 (0) | 2 (2) | 1 (3) | 0 (4) | 0 (4) ENZA: 15 (0) | 10 (5) | 5 (9) | 4 (10) | 0 (13) Survival table Endpoint: OS from SCNEPC diagnosis ENZA | 15 | 9 months NSAA | 4 | 7 months All patients | 19 | 9 months (IQR 2-18) P: 0.3 [Slide 3] Fig 3: Participants with 1, 2, or 3 DECIPHER Neuroendocrine (NE) classifiers DECIPHER Grid Classifier | NE-positive | Alternative category smallcell_3 | "Neuroendocrine-like" | "Adenocarcinoma" Lotan2016 | "Neuroendocrine carcinoma" | "Adenocarcinoma" Beltran2016 | "Neuroendocrine carcinoma" | "Adenocarcinoma" [Stacked bar chart; y-axis Percentage (%)] Participants with SCNEPC: 70% (95% CI 35-93) 1-3 NE Classifiers (n=7/10) Participants without SCNEPC: 43% (95% CI 39-47) 1-3 NE Classifiers (n=268/624) Legend: # NE-Positive Classifiers 0, 1, 2, 3

UCL Innovative Clinical Trials Unit Inclusion Criteria (TBC) • >18y • WHO PS 0 or 1 • No other malignancy • No previous IO • Biopsy proven resectable ccRCC AND intermediate/high risk disease based on imaging and biopsy: • cT1b/2a G4 cN0 • cT2b any grade cN0 • cT3 any grade (cT3a without VTT should be >4cm) cN0 • cT4 any grade cN0 • cN1 or M1 (fully resectable) any cT Randomisation Platform design will allow inter-arm comparisons with randomization for neoadjuvant therapy 4:1 with SoC RANDOMISE SoC → Surgery → 12/12 Pembro 3/12 Pembro → Surgery → 9/12 Pembro 3/12 Pembro → Surgery → No Pembro 3/12 NeoAdj B i.e. bispecific → Surgery → 9/12 Pembro 3/12 NeoAdj C i.e. HIF2ai + IO → Surgery → 9/12 Pembro CT CT CT CT CT Sample size 45 patients/arm to demonstrate a reduction of the 2-year event rate from 22% to 10% with a power of 80%. Primary endpoint 2-year Event Free Survival [EFS=disease progression before surgery, surgical delay or ineligibility, recurrence after surgery or death] Secondary /exploratory endpoints Recurrence free survival Overall survival Pathologic response rate Adverse events Quality of Life

[Slide 2] Why have ADCs struggled in RCC? —ADC Efficacy Chain— Pharmacokinetics | Tumor penetration | Target recognition | Intracellular trafficking | Payload | TME modulation Mosele et al. Nat Med (32)2774-2785, 2026 Breaking any link in the ADC efficacy chain leads to a loss of ADC efficacy BARRIER 1 Delivery | BARRIER 2 Targeting | BARRIER 3 Payload [Slide 3] Targeting the stroma FAP+CAFs support RCC progression and poor treatment outcomes. Why target FAP+ CAFs in RCC? Tumor EC 1; FAP+ Fibroblast; Pericyte 2 • FAP+ CAFs localized near tumor endothelial cells. • Patients-derived FAP+ CAFs promote RCC cell growth and invasion JAVELIN Renal 101: Progression-free survival Sunitinib: FAP+ Fibroblast High vs Low; p = 0.0051 Avelumab+Axitinib: FAP+ Fibroblast High vs Low; p = 0.0041 • High FAP+ CAF signature score are associated with shorter PFS both in sunitinib and Avelumab plus axitinib arm. Ma J. et al. Nat Commun 2025;16:9387 Can an ADC exploit this target? FAP-targeting ADC (e.g., OMTX705); Antibody–drug conjugate (ADC); FAP; FAP+ CAF; Promotes tumor growth and invasion; CAF depletion / modulation; Tumor cells; Potential reduction of stromal support Therapeutic hypothesis: overcome CAF-associated resistance What remains to be demonstrated • Efficacy of FAP-targeting ADCs in RCC • Improved delivery or treatment sensitivity

[Slide 2] A Consensus Framework For The Pathology Workup Of Metastatic Castration-Resistant Prostate Cancer Biopsies Goal: A simple standardized diagnostic framework Morphology Classifier (Choose One) | IHC Classifier* (Include the status of each marker) | Proliferation Classifier (Choose One) Adenocarcinoma | AR positive (based on AR and NKX3.1 IHC) | High (>60%) Poorly differentiated carcinoma | AR negative (based on AR and NKX3.1 IHC) | Medium (20-60%) Small cell carcinoma | NE positive (based on SYP and INSM1 IHC) | Low (<20%) Other | NE negative (based on SYP and INSM1 IHC) | *Biomarker positivity is defined as ≥1+ reactivity in ≥ 20% of cells with. AR-positive tumors can show either AR or NKX3.1 reactivity, meeting this threshold. For NE-positive tumors, ≥1+ reactivity in ≥ 20% for either SYP or INSM1 is required. "Prostatic adenocarcinoma, AR-positive, NE-negative, Ki-67 20%" "Small cell carcinoma, AR-negative, NE-positive, Ki-67 70%" Haffner et al., CCR 2025 [Slide 4] Summary • NEPC is a heterogenous disease - treatment should be tailored to clinical and pathologic features • Platinum chemotherapy and SCLC regimens could be considered for small cell/NEPC based on guidelines • DLL3 is very specific to NE carcinomas - careful patient selection for tarlatamab and other emerging DLL3-therapies critical (not enough to have TP53/RB1/PTEN) • Other potential targets in clinical trials: cell surface (eg, SEZ6, B7H3), epigenetic (eg, EZH2), rational combination strategies

[Slide 1] POTOMAC: Disease-Free Survival in Patients With Papillary Only and With T1 Tumours D+BCG (I+M) showed a consistent benefit across papillary only and T1 subgroups Papillary only (64%-65% of ITT population)a,b,1 D+BCG (I+M) N=217 | BCG (I+M) N=220 Number of events, n (%): 37 (17) | 65 (30) Median DFS, months (95% CI): NR (NR-NR) | NR (74.0-NR) Median follow-up in censored patients: 60.7 months Landmarks on curve (24/36/48 mo dashed lines): D+BCG 87% / 84% / 84%; BCG 81% / 76% / 74% HR 0.56 (95% CI, 0.37-0.84) T1 subset of papillary only (40%-43% of ITT population)c,d D+BCG (I+M) N=137 | BCG (I+M) N=145 Number of events, n (%): 22 (16) | 45 (31) Median DFS, months (95% CI): NR (NR-NR) | NR (69.3-NR) Median follow-up in censored patients: 60.7 months Landmarks on curve (24/36/48 mo dashed lines): D+BCG 88% / 85% / 85%; BCG 79% / 75% / 73% HR 0.48 (95% CI, 0.28-0.79) Probability of DFS / Time from randomisation (months) No. of patients at risk (papillary only): D+BCG (I+M): 217 204 196 190 184 183 178 177 176 174 168 167 164 159 154 150 146 144 135 127 102 98 69 57 38 19 9 1 0 BCG (I+M): 220 207 195 188 180 177 172 167 163 159 155 150 146 145 142 141 140 130 122 99 92 63 53 32 17 7 0 0 No. of patients at risk (T1): D+BCG (I+M): 137 130 120 117 117 114 114 112 109 108 106 103 101 99 96 90 87 70 68 52 45 30 17 8 1 0 BCG (I+M): 145 136 128 125 117 114 111 106 103 102 96 95 92 91 91 90 90 87 68 61 43 35 21 11 4 0 0 aBased on stratified analysis. bPrespecified sensitivity analysis. cBased on unstratified analysis. dPost-hoc exploratory analysis. DFS is defined as the time to the first: 1) recurrence of high-risk disease (recurrence of high-grade Ta, T1, or CIS; presentation with MIBC and/or metastatic disease or persistent CIS at 6 months); 2) death by any cause in the absence of recurrence. T1 disease: papillary disease stage T1, without CIS. Data cutoff 03 April 2025. 1. De Santis M, et al. Lancet. 2025;406:2221-2234. BCG, bacillus Calmette-Guerin; CI, confidence interval; CIS, carcinoma in situ; D, durvalumab; DFS, disease-free survival; HR, hazard ratio; I, induction; ITT, intent-to-treat; M, maintenance; MIBC, muscle-invasive bladder cancer; NR, not reached. Survival table Endpoint: DFS (papillary-only and T1 subgroups) Arm | N | Median | Median CI Papillary only - D+BCG (I+M) | 217 | NR | NR-NR Papillary only - BCG (I+M) | 220 | NR | 74.0-NR T1 - D+BCG (I+M) | 137 | NR | NR-NR T1 - BCG (I+M) | 145 | NR | 69.3-NR HR: Papillary only 0.56; T1 0.48 HR CI: Papillary only 0.37-0.84; T1 0.28-0.79 Landmark rates: 24 mo: Papillary only - D+BCG (I+M) 87% | Papillary only - BCG (I+M) 81% | T1 - D+BCG (I+M) 88% | T1 - BCG (I+M) 79% 36 mo: Papillary only - D+BCG (I+M) 84% | Papillary only - BCG (I+M) 76% | T1 - D+BCG (I+M) 85% | T1 - BCG (I+M) 75% 48 mo: Papillary only - D+BCG (I+M) 84% | Papillary only - BCG (I+M) 74% | T1 - D+BCG (I+M) 85% | T1 - BCG (I+M) 73% [Slide 3] POTOMAC: AE Summary for Patients With Papillary Only Tumours The safety profile of D+BCG (I+M) in patients with papillary only tumours was consistent with the overall safety population Events, n (%) | Papillary only1: D+BCG (I+M) N=215 | BCG (I+M) N=219 | Overall safety population2: D+BCG (I+M) N=336 | BCG (I+M) N=339 AEs of any cause | 208 (97) | 193 (88) | 325 (97) | 307 (91) Possibly related to any treatment | 189 (88) | 146 (67) | 298 (89) | 245 (72) Maximum grade 3 or 4 AEs | 73 (34) | 35 (16) | 113 (34) | 56 (17) Possibly related to any treatment | 49 (23) | 6 (3) | 71 (21) | 13 (4) Serious AEs | 64 (30) | 38 (17) | 108 (32) | 63 (19) Possibly related to any treatment | 27 (13) | 6 (3) | 45 (13) | 13 (4) AEs leading to death | 5 (2) | 4 (2) | 6 (2) | 4 (1) Possibly related to any treatment | 0 | 0 | 0 | 0 AEs leading to discontinuation of study treatment | 63 (29) | 38 (17) | 105 (31) | 68 (20) Possibly related to durvalumab | 35 (16) | 1 (0.5)a | 54 (16) | 1 (0.3)a Possibly related to BCG | 32 (15) | 27 (12) | 55 (16) | 55 (16) Any-grade imAEs | 58 (27) | 4 (2) | 91 (27) | 4 (1) Maximum grade 3 or 4 imAEs | 20 (9) | 1 (0.5) | 27 (8) | 1 (0.3) aPatient assigned to the BCG (I+M) arm due to a data-entry error. The papillary tumours only safety population includes all patients with papillary only tumours who received treatment. Causality per investigator assessment. Data cutoff 03 April 2025. 1. Shore ND, et al. J Urol. 2026;215:e3. 2. De Santis M, et al. Lancet. 2025;406:2221-2234. AE, adverse event; BCG, bacillus Calmette-Guerin; D, durvalumab; I, induction; imAE, immune-mediated adverse event; M, maintenance.

Decipher score is predictive of distant metastasis benefit of abiraterone in STAMPEDE node negative (N0) patients STAMPEDE All: Treatment x Decipher Interaction for DM: p = 0.11 STAMPEDE N0: Treatment x Decipher Interaction for DM: p = 0.01 Y-axis: 10-year distant metastasis (%) (0, 25, 50); X-axis: Decipher (0.25, 0.50, 0.75, 1.00) Legend: RT + ADT | RT + ADT + Abi Rug rows: RT + ADT + Abi | RT + ADT Sarah Howlett

[Slide 1] Figure 1: TheraPb Phase 1 study design Phase 1b - Dose Escalation | Interval 3+3 dose-escalation incorporating dose optimization strategies N=22 mAPMR: Post ≥1 ARPi; Post taxane or declined; RPT naïve; PSMA+ (>liver) ADVC001 [radioactive syringe icon] Cohort 1: 60 MBq Q6W, 4 doses Cohort 2a: 120 MBq Q4W, 4-6 doses Cohort 3a: 160 MBq Q4W, 4-6 doses Cohort 3b: 160 MBq Q2W, 4-6 doses Cohort 4a: 200 MBq Q4W, 4-6 doses Cohort 4b: 200 MBq Q2W, 4-6 doses Cohort 4c: 200 MBq Q1W, 4-6 doses Key endpoints Safety and Tolerability: DLTs; AEs, SAEs and rAESIs; PRO-CTCAE Efficacy: PSA response; ORR, rPFS, OS PK, Metabolites and Dosimetry: Organ absorbed dose; Clearance, Cmax, metabolites Exploratory Imaging and Biomarkers: 212Pb-PET; Predictors for toxicity, efficacy Recommended Phase 2 Dose and Regimen [RP2D icon] Timeline: Eligibility -> Treatment period 6-24 weeks -> SPECT/CT, Contrast CT, Bone scan, PSMA PET/CT, Safety labs / PK / PSA, TEAEs and TRAEs (from C1D1 to 12 weeks post last dose) -> Follow-up up to 5 yrs from C1D1 Clinical: related SAEs, rAESIs, clinical progression, subsequent anti-cancer tx Laboratory: safety labs, PSA Q6M Imaging: contrast CT / bone scan Q3M until progression [Slide 2] Figure 2: Phase 1b Swimmer plot Patients 1-22 (rows); X axis: Months since C1D1 (0-24) Legend: Study completion; Withdrawal; Death; Follow-up on-going; PR or CR (per RECIST v1.1); PSA50 (first instance); Radiographic or clinical progression; Commencement of new systemic anti-cancer therapy Cohort colors: Cohort 1 60 MBq Q6W; Cohort 2a 120 MBq Q4W; Cohort 3a 160 MBq Q4W; Cohort 3b 160 MBq Q2W; Cohort 4a 200 MBq Q4W; Cohort 4b 200 MBq Q2W; Cohort 4c 200 MBq Q1W; Follow-up [Slide 3] Table 2: Overall safety including long-term follow-up Incidence of AEs by number of events (pts) | Overall N = 22 | RP2D n = 15 Treatment-emergent AEs (TEAEs) | 210 (22) | 138 (15) Grade ≥ 3 | 26 (9) | 19 (5) SAEs | 16 (9) | 10 (5) Dose-limiting toxicity | 0 | 0 Leading to treatment discontinuation or dose reduction | 0 | 0 Leading to dose delay | 4 (4) | 1 (1) Treatment-related AEs (TRAEs) | 94 (20) | 64 (15) Grade ≥ 3 [a] | 2 (2) | 2 (2) SAEs | 0 | 0 Leading to treatment discontinuation or dose reduction | 0 | 0 Leading to dose delay | 1 (1) | 1 (1) Late radiation AEs of special interest (rAESIs) | 0 | 0 [Slide 4] Figure 3a: Best PSA response as percentage change from baseline in all pts (N = 22) Y axis: Best percentage change from baseline PSA (60 to -100); PSA50 reference line Legend: 60 MBq (Q6W); 120 MBq (Q4W); 160 MBq (Q4W, Q2W); 200 MBq (Q4W, Q2W, Q1W) Figure 3b: Best percentage change in target lesion sum of diameters from baseline in RECIST-evaluable pts (n = 6), by BICR Y axis: Best percentage change in SOD from baseline (20 PD, -30 PR, -100 CR) Bars: cPR -42%; cPR -67%; cPR -77%; uCR -82%; cCR -100%; uCR [a] Legend: 200 MBq (Q4W, Q2W, Q1W); 160 MBq (Q4W, Q2W); 120 MBq (Q4W)

[Slide 4] Patient selection across emerging immunotherapies Molecular biomarker, antigen expression and disease burden define three different treatment questions dMMR / MSI-H · ~3% | Antigen-positive | Low burden Checkpoint inhibition, dual where tolerated | T-cell engager: STEAP1, KLK2, PSMA with or without co-stimulation | Vaccines and in situ immunotherapy SETTINGS: Localized peri-operative, adjuvant; mHSPC; mCRPC | SETTINGS: Metastatic setting; High-risk localized; Neo adjuvant | SETTINGS: Localized; Oligometastatic? [Small italic notes under each column illegible] Prognostic biomarkers across all emerging immunotherapies? | ctDNA%, tumor burden on conventional imaging or PSMA-PET mean total volume

[Slide 1] WUTSUP-02: Study Design • This is a prospective, multi-center, phase II study evaluating neoadjuvant DV with perioperative Tislelizumab in high risk, non-metastatic UTUC (ChiCTR2300067836) Key inclusion criteria: • Age 18-75 • Histologically confirmed UTUC • cT2-4N0-2M0 or cTanyN1-2M0 • HER2 1+/2+/3+ • No contraindications to RNU or systemic therapy Disitamab vedotin 2.0 mg/kg, Q3W + Tislelizumab 200 mg, Q3W (4 cycles)(a) -> RNU(b) -> Tislelizumab 200 mg, Q3W (8 cycles) Primary endpoints: • Pathological complete response (pCR) Secondary endpoints: • Objective response rate (ORR) • Overall survival (OS) • Metastasis-free survival (MFS) • Treatment-related adverse events (TRAEs) a. After 2 cycles, interim assessment (contrast-enhanced CT, ureteroscopy, urine cytology/FISH); if progression, switch to RNU. b. Ipsilateral adrenalectomy and lymph node dissection (as indicated). Yige Bao [Slide 2] WUTSUP-02: Patient Characteristics • By May 2026, 42 patients were enrolled, with 39 having completed RNU* • The median age was 65 years, with 88.1% (37/42) high-grade tumors • HER2 expression was 3+ in 19.1% (8/42), 2+ in 61.8% (26/42), and 1+ in 19.1% (8/42) *An additional 3 patients achieved a clinical complete response and underwent kidney-sparing management Patient Characteristics | N=42 Median age (IQR), years | 65.0 (58.0, 71.0) Male sex, n (%) | 23 (54.8) ECOG PS, n (%) 0 | 33 (78.6) 1 | 9 (21.4) Tumor location, n (%) Ureter | 23 (54.8) Renal pelvis | 19 (45.2) Median tumor size (IQR), cm | 2.1 (1.5-2.7) High grade from biopsy, n (%) | 37 (88.1) HER2-positive, n (%) | 34 (80.9) Cisplatin-Ineligible, n (%) | 12 (28.6) Yige Bao [Slide 3] WUTSUP-02: Tumor Response • Pathological response: the pCR rate was 28.2% (11/39) • Radiographic response: the ORR was 81.0% (34/42) CR: 8 | PR: 26 | SD: 6 | PD: 2 • Subgroup response: the HER2 2+/3+ showed higher pCR (31.3% vs 14.3%) and ORR (87.5% vs 42.9%, p=0.03) than the HER2 1+ | ITT patients (N=42) | RNU patients (N=39) pCR rate(a) | 26.2 (12.9-39.5) | 28.2 (14.1-42.3) ORR | 81.0 (65.9-89.7) | 79.5 (66.8-92.2) a Excluding ypTis, ypTa [Waterfall plot: Primary Tumor Diameter Change (%) by patient 1-42, +20% and -30% reference lines] [Bar charts: pCR and ORR by HER2 2+/3+ vs HER2 1+; in-chart labels too small to read reliably] Yige Bao [Slide 4] WUTSUP-02: Survival and Safety • With a median follow-up of 24 months (IQR, 16–30), the 2-year OS rate was 87.1%, and 2-year MFS rate was 83.9% • The most common TRAEs included ALT/AST increased, hypoesthesia, alopecia, pruritus, and hyperglycemia | Any grade | Grade ≥ 3 Any TRAEs | 38 (90.5%) | 20 (47.6%) ALT/AST increased | 21 (50.0%) | 2 (4.8%) Hypoesthesia | 17 (40.5%) | 2 (4.8%) Alopecia | 17 (40.5%) | 0 (0.0%) Hyperglycemia | 13 (31.0%) | 2 (4.8%) Pruritus | 13 (31.0%) | 1 (2.4%) [KM curve: Overall Survival — Survival Probability vs Time (Months) 0-30] [KM curve: Metastasis-free Survival — Survival Probability vs Time (Months) 0-30] Yige Bao Survival table Endpoint: Overall survival and metastasis-free survival (single arm) DV + tislelizumab (single arm) | 42 Landmark rates: 2 years (OS): DV + tislelizumab 87.1% 2 years (MFS): DV + tislelizumab 83.9%

[Slide 3] PAnTHa: 225Ac-PSMA-Trillium in mCRPC Phase 1 dose escalation (n=50), after ≥1 ARPI and 1-2 taxanes if eligible. No prior radiopharmaceutical PSMA PET: uptake greater than liver in ≥1 lesion. Treatment every 6 weeks for up to 4 doses. Dose range 75-150 kBq/kg. Recommended dose for expansion (RDE): 125 kBq/kg. PSA responses by dose population: PSA50 Overall 58%, RDE 125 kBq/kg 83%; PSA90 Overall 36%, RDE 58% Overall n=50. RDE outcomes are a dose-cohort subgroup. Objective response 46% - Measurable disease, n=24 Disease control 83% - Measurable disease, n=24 Common treatment-emergent adverse events: Dry mouth 76%; Fatigue 48%; Nausea 44% Dry mouth: grade 1 in 52%, grade 2 in 24%, no grade ≥3. Grade ≥3 TEAEs: 38%, including lymphopenia in 20%. Serious TEAEs: 16%. TEAE-related discontinuation: 4%. No dose limiting toxicities or treatment-related deaths. PSA-PFS, rPFS and OS not reported. Preliminary dose escalation results, cutoff 18 September 2025. [reference footnote illegible] [Slide 4] VIOLET: 161Tb-PSMA-I&T in mCRPC Single-arm phase 1/2 interim analysis (n=30), after ARPI and taxane unless unsuitable. PSMA PET: SUVmax >20 in ≥1 lesion and ≥10 in measurable soft tissue, with no FDG-discordant disease. Up to 6 cycles every 6 weeks. Recommended starting dose 7.4 GBq, reducing by 0.4 GBq per cycle. PSA responses 70% 40% PSA50 21/30 PSA50 (95% CI 51-85) PSA90 12/30 PSA90 (23-59) Median PSA-PFS 9.0 months 95% CI 7-15.1 Median rPFS 11.1 months 95% CI 8.6-11.7 Median follow up 11.5 months Few severe treatment-related adverse events Grade 3 TRAEs: 7% (2/30) Pain and lymphopenia (one patient each) No grade 4 TRAEs No dose-limiting toxicity No treatment-related deaths, dose reductions or discontinuations Dry mouth 70% Anemia 67% Lymphopenia 63% Fatigue 40% Any grade. Dry mouth was grade 1 in all cases Buteau et al. Lancet Oncol 2026 [illegible] [illegible]

[Slide 1] 1. Impact of GLP1 Agonists on Cancer Incidence KEY ISSUE Growing global obesity burden 13 human malignancies associated with obesity and identified as obesity-associated cancers (OACs). KEY INSIGHT GLP-1 receptor agonists are associated with decreased cancer incidence in diabetic and obese patients. ESMO COMMITMENT ESMO advocates for increased investment in cancer prevention research to reduce the risk of OACs. GLP-1 receptor agonist use and cancer risk in obese nondiabetic adults Hsu A. et al., Annals of Oncology (2026) [Slide 2] 3. "Most Favoured Nation (MFN)" KEY ISSUE High U.S. drug prices compared with other high-income countries. MFN pricing policies aim to align U.S. medicine prices with the lowest prices paid in comparable markets. KEY INSIGHT MFN pricing may also lead manufacturers to alter global pricing and launch strategies, with potential consequences for patient access to medicines worldwide. ESMO COMMITMENT ESMO advocates for fair, sustainable pricing and reimbursement of cancer medicines, promoting value-based access and affordability while safeguarding timely patient access to innovative cancer treatments worldwide. Most-Favoured-Nation pricing for prescription drugs in US Medicare: a cohort study Thomas J Hwang et al., The Lancet (2026) [Slide 3] 6. AI for Clinical Decision-Making KEY ISSUE AI-driven clinical decision-making in oncology. Autonomous AI agents are increasingly capable of supporting complex oncology clinical decision-making. KEY INSIGHT AI agents can perform oncology decision-making tasks at near expert level in selected scenarios but require rigorous validation, governance, and human oversight is needed. ESMO COMMITMENT ESMO promotes safe and responsible AI adoption in cancer care through its Artificial Intelligence and Digital Oncology (AIDO) Committee and ESMO Basic Requirements for AI-Based Biomarkers in Oncology (EBAI), advancing real-world evidence, validation standards, transparency and human oversight. Development and validation of an autonomous artificial intelligence agent for clinical decision-making in oncology Ferber D. et al., Nature (2026). [Slide 4] 10. Oncology Workforce Crisis KEY ISSUE Growing global cancer workforce shortage. An estimated 100 million healthcare worker shortfall by 2050 threatens our ability to deliver quality care. KEY INSIGHT Without urgent investment in training, retention, wellbeing, and multidisciplinary cancer teams, workforce shortages will continue to worsen while cancer incidence rises. ESMO COMMITMENT ESMO advocates for policy solutions to the cancer workforce crisis and supports the EU's efforts to reduce burnout by strengthening psychosocial risk prevention. ESMO Resilience Task Force Reaccommodation's to Manage Psychosocial Risks, Optimise Well-Being and Reduce Burnout in Oncology. Lim KHJ et al., ESMO Open (2024).

[Slide 1] Study design (reconstructed individual survival data) KEYNOTE-564 Phase 3, randomised, placebo-controlled ■ Pembrolizumab (A) n = 496 ■ Placebo (B) n = 498 LITESPARK-022 Phase 3, randomised, placebo-controlled ■ Pembro + belzutifan (C) n = 921 ■ Pembro + placebo (D) n = 920 Digitisation of published Kaplan–Meier curves (DFS and OS) WebPlotDigitizer v5.7 · numbers at risk, events and censoring Reconstruction of individual time-to-event data Guyot iterative algorithm · R v4.5.3 Validation against published curves, HRs and event counts No difference between pembrolizumab arms (A) vs (D) Pooled pseudo-individual dataset · N = 2,835 Indirect comparison: Pembro + belzutifan (C) vs placebo (B) Cox proportional hazards models · DFS (primary) and OS (secondary) [Slide 2] 2,835 patients in the pooled reconstructed dataset 994 KEYNOTE-564 arms A + B 1,841 LITESPARK-022 arms C + D ~85% M0 intermediate-to-high risk in all arms Comparable populations: age, sex, ECOG PS and risk category were similar (all P ≥ 0.05). Race/ethnicity differed (P < 0.01) and PD-L1 data were more often missing in LITESPARK-022. Reconstructed pembrolizumab curves did not differ between trials (DFS and OS). Table 1. Baseline characteristics across KEYNOTE-564 and LITESPARK-022 Characteristic | KEYNOTE-564: Pembrolizumab (A) n = 496 | Placebo (B) n = 498 | LITESPARK-022: Pembro + belzutifan (C) n = 921 | Pembro + placebo (D) n = 920 | P value (B) vs (C) | (A) vs (D) Median age, years (range) | 60 (27–81) | 60 (25–84) | 59 (20–91) | 60 (24–86) | – | – Male sex, n (%) | 347 (70%) | 359 (72%) | 673 (73%) | 642 (70%) | 0.74 | 0.99 ECOG PS 1, n (%) | 75 (15%) | 72 (15%) | 133 (14%) | 143 (16%) | 0.99 | 0.90 Radical nephrectomy, n (%) | 459 (93%) | 459 (92%) | NA | NA | – | – Race or ethnic group, n (%) | | | | | <0.01 | <0.01 American Indian or Alaska native | 10 (2%) | 2 (0%) | – | – Asian | 63 (13%) | 75 (15%) | – | – Black | 7 (1%) | 5 (1%) | – | – Multiple | 8 (2%) | 5 (1%) | – | – White | 372 (75%) | 376 (76%) | 576 (63%) | 581 (63%) Missing data | 36 (7%) | 35 (7%) | 42 (5%) | 24 (3%) PD-L1 CPS, n (%) | | | | | 0.59 | 0.60 <1 | 124 (25%) | 113 (23%) | 222 (24%) | 217 (24%) ≥1 | 365 (74%) | 383 (77%) | 547 (59%) | 502 (55%) Missing | 7 (1%) | 2 (0%) | 152 (17%) | 201 (22%) Risk category, n (%) | | | | | 0.057 | 0.058 M0 intermediate-to-high risk | 422 (85%) | 433 (87%) | 779 (85%) | 782 (85%) M0 high risk | 40 (8%) | 37 (7%) | 58 (6%) | 53 (6%) M0 other | 5 (1%) | 0 (0%) | – | – M1 NED | 29 (6%) | 28 (6%) | 84 (9%) | 84 (9%) Sarcomatoid features, n (%) | | | | | 0.75 | 0.46 Present | 52 (10%) | 52 (12%) | 104 (11%) | 84 (9%) Absent | 414 (84%) | 415 (83%) | 730 (79%) | 724 (79%) Unknown | 30 (6%) | 24 (5%) | 87 (9%) | 112 (12%) P values by χ2 test. ECOG PS, Eastern Cooperative Oncology Group performance status; NA, not available; NED, no evidence of disease. CPS, combined positive score. Percentages as reported in the publication.5 [Slide 3] HR 0.55 — DFS, (C) vs (B) — 95% CI 0.45–0.69; P < 0.001 79.4% — 24-mo DFS, (C) vs 67.8% with placebo (B) +11.6 — percentage points absolute DFS gain at 24 months HR 0.56 — OS, (C) vs (B) — 95% CI 0.35–0.89; P = 0.014 A Disease-free survival: pembrolizumab + belzutifan vs placebo Legend: Placebo (n = 498); Pembrolizumab + belzutifan (n = 921) No. of Events Pembrolizumab + belzutifan 186 Pembrolizumab 151 Hazard ratio, 0.55 (95% CI, 0.45–0.69) P < .001 Y axis DFS (%); X axis Months No. at risk (0,5,...,40 months) Placebo: 498 439 389 341 209 145 56 19 1 Pembrolizumab + belzutifan: 921 867 824 735 599 417 153 60 0 Survival table Endpoint: DFS, pembrolizumab + belzutifan (C) vs placebo (B), reconstructed indirect comparison Pembrolizumab + belzutifan (C) | 921 Placebo (B) | 498 HR: 0.55 HR CI: 0.45-0.69 P: <0.001 Landmark rates: 24 months: Pembrolizumab + belzutifan (C) 79.4% | Placebo (B) 67.8% [Slide 4] B Overall survival: pembrolizumab + belzutifan vs placebo Legend: Placebo (n = 498) | Pembrolizumab + belzutifan (n = 921) Y axis: OS (%) ; X axis: Months No. of Events Pembrolizumab + belzutifan 38 Pembrolizumab 33 Hazard ratio, 0.56 (95% CI, 0.35-0.89) P = .014 No. at risk (months 0, 5, 10, 15, 20, 25, 30, 35, 40): Placebo: 498 494 485 480 336 209 117 48 3 Pembrolizumab + belzutifan: 921 915 905 845 746 593 319 97 0 Survival table Endpoint: OS Placebo | 498 Pembrolizumab + belzutifan | 921 HR: 0.56 HR CI: 0.35-0.89 P: .014

[Slide 1] HRR gene alteration prevalence Total prostate cancer patients, N = 18691 Y-axis: Prevalence among prostate cancer patients ATM 13.5% (n=1923) CHEK2 10.3% (n=1922) BRCA2 6.4% (n=1193) CDK12 3.7% (n=685) BRCA1 1.6% (n=296) PALB2 0.9% (n=164) BRIP1 0.5% (n=95) RAD54L 0.3% (n=63) BARD1 0.3% (n=58) RAD51B 0.2% (n=45) CHEK1 0.2% (n=38) FANCL 0.2% (n=34) RAD51D 0.2% (n=33) RAD51C 0.2% (n=30) Overall HRR 32% (n=5975) [Slide 2] Variant origin prediction by gene (mutations only) Patients with an HRR gene mutation, N = 5632 Y-axis: Patients (%) Gene (n): Germline mutation / Tumor-somatic mutation(s), no germline / CH-predicted mutation(s) only ATM (n=2423): 9% / 15% / 76% CHEK2 (n=1905): 8% / 3% / 87% [small grey indeterminate segment, unlabeled] BRCA2 (n=957): 40% / 53% / 7% CDK12 (n=644): 2% / 80% / 18% BRCA1 (n=260): 21% / 47% / 32% PALB2 (n=156): 31% / 46% / 24% BRIP1 (n=89): 49% / 33% / 18% RAD54L (n=57): 74% / 21% / 5% BARD1 (n=53): 34% / 62% / 4% RAD51B (n=33): 52% / 33% / 15% CHEK1 (n=33): 14% / 69% / 17% FANCL (n=31): 58% / 42% / — RAD51D (n=28): 43% / 39% / 18% RAD51C (n=22): 55% / 23% / 23% Overall HRR (n=5632): 18% / 26% / 56% Legend: Germline mutation; Tumor-somatic mutation(s), no germline; CH-predicted mutation(s) only; Indeterminate origin [Slide 3] TTNT: CH-only vs. germline/tumoral HRR Median: 6.7 vs. 5.8 HR 3.24 (95%CI 1.18-8.87); p = 0.0224 CH-predicted mutation(s) only (orange) Germline/tumoral HRR alterations (blue) Y-axis: Probability of Next Treatment; X-axis: Months from Cohort Eligibility Number at risk (0, 6, 12, 18, 24, 30, 36 months): Germline/tumoral HRR (blue): 53, 28, 12, 5, 0, 0, 0 CH-predicted only (orange): 8, 3, 0, 0, 0, 0, 0 rwOS: CH-only vs. germline/tumoral HRR Median: 15.3 vs. 8.3 HR 3.52 (95%CI 1.09-11.38); p = 0.036 CH-predicted mutation(s) only (orange) Germline/tumoral HRR alterations (blue) Y-axis: Probability of Survival; X-axis: Months from Cohort Eligibility Number at risk (0, 6, 12, 18, 24, 30, 36 months): Germline/tumoral HRR (blue): 52, 40, 26, 20, 13, 9, 8 CH-predicted only (orange): 8, 4, 2, 1, 1, 0, 0 Survival table Endpoint: TTNT and rwOS (two panels) Germline/tumoral HRR alterations | 53 CH-predicted mutation(s) only | 8 HR: TTNT HR 3.24; rwOS HR 3.52 HR CI: TTNT 1.18-8.87; rwOS 1.09-11.38 P: TTNT p = 0.0224; rwOS p = 0.036 [Slide 4] CH-predicted mutations pre- and post- PARPi Collected before PARPi exposure (n=366) | Collected after PARPi exposure (n=66) Any CH gene (OR 0.82): 79.2% | 75.8% TP53 (OR 1.35): 32.5% | 39.4% ATM (OR 1.01): 27.0% | 27.3% DNMT3A (OR 0.56): 34.4% | 22.7% TET2 (OR 0.67): 24.9% | 18.2% CHEK2 (OR 0.81): 18.0% | 15.2% ASXL1 (OR 0.45): 15.6% | 7.6% TERT (OR 1.35): 5.7% | 7.6% NF1 (OR 1.79): 4.4% | 7.6% CBL (OR 2.13): 2.2% | 4.5% U2AF1 (OR 1.40): 2.2% | 3.0% SF3B1 (OR 0.17): 8.5% | 1.5% RAD21 (OR 0.50): 3.0% | 1.5% GNAS (OR 0.69): 2.2% | 1.5% IDH2 (OR 1.11): 1.4% | 1.5% BRAF (OR 1.86): 0.8% | 1.5% STAG2 (OR 2.79): 0.5% | 1.5% MYD88 (OR 2.79): 0.5% | 1.5% EZH2 (OR 2.79): 0.5% | 1.5% MLL2 (OR 5.58): 0.3% | 1.5% KDM6A (OR 5.58): 0.3% | 1.5% JAK2 (OR 0.00): 1.9% | 0.0% PTPN11 (OR 0.00): 0.8% | 0.0% MPL (OR 0.00): 0.5% | 0.0% BCOR (OR 0.00): 0.3% | 0.0% KRAS: 0.0% | 0.0% 9 patients had paired samples and contributed to both groups. None were statistically significant after correcting for multiple comparisons.

[Slide 1] Figure 1: Overview of study First Regimen of 177Lu-PSMA Therapy - >=4 cycles of 177Lu-PSMA therapy - Favorable response (e.g. PSA decline >=50%, imaging-based favorable response) Disease progression Second Regimen of 177Lu-PSMA Therapy "Re-Challenge Therapy" Baseline: PSMA PET/CT Up to 6 additional cycles of 177Lu-PSMA therapy: 7.4 GBq (+/- 10%) every 6 weeks (+/- 1 week) LTFU Cycles 1,3,5 | Cycles 2,4,6 177Lu-PSMA injection; SPECT/CT (Cx D1-2); Labs (w3); Labs (w5) 177Lu-PSMA injection; SPECT/CT (Cx D1-2); Labs & PSMA PET on (w3); Labs (w5) [Slide 2] Table 1: Baseline Characteristics Characteristics | Median (Q1-Q3) / n (%) Age, years | 72.0 (62.5-81.5) Gleason score >=8 | 11 (61.1%) ECOG performance status 0-1 | 18 (90.0%) 2 | 2 (10.0%) Prior radical prostatectomy | 6 (30.0%) Prior definitive radiation therapy | 6 (30.0%) Prior other radiation therapy | 16 (80.0%) Prior ARPI, lines 1 | 8 (40.0%) >=2 | 12 (60.0%) Prior chemotherapy, lines 0 | 1 (5.0%) 1 | 13 (65.0%) >=2 | 6 (30.0%) No. of cycles during initial LuPSMA 4 | 1 (5.0%) 6 | 19 (95.0%) Interval initial LuPSMA completion to Re-LuPSMA initiation, [mo] | 9.2 (5.3-21.2) Intervening systemic therapy between LuPSMA courses | 7 (38.9%) [Slide 3] Table 3: Overview of Related Adverse Events . | Overall | % | >= Grade 3 | % Hematological disorders Lymphocytopenia | 14 | 70.0 | 7 | 35.0 Leukopenia | 11 | 55.0 | 1 | 5.0 Anemia | 8 | 40.0 | 7 | 35.0 Thrombocytopenia | 8 | 40.0 | 2 | 10.0 Neutropenia | 4 | 20.0 | 0 | 0.0 Renal disorders Creatinine increased | 3 | 15.0 | 1 | 5.0 GI disorders Nausea | 7 | 35.0 | 0 | 0.0 Vomiting | 4 | 20.0 | 0 | 0.0 Constipation | 2 | 10.0 | 0 | 0.0 Diarrhea | 1 | 5.0 | 0 | 0.0 General disorders Dry mouth | 10 | 50.0 | 0 | 0.0 Dry eyes | 2 | 10.00 | 0 | 0.0 Fatigue | 8 | 40.0 | 0 | 0.0 Bone pain | 1 | 5.0 | 0 | 0.0 Decreased appetite | 1 | 5.0 | 0 | 0.0

[Slide 1] FACT-P Total Score(T) * Indicates p<0.05 [Line chart: Mean change from baseline vs Weeks 6, 12, 18, 24, 30, 36; legend LuP, Doc; asterisks at weeks 24, 30, 36] Trial Outcomes Index (TOI)(T) [Line chart: Mean change from baseline vs Weeks 6, 12, 18, 24, 30, 36; legend LuP, Doc; asterisks at weeks 24, 30] [Slide 2] FACT-G * Indicates p<0.05 [Line chart: Mean change from baseline vs Weeks 6, 12, 18, 24, 30, 36; legend LuP, Doc; asterisks at weeks 12, 24, 30, 36] Physical Well Being [Line chart: weeks 6-36; legend LuP, Doc; asterisks at weeks 24, 30] [Slide 3] FACT-P Total Score Improved (%) | Stable (%) | Worse (%) | Missing (%) Week 12 LU-P: 28 | 30 | 28 | 14 Week 12 DOCE: 22 | 36 | 31 | 10 Week 24 LU-P: 29 | 34 | 18 | 19 Week 24 DOCE: 15 | 27 | 32 | 26 % patients Trials Outcomes Index (TOI) Improved (%) | Stable (%) | Worse (%) | Missing (%) Week 12 LU-P: 16 | 51 | 21 | 12 Week 12 DOCE: 12 | 50 | 29 | 9 Week 24 LU-P: 13 | 58 | 10 | 19 Week 24 DOCE: 6 | 42 | 26 | 26 % patients [Slide 4] GP5 Item - Side Effect Bother(+) (+)Higher scores indicate worse side effect bother [Line chart: Mean change from baseline vs weeks 6, 12, 18, 24, 30, 36; legend LuP, Doc; asterisks at weeks 6, 12, 24, 30, 36] Worst Recorded GP5 'Side Effect Bother' Score 0 | 1 to 2 | 3 to 4 LU-P: 28.6% | 56.0% | 15.4% DOC: 9.9% | 49.5% | 40.7% Kruskal-Wallis P < 0.001 % patients

[Slide 1] Anti-PD(L)1 checkpoint inhibition in dMMR / MSI-H mAPMR International RWD patients with dMMR mAPMR on anti-PD-(L)1 46% ORR 60% PSA50 response PFS across the entire cohort mPFS 7.7 months (95% CI 5.2-11.7) [PFS KM curve: Progression-free survival (proportion) vs Time in months 0-36; curve annotations 33%, 27%, 26%] Number at risk: [illegible — low resolution] [OS KM curve: Overall survival (proportion) vs Time (mo) 0-72] Median OS 27 mo (95% CI 17.7-43.5) Number at risk: [illegible — low resolution] van Wilpe Eur Urol Oncol 2025 [partly illegible] Speaker panel: Niven Mehra — Emerging immune-based therapies Survival table Endpoint: PFS and OS (real-world, anti-PD-(L)1 in dMMR/MSI-H mAPMR) Entire cohort — PFS | 7.7 months | 5.2-11.7 Entire cohort — OS | 27 mo | 17.7-43.5 [Slide 2] KLK2 × CD3 T-cell engager: pasritamig 174 patients (phase 1) 42.4% PSA50 at RP2D Stein et al. J Clin Oncol 2025 — pasritamig (JNJ-78278343) First-in-class KLK2 × CD3 bispecific. RP2D: 3.5 mg day 1, 18 mg day 8, 300 mg day 15, then every 6 weeks. Largest T-cell-engager dataset in prostate cancer to date. [Waterfall plot — axis labels illegible] [Safety table: columns All-Treated Population / RP2D Safety Population (IV Only); rows Participants with ≥1 TRAE, n (%); Serious TRAEs, n (%); Grade ≥3 TRAEs, n (%); TRAEs leading to treatment discontinuation, n (%) — cell values illegible at this resolution] KLK2 is more prostate-restricted than PSMA, which may widen the therapeutic window. Stein MN et al. J Clin Oncol 2025 [remainder illegible] [Slide 3] Next-generation engager design Engineering approaches to limit cytokine release, antigen escape and on-target/off-tumour toxicity MASKING / CONDITIONAL — Engager activated by tumour-associated proteases; systemic cytokine release markedly reduced. — VIR-5500 (PSMA (masked) × CD3 masked); JANX007 (PSMA × CD3 masked) COSTIMULATION (signal 2) — CD28 or 4-1BB engager alongside a CD3 engager, providing co-stimulation and limiting exhaustion. — Nezastomig (REGN5678, PSMA × CD28); JANX007+ JANX013 (CD28 costimulation) [small text, partly illegible] ALTERNATIVE EFFECTORS — Recruit γδ T cells or NK cells instead of conventional αβ T cells. — γδ-directed engagers; DLL3 and B7-H3 for NEPC [small text, partly illegible] [References — illegible] [Slide 4] Does masking widen the therapeutic window? First clinical signals with JANX007 and VIR-5500 MASKING / CONDITIONAL — Engager activated by tumour-associated proteases; systemic cytokine release markedly reduced. VIR-5500 (PSMA (masked) × CD3 masked) — 1/51 (2%) with CRS ≥ G3 [small text, low resolution] JANX007 (PSMA × CD3 masked) — 0/109 with CRS ≥ G3 [small text, low resolution] JANX007: NR CRS; 0% CRS ≥ grade 3 [JANX007 schematic: PSMA BD, CD3ε BD, Cleavable linker, Mask, Albumin BD; BD = binding domain; Designed with PK extended half-life] 28% CRS; 2% CRS ≥ grade 3 VIR 5500 [schematic with labeled domains — illegible] ≥ 3000 µg/kg Q3W: 45% ORR ≥ 3000 µg/kg Q3W: 82% PSA50 response [References — illegible]

nature medicine articles > article Article | Open access | Published: 24 October 2026 Neoadjuvant nivolumab with or without visugromab in muscle-invasive bladder cancer: the phase 2a GDFATHER-NEO trial Andrea Necchi, Chiara Mercinelli, Brigida A. Maiorano, Michela Piacentini, Antonio Cigliola, Roberto Iacovelli, Chiara Ciccarese, Chiara Ligato, Patrizia Giannatempo, Alessandro Rametta, Paolo Gontero, Francesco Soria, Christine Schuberth-Wagner, Joerg Wischhusen, Bernhard Rössler, ...

[Slide 1] Study design Key Eligibility Criteria • Histologically confirmed unresectable, advanced RCC with clear cell component • Measurable disease per RECIST v1.1 • Disease progression during or after anti-PD-(L)1 therapy and VEGF-TKI in sequence or in combination • Received ≤3 prior systemic regimens for metastatic disease N ≈758 -> R 1:1 Belzutifan 120 mg PO QD + zanzalintinib 60 mg PO QD Belzutifan 120 mg PO QD + placebo PO QD Treatment will continue until unacceptable toxicity, disease progression, or participant or physician decision to withdraw Stratification factors • IMDC risk (favorable vs intermediate vs poor) • Number of prior TKI (1 vs 2 or 3) • Prior cabozantinib (yes vs no) Primary end points: • PFS per RECIST v1.1 by BICR and OS Key secondary end point: • ORR per RECIST v1.1 by BICR Secondary end points: • DOR per RECIST v1.1 by BICR, safety, and PROs IMDC, International Metastatic RCC Database Consortium; PO, by mouth; QD, once daily; R, randomization. [Slide 2] Status [World map with enrolling countries shaded] Sites currently open to enrollment for LITESPARK-034 • The study is currently enrolling in Argentina, Australia, Brazil, Chile, Czechia, Denmark, France, Germany, Greece, Israel, Italy, Netherlands, Poland, Republic of Korea, Spain, Türkiye, United Kingdom, and the United States

[Slide 1] NIAGARA: Mutational Landscape & Association of Driver Alterations Select DDR-related gene alterations were associated with higher pCR rates in analysis of the combined-treatment arms Top mutated cancer driver genes TP53 50.1% | ARID1A 21.9% | KDM6A 18.4% | RB1 17.0% | KMT2D 14.0% | PIK3CA 13.3% | FGFR3 10.8% | ERBB2 6.9% | TSC1 6.6% | ELF3 6.4% | CDKN1A 6.4% | STAG2 5.7% | CREBBP 5.2% | EP300 4.7% | ERCC2 4.7% | RBM10 4.7% | KMT2C 4.4% | FAT1 4.4% | CDKN2A 4.2% | ZFP36L1 4.2% TP53 (50%) represented the most frequent mutation, significantly co-occurring with RB1, and is mutually exclusive with FGFR3 TMB status: High | Low. Alterations: Missense | Truncating | Splice site | In-frame indel | Multiple Association of gene mutation with probability of pCR (combined-arm analysis) [forest plot, OR (log OR); P<0.05 vs Non-significant; Favours pCR | Favours non-pCR] Increased pCR rates in patients with ERCC2, KMT2C, and FAT1 mutations ERCC2 (12/16, 75%) | KMT2C (11/15, 73%) | FAT1 (14/18, 78%) pCR rates in patients with gene mutation DDR, DNA damage response; OR, odds ratio; pCR, pathological complete response; TMB, tumour mutational burden [Slide 2] NIAGARA: EFS Benefit in Mutation-Defined Subgroups Addition of perioperative durvalumab to NAC consistently improved EFS across mutation-defined subgroups, with numerical increase of benefit for TMB-high and TP53- and FGFR3-mutated subgroups Population/subgroup | D arm | C arm | HR (95% CI) ITT (N=1063) | 533 | 530 | 0.68 (0.56-0.82) BEP (n=407) | 209 | 198 | 0.64 (0.45-0.89) TMB-high | 84 | 77 | 0.44 (0.25-0.76) TMB-low | 125 | 121 | 0.82 (0.53-1.26) TP53 mutated | 107 | 97 | 0.51 (0.32-0.81) TP53 WT | 102 | 101 | 0.81 (0.49-1.31) FGFR3 mutated | 20 | 24 | 0.24 (0.07-0.87) FGFR3 WT | 189 | 174 | 0.70 (0.49-0.99) TP53CC mutated (pathway) | 129 | 125 | 0.58 (0.38-0.88) RTK mutated (pathway) | 86 | 82 | 0.49 (0.28-0.85) DDR mutated (pathway) | 50 | 48 | 0.72 (0.36-1.43) Histone-mod mutated (pathway) | 83 | 85 | 0.63 (0.37-1.08) Chromatin remodelling (pathway) | 54 | 44 | 0.48 (0.23-0.98) EFS HR (log scale): Favours D arm | Favours C arm Graph depicts EFS HR (square) and 95% confidence intervals (lines), and was assessed by the Cox proportional hazards model, with treatment as the only covariate and ties handled by the Efron approach. BEP, biomarker-evaluable population; C, comparator; CC, cell cycle; CI, confidence interval; D, durvalumab; DDR, DNA damage response; EFS, event-free survival; HR, hazard ratio; ITT, intention-to-treat population; NAC, neoadjuvant chemotherapy; RTK, receptor tyrosine kinase; TMB, tumour mutational burden; WT, wild-type.

Acknowledgements • Special thanks to Gabriel Malouf, Pavlos Msaouel, Camillo Porta, Matteo Brunelli, Sebastiano Buti, Veronica Huber, Licia Rivoltini, Francesca Vignani, Veronica Prati, Marco Barella, Elena Verzoni and Marco Stellato who contributed to the preparation of this abstract and to conduct the REPRINT trial

TRAE • Ivonescimab monotherapy showed a manageable safety profile • The most common TRAEs were proteinuria, rhinitis, pruritus • No treatment-related deaths occurred N=26 | All Grade | Grade 3-4 TRAE, n (%) | 25 (96.2%) | 7 (26.9%) TRAE(≥20%) PT Term, n (%) | All Grade | Grade 3-4 Proteinuria | 11 (42.3%) | 4 (15.4%) Rhinitis | 10 (38.5%) | 0 Pruritus | 9 (34.6%) | 0 Hypertension | 7 (26.9%) | 0 Hypertriglyceridemia | 7 (26.9%) | 1 (3.8%) Arthralgia | 6 (23.1%) | 0 Data cutoff date: July 14, 2026 Jun Guo, MD

[Slide 1] KEYNOTE-564 Study Design (NCT03142334) Median study follow-up: 85.9 months (range, 76.6-103.2) Key Eligibility Criteria • Age ≥18 years • Histologically confirmed clear cell RCC • No prior systemic therapy for advanced disease • Surgery ≤12 weeks prior to randomization • Postnephrectomy intermediate-high risk of recurrence (M0): – pT2, grade 4 or sarcomatoid, N0 – pT3, any grade, N0 • Postnephrectomy high risk of recurrence (M0): – pT4, any grade, N0 – Any pT, any grade, N+ • Postnephrectomy and complete resection of metastasis (M1 NED) • ECOG PS score of 0 or 1 R 1:1 N = 496: Pembrolizumab 200 mg IV Q3W for ~1 year (≤17 cycles) N = 498: Placebo IV Q3W for ~1 year (≤17 cycles) Stratification Factors • M stage (M0 vs M1 NED) • M0 group further stratified: – ECOG PS score of 0 vs 1 – US vs rest of the world End Points Primary: DFS by investigator assessment Key secondary: OS Other secondary: Safety Exploratory: DFS and OS by protocol-defined risk subgroups and sarcomatoid features Data cutoff date: February 4, 2026. [Slide 2] DFS by Protocol-Defined Risk Subgroup Intermediate-high: HR 0.71 (95% CI, 0.58-0.87) Pembro landmarks 63.5% | 62.1% | 59.8%; Placebo 55.5% | 52.3% | 48.3% No. at risk (0–108 mo, every 12): Pembro 422 348 314 287 253 228 209 95 21 0; Placebo 433 340 299 274 239 211 190 88 15 0 High: HR 0.54 (95% CI, 0.31-0.94) Pembro landmarks 44.6% | 44.6% | 44.6%; Placebo 29.1% | 26.2% | 19.4% No. at risk: Pembro 40 26 18 18 17 16 15 8 1 0; Placebo 37 19 14 12 10 10 9 3 0 0 M1 NED: HR 0.50 (95% CI, 0.26-0.96) Pembro landmarks 48.9% | 36.2% | 36.2%; Placebo 27.6% | 23.0% | 23.0% No. at risk: Pembro 29 23 22 18 15 12 8 2 1 0; Placebo 28 13 10 8 8 6 5 3 0 0 Data cutoff date: February 4, 2026. Survival table Endpoint: DFS by protocol-defined risk subgroup (investigator) Intermediate-high: Pembro | 422 Intermediate-high: Placebo | 433 High: Pembro | 40 High: Placebo | 37 M1 NED: Pembro | 29 M1 NED: Placebo | 28 HR: Int-high 0.71; High 0.54; M1 NED 0.50 HR CI: 0.58-0.87; 0.31-0.94; 0.26-0.96 [Slide 4] OS by Protocol-Defined Risk Subgroup Intermediate-high: HR 0.66 (95% CI, 0.48-0.91) Pembro landmarks 89.2% | 87.5% | 84.7%; Placebo 83.7% | 80.9% | 78.0% No. at risk: Pembro 422 413 405 396 385 371 362 201 46 0; Placebo 433 425 410 [illegible] 375 358 346 180 42 0 High: HR 0.92 (95% CI, 0.44-1.90) Pembro landmarks 75.0% | 72.5% | 61.4%; Placebo 73.0% | 70.3% | 67.6% No. at risk: Pembro 40 38 36 35 32 30 29 16 1 0; Placebo 37 35 32 29 27 27 26 14 2 0 M1 NED: HR 0.54 (95% CI, 0.19-1.53) Pembro landmarks 86.1% | 82.5% | 78.1%; Placebo 70.6% | 66.9% | 66.9% No. at risk: Pembro 29 29 27 26 25 24 23 9 2 0; Placebo 28 24 22 21 21 19 18 10 1 0 Data cutoff date: February 4, 2026. Survival table Endpoint: OS by protocol-defined risk subgroup Intermediate-high: Pembro | 422 Intermediate-high: Placebo | 433 High: Pembro | 40 High: Placebo | 37 M1 NED: Pembro | 29 M1 NED: Placebo | 28 HR: Int-high 0.66; High 0.92; M1 NED 0.54 HR CI: 0.48-0.91; 0.44-1.90; 0.19-1.53

[Slide 1] De Novo vs Metachronous Metastatic Prostate Cancer (mPC): A Comprehensive Clinicogenomic Analysis ESMO 2026 Poster #2225 Panagiotis J. Vlachostergios, Anthony V. Serritella, Amy K. Taylor, John M. Floberg, Yingying Yu, Stamatina Fragkogianni, Kayla Viets Layng, Rana R. McKay, David F. Jarrard, Glenn Liu, Joshua M. Lang, Christos E. Kyriakopoulos [affiliations: University of Ioannina; Weill Cornell Medicine; University of Wisconsin; Tempus AI; UC San Diego; Henry Dunant Hospital Center, Athens — small text] INTRODUCTION ❑ The differences between de novo and metachronous mPC are incompletely characterized. ❑ This study compared demographic, clinical, molecular, and immunologic features between de novo and metachronous mPC. METHODS The Tempus Lens Platform was used to query the multimodal de-identified database to establish and subsequently analyze a cohort of patients diagnosed with mPC who underwent genomic profiling with xT (DNA) and xR (RNA) (n = 6,306). Patients were classified as de novo (Stage IV at diagnosis) or metachronous (Stage I–III at diagnosis with progression to Stage IV >180 days later). De Novo n = 6,171 Metachronous n = 135 ❑ Immune cell proportions were estimated from RNA using quanTIseq. ❑ Differential gene expression analyses were conducted using Wilcoxon rank-sum tests with Benjamini-Hochberg correction. ❑ Gene set enrichment analysis was conducted using the Hallmark libraries. ❑ Groups were compared using Wilcoxon rank-sum tests or chi-squared tests as appropriate. RESULTS Figure 1. Treatment with prostatectomy or radiation was more common in patients with metachronous disease (prostatectomy: 57% vs 5.2%, radiotherapy: 92% vs 22%, p<0.001). Figure 2. Treatment with upfront chemotherapy with ADT +/- ARPI was more common in de-novo metastatic pts (15.7% vs 8.1%, p=0.037). Figure 3. Metachronous pts were enriched for APLNR (13% vs 4.7%, p<0.001) and ZMYM3 somatic alterations (11% vs 4.8%, p<0.001), while most canonical prostate cancer drivers (e.g., TP53, PTEN, ERG) had a similar frequency between groups. [Fig 3 bar labels, de novo/metachronous %: APLNR 4.5/12.6; ZMYM3 4.7/11.1; LRP1B 9.4/14.8; FOXA1 13.9/20.0; SPOP 10.6/10.4; TP53 32.2/34.1; PTEN 25.6/26.7; ERG 36.1/39.3 — small text, verify against source] Figure 4. Metachronous pts demonstrated increased regulatory T cell and natural killer cell infiltration as estimated by quanTIseq (p<0.001). Figure 5. Inflammatory signaling (TNFα/NF-κB) and KRAS-related pathways were enriched in metachronous pts. SUMMARY ❑ Metachronous mPC is a distinct entity characterized by lower PSA at diagnosis, greater prior local therapy exposure, unique genomic alterations, and a more immunologically suppressed tumor microenvironment compared to de novo disease. ❑ Future work should investigate therapeutic vulnerabilities and differences in survival associated with metachronous vs. de novo disease. ACKNOWLEDGMENTS: We thank Vanessa Nepomuceno, PhD for assistance with poster preparation. [Slide 2] Clinicogenomic Differences Between Early-Onset (EO-PCa) and Average-Onset Prostate Cancer (AO-PCa): A Large-Scale Retrospective Cohort Study ESMO 2026 Poster #2256 Christos E. Kyriakopoulos, Panagiotis J. Vlachostergios, Anthony V. Serritella, Amy K. Taylor, John M. Floberg, Yingying Yu, Stamatina Fragkogianni, Kayla Viets Layng, Rana R. McKay, David F. Jarrard, Glenn Liu, Joshua M. Lang [affiliations: UW–Madison; Henry Dunant Hospital Center, Athens; University of Ioannina; Weill Cornell; Tempus AI; UC San Diego — small text] INTRODUCTION ❑ Early-onset prostate cancer (EO-PCa) is increasingly recognized as biologically distinct, but its clinical behavior, molecular features, and outcomes relative to average-onset (AO-PCa) disease remain incompletely characterized. SUMMARY • EO-PCa exhibits distinct clinical and molecular characteristics compared to AO-PCa. • OS appears to be driven by a combination of age and tumor biology. METHODS The Tempus Lens platform was used to query the de-identified, multimodal database for patients diagnosed with metastatic prostate cancer (N= 6,967) tested with Tempus xT (DNA-seq) and xR (RNA-seq). Eligible de novo metastatic patients had a sample collected within 60 days of diagnosis and were stratified by onset age (EO-PCa ≤55 years; AO-PCa >55 years). ❑ Clinicogenomic characteristics were compared using Wilcoxon rank-sum tests or chi-squared tests as appropriate. ❑ Real-world overall survival (rwOS) was defined as the time from sample collection to death or loss to follow up. ❑ Median rwOS (mOS) was estimated using Kaplan-Meier (KM) curves and compared using log-rank tests. Hazard ratios (HR) were estimated using Cox proportional hazards models. Risk set adjustment was used to account for immortal time bias. RESULTS Figure 1. Frequency of primary radiotherapy or prostatectomy treatment in EO-PCa vs. AO-PCa. Patients with EO-PCa were more likely to undergo primary treatment (radiotherapy: 37.2% vs 26.4%; prostatectomy: 18.7% vs 10.4%, p<0.001). Figure 2. Selected molecular differences between EO-PCa and AO-PCa. EO-PCa was enriched for germline BRCA2 variants (4.8% vs 2.6%, p=0.019), TMPRSS2 fusions (39.7% vs 34.7%, p=0.015), and ERG fusions (41.1% vs 36.6%, p=0.031). Somatic FOXA1 (15.4% vs 9.4%, p<0.001) and SPOP mutations (11.1% vs. 8.2%, p=0.033) were more common in AO-PCa. Figure 3. Median real-world OS for EO-PCa vs. AO-PCa. EO-PCa was associated with longer mOS in the overall cohort [42 vs 32 months, HR = 1.51 (95% CI 1.26-1.81), p<0.001] and in SPOP-alt and TP53-alt sub-cohorts. [Fig 3 bars, EO/AO months: Overall 42.0/32.0; SPOP alt 67.1/52.0; TP53 alt 33.8/25.9] Figure 4. Association of SPOP and TP53 alterations with real-world OS in AO-PCa and EO-PCa. SPOP alterations were prognostic for favorable mOS compared to wild-type (AO-PCa: 52 months vs 30.7 months, p<0.001; EO-PCa: 67.1 months vs 41.7 months, p= 0.41). TP53 alterations were prognostic for worse mOS compared to wild-type (AO-PCa: mOS 25.9 months vs 37.9 months, p<0.001; EO-PCa: mOS 33.8 vs 55.9 months, p<0.001). [Fig 4: four KM panels — A. SPOP Alterations in AO-PCa (Age >55); B. SPOP Alterations in EO-PCa (Age <=55); C. TP53 Alterations in AO-PCa (Age >55); D. TP53 Alterations in EO-PCa (Age <=55); panel p-values printed: A p = 1.79e-09, B p = 0.414, C p = 2.4e-25, D p = 0.000649 — small text] ACKNOWLEDGMENTS: We thank Vanessa Nepomuceno, PhD for assistance with poster preparation.

[Slide 1] Emerging Targets for Theranostics in Solid Tumors [Diagram: y-axis Expression (Tumor cells), x-axis Expression (CAFs/Stroma), Immune cells at origin] Established (highlighted): NIS – Radioiodine (131I) SSTR2 – Lutathera® (177Lu-Dotatate) NET – Metaiodobenzylguanidine (131I-MIBG) PSMA – Pluvicto® (177Lu-PSMA) +endothelium Clinical/Preclinical (italic): Her2, GRPR, DLL3, CAIX, Frα, Integrins (αvβ6), GD2 EphA2 B7-H3 MT1-MMP Integrins (αvβ3) endothelium CXCR4 – 177Lu-PentixaTher PD-L1 CD8, Granzyme B LRRC15 FAP – different FAP-Inhibitors (FAPIs) Legend: Established | Clinical | Preclinical Rainer Hamacher [Slide 2] A decade of lessons 2013 — Lete-cel (NYESO-1) TCR in synovial sarcoma: ORR 20-50%; DOR 15-30 weeks • Higher LDR impacts efficacy (Flu 120mg/m2 + Cy 2700mg/m2) • Bridging therapy is essential option • Correlates of response: expansion, persistence, IL15 levels, depletion of lymphocytes • Correlates of resistance: loss of HLA expression & Ag presenting machinery 2016 — Lete-cel (NYESO-1) TCR in myxoid round cell liposarcoma: ORR 20-40%; DOR 5-7m • Correlates of response: Expansion, persistence, IL15 levels, depletion of lymphocytes • Tocilizumab doesn't appear to impact efficacy 2019 — Afami-cel (MAGE-A4 TCR) in synovial sarcoma + MRCLS: ORR 39% (SS), 25% (MRCLS); DOR 12m (SS), 4m (MRCLS) • Clinical correlates: Lower disease burden, higher MAGEA4 expression, lack of bridging therapy • Higher cell dose may impact efficacy • Integrating cells early is likely better FDA approved in 2024 2020 — Lete-cel (NYESO-1) TCR in synovial sarcoma + MRCLS: ORR 43%; DOR 12.2m • Primary endpoint met BLA planned for late 2026 Sandra P. D'Angelo, MD D'Angelo SP Cancer Discovery 2018, D'Angelo SP JCO 2024, D'Angelo SP Lancet 2024 [Slide 3] The real-world eligibility funnel [Inverted funnel graphic, top to bottom:] Sarcoma population: histology + clinical status Biology: tumor antigen expression Restriction: compatible HLA Feasibility: manufacturing time, bridging, fitness Outcome: response durability [CONSORT flow diagram:] 373 patients pre-screened for HLA eligibility, MAGE-A4 expression, or both -> 268 not positive for HLA or MAGE-A4 105 positive for both HLA and MAGE-A4 -> 42 not eligible or did not wish to enrol 63 enrolled into cohort 1 and underwent leukapheresis (ITT population) -> 10 did not receive lymphodepletion chemotherapy 6 died due to disease progression 3 not eligible 1 withdrawn at investigator discretion 53 received lymphodepletion chemotherapy -> 1 did not receive afami-cel (withdrew consent) 52 received afami-cel (modified ITT) -> 6 ongoing in interventional phase at data cutoff (Aug 30, 2023) 46 completed interventional phase -> 35 discontinued study 31 died 1 non-adherent to study requirements 2 withdrew consent 1 lost to follow-up 11 ongoing in long-term follow-up Sandra P. D'Angelo, MD [Slide 4] Conclusions Afamicel and letecel have shown promise in synovial sarcoma and myxoid round cell liposarcoma • There is an on-going need to identify biomarkers of response/resistance • Consideration of synergistic combinations may optimize efficacy (mRNA vaccines) Leveraging cancer testis antigens as promising therapeutic targets through alternate therapeutic approaches, ie bispecific antibodies will continue to expand the field via off the shelf-options Exploring alternate adoptive cell approaches, ie TIL, CAR-T as well as new targets, ie fusions will contribute to novel and exciting therapeutic options in the future and expand beyond the HLA restriction @sandrapdangelo dangelos@mskcc.org Sandra P. D'Angelo, MD

CONCLUSIONS • There is a high prevalence of predicted CH among HRR gene mutations in liquid-biopsy samples (56% overall), including in BRCA1 (32%) and PALB2 (24%). • Patients with predicted CH-only HRR alterations do not appear to benefit from PARPi monotherapy, highlighting the importance of accounting for HRR mutation origin in therapy selection. • Some CH-predicted mutations are enriched in samples after PARPi exposure, including TP53, NF1, pTERT and CBL. Emmanuel S. Antonarakis, MD

[Slide 1] Figure 1. Age-specific death rates and 95% CI by disease state at enrollment. Legend: mAPMS | APMR Death Rate (per 1,000 person-years) Age group at enrollment (in years): <55 | 55-59 | 60-64 | 65-69 | 70-74 | 75-79 | 80-84 | 85+ [Slide 2] Figure 3. Hazard ratios (95% CI) for the association between age at enrollment and overall survival. Panels: Disease state-stratified | +country, stage M1, PSA, treatment | +hyperlipidemia, diabetes, hypertension, smoking, frailty Hazard Ratio (95% CI) [log axis 1, 2, 3; reference 60-64] Age group at enrollment (in years): <55 | 55-59 | 60-64 | 65-69 | 70-74 | 75-79 | 80-84 | 85+ [Slide 3] Figure 4. Overall survival in 20 patients with mAPMS by the (A) Hannum epigenetic clock and (B) clonal hematopoiesis (CH). A: Overall survival; Time from Enrollment (years) 0-5; Hannum clock: Low | High B: Overall survival; Time from Enrollment (years) 0-5; CH: CH- | CH+ Survival table Endpoint: Overall survival (exploratory cohort, n=20 mAPMS) Hannum clock Low Hannum clock High CH- CH+

[Slide 1] Decipher score associates with higher rates of distant metastases STAMPEDE ADT HR (95% CI): 1.72 (1.34 - 2.21) p-value: <0.001* Distant metastasis (%) 0 25 50 Years since randomisation 0 1 2 3 4 5 6 7 8 Number at risk (% distant metastasis) GC > 0.85: 101 | 91 (10%) | 64 (27%) | 9 (39%) GC 0.61 - 0.85: 65 | 62 (5%) | 50 (12%) | 13 (20%) GC ≤ 0.60: 42 | 40 (2%) | 36 (5%) | 10 (5%) STAMPEDE ADT + abi HR (95% CI): 1.27 (0.96 - 1.67) p-value: 0.10 Number at risk (% distant metastasis) GC > 0.85: 105 | 96 (4%) | 69 (16%) | 17 (19%) GC 0.61 - 0.85: 58 | 57 (0%) | 53 (2%) | 14 (11%) GC ≤ 0.60: 30 | 30 (0%) | 27 (3%) | 6 (14%) Decipher risk group: GC ≤ 0.60 | GC 0.61 - 0.85 | GC > 0.85 Sarah Howlett Survival table Endpoint: Distant metastasis (STAMPEDE ADT panel) GC > 0.85 | 101 GC 0.61 - 0.85 | 65 GC ≤ 0.60 | 42 HR: 1.72 HR CI: 1.34 - 2.21 P: <0.001* Landmark rates: at-risk column 2 (time not printed): GC > 0.85 10% | GC 0.61 - 0.85 5% | GC ≤ 0.60 2% at-risk column 3 (time not printed): GC > 0.85 27% | GC 0.61 - 0.85 12% | GC ≤ 0.60 5% at-risk column 4 (time not printed): GC > 0.85 39% | GC 0.61 - 0.85 20% | GC ≤ 0.60 5% [Slide 2] Decipher score associates with worse prostate cancer specific mortality STAMPEDE ADT HR (95% CI): 1.43 (1.14 - 1.81) p-value: 0.002* Prostate cancer specific mortality (%) 0 25 50 Years since randomisation 0 1 2 3 4 5 6 7 8 Number at risk (% PCSM) GC > 0.85: 101 | 98 (2%) | 80 (15%) | 49 (26%) GC 0.61 - 0.85: 65 | 65 (0%) | 55 (8%) | 43 (12%) GC ≤ 0.60: 42 | 41 (2%) | 38 (5%) | 34 (7%) STAMPEDE ADT + abi HR (95% CI): 1.60 (1.16 - 2.21) p-value: 0.004* Number at risk (% PCSM) GC > 0.85: 105 | 97 (4%) | 78 (10%) | 60 (22%) GC 0.61 - 0.85: 58 | 57 (0%) | 53 (2%) | 47 (5%) GC ≤ 0.60: 30 | 30 (0%) | 28 (3%) | 22 (3%) Decipher risk group: GC ≤ 0.60 | GC 0.61 - 0.85 | GC > 0.85 Sarah Howlett Survival table Endpoint: Prostate cancer-specific mortality (STAMPEDE ADT panel) GC > 0.85 | 101 GC 0.61 - 0.85 | 65 GC ≤ 0.60 | 42 HR: 1.43 HR CI: 1.14 - 1.81 P: 0.002* Landmark rates: at-risk column 2 (time not printed): GC > 0.85 2% | GC 0.61 - 0.85 0% | GC ≤ 0.60 2% at-risk column 3 (time not printed): GC > 0.85 15% | GC 0.61 - 0.85 8% | GC ≤ 0.60 5% at-risk column 4 (time not printed): GC > 0.85 26% | GC 0.61 - 0.85 12% | GC ≤ 0.60 7% [Slide 4] Decipher score identifies RWD cases at higher distant metastasis risk RWD:NCCN high risk Distant metastasis (%) 0 10 20 30 40 50 Years since randomisation 0 1 2 3 4 5 6 7 8 Number at risk (% distant metastasis) GC > 0.85: 2230 | 982 (9%) | 52 (14%) | 15 (16%) GC 0.61 - 0.85: 1565 | 745 (4%) | 43 (5%) | 21 (5%) GC ≤ 0.60: 1402 | 659 (2%) | 54 (3%) | 23 (5%) RWD:NCCN very high risk Number at risk (% distant metastasis) GC > 0.85: 517 | 191 (25%) | 10 (29%) | 4 (36%) GC 0.61 - 0.85: 157 | 76 (7%) | 6 (10%) | 2 (10%) GC ≤ 0.60: 85 | 40 (9%) | 2 (12%) | 1 (12%) Decipher risk group: GC ≤ 0.60 | GC 0.61 - 0.85 | GC > 0.85 Sarah Howlett Survival table Endpoint: Distant metastasis (RWD NCCN high risk panel) GC > 0.85 | 2230 GC 0.61 - 0.85 | 1565 GC ≤ 0.60 | 1402 Landmark rates: at-risk column 2 (time not printed): GC > 0.85 9% | GC 0.61 - 0.85 4% | GC ≤ 0.60 2% at-risk column 3 (time not printed): GC > 0.85 14% | GC 0.61 - 0.85 5% | GC ≤ 0.60 3% at-risk column 4 (time not printed): GC > 0.85 16% | GC 0.61 - 0.85 5% | GC ≤ 0.60 5%

[Slide 2] RESULTS: Duration of Disease Control Rows: FL | PCa | PCa | NSCLC | NB | PCa | PDAC | BLCA | LCC | ACC | GC | PDAC OS STATUS (per row): x | o | o | o | x | o | o | o | o | o | x | x Duration (months): 0 | 5 | 10 | 15 | 20 | 25 | 30 | 35 According to RECIST 1.1: CR, n=4 (33.3%) PR, n=4 (33.3%) SD, n=4 (33.3%) PD, n=0 (0%) -> Ongoing X Death (non-cancer related) x Death o Censored Median duration of response in patients with CR/PR has not been reached (abbreviation legend as on prior slide: ACC, BLCA, FL, GC, LCC, NB, NSCLC, PCa, PDAC) [Slide 3] CONCLUSIONS SL-28 represents a novel class of cell therapy, demonstrating antitumor activity across multiple biologically distinct cancer types CR: 33%; PR: 33%; SD: 33% SL-28 (TELLS) cells are not genetically modified. Their antitumor activity is enhanced through a novel TezR-mediated activation approach, representing the first therapeutic application of this Novel Receptor System SL-28 is an off-the-shelf, allogeneic cell therapy that does not require HLA matching or immunosuppressive conditioning. Clinical experience to date suggests a favorable safety profile, with no observed GvHD, ICANS, or CRS George Tetz, MD, PhD g.tetz@hmi-us.com

[Slide 1] Figure 4. Study design Key eligibility criteria: • Metastatic adenocarcinoma of the prostate • Docetaxel-naïve in the mCRPC setting (prior docetaxel allowed for mHSPC) • 1–2 prior ARPIs • Prior Lu 177-RLT allowed R (1:1) N≈736 Stratification factors: • Prior Lu 177-RLT (yes/no) • Prior taxane in mHSPC (yes/no) • Site of metastasis (liver/bone/other) Elfetabart drozuntecan (Elfe-D)a 6 mg/kg IV Q3W Docetaxel 75 mg/m2 IV Q3Wa,b (up to 10 cycles) Treatment until PCWG3-modified RECIST v1.1-defined progressive disease, unacceptable toxicity, withdrawal of consent, or other discontinuation criteria being met a Patients treated with an LHRH agonist or antagonist must continue such treatment throughout the trial. b Patients randomised to docetaxel will also receive prednisone or prednisolone. [Slide 2] Table 1. Key eligibility criteria Key inclusion criteria • Men aged ≥18 years • ECOG PS 0 or 1 • Histological, pathological, and/or cytological confirmation of prostate cancer (adenocarcinoma histology only) • ≥1 metastatic lesion on baseline CT/MRI, or ≥2 metastatic lesions on bone scan imaging • Prior orchiectomy and/or ongoing ADT with castrate serum/plasma testosterone levels (<50 ng/dL or <1.7 nmol/L) • Progressive disease based on ≥1 of the following: • Serum/plasma PSA progression and PSA ≥1.0 ng/mL at screening • Radiographic soft tissue progression (per PCWG3-modified RECIST v1.1) • Radiographic progression of bone disease: evaluable disease or new bone lesion(s) by bone scan per PCWG3 criteria • No prior systemic cytotoxic chemotherapy (including taxane-based) for mCRPC. Prior docetaxel for initial mHSPC treatment allowed if no radiographic progression and last dose ≥6 months pre-trial • Previous treatment with 1–2 ARPIs and progressive disease during or after ≥8 weeks of ARPI (≥12 weeks for patients with bone progression) • Prior Lu 177-RLT (or investigational PSMA-targeted radioligand) permitted if >8 weeks from completing treatment • Patients with known BRCA mutations should have received previous treatment with PARPi, if available Key exclusion criteria • Uncontrolled and/or untreated brain metastasis • Prior treatment with B7H3-targeted therapy, including B7H3 ADC • Prior treatment with ADC with topoisomerase I inhibitor payload • History of (non-infectious) ILD/pneumonitis that requires steroids, or current ILD/pneumonitis, or suspected ILD/pneumonitis that cannot be ruled out by imaging at screening • Has uncontrolled or significant cardiovascular disease [Slide 3] Table 2. Study endpoints Dual primary • rPFS (by BICR per PCWG3-modified RECIST v1.1) • OS Secondary • TFST, ORR (BICR), DOR (BICR), time to pain progression, investigator-assessed rPFS, time to first symptomatic skeletal-related event, time to PSA progression, PSA response rate, safety and tolerability Exploratory • Health-related quality of life (functional assessment of cancer therapy – prostate)

[Slide 1] Key eligibility criteria • Histologically or cytologically confirmed RCC with mainly clear cell carcinoma with advanced disease • IMDC favourable risk (IMDC score 0) • No prior systemic therapy • ECOG PS 0-1 Ivonescimab 20mg/kg Q3W N=3~6 Ivonescimab 30mg/kg Q3W N=3~6 → Ivonescimab 30mg/kg Q3W (N=30) No DLT was observed. RP2D of ivonescimab was 30 mg/kg Q3W. Primary endpoint: ORR Secondary endpoints: PFS, OS, DCR, DoR, TTR, safety [Slide 2] Ivonescimab (n=23) n (%) Best Overall Response (BOR) CR | 0 (0.0%) PR | 15 (65.2%) SD | 7 (30.4%) PD | 1 (4.3%) Objective Response Rate (ORR) | 65.2% Disease Control Rate (DCR) | 95.7% [Waterfall plot: Best percentage of change from baseline (%), PR / SD / PD] [Slide 3] Efficacy-PFS • Median follow-up was 13.3 months • The median PFS was not reached. The 12-month and 24-month PFS rates were 85.5% and 55.4%, respectively • No deaths have occurred to date 12m-rate 85.5% 24m-rate 55.4% AK112 (months 0, 3, 6, 9, 12, 15, 18, 21, 24, 27) At risk: 23 22 18 14 9 7 5 2 1 0 Censored: 0 0 3 6 11 11 12 15 16 17 Events: 0 1 2 3 3 5 6 6 6 6 Survival table Endpoint: Progression-free survival Ivonescimab (AK112) 30 mg/kg Q3W | 23 | Not reached Landmark rates: 12 months: Ivonescimab 85.5% 24 months: Ivonescimab 55.4% [Slide 4] N=26 | All Grade | Grade 3-4 TRAE, n (%) | 25 (96.2%) | 7 (26.9%) TRAE(≥20%) PT Term, n (%) | All Grade | Grade 3-4 Proteinuria | 11 (42.3%) | 4 (15.4%) Rhinitis | 10 (38.5%) | 0 Pruritus | 9 (34.6%) | 0 Hypertension | 7 (26.9%) | 0 Hypertriglyceridemia | 7 (26.9%) | 1 (3.8%) Arthralgia | 6 (23.1%) | 0

[Slide 1] Table 1. Key Eligibility Criteria Inclusion • Male patients aged ≥18 years • Histologically or cytologically confirmed prostate adenocarcinoma without small cell features • Progressive disease in the setting of surgical or medical castration, with evidence of disease progression as per PCWG3 on treatment with abiraterone ≥12 weeks in the mCSPC setting or first-line mCRPC setting • Metastatic disease in bone documented on bone scan, or in soft tissue documented on CT/MRI scan • ECOG PS of 0 or 1 • Adequate organ function Exclusion • Prior systemic anticancer therapy, beyond protocol-permitted abiraterone-based treatment and docetaxel in the mCSPC setting • Radiotherapy within 2 weeks before randomization • Known or suspected brain metastases or active leptomeningeal disease • Clinically significant cardiovascular disease • Active inflammatory gastrointestinal disease, chronic diarrhea, or previous gastric resection or lap-band surgery [Slide 2] Figure 1. Himalayas-1 Study Design Select Eligibility Criteria • Patients with mCRPC • Prior abiraterone in mCSPC or mCRPC setting • May have received docetaxel in mCSPC setting R 1:1 Rinzimetostat 400 mg QD + Darolutamide 600 mg BID (n=300) Physician's choice of control (ARPI or docetaxel) (n=300) R 1:1 -> Darolutamide 600 mg BID / Enzalutamide 160 mg QD Docetaxel 75 mg/m2 Q3W for up to 10 cycles Endpoints Primary Endpoint • rPFS Key Secondary Endpoint • OS

[Slide 1] Figure 1. Xerostomia incidence Dose (kBq/kg) — Grade 1 / Grade 2 (Participants (%)) 150: 31 / 62 125: 58 / 25 100: 77 / 15 75: 50 / 25 Legend: Grade 1, Grade 2 [Slide 2] Figure 5. Changes in sAA level by dose Panels: 75 kBq/kg | 100 kBq/kg | 125 kBq/kg | 150 kBq/kg Y-axis: Salivary amylase (U/L) 0–1000 X-axis: Treatment day (-42 to 231) [Spaghetti plots of individual patient sAA levels; transient early peaks, highest in 150 kBq/kg panel] [Slide 3] Figure 6. PSA response by xerostomia grade Y-axis: Best % PSA change; X-axis: Xerostomia grade (Grade 0, Grade 1, Grade 2) Dose level legend: 75 kBq/kg, 100 kBq/kg, 125 kBq/kg, 150 kBq/kg [Box plots with individual patient points] | | Grade 0 (n=7) | Grade 1 (n=27) | Grade 2 (n=16) | | PSA50 response, n (%) | 2 (29) | 19 (70) | 10 (63) |

[Slide 1] Clinical applications of plasma ctDNA profiling [Diagram: human silhouette with plasma tube labelled 'Plasma'] ctDNA detection (greyed) ctDNA quantification — Single and serial sampling (greyed; ctDNA% tube and line/curve icons) Prognostication and response prediction (greyed) ctDNA characterisation Genomic profiling — Mutations, structural rearrangements and copy number alterations [chromosome icons] Guiding precision oncology treatments and characterising acquired resistance Signal or noise: tumour, or white blood cells? Edmond M. Kwan, MBBS PhD [Slide 2] ctDNA abundance impacts alteration detection sensitivity ctDNA%: tumor ctDNA -> normal cfDNA; >60% | 30 - 10% | 5 - 1% | <0.1% Examples in mCRPC: SPOP p.F133L, AR p.T878A — Point mutations / most rearrangements AR amplification — Multi-copy gene amplification Chromosome 8p loss, Chromosome 13q loss — Chromosome and arm-level monoallelic CNV Loss of heterozygosity (e.g. BRCA2, MSH2/6, TP53) — Focal monoallelic CNV Homozygous loss (e.g. BRCA2, RB1, TP53, PTEN) — Focal biallelic deletion Detection sensitivity: Targeted (High–Low), Broad, Low-pass [gradient bars] Herberts C et al, Trends Cancer 2021 Edmond M. Kwan, MBBS PhD [Slide 3] Plasma detected almost every BRCA mutation BRCA1/2 mutations — 14 of 15 detected by plasma BRCA1/2 biallelic loss — 3 of 6 detected by plasma Non-BRCA HRR mutations — 29 of 36 detected by plasma [Legend] Filled square: Detected in tissue and plasma; Empty square: Tissue only (plasma miss) Edmond M. Kwan, MBBS PhD

Efficacy and Safety of AK112 (Ivonescimab), a PD-1/VEGF Bispecific Antibody, as First-Line Treatment for Patients with IMDC Favorable-Risk Advanced Renal Cell Carcinoma (aRCC): Results from a Phase Ib/II Study Xieqiao Yan, Xinan Sheng, Juan Li, Simin Li, Bixia Tang, Huichun Tian, Huayan Xu, Yan Kong, Jun Guo Presenter: Jun Guo, MD Peking University Cancer Hospital & Institute, Beijing, China Jun Guo

Replace, or complement? The answer differs by pathway BRCA1/2 + HRRm [Venn: Tissue NGS / Plasma ctDNA] 11 tissue only | 46 both | 4 plasma only n = 61 altered patients · circle and overlap areas proportional to patient numbers why discordant 11 tissue-only: [bar] 7 | 4 4 plasma-only: [bar] 2 | 1 | 1 Legend: not evaluable — ctDNA fraction below the class floor; assessable, still discordant — biology; technical — coverage / SNP filter or annotation; no explanation identified Plasma alone identifies 50 of 61 altered patients (82%); as an adjunct it adds 4 patients tissue missed (+7%). Plasma-first is defensible for PARP-inhibitor eligibility. PI3K / AKT [Venn: Tissue NGS / Plasma ctDNA] 55 tissue only | 20 both | 13 plasma only n = 88 altered patients · circle and overlap areas proportional to patient numbers why discordant 55 tissue-only: [bar] 42 | 13 13 plasma-only: [bar] 9 | 4 Plasma alone identifies 33 of 88 altered patients (38%). Plasma-first is not defensible; PI3K/AKT stays tissue-anchored. Niven Mehra

[Slide 1] Resected RCC at intermediate-high risk of recurrence, or completely resected metastatic disease Randomisation Allocation ratio 3:2:2 N=790 Arm A active monitoring (AM) N=340 Arm B durvalumab (D) 1500mg q4W for 1 year N=225 Arm C durvalumab 1500mg q4w for 1 year and tremelimumab (DT) 75mg at day 1 and week 4 N=225 1. Larkin J et al. RAMPART: first results. ESMO 2025. 2. Larkin J et al. Durvalumab monotherapy. ASCO 2026. 3. Merrick S et al. Patient reported outcomes. GU ASCO 2026. [Slide 2] OHQL & Functioning Primary QoL outcome: Change in OHQL from baseline to Month 15 showed no evidence of a difference between arms. Week 16 Adjusted mean difference from baseline** (durvalumab - AM) Durvalumab better Durvalumab worse -7.4* OHQL Physical functioning -7.1* Role functioning Social functioning Emotional functioning -5.3* Cognitive functioning Month 15 OHQL Physical functioning Role functioning Social functioning Emotional functioning -7.1* Cognitive functioning Higher scores = better OHQL/functioning with durvalumab. * Statistically significant and exceeded predefined domain-specific CMD thresholds: OHQL ≥4; Role function ≥6; Cognitive function ≥3. ** Adjusted for baseline QoL score and sex; all analyses were exploratory. [Slide 3] Symptoms & Financial Difficulties Primary QoL outcome: Change in OHQL from baseline to Month 15 showed no evidence of a difference between arms. Week 16 Adjusted mean difference from baseline** (Durvalumab - AM) Durvalumab worse Durvalumab better +5.5* Fatigue Nausea and vomiting Pain Dyspnoea Insomnia +5.8* Appetite loss Constipation Diarrhoea Financial difficulties Month 15 Fatigue Nausea and vomiting Pain +7.5* Dyspnoea Insomnia +7.4* Appetite loss +12.2* Constipation Diarrhoea Financial difficulties Higher scores = worse symptoms/financial difficulties with Durvalumab. * Statistically significant and exceeded predefined domain-specific CMD thresholds: Appetite loss ≥5; Dyspnoea ≥4; Constipation ≥5. ** Adjusted for baseline QoL score and sex; all analyses were exploratory.

[Slide 2] Neoantigen vaccination is highly immunogenic in RCC Heatmap: Patient ID 101, 102, 104, 105, 106, 107, 108, 109, 110; Vaccination timepoints Week 0, 3, 8, 12, 16, 20, 24; Vaccination pool A-D; Immunizing peptides Max values printed: 101: 1051; 102: 201; 104: 88; 105: 95; 106: 793; 107: 137; 108: 477; 109: 570; 110: 487 Normalized IFNγ ELISpot (Min per patient response to Max per patient response) Week 16: Immunogenic / Not immunogenic; No. of peptides per pool Ex vivo T cell response in all patients Patient 101; Vaccine pool: A, B, C, D; IFNγ spot-forming cells (per 10^6 PBMCs) vs Time (weeks) Deconvolution of immunogenic neoantigens 101; DMSO (no-restimulation); Pool A; Pool B; Pool C; Pool D IFNγ spot-forming cells (per 3×10^4 PBMCs); Mutation encoded by vaccine peptide Braun et al, Nature, 2025. [Slide 3] Driver mutations elicit strong immune resposnes Strong response against VHLmut peptide Patient 101: VHLmut - T cells reacting against VHL neoAg; DMSO - Negative control; HIV peptide - Negative control; PHA - Positive control (non-specifically activate T cells) Immune responses against common RCC driver mutations No. of patients vaccinated against driver mutations: VHL, PBRM1, BAP1, KDM5C, PIK3CA (Immunogenic / Not immunogenic) No. of peptides: Driver mutations, Passenger mutations (Immunogenic / Not immunogenic) [Slide 4] Vaccination induces durable T cell expansion and anti-tumor immunity Durable expansion of vaccine-reactive T cells Prime, Boost, Boost; % TCR β-chain vs Weeks after first vaccination (0-160); n = 9 patients Viral | Tumor | Vaccine Vaccine-reactive T cells can recognize tumor cells Pt. 101 week 16 PBMCs, in vitro expanded against PIK3CAmut peptide: Autologous whole tumor cells; DMSO Pt. 109 week 16 PBMCs, in vitro expanded against PBRM1mut peptide: Autologous whole tumor cells; DMSO IFNγ ELIspot Test vaccine-reactive T cells against patient's own tumor Patient ID 101-110; No. of vaccine peptide pools generating anti-tumor reactivity (Anti-tumor reactivity / No anti-tumor reactivity) [Br]aun et al, Nature, 2025.

[Slide 1] Incidence and patterns of co-occurrence of targetable genomic alterations in metastatic Prostate Cancer (mPC): Implications for personalized treatment stratification FINAL PUBLICATION NUMBER FPN: 2227P RESULTS - TARGETABLE PROFILES: 72% of evaluable pts (n=215) harbored a potentially targetable profile RESULTS - CO-OCCURRENCE: 92%; 33% [remaining poster text partly obscured by people; full poster transcribed in 2114385655375728819_s2] Adjacent poster 2228P: A phase 1 study of SHR-4394, a PSMA-targe[ted ...] [illegible] [Slide 2] Incidence and patterns of co-occurrence of targetable genomic alterations in metastatic Prostate Cancer (mPC): Implications for personalized treatment stratification FINAL PUBLICATION NUMBER FPN: 2227P Arduini D¹, Ciccarese C¹,², Occhipinti D², Belletto R², Neri A², Colasanti F², Di Maio L³, Dragonetti P³, Duranti S⁵, Nero C⁴, Giacò L⁷, Marino I⁵, Piermattei A⁵, Panfili A⁵, Rocco BCM⁸, Minucci A⁶, Tagliaferri L³, Gasbarrini A², Tortora G¹,², Iacovelli R¹,² ¹Medical Oncology, Fondazione Policlinico Universitario A. Gemelli IRCCS, Rome; ²Università Cattolica del Sacro Cuore, Rome; ³Diagnostic Imaging, Radiotherapy Oncology and Hematology, Gemelli IRCCS, Rome; ⁴Gynecologic Oncology Unit, Gemelli IRCCS and Catholic University, Rome; ⁵Scientific Directorate, Gemelli IRCCS, Rome; ⁶Molecular and Genomic Diagnostics, Genomics Core Facility, G-STeP, Gemelli IRCCS, Rome; ⁷Bioinformatics Research Core Facility, G-STeP, Gemelli IRCCS, Rome; ⁸Urology, Gemelli IRCCS, Rome, Italy BACKGROUND In mPC, genomic alterations (GenAlt) may act as predictive biomarkers for treatment selection. HRR gene alterations identify candidates for PARP inhibitors (PARPi); activation of the PI3K/AKT pathway may support AKT inhibitors (capivasertib). Two among TP53, RB1, PTEN alterations are associated with aggressive variant PC (AVPC) and the greatest benefit from chemotherapy. SPOP mutations may predict sensitivity to androgen receptor-targeting agents, whereas the role of TMPRSS2-ERG remains controversial. Aim: to challenge the single-driver paradigm in mPC by defining co-occurring molecular profiles to guide treatment strategies in the precision-medicine era. METHODS 342 mPC pts enrolled in the FPG500 program (NCT06020625) 127 (37.1%) excluded: insufficient / damaged biopsy 215 pts with evaluable CGP (TSO500 HT assay) Pts were stratified into potentially therapeutic profiles: PARPi - ≥1 HRR gene alteration AVPC - ≥2 alterations among PTEN, TP53, RB1 AKT-inhibitor - PTEN and/or PIK3CA alterations Hormone-sensitive - SPOP TMPRSS2-ERG - TMPRSS2-ERG fusion RESULTS - TARGETABLE PROFILES 72% of evaluable pts (n=215) harbored a potentially targetable profile Prevalence of each profile (% of evaluable pts) TMPRSS2-ERG | 26% AKT-inhibitor | 22.8% PARPi | 20% SPOP | 11.6% AVPC | 6% TMPRSS2-ERG was the most frequent profile, followed by the AKT-inhibitor and PARPi profiles. AVPC was identified in 6% (n=13) of pts. SPOP (11.6%) showed strict mutual exclusivity with TMPRSS2-ERG - two distinct, non-overlapping molecular groups. RESULTS - CO-OCCURRENCE Overlap of the AVPC profile (n=13) with other profiles AKT-inhibitor | 92% TMPRSS2-ERG | 54% 92% of AVPC pts also qualified for the AKT-inhibitor profile and 54% overlapped with the TMPRSS2-ERG profile. 33% overall co-occurrence among the other groups, mainly TP53 with PTEN or TMPRSS2 CONCLUSIONS • Comprehensive genomic profiling identified clinically targetable alterations in the majority of mPC pts (72%). • The high rate of genomic co-occurrence challenges the single-driver paradigm. • Findings support the potential role of routine molecular profiling to optimize treatment selection in mPC. Abbreviations: mPC, metastatic prostate cancer; CGP, comprehensive genomic profiling; HRR, homologous recombination repair; PARPi, PARP inhibitor; AVPC, aggressive variant prostate cancer. DISCLOSURE (presenting author) "No conflicts of interest to declare" FPG500 program, NCT06020625 CONTACT: daniela.arduini@guest.policlinicogemelli.it DanielaArduini3

[Slide 1] omHSPC · 1–5 metastases · ≥1 bone lesion · ECOG 0–2 · all lesions amenable to ablative RT 565 RANDOMIZED 1:1 2020–2025 ARM A: SOC* + MDT, n = 282, SBRT to all oligometastatic lesions ARM B: SOC*, n = 283, Standard-of-care treatment* TREATED n=279 SBRT n=245 | TREATED n=282 STRATIFIED BY METASTATIC BURDEN: 1–2 vs 3–5 SYSTEMIC INTENSIFICATION: None · docetaxel · NHA DISEASE PRESENTATION: De novo vs recurrent [Slide 2] | Characteristic | Overall N=565 | Arm A (SOC+MDT) N=282 | Arm B (SOC) N=283 | Age, years — median (IQR) | 71 (65–75) | 71 (65–75) | 72 (66–76) ECOG performance status 0 | 486 (86.0%) | 241 (85.5%) | 245 (86.6%) Number of metastases 1–2 | 421 (74.5%) | 211 (74.8%) | 210 (74.2%) 3–5 | 144 (25.5%) | 71 (25.2%) | 73 (25.8%) Disease presentation De novo | 289 (51.2%) | 145 (51.4%) | 144 (50.9%) Recurrent | 276 (48.8%) | 137 (48.6%) | 139 (49.1%) Metastatic sites Bone | 558 (98.8%) | 279 (98.9%) | 279 (98.6%) Lymph nodes | 209 (37.0%) | 104 (36.9%) | 105 (37.1%) Staging imaging Choline PET/CT | 354 (62.7%) | 178 (63.1%) | 176 (62.2%) PSMA PET/CT | 211 (37.3%) | 104 (36.9%) | 107 (37.8%) Additional systemic treatment None | 85 (15.0%) | 42 (14.9%) | 43 (15.2%) Docetaxel | 2 (0.4%) | 1 (0.4%) | 1 (0.4%) New-generation hormonal therapy | 478 (84.6%) | 239 (84.8%) | 239 (84.5%) Baseline PSA, ng/mL — median (IQR) | 6 (1–22) | 7 (2–22) | 6 (1–22) [Slide 3] Acute AEs within 6 months by randomized arm, n (%) | Maximum grade | ARM A (SOC + MDT) n=279 | ARM B (SOC) n=282 | Any acute AE | 253 (90.7) | 277 (98.2) Grade 1 | 92 (33.0) | 108 (38.3) Grade 2 | 115 (41.2) | 126 (44.7) Grade 3 | 39 (14.0) | 37 (13.1) Grade 4 | 3 (1.1) | 4 (1.4) Grade 5 | 4 (1.4) | 2 (0.7) [Slide 4] Brief Pain Inventory Scores in patients with bone pain at baseline [Left panel: Brief Pain interference score; BPI Score 0–10 by Study Visit; SOC + MDT (blue) vs SOC (red)] Number of patients at visit (Baseline / Week 6 / Month 3 / Month 6) SOC + MDT 90 / 84 / 87 / 101 SOC 93 / 73 / 80 / 88 [Right panel: Brief Pain severity score; BPI Score 0–10 by Study Visit; SOC + MDT (blue) vs SOC (red)] Number of patients at visit (Baseline / Week 6 / Month 3 / Month 6) SOC + MDT 93 / 83 / 89 / 100 SOC 94 / 75 / 79 / 84

[Slide 1] Week: 0 12 24 48 Patients with mCRPC nmCRPC or mHSPC N=111 RANDOMIZE Darolutamide -> Meets crossover criteria (Y -> Enzalutamide; N -> Darolutamide) -> Meets crossover criteria (Y -> Enzalutamide; N -> Darolutamide) Enzalutamide -> Meets crossover criteria (N -> Enzalutamide; Y -> Darolutamide) -> Meets crossover criteria (N -> Enzalutamide; Y -> Darolutamide) Week 0: CANTAB modules, Blood (PHS, AR), PROMs, TUG Week 12: CANTAB modules, PROMs, TUG Week 24: CANTAB modules, PROMs, TUG Week 48: CANTAB modules, PROMs, TUG Stratification: Age (< 65, 65-80, > 80) [Slide 2] | Darolutamide (N=38) | Enzalutamide (N=41) | Total (N=79) Age Median (IQR) | 70.2 (66.4, 76.0) | 69.5 (66.4, 74.2) | 70.1 (66.4, 76.0) Range | 58.9, 86.2 | 57.7, 84.5 | 57.7, 86.2 Age Category, n (%) <65 | 8 (21.1%) | 9 (22.0%) | 17 (21.5%) 65-80 | 26 (68.4%) | 29 (70.7%) | 55 (69.6%) >80 | 4 (10.5%) | 3 (7.3%) | 7 (8.9%) Gender, n (%) Male | 38 (100.0%) | 41 (100.0%) | 79 (100.0%) Race, n (%) American Indian or Alaska Native | 0 (0.0%) | 1 (2.4%) | 1 (1.3%) Black or African American/African Heritage | 0 (0.0%) | 4 (9.8%) | 4 (5.1%) White | 36 (94.7%) | 33 (80.5%) | 69 (87.3%) Not reported | 1 (2.6%) | 3 (7.3%) | 4 (5.1%) Unknown | 1 (2.6%) | 0 (0.0%) | 1 (1.3%) Ethnicity, n (%) Hispanic or Latino | 1 (2.6%) | 1 (2.4%) | 2 (2.5%) Not Hispanic or Latino | 35 (92.1%) | 38 (92.7%) | 73 (92.4%) Not reported | 1 (2.6%) | 2 (4.9%) | 3 (3.8%) Unknown | 1 (2.6%) | 0 (0.0%) | 1 (1.3%) Level of education, n (%) Left formal education at age 17-18 | 10 (26.3%) | 5 (12.2%) | 15 (19.0%) Undergraduate degree or equivalent | 18 (47.4%) | 30 (73.2%) | 48 (60.8%) Master's degree or equivalent | 4 (10.5%) | 3 (7.3%) | 7 (8.9%) PhD or equivalent | 2 (5.3%) | 1 (2.4%) | 3 (3.8%) Missing | 4 (10.5%) | 2 (4.9%) | 6 (7.6%) Known family history of dementia, n (%) Yes | 4 (10.5%) | 5 (12.2%) | 9 (11.4%) No | 34 (89.5%) | 36 (87.8%) | 70 (88.6%) [Slide 3] Figure 4: Distribution of PHS among Evaluable patients (N=79) [Horizontal histogram of PHS (y-axis 'phs', bins -0.75 to 2.75 in 0.5 steps; x-axis count 0-30) with adjacent box plot] [Slide 4] Table 5. Change in FACT-Cog at 24 Weeks MCCD at 24 Weeks by PHS Quartile % Change from Baseline | Quartile 1 (N=9) | Quartile 2 (N=9) | Quartile 3 (N=7) | Quartile 4 (N=9) | Total (N=34) MCCD | DAROLUTAMIDE Median | -14.3 | -16.7 | -22.2 | -33.3 | -17.4 (IQR) | (-20.0, -5.3) | (-37.5, 0.0) | (-40.0, -7.1) | (-44.4, -12.4) | (-40.0, -5.3) Range | -122.2, 5.4 | -92.3, 12.6 | -90.0, 7.0 | -240.0, 6.1 | -240.0, 12.6 MCCD | ENZALUTAMIDE Median | -33.5 | -50 | -33.3 | -31.4 | -34.7 (IQR) | (-266.7, -24.9) | (-73.3, -25.0) | (-75.0, -10.5) | (-36.8, -9.4) | (-73.3, -16.7) Range | -1000.0, -16.7 | -120.0, -0.2 | -500.0, 1.1 | -50.0, 0.0 | -1000.0, 1.1

[Slide 2] Cell therapy offers another selective way to deploy cytokines • Cytokines shape therapeutic product during manufacturing (IL-2, IL-7, IL-15, IL-21) • "Armored" cell therapy products extend cytokine support in vivo beyond cell infusion 1 Tethered Retain cytokines at the cell surface Membrane tether Engineered T cell Cytokine receptor CELL-ASSOCIATED EXPOSURE 2 Constitutively secreted Release cytokines from the cell product Engineered T cell Cytokine receptor LOCAL CYTOKINE SOURCE 3 Antigen-induced Couple production to antigen sensing Target cell Antigen Cytokine gene induced Engineered T cell Cytokine receptor CAR-NFAT / synNotch ANTIGEN-LINKED EXPOSURE [Slide 4] Engineering cytokines—and the T cells that receive them 1. Engineering expands how cytokines can be delivered and controlled. Exposure, cell selectivity, and potency are complementary design opportunities. 2. Cytokine-armored T cells show clinical promise—but how do we choose the right cytokine signal? Synthetic receptors enable systematic testing of alternative cytokine programs and nominating candidates for translation. 3. Repurposing innate cytokine signals expands the possibilities for T-cell therapy.. IL-9R, GCSFR and IL-22R signaling support distinct functions and improved antitumor activity in preclinical models. The opportunity is not only to control which cells receive cytokines—but to choose what those signals instruct therapeutic T cells to do.

THE PATH FORWARD RETAIN-3: Clinical and Molecular Response Guide De-escalation Eligibility: cT2–T3 N0, predominant urothelial carcinoma, ECOG 0–1 TURBT EV + P x3 cycles Restaging Cystoscopy + biopsies CT chest/abdomen Pelvic MRI Urine cytology ctDNA, real time seeT0 + ctDNA− seeT0 + ctDNA+ No cCR Active surveillance Maintenance pembrolizumab q6 wk x7 ctDNA clearance EV + P x3 then repeat ctDNA Persistent ctDNA+ Definitive local therapy Cystectomy or chemoradiation Proposed: N=120; primary endpoint 2-year MFS. Clinical and molecular response guide local and systemic de-escalation. EV, enfortumab vedotin; P, pembrolizumab. Slide adapted from P. Ghatalia, MD Alexander Kutikov, MD, FACS @uretericbud Fox Chase Cancer Center Temple Health The Urologic Institute

4505P KIM-1 dynamics and baseline CXCR3+ CD8+ T cells are associated with outcomes of ICI-based therapy in metastatic renal cell carcinoma Kohji Takemura, Hirotake Tsukamoto, Takanobu Motoshima, Shinichiro Watanabe, Shoichiro Kuroda, Masaya Baba, Yuki Sugiura, Kazushige Sakaguchi, Shinji Urakami, Yuji Miura 1) Department of Advanced Medical Development, Cancer Institute Hospital of JFCR. 2) Division of Clinical Immunology and Cancer Immunotherapy, Center for Cancer Immunotherapy and Immunobiology, Kyoto University. 3) Department of Urology, Saiseikai Kumamoto Hospital. 4) Department of Urology, Kumamoto University. 5) Multiomics platform, Center for Cancer Immunotherapy and Immunobiology, Kyoto University. 6) Department of Urology, Toranomon Hospital. 7) Department of Medical Oncology, Cancer Institute Hospital of JFCR. Contact E-mail: kohji.takemura@jfcr.or.jp Background KIM-1 has emerged as a promising biomarker for RCC. Although current evidence from recent phase III trials supports the prognostic value of circulating KIM-1 in mRCC, the biological role of KIM-1 remains poorly understood. We investigated the prognostic value of plasma KIM-1 in patients with metastatic RCC treated with immune-based therapies and explored its association with peripheral immune cell phenotypes. Methods [Schema: Patients with mRCC → Baseline plasma/PBMC (C1D1) → ICI-based therapy (ICI mono/dual or ICI-TKI) → On-treatment plasma/PBMC (C2D1) → Plasma KIM-1, T cell subsets in PBMC → PFS/OS] Study design: Multicenter retrospective study in Japan Participants: Patients who initiated ICI-based therapy for mRCC between October 2018 and July 2024 Treatment regimens: ICI mono/dual: Nivolumab / Ipilimumab + Nivolumab; ICI-TKI: ICI + tyrosine kinase inhibitor (TKI) combinations Measurements: Plasma KIM-1 and T-cell subsets in peripheral blood mononuclear cells (PBMC) were evaluated by ELISA and flow cytometry, respectively. KIM-1 was measured at baseline and on-treatment (mainly at the second ICI dose). Endpoints: Associations in plasma KIM-1 levels and T-cell subsets in PBMC of baseline and on-treatment samples with clinical outcomes were evaluated. Analyses were performed using available data. Results Table 1. Baseline characteristics (Total n=57) Sex Male 47 (82.5%) Age (median, range) 69 (28-90) IMDC risk category: Favorable 14 (24.6%); Intermediate 31 (54.4%); Poor 11 (19.3%) Histology: ccRCC 40 (70.2%); nccRCC 14 (24.6%) Prior nephrectomy 42 (73.7%) Treatment category: ICI mono/dual 36 (63.2%); - Ipilimumab + Nivolumab 25 (43.9%); - Nivolumab monotherapy 11 (19.3%); ICI-TKI 21 (36.8%) Table 2. Association between baseline KIM-1 and T-cell subsets in PBMC (class / subset / n / Spearman ρ / P value) CD8 CD3+CD8+ 54 −0.275 0.0440 CD3+CD8+CXCR3+ 54 −0.457 0.00051 CD3+CD8+ICOS+ 54 −0.363 0.00694 CD4 CD3+CD4+ 54 −0.397 0.00297 Th1; CD3+CD4+CXCR3+ 54 −0.350 0.00940 Th17; CD3+CD4+CCR6+CXCR3− 54 −0.295 0.0304 Tfh; CD3+CD4+CXCR5+ 54 −0.272 0.0467 CD4+CCR6+CCR4+ 39 −0.319 0.0476 CD3+CD4+ICOS+ 54 −0.377 0.00490 Only subsets with nominal p < 0.05 are shown. P values are unadjusted for multiple comparisons. Figure 1. Association between KIM-1 dynamics (delta KIM-1), baseline CXCR3+CD8+ T cells and clinical outcomes. [KM panels: Delta KIM-1 / PFS (KIM-1 Decrease vs KIM-1 Increase, HR 1.82 [illegible CI] p=0.08); Delta KIM-1 / OS (HR 3.98 [illegible CI] p=0.005); Baseline CXCR3+CD8+ T cell / PFS; Baseline CXCR3+CD8+ T cell / OS (High vs Low)] Figure 2. Association between KIM-1 dynamics and clinical outcomes stratified by baseline CXCR3+CD8+ T cells in PBMC. [KM panels PFS and OS; groups KIM-1 Decrease / CXCR3+CD8+ T cells High, KIM-1 Decrease / Low, KIM-1 Increase / High, KIM-1 Increase / Low; HR values illegible] Discussion ✓ A decrease in circulating KIM-1 during ICI-based therapy was significantly associated with longer OS, whereas the association with PFS did not reach significance. ✓ Higher baseline KIM-1 levels correlated with lower proportions of several peripheral T-cell subsets. The strongest unadjusted correlation was observed for CXCR3+CD8+ T cells, suggesting an association between KIM-1 and peripheral immune composition. ✓ Combining KIM-1 dynamics with baseline CXCR3+CD8+ T-cell proportions may help stratify patients by prognosis. This exploratory finding requires validation, as formal interaction tests were not significant and incremental prognostic value was not established. ✓ The retrospective design, small sample size, multiple comparisons, and variation in on-treatment sampling times limit interpretation. Take Home Message 1. Lower baseline KIM-1 levels are associated with higher frequencies of T-cell subsets implicated in antitumor immunity, particularly CXCR3+CD8+ T cells. 2. Combining KIM-1 dynamics with baseline CXCR3+CD8+ T cell status may improve prognostic stratification, warranting further validation. [References and conflict of interest text small]

[Slide 1] Final Protocol-Prespecified Analysis of the Phase 3 KEYNOTE-564 Trial of Adjuvant Pembrolizumab Versus Placebo in Clear Cell Renal Cell Carcinoma Toni K. Choueiri, Piotr Tomczak, Se Hoon Park, Balaji Venugopal, Thomas Ferguson, Naomi B. Haas, Yen-Hwa Chang, Stefan N. Symeonides, Jae Lyun Lee, Piotr Sawrycki, Naveed Sarwar, Mauricio Mahave, Howard Gurney, Marine Gross-Goupil, Gurjyot K. Doshi, John M. Burke, Rui Kang, Joseph E. Burgents, Priyanka Chablani, Thomas Powles [affiliations: small print, partly illegible] [Slide 2] OS, ITT Population Pembro: 96.3% | 93.9% | 91.2% | 87.7% | 85.9% | 82.3% Placebo: 93.9% | 89.5% | 85.8% | 82.2% | 79.3% | 76.6% | Pembro (N = 496) | Placebo (N = 498) Events, n (%) | 86 (17.3) | 118 (23.7) Median (95% CI), months | NR (NR-NR) | NR (NR-NR) HR 0.70 (95% CI, 0.53-0.92) Months No. at risk Pembro 496 489 485 479 472 468 460 452 445 440 428 422 417 383 227 131 49 1 0 Placebo 498 492 484 478 464 455 442 429 423 416 404 398 390 358 204 119 45 2 0 [...] end point was met at IA3 and was not formally statistically tested thereafter. Data cutoff date: February 4, 2026. Survival table Endpoint: Overall survival (ITT), final analysis Arm | N | Median | Median CI Pembro | 496 | NR | NR-NR Placebo | 498 | NR | NR-NR HR: 0.70 HR CI: 0.53-0.92 Landmark rates: dashed-line landmarks 1-6 (timepoints not labeled on slide): Pembro 96.3% / 93.9% / 91.2% / 87.7% / 85.9% / 82.3% same: Placebo 93.9% / 89.5% / 85.8% / 82.2% / 79.3% / 76.6%

[Slide 1] Concordance is high for HRR — PI3K/AKT biallelic loss is the outlier | SNV / indel | Biallelic loss BRCA1/2 | PPV 82%; PPA 93% (14/15) · 95% CI 70–99% | PPV 100%; PPA 50% (3/6) · 95% CI 19–81% HRR (non-BRCA) | PPV 97%; PPA 81% (29/36) · 95% CI 65–90% | n/a; n = 1 — not evaluable PI3K / AKT | PPV 65%; PPA 57% (13/23) · 95% CI 37–74% | PPV 50%; PPA 14% (8/58) · 95% CI 7–25% Accuracy fails predominantly for detection of copy-number loss in PI3K/AKT — PTEN-dominated Tissue as comparator, n = 301. PPA = positive percent agreement (plasma-positive among tissue-positive), Wilson 95% CI; PPV = positive predictive value (tissue-positive among plasma-positive). Tissue is a comparator. Updated after re-review of 10 PTEN homozygous-deletion calls since abstract submission; patients carrying both a BRCA and a non-BRCA HRR alteration (n = 3) are assigned to the BRCA row. Niven Mehra [Slide 3] Not all discordance is technical — and only plasma-detected PTEN loss appears prognostic Legend: Neither n=136, mOS 32 mo; Tissue-only n=58, mOS 29 mo; ctDNA-only n=23, mOS 11 mo; Both n=28, mOS 17 mo Overall survival; Months from first blood draw No. at risk (0, 12, 24, 36, 48, 60): Neither: 136, 121, 91, 59, 45, 19 Tissue-only: 58, 51, 36, 22, 16, 7 ctDNA-only: 23, 9, 3, 1, 0, 0 Both: 28, 21, 6, 4, 2, 1 Liquid NGS misses PTEN deletions, not PTEN mutations Tukachinsky, Clin Cancer Res 2021 · Chaki, Breast Cancer Res 2025 · Spatial and temporal heterogeneity of PI3K (PTEN loss) Ferraldeschi, Eur Urol 2015 · Punnoose & Ferraldeschi, Br J Cancer 2015 · Warner, Nature Cancer 2024 · PTEN loss found only in tissue is not prognostic 58 of 245 patients are tissue-only, and their survival is comparable to wild-type — 29 vs 32 months, HR 1.16 (0.81–1.66), p = 0.43. When PTEN loss is visible in plasma, median survival is 11–17 months: HR 3.50, and 2.11 (1.38–3.24) after adjusting for ctDNA fraction. Niven Mehra Survival table Endpoint: Overall survival from first blood draw, by PTEN-loss detection (mCRPC) Neither | 136 | 32 mo Tissue-only | 58 | 29 mo ctDNA-only | 23 | 11 mo Both | 28 | 17 mo HR: 1.16 (tissue-only vs wild-type) HR CI: 0.81–1.66 P: 0.43 [Slide 4] Conclusions and take-home message 1 BRCA1/2 and HRR Plasma is a credible standalone test for PARP-inhibitor eligibility when tumour fraction is adequate — 82% of all altered patients, and 100% of tissue-positive SNVs above 5% tumour fraction. As an adjunct it adds 7% over tissue alone. 2 PI3K / AKT Neither compartment is a gold standard. PTEN loss is spatially heterogeneous and discordance runs in both directions — relevant if the capivasertib biomarker moves from IHC to NGS. 3 Reporting standard A plasma result below the tumour-fraction threshold should be reported as not informative (not as negative). Matched leukocyte DNA is essential not optional — 10 apparent HRR/PI3K findings were clonal haematopoiesis. Proposed pathway for BRCA1/2 and HRR: plasma first at APMR -> reflex to tissue if tumour fraction is below threshold Niven Mehra

[Slide 1] Mechanism of action of ABBV-969 Top1i Dual binding: STEAP1 binding | PSMA binding STEAP1 | COOH | NH2 | PSMA Tumor cell -> Cell death PSMA, prostate-specific membrane antigen; STEAP1, six-transmembrane epithelial antigen of the prostate 1; Top1i, topoisomerase 1 inhibitor. [Slide 2] ABBV-969 achieved deep, durable PSA responses in patients with prior taxane and radioligand therapy exposure, bone/soft-tissue lesions, and diverse deleterious genomic alterations [Waterfall plot] Y-axis: PSA Change from Baseline (%), 30 to -100; dashed lines at -50, -70, -90 Annotation rows: PSMA IHC | STEAP1 IHC | AR Alteration | TP53 Alteration | RB1 Alteration | PTEN Alteration | BCOR Dose: 1 mg/kg; 2 mg/kg; 3 mg/kg; 5 mg/kg; 8 mg/kg; 10 mg/kg; 12.5 mg/kg PSMA IHC: [0,100); [100,200); [200,300]; NA STEAP1 IHC: [0,100); [100,200); [200,300]; NA AR: AR-Altered; AR-Non-Altered TP53: TP53-Altered; TP53-Non-Altered RB1: RB1-Altered; RB1-Non-Altered PTEN: PTEN-Altered; PTEN-Non-Altered BCOR: CR; PR; SD; NA Data cutoff: April 2, 2026. * Patients with measurable liver lesions. AR, androgen receptor; BCOR, best confirmed overall response; CR, complete response; IHC, immunohistochemistry; NA, not available; PR, partial response; PSA, prostate-specific antigen; PSMA, prostate-specific membrane antigen; PTEN, phosphatase and tensin homolog; RB1, retinoblastoma 1; SD, stable disease; STEAP1, six-transmembrane epithelial antigen of the prostate 1; TP53, tumor protein p53. [Slide 3] ABBV-969 triggered rapid and deep CTC reduction A. Abundant CTC shedding was observed at baseline n = 34; Median: 5.44; y-axis: log2(Baseline Total CTC Count + 1); All Doses Median (IQR) CTC at baseline: 43 (11, 120) B. Rapid and deep CTC reduction Kruskal–Wallis Test P-value <0.001 Week 1 (C1D1): n = 34, Median: 0 | Week 4 (C2D1): n = 20, Median: -89.97 | Week 7 (C3D1): n = 30, Median: -99.79; P<0.001 y-axis: Percentage Change in Total CTC; x-axis: Visit CTC conversion rate: • From ≥5 at baseline to <5 at C3D1: 78% • 100% clearance at C3D1: 47% C, cycle; CTC, circulating tumor cell; D, day; IQR, interquartile range. PSA50 responses were observed in patients with PSMA-negative mAPMR prostate cancer (mCRPC) at baseline (per VISION criteria)(8) | No PSA50 Response | PSA50 Response | Total | Percent Yes PSMA Negative Per VISION | 0 | 2 | 2 | 100 PSMA Positive Per VISION | 9 | 18 | 27 | 67 mAPMR, metastatic androgen pathway modulation resistant; PSA50, ≥50% decline in prostate specific antigen from baseline; PSMA, prostate-specific membrane antigen; mCRPC, metastatic castration resistant prostate cancer. • Two of 29 patients with baseline PSMA PET from ≥3 mg/kg doses were PSMA PET negative per VISION study criteria at baseline • Both achieved PSA50 with stable disease and confirmed partial response respectively per RECIST v1.1 [Slide 4] ABBV-969 triggered deep and rapid ctDNA reduction; reduction at Cycle 3, Day 1 was associated with radiographic progression-free survival (rPFS) A. Abundant ctDNA shedding was observed at baseline n = 49; Median: 0.29; y-axis: Baseline ctDNA Fraction; All Doses Median (IQR) baseline ctDNA fraction: 0.29 (0.05, 0.61) B. ABBV-969 triggered rapid and deep ctDNA reduction Kruskal-Wallis Test P-value <0.001 Week 1 (C1D1): n = 49, Median: 0 | Week 4 (C2D1): n = 19, Median: -88.56 | Week 7 (C3D1): n = 45, Median: -95.82 y-axis: ctDNA Percentage Change; x-axis: Visit ctDNA reduction rate at C3D1 (n=113 [small text; as rendered]): • ≥50%: 82% • ≥95%: 58% C. Deep ctDNA reduction at C3D1 was associated with radiographic PFS Legend: ≥95% Reduction; <95% Reduction y-axis: Survival Probability; x-axis: Time (Months) 0, 5, 10, 15, 20 Number at risk: ≥95% Reduction: 23, 22, 11, 6, 0 <95% Reduction: 19, 9, 3, 2, 0 rPFS | Overall (n=42) | <95% reduction (n=19) | ≥95% reduction (n=23) Median rPFS, months | 11.9 | 8.3 | NR HR (95% CI): 3.7 (1.38–9.93); log-rank P-value: 0.0055 Survival table Endpoint: rPFS by ctDNA reduction at C3D1 Overall | 42 | 11.9 months <95% reduction | 19 | 8.3 months ≥95% reduction | 23 | NR HR: 3.7 HR CI: 1.38–9.93 P: 0.0055 (log-rank)

RENAL CANCER 3638TiP FPN: 3638TiP RAPSTAR: randomized phase III trial comparing robotic-assisted partial nephrectomy (RAPN) versus stereotactic ablative radiotherapy (SABR) in localized renal cell carcinoma Bartłomiej Tomasik¹, Michał Czarnogórski², Krzysztof Konopa¹, Jan Adamowicz², Paweł Teterycz³,⁴, Krzysztof Koper⁵, Rafał Dziadziuszko¹, Tomasz Drewa² ¹Department of Oncology and Radiotherapy, Faculty of Medicine, Medical University of Gdańsk, Gdańsk, Poland; ²Department and Chair of Urology and Andrology, Ludwik Rydygier Collegium Medicum, Nicolaus Copernicus University, Bydgoszcz, Poland; ³Digital Medicine Center, Maria Sklodowska-Curie National Research Institute of Oncology, Warsaw, Poland; ⁴Department of Soft Tissue/Bone Sarcoma and Melanoma, Maria Sklodowska-Curie National Research Institute of Oncology, Warsaw, Poland; ⁵Department of Oncology, Ludwik Rydygier's Collegium Medicum in Bydgoszcz, Nicolaus Copernicus University in Toruń, Poland. [Logos: Medical University of Gdańsk; Nicolaus Copernicus University in Toruń; Medical Research Agency] INTRODUCTION Partial nephrectomy, performed as robotic-assisted partial nephrectomy (RAPN), is the standard treatment for localized renal cell carcinoma (RCC), providing excellent outcomes while preserving renal function (1). The phase II FASTRACK II trial demonstrated promising local control and acceptable toxicity with stereotactic ablative radiotherapy (SABR) (2). However, randomized evidence comparing SABR with surgery in operable patients is lacking. The RAPSTAR trial aims to address this evidence gap by comparing RAPN with SABR, with a primary focus on renal function preservation. Beyond renal function, the trial will also assess oncological outcomes, treatment-related toxicity, quality of life, and cost-effectiveness. METHODS RAPSTAR is a multicenter, randomized, open-label phase III trial comparing RAPN and SABR (1:1) in patients with localized RCC eligible for surgery and radiotherapy. A total of 164 patients (82 per arm) will be enrolled over 4 years, with post-treatment follow-up at 6, 12, 18, and 24 months. The sample size was calculated using the approach described by Borm et al. (3), assuming a clinically relevant between-group difference of 5 mL/min/1.73 m² in eGFR, a standard deviation of 15 mL/min/1.73 m², R² ≥ 0.5, a two-sided α of 0.05, and 80% power. The calculated sample size of 142 evaluable patients was increased to 164 to account for potential attrition. The primary analysis is planned in the intention-to-treat population, with a supportive per-protocol analysis. Inclusion criteria: • Age ≥18 years • Biopsy-confirmed diagnosis of RCC with no more than one lesion in one kidney (bilateral RCC allowed) • Stage I disease (tumour ≤7 cm in diameter) • ECOG performance status ≤2 • Eligible for both surgery and radiotherapy • Provision of written informed consent • Multidisciplinary decision justifying the need for active local treatment N = 164, 1:1 randomisation, stratification by: ✓ tumour size (T1a vs T1b) ✓ randomising centre SABR: 26 Gy in 1 fraction for tumors <4 cm; 42 Gy in 3 fractions for tumors 4-7 cm RAPN Figure 1. RAPSTAR trial design. ENDOPOINTS Between-group difference in the change in estimated glomerular filtration rate (eGFR) from baseline to 24 months after treatment (mL/min/1.73 m²). FREEDOM FROM LOCAL PROGRESSION · FREEDOM FROM DISTANT FAILURE · OVERALL SURVIVAL · PROGRESSION-FREE SURVIVAL · QUALITY OF LIFE · COST-EFFECTIVENESS EXPLORATORY ANALYSES 1. Evaluation of multiparametric magnetic resonance imaging (mpMRI) as a predictor of treatment response. 2. Analysis of circulating tumour DNA (ctDNA) to identify genetic alterations, mutational profiles, and tumour-specific DNA fragments associated with treatment response and disease progression. 3. Evaluation of augmented reality (AR) technologies for improving intraoperative navigation. 4. Intraoperative tumour localization using indocyanine green (ICG) fluorescence imaging and/or robot-assisted intraoperative ultrasonography. 5. Validation and comparative assessment of the R.E.N.A.L., PADUA, and SPARE nephrometry scoring systems for predicting the anatomical complexity of partial nephrectomy. FOLLOW-UP Post-treatment follow-up: 6, 12, 18 and 24 months Clinical assessment: physical examination, ECOG performance status. Imaging: chest CT and abdominal/pelvic MRI. Renal function: serum creatinine, eGFR, urinalysis; renal scintigraphy (EC/MAG3) at 12 and 24 months. Quality of life: EORTC QLQ-C30. Safety: adverse event and toxicity assessment. Exploratory biomarkers: ctDNA sampling at 12 and 24 months. Disease outcomes: progression, subsequent treatment and survival follow-up. DISCLOSURE STATEMENT Bartlomiej Tomasik declares no conflicts of interest. The RAPSTAR study is funded by the Polish government through the Medical Research Agency (2024/ABM/02/00045). REFERENCES (1) Bex A et al. "European association of urology guidelines on renal cell carcinoma: the 2025 update" European urology 87.6 (2025): 683-696. (2) Siva S et al. "Long-term outcomes of stereotactic ablative body radiotherapy for primary kidney cancer (TROG 15.03 FASTRACK II): a multicentre, non-randomised, phase 2 study" The Lancet Oncology 27.6 (2026): 711-720. (3) Borm GF, Fransen J, Lemmens WAJG. "A simple sample size formula for analysis of covariance in randomized clinical trials" Journal of clinical epidemiology 60.12 (2007): 1234-1238. Contact: bartlomiej.tomasik@gumed.edu.pl

[Slide 1] Figure 1: LIBERTAS Study Design Baseline + ITP (6 mos) Newly diagnosed mCSPC • Metastasis detected by conventional imaging and/or next-generation imaging • ECOG PS 0-1 (up to 2-3)a APA 240 mg/d + ADT + Ametris CentrePoint Insight Watchb Stratification • Tumor volume • Prior localized prostate cancer treatment MTP (≥18 mos) + ≤24-mos follow-up PSA undetectable <0.2 ng/mL -> R 1:1 -> Arm A: APA + Intermittent ADT / Arm B: APA + Continuous ADT PSA detectable ≥0.2 ng/mL -> Continue SOC off study Dual primary endpoints • 18-mo radiographic PFS rate • 18-mo % change in severity-adjusted HF score [Slide 2] Figure 2: Daily life activity behavior in the ITP overall population. Low HF severity | High HF severity Top panels - Y-axis: Mean ± SEM awake sedentary time (min/d); axis 430-470 Bottom panels - Y-axis: Mean ± SEM awake MVPA time (%); axis 9.0-12.5 X-axis timepoints: Baseline, C4D1, C7D1 [Line plots with error bars; no data values printed. Low HF severity sedentary time plotted below ~445 min/d; High HF severity plotted ~456-465 min/d. MVPA: low HF ~11-12%, high HF ~9.4-9.8% (read from plot, not printed).] [Slide 3] Figure 5: Patient-reported PHQ-9 fatigue (A) and sleep disturbance (B) in the ITP overall population. A - Low HF severity: Baseline (n=145), C4D1 (n=142), C7D1 (n=141) A - High HF severity: Baseline (n=148), C4D1 (n=136), C7D1 (n=138) B - Low HF severity: Baseline (n=145), C4D1 (n=142), C7D1 (n=141) B - High HF severity: Baseline (n=148), C4D1 (n=136), C7D1 (n=138) Y-axis: % of patients Categories: Not at all / Several days / More than half the days / Nearly every day [Grouped bar charts; no bar values printed.] [Footnote line at bottom cut off / illegible]

A Study or Subgroup | log[Hazard Ratio] | SE | Weight | Hazard Ratio IV, Random, 95% CI Lonati 2022 | -0.32 | 0.54 | 36 % | 0.73 [0.25, 2.09] Spaliviero 2014 | 0.05 | 0.86 | 17 % | 1.05 [0.19, 5.67] Willis 2015 | 0.84 | 0.45 | 45 % | 2.32 [0.96, 5.60] Total (95% CI) | | | 100 % | 1.32 [0.60, 2.88] Heterogeneity: Tau² = 0.14; Chi² = 2.84, df = 2 (P = 0.2 ); I² = 30% Test for overall effect: Z = 0.69 (P = 0. 5 ) Favors BP Favors RC B Study or Subgroup | log[Hazard Ratio] | SE | Weight | Hazard Ratio IV, Random, 95% CI Dursun 2022 | -0.17 | 0.12 | 94 % | 0.84 [0.67, 1.07] Lonati 2022 | -0.07 | 0.49 | 5 % | 0.93 [0.36, 2.44] Total (95% CI) | | | 100 % | 0.85 [0.68, 1.07] Heterogeneity: Tau² = 0.00; Chi² = 0.04, df = 1 (P = 0.8 ); I² = 0% Test for overall effect: Z = 1.41 (P = 0.16) Favors BP Favors RC

Figure 1: Study design 6 to 8 months after standard of care initiation | 28 days max after informed consent form | Treatment phase (until CRPC stage) Pre-screening: Standard of care initiation Screening phase: - De novo metastatic disease at initial diagnosis - Serum PSA ≥ 0,2 ng/ml at 6-8 mo after standard of care initiation - Stable or declining PSA but not a rising one - Testosterone level < 50ng/dl or <1,7 nmol/L R1:1 Arm A: Standard of care Arm B: Standard of care + ¹⁷⁷Lu-PSMA-617 -> Arm B: Standard of care 4 cycles of ¹⁷⁷Lu-PSMA-617 (every 6 weeks) Follow-up phase: Up to 10 years after randomization

TEAEs and TRAEs Experienced by ≥20% of Patients • Ten (90.9%) patients experienced a TEAE of any grade • Seven (63.6%) patients experienced a G ≥3 TEAE • Three patients experienced reversible G4 neutropenia • No febrile neutropenia was observed • Two patients with G1 neuropathy at baseline experienced G3 peripheral sensory neuropathy • No G5 events were observed [Butterfly bar chart: TEAE (left) | TRAE (right), Patients, % (0-60), by Grade 1-4] AEs listed: Constipation; Arthralgia; Back pain; Nausea; Vomiting; Fatigue; Neutropenia; Peripheral neuropathy, NOS; Peripheral sensory neuropathy; Peripheral motor neuropathy; Alopecia; Cough; Decreased appetite; Insomnia The safety profile of Ixo-V continues to be manageable, associated with predominately low-grade TEAEs, and consistent with prior reported experience. G, grade; Ixo-V, ixotatug vedotin; NOS, not otherwise specified; TEAEs, treatment-emergent adverse events; TRAEs, treatment-related adverse events.

First results of SWOG S2200, a randomized, phase II study comparing cabozantinib with or without atezolizumab in metastatic papillary renal cell carcinoma (mPRCC) Abstr 3591 – PAPMET 2 trial PLUS - We deal with a randomised clinical trial in a rare cancer population - SWOG organised trial – consortium - ORR of 47% is promising and in line with other combi trials in this population -At first sight no new safety signals -Most pts were first line MINUS -Inclusion of a limited number of patients -23pts, slow inclusion rate (200 patients) prematurely closed -PFS is primary endpoint (OS, ORR, safety = secondary) but seems comparable with other combi trials in this setting, PR in 8/17 pts = 47% -No MET information is provided, biomarkers ? (Sarcomatoid expression) -2 grade 5 events are reported : 1 in each group : high !

[Slide 1] Results Mean change over time †Higher scores indicate better QOL FACT-P Total Score† * Indicates p<0.05 Trial Outcomes Index (TOI)† [Two line charts: Mean change from baseline vs Weeks (6, 12, 18, 24, 30, 36); legend LuP / Doc; asterisks at weeks 24, 30, 36 (FACT-P) and weeks 24, 30 (TOI)] Weeks Mixed Model Assessment week x treatment arm interaction: p = 0.002 Weeks Mixed Model Assessment week x treatment arm interaction: p < 0.001 Presenter: Dr. Daniel J. Khalaf [Slide 2] Results GP5 item: 'I am bothered by the side effects of treatment' GP5 Item – Side Effect Bother † †Higher scores indicate worse side effect bother [Line chart: Mean change from baseline vs Week (6, 12, 18, 24, 30, 36); legend LuP / Doc; asterisks at weeks 6, 12, 24, 30, 36] Week Mixed Model Assessment week x treatment arm interaction: p = 0.007 Worst Recorded GP5 'Side Effect Bother' Score Legend: 0 | 1 to 2 | 3 to 4 LU-P: 28.6% | 56.0% | 15.4% DOC: 9.9% | 49.5% | 40.7% Kruskal-Wallis P < 0.001 % patients Score 0: Not at all Score 1: A little bit Score 2: Somewhat Score 3: Quite a bit Score 4: Very much Presenter: Dr. Daniel J. Khalaf

[Slide 1] Table 1: Demographic, clinical, and treatment characteristics | ≤2-lines (n=133) | ≥3-lines (n=91) | Total (n=224) Age at start Ra-223; years, median (range) | 77.0 (53.0-92.0) | 72.0 (45.0-88.0) | 75.0 (45.0-92.0) ECOG PS (n=162); n (%) | | | 0 | 20 (20.4) | 11 (17.2) | 31 (19.1) 1 | 70 (71.4) | 44 (68.8) | 114 (70.4) 2 | 8 (8.2) | 9 (14.1) | 17 (10.5) Extent of bone disease > 10 lesions; n (%) | 67 (50.4) | 56 (61.5) | 123 (54.9) Other metastatic locations; n (%) | | | Lymph node only | 16 (12.0) | 8 (0.8) | 24 (10.7) Lung | 4 (3.0) | 1 (1.1) | 5 (2.2) Liver | 1 (0.8) | 1 (1.1) | 2 (0.9) Prior treatments; n (%) | | | ARPIs | 119 (92.2) | 89 (97.8) | 208 (92.9) Chemotherapy | 61 (47.3) | 91 (100.0) | 152 (67.9) 5-6 injections of Ra-223; n (%) | 91 (68.4) | 49 (53.8) | 140 (62.5) Bone supportive agents during Ra-223; n (%) | 39 (29.3) | 25 (27.2) | 64 (28.6) [Slide 2] Figure 3: OS by No. mPC lines prior to Radium-223 No. mPC lines prior Radium-223 Early Line (0-2L) Late line (≥3L) Median: 12.4 months 95% CI: 11.1 – 13.8 Median: 9.3 months 95% CI: 7.7 – 11.0 N. at Risk | 0 | 6 | 12 | 18 | 24 | 30 | 36 | 42 | 48 Early line | 133 | 110 | 69 | 36 | 16 | 8 | 2 | 1 | 0 Later line | 91 | 66 | 35 | 19 | 8 | 3 | 0 Follow-up time (months) Cumulative survival Survival table Endpoint: Overall survival by number of mPC lines prior to Radium-223 Arm | N | Median | Median CI Early Line (0-2L) | 133 | 12.4 months | 11.1 - 13.8 Late line (≥3L) | 91 | 9.3 months | 7.7 - 11.0 [Slide 3] Figure 3: OS by No. mPC lines prior to Radium-223 No. mPC lines prior Radium-223 Early Line (0-2L) Late line (≥3L) Median: 12.4 months 95% CI: 11.1 – 13.8 Median: 9.3 months 95% CI: 7.7 – 11.0 N. at Risk | 0 | 6 | 12 | 18 | 24 | 30 | 36 | 42 | 48 Early line | 133 | 110 | 69 | 36 | 16 | 8 | 2 | 1 | 0 Later line | 91 | 66 | 35 | 19 | 8 | 3 | 0 Follow-up time (months) Cumulative survival Survival table Endpoint: Overall survival by number of mPC lines prior to Radium-223 Arm | N | Median | Median CI Early Line (0-2L) | 133 | 12.4 months | 11.1 - 13.8 Late line (≥3L) | 91 | 9.3 months | 7.7 - 11.0 [Slide 4] Figure 4: Sankey diagram: Sequence of treatment lines for mPC prior to Radium-223 therapy L1-ARPi N=148 L1-Docetaxel N=70 31.8% L1-Triplet therapy N=2 0.9% L2-Radium-223 N=67 30.5% L2-ARPi N=71 32.3% L2-Docetaxel N=69 31.4% L2-Cabazitaxel N=6 2.7% L2-Others N=4 1.8% L2-Lu-PSMA N=2 0.9% L2-PARPi N=1 0.5% L3-Radium-223 N=62 40.5% L3-Cabazitaxel N=37 24.2% L3-ARPi N=21 13.7% L3-Docetaxel N=20 13.1% L3-Others N=7 4.6% L3-PARPi N=4 2.6% L3-Carboplatin N=2 1.3% L4-Radium-223 N=68 74.7% L4-Cabazitaxel N=12 13.2% L4-ARPi N=4 4.4% L4-Others N=3 3.3% L4-Carboplatin N=2 2.2% L4-Docetaxel N=1 1.1% L4-Lu-PSMA N=1 1.1% L5-Radium-223 N=18 78.3% L5-Carboplatin N=2 8.7% L5-ARPi N=1 4.3% L5-PARPi N=1 4.3% L5-Others N=1 4.3% L6-Radium-223 N=3 60.0% L6-Carboplatin N=1 20.0% L6-Others N=1 20.0% L7-Radium-223 N=2 100.0%

[Title partially cropped] Identify... Co-expression of targets – identify[ing...] Target co-expression (protein): All ROIs Matrix axes (proteins): Axl, c-Kit, CD44v6, CD47, CD70, CD73, CD99, EGFR, ErbB2_HER2, Fibronectin, IGF1 Receptor, Met (c-Met), MUC1, NCAM1, Nectin-4, PDGFR alpha, PDGFR alpha + PDGFR beta, PSMA, Tissue Factor, TROP2 [Dot matrix with boxed co-expression pairs; circled highlights:] ERBB2 / AXL EGFR / CD70 Nectin-4 / KIT PSMA / KIT Breelyn A. Wilky, MD – University of Colorado Anschutz School of Medicin[e]

Progression Free Survival Progression-Free Probability PFS Time (Months) PFS Events: 14 PFS Events: 9 Treatment Arm Cabo Only Cabo + Atezo Number at risk Cabo 10 months Cabo/Atezo 25 months HR 0.46 95% CI 0.22 – 0.97 One-sided P-value 0.043 Survival table Endpoint: PFS Cabo only | 16 | 10 months Cabo + Atezo | 17 | 25 months HR: 0.46 HR CI: 0.22–0.97 P: 0.043 (one-sided)

ESMO 20[26] Genomic biomarkers of response to Extreme Bipolar Androgen Therapy (BAT) Alternating Darolutamide and Testosterone Cypionate in Patients with Metastatic Castration-Resistant Prostate Cancer (mCRPC) - ExBAT trial/LACOG 0620 Authors: Diogo Assed Bastos, MD, PhD; Daniel Girardi, MD; Andrey Soares, MD; Fernando Cotait Maluf, MD, PhD; Denis Leonardo Jardim, MD, PhD; Igor Morbeck, MD, MSc; Mariane Sousa Fontes, MD; Vinicius Carrera Souza, MD; Rafaela Gomes de Jesus, BSc, MSc; Pedro Isaacsson Velho, MD, PhD Abstract No: [illegible] | Poster No: [illegible] INTRODUCTION - Sequencing androgen receptor-pathway inhibitors (ARPI) have limited activity in mCRPC. - Bipolar androgen therapy (BAT), which cycles serum testosterone between supraphysiological and castrate levels, can resensitize prostate cancer to androgen receptor pathway inhibitors (ARPIs) in patients who previously progressed on them. - After a course of BAT, enzalutamide rechallenge has yielded meaningful PSA and radiographic responses, with PSA50 rates above 50%. The mechanism is incompletely understood but plausibly reflects BAT's demonstrated ability to downregulate AR and reverse AR-V7 expression in circulating tumor cells. - To report exploratory genomic biomarker analyses to identify predictors of response. METHODS [Figure 1 schema - small text largely illegible: Phase 2, single-arm, multicenter (9 institutions), open-label trial - N = 51 treated patients; Testosterone cypionate (BAT) / Darolutamide / Washout; Endpoints] - Study Design: phase II, single-arm trial in mCRPC pts after progression to abiraterone (Figure 1). - Population & Settings: Pts received testosterone cypionate 400 mg IM (day 1) followed by darolutamide 1200 mg/day (days 29-56) in 63-day cycles. - Data Collection: ctDNA was collected at baseline and end of treatment and evaluated for evaluation by next-generation sequencing (NGS) of AR alterations, including amplification, and mutations as potential biomarkers of rPFS and OS. - Study Registration: NCT04558866 (ClinicalTrials.gov). RESULTS rPFS by AR Amplification Status - AR amplification was significantly associated with inferior rPFS (median 8.3 vs. 2.9 months; HR 0.39; p = 0.0284), compared to non-amplified status (Figure 2). [KM] Median (95% CI) 8.3m (3.9-NE) [Non-Amplified]; Median (95% CI) 2.9m (1.8-NE) [Amplified] P-value: 0.0284; HR: 0.39 AR | Events/Total | % 12m rPFS (95% CI) | HR (95% CI) Amplified | 7/8 | 0.0 (NE-NE%) | Reference Non-Amplified | 23/38 | 42.3 (26.2-67.6%) | 0.39 (0.16-0.94) At Risk (Months 0,2,4,6,8,10,12,14): Amplified 8, 6, 2, 2, 2, 0; Non-Amplified 38, 30, 22, 18, 18, 16, 15, 0 Figure 2. Radiographic Progression-Free Survival* by AR amplification Status. 12m Overall Survival by AR Amplification Status - Overall survival was significantly longer in patients with non-amplified AR compared to those with AR amplification (HR 0.19, 95% CI 0.07-0.55; p=0.001) (Figure 3). [KM] 12m OS (95% CI) 88.9% (73.0-95.7%) [Non-amplified]; 12m OS (95% CI) 50% (15.2-77.5%) [Amplified] P-value: 0.001; HR: 0.19 AR | Events/Total | Median (95% CI) | HR (95% CI) Amplified | 6/8 | 11.6 (3.7-NE) | Reference Non-Amplified | 9/38 | NR (21.6-NE) | 0.19 (0.07-0.55) At Risk (Months 0-14): Amplified 8, 8, 7, 6, 6, 5, 3, 2; Non-amplified 38, 38, 37, 36, 34, 32, 31, 25 Figure 3. Overall Survival by AR Amplification Status. Genomic Profile and Individual Clinical Outcomes [Oncoprint heatmap + swimmer plot; gene labels and keys largely illegible] Figure 4: Swimmer plot with available molecular somatic profiling results [Footnotes on radiographic progression definitions and **Other gene list (e.g. SF3B1, CTNNB1, ARID2, SPOP, ZFHX3, APC, SRSF2, PIK3CA, SETD2, ACVR1, RAD50, FANCI, BARD1, AXIN2, PRKN, AKT1, FOXA1, MYC, GNAS, EPCAM, FANCA, RET, LRP1B, MGA, INPP4A, MSH6, NBN, RB1, STAG2, KDM6A, PMS1, RNF43, EPHA3, KEAP1, CBL, PALB2, BRAF, RPS6KB1, FOXL2, MEN1); DNMT3A, TET2, ASXL1 and PPM1D are recognized clonal hematopoiesis (CHIP)-associated genes ... PTEN alterations were not detected in this baseline ctDNA cohort. - partly illegible] CONCLUSION - Alternating bipolar androgen therapy and darolutamide (ExBAT) showed durable rPFS and a manageable safety profile, in men with mCRPC who progressed on abiraterone. - Baseline AR amplification emerged as a potential biomarker of resistance to this strategy, identifying a subgroup with markedly worse rPFS and OS and warranting prospective validation as a tool for treatment selection. These findings support further evaluation of BAT-based ARPI sequencing, potentially guided by AR status, as an alternative to conventional therapy in this setting. Survival table Endpoint: rPFS by AR amplification status (Figure 2) Arm | N | Median | Median CI AR non-amplified | 38 | 8.3 mo | 3.9-NE AR amplified | 8 | 2.9 mo | 1.8-NE HR: 0.39 (non-amplified vs amplified reference) HR CI: 0.16-0.94 P: 0.0284 Landmark rates: 12 mo rPFS: AR non-amplified 42.3% (CI 26.2-67.6%) | AR amplified 0.0% (CI NE-NE) 12 mo OS: AR non-amplified 88.9% (CI 73.0-95.7%) | AR amplified 50% (CI 15.2-77.5%)

[Slide 1] COSMIC-313 Is a Randomized First-Line Trial in Advanced ccRCC Comparing C+N+I vs N+I With Serial Plasma KIM-1 Collection Advanced RCC (N=855) • No prior systemic therapy • Clear cell component • Intermediate or poor risk • Measurable disease (RECIST v1.1) • Karnofsky performance status ≥70% R 1:1 • IMDC risk • Region Cabozantinib + Nivolumab + Ipilimumab (C+N+I) Weeks 0–12: Cabozantinib 40 mg PO QD, Nivolumab 3 mg/kg Q3W, Ipilimumab 1 mg/kg Q3W Weeks 13+: Cabozantinib 40 mg PO QD, Nivolumab 3 mg/kg Q3W Placebo + Nivolumab + Ipilimumab (N+I) Weeks 0–12: Placebo, Nivolumab 3 mg/kg Q3W, Ipilimumab 1 mg/kg Q3W Weeks 13+: Placebo, Nivolumab 3 mg/kg Q3W Weeks 0 3 6 9 13 n=703 n=635 n=583 n=533 n=495 KIM-1 test results (n) C+N+I n=356 n=272 n=208 n=158 n=146 N+I n=347 n=295 n=253 n=203 n=139 Plasma samples, collected at baseline and post treatment, were profiled for KIM-1 levels using an enzyme-based electrochemiluminescence assay [Slide 2] Higher Baseline KIM-1 Levels Are Associated With IMDC Poor-Risk Disease, Absence of Prior Nephrectomy, and Greater Tumor Burden IMDC Risk Wilcoxon, P = 1e−10 KIM−1 (pg/mL) Intermediate | Poor Prior Nephrectomy Wilcoxon, P < 2.2e−16 No | Yes Radiographic Tumor Burden R = 0.58, P < 2.5e−16 Week 1 KIM−1 (pg/mL) Sum of diameters at week 1 [Slide 3] Early KIM-1 Decrease Is Predictive of Clinical Outcome in IO Arm (N+I) Key Finding: • A >30% decrease at week 3 strongly predicts PFS, OS, and ORR (59% vs. 31%) with N+I • Under triplet therapy, this threshold does not stratify outcomes, indicating distinct biology Distribution by Arm Density Treatment C+N+I / N+I 30% Decrease Percent Change from W1D1 to W4D4 Placebo + Nivolumab + Ipilimumab (N+I) Cabozantinib + Nivolumab + Ipilimumab (C+N+I) PFS OS N+I >30% Decrease (n=109) N+I ≤30% Decrease (n=212) C+N+I >30% Decrease (n=246) C+N+I ≤30% Decrease (n=68) Progression-free survival probability Overall survival probability Months Survival table Endpoint: PFS and OS by week-3 KIM-1 decrease (>30% vs ≤30%) N+I >30% decrease | 109 N+I ≤30% decrease | 212 C+N+I >30% decrease | 246 C+N+I ≤30% decrease | 68 [Slide 4] Machine Learning Model to Predict Week-3 KIM-1 Decrease Captured Metabolic Signatures in Addition to Immune-Inflamed Programs Key Finding: • AI/machine learning model trained on baseline N+I RNA-seq to predict N+I-mediated KIM-1 decrease • The model captured metabolic signatures, which were enriched in non-decreasers, have been implicated in immunosuppressive tumors states, and may define settings in which VEGFR-targeted therapy provides incremental benefit N+I Placebo PO QD + nivolumab 3 mg/kg IV Q3W x 4 + ipilimumab 1 mg/kg IV Q3W x 4 ≤30% Decrease | >30% Decrease DGE using N+I bulk RNA-seq Train a random forest classifier to predict N+I-associated 30% KIM-1 decrease Random Forest Classifier Model Performance Accuracy 88.6% Sensitivity 81.9% Specificity 100 % Classify all patients into N+I-associated 30% KIM-1 decreaser Predicted / Observed [illegible] Patients | [illegible] Patients | 21 Patients | [illegible] Patients C+N+I | N+I No N+I-associated 30% KIM-1 decr. | N+I-associated 30% KIM-1 decr. z-score IMDC: Intermediate / Poor PDL1 TPS: <1% / ≥1% Original KIM-1 decrease: ≤30% Decrease / >30% Decrease CD8+ Tem pDC Class-switched memory B-cells ImmuneScore MicroenvironmentScore CD4+ memory T-cells CD8+ T-cells CD8+ Tcm aDC HALLMARK_INFLAMMATORY_RESPONSE HALLMARK_IL6_JAK_STAT3_SIGNALING HALLMARK_COMPLEMENT HALLMARK_ALLOGRAFT_REJECTION HALLMARK_INTERFERON_GAMMA_RESPONSE Macrophages M1 Macrophages Tregs DC HALLMARK_SPERMATOGENESIS HALLMARK_BILE_ACID_METABOLISM HALLMARK_ADIPOGENESIS HALLMARK_FATTY_ACID_METABOLISM HALLMARK_PEROXISOME Immune Signatures Metabolic Signatures C, cabozantinib; decr, decrease; DGE, differential gene expression; I, ipilimumab; IMDC, International mRCC Database Consortium; KIM-1, kidney injury molecule-1; N, nivolumab; PDL1, programmed death-ligand 1; PO, by mouth; TPS, tumor proportion score; VEGFR, vascular endothelial growth factor receptor; QD, once daily; Q3W, every 3 weeks.

[Slide 2] DFS by Protocol-Defined Risk Subgroup Intermediate-high HR 0.71 (95% CI, 0.58-0.87) High HR 0.54 (95% CI, 0.31-0.94) M1 NED HR 0.50 (95% CI, 0.26-0.96) DFS, % Months Pembro Placebo No. at risk Survival table Endpoint: DFS by protocol-defined risk subgroup (exploratory) Pembro, intermediate-high | 422 Placebo, intermediate-high | 433 Pembro, high | 40 Placebo, high | 37 Pembro, M1 NED | 29 Placebo, M1 NED | 28 HR: Intermediate-high 0.71; High 0.54; M1 NED 0.50 HR CI: 0.58-0.87; 0.31-0.94; 0.26-0.96 Intermediate-high 60/72/84 mo: pembro 63.5% / 62.1% / 59.8%; placebo 55.5% / 52.3% / 48.3% (as labelled) High 60/72/84 mo: pembro 44.6% / 44.6% / 44.6%; placebo 29.1% / 26.2% / 19.4% (as labelled) M1 NED 60/72/84 mo: pembro 48.9% / 36.2% / 36.2%; placebo 27.6% / 23.0% / 23.0% (as labelled) [Slide 3] OS by Protocol-Defined Risk Subgroup Intermediate-high HR 0.66 (95% CI, 0.48-0.91) High HR 0.92 (95% CI, 0.44-1.90) M1 NED HR 0.54 (95% CI, 0.19-1.53) OS, % Months Pembro Placebo No. at risk Survival table Endpoint: OS by protocol-defined risk subgroup (exploratory) Pembro, intermediate-high | 422 Placebo, intermediate-high | 433 Pembro, high | 40 Placebo, high | 37 Pembro, M1 NED | 29 Placebo, M1 NED | 28 HR: Intermediate-high 0.66; High 0.92; M1 NED 0.54 HR CI: 0.48-0.91; 0.44-1.90; 0.19-1.53 Intermediate-high 60/72/84 mo: pembro 89.2% / 87.5% / 84.7%; placebo 83.7% / 80.9% / 78.0% (as labelled) High 60/72/84 mo: pembro 75.0% / 72.5% / 61.4%; placebo 73.0% / 70.3% / 67.6% (as labelled) M1 NED 60/72/84 mo: pembro 86.1% / 82.5% / 78.1%; placebo 70.6% / 66.9% / 66.9% (as labelled)

[Slide 1] MADRID 2026 ESMO congress Clinical impact of AI Models in Medical Imaging: Evidence from the ChAImeleon Multi-Cancer Project Gloria Ribas 23-27 October 2026 Gloria Ribas Despuig The in-silico clinical validation CHAIMELEON platform: A standardized multi-center framework to validate the impact of human-AI interaction across five cancers [Slide 2] CLINICAL VALIDATION PIPELINE WORKFLOW OF THE CLINICAL VALIDATION DEVELOPED INPUTS FROM AI MODELS - CHAIMELEON PROJECT OVERALL SURVIVAL STAGE : TNM LUNG COLON BREAST PROSTATE RECTUM HIGH VS LOW RISK HISTOLOGY SUBTYPES INVASION: EVMI &/OR MFI chAImeleon DICOM IMAGES MMG MRI CT STRUCTURED CLINICAL DATA DEVELOPMENT OF THE PLATFORM FOR IN-SILICO CLINICAL VALIDATION 1 CLINICAL EVALUATION 2 AI TRUST ASSESSMENT 3 FINAL CHECK GT: Ground Truth CLINICIAN RECRUITMENT & PLATFORM TESTING 46 CLINICIANS RECRUITED EACH CLINICIAN EVALUATED 20 UNIQUE CASES PER ASSIGNED CANCER TYPE 1 CLINICAL EVALUATION 30-DAY WASHOUT PERIOD 2 AI TRUST ASSESSMENT 3 FINAL CHECK USER FEEDBACK LIKERT QUESTIONNAIRE EACH CLINICIAN EVALUATED THE SYSTEM USING A 5-POINT LIKERT SCALE ACROSS 4 DIMENSIONS: PLATFORM USABILITY TRUST IN AI AI UTILITY AI IMPACT ON WORKFLOW [Slide 3] PERFORMANCE TIME DX - IMPROVEMENT WITH AI SUPPORT Phase I — 2 months — Phase II Diagnosis time Lung p < 0.001 | Breast p < 0.001 | Prostate p < 0.001 | Colon p < 0.001 | Rectum p < 0.001 | Global p < 0.001 Diagnosis made by the clinician without AI assistance | Diagnosis made by the clinician with AI assistance AI, Artificial Inteligence. Accuracy (without AI | with AI) Overall survival: Lung 40.0% | 51.2% Histology types: Breast 43.3% | 44.3% Risk group: Prostate 75.1% | 75.3% TNM stages Size, node involment, metastasis: Colon (T) 35.9% | 55.0%; Colon (N) 36.8% | 40.5%; Colon (M) 80.5% | 94.1%; Colon (global) 17.7% | 23.6% Invasion (IVEM or/and IMF) extramural vascular or/& fascia mesorrectal: Rectum (EVI) 61.3% | 66.7%; Rectum (MFI) 69.0% | 78.3%; Rectum (global) 56.3% | 64.3% Overall: ChAImeleon 54.3% | 58.3% COLON 14% | LUNG 22% | RECTUM 20% | PROSTATE 26% | BREAST 18% Improved 55% | Unchanged 17% | Worsened 28%

[Slide 1] 4737RO Durvalumab (D) in Combination With BCG Induction and Maintenance (I+M) for BCG-Naïve, High-Risk Non-Muscle-Invasive Bladder Cancer (NMIBC): Early Recurrence and Cystectomy-Free Survival in Patients With Papillary Only Tumours From POTOMAC Maria De Santis,1,2 Neal D. Shore,3 Joan Palou Redorta,4 Michał Krawczyński,5 Artur Seyitkuliev,6 Andrey Novikov,7 Félix Guerrero-Ramos,8 Minoru Kato,9 Lieven Goeman,10 Eva Hellmis,11 Thomas Powles,12 Kilian M. Gust,13 Robert Zielinski,14 Julien Deturmeny,15 Girish Kulkarni,16 Jarmo Hunting,17 Ana Quartilho,18 Suliman Boulos,18 Hiroyuki Nishiyama19 1Charité Universitätsmedizin, Berlin, Germany; 2Medical University of Vienna, Vienna, Austria; 3START Carolinas/Carolina Urologic Research Center, Myrtle Beach, SC, USA; 4Fundació Puigvert, Universitat Autònoma de Barcelona, Barcelona, Spain; 5Clinical Research Centre Sp., Poznan, Poland; 6St. Petersburg Hospital of the Russian Academy of Sciences, St. Petersburg, Russia; 7North-Western State Medical University, St. Petersburg, Russia; 8Hospital Universitario 12 de Octubre, Madrid, Spain; 9Osaka Metropolitan University, Osaka, Japan; 10University Hospitals Leuven, Leuven, Belgium; 11Urologicum-Duisburg, Duisburg, Germany; 12Barts Cancer Institute Biomedical Research Centre, Queen Mary University of London, London, UK; 13Medical University of Vienna, Vienna, Austria; 14Western Sydney University, Orange Hospital, NSW, Australia; 15Hôpital Européen, Marseille, France; 16Princess Margaret Cancer Centre, Toronto, ON, Canada; 17St. Antonius Hospital, Utrecht, The Netherlands; 18AstraZeneca, Cambridge, UK; 19University of Tsukuba, Ibaraki, Japan [Slide 2] POTOMAC: Baseline Characteristics for Patients With Papillary Only Tumoursa Demographic and disease characteristics were generally well balanced between the arms Characteristics | D+BCG (I+M) N=217 | BCG (I+M) N=220 [Values as printed, top to bottom; row labels appear vertically offset from values on the projected slide] Median age, years (range) | 67 (24–86) | 67 (32–86) Sex, n (%) Male | 170 (78) | 173 (79) Region, n (%) Western Europe | 85 (39) | 74 (34) ROW | 132 (61) | 146 (66) Race, n (%) White | 178 (82) | 180 (82) Asian | 33 (15) | 33 (15) Other | 3 (1) | 5 (2) Missing | 3 (1) | 2 (1) ECOG PS, n (%) 0 | 186 (86) | 195 (89) 1 | 31 (14) | 25 (11) Smoking, n (%) Current | 42 (19) | 41 (19) Former | 102 (47) | 109 (50) Never | 73 (34) | 70 (32) Disease stage, n (%) T1 [red box] | 137 (63) | 145 (66) Ta | 79 (36) | 75 (34) Other | 1 (0.5) | 0 Higher-risk papillary disease,b n (%) Yes | 114 (53) | 120 (55) aDefined as any grade T1 or Ta disease without CIS, that met the high-risk disease enrollment criteria. bDefined as T1G3/T1 high-grade or multiple and recurrent and large tumours (those with a diameter of ≥3 cm). Data cutoff 03 April 2025. BCG, Bacillus Calmette-Guérin; D, durvalumab; ECOG PS, Eastern Cooperative Oncology Group performance status; I, induction; M, maintenance; ROW, rest of world.

[Slide 1] Table 1. Baseline CTC Counts (ITT Population) | Talazoparib + Enzalutamide (n=300) | Placebo + Enzalutamide (n=299) Patients with evaluable CTC count, n | 182 | 175 Baseline CTC count | | Median (range), cells/7.5 mL | 0 (0–105) | 0 (0–1486) Mean (SD), cells/7.5 mL | 3.4 (12.3) | 14.2 (113.7) ≥5 CTC/7.5 mL at baseline, n (%) | 30 (10.0) | 36 (12.0) <5 CTC/7.5 mL at baseline, n (%) | 152 (50.7) | 139 (46.5) =0 CTC/7.5 mL at baseline, n (%) | 121 (40.3) | 112 (37.5) >0 CTC/7.5 mL at baseline, n (%) | 61 (20.3) | 63 (21.1) Counts reported for patients who had a baseline CTC assessment and at least one post-baseline CTC assessment. The denominator for patients with CTC counts at baseline is the ITT population. CTC=circulating tumor cell; ITT=intent-to-treat; SD=standard deviation. [Slide 2] Table 2. Post-Baseline CTC Conversion at Any Time (ITT Population) | Talazoparib + Enzalutamide (n=300) | Placebo + Enzalutamide (n=299) ≥5 CTC/7.5 mL at baseline, n (%) | 30 (10.0) | 36 (12.0) CTC conversion from ≥5 to <5, n | 27 | 23 CTC conversion* from ≥5 to <5, % (95% CI) | 90.0% (72.3–97.4) | 63.9% (46.2–78.7) >0 CTC/7.5 mL at baseline, n (%) | 61 (20.3) | 63 (21.1) CTC conversion from >0 to 0, n | 42 | 38 CTC conversion* from >0 to 0, % (95% CI) | 68.9% (55.6–79.8) | 60.3% (47.2–72.2) <5 CTC/7.5 mL at baseline, n (%) | 152 (50.7) | 139 (46.5) CTC increase from baseline <5 value at any time, n | 119 | 106 CTC increase* from baseline <5 value at any time, % (95% CI) | 78.3% (70.7–84.4) | 76.3% (68.2–82.9) Counts reported for patients who had a baseline CTC assessment and at least one post-baseline CTC assessment. The denominator for patients with CTC counts at baseline is the ITT population. *Conversion counts are from blood collection time points at Weeks 9, 17, and 25. The denominator for conversion/increase calculations is patients who have CTC counts at baseline under each category. Two-sided 95% CI was calculated using the Wilson score interval with continuity correction. CI=confidence interval; CTC=circulating tumor cell; ITT=intent-to-treat. [Slide 3] Figure 1. rPFS in CTC Evaluable Population by Treatment and A) CTC=0 and B) CTC <5 at Baseline A. | Events/n | Median rPFS (95% CI), mo TALA + ENZA, CTC=0 | 25/121 | NR (NR–NR) TALA + ENZA, CTC>0 | 20/61 | NR (34.3–NR) PBO + ENZA, CTC=0 | 35/112 | NR (45.8–NR) PBO + ENZA, CTC>0 | 36/63 | 22.7 (17.2–NR) Unstratified TALA + ENZA BL CTC=0 vs CTC >0 HR=0.53; 95% CI, 0.29–0.95; 2-sided p=0.03 PBO + ENZA BL CTC=0 vs CTC >0 HR=0.42; 95% CI, 0.26–0.68; 2-sided p<0.001 B. | Events/n | Median rPFS (95% CI), mo TALA + ENZA, CTC <5 | 33/152 | NR (NR–NR) TALA + ENZA, CTC ≥5 | 12/30 | NR (18.4–NR) PBO + ENZA, CTC <5 | 48/139 | NR (45.8–NR) PBO + ENZA, CTC ≥5 | 23/36 | 18.4 (12.9–29.3) Unstratified TALA + ENZA BL CTC <5 vs CTC ≥5 HR=0.41; 95% CI, 0.21–0.80; 2-sided p=0.007 PBO + ENZA BL CTC <5 vs CTC ≥5 HR=0.35; 95% CI, 0.21–0.58; 2-sided p<0.001 rPFS, % (y-axis); rPFS, months (x-axis 0–52) No. at risk [rows printed but too small to transcribe reliably] rPFS based on investigator assessment. Unstratified hazard ratio based on Cox proportional hazards model. P-value from an unstratified log-rank test. Month 0 is the time of randomization. BL=baseline; CI=confidence interval; CTC=circulating tumor cell; ENZA=enzalutamide; HR=hazard ratio; NR=not reached; PBO=placebo; rPFS=radiographic progression-free survival; TALA=talazoparib. Survival table Endpoint: rPFS (investigator-assessed), CTC-evaluable population, by baseline CTC Arm | N | Median | Median CI Panel A: TALA + ENZA, CTC=0 | 121 | NR | NR–NR Panel A: TALA + ENZA, CTC>0 | 61 | NR | 34.3–NR Panel A: PBO + ENZA, CTC=0 | 112 | NR | 45.8–NR Panel A: PBO + ENZA, CTC>0 | 63 | 22.7 mo | 17.2–NR Panel B: TALA + ENZA, CTC <5 | 152 | NR | NR–NR Panel B: TALA + ENZA, CTC ≥5 | 30 | NR | 18.4–NR Panel B: PBO + ENZA, CTC <5 | 139 | NR | 45.8–NR Panel B: PBO + ENZA, CTC ≥5 | 36 | 18.4 mo | 12.9–29.3 HR: TALA+ENZA BL CTC=0 vs >0: 0.53; PBO+ENZA BL CTC=0 vs >0: 0.42; TALA+ENZA BL CTC<5 vs ≥5: 0.41; PBO+ENZA BL CTC<5 vs ≥5: 0.35 HR CI: 0.29–0.95; 0.26–0.68; 0.21–0.80; 0.21–0.58 P: p=0.03; p<0.001; p=0.007; p<0.001 (2-sided) [Slide 4] Figure 2. rPFS in CTC Evaluable Population by Treatment and A) CTC Conversion and B) CTC0 at Week 9 A. | Events/n | Median rPFS (95% CI), mo TALA + ENZA, CTC conversion | 10/22 | 32.0 (18.4–NR) TALA + ENZA, no CTC conversion | 1/6 | NR (NR–NR) PBO + ENZA, CTC conversion | 16/21 | 17.2 (6.5–26.5) PBO + ENZA, no CTC conversion | 5/12 | NR (15.4–NR) Unstratified TALA + ENZA CTC conversion vs no CTC conversion HR=2.81; 95% CI, 0.36–21.94; 2-sided p=0.30 PBO + ENZA CTC conversion vs no CTC conversion HR=2.31; 95% CI, 0.84–6.35; 2-sided p=0.09 B. | Events/n | Median rPFS (95% CI), mo TALA + ENZA, CTC0 | 6/22 | NR (NR–NR) TALA + ENZA, no CTC0 | 9/26 | NR (29.3–NR) PBO + ENZA, CTC0 | 9/19 | NR (22.7–NR) PBO + ENZA, no CTC0 | 12/24 | 29.5 (15.6–NR) Unstratified TALA + ENZA CTC0 vs no CTC0 HR=0.62; 95% CI, 0.22–1.75; 2-sided p=0.36 PBO + ENZA CTC0 vs no CTC0 HR=0.72; 95% CI, 0.30–1.73; 2-sided p=0.47 rPFS, % (y-axis); rPFS, months (x-axis 0–52) No. at risk [rows printed but too small to transcribe reliably] rPFS based on investigator assessment. Unstratified hazard ratio based on Cox proportional hazards model. P-value from an unstratified log-rank test. Month 0 is the time of randomization. CI=confidence interval; CTC=circulating tumor cell; CTC0=CTC count reduction from >0 to 0 per 7.5 mL of blood; ENZA=enzalutamide; HR=hazard ratio; NR=not reached; PBO=placebo; rPFS=radiographic progression-free survival; TALA=talazoparib. Survival table Endpoint: rPFS (investigator-assessed), CTC-evaluable population, by Week 9 CTC conversion / CTC0 Arm | N | Median | Median CI Panel A: TALA + ENZA, CTC conversion | 22 | 32.0 mo | 18.4–NR Panel A: TALA + ENZA, no CTC conversion | 6 | NR | NR–NR Panel A: PBO + ENZA, CTC conversion | 21 | 17.2 mo | 6.5–26.5 Panel A: PBO + ENZA, no CTC conversion | 12 | NR | 15.4–NR Panel B: TALA + ENZA, CTC0 | 22 | NR | NR–NR Panel B: TALA + ENZA, no CTC0 | 26 | NR | 29.3–NR Panel B: PBO + ENZA, CTC0 | 19 | NR | 22.7–NR Panel B: PBO + ENZA, no CTC0 | 24 | 29.5 mo | 15.6–NR HR: TALA+ENZA conversion vs no: 2.81; PBO+ENZA conversion vs no: 2.31; TALA+ENZA CTC0 vs no CTC0: 0.62; PBO+ENZA CTC0 vs no CTC0: 0.72 HR CI: 0.36–21.94; 0.84–6.35; 0.22–1.75; 0.30–1.73 P: p=0.30; p=0.09; p=0.36; p=0.47 (2-sided)

MADRID 2026 ESMO congress 2238P A phase I/II trial of combined cabazitaxel and lutetium-177 PSMA-617 (LuPSMA) in patients (pts) with metastatic castration-resistant prostate cancer (mCRPC): Results of the LuCAB trial L.K. Kostos1, J.P. Buteau2, M. Fahey3, G. Kong2, B. Tran1, P. Jackson2, L. Mcintyre4, M. Haskali5, B. Emmerson2, W. Harris1, L. MacFarlane2, J. Lewin1, D. Murphy6, L. Furic7, M. Crumbaker8, L. Emmett9, M. Hofman2, A.A. Azad1 1 Department of Medical Oncology, Peter MacCallum Cancer Center, Melbourne, Australia, 2 Department of Cancer Imaging, Peter MacCallum Cancer Center, Melbourne, Australia, 3 Department of Biostatistics And Clinical Trials (bact), Peter MacCallum Cancer Center, Melbourne, Australia, 4 Parkville Cancer Clinical Trials Unit, Peter MacCallum Cancer Center, Melbourne, Australia, 5 Department of Radiopharmacy, Peter MacCallum Cancer Center, Melbourne, Australia, 6 Department of Urology, Peter MacCallum Cancer Center, Melbourne, Australia, 7 Department of Research, Peter MacCallum Cancer Centre, Melbourne, Australia, 8 Department of Medical Oncology, St Vincent's Hospital Sydney, Darlinghurst, Australia9 Department of Theranostic and Nuclear Medicine, St Vincent's Hospital Sydney, Darlinghurst, Australia Background Cabazitaxel may improve LuPSMA efficacy by overcoming heterogeneous PSMA expression and enhancing radiation damage as a radiosensitiser. Methods We conducted a single-arm, investigator-initiated phase I/II trial at two centres. Eligible pts had progressive mCRPC, PSMA-PET positive disease (SUVmax ≥15) without discordance on FDG-PET, and prior treatment with docetaxel and an androgen receptor pathway inhibitor. LuPSMA (7.4 GBq) was administered IV 6-weekly (≤6 cycles), concurrently with cabazitaxel IV 3-weekly (≤12 doses), on days 2 and 23 of a 6-week cycle with G-CSF prophylaxis. Four dose levels of cabazitaxel (12.5mg/m2 -20mg/m2) were evaluated using a 3+3 design. The primary objective was to determine the maximum tolerated dose (MTD) and recommended phase II dose (RP2D). Key secondary endpoints included safety, ≥50% PSA response rate (PSA50-RR), PSA- and radiographic progression-free survival (PSA-PFS and rPFS), and overall survival (OS). Results Between 26 Jul 2022 and 13 Feb 2025, 35 eligible pts were enrolled. Median (IQR) age was 68.0 years (63.0 - 75.0), baseline PSA 50.2 ug/L (IQR 15.1 - 109.8), PSMA SUVmean 8.3 (6.5 - 10.2), PSMA volume 493.3 mL (119.0-874.7), and FDG volume 85.8 mL (30.0 - 246.9). 25 pts (71%) had de novo metastatic disease; 10 (29%) had M1c disease at baseline. No dose-limiting toxicities were observed. The MTD and RP2D was 20mg/m2 cabazitaxel 3-weekly with 7.4 GBq LuPSMA 6-weekly. 26 (74%) and 22 (63%) pts received all doses of LuPSMA and cabazitaxel, respectively. Clinically significant grade 3 treatment-related adverse events (TRAEs) occurred in 8 (24%) pts: neutropenia, anemia (both n=2), diarrhea, weight loss, urinary tract infection and fatigue (n=1 each). No febrile neutropenia, grade 4 or 5 TRAEs occurred. The efficacy analysis was conducted in pts treated at the MTD (n=26) with a median follow-up of 20.4 months. PSA50-RR was 73% (95% CI 54-86), the median PSA-PFS was 8.5 (95% CI 4.8-12.1), rPFS 10.9 (10.2-13.4) and OS 16.5 (13.4-21.8) months. Conclusions Combined LuPSMA and cabazitaxel was safe with encouraging anti-tumour activity in a heavily pre-treated cohort of mCRPC pts with unfavourable baseline characteristics. Clinical trial identification NCT05340374.

nature medicine Search Log in Content About Publish articles > article Download PDF Article | Open access | Published: 24 October 2026 Neoadjuvant nivolumab with or without visugromab in muscle-invasive bladder cancer: the phase 2a GDFATHER-NEO trial Andrea Necchi, Chiara Mercinelli, … Felix S. Lichtenegger + Show authors Nature Medicine (2026) Cite this article Save article View saved research Abstract

WUTSUP-02: Conclusions • The perioperative Tislelizumab with neoadjuvant Disitamab Vedotin demonstrates a positive antitumor activity and a manageable safety profile in high risk, non-metastatic UTUC • Higher HER2 expression may be associated with numerically greater response rates, but this exploratory finding requires further validation Yige Bao yige.bao@scu.edu.cn

[Slide 2] Baseline Immune Enrichment Is Associated With Early KIM-1 Decrease in N+I-Treated Patients Key Finding: • N+I-treated KIM-1 decreasers in COSMIC-313 exhibit baseline immune-inflamed signatures (immune cells) • Non-decreasers exhibited immune-depleted profiles Placebo PO QD + nivolumab 3 mg/kg IV Q3W + ipilimumab 1 mg/kg IV Q3W Baseline serum KIM-1 (prior to first dose) -> Week 3 Post-cycle 1 serum KIM-1 ≤30% Decrease | >30% Decrease • Compare using bulk RNA-seq from baseline • Differential gene analysis • Differential pathway analysis [Heatmap: ≤30% Decrease vs >30% Decrease; rows: IMDC; PDL1 TPS; MicroenvironmentScore; ImmuneScore; B-cells; Class-switched memory B-cells; DC; aDC; CD8+ T-cells; CD8+ Tcm; Tregs; CD4+ memory T-cells; Macrophages; Macrophages M1; HALLMARK_ALLOGRAFT_REJECTION; HALLMARK_INTERFERON_GAMMA_RESPONSE; HALLMARK_IL6_JAK_STAT3_SIGNALING; HALLMARK_INFLAMMATORY_RESPONSE; HALLMARK_TNFA_SIGNALING_VIA_NFKB; HALLMARK_COMPLEMENT; HALLMARK_IL2_STAT5_SIGNALING; HALLMARK_KRAS_SIGNALING_UP] z-score; IMDC: Intermediate, Poor; PDL1 TPS: <1%, ≥1% Heatmap downsampled to have equal representation from the two groups

Phase 2 Study 2870-002/SKB264-II-06 (NCT05642780) Key Eligibility Criteria for UC Cohort • Aged ≥18 y • Histologically or cytologically confirmed locally advanced or metastatic urothelial carcinoma of the renal pelvis, ureter, bladder, or urethra(a) • Not eligible for cisplatin-containing chemotherapy(b) • No prior systemic therapy for advanced or unresectable urothelial carcinoma(c) • Measurable lesion per RECIST v1.1 • ECOG PS 0 or 1 → Sac-TMT 4 mg/kg or 5 mg/kg IV Q2W(d) + Pembrolizumab 400 mg IV Q6W; N = 40 Endpoints Primary • Safety, including AEs and AEs leading to treatment discontinuation • ORR per RECIST v1.1 by investigator Secondary • DCR per RECIST v1.1 by investigator • DOR per RECIST v1.1 by investigator • PFS per RECIST v1.1 by investigator • OS [Footnotes a-d too blurry to transcribe reliably: concern mixed histology (urothelial component >50%, plasmacytoid <10%), cisplatin-ineligibility definitions (creatinine clearance <60 but >30 mL/min, grade ≥2 audiometric hearing loss, grade ≥2 peripheral neuropathy, or NYHA Class III heart failure), prior adjuvant/neoadjuvant therapy with recurrence >12 months permitted, and treatment duration (pembrolizumab up to 2 years) - [illegible detail]]

REPRINT Results 13 Patients Enrolled May 2025 – April 2026 4 Excluded Post-Review Central pathological review 9 Evaluable Patients Included in final analysis Diagnostic Exclusion Rate (30.8%) Centralized pathological review reallocated over 30% of initially suspected cases, underscoring the critical need for specific molecular markers.

[Slide 1] Efficacy & Survival Outcomes A [waterfall plot: Change from baseline (%); Best response CR / SD / PR / PD] B [spider plot: Change from baseline (%) at Baseline, After Cycle 2, After Cycle 4; Clinical response CR / SD / PR / PD] C [KM curve: Event-free survival (%) vs Time from enrolment (months)] Events: 8 (32%) Median EFS: Not reached 12-month EFS rate: 72.0 (95%CI 50.1-85.5) 24-month EFS rate: 67.8 (95%CI 45.7-82.4) Number at risk (0/6/12/18/24 mo): 25, 22, 18, 12, 5 D [KM curve: Overall survival (%) vs Time from enrolment (months)] Deaths: 3 (12%) Median OS: Not reached 12/24-month OS rate: 88.0 (95%CI 67.3-96.0) Number at risk (0/6/12/18/24 mo): 25, 23, 22, 16, 7 Best ORR: 84.0% 21/25 pts (2 CR, 19 PR) vs 50% of historical TIP chemo Pathological CR (pCR): 48.0% 12/25 pts vs 10% of historical TIP chemo Median Follow-up: 21.4 mo Median EFS/OS: Not reached 1-yr EFS rate: 72% 1-yr OS rate: 88% Full analysis set (FAS): all enrolled patients receiving at least one dose of study treatment. All efficacy, survival, and safety analyses were performed in the FAS. Survival table Endpoint: EFS and OS (single-arm), FAS Toripalimab + TIP (EFS) | 25 | Not reached Toripalimab + TIP (OS) | 25 | Not reached Landmark rates: 12 mo: EFS 72.0% (CI 50.1-85.5) 24 mo: EFS 67.8% (CI 45.7-82.4) 12/24 mo: OS 88.0% (CI 67.3-96.0) [Slide 2] Exploratory Stratification & Conclusions [KM curve: Event-free survival (%) vs Time from enrolment (months); HPV positive vs HPV Negative] HR: 0.31 (95% CI,0.06–1.58) Log-rank p=0.13 Number at risk (0/3/6/9/12/15/18/21/24 mo): HPV negative (blue): 12, 11, 10, 8, 7, 7, 5, 3, 2 HPV positive (red): 12, 12, 12, 12, 11, 9, 7, 7, 3 [KM curve: Overall survival (%) vs Time from enrolment (months); HPV positive vs HPV Negative] HR: 0.01 (95% CI,0–152.55) Log-rank p=0.07 Number at risk (0/3/6/9/12/15/18/21/24 mo): HPV negative (blue): 12, 12, 10, 9, 9, 9, 7, 5, 4 HPV positive (red): 12, 12, 12, 12, 12, 11, 9, 7, 3 HPV+ vs HPV-: Trend toward favorable outcomes (not statistically significant) • pCR rate: 66.7% vs 33.3%, p=0.22 • 1-yr OS rate: 100% vs 75%, p=0.070 • Neoadjuvant toripalimab + TIP demonstrates encouraging ORR/pCR and manageable safety in La-PSCC. • Longer follow-up and phase III RCTs are required to validate these preliminary findings. One patient did not undergo HPV genotyping and was excluded from these HPV-stratified analyses. Survival table Endpoint: EFS and OS by HPV status (exploratory) HPV positive | 12 HPV negative | 12 HR: EFS 0.31; OS 0.01 HR CI: EFS 0.06-1.58; OS 0-152.55 P: EFS log-rank 0.13; OS log-rank 0.07 Landmark rates: 1 yr: HPV positive (OS) 100% | HPV negative (OS) 75%

[Slide 1] 10:15 - 11:45 Proffered paper: Genitourinary tumours, prostate, penile and testis CHAIRS: CHRISTOPHER J SWEENEY, URSULA VOGL Decipher score is predictive of distant metastasis benefit of abiraterone in STAMPEDE node negative (N0) patients STAMPEDE All: Treatment x Decipher Interaction for DM: p = 0.11 STAMPEDE N0: Treatment x Decipher Interaction for DM: p = 0.01 [Two plots: y-axis 10-year distant metastasis (%) 0/25/50; x-axis Decipher 0.25/0.50/0.75/1.00; curves RT + ADT (solid purple) and RT + ADT + Abi (dashed orange), with Decipher score distribution strip plots per arm] Sarah Howlett [Slide 2] My summary: Important progress towards personalised treatment Strong prognostic validation and an encouraging signal for treatment selection PROGNOSTIC VALUE | Distant metastasis aHR 1.44 per 0.1 Decipher increse 95% CI 1.19-1.76; p < 0.001 REPRODUCED in RWD | Consistent associations in STAMPEDE and real-world cohorts PREDICTIVE SIGNAL | N0 treatment × Decipher interaction for DM: p = 0.01 The data is provocative but needs further validation Nicholas van As

Final pathological and radiological outcomes and initial EFS from a blinded, exploratory phase 2 trial of nivolumab with or without visugromab in MIBC (GDFATHER-NEO) Andrea Necchi1,2, Chiara Mercinelli1,2, Brigida A. Maiorano1, Antonio Cigliola1,2, Roberto Iacovelli3,4, Chiara Ciccarese3,4, Chiara Ligato3,4, Patrizia Giannatempo5, Alessandro Rametta5, Paolo Gontero6,7, Francesco Soria6, Aasim Amin8, Lena Lemke8, Frank Hermann8, Ana Kamova8, Kathrin Klar8, Alexandra Auckenthaler8, Sujata Rao8, Eugen Leo8, Felix S. Lichtenegger8 1. Department of Medical Oncology, IRCCS San Raffaele Hospital, Comprehensive Cancer Center, Milan, Italy; 2. Vita-Salute San Raffaele University, Milan, Italy; 3. Department of Oncology, Fondazione Policlinico Universitario Agostino Gemelli, Rome, Italy; 4. Translational Medicine and Surgery, Università Cattolica del Sacro Cuore, Rome, Italy; 5. Department of Medical Oncology, Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy; 6. Division of Urology, Department of Surgical Sciences, San Giovanni Battista Hospital, Turin, Italy; 7. University of Turin, Turin, Italy; 8. CatalYm GmbH, Munich, Germany Rapid Oral Session I, GU Kidney & Bladder, Saturday October 24, 10:15-11:45 CET

Poster 3656 Enfortumab Vedotin Plus Pembrolizumab for Previously Untreated Locally Advanced or Metastatic Urothelial Carcinoma: Updated Exploratory Analysis of EV-302 in Older Patients and Those With Comorbidities Srikala S. Sridhar, Michiel van der Heijden, Thomas Powles, Jens Bedke, Begona Perez Valderrama, Jean Hoffman-Censits, Christof Vulsteke, Gopakumar Iyer, Steffen Rausch, Evan Y. Yu, Ignacio Duran, Se Hoon Park, Jian-Ri Li, Eiji Kikuchi, Blanca H[illegible], Jasmine Lichfield, Xuesong Yu, Dana Wishengr[illegible] Patients and Baseline Characteristics • At data cutoff (October 6, 2025), median follow-up in the ITT population was 42.8 months (95% CI, 41.7-43.9) • Select baseline characteristics are shown in Table 1 Table 1. Key Demographic and Baseline Disease Characteristics Across Analyzed Subgroups Figure 2. OS in Subgroups of Interest A. Age ≥75 years B. BMI ≥30 kg/m2 C. Diabetes D. GFR <60 mL/min Table 2. Summary of TRAEs in Subgroups of Interest Table 3. Treatment-Related AEs of Special Interest for EV [remaining poster text illegible at this resolution]

[Left slide] Exploratory Stratification & Conclusions EFS by HPV status — HPV positive / HPV negative — HR 0.31 (95% CI 0.08–1.28); Log-rank p=0.10 Y: Event-free survival (%); X: Time from enrollment (months) 0,6,12,18,24 Number at risk: HPV+ 12 11 10 8 6; HPV− 13 9 6 4 3 OS by HPV status — HPV positive / HPV negative — HR 0.42 (95% CI 0.09–1.96); Log-rank p=0.24 Y: Overall survival (%); X: Time from enrollment (months) Number at risk: HPV+ 12 12 11 9 7; HPV− 13 11 8 6 4 HPV+ vs HPV−: Trend toward improved outcomes in HPV+ disease (not statistically significant) • pCR rate: 66.7% vs 33.3%, p=0.22 • 1-yr OS rate: 100% vs 75%, p=0.070 • Neoadjuvant toripalimab + TIP demonstrates encouraging ORR/pCR and manageable safety in La-PSCC. • Longer follow-up and phase III RCTs are required to validate these preliminary findings. One patient did not undergo HPV genotyping and was excluded from these HPV-stratified analyses. [Right slide] Safety & Postsurgical Complications Safety (CTCAE 5.0) • Any Grade: 100% (25/25) • Grade 3–4: 84% - Neutropenia: 72% - Febrile neutropenia: 8% - Hypothyroidism: 8% (Grade 1–2) • No patient discontinued treatment due to toxicity Postsurgical Complications (mClavien–Dindo) • 21 pts received consolidative surgery • 20 pts underwent simultaneous ILND + PLND • Any grade: 85.7% (18/21) - Lymphatic fistula: 15 pts, Grade I / II - Pneumonia: 1 pt Grade IVa, 1 pt Grade V * HPV status (N=25) • HPV positive: 12 pts (48%) • HPV negative: 13 pts (52%) HPV status and surgical outcomes • Complication rate (any grade): 83.3% (10/12) vs 87.5% (7/8) • No clear difference by HPV status * aspiration death (POD 30) in a patient with preexisting schizophrenia, after self-administered alprazolam/clozapine during a manic episode; unrelated to treatment. Survival table Endpoint: EFS and OS by HPV status (exploratory), neoadjuvant toripalimab + TIP, La-PSCC HPV positive | 12 HPV negative | 13 HR: 0.31 (EFS, HPV+ vs HPV−) HR CI: 0.08–1.28 P: 0.10 (log-rank) Landmark rates: 1 yr: HPV positive OS 100% | HPV negative OS 75%

[Left slide] Response Rates in Patients with GCTa,b [Waterfall plot] Y-axis: Best change from baseline, % [partially obscured]; X-axis: Patient ID: OC04 T01 T06 T04 T03 T07 T08 T02 T05 S07 Legend: 0.4 mg/kg; 1.1 mg/kg; 2.2 mg/kg; 2.4 mg/kg; 3.0 mg/kg; 4 mg/kg; Ovarian tumor; Mediastinal tumor; Received prior HDCT/ASCT; Tx ongoing Parameterc | All dose levels (N=11) BOR, n (%) PR 5 (45.5) SD 3 (27.3) PD 2 (18.2) NE 1 (9.1) ORR, % 45.5; 95% CI (16.7-76.6) DCR, % 72.7; 95% CI (39.0-94.0) mPFS, months 5.8; 95% CI (1.4-13.8) • Responses were observed at Ixo-V dose levels ranging from 1 mg/kg to 3.4 mg/kg, with a median follow-up of 13.8 months • Three of 6 (50.0%) patients who received prior HDCT/ASCT had confirmed responses to Ixo-V • One patient with testicular GCT has maintained a response for 47.4 months with treatment ongoing at time of data cutoff Ixo-V monotherapy exhibits preliminary antitumor clinical activity at doses ≥1 mg/kg in heavily pretreated patients with recurrent GCT, including those with disease refractory to ASCT BOR, best overall response; DCR, disease control rate; HDCT/ASCT, high-dose chemotherapy with autologous stem cell transplant; Ixo-V, ixotatug vedotin. aData cutoff was March [illegible, possibly 16], 2026. bPrimary sites for the 11 GCT patients include: testicular (n=9), ovarian (n=1) and mediastinal (n=1). [Right slide] Conclusions • Ixo-V monotherapy demonstrates meaningful clinical activity in heavily pretreated patients with recurrent GCT, including patients with disease refractory to HDCT/ASCT • Ixo-V has a manageable safety profile, characterized by predominantly low-grade TEAEs • These phase 1 data support ongoing clinical investigation of Ixo-V in GCT that have progressed on or after platinum-based chemotherapy GCT, germ cell tumors; HDCT/ASCT, high-dose chemotherapy with autologous stem cell transplant; TEAEs, treatment emergent adverse events

A neoantigen vaccine for kidney cancer Trial overview Tumor procurement → Target selection → Personal vaccine manufacture → Vaccine administration ccRCC Stage III-IV; No evidence of disease (NED) after surgery WES, RNAseq → Neoantigen prediction (HLAthena) → Identification of up to 20 personal neoantigens Braun et al, Nature, 2025.

[Slide 1] 13:30 - 15:00 Rapid oral: Genitourinary tumours, prostate, penile and testis CHAIRS: THOMAS HWANG, EDMOND KWAN, ALISON REID Background PR21 trial Study schema Key Eligibility - mCRPC post ARPI - Chemotherapy naive - Except prior Docetaxel for mHSCP allowed if >= 12 months prior - PSMA-PET positive (uptake > liver) - Excluded if < 50% of extra-osseous lesions or < 5 cm soft tissue lesion were PSMA negative Randomize 1:1 177Lu-PSMA-617 7.4 GBq IV every 6 weeks x 6 cycles | Docetaxel 75 mg/m2 IV every 3 weeks up to 12 cycles Baseline FACT-P; FACT-P every 6 weeks x 6; FACT-P every 12 weeks Radiographic Progression 1 -> crossover: Docetaxel 75 mg/m2 IV every 3 weeks up to 12 cycles | 177Lu-PSMA-617 7.4 GBq IV every 6 weeks x 6 cycles; FACT-P; FACT-P every 6 weeks x 6 Radiographic Progression 2 Primary Endpoint: rPFS 1 Additional Endpoints: rFPS 2, OS, Quality of Life* *Trials Outcomes Index (TOI) was a pre-specified Quality of Life outcome of interest. rPFS1(1) [small KM inset: 177Lu-PSMA-617 vs Docetaxel; Time in Months] HR 1.01 (90% CI 0.77-1.31) 1-sided p = 0.51 Presenter: Dr. Daniel J. Khalaf (1)Kim N. Chi et al, ESMO Congress 2025 [Speaker video inset] Daniel Khalaf Patient reported quality of life (QoL) for docetaxel (Doc) vs lutetium177-617-PSMA (LuP) for metastatic castration-resistant prostate cancer (mCRPC): Results from the Canadian cancer trials group (CCTG) PR21 trial BILBAO AUDITORIUM - HALL 3 [Slide 2] 13:30 - 15:00 Rapid oral: Genitourinary tumours, prostate, penile and testis CHAIRS: THOMAS HWANG, EDMOND KWAN, ALISON REID Results Study Population and Questionnaire Completion rates Randomized, N=199 n=100 -> 177Lutetium-PSMA-617 Received Protocol therapy: n = 94 n=99 -> Docetaxel Received protocol therapy: n = 95 Baseline questionnaires: LuP Baseline 91 received / 94 expected (96.8%); Docetaxel 91 received / 95 expected (95.8%) On-treatment questionnaire completion rates: LuP: Week 6 79.3%; Week 12 81.0%; Week 18 84.3%; Week 24 80.6%; Week 30 74.1%; Week 36 61.2% Docetaxel: Week 6 89.2%; Week 12 86.0%; Week 18 70.1%; Week 24 75.8%; Week 30 59.0%; Week 36 32.8% Mean baseline FACT-P Scores [bar chart, Lu-P vs Doce; categories Total, TOI, FACT-G, PWB, FWB, PCS, EWB, SWB; numerical score axis 0-120; bar values not printed] Presenter: Dr. Daniel J. Khalaf [Speaker video inset] Daniel Khalaf Patient reported quality of life (QoL) for docetaxel (Doc) vs lutetium177-617-PSMA (LuP) for metastatic castration-resistant prostate cancer (mCRPC): Results from the Canadian cancer trials group (CCTG) PR21 trial BILBAO AUDITORIUM - HALL 3 [Slide 3] 13:30 - 15:00 Rapid oral: Genitourinary tumours, prostate, penile and testis CHAIRS: THOMAS HWANG, EDMOND KWAN, ALISON REID Results MCID* Response analysis FACT-P Total Score (Improved % / Stable % / Worse % / Missing %) Week 12: LU-P 28 / 30 / 28 / 14; DOCE 22 / 36 / 31 / 10 Week 24: LU-P 29 / 34 / 18 / 19; DOCE 15 / 27 / 32 / 26 Trials Outcomes Index (TOI) (Improved % / Stable % / Worse % / Missing %) Week 12: LU-P 16 / 51 / 21 / 12; DOCE 12 / 50 / 29 / 9 Week 24: LU-P 13 / 58 / 10 / 19; DOCE 6 / 42 / 26 / 26 % patients Improved indicates increase from baseline by >= the MCID* Worse indicates decrease from baseline by >= the MCID MCID for Total FACT-P : 9 points MCID for TOI: 7 points *MCID: Minimal Clinically Important Difference Presenter: Dr. Daniel J. Khalaf [Speaker video inset] Daniel Khalaf Patient reported quality of life (QoL) for docetaxel (Doc) vs lutetium177-617-PSMA (LuP) for metastatic castration-resistant prostate cancer (mCRPC): Results from the Canadian cancer trials group (CCTG) PR21 trial BILBAO AUDITORIUM - HALL 3 [Slide 4] 13:30 - 15:00 Rapid oral: Genitourinary tumours, prostate, penile and testis CHAIRS: THOMAS HWANG, EDMOND KWAN, ALISON REID Results Mean change over time FACT-P Total Score(T) [line chart, LuP vs Doc, mean change from baseline, weeks 6-36; * at weeks 24, 30, 36] * Indicates p<0.05 Mixed Model Assessment week x treatment arm interaction: p = 0.002 Trial Outcomes Index (TOI)(T) [line chart, LuP vs Doc, weeks 6-36; * at weeks 24, 30] (T)Higher scores indicate better QOL Mixed Model Assessment week x treatment arm interaction: p < 0.001 Presenter: Dr. Daniel J. Khalaf [Speaker video inset] Daniel Khalaf Patient reported quality of life (QoL) for docetaxel (Doc) vs lutetium177-617-PSMA (LuP) for metastatic castration-resistant prostate cancer (mCRPC): Results from the Canadian cancer trials group (CCTG) PR21 trial BILBAO AUDITORIUM - HALL 3

a Assessed for eligibility n = 50 September 2023 to March 2025 Excluded (n = 19) Did not meet eligibility criteria (n = 17) Other reasons (n = 2) Enrolled and treated n = 31 Safety population Group A, n = 15; group B, n = 16 Excluded from efficacy analyses (n = 4) Group A: withdrawal, n = 1; dosing error, n = 1 (both excluded from both analyses) Group B: late withdrawal before RC, n = 1 (excluded from pathological analysis) Group B: not RECIST evaluable, n = 1 (excluded from radiographic analysis) Local pathological response population (pCR, MPR) n = 28 Group A, n = 13 Group B, n = 15 Local radiographic response population (ORR per RECIST 1.1) n = 28 Group A, n = 13 Group B, n = 15 b Eligible population Stage T2–T4a N0 M0 muscle-invasive bladder cancer; ineligible for or declining cisplatin-based chemotherapy; pure or predominant (≥50%) urothelial histology TURBT Baseline tumour biopsy, TNM staging, PD-L1 status (CPS) Sequential treatment assignment (1:1) Stratified by disease stage (T2 versus T3–T4a) and PD-L1 CPS (<10 versus ≥10) Alternating allocation (A→B→A→B...) within strata Participants per stratum Stratum | A | B T2/CPS <10 | 5 | 5 T2/CPS ≥10 | 4 | 5 T3–T4a/CPS <10 | 2 | 2 T3–T4a/CPS ≥10 | 4 | 4 Total | 15 | 16 Group A (n = 15) Nivolumab 480 mg i.v. + placebo i.v. Three cycles, every 4 weeks 4–8 weeks after the last trea[tment] [obscured by QR code] Restaging CT (RECIST v1.1) Radi[cal cystectomy] (RC) [obscured]

[Slide 2] [Photo of presenter beside poster; sign: POSTER AREA; board 3599P] #3599P Plasma Proteomics Reveals the Biological Landscape of Tyrosine Kinase Inhibitors (TKI) Sensitivity in Metastatic Renal Cell Carcinoma (mRCC) Liliana Ascione, Clara Steiner, Coren Lahav, Yehonatan Elon, Adam P. Dicker, Michal Harel, Marc Eid, Renee Maria Saliby, Karl Semaan, Eddy Saad, Razane El Hajj Chehade, Marc Machaalani, Mustafa Saleh, Wassim Daoud Khatoun, Pablo Barrios, Rashad Nawfal, Francois-Xavier Danlos, Laurence Albiges, Toni K. Choueiri, Wenxin Xu [affiliation superscripts] 1Dana-Farber Cancer Institute, Boston, MA, USA; European Institute of Oncology (IEO), IRCCS, Milan, Italy; 3 Department of Urology, University Hospital Leipzig, Leipzig, Germany; 4OncoHost, Binyamina, Israel; 5Thomas Jefferson University, Philadelphia, PA, USA; 6Institut Gustave Roussy, Villejuif, France; 7Harvard Medical School, Boston, MA, USA Background - We previously developed a plasma proteomic signature (TKI-POS) by selecting circulating proteins associated with response to TKI therapy in RCC. - This PRO-TKI score was associated with prolonged OS after treatment with TKI-containing regimens, versus immune-checkpoint inhibitors-only regimens, in an independent validation cohort. - A bioinformatic analysis of the model proteins was performed to characterize the biological mechanisms underlying the plasma proteins most predictive of TKI benefit. Methods [Schema: Pre-treatment Plasma Sample -> Aptamer-based Proteomic Assay -> Computational Analysis -> TKI-positive / TKI-negative] - Pre-treatment plasma proteomic data from 178 mRCC were generated using SomaScan 7K. - During the TKI-POS model development, proteins were selected from the 7K panel. - Protein localization and expression in different tissues were obtained from publicly available datasets. - Fisher's exact test (FDR<0.1) was used to identify significantly enriched functional groups. Figure A: Venn diagram - Intracellular 152; Membranal 29; Secreted 84; overlaps 52, 48, 14, 9 Figure B: protein network (legend: Angiogenesis; Blood vessel; Blood vessel morphogenesis; Endothelium development; HIF-1 signaling pathway; PDGF/VEGF domain; Regulation of angiogenesis; Signaling by PDGF; Signaling by VEGF; VEGFA-VEGFR2 signaling; VEGF ligand-receptor interactions; Cell adhesion; Cell-cell signaling; Extracellular matrix; Immune system process; Inflammatory response) Figure C: enrichment terms - Prognostic unfavorable in RCC; HoC: Inducing angiogenesis; NMF2 (angiogenesis; FA oxidation); Secreted; Extracellular region part; ECM core; NMF5 (cell cycle; FA synthesis); HoC: Tumor promoting inflammation; NMF3 (Complement; O oxidation); HoC: Replicative immortality; NMF1 (angiogenesis; stroma); Lipid transport; NMF7 (snoRNA) Figure A. Protein localization Venn diagram based on Human Protein Atlas data; B. Protein network of angiogenesis-related proteins; C. Enrichment analysis of the core proteins in the model Results - TKI-POS model is comprised of 416 proteins (393 non-redundant proteins). 169 proteins are considered as core predictive proteins, as they had higher weight in the model (i.e., they were selected >=10 times during model development). - Protein localization analyses support a composite circulating signal derived from tumor, immune and host stromal compartments (Figure A). - Among the 416 model proteins, 82 were associated with angiogenesis and vascular remodeling. - Additional proteins associated with TKI response were linked to tumor plasticity (e.g., B-catenin), inflammation (e.g., IL-8 and IL6), endothelial biology (e.g., PLXDC1 and PLXND1), and extracellular matrix remodeling (e.g., THSB2 and THSB4); Figure B. - Enrichment analysis further demonstrated significant overrepresentation of pathways related to angiogenesis, tumor-promoting inflammation, extracellular matrix remodeling, lipid transport, and poor prognosis in RCC (Figure C). Conclusions Integrative bioinformatic an[...] circulating proteomic model[...] response elucidates the [...] of TKI sensitivity in m[...] strategies to o[...] Prospective val[...] independe[...] [partly obscured by presenter] Liliana Ascione, MD | liliana.ascione@ieo.it; Toni K. Choueiri | toni_choueiri@dfci.harvard.edu; Wenxin (Vincent) Xu | wenxin_xu1@dfci.harvard.edu *Liliana Ascione has no conflicts of interest to declare. [Logos: Dana-Farber Cancer Institute, Gustave Roussy, OncoHost] [Slide 3] [Photo of two presenters (badges: Paloma Cejas; Joaquim Bellmunt) in front of poster; board 4743; much of poster obscured] Functional epigenomic liquid biopsy to guide treatment at EV/P resistance in metastatic urothelial carcinoma (mUC): Emerging role of ERBB3 Bellmunt J1, Li R2, El Sarout S1, Xie Y2, Gulati G1, Epstein IB1, Kukreja S2, Rodriguez L2, Berg SA1, Mantia Ch1, McGregor B1, Kaymakcalan1, Nawfal R1, Baca SC1,2, Choueiri TK1, Freedman ML1, Long HW1,2, Cejas P1,2,3 1. Dana-Farber Cancer Institute, Brigham and Women's Hospital, Boston, MA, USA. 2. Center for Functional Cancer Epigenetics, Dana-Farber Cancer Institute, Boston, MA 02215, USA. 3. Translational Oncology Laboratory, Hospital La Paz Institute for Health Research (IdiPAZ), Madrid, Spain BACKGROUND - There is no validated strategy to guide personalized treatment selection following 1st line EV/P failure in mUC - Liquid biopsy approaches enable real-time functional tumor profiling and may identify emerging actionable targets to inform rational treatment sequencing - We developed an epigenomic chromatin-based liquid biopsy platform interrogating tumor-derived circulating chromatin using H3K4me2 to capture active enhancer[s] and promoters METHODS 48 Patients: 7 CR; 20 PR; 11 SD; 10 PD Baseline (Pre-treatment); Mid (During treatment); Post (After treatment) - Liquid Biopsy Step 1 Liquid Biopsy; Step 2 Chromatin IP; Step 3 Library prep; Step 4 Sequencing; Step [5] An[alysis] Fig 1. Study design and w[orkflow] RESULTS - Forty-eight pts contributed 96 longitudinal plasma samples. - Best responses included: CR 7 pts, PR 22, SD 9, and PD 10 - No baseline chromatin features significantly correlated with response, although responders showed a trend toward higher [...] [...]reme phenotypes (CR vs PD) demonstrated [...] of cancer-specific regulatory elements at PD, [...]B3, which was consistently upregulated at PD [...]atory activity was not enriched in refractory [...] [...]ts included ERCC2 and CD276, while [...] of FGFR3 and ERBB2 regulatory activity [...]ding known somatic alterations [...]TD2, WEE1, and PPARG [...] [Figure 2b ERBB3 normalized RPM Pre vs Post by patient (P37, P38, P18, P22, P48, P32, P4...); radar plots Patient P48 (PR) and Patient P26 (PD) with ERCC2, ERBB3, CD276, WEE1, BRCA2 axes; heatmap of actionable genes - partly obscured] [...]emodeling and emergent therapeutic vulnerabilities during EV/P [...]-fraction-corrected signal increase in 4/5 patients with PD Abstract #4743P ESMO 2026 [Slide 4] [Photo of two presenters beside poster; board 2265P] FPN: 2265P ESMO Congress 2026 Baseline and dynamic changes in circulating tumor DNA (ctDNA) as correlates of intrinsic and acquired resistance to 177Lu-PSMA-617 (LuPSMA) in metastatic castration-resistant prostate cancer (mCRPC) Praful Ravi1, Guanglong Jiang2, Chao Dai2, Brady James1, Megan Lee1, Hailey A. Stoltenberg1, Gwo-Shu Mary Lee1, Heather Jacene1, Jacob E. Berchuck3, Sylvan Baca1, Yong Huang2, Pan Du2, Shidong Jia2, Alok K. Tewari1 1Dana-Farber Cancer Institute, Boston, MA, USA; 2Predicine Inc., Hayward, CA, USA; 3Winship Cancer Institute of Emory University, Atlanta, GA, USA [Logos: Dana-Farber Cancer Institute; Harvard Medical School] Background - Intrinsic or acquired resistance to LuPSMA is seen in nearly all patients (pts) with mCRPC. - Biomarkers to select pts for LuPSMA monotherapy, combination strategies or alternative therapies are lacking. - Serial ctDNA profiling before, during and at progression (PD) on LuPSMA may build on prior publications to elucidate genomic and epigenomic associations with resistance. Methods - mCRPC pts treated with >=2 cycles of LuPSMA at Dana-Farber Cancer Institute (Jul 2022 - Aug 2024) with research plasma at cycle 1 (C1D1) and at C3D1 and/or PD. [LuPSMA treatment cycles C1-C6; C1D1 (n=63); C3D1 (n=50); PD (n=41); PredicineCARE: Targeted NGS: SNVs, CNVs, fusions; PredicineSCORE: LP-WGS: CNB score, tumor fraction (TF), ARBS nucleosome profiling] [Clinical outcome definitions table partly obscured: ...response Best PSA decline >90% 17/72; ...tional response PSA90 or OS >18 mo 24/72; ...c resistance Clinical/radiographic PD <=3 mo or OS <=9 mo 25/72; ...d resistance PD after 3 mo 54/72] [Stat tiles: 72 treated LuPSMA; 63 with baseline ctDNA NGS; 154 LP-WGS samples (52 pts); 35% intrinsic resistance] Median age 71 yrs; median PSA 41 ng/mL; visceral metastases 33%. Self-reported Race: 60 White, 6 Black, 4 Hispanic, 1 Asian, 1 N/A Results: Tumor Fraction & ARBS Baseline ctDNA tumor fraction is prognostic [KM: p < 0.0001; table Continuous TF: OS HR 25.6 P 2.6x10^-x; CR-PFS 6.5, 2.3x10^-x; PSA-PFS 6.4, 5.3x10^-x - exponents illegible] Figure 2. (Left) OS by baseline TF >2% vs <=2%. (Right) Baseline combo TF by response group (Wilcoxon). *Cox HR per unit TF (0-1). - TF >2% (n=46) vs <=2% (n=17): shorter OS (log-rank p<0.0001). - Mean TF lower in exceptional responders (0.08 vs 0.33; P=1.3x10^-x) and higher with intrinsic resistance (0.41 vs 0.15; P=1.7x10^-x). Integrated Analysis of PSA, tumor fraction and AR binding sites during LuPSMA treatment Figure 3. (A) ARBS abnormality score vs CNV-inferred TF (R=0.81); ... 105/154 samples (68%) CNB-positive; 78/154 (51%) AR-positive by ARBS profiling. In pts 54 and 55, TF and ARBS abnormality fell at C3D1 and rebounded at PD; in pt 54, PSA remained below baseline. Results: Intrinsic Resistance Baseline alterations in the 9p21 locus (MTAP, CDKN2A/B), the PI3K/AKT pathway, AR gain, MDM2 gain and BRCA2 were enriched in pts with intrinsic resistance. [Forest table, Gene | Resistant (n=25) | Non-resist. (n=38) | OR (95% CI) | FDR - e.g. MTAP 11 (44%) | 1 (3%) | 27.5 (3.4->400) | 0.007; CDKN2B 11 (44%) | 2 (5%) | 13.5 (2.5-140) | 0.007; remaining rows small, partially legible] Figure 4. Baseline gene alterations (SNV/CNV/fusion) associated with intrinsic resistance (all FDR <0.025; n=63). - PTEN alteration (OR 0.08, FDR 0.034) and MYC gain (OR 0.09, FDR 0.040) were less frequent in exceptional responders. - Associations are univariable and not adjusted for TF; many CNV associations may reflect higher tumor burden. Results: Acquired Resistance Of 41 pts with PD samples, 18 had treatment-emergent or clonally enriched alterations, spanning TP53, DNA repair, PI3K/RAS/MAPK, WNT and AR-axis genes. Figure 5. Alterations emergent or enriched (VAF increase >10%) at PD vs [...] patient. - Emergent AR L702H; BRCA2, POLE, MRE11 and P[...] alterations in 5 pts; TP53 in 6 pts. - 4 clones declined >10% at PD (TP53 x2, RB1, RA[...]) Conclusions - Baseline ctDNA tumor fraction is strongly prognostic o[...] identifies pts with markedly shorter OS. - Baseline 9p21 (MTAP/CDKN2A/B) loss, PI3K/AKT alte[...] MDM2 gain and BRCA2 alterations are associated wi[th intrinsic] resistance. - Acquired resistance is genomically heterogeneous, wi[...] suggests AR signaling is retained at PD in most patie[nts]. - Integrating ctDNA with clinical features may help sele[ct ...] combinations or alternative therapies; validation in la[rger cohorts is] warranted. - Longitudinal ctDNA tumor fraction dynamics and AR [...] molecular insights beyond PSA alone during prosta[...] - Plasma cfRNA analysis is ongoing. Acknowledgements: Study sponsored in part by Predicine Inc. Disclosure: The presenting author has no relevant disclosures

Background • Adjuvant pembrolizumab is a standard of care for patients with clear cell renal cell carcinoma (ccRCC) at increased risk of recurrence postsurgery based on significantly improved disease-free survival (DFS) and overall survival (OS) versus placebo in the primary results of the KEYNOTE-564 trial (NCT03142334)1 • With a minimum of 5 years of follow-up, adjuvant pembrolizumab continued to show improved DFS and OS versus placebo; this benefit was also observed across key protocol-defined risk subgroups2 • We present updated data from the final protocol-prespecified analysis of KEYNOTE-564 with 7 years of follow-up 1. Choueiri TK et al. N Engl J Med. 2021;385:683-694. 2. Haas NB et al. J Clin Oncol. 2025;43:4514.

CRPC-FS Deep PSA response 88m Non-deep response 13.3m p<0.001 Survival probability Time (months) Number at risk CRPC-FS Multivariate analysis PSA<0.2 p<0.001 Stage IV at diagnosis High volume p 0.009 ADT + D vs ADT p 0.03 ADT + ARPI vs ADT p 0.02 Triplet vs ADT p 0.036 Hazard Ratio (95% CI) OS Deep PSA response 83.5m Non-deep response 31.5m p<0.001 OS Multivariate analysis PSA<0.2 p<0.001 Stage IV at diagnosis High volume p<0.001 ADT + D vs ADT ADT + ARPI vs ADT Triplet vs ADT Hazard Ratio (95% CI) Survival table Endpoint: CRPC-free survival and overall survival by PSA response Deep PSA response (CRPC-FS) | 390 | 88m Non-deep response (CRPC-FS) | 211 | 13.3m Deep PSA response (OS) | 392 | 83.5m Non-deep response (OS) | 211 | 31.5m P: p<0.001 (both CRPC-FS and OS)

[Slide 1] Low-, intermediate- or high-risk localised prostate cancer Unsuitable (or unwilling) for radical treatment or active surveillance → WW with delayed ADT [I, A] Low risk → Active surveillance [I, A] Favourable intermediate risk → Active surveillance [II, A] | RP [I, B] | SBRT [I, A] Brachytherapy [I, B] EBRT [I, B] Unfavourable intermediate risk → RP [I, B] | EBRT ± brachytherapy [I, B] → ADT for 4-6 months [I, A] High-risk localised or locally advanced disease → RP [I, B] RP + PLND [III, B] | EBRT ± brachytherapy [I, B] → ADT for 18-36 months [I, A] Monitor PSA levels [II, A] Undetectable PSA → Surveillance [I, A] | pN1 → If unfavourable or high-risk: RT ± ADT [II, B] Confirmed persistent PSA → M0 at restaging → Salvage RT ± ADT for 6-24 months [II, B] Ann Oncol. 2026 Apr;37(4):453-469. [Slide 2] Very high-risk or cN1 prostate cancer Very high-risk localised disease → WW with delayed ADT [I, A] | EBRT ± brachytherapy to prostate + ADT (3 years) + abiraterone (2 years) [I, A] | RP [I, B] RP + PLND [III, B] cN1 disease (by conventional imaging) → EBRT ± brachytherapy to prostate + ADT (3 years) + abiraterone (2 years) [I, A] Monitor PSA levels [II, A] Undetectable PSA → Surveillance [I, A] | pN1 → Surveillance [I, A] | RT ± ADT [IV, C] Confirmed persistent PSA → M0 at restaging → Salvage RT ± ADT for 6-24 months [II, B] Ann Oncol. 2026 Apr;37(4):453-469. STAMPEDE pts [Slide 3] T recovery to ≥200 ng/dL: Median 20.7 months (~9 months from EOT) APA + ADT Median not reached RP + PLND alone Median 19.1 months 63.0% (95% CI, 52.2-71.9) 41.4% (95% CI, 35.6-47.1) HR, 0.38a 95% CI, 0.26-0.55 p<0.0001b Event-Free Survival (%) Months from Randomization No. at risk APA + ADT 102 99 96 91 82 71 38 10 RP + PLND alone 306 195 169 146 131 121 74 15 Survival table Endpoint: Event-free survival APA + ADT | 102 | NR RP + PLND alone | 306 | 19.1 HR: 0.38 HR CI: 0.26-0.55 P: <0.0001 36 mo: 63.0% (95% CI, 52.2-71.9) APA + ADT vs 41.4% (95% CI, 35.6-47.1) RP + PLND alone [Slide 4] The PACE umbrella Localised prostate cancer CI Dr Nick van As The Royal Marsden Cancer Charity PACE A 234 → Surgery | 5 days SBRT PACE B 874 → 4 weeks | 5 days PACE C 1182 patients → 4 weeks and ADT | 5 days and ADT PACE C CI Alison Tree

Can We Emulate a KIM-1-Guided Precision Medicine Trial Testing the Benefit of Adding a TKI in Patients With IO-Resistant ccRCC? Hypothetical Trial Design Metastatic ccRCC -> Biomarker sampling (baseline) -> Nivolumab + ipilimumab -> Biomarker sampling (Day 21) KIM-1 ≤30% decrease (~70% of patients) -> R -> Zanzalintinib* + nivolumab / Nivolumab + ipilimumab (Improved outcomes?); mPFS <8 months KIM-1 >30% decrease (~30% of patients) -> Nivolumab + ipilimumab; mPFS >70 months COSMIC-313 In Silico Data Metastatic ccRCC -> Predict KIM-1 nivolumab + ipilimumab suboptimal decrease based on transcriptomic model -> R -> Cabozantinib + nivolumab + ipilimumab / Nivolumab + ipilimumab Can we predict the benefit of adding TKI to IO among patients who are unlikely to have a ≤30% decrease in KIM-1 on ipi/nivo? • What is the biology of KIM-1 decreasers? • Can predictive modeling be used to predict patients who will have IO-based KIM-1 decrease across treatment arms in COSMIC-313? *Zanzalintinib is a novel, oral small-molecule inhibitor of multiple kinases, including the TAM kinases (TYRO3, AXL, and MER), MET, and VEGF receptors, which are involved in oncogenic processes including tumor cell proliferation, metastasis, angiogenesis, drug resistance, and evasion of antitumor immunity. ccRCC, clear cell renal cell carcinoma; KIM-1, kidney injury molecule-1; IO, immuno-oncology; mPFS, median progression-free survival; R, randomized; TKI, tyrosine kinase inhibitor.

[Slide 2] Figure 1A EFS, reTURBT group (N = 53) Figure 1B OS, reTURBT group (N = 53) TURBT only Event-free survival (%) / Overall survival (%) Time from therapy (months) Number at risk Cumulative number of censoring 36-mo EFS 66.7% (95% CI 54.5–81.7); MFS 65.9% (50.9–85.4) 36-mo OS 82.7% (95% CI 72.4–94.6) Survival table Endpoint: EFS, MFS and OS (reTURBT group) reTURBT (TURBT only) | 53 36-mo EFS 66.7% (95% CI 54.5–81.7) 36-mo MFS 65.9% (50.9–85.4) 36-mo OS 82.7% (95% CI 72.4–94.6) [Slide 3] Figure 2A EFS by pathological response Figure 2B OS by pathological response TURBT no RD, TURBT RD, RC pCR, RC no pCR Event-free survival (%) / Overall survival (%) Time from therapy (months) Number at risk Cumulative number of censoring 24-mo EFS, reTURBT vs RC: pCR 87.6% vs HR 1.96 (0.53–7.27); non-pCR 19.3% vs 67.8% 24-mo OS, reTURBT vs RC: pCR 96.7%, HR 0.90 (0.10–8.40); non-pCR 50.4% vs 82.5% TURBT no RD / RD: ypT0Nx / residual disease at reTURBT; RC pCR: ypT0N0. Log-rank p<0.001 across the four groups for EFS, MFS and OS. Survival table Endpoint: EFS and OS by pathological response TURBT no RD | 35 TURBT RD | 18 RC pCR | 72 RC no pCR | 141 HR: pCR reTURBT vs RC: EFS HR 1.96; OS HR 0.90 HR CI: EFS 0.53–7.27; OS 0.10–8.40 P: Log-rank p<0.001 across the four groups 24-mo EFS pCR reTURBT 87.6%; non-pCR 19.3% vs 67.8% 24-mo OS pCR reTURBT 96.7%; non-pCR 50.4% vs 82.5% [Slide 4] Figure 2C Swimmer plot, reTURBT group (N = 53) Patient ID Time (months) ypT0 status: Non-ypT0, ypT0 Events: TURBT, Delayed RC, Delayed RT, Relapse, Maintenance therapy, Death

[Slide 2] Potential barriers in ccRCC Mosele et al. Nat Med (32)2774-2785, 2026 1. Hypervascularity Ü uniform ADC delivery Yuan Q. et al. Clin Genitourin Cancer 14(6); e585-e594, 2016 • Intratumoral perfusion is spatially heterogeneous even in hypervascular ccRCC. • Heterogenous perfusion may result in spatially uneven ADC delivery to tumor cells. [Slide 4] BARRIER 1: Delivery Can the ADC reach the tumor cell? Pharmacokinetics | Tumor penetration | Target recognition | Intracellular trafficking | Payload | TME modulation Mosele et al. Nat Med (32)2774-2785, 2026 Potential barriers in ccRCC ADC delivery = Balance between Perfusion and Permeability Antibody Drug Conjugates tumor [g] Tumor Vascular density [1/cm] Vessel wall pore radius [nm] ×10^-8 Harkos C. et al. J Immunother Cancer 2026;14:e015357 Functional vascular density Amount of perfused, functional vessels High functional density Low functional density • Determines how much ADC can be delivered to the tumor. • Heterogeneous and inefficient perfusion in ccRCC. Vessel wall pore radius Vascular permeability for macromolecules Large pore radius (≥ ~40 nm) Small pore radius (< ~40 nm) • Controls ADC extravasation across the vessel wall. • Excessive normalization may reduce permeability. Optimal ADC delivery requires both sufficient functional vessel density and adequate vascular pore size.

[Slide 1] Perioperative Study population • Adults • Cisplatin-eligible patients with MIBC (cT2–T4aN0/1M0) • UC or UC with DDHS • Evaluated and confirmed for RC • CrCl of ≥40 mL/min R 1:1 Neoadjuvant Q3W, 4 cycles; Adjuvant Q4W, 8 cycles D arm N=533: D 1500 mg IV + Gemcitabine + cisplatin → RC → D 1500 mg IV C arm N=530: Gemcitabine + cisplatin → RC → No treatment Dual primary endpoints • pCR and EFS Key secondary endpoint • OS Secondary endpoints • MFS, DSS, and DFS • Safety First patient enrolled: Nov 2018 Last patient enrolled: Jul 2021 Last RC: Nov 2021 Full study design details have been previously published and are available at ClinicalTrials.gov, NCT03732677; EudraCT number, 2018-001811-59. C, comparator; CrCl, creatinine clearance; D, durvalumab; DFS, disease-free survival; DSS, disease-specific survival; IV, intravenously; MFS, metastasis-free survival; Q3W, every 3 weeks; Q4W, every 4 weeks; R, randomised; RC, radical cystectomy. [Slide 2] Figure 1. Histologic subtypes in NIAGARA UC with DDHS = 16% of ITT^6 84% Invasive UC, not otherwise specified 8% UC with squamous differentiation 2% UC with glandular differentiation 5% UC with other histologic subtype N=1063 Squamous differentiation, n=87; glandular differentiation, n=25; other histologic subtype, n=53. [Slide 3] Figure 2. pCR, EFS, and OS in patients with UC with DDHS A pCR rate^6 Odds ratio, 1.43 (95% CI, 0.75–2.70) D arm n=76: 41% (31/76); C arm n=89: 33% (29/89) ITT^6,13 | D arm N=533 | C arm N=530 | Odds ratio (95% CI) pCR rate, n | 199 (37%) | 146 (28%) | 1.60 (1.23–2.08) B EFS^6 Probability of EFS / Time from randomisation (months); D arm, C arm | UC with DDHS D arm n=76 | C arm n=89 | ITT^6 D arm N=533 | C arm N=530 EFS events, n (%) | 24 (32) | 49 (55) | 187 (35) | 246 (46) Median EFS (95% CI), months | NR (NR–NR) | 32.2 (16.1–NR) | NR (NR–NR) | 46.1 (32.2–NR) HR (95% CI) | 0.52 (0.32–0.84) | 0.68 (0.56–0.82) C OS^6 Probability of OS | UC with DDHS D arm n=76 | C arm n=89 | ITT^6 D arm N=533 | C arm N=530 Deaths, n (%) | 19 (25) | 39 (44) | 136 (26) | 169 (32) Median OS (95% CI), months | NR (NR–NR) | NR (35.8–NR) | NR (NR–NR) | NR (NR–NR) HR (95% CI) | 0.53 (0.30–0.91) | 0.75 (0.59–0.93) Data cutoff: 29 April 2024. pCR results assessed by central pathology review are presented as odds ratio with corresponding CIs obtained using logistic regression (adjusted for stratification factors in the ITT and in the subgroup stratified by treatment and UC with DDHS only). EFS was assessed by BICR, or by central pathology review when a biopsy was required to evaluate a suspected new lesion. Median EFS and OS and corresponding 95% CIs were estimated using the Kaplan–Meier method. For the subgroup analyses, HR was based on an unstratified Cox proportional hazards model with ties handled by the Efron approach. BICR, blinded independent central review; CI, confidence interval; HR, hazard ratio; NR, not reached. Survival table Endpoint: EFS (UC with DDHS subgroup) Arm | N | Median | Median CI D arm (durvalumab + gem/cis) | 76 | NR | NR-NR C arm (gem/cis) | 89 | 32.2 | 16.1-NR HR: 0.52 HR CI: 0.32-0.84 [Slide 4] Figure 3. Additional efficacy outcomes in patients with UC with DDHS A MFS B DSS C DFS (RC + disease-free) Probability of MFS / DSS / DFS; Time from randomisation (months); Time from radical cystectomy (months); D arm, C arm MFS | UC with DDHS D arm n=76 | C arm n=89 | ITT D arm N=533 | C arm N=530 MFS events, n (%) | 23 (30) | 41 (46) | 152 (29) | 201 (38) Median MFS (95% CI), months | NR (NR–NR) | 43.8 (30.6–NR) | NR (NR–NR) | NR (48.0–NR) HR (95% CI) | 0.63 (0.37–1.04) | 0.67 (0.54–0.83) DSS | UC with DDHS D arm n=76 | C arm n=89 | ITT D arm N=533 | C arm N=530 Death due to bladder cancer, n (%) | 9 (12) | 25 (28) | 85 (16) | 114 (22) Median DSS (95% CI), months | NR (NR–NR) | NR (NR–NR) | NR (NR–NR) | NR (NR–NR) HR (95% CI) | 0.39 (0.17–0.81) | 0.69 (0.52–0.91) DFS | UC with DDHS D arm n=49 | C arm n=54 | ITT D arm N=352 | C arm N=337 DFS events, n (%) | 8 (16) | 19 (35) | 78 (22) | 102 (30) Median DFS (95% CI), months | NR (NR–NR) | NR (32.6–NR) | NR (NR–NR) | NR (51.3–NR) HR (95% CI) | 0.45 (0.19–0.99) | 0.69 (0.51–0.93) Data cutoff: 29 April 2024. Median MFS, DSS, and DFS and corresponding 95% CIs were estimated using the Kaplan–Meier method. For the subgroup analyses, HR was based on an unstratified Cox proportional hazards model with ties handled by the Efron approach. For the ITT, HR was based on a stratified Cox proportional hazards model, and the stratification factors were tumour stage (T2N0 vs >T2N0), renal function (adequate vs borderline), and PD-L1 status (high vs low/negative) per interactive voice response system, with ties handled by the Efron approach. Survival table Endpoint: MFS (UC with DDHS subgroup) Arm | N | Median | Median CI D arm (durvalumab + gem/cis) | 76 | NR | NR-NR C arm (gem/cis) | 89 | 43.8 | 30.6-NR HR: 0.63 HR CI: 0.37-1.04

[Slide 2] Arrival from prior therapy mCRPC Slowly rising PSA (Sustained response) | Indolent disease course Continue monitoring Oligoprogressive disease | Consider SBRT while maintaining systemic control Progression requiring change in therapy | Select next mechanism based on prior exposure and biomarkers Rapid progression | Symptomatic/visceral disease Consider prompt systemic therapy Aggressive variant / neuroendocrine transformation | Consider platinum-based therapy Clinical trials [Slide 3] CLINICAL PRACTICE GUIDELINES ESMO Advanced and Metastatic Prostate Cancer Living Guideline v1.0 June 2026 Metastatic CRPC aBPAs are administered concomitantly with anticancer treatments. Fizazi K, Attard G, Azad A et al Advanced and metastatic prostate cancer: ESMO Clinical Practice Guideline for diagnosis, treatment and follow-up Annals of Oncology, 2026; 37, 590-607 mCRPC 1 First, think about bone protective agents in all patients with bone metastasesa Patients with bone metastases BPAa [I, A] Biomarkers testing Patient without detected genetic alterations Presence of detected genetic alterations [Slide 4] REDUSE 96/12: Denosumab Q12w vs Q4w Phase III, randomized, non-inferiority trial in patients with bone metastases from mBC or mCRPC Trial design N = 1,380 Patients with bone metastases from mBC or mCRPC Randomised, phase III, double-blind Denosumab 120 mg SC every 4 weeks (Q4w) Denosumab 120 mg SC every 12 weeks (Q12w) after induction phase All patients: 500 mg calcium + 400 IU vitamin D daily Key efficacy result Q12w non-inferior to Q4w for time to first symptomatic skeletal event (SSE) HR 1.023 (90% CI 0.874–1.197) Key safety results Lower risk of hypocalcaemia significantly less with Q12w Lower risk of ONJ/tooth infection less with Q12w Economic impact Drug costs reduced by 53% per patient with Q12w

[Slide 1] 1 Patient cohort 9 patiens enrolled 7 male, 2 female 73.5 y median age Tumor entities: ❖ 7 Urothelial cancer ❖ 1 Head and neck cancer ❖ 1 Cervical cancer 2 Baseline imaging [89Zr]Zr-DFO-EV PET/CT Dose: [37 MBq] [18F]FDG PET/CT Dose: [2 MBq/kg] • EV PET scan 5-7 d post injection • Additional FDG PET within 4 d prior to EV PET in 6 patients 3 Treatment EV P 4 Follow-up imaging [89Zr]Zr-DFO-EV PET/CT Dose: [37 MBq] [18F]FDG PET/CT Dose: [2 MBq/kg] Imaging analysis: • Lesion detection rates (EV vs. FDG vs. ceCT) • Tracer uptake (SUVmax) • Tumor-to-background ratio 5 Response assessment RECIST 1.1 Complete response Partial response Stable disease Progressive disease

[Slide 1] RESULTS: Baseline tumor fraction distributions Figure 1: Baseline ctDNA TF Captures Disease Burden Across Key Subgroups A P = 0.047 13% (n=50) No | 30% (n=49) Yes Prior Docetaxel B P < 0.001 12% (n=54) ≤20 | 33% (n=45) >20 Bone Lesions C P = 0.003 8% (n=34) 0 | 28% (n=65) 1+ ECOG Performance Status D P < 0.001 5% (n=49) <Median | 33% (n=50) ≥Median PSA (Median split) Tumor Fraction (%) Wilcoxon rank-sum test | Significant p-values (< 0.05) shown in red bold Continuous tumor fraction values correlate with adverse clinical events previously demonstrated to associate to lack of response to combination Arm A (>20 bone lesions, prior docetaxel). [Slide 2] Table 1: Baseline ctDNA detection and TF magnitude as prognostic factors for rPFS, adjusted for clinical variables, by arm. Detection and TF ≤15% overlap and were therefore tested in separate adjusted models. Arm | Covariate | n | Univariate HR (95% CI) | p | Multivariate Model 1: TF non-detected HR (95% CI) | p | Multivariate Model 2: TF <=15 HR (95% CI) | p Arm A: Combo | ctDNA TF Not Detected | 47 | 0.37 (0.15-0.92) | 0.032 | 0.45 (0.17-1.19) | 0.107 | | | ctDNA TF ≤15% | 47 | 0.12 (0.05-0.30) | <0.001 | | | 0.14 (0.04-0.46) | 0.001 | PSA ≥55.1 ng/mL | 47 | 3.07 (1.41-6.71) | 0.005 | 1.96 (0.83-4.63) | 0.124 | 1.02 (0.36-2.87) | 0.970 | Bone lesions >20 | 47 | 2.23 (1.08-4.62) | 0.031 | 1.79 (0.81-3.93) | 0.148 | 1.27 (0.54-3.00) | 0.579 | Prior docetaxel | 47 | 2.00 (0.94-4.29) | 0.074 | 1.65 (0.75-3.64) | 0.217 | 1.26 (0.52-3.04) | 0.604 Arm B: Monotherapy | ctDNA TF Not Detected | 46 | 0.36 (0.14-0.95) | 0.038 | 0.31 (0.12-0.85) | 0.023 | | | ctDNA TF ≤15% | 43 | 0.46 (0.23-0.91) | 0.027 | | | 0.46 (0.22-0.99) | 0.047 | PSA ≥55.1 ng/mL | 46 | 2.34 (1.14-4.80) | 0.021 | 2.10 (0.97-4.56) | 0.059 | 1.93 (0.81-4.57) | 0.137 | Bone lesions >20 | 46 | 0.64 (0.33-1.26) | 0.195 | 0.47 (0.21-1.08) | 0.075 | 0.49 (0.21-1.13) | 0.096 | Prior docetaxel | 46 | 1.02 (0.53-1.94) | 0.955 | 1.23 (0.55-2.74) | 0.618 | 1.25 (0.53-2.94) | 0.615 PSA dichotomized at 55.1 ng/mL. TF dichotomized at 15% (TF >15% as reference). Multivariate models adjust for PSA, bone lesion burden, and prior docetaxel within each arm. Cohort includes all patients with a ctDNA detection result (Detected or Not Detected). Univariate analyses use per-variable complete cases. Multivariate analyses use complete cases on all adjustment variables. Each multivariate model includes one TF covariate; blank cells indicate the covariate was not in that model [Slide 3] RESULTS: Tumor fraction trajectories restricted to the SOC Mono arm B Figure 2. Early radiographic progression events clustered predominantly within the first 6 months from baseline (C1D1) and C2D1, supporting rPFS at 6 months, the scheduled duration of the radium-223 monotherapy course, as the classification endpoint for the concordance diagnostic analyses ctDNA trajectories at baseline and C2D1 ctDNA dynamics — Arm B: Radium-223 (Baseline + C2 union cohort, N = 45) Timepoint: Cycle 1 (Baseline); Cycle 2 ctDNA TF: ≤ 15%; > 15% ctDNA detection: Detected; Not Detected PFS endpoint: Event ≤6 mo from anchor; Event >6 mo from anchor; Censored Months from C1D1 (treatment start) TABLE 3: TF magnitude at baseline and persistent on-treatment ctDNA TF detection, captures the patients with near-term progression within the radium-223 treatment window restricted to patients evaluable for 6-month status Biomarker | Timepoint | N | Sens | Spec | PPV | NPV TF > 15% | Baseline | 40 | 69.0% | 72.7% | 87.0% | 47.1% | C2 | 41 | 60.7% | 92.3% | 94.4% | 52.2% | C4 | 21 | 50.0% | 85.7% | 87.5% | 46.2% ctDNA detection | Baseline | 43 | 96.6% | 42.9% | 77.8% | 85.7% | C2 | 41 | 82.1% | 84.6% | 92.0% | 68.8% | C4 | 21 | 78.6% | 85.7% | 91.7% | 66.7% Patient-level ctDNA trajectories at baseline and C2D1, aligned to rPFS outcomes, supported rPFS at 6 months as the classification endpoint for the concordance diagnostic analyses. A baseline TF threshold of >15% identified patients at risk for 6-month progression with strong positive prognostic value but limited sensitivity (that did not improve substantially at C2D1 and at C4). In contrast, ctDNA detection status in Arm B was highly concordant with 6-month progression beginning at C2D1 and remained concordant at C4, with detection at C2D1 also outperforming the baseline detection call. [Slide 4] RESULTS: Genomic composition of persistent ctDNA in the SOC Arm B progressors. Figure 4. Alteration dynamics at C2D1 in True-positive progressors on radium-223 monotherapy (Arm B, n=23). (A) Patient-by-gene heatmap showing baseline-to-C2D1 dynamics for genes altered in ≥2 patients. Each row represents one patient, each column one gene; tiles are colored by alteration dynamic (dark blue = persistent at both timepoints; orange = emergent at C2D1; gray = cleared at C2D1). Genes are grouped by pathway (AR axis, tumor suppressor, cell cycle, HRR, other). (B) Frequency of persistent alterations (present at both baseline and C2D1) among progressors, restricted to genes persistent in ≥2 patients; bars colored by pathway. Short variants with tumor-somatic probability >0.80 were retained regardless of functional annotation, including variants of uncertain significance (VUS), and the same criteria were applied at baseline and C2D1, so that all tumor-derived signal contributing to ctDNA persistence was captured. A. Alteration dynamics in true-positive progressors (Arm B, n = 23) AR axis | Tumor suppressor | Cell cycle | HRR | Other Dynamic (baseline → C2D1): Persistent; Emergent; Cleared B. Persistent alterations (baseline + C2D1) in TP progressors Genes persistent in ≥2 of 23 patients AR 15/23 (65%) TMPRSS2 12/23 (52%) PTEN 9/23 (39%) TP53 6/23 (26%) CTNNB1 3/23 (13%) FAS 3/23 (13%) MLL2 3/23 (13%) MYC 3/23 (13%) NBN 3/23 (13%) RAD21 3/23 (13%) AKT1 2/23 (9%) AXIN1 2/23 (9%) BRAF 2/23 (9%) BRCA2 2/23 (9%) CARD11 2/23 (9%) CCNE1 2/23 (9%) CDKN1B 2/23 (9%) CIC 2/23 (9%) EPHB1 2/23 (9%) HGF 2/23 (9%) IRS2 2/23 (9%) PIK3CA 2/23 (9%) RB1 2/23 (9%) % of TP progressors with persistent alteration

[Slide 2] High-risk localised / locally advanced or node-positive prostate cancer *RT + ADT (3 years) + abiraterone (2 yrs) *RT + ADT + enzalutamide (2 yrs) *ADT + apalutamide (6 mo) + surgery + ADT + apalutamide (6 mo) *RT + ADT + darolutamide (24 mo) Colour code: treatment modality Hormonal / ARPI Chemotherapy Triplet: chemo + ARPI PARP inhibitor Radioligand / radium-223 Immunotherapy / targeted therapy Radiotherapy * Not EMA approved Hormone-sensitive prostate cancer HSPC/Androgen pathway modulation-sensitive/naive APMS/APMN High-risk nmHSPC (EMBARK): enzalutamide ± ADT ADT + ARPI (abiraterone, apalutamide, darolutamide, enzalutamide) ADT + docetaxel ADT + docetaxel + abiraterone or darolutamide ADT + abiraterone + niraparib (BRCA alteration) ADT + RT to primary tumour *ADT + enzalutamide + talazoparib (HRR alteration) *ADT + ARPI + 177Lu-PSMA-617 *ADT + abiraterone + capivasertib (PTEN deficiency) Castration-resistant prostate cancer CRPC/Androgen pathway modulation-resistant APMR High-risk nmCRPC: ADT + enzalutamide, apalutamide or darolutamide ADT + abiraterone or enzalutamide ADT + docetaxel or cabazitaxel ADT + radium-223 (± enzalutamide) ADT + 177Lu-PSMA-617 or 177Lu-PSMA-I&T ADT + abiraterone + olaparib or niraparib (BRCA1/2) ADT + enzalutamide + talazoparib (HRR alteration) *ADT + checkpoint inhibition (dMMR and/or MSI-high) or targeted therapy in case of a relevant alteration *Sipuleucel-T (USA) [Slide 3] A PSA ≤0.2 ng/mL B PSA >0.2 to 1.0 ng/mL C PSA >1.0 to 3.0 ng/mL D PSA >3.0 ng/mL MON0 | MON+ | M1LV | M1HV Overall survival (%) Time from landmark (months) Number at risk (censored) Survival table Endpoint: Overall survival from landmark A (PSA ≤0.2): MON0 | 719 A (PSA ≤0.2): MON+ | 417 A (PSA ≤0.2): M1LV | 620 A (PSA ≤0.2): M1HV | 428 B (PSA >0.2–1.0): MON0 | 343 B (PSA >0.2–1.0): MON+ | 207 B (PSA >0.2–1.0): M1LV | 391 B (PSA >0.2–1.0): M1HV | 417 C (PSA >1.0–3.0): MON0 | 189 C (PSA >1.0–3.0): MON+ | 163 C (PSA >1.0–3.0): M1LV | 249 C (PSA >1.0–3.0): M1HV | 269 D (PSA >3.0): MON0 | 125 D (PSA >3.0): MON+ | 104 D (PSA >3.0): M1LV | 294 D (PSA >3.0): M1HV | 776 [Slide 4] MIGRATION BETWEEN CONVENTIONAL IMAGING AND PSMA PET Conventional imaging (CI) (CT/MRI + Bone Scan) → PSMA-PET High-Volume Disease (HVD-CI) n = 17 Low-Volume Disease (LVD-CI) n = 50 94% 22% 48% 30% High-Volume Disease (HVD-PET) n = 27 Low-Volume Disease (LVD-PET) n = 24 Non-Metastatic Disease (M0-PET) n = 16

[Slide 1] Figure 1: Conditional adjusted cumulative incidence of outcomes according to recurrence status at 3-years (A) PCSM Recurrence No 5-year PCSM: 2.5% (2.2%,2.7%) Yes 5-year PCSM: 18% (17%,20%) PCSM (%) Years post landmark timepoint At Risk (B) ACM Recurrence No 5-year ACM: 17% (16%,17%) Yes 5-year ACM: 34% (32%,35%) ACM (%) Years post landmark timepoint At Risk Survival table Endpoint: Prostate cancer-specific mortality (conditional cumulative incidence, landmark at 3 years) No recurrence at 3 years | 14443 Recurrence at 3 years | 2818 5-year PCSM: No 2.5% (2.2%, 2.7%) vs Yes 18% (17%, 20%) 5-year ACM: No 17% (16%, 17%) vs Yes 34% (32%, 35%) [Slide 2] Table 2: Conditional adjusted 5-year PCSM and ACM rates in clinical subgroups, according to recurrence status at 3 years. Recurrence at 3 years | N | Conditional 5-year PCSM (95% CI) | Conditional 5-year ACM (95% CI) No. of NCCN high-risk factors 0 factor | No | 8515 | 1.3 (1.1, 1.5) | 15 (15,16) 0 factor | Yes | 1078 | 12 (9.7, 14) | 25 (23,27) 1 factor | No | 3916 | 3.0 (2.6, 3.5) | 17 (16,18) 1 factor | Yes | 851 | 22 (20, 25) | 37 (34,40) 2-3 factors | No | 1959 | 5.3 (4.5, 6.1) | 20 (19,22) 2-3 factors | Yes | 879 | 31 (28, 35) | 46 (43,49) Planned ADT duration for high-risk disease None | No | 315 | 2.8 (1.5, 4.5) | 19 (14,23) None | Yes | 197 | 25 (19, 32) | 40 (33,48) 3-8 months | No | 3233 | 3.4 (2.9,3.9) | 18 (17,20) 3-8 months | Yes | 1152 | 26 (23,28) | 40 (37,43) ≥24 months | No | 2375 | 4.6 (3.9,5.3) | 18 (16,19) ≥24 months | Yes | 391 | 26 (22,30) | 43 (39,48) Age at randomization <70 years | No | 7511 | 2.7 (2.4,3.0) | 12 (12,13) <70 years | Yes | 1558 | 20 (18,22) | 29 (27,32) ≥70 years | No | 6930 | 2.2 (1.8,2.5) | 21 (20,22) ≥70 years | Yes | 1260 | 17 (15,20) | 38 (36,40) [Slide 3] Table 3: Summary of trial-level surrogacy in the overall population (25 treatment comparisons), as well as in patients with high-risk disease (22 treatment comparisons, excluding trials with fewer than 50 high-risk patients), and age under 70 years old (24 treatment comparisons, excluding trials with fewer than 50 patients under the age of 70) Overall population: No. of trials | R2 (95% CI) || High-risk population: No. of trials | R2 (95% CI) || Age <70 years: No. of trials | R2 (95% CI) A. Non-prostate cancer death as an event Δ (5-yr OS) on Δ (2-yr RFS) | 25 | 0.09 (0.00,0.33) | 22 | 0.03 (0.00,0.25) | 24 | 0.02 (0.00,0.21) Δ (8-yr OS) on Δ (3-yr RFS) | 21A | 0.29 (0.02,0.53) | 18A | 0.22 (0.00,0.49) | 20 | 0.06 (0.00,0.31) Δ (10-yr OS) on Δ (5-yr RFS) | 19B | 0.27 (0.00,0.53) | 16B | 0.12 (0.00,0.42) | 18 | 0.13 (0.00,0.41) Log (OS HR) on Log (RFS HR) | 25 | 0.30 (0.03,0.53) | 22 | 0.16 (0.00,0.42) | 24 | 0.28 (0.02,0.52) B. Non-prostate cancer death as a competing risk Δ (5-yr PCSM) on Δ (2-yr TTR) | 23C | 0.45 (0.12,0.64) | 21C | 0.52 (0.17,0.69) | 23C | 0.39 (0.08,0.60) Δ (8-yr PCSM) on Δ (3-yr TTR) | 20AC | 0.56 (0.21,0.73) | 18A | 0.51 (0.13,0.70) | 20AC | 0.53 (0.17,0.70) Δ (10-yr PCSM) on Δ (5-yr TTR) | 18BC | 0. 52 (0.14,0.70) | 16B | 0.48 (0.08,0.69) | 18BC | 0.42 (0.06,0.64) Log (PCSM HR) on Log (TTR HR) | 23C | 0.38 (0.07,0.60) | 21C | 0.08 (0.00,0.33) | 23C | 0.38 (0.07,0.59) RFS – relapse-free survival; OS – overall survival; TTR – time to recurrence; PCSM – prostate cancer-specific mortality A Excluding studies with median follow-up < 6 years; B Excluding studies with median follow-up < 8 years; C Excluding studies with number of prostate cancer death < 3 [Slide 4] Figure 2: Bubble plots and regression at the trial level (A) Difference in 8-year OS on difference in 3-year RFS (Positive value means better outcome in the exp arm) Difference in 8-year OS Difference in 3-year RFS R2 =0.29(95%CI:0.02-0.53) (B) Difference in 8-year PCSM on difference in 3-year TTR (Negative value means better outcome in the exp arm. Difference in 8-year PCSM Difference in 3-year TTR R2 =0.56(95%CI:0.21-0.73)

[Slide 1] 1 unselected mCRPC Prior to 1L or 2L therapy 2 NGS Archival tissue Fresh biopsy TruSightOncology 500 with TC of >30% Cell-free DNA White-blood cells Custom ctDNA Prostate Panel Chip-corrected ctDNA 3 Concordance [Slide 2] ctDNA ≥20% = 56/174 (32%) FRESH (n=102) ARCHIVAL (n=72) ctDNA fraction % biopsy 174 (58%) pts with HRR/PI3K pathway alterations Median ctDNA = 5% ctDNA <2% = 52/174 (30%) [Slide 4] BRCA1/2 mutations | 14 of 15 detected by plasma | +3 (plasma only) BRCA1/2 biallelic loss | 3 of 6 detected by plasma Non-BRCA HRR mutations | 29 of 36 detected by plasma | +1 (plasma only) Legend: Detected in tissue and plasma | Tissue only (plasma miss) | Plasma only (tissue miss) What proportion of patients with unsuccessful tissue NGS or no tissue could be salvaged with ctDNA?

[Slide 1] TRIAL DESIGN 35 patients treated in total Key eligibility • Progressive mCRPC • PSMA-PET positive disease (SUVmax ≥15) with no discordance on FDG-PET • Prior exposure to docetaxel and ≥1 androgen receptor pathway inhibitor • Adequate bone marrow, renal and hepatic function • ECOG performance status 0-1 Phase I | 15 patients | 3+3 design Dose escalation Up to 6 cycles 177Lu-PSMA 7.4 GBq IV 6 weekly + cabazitaxel Up to 12 doses IV 3-weekly Prophylactic G-CSF given after each dose First dose of cabazitaxel one day after 177Lu-PSMA 4 planned dose levels: 12.5 mg/m² (n=3); 15 mg/m² (n=3); 17.5 mg/m² (n=3); 20 mg/m² (n=6) No dose-limiting toxicity cabazitaxel RP2D: 20 mg/m² Phase II | 20 patients | Dose expansion Recruitment completed: Feb. 2025 Up to 6 cycles 177Lu-PSMA 7.4 GBq IV 6 weekly + cabazitaxel IV 3-weekly 20 mg/m² Prophylactic G-CSF given after each dose [Slide 2] RESULTS Median follow-up: 20.4 months Safety analysis: n=34 • Efficacy analysis: n=26 (RP2D cohort) 71% had de novo metastatic disease 29% had M1c disease at baseline No dose-limiting toxicities 24% experienced clinically significant Grade ≥3 treatment-related adverse events • neutropaenia (n=2) • anaemia (n=2) • weight loss (n=1) • diarrhoea (n=1) • fatigue (n=1) • urinary tract infection (n=1) • No deaths PSA50-RR 73% [95% CI 54 – 86] PSA90-RR 42% [95% CI 26 – 61] Median PSA-PFS 8.5 months [95% CI 4.8 – 12.1] Radiographic PFS 10.9 months [95% CI 10.2 – 13.4] Overall Survival 16.5 months [95% CI 13.4 – 21.8] [Slide 3] The combination of cabazitaxel and 177Lu-PSMA is feasible and safe in heavily pre-treated patients with mCRPC, with encouraging anti-tumour activity. Cabazitaxel treats both PSMA-positive and PSMA-negative cells radiosensitises cells 177Lu-PSMA β- particles Cell cycle: G0 phase (resting), G1, S, G2, M The findings support further evaluation in a randomised clinical trial. Sponsor: Pet[er MacCallum Cancer Centre] PSMA-617 was supplied by [...]

Darolutamide 600 mg BID + ADT ARASEC match-eligible population (N=223) ADT CHAARTED match-eligible population (N=393)* Key eligibility - mHSPC - ECOG performance status 0-2 - ADT started <=120 days before enrollment - No prior chemotherapy or hormone therapy 1:1 propensity score matching with a caliper (maximum difference between scores) of 0.1 [Illustration: score pairs 0.25 vs 0.52 (rejected, X); 0.10 vs 0.12 (matched, check)] Propensity scores calculated from prognostic baseline variables - Age (<68 years, >=68 years) - ECOG performance status (0, >0) - Extent of disease (high volume with visceral metastases, high volume without visceral metastases, low volume) - Prior local therapy (yes/no, proxy for recurrent vs de novo disease) - Gleason score (<8, >=8, unknown) - Baseline PSA (<=10 ng/mL, >10 ng/mL) ARASEC matched population (n=160) CHAARTED matched population (n=160) Outcomes Primary endpoint - PFS Secondary endpoints - OS - Time to mCRPC - rPFS - PSA <0.2 ng/mL response rate - Safety(dagger) Median follow-up Darolutamide arm: 26 months CHAARTED-matched arm: 28 months

[Slide 1] [177Lu]Lu-PSMA | [223Ra]RaCl2 Target | All PSMA-positive metastasis | Bone metastasis only Treatment | 4-6 cycles of 6 weeks | 6 cycles of 4 weeks Patient impact | ↑ | ↓ Cost | ↑ | ↓ Phase III studies (median OS in months) | VISION (2021) After taxane and ARPI 11.3 → 15.3 | ALSYMPCA (2013) Pre-ARPI, pre-PSMA PET era 11.3 → 14.9 OS = overall survival [177Lu]Lu-PSMA [223Ra]RaCl2 [Slide 2] Unweighted | Weighted Characteristic | [177Lu]Lu-PSMA | [223Ra]RaCl2 | SMD | SMD Age, years | 72.6 ± 8.0 | 71.9 ± 8.3 | 0.08 | 0.14 ECOG performance status ≥2 | 28.3% | 8.5% | 0.58 | 0.23 Gleason score ≥9 | 48.3% | 34.9% | 0.37 | 0.19 >6 bone metastases | 81.5% | 89.2% | 0.49 | 0.08 Lymphadenopathy (≥1cm) | 51.1% | 26.9% | 0.51 | 0.07 PSA (µg/L) | 606.8 ± 1023.8 | 245.8 ± 556.9 | 0.44 | 0.04 ALP (U/L) | 293.4 ± 457.7 | 225.2 ± 277.5 | 0.18 | 0.16 Haemoglobin (mmol/L) | 7.4 ± 1.0 | 7.7 ± 1.0 | 0.38 | 0.06 LDH (U/L) | 339.7 ± 429.1 | 283.2 ± 218.5 | 0.166 | 0.05 Previous treatments ≥4 | 16.3% | 5.9% | 0.88 | 0.37 SMD = standardised mean difference, ECOG = Eastern Cooperative Oncology Group, PSA = prostate specific antigen, ALP = alkaline phosphatase, LDH = lactate dehydrogenase [Slide 3] Overall survival before and after sIPTW weighting: Left panel: Median (months) [177Lu]Lu-PSMA 8.9; [223Ra]RaCl2 12.9; HR = 1.27 (95% CI 0.95 - 1.70); p = 0.11 Right panel: Median (months) [177Lu]Lu-PSMA 12.6; [223Ra]RaCl2 11.2; HR = 0.97 (95% CI 0.66 - 1.44); p = 0.89 Overall survival probability (%) Time (months) Number at risk Weighed number at risk CONCLUSIONS Survival table Endpoint: Overall survival (real-world, before and after sIPTW weighting) [177Lu]Lu-PSMA (unweighted) | 92 | 8.9 months [223Ra]RaCl2 (unweighted) | 323 | 12.9 months [177Lu]Lu-PSMA (sIPTW-weighted) | 75.4 (weighted) | 12.6 months [223Ra]RaCl2 (sIPTW-weighted) | 315 (weighted) | 11.2 months HR: Unweighted 1.27; weighted 0.97 HR CI: Unweighted 0.95 - 1.70; weighted 0.66 - 1.44 P: Unweighted 0.11; weighted 0.89

[Slide 1] Primary Endpoint: rPFS (ITT) HR: 1.01 (90% CI: 0.77, 1.31); 1-sided p = 0.51 177Lu-PSMA-617: median 8.6 months (90% CI: 7.1, 10.6) Docetaxel: median 10.7 months (90% CI: 7.8, 11.1) rPFS Probability Time in Months Number at risk (months 0, 6, 12, 18, 24, 30) Docetaxel: 99 59 27 7 3 1 177Lu-PSMA-617: 100 56 25 10 2 2 Survival table Endpoint: rPFS (ITT), primary endpoint Arm | N | Median | Median CI 177Lu-PSMA-617 | 100 | 8.6 months | 90% CI 7.1-10.6 Docetaxel | 99 | 10.7 months | 90% CI 7.8-11.1 HR: 1.01 HR CI: 90% CI 0.77-1.31 P: 1-sided p = 0.51 [Slide 4] FACT-G * Indicates p<0.05 Mean change from baseline Weeks: 6 12 18 24 30 36 Legend: LuP | Doc [* markers at weeks 12, 24, 30, 36] Physical Well Being Weeks: 6 12 18 24 30 36 Legend: LuP | Doc [* markers at weeks 24, 30]

[Slide 1] Ixo-V CLDN6 endocytosis MMAE payload released CANCER CELL Apoptosis [Slide 2] CLDN6 mRNA Expression Across Tumor Types^2 Log(FKPM)-RNA-seq AML n=173 | Large B-cell lymphoma n=48 | Prostate cancer n=550 | Melanoma n=474 | Colon cancer n=329 | Rectal cancer n=105 | Bladder cancer n=426 | Liver cancer n=423 | Kidney clear cell carcinoma n=606 | Breast cancer n=1218 | Cervical squamous cell carcinoma n=308 | Head and neck cancer n=566 | Kidney chromophobe n=91 | Sarcoma n=265 | Lower grade glioma n=530 | Adrenocortical carcinoma n=92 | Glioblastoma n=172 | Kidney papillary cell carcinoma n=323 | Lung squamous cell carcinoma n=533 | Esophageal cancer n=198 | Thymoma n=122 | Stomach adenocarcinoma n=450 | Cholangiocarcinoma n=45 | Thyroid carcinoma n=572 | Lung adenocarcinoma n=576 | Pancreatic cancer n=183 | Endometrioid cancer n=201 | Uterine carcinosarcoma n=57 | Ovarian cancer n=308 | Testicular cancer n=156 (highlighted)

[Slide 1] Patients treated with LuPSMA n = 509 Excluded (n=319) No subsequent therapy (n=227) Insufficient follow-up (n=71) Intercurrent therapy with LuPSMA (n=21) Included (n=190) LuPSMA retreatment (n=90) Other subsequent therapy (n=100) [Slide 2] Other subsequent therapies following initial LuPSMA Cabazitaxel 59% Docetaxel 15% Carboplatin +/- etoposide 14% Clinical trial 4% Olaparib 3% Mitoxantrone 2% Cabazitaxel/Carboplatin 1% Capecitabine 1% Enzalutamide 1% [Slide 3] Outcomes for subsequent therapies following progression on LuPSMA LuPSMA retreatment (n=90)† | Other subsequent systemic therapy (n=100)† Median PSA-PFS of initial LuPSMA course | 11 (10.6-12.6) | 4.1 (3.8-5.2) Median months from last LuPSMA cycle to subsequent therapy | 8.7 (IQR 5.8-12.8) | 2.5 (IQR 1.6-5.1) Median baseline PSA prior to subsequent therapy (ug/L) | 32 (IQR 10-106) | 112 (IQR 39 – 408) PSA50-RR to subsequent therapy | 40 (44%) | 20 (20%) PSA90-RR to subsequent therapy | 14 (16%) | 6 (6%) Median PSA-PFS of subsequent therapy | 4.1 (3.8 – 5.2) | 2.9 (2.6 – 3.7) Median OS from start of LuPSMA retreatment or other systemic therapy | 16 (12.3-20.0) | 7.5 (6.6-9.4) Median OS from start of initial LuPSMA course | 32.3 (28.4-37.5) | 14.7 (13.2-17.1) †n (%); months (95% CI)

Baseline Characteristics • As of July 14, 2026, a total of 26 patients with untreated IMDC Favorable-Risk advanced RCC were enrolled Characteristic | Patients, N=26 (%) Age, median(range), years | 56.5 (44-80) Male | 23 (88.5%) ECOG PS | 0 | 25 (96.2%) 1 | 1 (3.8%) IMDC risk category | Favorable | 26 (100.0%) PD-L1 expression (CPS) | CPS ≥1 | 4 (15.4%) CPS <1 | 11 (42.3%) Unknown | 11 (42.3%) Number of metastatic sites, n (%) | ≥2 metastatic sites | 11 (42.3%) History of Prior Nephrectomy | Yes | 26 (100.0%) [Jun] Guo, MD

[Survival curve figure] Y-axis: Survival Proportion in ADGC phase 1 X-axis: Age (60-100) Legend: 1st PHS Quartile 2nd PHS Quartile 3rd PHS Quartile 4th PHS Quartile Uncorrected baseline Survival table Endpoint: Survival proportion (dementia-free, ADGC phase 1) by age, by polygenic hazard score quartile 1st PHS Quartile 2nd PHS Quartile 3rd PHS Quartile 4th PHS Quartile Uncorrected baseline

[Slide 1] PATIENT AND SAMPLE OVERVIEW 349 patients 553 NGS reports 69 Median age 22.6% Tissue failure rate 5.3% ctDNA failure rate Ethnicity distribution White 235 (67%) Black 67 (19%) †Other / mixed / not specified 47 (14%) †Includes South Asian, East Asian, mixed, other ethnic group, and not specifie[illegible] NGS reports by assay platform ctDNA (N = 319) FoundationOne Liquid AB1 159 (50%) FoundationOne Liquid CDx 131 (41%) Guardant360 29 (9%) Tissue (N = 234) FoundationOne CDx 182 (78%) OncoPlex 52 (22%) FoundationOne Liquid AB1: TRITON3. FoundationOne Liquid CDx: TRITON2, GSTT ctDNA Pilot Project, TARGET. Guardant360: free kits/self-payers. FoundationOne CDx: TRITON3, TRITON2. OncoPlex: NEPTUNES. [Slide 2] Alteration prevalence (%) by ethnicity White Black Others % of patients TP53: White 103 (43.8%) | Black 23 (34.3%) | Others 19 (40.4%) PTEN: White 47 (20.0%) | Black 13 (19.4%) | Others 6 (12.8%) RB1: White 13 (5.5%) | Black 4 (5.9%) | Others 6 (12.8%) AVPC-m: White 24 (10.2%) | Black 11 (16.4%) | Others 8 (17.0%) BRCA1/2: White 16 (6.8%) | Black 8 (11.8%) | Others 4 (8.5%) Labels show n and %. Patient-level, tissue/ctDNA combined (White n=235, Black n=67, Others n=47). [Slide 3] HRD (non-BRCA) alterations breakdown by ethnicity White Black Others % of patients ATM: White 29 (12%) | Black 5 (7%) | Others 5 (11%) CDK12: White 9 (4%) | Black 6 (9%) | Others 4 (9%) CHEK2: White 14 (6%) | Black 2 (3%) | Others 1 (2%) PALB2: White 0 (0%) | Black 1 (1.5%) | Others 1 (2%) RAD54L: White 1 (0.4%) | Black 1 (1.5%) | Others 0 (0%) BRIP1: White 0 (0%) | Black 1 (1.5%) | Others 0 (0%) Labels show n and %. Patient-level, tissue/ctDNA combined (White n=235, Black n=67, Others n=47).

[Slide 1] LuPSMA treatment cycles C1 C2 C3 C4 C5 C6 C1D1 (n=63) C3D1 (n=50) PD (n=41) PredicineCARE: Targeted NGS: SNVs, CNVs, fusions PredicineSCORE: LP-WGS: CNB score, tumor fraction (TF), ARBS nucleosome profiling Figure 1. Study schema and ctDNA assays. Endpoint | Definition | n / N PSA90 response | Best PSA decline >90% | 17 / 72 Exceptional response | PSA90 or OS >18 mo | 24 / 72 Intrinsic resistance | Clinical/radiographic PD ≤3 mo or OS ≤9 mo | 25 / 72 Acquired resistance | PD after 3 mo | 54 / 72 [caption below table cut off] [Slide 2] Baseline ctDNA tumor fraction is prognostic Combo TF > 2%: High / Low [KM curve: Survival probability vs Months] p < 0.0001 Number at risk (TF > 2%) (0, 12, 24 months) High 46 15 2 Low 17 9 5 Exceptional Responder — Combo TF — Wilcoxon P = 0.000127; FALSE mean = 0.334 [partially obscured]; TRUE mean = 0.0753 [partially obscured] Intrinsic Resistance — Combo TF — Wilcoxon P = 1.66e-05; FALSE mean = 0.151; TRUE mean = 0.406 Continuous TF | HR* | P OS | 25.6 | 2.6×10^-6 CR-PFS | 6.5 | 2.3×10^-4 PSA-PFS | 6.4 | 5.3×10^-4 Figure 2. (Left) OS by baseline TF >2% vs ≤2%. (Right) Baseline combo TF by response group (Wilcoxon). *Cox HR per unit TF (0–1). Survival table Endpoint: OS by baseline ctDNA tumor fraction (>2% vs ≤2%) TF >2% (High) | 46 TF ≤2% (Low) | 17 P: <0.0001 [Slide 3] Integrated Analysis of PSA, tumor fraction and AR binding sites during LuPSMA treatment A All samples (n=154) — High risk / Low risk R = 0.81 CNV-inferred TF vs ARBS abnormality score B Patient 54 / C Patient 55 — C1D1, C3D1, Normal cfDNA, PD Fragment coverage vs Distance from ARBS (bp) D PSA | Tumor fraction | ARBS abnormality score — Pt 54 (C1D1–C6D1; C1D1, C3D1, PD), Pt 55 (C1D1–C4D1; C1D1, C3D1, PD) Figure 3. (A) ARBS abnormality score vs CNV-inferred TF (R=0.81); samples classified by copy number burden. (B, C) ARBS nucleosome depletion at C1D1, reduced at C3D1, regained at PD. (D) PSA, TF and ARBS over time.

394P TRANSCRIPTOMIC CHARACTERIZATION OF BASELINE IMMUNE PROGRAMS IN BCG-TREATED NON-MUSCLE INVASIVE BLADDER CANCER Pablo Sanchis1,2*, Rocío Solberg2, Ivan Macharashvili3, Agustina Sabater2,4, Hernán García Rivello5, Geraldine Gueron1, Federico Cayol3 1 Department of Biological Chemistry, IQUIBICEN (UBA-CONICET), Buenos Aires, Argentina; 2 Universidad Argentina de la Empresa (UADE), Instituto de Tecnología (INTEC), Buenos Aires, Argentina; 3 Department of Oncology, Hospital Italiano de Buenos Aires, Buenos Aires, Argentina; 4 Department of Genitourinary Medical Oncology, MD Anderson Cancer Center, Houston, United States of America; 5 Department of Pathology, Hospital Italiano de Buenos Aires, Buenos Aires, Argentina. *Contact: pabloasanchis@gmail.com [Logos: IQUIBICEN, Hospital Italiano de Buenos Aires, UADE, International Cancer Foundation, ESMO] 1. BACKGROUND & RATIONALE High Clinical Burden of BCG Failure: Intravesical Bacillus Calmette–Guérin (BCG) immunotherapy is the established standard-of-care for high-risk non–muscle invasive bladder cancer (NMIBC). However, up to 50% of patients experience recurrence or progression, frequently requiring radical cystectomy. Unmet Need for Predictive Biomarkers: There are currently no validated baseline biomarkers to identify tumors inherently resistant to BCG prior to treatment initiation. Study Hypothesis & Objective: We hypothesized that baseline primary NMIBC tumors harbor pre-existing transcriptomic and immunological rewiring that dictates therapeutic failure prior to BCG exposure. 2. METHODOLOGICAL WORKFLOW Phase 1: Multi-Cohort Harmonization & Quality Control Transcriptomics datasets (GSE13507, GSE176178, GSE224248) were downloaded from Gene Expression Omnibus. Patient clinical annotations were standardized for recurrence, progression, and BCG response. Phase 2: Differential Expression & Pathway Scoring (GSVA & GSEA) Differential gene expression was performed. Gene Set Variation Analysis (GSVA) and GSEA evaluated functional activity across 10,488 GO Biological Processes, KEGG, and HALLMARK pathways. Phase 3: Immune Microenvironment Deconvolution & Subtyping xCell algorithms quantified 64 immune and stroma cell types. Unsupervised hierarchical clustering and PCA identified distinct pre-treatment resistance phenotypes among BCG non-responders. 3. SUMMARY TABLE OF ANALYZED COHORTS Dataset | Total (n) | BCG (n) | Recur. | Prog. | BCG Resp. | RFS | PFS | OS GSE13507 | 256 | 56 | Yes | Yes | No | No | No | Yes GSE176178 | 44 | 44 | No | No | Yes | No | No | No GSE224248 | 49 | 49 | No | Yes | No | No | Yes | No E-MTAB | 476 | 88 | No | Yes | No | Yes | Yes | Yes BRS Cohort | 327 | 132 | Yes | Yes | No | [illegible] | [illegible] | [illegible] The presenting author declares no conflict of interest related to this work. | ESMO 2026 Congress, Madrid, Spain 4. BASELINE TRANSCRIPTIONAL DYSREGULATION [Volcano plot: Baseline Differential Gene Expression (BCG Recurrent vs Non-Recurrent); Venn diagram, overlap 45; functional category table - small text largely illegible] Figure 1. Baseline transcriptional dysregulation in BCG-recurrent tumors. (A) Baseline Differential Gene Expression (Volcano Plot): Differential gene expression in pre-treatment primary NMIBC tumors comparing patients who subsequently recurred following BCG vs. non-recurrent controls in GSE13507 (n = 56). (B) Venn's diagram showcasing common genes significantly deregulated in the GSE13507 and GSE176178 datasets. Significant upregulation of key inflammatory, hypoxia, and metabolic mediators (including MARCO, CSF1, VEGFA, S100A12, HMOX1, HK2, and MTHFD2; P < 0.05, log2 fold change > 0.5) indicates coordinated baseline molecular rewiring in recurrent tumors at the time of diagnosis. 5. UNSUPERVISED PATHWAY SUBTYPING (GSVA) [Heatmap; GSVA boxplots: Antigen processing and presentation of peptide antigen via MHC class I; Natural Killer cell activation involved in immune response; GSEA T cell activation pathway NES = -1.49, p = 0.00229 (small, partly legible)] Figure 2. Baseline immune suppression in BCG-recurrent tumors. (A) Heatmap showcasing an unsupervised clustering analysis based on the enrichment levels of biological pathways profiled using GSVA. (B) Normalized GSVA enrichment scores in non-recurrent and recurrent tumors from the GSE13507 dataset. (C) GSEA plot illustrating significant downregulation of the T Cell Activation Pathway prior to BCG treatment in patients that subsequently recurred (NES = [illegible] normalized enrichment score). * P < 0.05. 6. BIOLOGICAL HETEROGENEITY [partly obscured by presenter] [PCA plot and heatmap, obscured] [Presenter Pablo Sanchis standing beside poster, ESMO badge 'Poster Presenter']

[Slide 1] POTOMAC: Study Design and Primary Endpoint Durvalumab + BCG (I+M) reduced the risk of high-risk disease recurrence or death by 32% in the ITT population1 1018 patients with BCG-naive, high-risk NMIBC -> R 1:1:1 D+BCG (I+M) N=339: Durvalumab 1500 mg IV (Q4W × 13 cycles); BCG Induction Weekly × 6 weeks; BCG Maintenance 3 weekly doses at 3m 6m 12m 18m 24m D+BCG (I only) N=339: Durvalumab 1500 mg IV (Q4W × 13 cycles); BCG Induction Weekly × 6 weeks BCG (I+M) N=340: BCG Induction Weekly × 6 weeks; BCG Maintenance 3 weekly doses at 3m 6m 12m 18m 24m First patient enrolled: June 2018 | Last patient enrolled: October 2020 | Last D dose: January 2022 | Last BCG dose: January 2023 | Data cutoff: 03 April 2025 Primary endpoint: Disease-free survival D+BCG (I+M) N=339 | BCG (I+M) N=340 Events, n (%): 67 (20) | 98 (29) Median DFS, months (95% CI): NR (NR–NR) | NR (74.0–NR) Landmarks (12/24/36/48 months): D+BCG 92% / 87% / 82% / 80%; BCG 87% / 82% / 77% / 75% HR 0.68 (95% CI, 0.50–0.93) Stratified log-rank P value = 0.0154 No. of patients at risk: [illegible at this resolution] Median follow-up: 60.7 months For additional information, please refer to: De Santis M, et al. Lancet. 2025;406:2221–2234. Survival table Endpoint: Disease-free survival (primary), ITT Arm | N | Median | Median CI D+BCG (I+M) | 339 | NR | NR–NR BCG (I+M) | 340 | NR | 74.0–NR HR: 0.68 HR CI: 0.50–0.93 P: 0.0154 (stratified log-rank) Landmark rates: 12 months: D+BCG (I+M) 92% | BCG (I+M) 87% 24 months: D+BCG (I+M) 87% | BCG (I+M) 82% 36 months: D+BCG (I+M) 82% | BCG (I+M) 77% 48 months: D+BCG (I+M) 80% | BCG (I+M) 75% [Slide 2] POTOMAC: Disease-Free Survival in Patients With Papillary Only and With T1 Tumours D+BCG (I+M) showed a consistent benefit across papillary only and T1 subgroups Papillary only (64%–65% of ITT population)a,b,1 D+BCG (I+M) N=217 | BCG (I+M) N=220 Number of events, n (%): 37 (17) | 65 (30) Median DFS, months (95% CI): NR (NR–NR) | NR (74.0–NR) Median follow-up in censored patients: 60.7 months Landmarks (24/36/48 months): D+BCG 87% / 84% / 84%; BCG 81% / 76% / 74% HR 0.56 (95% CI, 0.37–0.84) T1 subset of papillary only (40%–43% of ITT population)c,d D+BCG (I+M) N=137 | BCG (I+M) N=145 Number of events, n (%): 22 (16) | 45 (31) Median DFS, months (95% CI): NR (NR–NR) | NR (69.3–NR) Median follow-up in censored patients: 60.7 months Landmarks (24/36/48 months): D+BCG 88% / 85% / 85%; BCG 79% / 75% / 73% HR 0.48 (95% CI, 0.28–0.79) No. of patients at risk: [illegible at this resolution] Survival table Endpoint: DFS, papillary-only and T1 subsets Arm | N | Median | Median CI D+BCG (I+M) papillary only | 217 | NR | NR–NR BCG (I+M) papillary only | 220 | NR | 74.0–NR D+BCG (I+M) T1 | 137 | NR | NR–NR BCG (I+M) T1 | 145 | NR | 69.3–NR HR: 0.56 (papillary only); 0.48 (T1) HR CI: 0.37–0.84; 0.28–0.79 Landmark rates: 24 months: D+BCG papillary only 87% | BCG papillary only 81% | D+BCG T1 88% | BCG T1 79% 36 months: D+BCG papillary only 84% | BCG papillary only 76% | D+BCG T1 85% | BCG T1 75% 48 months: D+BCG papillary only 84% | BCG papillary only 74% | D+BCG T1 85% | BCG T1 73% [Slide 4] POTOMAC: Cystectomy-Free Survival in Patients With Papillary Only and With T1 Tumours Cystectomy-free survival showed a favourable trend for D+BCG (I+M) vs BCG (I+M)a Papillary only D+BCG (I+M) N=217 | BCG (I+M) N=220 Total events, n (%): 31 (14) | 52 (24) Median CFS, months (95% CI): NR (NR–NR) | NR (78.3–NR) HR 0.59 (95% CI, 0.38–0.92) T1 D+BCG (I+M) N=137 | BCG (I+M) N=145 Total events, n (%): 18 (13) | 39 (27) Median CFS, months (95% CI): NR (NR–NR) | 78.3 (78.3–NR) HR 0.45 (95% CI, 0.25–0.78) No. of patients at risk: [illegible at this resolution] Survival table Endpoint: Cystectomy-free survival, papillary-only and T1 subsets Arm | N | Median | Median CI D+BCG (I+M) papillary only | 217 | NR | NR–NR BCG (I+M) papillary only | 220 | NR | 78.3–NR D+BCG (I+M) T1 | 137 | NR | NR–NR BCG (I+M) T1 | 145 | 78.3 | 78.3–NR HR: 0.59 (papillary only); 0.45 (T1) HR CI: 0.38–0.92; 0.25–0.78

[Slide 1] Figure 1. SUVmax: N188 vs FDG by lesion site T (Primary tumor) Spearman r=-0.096 p=0.695, n=19 95% CI [-0.53, 0.37] N (Lymph nodes) Spearman r=0.657 p<0.001, n=76 95% CI [0.51, 0.77] M (Visceral) Spearman r=0.210 p=0.148, n=49 95% CI [-0.08, 0.46] Bone (Osseous) Spearman r=-0.001 p=0.994, n=33 95% CI [-0.34, 0.34] Axes: Nectin4 SUVmax / FDG SUVmax Figure 2. TBR (SUVmax/BP): N188 vs FDG by lesion site T (Primary tumor) Spearman r=0.058 p=0.820, n=18 95% CI [-0.42, 0.51] N (Lymph nodes) Spearman r=0.635 p<0.001, n=76 95% CI [0.48, 0.75] M (Visceral) Spearman r=0.466 p<0.001, n=49 95% CI [0.21, 0.66] Bone (Osseous) Spearman r=0.298 p=0.092, n=33 95% CI [-0.05, 0.58] Axes: Nectin4 TBR (SUVmax/BP) / FDG TBR (SUVmax/BP) Lines: OLS fit with 95% CI band, for visualisation only; statistics are Spearman rank-based. [Slide 3] Figure 4. Case presentation: [18F]-FDG (a-c) vs [68Ga]-N188 (d-f) [18F]-FDG-PET/CT [68Ga]-N188-PET/CT a b c d e f

[Slide 2] Randomized (n=36) 1:1 Allocation Cabozantinib alone Allocated to intervention (n=18) Ineligible (n=2) Reasons: No evidence of metastatic papillary RCC (n=1), insufficient baseline procedures (n=1) Cabozantinib + Atezolizumab Allocated to intervention (n=18) Ineligible (n=1) Reason: Insufficient baseline procedures (n=1) Follow-up Eligible (n=16) Response Assessment: Determinable (n=15) Not Determinable (n=1) Consent withdrawn (n=1) Eligible (n=17) Response Assessment: Determinable (n=14) Not Determinable (n=3) Analysis Analyzed for primary outcome (n=16) Analyzed for safety (n=16) Analyzed for primary outcome (n=17) Analyzed for safety (n=17) Activated September 19, 2022 Closed January 16, 2026 Data cut off July 20, 2026 [Slide 4] Adverse Events Two grade 5 events were recorded – Respiratory failure (cabo, n=1) Lower gastrointestinal hemorrhage (cabo/atezo, n=1) Adverse Event | Cabozantinib (n=16) All Grade | Grade 3/4 | Cabozantinib/Atezolizumab (n=17) All Grade | Grade 3/4 Total | 16 (100%) | 12 (75%) | 17 (100%) | 11 (65%) Diarrhea | 11 (69%) | 0 (0%) | 11 (65%) | 1 (6%) Fatigue | 10 (62%) | 3 (19%) | 12 (71%) | 0 (0%) AST increased | 12 (75%) | 1 (6%) | 9 (53%) | 0 (0%) Hypertension | 9 (56%) | 1 (6%) | 10 (59%) | 6 (35%) Hand-Foot syndrome | 9 (56%) | 1 (6%) | 9 (53%) | 2 (12%) ALT increased | 9 (56%) | 1 (6%) | 6 (35%) | 2 (12%) Nausea | 6 (38%) | 1 (6%) | 7 (41%) | 0 (0%) TSH increased | 8 (50%) | 0 (0%) | 5 (29%) | 0 (0%) Anemia | 5 (31%) | 0 (0%) | 7 (41%) | 0 (0%) Anorexia | 6 (38%) | 3 (19%) | 6 (35%) | 0 (0%) Mucositis oral | 6 (38%) | 0 (0%) | 6 (35%) | 0 (0%) Dysgeusia | 6 (38%) | 0 (0%) | 5 (29%) | 0 (0%) Hyponatremia | 4 (25%) | 0 (0%) | 7 (41%) | 0 (0%) Proteinuria | 5 (31%) | 1 (6%) | 6 (35%) | 0 (0%) Hypoalbuminemia | 5 (31%) | 0 (0%) | 4 (24%) | 0 (0%) Hypocalcemia | 4 (25%) | 0 (0%) | 5 (29%) | 0 (0%) Hypothyroidism | 5 (31%) | 0 (0%) | 4 (24%) | 0 (0%) Platelet count decreased | 6 (38%) | 0 (0%) | 3 (18%) | 0 (0%) Weight loss | 5 (31%) | 1 (6%) | 4 (24%) | 1 (6%) Hoarseness | 5 (31%) | 0 (0%) | 3 (18%) | 0 (0%) Neutrophil count decreased | 6 (38%) | 1 (6%) | 2 (12%) | 0 (0%) Creatinine increased | 5 (31%) | 0 (0%) | 2 (12%) | 0 (0%) Hyperglycemia | 4 (25%) | 0 (0%) | 3 (18%) | 0 (0%) Eosinophilia | 4 (25%) | 0 (0%) | 2 (12%) | 0 (0%) Muscle cramp | 4 (25%) | 0 (0%) | 2 (12%) | 1 (6%) Hair color changes | 4 (25%) | 0 (0%) | 1 (6%) | 0 (0%) Headache | 4 (25%) | 0 (0%) | 1 (6%) | 0 (0%) Vomiting | 4 (25%) | 0 (0%) | 1 (6%) | 0 (0%) Rash maculo-papular | 0 (0%) | 0 (0%) | 4 (24%) | 0 (0%)

Figure 3a: Best PSA response as percentage change from baseline in all pts (N = 22) Best percentage change from baseline PSA PSA50 60 MBq (Q6W) | 120 MBq (Q4W) | 160 MBq (Q4W, Q2W) | 200 MBq (Q4W, Q2W, Q1W) Figure 3b: Best percentage change in target lesion sum of diameters from baseline in RECIST-evaluable pts (n = 6), by BICR cPR -42% | cPR -67% | cPR -77% | uCR -82% | cCR -100% | uCR [a] Best percentage change in SOD from baseline PD PR CR 200 MBq (Q4W, Q2W, Q1W) | 160 MBq (Q4W, Q2W) | 120 MBq (Q4W)

[World map of study sites] Legend: Active (dark blue) - United States, Canada, South Korea; Planned (light blue) - Brazil, Argentina, United Kingdom, France, Spain, Italy, Germany/central Europe, Turkey, China, Japan [country shading approximate, read from map] Active Planned

[Slide 1] PATIENT 9 | INTRAPATIENT DLL3 HETEROGENEITY Strong DLL3 expression on IHC (H-score 280) contrasted with limited PET uptake, with only 10/41 lesions (24%) DLL3-avid, highlighting marked intrapatient heterogeneity. Osseous metastases involving the ribs with soft tissue component showing FDG and DLL3 positivity (red arrow) while lung metastasis shows FDG avidity without DLL3 avidity (green arrow). DLL3 FDG PSMA DLL3 FDG [Slide 2] PATIENT 13 | DLL3 PET-GUIDED BIOPSY After multiple lines of systemic therapy for prostate adenocarcinoma, DLL3 PET revealed a heterogeneous pattern of target expression, with only 4 of 40 assessed lesions (10%) demonstrating detectable DLL3 uptake. A DLL3-avid lesion was subsequently selected for biopsy, which confirmed neuroendocrine differentiation and high DLL3 expression by immunohistochemistry (H-score 170). The concordance between imaging and tissue illustrates how DLL3 PET may help identify lesions that provide clinically meaningful information about tumor biology. This case supports the potential role of whole-body DLL3 imaging in complementing conventional tissue sampling by directing biopsy toward lesions with evidence of the target of interest. DLL3 FDG

Aim 1. Patients with prostate cancer and F1LCDx N = 18,691 Filtering: 15,766 no clinical data Clinical data available N = 2,925 Filtering: 1,846 did not receive PARPi Received PARPi N = 430 Filtering 7 missing sample collection date Aim 3. Sample before PARPi N = 366 Sample after PARPi N = 66* *9 patients had paired samples Filtering: 20 mHSPC; 61 missing information; 3 entered after PARPi therapy; 194 with no BRCA1/2, PALB2, RAD51C/D, BARD1 alteration**; 27 PARPi combo Aim 2. PARPi monotherapy and mCRPC with HRR alteration N = 61

[Slide 1] WHY WAS NEOADJUVANT THERAPY GIVEN ? Downstage vena cava thrombus 41% Facilitate resecability 47% Facilitate nephron-sparing surgery 12% [Slide 2] SURGICAL OUTCOMES Type of surgery | Number of patients (N = 17) Radical nephrectomy | 10 Partial nephrectomy | 5 Radical nephrectomy + controlateral partial nephrectomy | 1 Cryotherapy | 1 R0 resection 93% 2 postoperative complications (12.5%) : Urinary fistula, Pseudoaneurysm [Slide 3] TUMOR RESPONSE (16 patients evaluable) ORR 68% Best response (RECIST v1.1), n PR 11 SD 5 CR 0 [Slide 4] PATHOLOGICAL OUTCOMES AFTER NEOADJUVANT Figure 1 : Pathological stage after neoadjuvant treatment Number of patients Pathological stage pT0 pT1a pT1b pT2a pT3a pTx Pathological features, n (%) Tumor necrosis | 9 (53) Venous tumor thrombus | 7 (41) Sarcomatoid component | 3 (17)

[Slide 1] CURRENT STANDARD Dual checkpoint blockade: a durable tail at 9 years Anti-CTLA-4 Ipilimumab: releases T-cell priming in the lymph node Anti-PD-1 Nivolumab: restores effector function in the tumour CheckMate 214 N = 1,096 · first-line aRCC NIVO + IPI ×4 → NIVO vs sunitinib Median follow-up 9.3 years 31.4% vs 19.5% OS at 9 years · HR 0.71 (0.62–0.82) 22.7% vs 9.0% PFS at 8 years: the plateau But: ORR 39.5% and median PFS 12.4 vs 12.3 mo: the tail is a minority. ~70–75% relapse on combination therapy. A Overall survival B Progression-free survival per IRRC NIVO + IPI SUN Months Choueiri TK et al. Ann Oncol 2026;37:960. Burgers FH, van der Mijn JCK et al. Nat Rev Nephrol 2025. Survival table Endpoint: OS (CheckMate 214, ITT) NIVO + IPI Sunitinib HR: 0.71 HR CI: 0.62–0.82 9-year OS 31.4% vs 19.5% 8-year PFS 22.7% vs 9.0% [Slide 2] RESISTANCE Why most patients still progress on checkpoint inhibitors Cancer cell characteristics Intratumoural heterogeneity Low tumour mutational burden Tumour microenvironment Few B cells and TLS Dysfunctional DCs outside TLS B cell T cell Immature DC High infiltration of dysfunctional T cells ↑PD-1 ↑LAG-3 ↑CTLA-4 ↑TIM-3 Immunosuppressive cell subsets M2-likeTAMs MDSCs Treg cell IL-10, TGFβ NEW: HYPOXIA VEGFR TKI/IO responders ccRCC ↑VEGF Vhl-/- pseudohypoxia VEGFR TKI chronic VEGFR TKI SPP1 TAM True TME hypoxia metastasis • VHL loss = pseudohypoxia; normoxic HIF signaling • High pretreatment hypoxia signatures predict worse IO/TKI outcome (JAVELIN Renal 101, IMmotion150) • TKI-induced hypoxia recruits SPP1+ TAMs [Slide 3] NEW CHECKPOINTS LAG-3: the only new checkpoint with a phase 3 win RELATIVITY-047 · PFS · 1L melanoma · n = 714 Progression-free Survival (%) Relatlimab–nivolumab Nivolumab Months No. at Risk PFS 10.1 vs 4.6 mo, HR 0.75 (0.62–0.92) OS 51.0 vs 34.1 mo, HR 0.80 LAG-3 agents beyond relatlimab Agent | Latest result Fianlimab | Phase 3 1L melanoma (+ cemiplimab vs pembro): PFS 11.5 vs 6.4 mo, HR 0.85, not significant (May 2026); head-to-head vs nivo/rela ongoing Favezelimab | KEYFORM-007 (MSS CRC) stopped for futility (2024) Eftilagimod alfa Soluble LAG-3, APC agonist | TACTI-002 NSCLC ORR 48%; phase 3 TACTI-004 stopped for futility (Mar 2026) TIGIT PHASE 3 FAILURES Tiragolumab, vibostolimab, belrestotug, domvanalimab: all negative or futile (2024–26) TIM-3 PHASE 3 FAILURE Cobolimab: COSTAR Lung missed OS (2025); sabatolimab stopped in MDS VISTA PHASE 1 KVA12123, SNS-101: tolerable, stable disease only Tawbi HA et al. NEJM 2022;386:24. Fianlimab NCT05352672. KEYFORM-007. TACTI-004 (Mar 2026). TIGIT/TIM-3: Clinical Trials Arena 2025; ApexOnco 2026. Survival table Endpoint: PFS (RELATIVITY-047, 1L melanoma) Relatlimab–nivolumab | 355 | 10.1 mo Nivolumab | 359 | 4.6 mo HR: 0.75 HR CI: 0.62–0.92 [Slide 4] In RCC, LAG-3 and TIGIT have not improved on the backbone Confirmed ORR (%) · KEYMAKER-U03 substudy 03A, first-line ccRCC Pembrolizumab + lenvatinib (reference) 80.6 Belzutifan + pembro + lenvatinib 77.5 Quavonlimab (CTLA-4)/pembro + lenvatinib 71.3 Favezelimab (LAG-3)/pembro + lenvatinib 62.7 Vibostolimab (TIGIT)/pembro + belzutifan 42.5 Substudy 03B · previously treated Favezelimab (LAG-3) + pembrolizumab: no objective responses (6 arms, n = 263) VISTA · KVA12123 ± pembrolizumab Phase 1: one ICI-refractory RCC patient with stable disease (−24%) TIM-3 Strong rationale (TIM-3+ dysfunctional CD8+ TILs), but no RCC efficacy data reported Suárez C et al. Ann Oncol 2026. KVA12123: Kineta, Mar 2024. Burgers FH, van der Mijn JCK et al. Nat Rev Nephrol 2025.

[Slide 2] IS PRECISION MEDICINE POSSIBLE BASED ON TCGA GENE EXPRESSION SUBTYPES? Lerner,…Kwiatkowski; TCGA (N=412), ASCO 2017, CELL 2017 Luminal: KRT20+, GATA3+, FOXA1+ Basal/Squamous: KRT5,6,14+, GATA3-, FOXA1- Luminal-papillary: FGFR3 mut, fusion, amp; Papillary histology; SHH+; Low CIS → Low risk; NAC*; FGFR3 inhibitors * Low predicted likelihood of response, based on preliminary data Luminal-infiltrated: Low purity; EMT markers (TWIST1, ZEB1); miR-200 family; Medium CD274 (PD-L1), CTLA-4; Myofibroblast markers; Wild type p53 → Anti-PD-L1, PD-1, CTLA-4; Cisplatin-based NAC** ** Low response rate Luminal: UPKs; KRT20; SNX31 → Targeted therapy? Basal/Squamous: Female; Squamous differentiation; Basal keratin markers; High CD274 (PD-L1), CTLA4; Immune infiltrates → Anti-PD-L1, PD-1, CTLA-4; Cisplatin-based NAC Neuronal: SOX2; DLX6; MSI1; PLEKHG4B; E2F3/SOX4 amp; High cell cycle → Etoposide/Cisplatin NAC [Slide 3] There is a relationship between molecular subtypes and response to cisplatin based Neo-adjuvant in MIBC. Non-NAC TCGA dataset (n = 397) | Discovery NAC dataset (n = 200) UNC: p = 0.009 | p = 0.014 (legend: Claudin, Basal, Luminal) → Basal ↑ Lund: p = 0.038 | p = 0.3 → Urobasal B ↑ TCGA: p < 0.0001 | p = 0.09 (legend: Cluster IV, Cluster III, Cluster II, Cluster I) Time (mo) Seiler et al. European Urology 2017 Survival table Endpoint: Survival by molecular subtype (Seiler et al. Eur Urol 2017; endpoint not labelled) P: UNC 0.009 / 0.014; Lund 0.038 / 0.3; TCGA <0.0001 / 0.09 (non-NAC / NAC datasets) [Slide 4] Molecular types and I-O B T-effector gene expression vs response (Progressive disease, Stable disease, Partial response, Complete response) CXCL9 p=0·0057; CXCL10 p=0·0079 Log2 gene expression Interferon-γ-inducible T-helper-1-type chemokines D CD8 infiltration vs response p=0·0265 % CD8 in tumour area: Stable/progressive disease vs Complete/partial response CD8+ T-cell infiltration E IC score by TCGA subtype (IC0, IC1, IC2, IC3): Cluster I n=72, Cluster II n=50 (Luminal); Cluster III n=38, Cluster IV n=35 (Basal) G Response by TCGA subtype (Progressive disease, Stable disease, Partial response, Complete response): Cluster I n=72, Cluster II n=50, Cluster III n=38, Cluster IV n=35 Discussant : Prof. Syed A Hussain, University of Sheffield Rosenberg et al. Lancet 2016; 387: 1909–20

Figure 2. Global Himalayas-1 Study Footprint [World map with participating countries shaded, including North America, Mexico, Brazil, Argentina, Western Europe, Turkey, India, China, Japan, South Korea, Southeast Asia, Australia] himalayasprostatestudy.com

[Slide 1] NIAGARA: Study Design Study population • Adults • Cisplatin-eligible MIBC (cT2–T4aN0/1M0) • UC or UC with divergent differentiation or histological subtypes • Evaluated and confirmed for RC • CrCl of ≥40 mL/min R 1:1 BEPs: WES n=407; RNA-seq n=402 Durvalumab arm N=533 Neoadjuvant Q3W, 4 cycles: Durvalumab 1500 mg IV + Gemcitabine + cisplatin RC Adjuvant Q4W, 8 cycles: Durvalumab 1500 mg IV Comparator arm N=530: Gemcitabine + cisplatin → No treatment Primary endpoints • EFS • pCR Key secondary endpoint • OS Exploratory analyses • MIBC molecular drivers and subtypes in relation to treatment benefit • Dual DNA/RNA was extracted from baseline tumours for WES and RNA-seq analysis by Natera Inc. (Austin, TX, USA)a • Exploratory, non-alpha controlled assessment of molecular drivers and outcomes was conducted [Slide 2] NIAGARA: Biomarker-Evaluable Population Baseline Characteristics Baseline characteristics were generally similar between the BEPs and the ITT Baseline characteristicsa | WES BEP (n=407) | RNA-seq BEP (n=402) | ITT (N=1063)1 Treatment arm D+NAC | 51% | 52% | 50% Treatment arm NAC | 49% | 48% | 50% Age, years Median (range) | 66 (32–84) | 66 (32–84) | 65 (32–84) Sex Male | 82% | 83% | 82% Tumour stage >T2N0 | 60% | 60% | 57% Tumour stage T2N0 | 40% | 40% | 43% Renal function Adequate renal function | 81% | 81% | 81% Renal function Borderline renal function | 19% | 19% | 19% Regional lymph nodes N0 | 94% | 94% | 95% Regional lymph nodes N1 | 6% | 6% | 5% Histology UC | 84% | 85% | 85% Histology UC with divergent differentiation or histological subtypes | 16% | 15% | 15% PD-L1 expressionb High | 70% | 68% | 73% PD-L1 expressionb Low | 30% | 32% | 27% Cystectomyc Yes | 95% | 95% | 86% Cystectomyc No | 5% | 5% | 14% • Matched baseline WES and RNA-seq samples were available for n=352 patients [Slide 4] NIAGARA: EFS Benefit in Mutation-Defined Subgroups Addition of perioperative durvalumab to NAC consistently improved EFS across mutation-defined subgroups, with numerical increase of benefit for TMB-high and TP53- and FGFR3-mutated subgroups Population/subgroup | D arm | C arm | HR (95% CI) ITT (N=1063) | 533 | 530 | 0.68 (0.56–0.82) BEP (n=407) | 209 | 198 | 0.64 (0.45–0.89) TMB-high | 84 | 77 | 0.44 (0.25–0.76) TMB-low | 125 | 121 | 0.82 (0.53–1.26) TP53 mutated | 107 | 97 | 0.51 (0.32–0.81) TP53 WT | 102 | 101 | 0.81 (0.49–1.31) FGFR3 mutated | 20 | 24 | 0.24 (0.07–0.87) FGFR3 WT | 189 | 174 | 0.70 (0.49–0.99) TP53CC mutated (pathway) | 129 | 125 | 0.58 (0.38–0.88) RTK mutated (pathway) | 86 | 82 | 0.49 (0.28–0.85) DDR mutated (pathway) | 50 | 48 | 0.72 (0.36–1.43) Histone-mod mutated (pathway) | 83 | 85 | 0.63 (0.37–1.08) Chromatin remodelling (pathway) | 54 | 44 | 0.48 (0.23–0.98) EFS HR (log scale) Favours D arm | Favours C arm Graph depicts EFS HR (square) and 95% confidence intervals (lines), and was assessed by the Cox proportional hazards model, with treatment as the only covariate and ties handled by the Efron approach. BEP, biomarker-evaluable population; C, comparator; CC, cell cycle; CI, confidence interval; D, durvalumab; DDR, DNA damage response; EFS, event-free survival; HR, hazard ratio; ITT, intention-to-treat population; NAC, neoadjuvant chemotherapy; RTK, receptor tyrosine kinase; TMB, tumour mutational burden; WT, wild-type.

[Best Poster rosette] BioChiP: A phase II, open-label, biomarker-guided study of carboplatin efficacy in pretreated metastatic castration resistant prostate cancer Background [illegible] Design [illegible] Key elegibility criteria [illegible] Procedures [illegible] Biomarker analyses [illegible] Endpoints [illegible] Statistical analysis [illegible] Conclusions [illegible] Acknowledgments Personal information TAKE HOME MESSAGES • Without biomarker selection, carboplatin showed a combined response rate of 21.5%. • Higher response rates were associated with a RAD51-defined HRD phenotype and with BRCA1/2 alterations. Genomic landscape of study population Primary endpoint · Response Rate 21.5% 95% CI 12.3 - 33.5 BIOCHEMICAL RESPONSE RATE 16.9% 95% CI 8.3 - 28.3 RADIOGRAPHIC RESPONSE RATE 22.7% 95% CI 11.5 - 37.8 Key secondary endpoint RESPONSE RATE BY RAD51 [small table - largely illegible] Patients characteristics [illegible table] Secondary endpoints OVERALL POPULATION rPFS CLINICAL PFS OVERALL SURVIVAL Exploratory [illegible]

[Slide 1] Fig 1: Cumulative incidence of SCNEPC, by treatment group Cumulative Incidence of SCNEPC in ENZAMET (%) Enzalutamide Conventional NSAA Years Number at risk (event) Enzalutamide Conventional NSAA Survival table Endpoint: Cumulative incidence of SCNEPC Enzalutamide | 563 Conventional NSAA | 562 [Slide 2] Median OS (IQR) from SCNEPC diagnosis in ALL 19 pts 9 months (2-18), ENZA vs NSAA 9 vs 7, p=0.3 (see Fig 2) Fig 2: OS from diagnosis of SCNEPC, by treatment group Proportion Alive NSAA ENZA Months since SCNEPC diagnosis Number at risk (number of events) Survival table Endpoint: OS from SCNEPC diagnosis ENZA | 15 | 9 months NSAA | 4 | 7 months P: 0.3

[Slide 1] BARCODE2: Open-label, single arm phase II study Part 1: Genetic screening for actionable DRG PGV - Cohort: mCRPC with prior/ongoing treatment of docetaxel and/or abiraterone/enzalutamide - Screen using DRG panel (n=115) based on known associations of germline variants with mCRPC* → PGV absent: standard of care off study - PGV present: monitor until qualifying progression Part 2: Treatment with 3-weekly carboplatin (AUC5) - Progression after treatment of docetaxel and abiraterone/enzalutamide ← Direct entrants: mCRPC with eligible treatment history and prior confirmed PGV - Enter part 2 and start carboplatin - Follow-up including RECIST, PSA, bone scans and adverse events → Discontinue carboplatin after maximum 10 cycles or disease progression if occurring first [Slide 2] From May 2017 to June 2025: genetic profiling of 301/303 (99.3%) part 1 entrants - 7/301 (2.3%) part 2 ineligible mutation - 249/301 (82.7%) with no mutation - 45/301 (15.0%) with part 2 eligible variant Group 1 (BRCA): 9/301 (3.0%) [95% CI 1.4-5.6] → 9* entered trial; 2 did not → 6/9 (66.7%) started carboplatin Group 2 (Lynch syndrome): 3/301 (1.0%) [95% CI 0.2-2.9] → 1 entered trial; 2 did not → 1/1 (100%) started carboplatin Group 3 (other DRG): 33/301 (11.0%) [95% CI 7.7-15.1] → 14 entered trial; 19 did not → 12/14 started carboplatin (85.7%) [Slide 3] | | Group 1 (BRCA) (N=9) | Group 2 (Lynch syndrome) (N=1) | Group 3 (other DRG) (N=14) | | Target response in stage 1 & 2 | ≥1/6; ≥3/11 | ≥1/6; ≥3/11 | ≥2/12; ≥5/18 | | Stage 1 response rate (%) | 2/6 (33.3) [82% CI 8.7-67.7] | 0/1 (0) [91% 1-sided CI 0-91.0] | 0/11 (0) [95% 1-sided CI 0-23.8] | | Outcome | Target met; recruitment closed due to new treatment pathway | Insufficient data for analysis; recruitment closed | Target not met; recruitment closed |

[Slide 1] STAMPEDE (n=401) | 18% | 31% | 51% Commercial testing (n = 27,607) NCCN very high risk (n = 2,845) | 13% | 23% | 64% NCCN high risk (n = 24,762) | 33% | 29% | 37% Decipher risk group* GC > 0.85 (purple) GC 0.61 - 0.85 (red) GC ≤ 0.60 (blue) (bar order: GC ≤ 0.60 | GC 0.61-0.85 | GC > 0.85) [Slide 3] Left panel: STAMPEDE ADT Prostate cancer specific mortality (%) vs Years since randomisation (0-8) HR (95% CI): 1.43 (1.14 - 1.81) p-value: 0.002* Number at risk (% PCSM) GC > 0.85: 101 | 98 (2%) | 80 (15%) | 49 (26%) GC 0.61-0.85: 65 | 65 (0%) | 55 (8%) | 43 (12%) GC ≤ 0.60: 42 | 41 (2%) | 38 (5%) | 34 (7%) Right panel: STAMPEDE ADT + abi HR (95% CI): 1.60 (1.16 - 2.21) p-value: 0.004* Number at risk (% PCSM) GC > 0.85: 105 | 97 (4%) | 78 (10%) | 60 (22%) GC 0.61-0.85: 58 | 57 (0%) | 53 (2%) | 47 (5%) GC ≤ 0.60: 30 | 30 (0%) | 28 (3%) | 22 (3%) Decipher risk group: GC ≤ 0.60 | GC 0.61 - 0.85 | GC > 0.85 Survival table Endpoint: Prostate cancer-specific mortality by Decipher risk group (STAMPEDE ADT and ADT+abi cohorts) ADT: GC > 0.85 | 101 ADT: GC 0.61-0.85 | 65 ADT: GC ≤ 0.60 | 42 ADT + abi: GC > 0.85 | 105 ADT + abi: GC 0.61-0.85 | 58 ADT + abi: GC ≤ 0.60 | 30 HR: ADT: 1.43; ADT + abi: 1.60 HR CI: ADT: 1.14 - 1.81; ADT + abi: 1.16 - 2.21 P: ADT: 0.002*; ADT + abi: 0.004* Landmark rates: last at-risk column (time not labelled): ADT: GC > 0.85 26% | ADT: GC 0.61-0.85 12% | ADT: GC ≤ 0.60 7% | ADT + abi: GC > 0.85 22% | ADT + abi: GC 0.61-0.85 5% | ADT + abi: GC ≤ 0.60 3% [Slide 4] STAMPEDE standardized sHR for distant metastasis (y-axis 1.0-3.0) RWD standardized sHR for distant metastasis (x-axis 1.0-2.0) Labelled points: Decipher; pTEN loss Saal; p53 loss; pTEN loss Liu; Zhang basal; PORTOS; Androgen response HM Legend: DNA repair | Hallmarks of cancer | Prostate subtyping | Risk models

[Slide 1] Table 2. HRR Gene Alteration Frequencies* Among Patients With Informative NGS Results in the TALAPRO-2 China Cohort | TALA + ENZA (n=31) | PBO + ENZA (n=32) | Total (N=63) ATM | 5 (16.1) | 2 (6.3) | 7 (11.1) ATR | 1 (3.2) | 1 (3.1) | 2 (3.2) BRCA1 | 1 (3.2) | 0 | 1 (1.6) BRCA2 | 5 (16.1) | 3 (9.4) | 8 (12.7) CDK12 | 3 (9.7) | 8 (25.0) | 11 (17.5) CHEK2 | 1 (3.2) | 1 (3.1) | 2 (3.2) FANCA | 1 (3.2) | 0 | 1 (1.6) MLH1 | 0 | 1 (3.1) | 1 (1.6) MRE11A | 0 | 0 | 0 NBN | 0 | 0 | 0 PALB2 | 0 | 1 (3.1) | 1 (1.6) RAD51C | 0 | 0 | 0 Data are presented as n (%), with percentages calculated using the number of patients with informative NGS results as the denominator. Analyses included all informative tumor tissue NGS results available for the China cohort, including prospectively available results and retrospectively analyzed results obtained after randomization. *Mutations detected in one or more genes from the 12-gene HRR panel, alone or in combination with mutations in other genes, among patients with informative NGS results in the TALAPRO-2 China cohort. Patients with alterations in more than one HRR gene were counted once for each affected gene; therefore, counts across genes were not mutually exclusive and should not be summed. ENZA=enzalutamide; HRR=homologous recombination repair; NGS=next-generation sequencing; PBO=placebo; TALA=talazoparib. [Slide 2] Figure 2. rPFS by BICR for Selected 12-Panel HRR Single-Gene Alterations Groups and Gene Alteration Clusters Based on Prospective Testing of Patients in the TALAPRO-2 China Cohort* Subgroup | TALA + ENZA No. of events/No. of patients | PBO + ENZA | TALA + ENZA Median (months) | PBO + ENZA | HR (95% CI) | p-value All HRR-deficient | 8/14 | 9/12 | 15.7 | 4.0 | 0.20 (0.07–0.61) | 0.0023 Only CHEK2† | 0/1 | 0/1 | NR | NR | NR | NR Only ATM† | 2/3 | 0/1 | 13.8 | NR | NR | NR Only BRCA2† | 2/4 | 0 | 27.7 | NR | NR | NR Only CDK12 | 2/3 | 5/6 | 15.7 | 4.0 | 0.14 (0.02–1.31) | 0.0513 BRCA cluster | 4/6 | 2/3 | 19.4 | 8.3 | 0.35 (0.05–2.62) | 0.2872 CDK12 cluster | 2/3 | 7/8 | 15.7 | 3.8 | 0.13 (0.02–1.10) | 0.0324 ATM cluster† | 2/4 | 0/1 | 13.8 | NR | NR | NR Other gene cluster† | 0/1 | 0 | NR | NR | NR | NR Hierarchy Favors TALA + ENZA | Favors PBO + ENZA HRs for all subgroups were based on an unstratified Cox model with treatment as the only covariate; p-values are 2-sided. For the single gene groupings, only patients bearing alteration(s) in the designated HRR gene and no other HRR genes are displayed. *HRR gene alteration status was based on prospective tumor tissue testing results available at or before randomization; †HR (95% CI) cannot be calculated. BICR=blinded independent central review; CI=confidence interval; ENZA=enzalutamide; HR=hazard ratio; HRR=homologous recombination repair; NR=not reached; PBO=placebo; rPFS=radiographic progression-free survival; TALA=talazoparib. [Slide 3] Figure 3. Heatmap of Tumor Mutations by Response Category in the Evaluable Biomarker Analysis Population of the TALAPRO-2 China Cohort (n=54) Treatment Response PTEN (29.6%) TP53 (29.6%) MLL2 (27.8%) ATM (24.1%) BRCA2 (18.5%) CDK12 (18.5%) APC (16.7%) PTCH1 (16.7%) IRS2 (14.8%) TET2 (14.8%) TMPRSS2 (14.8%) ATR (13.0%) FGFR3 (13.0%) LTK (13.0%) MSH3 (13.0%) RAD21 (13.0%) STK11 (13.0%) TSC2 (13.0%) ZNF217 (13.0%) BRAF (11.1%) KDM6A (11.1%) SPEN (11.1%) SPOP (11.1%) Treatment: PBO + ENZA; TALA + ENZA Response: CR; NCR/NPD; NE; PD; PR; SD Alterations: None; CN-Amplification; CN-Loss; Deletion; Insertion; RE-Deletion; RE-Fusion; RE-Other; RE-Truncation; SV; Multiple alterations [Slide 4] Figure 4. Exploratory Analysis of Associations Between Non-HRR Gene Alterations and rPFS in the TALAPRO-2 China Cohort TP53 MLL2 TET2 PTEN P-value (log scale) HR (log scale)