ESMO 2026 GU Cancer Slides

[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 data captured as a table — no point estimates read from the curve] 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 {'timepoint': '24 months (dashed line)', 'arm': 'EV + pembro, ctDNA−', 'value': '86.1%', 'ci': None} {'timepoint': '24 months (dashed line)', 'arm': 'Cis + gem, ctDNA−', 'value': '82.7%', 'ci': None} {'timepoint': '24 months (dashed line)', 'arm': 'EV + pembro, ctDNA+', 'value': '71.2%', 'ci': None} {'timepoint': '24 months (dashed line)', 'arm': 'Cis + gem, ctDNA+', 'value': '54.2%', 'ci': None} Note: Two HRs (one per ctDNA stratum), unstratified Cox with treatment as covariate. Landmark-to-arm assignment by curve/label colour (light teal=EV+P ctDNA−, pink=Cis+gem ctDNA−, dark teal=EV+P ctDNA+, dark red=Cis+gem ctDNA+); timepoint inferred from the dashed line at 24 months, not explicitly labelled '24-mo'. Medians not printed. Data cutoff Oct 27, 2025. [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 data captured as a table — no point estimates read from the curve] 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 {'timepoint': '24 months (dashed line)', 'arm': 'EV + pembro Negative-Negative', 'value': '94.6%', 'ci': None} {'timepoint': '24 months (dashed line)', 'arm': 'EV + pembro Positive-Negative', 'value': '81.1%', 'ci': None} {'timepoint': '24 months (dashed line)', 'arm': 'EV + pembro Positive-Positive', 'value': '33.3%', 'ci': None} {'timepoint': '24 months (dashed line)', 'arm': 'Cis + gem Negative-Negative', 'value': '87.8%', 'ci': None} {'timepoint': '24 months (dashed line)', 'arm': 'Cis + gem Positive-Negative', 'value': '63.2%', 'ci': None} {'timepoint': '24 months (dashed line)', 'arm': 'Cis + gem Positive-Positive', 'value': '35.3%', 'ci': None} Note: ctDNA clearance rate: EV + pembro 82.1% (69/84) vs Cis + gem 66.7% (68/102). No HR printed. Landmark-to-subgroup assignment by curve/label colour; timepoint inferred from dashed line at 24 months. Data cutoff Oct 27, 2025. [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 data captured as a table — no point estimates read from the curve] 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 {'timepoint': '24 months (dashed line)', 'arm': 'EV + pembro, ctDNA−', 'value': '89.9%', 'ci': None} {'timepoint': '24 months (dashed line)', 'arm': 'Cis + gem, ctDNA−', 'value': '82.2%', 'ci': None} {'timepoint': '24 months (dashed line)', 'arm': 'EV + pembro, ctDNA+', 'value': '10.0%', 'ci': None} {'timepoint': '24 months (dashed line)', 'arm': 'Cis + gem, ctDNA+', 'value': '10.0%', 'ci': None} Note: Unstratified Cox with treatment as covariate, HR per ctDNA stratum. Landmark assignment by colour; timepoint inferred from dashed line at 24 months. Medians not printed. Data cutoff Oct 27, 2025.

[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. [ESMO Congress Madrid 2026 logo; podium sign MADRID SPAIN 23-27 OCTOBER 2026; photo of speaker at podium and panel] [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. [ESMO Congress Madrid 2026 logo; speaker video panel; previous slide partially visible at left edge] [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 [ESMO Congress Madrid 2026 logo; speaker video panel; podium MADRID SPAIN 23-27 OCTOBER 2026]

[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). [ESMO Congress Madrid 2026 logo] [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. [ESMO Congress Madrid 2026 logo] [Survival data captured as a table — no point estimates read from the curve] 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) {'timepoint': '24 mo', 'arm': 'EV + pembro Negative-Negative', 'value': '94.6%', 'ci': None} {'timepoint': '24 mo', 'arm': 'EV + pembro Positive-Negative', 'value': '81.1%', 'ci': None} {'timepoint': '24 mo', 'arm': 'EV + pembro Positive-Positive', 'value': '33.3%', 'ci': None} {'timepoint': '24 mo', 'arm': 'Cis + gem Negative-Negative', 'value': '87.8%', 'ci': None} {'timepoint': '24 mo', 'arm': 'Cis + gem Positive-Negative', 'value': '63.2%', 'ci': None} {'timepoint': '24 mo', 'arm': 'Cis + gem Positive-Positive', 'value': '35.3%', 'ci': None} Note: 24-month landmark-to-arm mapping inferred from curve position/color (top curve = Negative-Negative, middle = Positive-Negative clearers, bottom = Positive-Positive). No medians or HRs printed. Footnote: EFS differences cannot be attributed to the neoadjuvant phase alone. [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. [ESMO Congress Madrid 2026 logo]

[Slide 2] [Photo of ESMO Congress Madrid 2026 stage screen; presenter video inset at right] 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. [MADRID 2026 ESMO congress logo]

[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 [Video inset: Jonathan Chatzkel - ctDNA analysis of the Phase 3 KEYNOTE-B15 trial: neoadjuvant and adjuvant (neo/adj) enfortumab vedotin (EV) + pembrolizumab (pembro) for participants (pts) with muscle-invasive bladder cancer (MIBC) who are eligible for cisplatin] [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 data captured as a table — no point estimates read from the curve] 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 {'timepoint': '24 months', 'arm': 'ctDNA- EV + pembro', 'value': '86.1%', 'ci': None} {'timepoint': '24 months', 'arm': 'ctDNA- Cis + gem', 'value': '82.7%', 'ci': None} {'timepoint': '24 months', 'arm': 'ctDNA+ EV + pembro', 'value': '71.2%', 'ci': None} {'timepoint': '24 months', 'arm': 'ctDNA+ Cis + gem', 'value': '54.2%', 'ci': None} Note: Unstratified Cox model with treatment as covariate. Medians not printed. DCO October 27, 2025. Landmark-to-arm mapping by label colour/position at the 24-mo dashed line. [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 data captured as a table — no point estimates read from the curve] 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 {'timepoint': '24 months', 'arm': 'EV + pembro Negative-Negative', 'value': '94.6%', 'ci': None} {'timepoint': '24 months', 'arm': 'EV + pembro Positive-Negative', 'value': '81.1%', 'ci': None} {'timepoint': '24 months', 'arm': 'EV + pembro Positive-Positive', 'value': '33.3%', 'ci': None} {'timepoint': '24 months', 'arm': 'Cis + gem Negative-Negative', 'value': '87.8%', 'ci': None} {'timepoint': '24 months', 'arm': 'Cis + gem Positive-Negative', 'value': '63.2%', 'ci': None} {'timepoint': '24 months', 'arm': 'Cis + gem Positive-Positive', 'value': '35.3%', 'ci': None} Note: ctDNA clearance EV+pembro 82.1% (69/84) vs cis+gem 66.7% (68/102). No HR/medians printed. DCO October 27, 2025.

[Slide 1] 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 data captured as a table — no point estimates read from the curve] Endpoint: INV-assessed DFS (ctDNA-negative post-RC, no treatment cohort) No treatment | 357 | NE | NE {'timepoint': '12 mo', 'arm': 'No treatment', 'value': '95.4%', 'ci': None} {'timepoint': '24 mo', 'arm': 'No treatment', 'value': '88.4%', 'ci': None} Note: Single-arm (observational) cohort; no comparator. 'Median 21.8 mo fro...' text is cropped at right edge (likely median follow-up from cystectomy) - do not interpret. 8 DFS events were deaths not clearly attributed to recurrence.

[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 data captured as a table — no point estimates read from the curve] Endpoint: EFS per BICR by baseline ctDNA status EV + pembro: ctDNA− | 121 Cis + gem: ctDNA− | 102 EV + pembro: ctDNA+ | 133 Cis + gem: ctDNA+ | 134 {'timepoint': '24 months', 'arm': 'EV + pembro: ctDNA−', 'value': '86.1%', 'ci': None} {'timepoint': '24 months', 'arm': 'Cis + gem: ctDNA−', 'value': '82.7%', 'ci': None} {'timepoint': '24 months', 'arm': 'EV + pembro: ctDNA+', 'value': '71.2%', 'ci': None} {'timepoint': '24 months', 'arm': 'Cis + gem: ctDNA+', 'value': '54.2%', 'ci': None} Note: Subgroup HRs: ctDNA− 0.68 (0.37-1.26); ctDNA+ 0.54 (0.37-0.80); unstratified Cox. Landmark-to-arm mapping by curve label color/position. Data cutoff October 27, 2025. [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 data captured as a table — no point estimates read from the curve] Endpoint: EFS per BICR by postsurgery ctDNA status EV + pembro: ctDNA− | 199 Cis + gem: ctDNA− | 193 EV + pembro: ctDNA+ | 10 Cis + gem: ctDNA+ | 30 {'timepoint': '24 months', 'arm': 'EV + pembro: ctDNA−', 'value': '89.9%', 'ci': None} {'timepoint': '24 months', 'arm': 'Cis + gem: ctDNA−', 'value': '82.2%', 'ci': None} {'timepoint': '24 months', 'arm': 'EV + pembro: ctDNA+', 'value': '10.0%', 'ci': None} {'timepoint': '24 months', 'arm': 'Cis + gem: ctDNA+', 'value': '10.0%', 'ci': None} Note: Subgroup HRs: ctDNA− 0.48 (0.29-0.81); ctDNA+ 0.63 (0.30-1.35). Data cutoff October 27, 2025.

[Slide 3] [Photo of main screen with presenter inset] 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. [ESMO Congress Madrid 2026 logo] [Survival data captured as a table — no point estimates read from the curve] 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 {'timepoint': '24 mo (dashed line)', 'arm': 'ctDNA− EV + pembro', 'value': '86.1%', 'ci': None} {'timepoint': '24 mo (dashed line)', 'arm': 'ctDNA− Cis + gem', 'value': '82.7%', 'ci': None} {'timepoint': '24 mo (dashed line)', 'arm': 'ctDNA+ EV + pembro', 'value': '71.2%', 'ci': None} {'timepoint': '24 mo (dashed line)', 'arm': 'ctDNA+ Cis + gem', 'value': '54.2%', 'ci': None} Note: Unstratified Cox HR with treatment as covariate. Medians not printed. Landmark percentages printed at a dashed line drawn at 24 months (timepoint not labeled in text). DCO Oct 27, 2025.

[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 4] 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 data captured as a table — no point estimates read from the curve] 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 {'timepoint': '24 months', 'arm': 'ctDNA- EV + pembro', 'value': '89.9%', 'ci': None} {'timepoint': '24 months', 'arm': 'ctDNA- Cis + gem', 'value': '82.2%', 'ci': None} {'timepoint': '24 months', 'arm': 'ctDNA+ EV + pembro', 'value': '10.0%', 'ci': None} {'timepoint': '24 months', 'arm': 'ctDNA+ Cis + gem', 'value': '10.0%', 'ci': None} Note: Unstratified Cox model. Medians not printed. DCO October 27, 2025. Both ctDNA+ arms labelled 10.0% at 24 mo (teal label for EV+P, dark label for cis+gem).

[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 4] 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 data captured as a table — no point estimates read from the curve] 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 {'timepoint': '24 mo', 'arm': 'Cis + gem, ctDNA−', 'value': '82.2%', 'ci': None} {'timepoint': '24 mo', 'arm': 'EV + pembro, ctDNA+', 'value': '10.0%', 'ci': None} {'timepoint': '24 mo', 'arm': 'Cis + gem, ctDNA+', 'value': '10.0%', 'ci': None} Note: EV + pembro ctDNA− 24-mo landmark label is washed out/illegible in the photo (omitted, not estimated). EV+pembro ctDNA− n and events are faint (199; 22 (11.1)) — verify against a clean copy. 82.2% label partially obscured by tick marks. DCO Oct 27, 2025.

[Slide 1] [ESMO Congress Madrid 2026 logo] 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 [ESMO logo] [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] [ESMO Congress Madrid 2026 logo] [Survival data captured as a table — no point estimates read from the curve] 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) {'timepoint': '24 mo', 'arm': 'KEYNOTE-B15 EV + pembro pre-chemo ctDNA+', 'value': '71.2%', 'ci': None} {'timepoint': '24 mo', 'arm': 'KEYNOTE-B15 Cis + gem pre-chemo ctDNA-', 'value': '82.7%', 'ci': None} {'timepoint': '24 mo', 'arm': 'KEYNOTE-B15 Cis + gem pre-chemo ctDNA+', 'value': '54.2%', 'ci': None} {'timepoint': '24 mo', 'arm': 'KEYNOTE-B15 EV + pembro pre-surgery Negative-Negative', 'value': '94.6%', 'ci': None} {'timepoint': '24 mo', 'arm': 'KEYNOTE-B15 EV + pembro pre-surgery Positive-Positive', 'value': '33.3%', 'ci': None} {'timepoint': '24 mo', 'arm': 'KEYNOTE-B15 Cis + gem pre-surgery Negative-Negative', 'value': '87.8%', 'ci': None} {'timepoint': '24 mo', 'arm': 'KEYNOTE-B15 Cis + gem pre-surgery Positive-Positive', 'value': '35.3%', 'ci': None} Note: EV+pembro pre-chemo ctDNA- 24-mo label illegible (not filled). n = number at risk at month 0. NIAGARA clearance HR too small to read. Faint ctDNA- at-risk rows are best readings. [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: [...] [ESMO Congress Madrid 2026 logo; speaker on video inset] [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. [ESMO Congress Madrid 2026 logo] [Survival data captured as a table — no point estimates read from the curve] 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 {'timepoint': '24 mo', 'arm': 'EV + pembro ctDNA-', 'value': '89.9%', 'ci': None} {'timepoint': '24 mo', 'arm': 'Cis + gem ctDNA-', 'value': '82.2%', 'ci': None} {'timepoint': '24 mo', 'arm': 'ctDNA+ (both arms)', 'value': '10.0% / 10.0%', 'ci': None} Note: KEYNOTE-B15 HRs from unstratified Cox model with treatment as covariate. Two 10.0% labels for ctDNA+ arms at 24 mo (arm mapping not distinguishable). Data cutoff October 27, 2025.

[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)

[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. [ESMO Congress Madrid 2026 logo; presenter inset at podium, 'MADRID SPAIN 23-27 OCTOBER 2026'] [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. [ESMO Congress Madrid 2026 logo; presenter inset]

[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).

[Slide 1] 16:30 - 18:15 Presidential Symposium I CHAIRS: FABRICE ANDRÉ, JAMES LARKIN Thomas Powles — Perioperative durvalumab with or without tremelimumab plus neoadjuvant enfortumab vedotin in muscle-invasive bladder cancer: Results from the phase 3 VOLGA trial ALICANTE AUDITORIUM - HALL 6 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 data captured as a table — no point estimates read from the curve] Endpoint: EFS by BICR (ITT) 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 {'timepoint': '12 mo', 'arm': 'D+EV', 'value': '73%', 'ci': None} {'timepoint': '12 mo', 'arm': 'Control', 'value': '50%', 'ci': None} {'timepoint': '24 mo', 'arm': 'D+EV', 'value': '64%', 'ci': None} {'timepoint': '24 mo', 'arm': 'Control', 'value': '45%', 'ci': None} Note: Stratified log-rank P; significance boundary 0.0218 (2.7% overall 2-sided alpha). Median follow-up 23 months. Data cutoff 02 March 2026. [Slide 2] 16:30 - 18:15 Presidential Symposium I CHAIRS: FABRICE ANDRÉ, JAMES LARKIN Thomas Powles — Perioperative durvalumab with or without tremelimumab plus neoadjuvant enfortumab vedotin in muscle-invasive bladder cancer: Results from the phase 3 VOLGA trial ALICANTE AUDITORIUM - HALL 6 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 data captured as a table — no point estimates read from the curve] Endpoint: OS (ITT) D+EV | 233 | NR | NC–NC Control | 229 | NR | 37.7–NC HR: 0.58 HR CI: 0.40–0.83 P: 0.0027 {'timepoint': '12 mo', 'arm': 'D+EV', 'value': '87%', 'ci': None} {'timepoint': '12 mo', 'arm': 'Control', 'value': '81%', 'ci': None} {'timepoint': '24 mo', 'arm': 'D+EV', 'value': '78%', 'ci': None} {'timepoint': '24 mo', 'arm': 'Control', 'value': '69%', 'ci': None} Note: Events are deaths. Stratified log-rank P; significance boundary 0.0104 (5% overall 2-sided alpha, Gamma family (-3) spending). Median follow-up 26 months. Data cutoff 02 March 2026.

[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 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] 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 data captured as a table — no point estimates read from the curve] Endpoint: EFS (durvalumab+EV arm) by ctDNA dynamics Persistent negative | 24 Clearance | 52 Persistent positive | 35 {'timepoint': '12 mo', 'arm': 'Persistent negative', 'value': '91%', 'ci': None} {'timepoint': '24 mo', 'arm': 'Persistent negative', 'value': '91%', 'ci': None} {'timepoint': '12 mo', 'arm': 'Clearance', 'value': '88%', 'ci': None} {'timepoint': '24 mo', 'arm': 'Clearance', 'value': '81%', 'ci': None} {'timepoint': '12 mo', 'arm': 'Persistent positive', 'value': '70%', 'ci': None} {'timepoint': '24 mo', 'arm': 'Persistent positive', 'value': '55%', 'ci': None} Note: Exploratory; n = at-risk at time 0; no medians/HR printed; ctDNA via Signatera Genome; landmark timepoints from dashed lines at 12 and 24 months

[Slide 4] 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.

[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 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

[Slide 2] VOLGA: EFS by BICR for Durvalumab+T+EV vs Control – ITT Durvalumab+T+EV also demonstrated a statistically significant improvement in EFS vs control | 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 valuea | 0.0004 Curve labels: 12 mo: 67% (D+T+EV), 50% (Control); 24 mo: 60% (D+T+EV), 45% (Control) Durvalumab+T+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+T+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 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 Lan-DeMets alpha spending function with O'Brien-Fleming boundary – with the observed number of events, the boundary for declaring statistical significance was 0.0270 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; T, tremelimumab. [Survival data captured as a table — no point estimates read from the curve] Endpoint: EFS by BICR (ITT) D+T+EV | 233 | NR | 27.9–NC Control | 229 | 12.1 | 7.6–27.0 HR: 0.61 HR CI: 0.47–0.80 P: 0.0004 {'timepoint': '12 mo', 'arm': 'D+T+EV', 'value': '67%', 'ci': None} {'timepoint': '12 mo', 'arm': 'Control', 'value': '50%', 'ci': None} {'timepoint': '24 mo', 'arm': 'D+T+EV', 'value': '60%', 'ci': None} {'timepoint': '24 mo', 'arm': 'Control', 'value': '45%', 'ci': None} Note: Stratified log-rank P; significance boundary 0.0270 (2.7% overall 2-sided alpha). Median follow-up 23 months. Data cutoff 02 March 2026. [Slide 3] VOLGA: OS for Durvalumab+T+EV vs Control – ITT OS did not reach statistical significance at this interim analysis, and follow-up is ongoing | 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 valuea | 0.1230 Curve labels: 12 mo: 82% (D+T+EV), 81% (Control); 24 mo: 75% (D+T+EV), 69% (Control) Durvalumab+T+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+T+EV | 233 | 216 | 200 | 193 | 186 | 177 | 150 | 130 | 106 | 77 | 60 | 41 | 19 | 11 | 6 | 2 | 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 0.0138 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; T, tremelimumab. [Survival data captured as a table — no point estimates read from the curve] Endpoint: OS (ITT), interim D+T+EV | 233 | NR | NC–NC Control | 229 | NR | 37.7–NC HR: 0.77 HR CI: 0.54–1.07 P: 0.1230 {'timepoint': '12 mo', 'arm': 'D+T+EV', 'value': '82%', 'ci': None} {'timepoint': '12 mo', 'arm': 'Control', 'value': '81%', 'ci': None} {'timepoint': '24 mo', 'arm': 'D+T+EV', 'value': '75%', 'ci': None} {'timepoint': '24 mo', 'arm': 'Control', 'value': '69%', 'ci': None} Note: NOT statistically significant at this interim analysis (boundary 0.0138, 5% overall 2-sided alpha). Events are deaths. Median follow-up 26 months. Data cutoff 02 March 2026.

[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 value(a) | <0.0001 Probability of EFS 73% | 64% | Durvalumab+EV arm 50% | 45% | Control arm Median follow-up(b): 23 months Time from randomisation (months): 0 | 3 | 6 | 9 | 12 | 15 | 18 | 21 | 24 | 27 | 30 | 33 | 36 | 39 | 42 | 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. (a)The 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. (b)In 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 data captured as a table — no point estimates read from the curve] Endpoint: EFS by BICR (ITT), D+EV vs control Durvalumab+EV (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 (stratified log-rank) {'timepoint': '12 months', 'arm': 'Durvalumab+EV', 'value': '73%', 'ci': None} {'timepoint': '24 months', 'arm': 'Durvalumab+EV', 'value': '64%', 'ci': None} {'timepoint': '12 months', 'arm': 'Control', 'value': '50%', 'ci': None} {'timepoint': '24 months', 'arm': 'Control', 'value': '45%', 'ci': None} Note: Landmarks printed at dashed 12- and 24-month lines (slide rounds to whole %; NEJM figure shows 73.1/63.5 vs 50.3/45.1). Significance boundary 0.0218. Median follow-up 23 months. DCO 02 March 2026. [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.

[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 data captured as a table — no point estimates read from the curve] Endpoint: EFS (ITT), durvalumab-EV vs control 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 {'timepoint': '12 months', 'arm': 'Durvalumab–EV', 'value': '73.1', 'ci': None} {'timepoint': '24 months', 'arm': 'Durvalumab–EV', 'value': '63.5', 'ci': None} {'timepoint': '12 months', 'arm': 'Control', 'value': '50.3', 'ci': None} {'timepoint': '24 months', 'arm': 'Control', 'value': '45.1', 'ci': None} Note: NEJM figure; dashed lines at 12 and 24 months per caption. n from at-risk row at month 0. DCO March 2, 2026. Median labelled NE here vs NR on the ESMO slide. [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 data captured as a table — no point estimates read from the curve] Endpoint: EFS (ITT), durvalumab-tremelimumab-EV vs control 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 {'timepoint': '12 months', 'arm': 'Durvalumab–tremelimumab–EV', 'value': '66.8', 'ci': None} {'timepoint': '24 months', 'arm': 'Durvalumab–tremelimumab–EV', 'value': '60.1', 'ci': None} {'timepoint': '12 months', 'arm': 'Control', 'value': '50.3', 'ci': None} {'timepoint': '24 months', 'arm': 'Control', 'value': '45.1', 'ci': None} Note: NEJM figure; dashed lines at 12 and 24 months per caption. DCO March 2, 2026. [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 data captured as a table — no point estimates read from the curve] Endpoint: OS (ITT), durvalumab-EV vs control 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 {'timepoint': '12 months', 'arm': 'Durvalumab–EV', 'value': '87.0', 'ci': None} {'timepoint': '24 months', 'arm': 'Durvalumab–EV', 'value': '78.4', 'ci': None} {'timepoint': '12 months', 'arm': 'Control', 'value': '80.9', 'ci': None} {'timepoint': '24 months', 'arm': 'Control', 'value': '69.4', 'ci': None} Note: NEJM figure; dashed lines at 12 and 24 months per caption. Medians not estimable in either arm. DCO March 2, 2026.

[Slide 2] 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 [affiliations too small to read in this photo] [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.

[Slide 4] 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.

[Slide 4] 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: 87% / 81% (12 months); 78% / 69% (24 months) Durvalumab+EV arm Control arm Median follow-upᵇ: 26 months Probability of OS Time from randomisation (months): 0 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 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 [footnote cropped] [Survival data captured as a table — no point estimates read from the curve] Endpoint: Overall survival (ITT) Durvalumab+EV | 233 | NR | NC–NC Control | 229 | NR | 37.7–NC HR: 0.58 HR CI: 0.40–0.83 P: 0.0027 (stratified log-rank) {'timepoint': '12 months', 'arm': 'Durvalumab+EV', 'value': '87%', 'ci': None} {'timepoint': '12 months', 'arm': 'Control', 'value': '81%', 'ci': None} {'timepoint': '24 months', 'arm': 'Durvalumab+EV', 'value': '78%', 'ci': None} {'timepoint': '24 months', 'arm': 'Control', 'value': '69%', 'ci': None} Note: Median follow-up 26 months; significance boundary 0.0104 (5% 2-sided). Data cutoff cropped on this image (02 March 2026 per same slide elsewhere).

[Slide 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 data captured as a table — no point estimates read from the curve] Endpoint: EFS Durvalumab–EV Control {'timepoint': '12 months', 'arm': 'Durvalumab–EV', 'value': '73.1', 'ci': None} {'timepoint': '24 months', 'arm': 'Durvalumab–EV', 'value': '63.5', 'ci': None} {'timepoint': '12 months', 'arm': 'Control', 'value': '50.3', 'ci': None} {'timepoint': '24 months', 'arm': 'Control', 'value': '45.1', 'ci': None} Note: NEJM figure; no HR/N printed in crop.

[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 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.

[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 data captured as a table — no point estimates read from the curve] 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 {'timepoint': '4 years', 'arm': 'RT + ADT + Darolutamide', 'value': '95%', 'ci': None} {'timepoint': '4 years', 'arm': 'RT + ADT + Placebo', 'value': '88%', 'ci': None} Note: Interim analysis triggered at 87th MFS-CI event; P=0.0001 met rejection boundary 0.012. Medians not reached/not printed. Event components: metastases 9 (33%) vs 46 (73%); death w/o prior metastases 18 (67%) vs 17 (27%). At-risk rows: blue = darolutamide, red = placebo (legend color mapping per header colors). [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

[Slide 1] 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)

[Slide 4] 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] 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 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 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» [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

[Slide 1] 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 Silke Gillessen — Invited Discussant LBA4 [references line partially legible:] Attard G et al., Lancet 2022; [illegible] et al., ESMO 2025; Williams S et al., J Clin Onc 2026, in press; [illegible] et al., ESMO 2026 [Slide 2] 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

[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 data captured as a table — no point estimates read from the curve] 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 {'timepoint': '4 years', 'arm': 'RT + ADT + Darolutamide', 'value': '95%', 'ci': None} {'timepoint': '4 years', 'arm': 'RT + ADT + Placebo', 'value': '88%', 'ci': None} Note: Follow-up 46 months; interim analysis at 67% of required events. OS immature: 4/21 (Daro) and 7/24 (Plac) deaths from prostate cancer. [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] 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

[Slide 2] 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

[Slide 2] 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 data captured as a table — no point estimates read from the curve] 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 {'timepoint': '4 years', 'arm': 'RT + ADT + Darolutamide', 'value': '95%', 'ci': None} {'timepoint': '4 years', 'arm': 'RT + ADT + Placebo', 'value': '88%', 'ci': None} Note: Medians not printed. Event components: metastases 9 (33%) vs 46 (73%); death w/o prior metastases 18 (67%) vs 17 (27%). At-risk rows: blue = darolutamide, magenta = placebo (per header colors).

[Slide 2] 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 data captured as a table — no point estimates read from the curve] 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 {'timepoint': '4 years', 'arm': 'RT + ADT + Darolutamide', 'value': '95%', 'ci': None} {'timepoint': '4 years', 'arm': 'RT + ADT + Placebo', 'value': '88%', 'ci': None} Note: Medians not printed. Event breakdown: metastases 9 (33%) vs 46 (73%); death w/o prior metastases 18 (67%) vs 17 (27%) (percentages are of events). Trial name not printed on slide; tweet caption identifies DASL-HiCaP (CCTG PR.22 / ANZUP 1801), interim results.

[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) | 0.80 | 0.80 | No | 17/395 | 38/395 | 0.44 (0.25 to 0.78) | | Previous radical prostatectomy | Yes | 1/85 | 10/86 | 0.1 (0.01 to 0.79) | 0.08 | 0.50 | No | 26/468 | 53/467 | 0.48 (0.3 to 0.77) | | Age | 70+ | 17/298 | 37/316 | 0.5 (0.28 to 0.88) | 0.42 | 0.80 | <70 | 10/255 | 26/237 | 0.34 (0.16 to 0.7) | | Gleason score | 9 or 10 | 18/411 | 50/395 | 0.35 (0.2 to 0.6) | 0.17 | 0.80 | Others | 9/142 | 13/158 | 0.71 (0.3 to 1.66) | | Brachytherapy used | Yes | 3/51 | 5/39 | 0.54 (0.13 to 2.26) | 0.72 | 0.80 | No | 24/502 | 58/514 | 0.41 (0.25 to 0.66) | | PET imaging at baseline | Yes | 14/309 | 30/304 | 0.45 (0.24 to 0.85) | 0.77 | 0.80 | No | 13/244 | 33/249 | 0.39 (0.21 to 0.75) | | Overall | All Patients | 27/553 | 63/553 | 0.42 (0.27 to 0.66) | | Axis: 0.12 | 0.25 | 0.50 | 1.0 | 2.0 ; Darolutamide benefit <- -> Placebo benefit *PET PSMA done at baseline was at discretion of treating MD and had no contribution to eligibility Tamim Niazi @Niazi_Dr7 #DASL-HiCaP

[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 [ESMO Congress Madrid 2026 logo; presenter video inset] [Survival data captured as a table — no point estimates read from the curve] Endpoint: OS NIVO+GC | 325 | NR | NE-NE GC | 328 | NR | NE-NE HR: 0.88 HR CI: 0.64-1.21 {'timepoint': '12 mo', 'arm': 'NIVO+GC', 'value': '91.6%', 'ci': None} {'timepoint': '12 mo', 'arm': 'GC', 'value': '92.8%', 'ci': None} {'timepoint': '24 mo', 'arm': 'NIVO+GC', 'value': '84.8%', 'ci': None} {'timepoint': '24 mo', 'arm': 'GC', 'value': '85.2%', 'ci': None} {'timepoint': '36 mo', 'arm': 'NIVO+GC', 'value': '79.8%', 'ci': None} {'timepoint': '36 mo', 'arm': 'GC', 'value': '80.5%', 'ci': None} Note: OS descriptive only (EFS testing condition not met). A second dashed line near ~39 mo is labeled NIVO+GC 79.8% / GC 78.4%; timepoint not printed so not tabulated. Arm-to-label color assignment per slide color (teal=NIVO+GC, grey=GC). [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 data captured as a table — no point estimates read from the curve] Endpoint: OS (key secondary), ITT Pembro + cis + gem | 453 | NR | NR-NR Placebo + cis + gem | 454 | NR | NR-NR HR: 1.00 HR CI: 0.79-1.27 {'timepoint': '24 mo', 'arm': 'Pembro + cis + gem', 'value': '82.8%', 'ci': None} {'timepoint': '24 mo', 'arm': 'Placebo + cis + gem', 'value': '81.7%', 'ci': None} {'timepoint': '36 mo', 'arm': 'Pembro + cis + gem', 'value': '76.8%', 'ci': None} {'timepoint': '36 mo', 'arm': 'Placebo + cis + gem', 'value': '76.6%', 'ci': None} {'timepoint': '48 mo', 'arm': 'Pembro + cis + gem', 'value': '72.9%', 'ci': None} {'timepoint': '48 mo', 'arm': 'Placebo + cis + gem', 'value': '73.4%', 'ci': None} {'timepoint': '60 mo', 'arm': 'Pembro + cis + gem', 'value': '70.8%', 'ci': None} {'timepoint': '60 mo', 'arm': 'Placebo + cis + gem', 'value': '70.9%', 'ci': None} Note: HR not formally statistically tested. DCO January 26, 2026. Arm assignment of landmark labels by slide color (teal=pembro, maroon=placebo).

[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 data captured as a table — no point estimates read from the curve] Endpoint: EFS per BICR (co-primary) 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) {'timepoint': '12 mo', 'arm': 'NIVO+GC', 'value': '82.6%', 'ci': None} {'timepoint': '12 mo', 'arm': 'GC', 'value': '78.6%', 'ci': None} {'timepoint': '24 mo', 'arm': 'NIVO+GC', 'value': '72.3%', 'ci': None} {'timepoint': '24 mo', 'arm': 'GC', 'value': '71.6%', 'ci': None} {'timepoint': '36 mo', 'arm': 'NIVO+GC', 'value': '68.1%', 'ci': None} {'timepoint': '36 mo', 'arm': 'GC', 'value': '68.5%', 'ci': None} Note: A further dashed line ~39 mo labeled 67.3% (NIVO+GC) / 65.8% (GC); timepoint not printed so not tabulated. Label-to-arm by color (blue=NIVO+GC, grey=GC).

[Slide 1] CHAIRS: AXEL BEX, ANJA LORCH [ESMO Congress Madrid 2026 logo] 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 data captured as a table — no point estimates read from the curve] Endpoint: EFS and OS by pre-cystectomy ctDNA status (pooled across arms) Negative C1D1+C4D1 | 147 Cleared by C4D1 | 153 Positive at C4D1 | 46 Note: No medians, HRs, or p-values printed. n = first number-at-risk value. Number-at-risk digits from a photo of a screen; small and partly blurry, minor digit uncertainty. Assay: Signatera (Natera). [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 data captured as a table — no point estimates read from the curve] Endpoint: EFS per BICR (co-primary) 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 {'timepoint': '12 mo', 'arm': 'NIVO+GC', 'value': '82.6%', 'ci': None} {'timepoint': '12 mo', 'arm': 'GC', 'value': '78.6%', 'ci': None} {'timepoint': '24 mo', 'arm': 'NIVO+GC', 'value': '72.3%', 'ci': None} {'timepoint': '24 mo', 'arm': 'GC', 'value': '71.6%', 'ci': None} {'timepoint': '36 mo', 'arm': 'NIVO+GC', 'value': '68.1%', 'ci': None} {'timepoint': '36 mo', 'arm': 'GC', 'value': '68.5%', 'ci': None} {'timepoint': 'unlabelled 4th dashed line (~38-39 mo)', 'arm': 'NIVO+GC', 'value': '67.3%', 'ci': None} {'timepoint': 'unlabelled 4th dashed line (~38-39 mo)', 'arm': 'GC', 'value': '65.8%', 'ci': None} Note: Arm assignment of landmark labels by color (teal = NIVO+GC, grey = GC). Significance threshold p=0.0426; not met. Fourth landmark timepoint is not printed - verify before use.

[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 [ESMO Congress Madrid 2026 logo] [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 [ESMO Congress Madrid 2026 logo] [Survival data captured as a table — no point estimates read from the curve] Endpoint: EFS and OS by pretreatment (C1D1) ctDNA status, pooled across NIVO+GC and GC arms C1D1 ctDNA positive | 229 C1D1 ctDNA negative | 187 Note: No HR, medians, or landmark rates printed; n taken from month-0 number at risk. Prognostic pooled analysis, not a treatment comparison. [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 [ESMO Congress Madrid 2026 logo] [Survival data captured as a table — no point estimates read from the curve] 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 Note: No HR, medians, or landmark rates printed; n from month-0 number at risk. Prognostic pooled analysis.

[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 data captured as a table — no point estimates read from the curve] Endpoint: Overall survival (secondary; descriptive) NIVO+GC | 325 | NR | NE-NE GC | 328 | NR | NE-NE HR: 0.88 [blurry] HR CI: 0.64-1.21 [blurry] {'timepoint': '12 mo', 'arm': 'NIVO+GC', 'value': '91.6%', 'ci': None} {'timepoint': '12 mo', 'arm': 'GC', 'value': '92.8%', 'ci': None} {'timepoint': '24 mo', 'arm': 'NIVO+GC', 'value': '84.8%', 'ci': None} {'timepoint': '24 mo', 'arm': 'GC', 'value': '85.2%', 'ci': None} {'timepoint': '36 mo', 'arm': 'NIVO+GC', 'value': '79.8%', 'ci': None} {'timepoint': '36 mo', 'arm': 'GC', 'value': '80.5%', 'ci': None} {'timepoint': '~39 mo (dashed line, exact month not labeled)', 'arm': 'NIVO+GC', 'value': '79.8%', 'ci': None} {'timepoint': '~39 mo (dashed line, exact month not labeled)', 'arm': 'GC', 'value': '78.4%', 'ci': None} Note: Low-resolution source; table cells (events, HR/CI) blurry - VERIFY against primary source before use. Arm assignment of landmark labels by color (teal=NIVO+GC, grey=GC). Some landmark digits low confidence. [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 data captured as a table — no point estimates read from the curve] 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 Note: No HR/medians printed. Low-res source; GC rows faint (GC C1D1 negative row illegible; n possibly 103/101). Pooled-slide n (C1D1 positive 229 = 108+121; negative 187 = 84+103) suggests GC negative n=103, but inferred, not read.

[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 data captured as a table — no point estimates read from the curve] Endpoint: OS (secondary, descriptive) NIVO+GC | 325 | NR | NE-NE GC | 328 | NR | NE-NE HR: 0.88 HR CI: 95% CI 0.64-1.21 {'timepoint': '12 mo', 'arm': 'NIVO+GC', 'value': '91.6%', 'ci': None} {'timepoint': '12 mo', 'arm': 'GC', 'value': '92.8%', 'ci': None} {'timepoint': '24 mo', 'arm': 'NIVO+GC', 'value': '84.8%', 'ci': None} {'timepoint': '24 mo', 'arm': 'GC', 'value': '85.2%', 'ci': None} {'timepoint': '36 mo', 'arm': 'NIVO+GC', 'value': '79.8%', 'ci': None} {'timepoint': '36 mo', 'arm': 'GC', 'value': '80.5%', 'ci': None} {'timepoint': 'unlabeled dashed line between 36 and 42 mo', 'arm': 'NIVO+GC', 'value': '79.8%', 'ci': None} {'timepoint': 'unlabeled dashed line between 36 and 42 mo', 'arm': 'GC', 'value': '78.4%', 'ci': None} Note: OS analyzed descriptively (EFS testing condition not met); no p-value printed.

[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. [MADRID 2026 ESMO congress logo; presenter video inset] [Survival data captured as a table — no point estimates read from the curve] Endpoint: EFS per BICR (co-primary endpoint) 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) {'timepoint': 'unlabeled dashed line (~12 mo)', 'arm': 'NIVO+GC', 'value': '82.6%', 'ci': None} {'timepoint': 'unlabeled dashed line (~12 mo)', 'arm': 'GC', 'value': '78.6%', 'ci': None} {'timepoint': 'unlabeled dashed line (~24 mo)', 'arm': 'NIVO+GC', 'value': '72.3%', 'ci': None} {'timepoint': 'unlabeled dashed line (~24 mo)', 'arm': 'GC', 'value': '71.6%', 'ci': None} {'timepoint': 'unlabeled dashed line (~36 mo)', 'arm': 'NIVO+GC', 'value': '68.1%', 'ci': None} {'timepoint': 'unlabeled dashed line (~36 mo)', 'arm': 'GC', 'value': '68.5%', 'ci': None} {'timepoint': 'unlabeled dashed line (~38-39 mo)', 'arm': 'NIVO+GC', 'value': '67.3%', 'ci': None} {'timepoint': 'unlabeled dashed line (~38-39 mo)', 'arm': 'GC', 'value': '65.8%', 'ci': None} Note: Landmark percentages are printed on the curves but their timepoints are NOT labeled; timepoints are positional approximations from dashed drop-lines — confirm before citing. Results table text is small and angled. [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. [MADRID 2026 ESMO congress logo; presenter video inset] [Survival data captured as a table — no point estimates read from the curve] Endpoint: EFS and OS by pretreatment (C1D1) ctDNA status, pooled across arms C1D1 ctDNA positive | 229 C1D1 ctDNA negative | 187 Note: No medians, HRs, p-values or landmark rates printed; only number-at-risk rows. N from time-0 at-risk.

[Slide 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. [ESMO Congress Madrid 2026 logo]

[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 data captured as a table — no point estimates read from the curve] Endpoint: EFS per BICR (co-primary) 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 {'timepoint': '12 mo', 'arm': 'NIVO+GC', 'value': '82.6%', 'ci': None} {'timepoint': '12 mo', 'arm': 'GC', 'value': '78.6%', 'ci': None} {'timepoint': '24 mo', 'arm': 'NIVO+GC', 'value': '72.3%', 'ci': None} {'timepoint': '24 mo', 'arm': 'GC', 'value': '71.6%', 'ci': None} {'timepoint': '36 mo', 'arm': 'NIVO+GC', 'value': '68.1%', 'ci': None} {'timepoint': '36 mo', 'arm': 'GC', 'value': '68.5%', 'ci': None} {'timepoint': '4th dashed line (timepoint not printed)', 'arm': 'NIVO+GC', 'value': '67.3%', 'ci': None} {'timepoint': '4th dashed line (timepoint not printed)', 'arm': 'GC', 'value': '65.8%', 'ci': None} Note: Landmark timepoints 12/24/36 from dashed lines at printed axis ticks; 4th dashed line falls between 36 and 42 mo and is not labeled. [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 (%)

[Slide 1] 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.

[Slide 4] 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 data captured as a table — no point estimates read from the curve] Endpoint: EFS by investigator assessment (primary), ITT 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) {'timepoint': '24 mo (dashed line at axis tick)', 'arm': 'Pembro + cis + gem', 'value': '76.7%', 'ci': None} {'timepoint': '24 mo (dashed line at axis tick)', 'arm': 'Placebo + cis + gem', 'value': '70.0%', 'ci': None} {'timepoint': '36 mo (dashed line at axis tick)', 'arm': 'Pembro + cis + gem', 'value': '69.2%', 'ci': None} {'timepoint': '36 mo (dashed line at axis tick)', 'arm': 'Placebo + cis + gem', 'value': '66.3%', 'ci': None} {'timepoint': '48 mo (dashed line at axis tick)', 'arm': 'Pembro + cis + gem', 'value': '66.0%', 'ci': None} {'timepoint': '48 mo (dashed line at axis tick)', 'arm': 'Placebo + cis + gem', 'value': '62.3%', 'ci': None} {'timepoint': '60 mo (dashed line at axis tick)', 'arm': 'Pembro + cis + gem', 'value': '63.5%', 'ci': None} {'timepoint': '60 mo (dashed line at axis tick)', 'arm': 'Placebo + cis + gem', 'value': '61.9%', 'ci': None} Note: Landmark timepoints inferred from dashed-line positions (aligned with 24/36/48/60 mo ticks); not printed as text. Primary EFS not statistically significant.

[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. [ESMO congress logo, partially cropped] [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. [ESMO congress logo, partially cropped] [Survival data captured as a table — no point estimates read from the curve] Endpoint: EFS by investigator assessment (ITT) 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) {'timepoint': '24 mo', 'arm': 'Pembro + cis + gem', 'value': '76.7%', 'ci': None} {'timepoint': '24 mo', 'arm': 'Placebo + cis + gem', 'value': '70.0%', 'ci': None} {'timepoint': '36 mo', 'arm': 'Pembro + cis + gem', 'value': '69.2%', 'ci': None} {'timepoint': '36 mo', 'arm': 'Placebo + cis + gem', 'value': '66.3%', 'ci': None} {'timepoint': '48 mo', 'arm': 'Pembro + cis + gem', 'value': '66.0%', 'ci': None} {'timepoint': '48 mo', 'arm': 'Placebo + cis + gem', 'value': '62.3%', 'ci': None} {'timepoint': '60 mo', 'arm': 'Pembro + cis + gem', 'value': '63.5%', 'ci': None} {'timepoint': '60 mo', 'arm': 'Placebo + cis + gem', 'value': '61.9%', 'ci': None} Note: DCO January 26, 2026. Landmark timepoints inferred from dotted lines at 24/36/48/60 months on x-axis. [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. [ESMO congress logo, partially cropped]

[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. [ESMO Congress Madrid 2026 logo; podium sign: MADRID SPAIN 23-27 OCTOBER 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 data captured as a table — no point estimates read from the curve] Endpoint: OS (ITT) Pembro + cis + gem | 453 | NR | NR-NR Placebo + cis + gem | 454 | NR | NR-NR HR: 1.00 HR CI: 0.79-1.27 {'timepoint': '24 mo', 'arm': 'Pembro + cis + gem', 'value': '82.8%', 'ci': None} {'timepoint': '24 mo', 'arm': 'Placebo + cis + gem', 'value': '81.7%', 'ci': None} {'timepoint': '36 mo', 'arm': 'Pembro + cis + gem', 'value': '76.8%', 'ci': None} {'timepoint': '36 mo', 'arm': 'Placebo + cis + gem', 'value': '76.6%', 'ci': None} {'timepoint': '48 mo', 'arm': 'Pembro + cis + gem', 'value': '72.9%', 'ci': None} {'timepoint': '48 mo', 'arm': 'Placebo + cis + gem', 'value': '73.4%', 'ci': None} {'timepoint': '60 mo', 'arm': 'Pembro + cis + gem', 'value': '70.8%', 'ci': None} {'timepoint': '60 mo', 'arm': 'Placebo + cis + gem', 'value': '70.9%', 'ci': None} Note: Not formally statistically tested per footnote. Low-res image; pembro 36-mo value uncertain (76.8 vs 76.6). [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. [Presenter video inset: Jens Bedke at ESMO podium] 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 [ESMO Congress Madrid 2026 logo] [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. [Presenter video inset: Jens Bedke at ESMO podium] 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 [ESMO Congress Madrid 2026 logo] [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 [QR code] 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] [Presenter video inset: Jens Bedke at ESMO podium] 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 [ESMO Congress Madrid 2026 logo]

[Slide 1] [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 data captured as a table — no point estimates read from the curve] Endpoint: EFS 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) {'timepoint': '24 mo', 'arm': 'Pembrolizumab group', 'value': '76.7%', 'ci': None} {'timepoint': '24 mo', 'arm': 'Placebo group', 'value': '70.0%', 'ci': None} {'timepoint': '36 mo', 'arm': 'Pembrolizumab group', 'value': '69.2%', 'ci': None} {'timepoint': '36 mo', 'arm': 'Placebo group', 'value': '66.3%', 'ci': None} {'timepoint': '48 mo', 'arm': 'Pembrolizumab group', 'value': '66.0%', 'ci': None} {'timepoint': '48 mo', 'arm': 'Placebo group', 'value': '62.3%', 'ci': None} {'timepoint': '60 mo', 'arm': 'Pembrolizumab group', 'value': '63.5%', 'ci': None} {'timepoint': '60 mo', 'arm': 'Placebo group', 'value': '61.9%', 'ci': None} Note: Journal pre-proof figure, not a congress slide. Same numbers as ESMO26 slide 2113598042091794452_s3 (which labels P as 1-sided).

[Slide 3] 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. [ESMO Madrid stage podium decoration]

[Slide 2] 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 data captured as a table — no point estimates read from the curve] Endpoint: EFS by investigator assessment (ITT) 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* {'timepoint': '24 mo', 'arm': 'Pembro + cis + gem', 'value': '76.7%', 'ci': None} {'timepoint': '36 mo', 'arm': 'Pembro + cis + gem', 'value': '69.2%', 'ci': None} {'timepoint': '48 mo', 'arm': 'Pembro + cis + gem', 'value': '66.0%', 'ci': None} {'timepoint': '60 mo', 'arm': 'Pembro + cis + gem', 'value': '63.5%', 'ci': None} {'timepoint': '24 mo', 'arm': 'Placebo + cis + gem', 'value': '70.0%', 'ci': None} {'timepoint': '36 mo', 'arm': 'Placebo + cis + gem', 'value': '66.3%', 'ci': None} {'timepoint': '48 mo', 'arm': 'Placebo + cis + gem', 'value': '62.3%', 'ci': None} {'timepoint': '60 mo', 'arm': 'Placebo + cis + gem', 'value': '61.9%', 'ci': None} Note: Primary endpoint not met (1-sided P=0.1366). Asterisk footnote not visible in crop. Landmark timepoints taken from dotted vertical lines at 24/36/48/60 months where values are printed.

[Slide 1] [ESMO Congress Madrid 2026 logo] 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]

[Slide 1] 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 data captured as a table — no point estimates read from the curve] Endpoint: EFS 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 {'timepoint': '24 mo', 'arm': 'Pembrolizumab group', 'value': '76.7%', 'ci': None} {'timepoint': '24 mo', 'arm': 'Placebo group', 'value': '70.0%', 'ci': None} {'timepoint': '36 mo', 'arm': 'Pembrolizumab group', 'value': '69.2%', 'ci': None} {'timepoint': '36 mo', 'arm': 'Placebo group', 'value': '66.3%', 'ci': None} {'timepoint': '48 mo', 'arm': 'Pembrolizumab group', 'value': '66.0%', 'ci': None} {'timepoint': '48 mo', 'arm': 'Placebo group', 'value': '62.3%', 'ci': None} {'timepoint': '60 mo', 'arm': 'Pembrolizumab group', 'value': '63.5%', 'ci': None} {'timepoint': '60 mo', 'arm': 'Placebo group', 'value': '61.9%', 'ci': None} Note: Landmark values printed on curve. Data cut not printed on this figure.

[Slide 1] 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] [ESMO Congress Madrid 2026 logo] 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] [ESMO Congress Madrid 2026 logo] 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] [ESMO Congress Madrid 2026 logo] 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] [ESMO Congress Madrid 2026 logo] 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 [Madrid skyline artwork] [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 [ESMO Congress Madrid 2026 logo] [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 [ESMO Congress Madrid 2026 logo]

[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 [ESMO Congress Madrid 2026 logos] [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 [ESMO Congress Madrid 2026 logo] BARCELONA AUDITORIUM - HALL 9

[Slide 2] 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 [ESMO Congress Madrid 2026 logo] [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

[Slide 2] 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

[Slide 2] 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 [Inset speaker video; MADRID 2026 ESMO congress logo; panel seated on stage] [Slide 4] Most Common Treatment-Emergent vs Treatment-Related AEs [Butterfly bar chart: TEAEs, % (left) vs TRAEs, % (right); Any grade / Grade ≥ 3] AE terms (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; Neutropeniaᵃ; Amylase increase; Rhinorrhea; Stomatitis; Rash maculopapular; Spinal fracture [Individual bar values too small/angled to transcribe reliably - 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 neutropeniaᵃ 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 [Inset speaker video; MADRID 2026 ESMO congress logo]

[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.

[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 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 [ESMO Congress Madrid 2026 logo]

[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 data captured as a table — no point estimates read from the curve] Endpoint: Event-free survival (ad hoc analysis ITT population) 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 {'timepoint': '12 mo', 'arm': 'EV + pembro', 'value': '77.2%', 'ci': None} {'timepoint': '12 mo', 'arm': 'Pembro', 'value': '66.4%', 'ci': None} {'timepoint': '12 mo', 'arm': 'Control', 'value': '55.0%', 'ci': None} {'timepoint': '24 mo', 'arm': 'EV + pembro', 'value': '75.8%', 'ci': None} {'timepoint': '24 mo', 'arm': 'Pembro', 'value': '55.8%', 'ci': None} {'timepoint': '24 mo', 'arm': 'Control', 'value': '41.0%', 'ci': None} {'timepoint': '36 mo', 'arm': 'EV + pembro', 'value': '69.5%', 'ci': None} {'timepoint': '36 mo', 'arm': 'Pembro', 'value': '49.6%', 'ci': None} {'timepoint': '36 mo', 'arm': 'Control', 'value': '37.4%', 'ci': None} Note: HRs descriptive only (not prespecified hypothesis test); no p-values printed. DCO 6 June 2025. Ad hoc 3-arm concurrently randomized population (n=255). [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 data captured as a table — no point estimates read from the curve] Endpoint: Overall survival (ad hoc analysis ITT population) 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 {'timepoint': '12 mo', 'arm': 'EV + pembro', 'value': '83.3%', 'ci': None} {'timepoint': '12 mo', 'arm': 'Pembro', 'value': '78.2%', 'ci': None} {'timepoint': '12 mo', 'arm': 'Control', 'value': '74.6%', 'ci': None} {'timepoint': '24 mo', 'arm': 'EV + pembro', 'value': '78.4%', 'ci': None} {'timepoint': '24 mo', 'arm': 'Pembro', 'value': '66.9%', 'ci': None} {'timepoint': '24 mo', 'arm': 'Control', 'value': '62.6%', 'ci': None} {'timepoint': '36 mo', 'arm': 'EV + pembro', 'value': '74.5%', 'ci': None} {'timepoint': '36 mo', 'arm': 'Pembro', 'value': '61.4%', 'ci': None} {'timepoint': '36 mo', 'arm': 'Control', 'value': '55.7%', 'ci': None} Note: HRs descriptive only; no p-values printed. DCO 6 June 2025. [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. [ESMO Congress Madrid 2026 logo; presenter video inset; podium: MADRID SPAIN 23-27 OCTOBER 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 [ESMO Congress Madrid 2026 logo; presenter video inset] [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 [ESMO Congress Madrid 2026 logo; presenter video inset]

[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 data captured as a table — no point estimates read from the curve] Endpoint: EFS (ad hoc ITT, 3 concurrently randomized arms) 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 {'timepoint': '12 mo', 'arm': 'EV + pembro', 'value': '77.2%', 'ci': None} {'timepoint': '12 mo', 'arm': 'Pembro', 'value': '66.4%', 'ci': None} {'timepoint': '12 mo', 'arm': 'Control', 'value': '55.0%', 'ci': None} {'timepoint': '24 mo', 'arm': 'EV + pembro', 'value': '75.8%', 'ci': None} {'timepoint': '24 mo', 'arm': 'Pembro', 'value': '55.8%', 'ci': None} {'timepoint': '24 mo', 'arm': 'Control', 'value': '41.0%', 'ci': None} {'timepoint': '36 mo', 'arm': 'EV + pembro', 'value': '69.5%', 'ci': None} {'timepoint': '36 mo', 'arm': 'Pembro', 'value': '49.6%', 'ci': None} {'timepoint': '36 mo', 'arm': 'Control', 'value': '37.4%', 'ci': None} Note: Descriptive comparison only (not prespecified hypothesis test). Low-res capture; digits read from upscaled image. [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 data captured as a table — no point estimates read from the curve] Endpoint: OS (ad hoc ITT, 3 concurrently randomized arms) 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 {'timepoint': '12 mo', 'arm': 'EV + pembro', 'value': '83.3%', 'ci': None} {'timepoint': '12 mo', 'arm': 'Pembro', 'value': '78.2%', 'ci': None} {'timepoint': '12 mo', 'arm': 'Control', 'value': '74.6%', 'ci': None} {'timepoint': '24 mo', 'arm': 'EV + pembro', 'value': '78.4%', 'ci': None} {'timepoint': '24 mo', 'arm': 'Pembro', 'value': '66.9%', 'ci': None} {'timepoint': '24 mo', 'arm': 'Control', 'value': '62.6%', 'ci': None} {'timepoint': '36 mo', 'arm': 'EV + pembro', 'value': '74.5%', 'ci': None} {'timepoint': '36 mo', 'arm': 'Pembro', 'value': '61.4%', 'ci': None} {'timepoint': '36 mo', 'arm': 'Control', 'value': '55.7%', 'ci': None} Note: Descriptive only. Low-res capture; digits read from upscaled image.

[Slide 1] 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).

[Slide 2] 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.

[Slide 1] [Photo of ESMO 2026 session hall: speaker at podium, discussant video inset on left of screen; podium sign 'ESMO congress MADRID SPAIN 23-27 OCTOBER 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. [ESMO Congress Madrid 2026 logo]

[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 data captured as a table — no point estimates read from the curve] Endpoint: Overall survival (ad hoc, ITT) 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 {'timepoint': '12 mo', 'arm': 'EV + pembro', 'value': '83.3%', 'ci': None} {'timepoint': '12 mo', 'arm': 'Pembro', 'value': '78.2%', 'ci': None} {'timepoint': '12 mo', 'arm': 'Control', 'value': '74.6%', 'ci': None} {'timepoint': '24 mo', 'arm': 'EV + pembro', 'value': '78.4%', 'ci': None} {'timepoint': '24 mo', 'arm': 'Pembro', 'value': '66.9%', 'ci': None} {'timepoint': '24 mo', 'arm': 'Control', 'value': '62.6%', 'ci': None} {'timepoint': '36 mo', 'arm': 'EV + pembro', 'value': '74.5%', 'ci': None} {'timepoint': '36 mo', 'arm': 'Pembro', 'value': '61.4%', 'ci': None} {'timepoint': '36 mo', 'arm': 'Control', 'value': '55.7%', 'ci': None} Note: Ad hoc, descriptive.

[Slide 2] 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 [ESMO Congress Madrid 2026 logo] [Survival data captured as a table — no point estimates read from the curve] Endpoint: OS (ADT+docetaxel population, n=710) 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 {'timepoint': '7 years', 'arm': 'SOC+Abi', 'value': '45.7%', 'ci': '40.3-51.2'} {'timepoint': '7 years', 'arm': 'SOC', 'value': '30.7%', 'ci': '25.8-36.1'} {'timepoint': '8 years', 'arm': 'SOC+Abi', 'value': '40.3%', 'ci': '34.3-46.[illegible]'} {'timepoint': '8 years', 'arm': 'SOC', 'value': '27.1%', 'ci': '22.0-32.9'} Note: Adjusted Cox HR (prostate RT, ECOG, castration type, disease burden). Curve annotations show 1.8 y median gain and 46% vs 31% (approx. 7-yr arrows).

[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 [ESMO Congress Madrid 2026 logo] [Slide 2] 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 [KM curve annotation: 1.5 y gap between curves] Restricted Mean Survival Time difference at 8 years = 0.48y (0.17-0.79) P value RMST test of equality = 0.0021 Time from randomization (in years) 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) *Cox model adjusted for prostate RT, ECOG, type of castration, docetaxel administration and disease burden Karim Fizazi [ESMO Congress Madrid 2026 logo] [Survival data captured as a table — no point estimates read from the curve] Endpoint: Overall survival (overall population, n=1172) SOC (+/- RXT)+Abi | 583 | 6.1 y | 5.3-6.6 (95.1% CI) SOC (+/- RXT) | 589 | 4.6 y | 4.2-5.2 (95.1% CI) HR: 0.75 (adjusted) HR CI: 0.65-0.87 (95.1% CI) P: 0.0001 {'timepoint': '7 years', 'arm': 'SOC (+/- RXT)+Abi', 'value': '44.9%', 'ci': '40.8-49.2'} {'timepoint': '7 years', 'arm': 'SOC (+/- RXT)', 'value': '34.2%', 'ci': '30.3-38.3'} {'timepoint': '8 years', 'arm': 'SOC (+/- RXT)+Abi', 'value': '40.3%', 'ci': '36.0-44.8'} {'timepoint': '8 years', 'arm': 'SOC (+/- RXT)', 'value': '29.5%', 'ci': '25.6-33.8'} Note: RMST difference at 8 years 0.48y (0.17-0.79), p=0.0021. HR Cox-adjusted for prostate RT, ECOG, castration type, docetaxel, disease burden. Curve annotation 1.5 y. [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 [ESMO Congress Madrid 2026 logo] [Survival data captured as a table — no point estimates read from the curve] Endpoint: rPFS by metastatic volume (ADT+Docetaxel population, n=710) 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) {'timepoint': '7 years', 'arm': 'High volume SOC+Abi', 'value': '33.4%', 'ci': '23.7-44.7'} {'timepoint': '7 years', 'arm': 'High volume SOC', 'value': '10.9%', 'ci': '5.7-19.8'} {'timepoint': '7 years', 'arm': 'Low volume SOC+Abi', 'value': '43.4%', 'ci': '29.6-58.2'} {'timepoint': '7 years', 'arm': 'Low volume SOC', 'value': '20.9%', 'ci': '10.9-36.4'} Note: Two panels on one slide (high vs low volume). CIs are 99.9%. Number-at-risk rows for high volume are small/partly blurry in photo.

[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 data captured as a table — no point estimates read from the curve] Endpoint: Overall survival, ADT + docetaxel population (triplet vs doublet) 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 {'timepoint': '7-year', 'arm': 'SOC+Abi', 'value': '45.7%', 'ci': '40.3-51.2'} {'timepoint': '7-year', 'arm': 'SOC', 'value': '30.7%', 'ci': '25.8-36.1'} {'timepoint': '8-year', 'arm': 'SOC+Abi', 'value': '40.3%', 'ci': '34.3-46. (as printed)'} {'timepoint': '8-year', 'arm': 'SOC', 'value': '27.1%', 'ci': '22.0-32.9'} Note: Median gain annotated 1.8 y; curve labels 46% (triplet) vs 31% (doublet). [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 data captured as a table — no point estimates read from the curve] Endpoint: Overall survival by disease volume (triplet vs doublet) 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 {'timepoint': '7-year', 'arm': 'High volume SOC+Abi', 'value': '40.9%', 'ci': '34.4-47.7'} {'timepoint': '7-year', 'arm': 'High volume SOC', 'value': '22.4%', 'ci': '17.2-28.7'} {'timepoint': '7-year', 'arm': 'Low volume SOC+Abi', 'value': '54.2%', 'ci': '45.0-63.1'} {'timepoint': '7-year', 'arm': 'Low volume SOC', 'value': '46.0%', 'ci': '36.9-55.4'} Note: Two subgroups on one slide; HR/p given per subgroup. Median gains annotated 1.1 y (high) and 1.6 y (low).

[Slide 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 [ESMO Congress Madrid 2026 logo]

[Slide 1] 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 data captured as a table — no point estimates read from the curve] Endpoint: Overall survival by metastatic volume (ADT+docetaxel population, n=710) 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 {'timepoint': '7 years', 'arm': 'High volume SOC+Abi', 'value': '40.9%', 'ci': '34.4-47.7'} {'timepoint': '7 years', 'arm': 'High volume SOC', 'value': '22.4%', 'ci': '17.2-28.7'} {'timepoint': '7 years', 'arm': 'Low volume SOC+Abi', 'value': '54.2%', 'ci': '45.0-63.1'} {'timepoint': '7 years', 'arm': 'Low volume SOC', 'value': '46.0%', 'ci': '36.9-55.4'} Note: 95.1% CIs as printed. No HRs printed on this slide. Blurry capture; n values should be verified.

[Slide 1] 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 data captured as a table — no point estimates read from the curve] 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 {'timepoint': '~7 y (arrow position)', 'arm': 'High volume Triplet', 'value': '3[?]%', 'ci': None} {'timepoint': '~7 y (arrow position)', 'arm': 'High volume Doublet', 'value': '11%', 'ci': None} {'timepoint': '~7 y (arrow position)', 'arm': 'Low volume Triplet', 'value': '43%', 'ci': None} {'timepoint': '~7 y (arrow position)', 'arm': 'Low volume Doublet', 'value': '21%', 'ci': None} Note: No medians/HR printed. High-volume triplet label digit uncertain. n=224 from abiraterone at-risk row start (high volume).

[Slide 3] 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 [ESMO Congress logo] [Survival data captured as a table — no point estimates read from the curve] Endpoint: OS by metastatic volume, ADT+docetaxel population 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 {'timepoint': '7 years', 'arm': 'High volume SOC+Abi', 'value': '40.9%', 'ci': '34.4-47.7'} {'timepoint': '7 years', 'arm': 'High volume SOC', 'value': '22.4%', 'ci': '17.2-28.7'} {'timepoint': '7 years', 'arm': 'Low volume SOC+Abi', 'value': '54.2%', 'ci': '45.0-63.1'} {'timepoint': '7 years', 'arm': 'Low volume SOC', 'value': '46.0%', 'ci': '36.9-55.4'} Note: Cox models adjusted for prostate RT, ECOG and type of castration.

[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 data captured as a table — no point estimates read from the curve] 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 {'timepoint': '36 months', 'arm': 'APA + ADT', 'value': '63.0%', 'ci': '52.2-71.9'} {'timepoint': '36 months', 'arm': 'RP + PLND alone', 'value': '41.4%', 'ci': '35.6-47.1'} Note: Landmark % printed at the 36-month dashed line. T recovery to >=200 ng/dL: median 20.7 months (~9 months from EOT). Numbers at risk 0/6/12/18/24/30/36/42: APA+ADT 102, 99, 96, 91, 82, 71, 38, 10; RP+PLND 306, 195, 169, 146, 131, 121, 74, 15. [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 data captured as a table — no point estimates read from the curve] 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 Note: No medians or landmark % printed. Metastasis detection method n (%): PSMA PET 9 (8.8) vs 50 (16.3); Conventional imaging 1 (1.0) vs 6 (2.0); Histopathology 1 (1.0) vs 0. Numbers at risk 0/6/12/18/24/30/36/42: APA+ADT 102, 99, 97, 94, 90, 85, 52, 13; RP+PLND 306, 262, 255, 241, 231, 220, 130, 25. [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 data captured as a table — no point estimates read from the curve] 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 {'timepoint': '36 months', 'arm': 'APA + ADT', 'value': '64.6%', 'ci': '54.1-73.3'} {'timepoint': '36 months', 'arm': 'RP + PLND alone', 'value': '32.0%', 'ci': '26.7-37.5'} Note: Landmark timepoint inferred from dashed vertical line at 36 months; footnotes a/b not shown on crop. N from numbers-at-risk row.

[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 data captured as a table — no point estimates read from the curve] 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 Note: N taken from number-at-risk at month 0 (102 vs 306) as printed. No medians printed. Investigator-assessed MFS HR 0.38 (95% CI 0.22-0.66), nominal p=0.0004. Detection-method table: PSMA PET 9 (8.8) vs 50 (16.3); conventional 1 (1.0) vs 6 (2.0); histopathology 1 (1.0) vs 0.

[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 data captured as a table — no point estimates read from the curve] 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 {'timepoint': '36 months', 'arm': 'APA + ADT', 'value': '63.0%', 'ci': '52.2-71.9'} {'timepoint': '36 months', 'arm': 'RP + PLND alone', 'value': '41.4%', 'ci': '35.6-47.1'} Note: Stratified HR; stratified log-rank p. Landmark timepoint inferred from dashed line at 36 mo and corroborated by EFS-over-time table slide (63 vs 41 at 36 months). T recovery median 20.7 months. [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 data captured as a table — no point estimates read from the curve] 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 {'timepoint': '36 months', 'arm': 'APA + ADT', 'value': '64.6%', 'ci': '54.1-73.3'} {'timepoint': '36 months', 'arm': 'RP + PLND alone', 'value': '32.0%', 'ci': '26.7-37.5'} Note: Landmark timepoint inferred from dashed line at 36 months on x-axis. Numbers at risk at month 0 printed as 102 (APA+ADT) and 306 (RP+PLND alone) as seen in photo; arm n not printed separately - verify against source. Exploratory endpoint.

[Slide 4] 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

[Slide 1] 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 [affiliations list - small print, largely illegible] Presented by A Briganti at ESMO; October 23-27, 2026; Madrid, Spain [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

[Slide 1] 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] 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 [ESMO Congress Madrid 2026 logo] [Survival data captured as a table — no point estimates read from the curve] Endpoint: OS (updated DCO) Capivasertib + abiraterone | 507 Placebo + abiraterone | 505 HR: 0.82 HR CI: 0.68, 0.99 Note: DCO 9 Mar 2026. Subgroups: high w/ visceral mets HR 0.74 (0.49-1.11); high w/o visceral mets 0.89 (0.70-1.14); low 0.70 (0.42-1.15). OS HR at primary rPFS analysis 0.90 (0.71-1.15). No medians or p-value printed.

[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 [ESMO Congress Madrid 2026 logo] [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 [ESMO Congress Madrid 2026 logo] [Survival data captured as a table — no point estimates read from the curve] Endpoint: Overall survival (CAPItello-281 updated DCO, 9 Mar 2026) 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) Note: No landmark rates printed. Footnote: OS HR at primary rPFS analysis was 0.90 (0.71, 1.15). [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 [ESMO Congress Madrid 2026 logo] [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 [ESMO Congress Madrid 2026 logo]

[Slide 1] 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 data captured as a table — no point estimates read from the curve] Endpoint: Overall survival (updated descriptive analysis, 42% maturity) 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 Note: Updated DESCRIPTIVE analysis (not formal hypothesis test); data cut-off March 9, 2026; 42% OS maturity.

[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 data captured as a table — no point estimates read from the curve] Endpoint: Investigator-assessed rPFS (primary DCO 7 Oct 2024) 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 Note: Primary rPFS analysis; 7.5-month median difference annotated on slide. [Slide 3] 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) Number of patients at risk (0,3,...,66 months) 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 data captured as a table — no point estimates read from the curve] Endpoint: OS (updated DCO) 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 (descriptive) Note: Descriptive OS analysis: alpha controlled, did not meet statistical significance. Updated DCO date not printed. [Slide 4] 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. Annotation: 26.0 months — 9.5 months — 35.5 months Legend: Capivasertib + abiraterone; Placebo + abiraterone Y axis: Probability of rPFS; X axis: Time from randomisation (months) Number of patients at risk (0,3,...,63 months) 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 data captured as a table — no point estimates read from the curve] Endpoint: rPFS (updated DCO, descriptive) Capivasertib + abiraterone | 507 | 35.5 months | 29.7, 40.5 Placebo + abiraterone | 505 | 26.0 months | 22.4, 31.3 HR: 0.83 HR CI: 0.70, 0.99 P: 0.033 (nominal) Note: Descriptive rPFS analysis, not formally tested; different cut from primary rPFS (33.2 vs 25.7, HR 0.81) on s2.

[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 [ESMO Congress Madrid 2026 logo; presenter video inset 'MADRID SPAIN 23-27 OCTOBER 2026'] [Survival data captured as a table — no point estimates read from the curve] Endpoint: Overall survival (updated DCO) 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) Note: DCO 9 Mar 2026. OS HR at primary rPFS analysis was 0.90 (0.71, 1.15). [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 [ESMO Congress Madrid 2026 logo; presenter video inset 'MADRID SPAIN 23-27 OCTOBER 2026'] [Survival data captured as a table — no point estimates read from the curve] 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 Note: Post hoc exploratory. DCO 9 Mar 2026. No p-values printed for subgroups. [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 [ESMO Congress Madrid 2026 logo; presenter video inset 'MADRID SPAIN 23-27 OCTOBER 2026'] [Survival data captured as a table — no point estimates read from the curve] Endpoint: Symptomatic skeletal event-free survival (SSE-FS, updated DCO) 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) Note: DCO 9 Mar 2026. Median difference 7.9 months printed.

[Slide 1] 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 data captured as a table — no point estimates read from the curve] Endpoint: PFS by TROP2 H-score (left panel) TROP2 H-score <median | 18 | 11.2 mo | 3.7–NR TROP2 H-score ≥median | 19 | 11.9 mo | 5.6–21.1 Note: Right panel PD-L1: CPS <10 n=32, median 15.6 mo (7.2–NR); CPS ≥10 n=8, median 5.7 mo (0.5–NR). Overall cohort median PFS 11.2 (5.8–NR) mo; median OS NR (17.2–NR). No HRs printed. Data cutoff Jan 26, 2026; median follow-up 22.8 (19.8–30.0) mo.

[Slide 1] [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 [ESMO Congress Madrid 2026 logo; podium MADRID SPAIN 23-27 OCTOBER 2026] 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] 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.

[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 [ESMO Congress logo]

[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% [ESMO Congress Madrid 2026 logo] [Survival data captured as a table — no point estimates read from the curve] Endpoint: EFS and OS by pre-treatment ctDNA status (pooled); NIAGARA EFS by baseline ctDNA ctDNA positive ctDNA negative Note: Low-res image; no values legible on curves. Baseline ctDNA+ rates 51% (gem-cis) vs 59% (gem-cis nivolumab) in CA017-078, low-confidence read. [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] [ESMO Congress Madrid 2026 logo] [Survival data captured as a table — no point estimates read from the curve] Endpoint: EFS by ctDNA clearance status (C1D1 to C4D1) C1D1 to C4D1 cleared C1D1 to C4D1 not cleared Note: No printed medians, HR or landmark values legible; image low-res. [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 [ESMO Congress Madrid 2026 logo] [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) [ESMO Congress Madrid 2026 logo]

[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 data captured as a table — no point estimates read from the curve] Endpoint: EFS, KEYNOTE-905/EV-303 (EV + pembro vs control), as reproduced on discussant slide 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 {'timepoint': '12 months', 'arm': 'EV + pembro', 'value': '77.8%', 'ci': None} {'timepoint': '12 months', 'arm': 'Control', 'value': '55.1%', 'ci': None} {'timepoint': '24 months', 'arm': 'EV + pembro', 'value': '74.7%', 'ci': None} {'timepoint': '24 months', 'arm': 'Control', 'value': '39.4%', 'ci': None} Note: Discussant reproduction of KEYNOTE-905/EV-303 data alongside VOLGA D+EV EFS (HR 0.51, 0.38–0.67). pCR bars: VOLGA inv 45%, VOLGA central 40%, KN-905 57%, KN-B15 56%. [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] [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 data captured as a table — no point estimates read from the curve] Endpoint: OS, KEYNOTE-905/EV-303 (EV + pembro vs control), as reproduced on discussant slide 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 {'timepoint': '12 months', 'arm': 'EV + pembro', 'value': '86.3%', 'ci': None} {'timepoint': '12 months', 'arm': 'Control', 'value': '75.7%', 'ci': None} {'timepoint': '24 months', 'arm': 'EV + pembro', 'value': '79.7%', 'ci': None} {'timepoint': '24 months', 'arm': 'Control', 'value': '63.1%', 'ci': None} Note: Forest plot OS HRs: KN905/EV303 0.50 (0.33–0.74); VOLGA EV+Durva 0.58 (0.40–0.83); VOLGA EV+Durva+Treme 0.77 (0.54–1.07). VOLGA D+EV inset: 87%/78% vs 81%/69% at 12/24 mo.

[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 1] [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.

[Slide 2] 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.

[Slide 1] 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

[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 data captured as a table — no point estimates read from the curve] 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 {'timepoint': '12 mo', 'arm': 'EFS EV+pembro', 'value': '77.2%', 'ci': None} {'timepoint': '24 mo', 'arm': 'EFS EV+pembro', 'value': '75.8%', 'ci': None} {'timepoint': '36 mo', 'arm': 'EFS EV+pembro', 'value': '69.5%', 'ci': None} {'timepoint': '12 mo', 'arm': 'EFS Pembro', 'value': '66.4%', 'ci': None} {'timepoint': '24 mo', 'arm': 'EFS Pembro', 'value': '55.8%', 'ci': None} {'timepoint': '36 mo', 'arm': 'EFS Pembro', 'value': '49.6%', 'ci': None} {'timepoint': '12 mo', 'arm': 'EFS RC+PLND', 'value': '55.0%', 'ci': None} {'timepoint': '24 mo', 'arm': 'EFS RC+PLND', 'value': '41.0%', 'ci': None} {'timepoint': '36 mo', 'arm': 'EFS RC+PLND', 'value': '37.4%', 'ci': None} {'timepoint': '12 mo', 'arm': 'OS EV+pembro', 'value': '83.3%', 'ci': None} {'timepoint': '24 mo', 'arm': 'OS EV+pembro', 'value': '78.4%', 'ci': None} {'timepoint': '36 mo', 'arm': 'OS EV+pembro', 'value': '74.5%', 'ci': None} {'timepoint': '12 mo', 'arm': 'OS Pembro', 'value': '78.2%', 'ci': None} {'timepoint': '24 mo', 'arm': 'OS Pembro', 'value': '66.9%', 'ci': None} {'timepoint': '36 mo', 'arm': 'OS Pembro', 'value': '61.4%', 'ci': None} {'timepoint': '12 mo', 'arm': 'OS RC+PLND', 'value': '74.6%', 'ci': None} {'timepoint': '24 mo', 'arm': 'OS RC+PLND', 'value': '62.6%', 'ci': None} {'timepoint': '36 mo', 'arm': 'OS RC+PLND', 'value': '55.7%', 'ci': None} Note: Landmark values are small, low-resolution print; read via zoomed crop - verify against source before publishing (esp. pembro 36-mo EFS 49.6% vs 49.0% shown on next slide). EV-103 curves are digitised/approximate per slide footnote. [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] 1047P [Gustave Roussy Cancer Campus Grand Paris logo] 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 [Gustave Roussy logo] [ESMO logo] 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] [PEACE logo] [Gustave Roussy Paris Cancer Campus logo] [AFU Association Francaise d'Urologie logo] 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 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 | Number of patients 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: TMB-high | 84 | 77 | 0.44 (0.25–0.76) TMB: TMB-low | 125 | 121 | 0.82 (0.53–1.26) Gene mutations: TP53 mutated | 107 | 97 | 0.51 (0.32–0.81) Gene mutations: TP53 WT | 102 | 101 | 0.81 (0.49–1.31) Gene mutations: FGFR3 mutated | 20 | 24 | 0.24 (0.07–0.87) Gene mutations: FGFR3 WT | 189 | 174 | 0.70 (0.49–0.99) Top altered pathways: TP53CC mutated (pathway) | 129 | 125 | 0.58 (0.38–0.88) Top altered pathways: RTK mutated (pathway) | 86 | 82 | 0.49 (0.28–0.85) Top altered pathways: DDR mutated (pathway) | 50 | 48 | 0.72 (0.36–1.43) Top altered pathways: Histone-mod mutated (pathway) | 83 | 85 | 0.63 (0.37–1.08) Top altered pathways: Chromatin remodelling (pathway) | 54 | 44 | 0.48 (0.23–0.98) EFS HR (log scale) 0.125 0.25 0.5 1 2 4 8 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.

[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. [Podium: MADRID SPAIN 23-27 OCTOBER 2026] [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 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 data captured as a table — no point estimates read from the curve] Endpoint: rPFS per PCWG3-modified RECIST v1.1 by BICR, total population (subgroups in table) 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) {'timepoint': '6 mo', 'arm': 'Pembro + chemo (total)', 'value': '49%', 'ci': None} {'timepoint': '6 mo', 'arm': 'Chemo (total)', 'value': '26%', 'ci': None} Note: PD-L1 subgroups are post hoc. All HR CIs cross 1. DCO November 19, 2025. [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 data captured as a table — no point estimates read from the curve] Endpoint: OS, total population (subgroups in table) 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) {'timepoint': '6 mo', 'arm': 'Pembro + chemo (total)', 'value': '85%', 'ci': None} {'timepoint': '6 mo', 'arm': 'Chemo (total)', 'value': '59%', 'ci': None} Note: OS is a secondary endpoint; PD-L1 subgroups post hoc; small randomized cohort (N=70). No p-value printed. DCO November 19, 2025.

[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] 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] [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 [...] [ESMO OPEN - Science for optimal cancer care logo; QR code] 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.

[Slide 1] 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] [Cancer Research UK City of London Centre logo] [ECMC logo] 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 [Queen Mary University of London / Barts Cancer Institute logo] [ECMC logo] 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

[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 [Speaker video panel: Toni Choueiri, Perspectives from the medical oncologist] ALICANTE AUDITORIUM - HALL 6 [ESMO Congress Madrid 2026 logo] [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. [Speaker video panel: Toni Choueiri, Perspectives from the medical oncologist] ALICANTE AUDITORIUM - HALL 6 [ESMO Congress Madrid 2026 logo] [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 [Speaker video panel: Toni Choueiri, Perspectives from the medical oncologist] ALICANTE AUDITORIUM - HALL 6 [ESMO Congress Madrid 2026 logo] [Slide 4] 15:15 - 16:00 What oncologists should know about mRNA vaccines CHAIRS: YELENA JANJIGIAN, JAMES LARKIN [ACIR logo] 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 [Speaker video panel: Toni Choueiri, Perspectives from the medical oncologist] ALICANTE AUDITORIUM - HALL 6 [ESMO Congress Madrid 2026 logo]

[Slide 1] [Dana-Farber Cancer Institute logo] [Harvard Medical School logo] 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 [ESMO Congress Madrid 2026 logo] [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 [ESMO Congress Madrid 2026 logo]

[Slide 1] 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 [ESMO Congress Madrid 2026 logo; speaker at podium below screen] [Survival data captured as a table — no point estimates read from the curve] Endpoint: EFS by ctDNA clearance status (C1D1 to C4D1) C1D1 to C4D1 cleared | 153 C1D1 to C4D1 not cleared | 46 Note: No HR, medians or landmark values printed; n from number-at-risk at time 0.

[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 [City of Hope logo] 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 [City of Hope logo] 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 [QR code]

[Slide 1] [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 [UT Southwestern Medical Center logo; The University of Texas MD Anderson Cancer Center logo] 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] 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] 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). [ESMO Congress logo] 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). [ESMO Congress logo] 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). [ESMO Congress logo] Niven Mehra — Emerging immune-based therapies BILBAO AUDITORIUM - [HALL 3, partly obscured]

[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] 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 data captured as a table — no point estimates read from the curve] 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 {'timepoint': '36 mo', 'arm': 'APA + ADT', 'value': '64.6%', 'ci': '54.1-73.3'} {'timepoint': '36 mo', 'arm': 'RP + PLND alone', 'value': '32.0%', 'ci': '26.7-37.5'} Note: Landmark labels at 36-mo dashed line (timepoint inferred from dashed line position). RP+PLND at-risk row cropped.

[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 2] 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 data captured as a table — no point estimates read from the curve] 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 {'timepoint': 'at-risk column 2 (year not explicitly printed)', 'arm': 'NCCN high risk, GC > 0.85', 'value': '9%', 'ci': None} {'timepoint': 'at-risk column 3 (year not explicitly printed)', 'arm': 'NCCN high risk, GC > 0.85', 'value': '14%', 'ci': None} {'timepoint': 'at-risk column 4 (year not explicitly printed)', 'arm': 'NCCN high risk, GC > 0.85', 'value': '16%', 'ci': None} {'timepoint': 'at-risk column 2 (year not explicitly printed)', 'arm': 'NCCN high risk, GC 0.61-0.85', 'value': '4%', 'ci': None} {'timepoint': 'at-risk column 3 (year not explicitly printed)', 'arm': 'NCCN high risk, GC 0.61-0.85', 'value': '5%', 'ci': None} {'timepoint': 'at-risk column 4 (year not explicitly printed)', 'arm': 'NCCN high risk, GC 0.61-0.85', 'value': '5%', 'ci': None} {'timepoint': 'at-risk column 2 (year not explicitly printed)', 'arm': 'NCCN high risk, GC ≤ 0.60', 'value': '2%', 'ci': None} {'timepoint': 'at-risk column 3 (year not explicitly printed)', 'arm': 'NCCN high risk, GC ≤ 0.60', 'value': '3%', 'ci': None} {'timepoint': 'at-risk column 4 (year not explicitly printed)', 'arm': 'NCCN high risk, GC ≤ 0.60', 'value': '5%', 'ci': None} {'timepoint': 'at-risk column 2 (year not explicitly printed)', 'arm': 'NCCN very high risk, GC > 0.85', 'value': '25%', 'ci': None} {'timepoint': 'at-risk column 3 (year not explicitly printed)', 'arm': 'NCCN very high risk, GC > 0.85', 'value': '29%', 'ci': None} {'timepoint': 'at-risk column 4 (year not explicitly printed)', 'arm': 'NCCN very high risk, GC > 0.85', 'value': '36%', 'ci': None} {'timepoint': 'at-risk column 2 (year not explicitly printed)', 'arm': 'NCCN very high risk, GC 0.61-0.85', 'value': '7%', 'ci': None} {'timepoint': 'at-risk column 3 (year not explicitly printed)', 'arm': 'NCCN very high risk, GC 0.61-0.85', 'value': '10%', 'ci': None} {'timepoint': 'at-risk column 4 (year not explicitly printed)', 'arm': 'NCCN very high risk, GC 0.61-0.85', 'value': '10%', 'ci': None} {'timepoint': 'at-risk column 2 (year not explicitly printed)', 'arm': 'NCCN very high risk, GC ≤ 0.60', 'value': '9%', 'ci': None} {'timepoint': 'at-risk column 3 (year not explicitly printed)', 'arm': 'NCCN very high risk, GC ≤ 0.60', 'value': '12%', 'ci': None} {'timepoint': 'at-risk column 4 (year not explicitly printed)', 'arm': 'NCCN very high risk, GC ≤ 0.60', 'value': '12%', 'ci': None} Note: Values are printed % distant metastasis beside numbers at risk; the timepoint of each at-risk column is not labeled on the slide.

[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. [ESMO Congress Madrid 2026 logo; presenter at podium] [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. [ESMO Congress Madrid 2026 logo; presenter at podium] [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. [ESMO Congress Madrid 2026 logo; presenter at podium] [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). [ESMO Congress Madrid 2026 logo; presenter at podium]

[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] 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 [affiliations and disclosure statement small print omitted]

[Slide 1] 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 [ESMO Congress Madrid 2026 logo] [Video inset of presenter at podium: MADRID SPAIN 23-27 OCTOBER 2026] 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)

[Slide 2] 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] 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 data captured as a table — no point estimates read from the curve] Endpoint: DFS (papillary-only subgroup; T1 subset reported in notes) 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) {'timepoint': '24 mo', 'arm': 'D+BCG papillary only', 'value': '87%', 'ci': None} {'timepoint': '24 mo', 'arm': 'BCG papillary only', 'value': '81%', 'ci': None} {'timepoint': '36 mo', 'arm': 'D+BCG papillary only', 'value': '84%', 'ci': None} {'timepoint': '36 mo', 'arm': 'BCG papillary only', 'value': '76%', 'ci': None} {'timepoint': '48 mo', 'arm': 'D+BCG papillary only', 'value': '84%', 'ci': None} {'timepoint': '48 mo', 'arm': 'BCG papillary only', 'value': '74%', 'ci': None} {'timepoint': '24 mo', 'arm': 'D+BCG T1 subset', 'value': '88%', 'ci': None} {'timepoint': '24 mo', 'arm': 'BCG T1 subset', 'value': '79%', 'ci': None} {'timepoint': '36 mo', 'arm': 'D+BCG T1 subset', 'value': '85%', 'ci': None} {'timepoint': '36 mo', 'arm': 'BCG T1 subset', 'value': '75%', 'ci': None} {'timepoint': '48 mo', 'arm': 'D+BCG T1 subset', 'value': '85%', 'ci': None} {'timepoint': '48 mo', 'arm': 'BCG T1 subset', 'value': '73%', 'ci': None} Note: Landmark % values printed on slide; timepoints assigned from dashed lines at 24/36/48 months. No p-value printed. Median follow-up in censored patients 60.7 months (both subgroups). [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 data captured as a table — no point estimates read from the curve] Endpoint: Cystectomy-free survival (papillary-only subgroup; T1 subgroup reported separately) 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) Note: Two subgroup KM panels on one slide; no p-values or landmark rates printed. Slide says 'favourable trend'.

[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. [ESMO Congress Madrid 2026 logo] [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. [ESMO Congress Madrid 2026 logo]

[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 data captured as a table — no point estimates read from the curve] 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) Note: Deep arm at-risk at time 0 differs between panels (390 CRPC-FS, 392 OS), so n left null for that arm. [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).

[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] 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 [i+12 Instituto de Investigación Hospital 12 de Octubre logo] 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 [Cris contra el cáncer logo]

[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 data captured as a table — no point estimates read from the curve] Endpoint: PFS 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 Note: PFS by disease volume. n = patients at risk at month 0. HR applies per volume subgroup, darolutamide + ADT vs ADT (CHAARTED). No p-values printed. [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 data captured as a table — no point estimates read from the curve] Endpoint: OS 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 Note: OS by disease volume; low-volume CI crosses 1. n = patients at risk at month 0. HR applies per volume subgroup, darolutamide + ADT vs ADT (CHAARTED). No p-values printed. [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 data captured as a table — no point estimates read from the curve] Endpoint: Time to mCRPC 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 Note: Time to mCRPC by disease volume. n = patients at risk at month 0. HR applies per volume subgroup, darolutamide + ADT vs ADT (CHAARTED). No p-values printed. [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 data captured as a table — no point estimates read from the curve] Endpoint: rPFS 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 Note: rPFS by disease volume. n = patients at risk at month 0. HR applies per volume subgroup, darolutamide + ADT vs ADT (CHAARTED). No p-values printed.

[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 data captured as a table — no point estimates read from the curve] Endpoint: PFS by investigator (INV) 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 {'timepoint': '24 mo', 'arm': 'Pooled DE/A/K EV+P', 'value': '39.9%', 'ci': '30.6, 49.1'} {'timepoint': '48 mo', 'arm': 'Pooled DE/A/K EV+P', 'value': '29.1%', 'ci': '20.1, 38.6'} {'timepoint': '60 mo', 'arm': 'Pooled DE/A/K EV+P', 'value': '24.3%', 'ci': '15.1, 34.7'} {'timepoint': '84 mo', 'arm': 'Pooled DE/A/K EV+P', 'value': '20.2%', 'ci': '10.6, 32.1'} {'timepoint': '24 mo', 'arm': 'EV monotherapy (Cohort K)', 'value': '13.7%', 'ci': '5.6, 25.5'} {'timepoint': '48 mo', 'arm': 'EV monotherapy (Cohort K)', 'value': '13.7%', 'ci': '5.6, 25.5'} {'timepoint': '60 mo', 'arm': 'EV monotherapy (Cohort K)', 'value': '13.7%', 'ci': '5.6, 25.5'} Note: No HR (study not designed to compare EV+P vs EV mono). Landmark timepoints inferred from drop-line positions on x-axis (not printed as labels). [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 data captured as a table — no point estimates read from the curve] Endpoint: Duration of response by investigator (INV) 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 {'timepoint': '24 mo', 'arm': 'Pooled DE/A/K EV+P', 'value': '53.7%', 'ci': '41.3, 64.6'} {'timepoint': '48 mo', 'arm': 'Pooled DE/A/K EV+P', 'value': '41.1%', 'ci': '28.5, 53.3'} {'timepoint': '60 mo', 'arm': 'Pooled DE/A/K EV+P', 'value': '37.0%', 'ci': '23.6, 50.4'} {'timepoint': '84 mo', 'arm': 'Pooled DE/A/K EV+P', 'value': '30.8%', 'ci': '16.2, 46.8'} {'timepoint': '24 mo', 'arm': 'EV monotherapy (Cohort K)', 'value': '29.4%', 'ci': '12.3, 48.8'} {'timepoint': '48 mo', 'arm': 'EV monotherapy (Cohort K)', 'value': '29.4%', 'ci': '12.3, 48.8'} Note: n = responders (ORR by INV 81/121 [66.9%] EV+P; 33/73 [45.2%] EV mono). No HR. Landmark timepoints inferred from drop-line positions. [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

[Slide 1] [ESMO Congress Madrid 2026 logo] 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] [ESMO Congress Madrid 2026 logo] 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] [ESMO Congress Madrid 2026 logo] 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] [ESMO Congress Madrid 2026 logo] 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] 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 data captured as a table — no point estimates read from the curve] 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) {'timepoint': '12 mo', 'arm': 'All (EFS)', 'value': '92.9%', 'ci': None} {'timepoint': '24 mo', 'arm': 'All (EFS)', 'value': '85%', 'ci': None} {'timepoint': '36 mo', 'arm': 'All (EFS)', 'value': '67.2%', 'ci': None} {'timepoint': '12 mo', 'arm': 'All (OS)', 'value': '96.7%', 'ci': None} {'timepoint': '24 mo', 'arm': 'All (OS)', 'value': '92.8%', 'ci': None} {'timepoint': '36 mo', 'arm': 'All (OS)', 'value': '86.6%', 'ci': None} Note: Cutoff Sep 1, 2026; median follow-up 31.6 months. RC vs non-RC subgroup comparison is non-randomized. [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] 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 urethraa • Not eligible for cisplatin-containing chemotherapyb • No prior systemic therapy for advanced or unresectable urothelial carcinomac • Measurable lesion per RECIST v1.1 • ECOG PS 0 or 1 → Sac-TMT 4 mg/kg or 5 mg/kg IV Q2Wd + 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 aParticipants with mixed histology were eligible provided that the urothelial component was >50% and the plasmacytoid component was <10% per local pathology assessment. Participants with non-urothelial cancer of the urinary tract were not eligible. bDefined as meeting ≥1 of the following criteria: creatinine clearance (calculated or measured) <60 mL/min but >30 mL/min, grade ≥2 audiometric hearing loss, grade ≥2 peripheral neuropathy, or NYHA Class III heart failure. cPlatinum-based therapy or nivolumab used in the adjuvant or neoadjuvant setting with recurrence >12 months since completion of therapy was permitted. Low-dose chemotherapy given concurrently with radiation to the primary tumor site was not considered systemic therapy. dParticipants received sac-TMT until PD or unacceptable toxicity and pembrolizumab until PD, unacceptable toxicity, or completion of 2 years of treatment. [Slide 2] ORR, DCR, and DOR by TROP2 and PD-L1 | Sac-TMT + pembro (N = 40) Confirmed ORR (95% CI), % | 68 (51–81) Subgroup, % (95% CI) [responders/N] TROP2 H-score <median | 72 (47–90) [13/18] TROP2 H-score ≥median | 68 (43–87) [13/19] PD-L1 CPS <10 | 72 (53–86) [23/32] PD-L1 CPS ≥10 | 50 (16–84) [4/8] Confirmed DCR (95% CI), % | 88 (73–96) Subgroup, % (95% CI) [responders/N] TROP2 H-score <median | 89 (65–99) [16/18] TROP2 H-score ≥median | 89 (67–99) [17/19] PD-L1 CPS <10 | 94 (79–99) [30/32] PD-L1 CPS ≥10 | 63 (25–92) [5/8] DOR, median (range), mo | 19.4 (2.0–25.8+) Subgroup, median, (range), mo [no. of responders] TROP2 H-score <median | NR (2.0–25.8+) [13] TROP2 H-score ≥median | 15.4 (3.6–22.2+) [13] PD-L1 CPS <10 | 19.4 (2.0–25.8+) [23] PD-L1 CPS ≥10 | 9.0 (3.9–22.1+) [4] Median study follow-up duration, 22.8 (range, 19.8–30.0) months. Median TROP2 H-score = 270. 95% CIs for ORR and DCR are based on the Clopper-Pearson method. Median DORs are based on the Kaplan-Meier method for censored data. Data cutoff, Jan 26, 2026. [Slide 3] PFS by TROP2 and PD-L1 [Left KM: PFS, % vs Time, mo] Median (95% CI), mo TROP2 H-score <median | 11.2 (3.7–NR) TROP2 H-score ≥median | 11.9 (5.6–21.1) No. at risk (0, 3, 6, 9, 12, 15, 18, 21, 24, 27, 30) <median | 18 16 11 9 7 7 7 6 3 1 0 ≥median | 19 17 12 10 9 8 6 4 0 0 0 [Right KM: PFS, % vs Time, mo] 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 (0, 3, 6, 9, 12, 15, 18, 21, 24, 27, 30) CPS <10 | 32 30 22 18 15 15 13 9 4 1 0 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 data captured as a table — no point estimates read from the curve] Endpoint: PFS by TROP2 H-score and PD-L1 CPS (single-arm sac-TMT + pembro, N=40) Overall cohort | 40 | 11.2 | 5.8-NR TROP2 H-score <median | 18 | 11.2 | 3.7-NR TROP2 H-score ≥median | 19 | 11.9 | 5.6-21.1 PD-L1 CPS <10 | 32 | 15.6 | 7.2-NR PD-L1 CPS ≥10 | 8 | 5.7 | 0.5-NR Note: Subgroup n from number-at-risk at time 0. Overall median OS NR (95% CI 17.2 mo-NR). No HRs printed (descriptive subgroups). Data cutoff Jan 26, 2026; median follow-up 22.8 mo.

[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 data captured as a table — no point estimates read from the curve] Endpoint: Overall survival Pooled DE/A/K EV+P | 121 | 27.7 mo | 20.9, 39.6 EV monotherapy | 73 | 22.8 mo | 16.3, 28.4 {'timepoint': '24 months [?]', 'arm': 'Pooled DE/A/K EV+P', 'value': '52.9%', 'ci': '43.6, 61.3'} {'timepoint': '48 months [?]', 'arm': 'Pooled DE/A/K EV+P', 'value': '38.8%', 'ci': '30.2, 47.4'} {'timepoint': '60 months', 'arm': 'Pooled DE/A/K EV+P', 'value': '33.0%', 'ci': '24.6, 41.6'} {'timepoint': '84 months [?]', 'arm': 'Pooled DE/A/K EV+P', 'value': '28.2%', 'ci': '19.6, 37.4'} {'timepoint': '24 months [?]', 'arm': 'EV monotherapy', 'value': '46.4%', 'ci': '34.6, 57.3'} {'timepoint': '48 months [?]', 'arm': 'EV monotherapy', 'value': '25.5%', 'ci': '16.0, 36.2'} {'timepoint': '60 months', 'arm': 'EV monotherapy', 'value': '23.9%', 'ci': '14.6, 34.5'} Note: Single-arm cohorts; no comparative statistics. 60-mo (5-yr) values corroborated by conclusions slide; other timepoints inferred from dashed lines. Data cutoff Feb 20, 2026.

[Slide 1] [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 [Author list and affiliations - fine print illegible] [VHIO logo; Hospital Universitario 12 de Octubre logo] 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] 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

[Slide 1] [MADRID 2026 ESMO congress logo] # 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. [IAF logo] 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)] [Madrid 2026 ESMO Congress logo] [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 [Madrid 2026 ESMO Congress logo] [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

[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.

[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 data captured as a table — no point estimates read from the curve] Endpoint: Castration resistant-free survival (CRFS) TRIDA (darolutamide + ADT + docetaxel, real-world) | 402 {'timepoint': '12 months', 'arm': 'TRIDA', 'value': '83.8%', 'ci': '79.5-88.1'} {'timepoint': '18 months', 'arm': 'TRIDA', 'value': '71.6%', 'ci': '64.2-79.0'} {'timepoint': '12 months', 'arm': 'ARASENS (reference, approximate)', 'value': '~78%', 'ci': None} {'timepoint': '18 months', 'arm': 'ARASENS (reference, approximate)', 'value': '~86', 'ci': None} Note: Single-arm real-world cohort; ARASENS values are approximate cross-trial reference figures printed with '~' (18m value printed without % sign). Median not reached/not printed. [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 data captured as a table — no point estimates read from the curve] Endpoint: Overall survival TRIDA (darolutamide + ADT + docetaxel, real-world) | 402 {'timepoint': '12 months', 'arm': 'TRIDA', 'value': '90.4%', 'ci': '86.9-93.9'} {'timepoint': '18 months', 'arm': 'TRIDA', 'value': '86.5%', 'ci': '81.0-92.0'} {'timepoint': '12 months', 'arm': 'ARASENS (reference, approximate)', 'value': '~93%', 'ci': None} {'timepoint': '18 months', 'arm': 'ARASENS (reference, approximate)', 'value': '~88', 'ci': None} Note: Single-arm real-world cohort; median OS not printed (caption says NR). ARASENS values approximate cross-trial reference. [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

[Slide 1] 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] 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 data captured as a table — no point estimates read from the curve] 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 Note: HR is per Decipher risk group comparison (continuous/grouped, as printed). Cumulative % DM printed beside numbers at risk at three unlabeled timepoints (columns approx. years 2, 5, 8 by position - not printed); not recorded as landmarks. ADT: GC>0.85 10%/27%/39%, GC0.61-0.85 5%/12%/20%, GC≤0.60 2%/5%/5%. ADT+abi: GC>0.85 4%/16%/19%, GC0.61-0.85 0%/2%/11%, GC≤0.60 0%/3%/14%. [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 data captured as a table — no point estimates read from the curve] 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* Note: % PCSM printed beside numbers at risk at three unlabeled timepoints; not recorded as landmarks. ADT: GC>0.85 2%/15%/26%, mid 0%/8%/12%, low 2%/5%/7%. ADT+abi: GC>0.85 4%/10%/22%, mid 0%/2%/5%, low 0%/3%/3%. [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] 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] 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 data captured as a table — no point estimates read from the curve] Endpoint: Cumulative incidence of treatment-emergent SCNEPC Enzalutamide | 563 Conventional NSAA | 562 Note: Cumulative-incidence (not survival) curve; no HR/IRR/p printed on this figure. Event counts are from the number-at-risk (event) rows. [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 data captured as a table — no point estimates read from the curve] Endpoint: OS from SCNEPC diagnosis ENZA | 15 | 9 months NSAA | 4 | 7 months All patients | 19 | 9 months (IQR 2-18) P: 0.3 Note: Medians printed in header box; events from final at-risk column. Very small n. [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

[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 [ESMO Congress Madrid 2026 logo] 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 [ESMO Congress logo] [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] [ESMO Congress logo] [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 [ESMO Congress logo]

[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 [ESMO Congress logo]

[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] [ESMO Congress Madrid 2026 logo] [Slide 4] VIOLET: 161Tb-PSMA-I&T in mCRPC [Peter Mac logo] [headshots: James Buteau, Louise Kostos, Michael Hofman, Arun Azad] 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] [ESMO Congress Madrid 2026 logo]

[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] 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 1] 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

[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 [ESMO Congress Madrid 2026 logo] [Survival data captured as a table — no point estimates read from the curve] 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 Note: Single-cohort RWD; no HR. PFS curve shows printed annotations 33%, 27%, 26% but their timepoints are not legibly labeled — not tabulated. Low-resolution screenshot; OS CI upper bound '43.5' read at low resolution. Cohort N illegible. [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] [ESMO Congress Madrid 2026 logo] [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] [ESMO Congress Madrid 2026 logo] [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] [ESMO Congress Madrid 2026 logo]

[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 data captured as a table — no point estimates read from the curve] 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 {'timepoint': '24 months', 'arm': 'Pembrolizumab + belzutifan (C)', 'value': '79.4%', 'ci': None} {'timepoint': '24 months', 'arm': 'Placebo (B)', 'value': '67.8%', 'ci': None} Note: Events box lists 'Pembrolizumab 151' (printed label; likely refers to the comparator/placebo arm events — not assigned). OS HR 0.56 (0.35-0.89), P=0.014 shown in header tile. Indirect comparison of reconstructed pseudo-IPD across two trials. [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 data captured as a table — no point estimates read from the curve] Endpoint: OS Placebo | 498 Pembrolizumab + belzutifan | 921 HR: 0.56 HR CI: 0.35-0.89 P: .014 Note: Medians not printed. Events row for comparator is labeled 'Pembrolizumab' on the figure although the legend arm is 'Placebo' - figure labeling inconsistency, reproduced verbatim. At-risk table printed at 5-month intervals.

[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. [ESMO Congress Madrid 2026 logo] [Survival data captured as a table — no point estimates read from the curve] Endpoint: DFS (papillary-only and T1 subgroups) 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 {'timepoint': '24 mo', 'arm': 'Papillary only - D+BCG (I+M)', 'value': '87%', 'ci': None} {'timepoint': '36 mo', 'arm': 'Papillary only - D+BCG (I+M)', 'value': '84%', 'ci': None} {'timepoint': '48 mo', 'arm': 'Papillary only - D+BCG (I+M)', 'value': '84%', 'ci': None} {'timepoint': '24 mo', 'arm': 'Papillary only - BCG (I+M)', 'value': '81%', 'ci': None} {'timepoint': '36 mo', 'arm': 'Papillary only - BCG (I+M)', 'value': '76%', 'ci': None} {'timepoint': '48 mo', 'arm': 'Papillary only - BCG (I+M)', 'value': '74%', 'ci': None} {'timepoint': '24 mo', 'arm': 'T1 - D+BCG (I+M)', 'value': '88%', 'ci': None} {'timepoint': '36 mo', 'arm': 'T1 - D+BCG (I+M)', 'value': '85%', 'ci': None} {'timepoint': '48 mo', 'arm': 'T1 - D+BCG (I+M)', 'value': '85%', 'ci': None} {'timepoint': '24 mo', 'arm': 'T1 - BCG (I+M)', 'value': '79%', 'ci': None} {'timepoint': '36 mo', 'arm': 'T1 - BCG (I+M)', 'value': '75%', 'ci': None} {'timepoint': '48 mo', 'arm': 'T1 - BCG (I+M)', 'value': '73%', 'ci': None} Note: Papillary-only HR stratified (prespecified sensitivity); T1 HR unstratified post-hoc exploratory. Median follow-up 60.7 mo in censored patients. Data cutoff 03 April 2025. [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. [ESMO Congress Madrid 2026 logo]

[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) [ESMO Congress Madrid 2026 logo] [Inset video feed of the speaker at the podium] [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) [ESMO Congress logo] [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). [ESMO Congress Madrid 2026 logo] [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). [ESMO Congress Madrid 2026 logo]

[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] 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 data captured as a table — no point estimates read from the curve] 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 Note: Medians printed as 'TTNT 6.7 vs. 5.8' and 'rwOS 15.3 vs. 8.3' without explicit arm attribution on the slide, so not assigned to arms here. rwOS at-risk n at time 0 is 52 (germline/tumoral) vs 8 (CH-only). Very small CH-only group (n=8). [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 2] 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 2] 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] 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.

[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] 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 [ESMO Congress Madrid 2026 logo]

[Slide 1] 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 [Slide 2] ORR, DCR, and DOR by TROP2 and PD-L1 | Sac-TMT + pembro (N = 40) Confirmed ORR (95% CI), % | 68 (51–81) Subgroup, % (95% CI) [responders/N] TROP2 H-score <median | 72 (47–90) [13/18] TROP2 H-score ≥median | 68 (43–87) [13/19] PD-L1 CPS <10 | 72 (53–86) [23/32] PD-L1 CPS ≥10 | 50 (16–84) [4/8] Confirmed DCR (95% CI), % | 88 (73–96) Subgroup, % (95% CI) [responders/N] TROP2 H-score <median | 89 (65–99) [16/18] TROP2 H-score ≥median | 89 (67–99) [17/19] PD-L1 CPS <10 | 94 (79–99) [30/32] PD-L1 CPS ≥10 | 63 (25–92) [5/8] | Sac-TMT + pembro (N = 40) DOR, median (range), mo | 19.4 (2.0–25.8+) Subgroup, median, (range), mo [no. of responders] TROP2 H-score <median | NR (2.0–25.8+) [13] TROP2 H-score ≥median | 15.4 (3.6–22.2+) [13] PD-L1 CPS <10 | 19.4 (2.0–25.8+) [23] PD-L1 CPS ≥10 | 9.0 (3.9–22.1+) [4] [Slide 3] 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) Time, mo: 0 | 3 | 6 | 9 | 12 | 15 | 18 | 21 | 24 | 27 | 30 No. at risk TROP2 <median | 18 | 16 | 11 | 9 | 7 | 7 | 7 | 6 | 3 | 1 | 0 TROP2 ≥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) Time, mo: 0 | 3 | 6 | 9 | 12 | 15 | 18 | 21 | 24 | 27 | 30 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 [Survival data captured as a table — no point estimates read from the curve] Endpoint: PFS by biomarker subgroup (single-arm, sac-TMT + pembrolizumab) TROP2 H-score <median | 18 | 11.2 mo | 3.7–NR TROP2 H-score ≥median | 19 | 11.9 mo | 5.6–21.1 PD-L1 CPS <10 | 32 | 15.6 mo | 7.2–NR PD-L1 CPS ≥10 | 8 | 5.7 mo | 0.5–NR Note: Biomarker subgroups within a single-arm cohort (N=40); no HR printed. n from numbers-at-risk at time 0. Low-resolution screenshot but values legible.

[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 data captured as a table — no point estimates read from the curve] Endpoint: Overall survival (exploratory cohort, n=20 mAPMS) Hannum clock Low Hannum clock High CH- CH+ Note: No medians, HRs, p-values, landmarks or numbers at risk printed; only curves.

[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 data captured as a table — no point estimates read from the curve] Endpoint: Overall survival and metastasis-free survival (single arm) DV + tislelizumab (single arm) | 42 {'timepoint': '2 years (OS)', 'arm': 'DV + tislelizumab', 'value': '87.1%', 'ci': None} {'timepoint': '2 years (MFS)', 'arm': 'DV + tislelizumab', 'value': '83.9%', 'ci': None} Note: Single-arm phase II; n=42 taken from enrollment on sibling slide (not printed on this slide). Median follow-up 24 months (IQR 16-30). No medians printed.

[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] 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] 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 [ESMO Congress Madrid 2026 logo] [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 [ESMO Congress Madrid 2026 logo] [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 [ESMO Congress Madrid 2026 logo]

[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 data captured as a table — no point estimates read from the curve] Endpoint: Time to definitive deterioration in GHS/QoL (EORTC QLQ-C30) 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 Note: Slide has 3 panels. Urinary symptoms (QLQ-PR25) TTDD: 42/240 NR (48.8–NR) vs 46/234 51.4 (48.8–NR), HR 0.94 (0.61–1.44). Worst pain (BPI-SF) TTD: 81/240 NR (32.9–NR) vs 93/234 NR (23.0–NR), HR 0.82 (0.61–1.11). Stratified HRs.

[Slide 1] [Collage of 4 photos: panel on stage, speaker at podium, three speakers posing, and a partially visible discussant slide] [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] 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 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] 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 [ESMO Congress Madrid 2026 logo]

[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 data captured as a table — no point estimates read from the curve] 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* {'timepoint': 'at-risk column 2 (time not printed)', 'arm': 'GC > 0.85', 'value': '10%', 'ci': None} {'timepoint': 'at-risk column 3 (time not printed)', 'arm': 'GC > 0.85', 'value': '27%', 'ci': None} {'timepoint': 'at-risk column 4 (time not printed)', 'arm': 'GC > 0.85', 'value': '39%', 'ci': None} {'timepoint': 'at-risk column 2 (time not printed)', 'arm': 'GC 0.61 - 0.85', 'value': '5%', 'ci': None} {'timepoint': 'at-risk column 3 (time not printed)', 'arm': 'GC 0.61 - 0.85', 'value': '12%', 'ci': None} {'timepoint': 'at-risk column 4 (time not printed)', 'arm': 'GC 0.61 - 0.85', 'value': '20%', 'ci': None} {'timepoint': 'at-risk column 2 (time not printed)', 'arm': 'GC ≤ 0.60', 'value': '2%', 'ci': None} {'timepoint': 'at-risk column 3 (time not printed)', 'arm': 'GC ≤ 0.60', 'value': '5%', 'ci': None} {'timepoint': 'at-risk column 4 (time not printed)', 'arm': 'GC ≤ 0.60', 'value': '5%', 'ci': None} Note: Second panel STAMPEDE ADT + abi: n 105/58/30; HR 1.27 (0.96 - 1.67), p 0.10; at-risk % GC>0.85 4%/16%/19%, GC0.61-0.85 0%/2%/11%, GC≤0.60 0%/3%/14%. Time points of at-risk columns not printed. HR comparison basis (e.g., per Decipher unit vs groups) not stated on slide. [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 data captured as a table — no point estimates read from the curve] 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* {'timepoint': 'at-risk column 2 (time not printed)', 'arm': 'GC > 0.85', 'value': '2%', 'ci': None} {'timepoint': 'at-risk column 3 (time not printed)', 'arm': 'GC > 0.85', 'value': '15%', 'ci': None} {'timepoint': 'at-risk column 4 (time not printed)', 'arm': 'GC > 0.85', 'value': '26%', 'ci': None} {'timepoint': 'at-risk column 2 (time not printed)', 'arm': 'GC 0.61 - 0.85', 'value': '0%', 'ci': None} {'timepoint': 'at-risk column 3 (time not printed)', 'arm': 'GC 0.61 - 0.85', 'value': '8%', 'ci': None} {'timepoint': 'at-risk column 4 (time not printed)', 'arm': 'GC 0.61 - 0.85', 'value': '12%', 'ci': None} {'timepoint': 'at-risk column 2 (time not printed)', 'arm': 'GC ≤ 0.60', 'value': '2%', 'ci': None} {'timepoint': 'at-risk column 3 (time not printed)', 'arm': 'GC ≤ 0.60', 'value': '5%', 'ci': None} {'timepoint': 'at-risk column 4 (time not printed)', 'arm': 'GC ≤ 0.60', 'value': '7%', 'ci': None} Note: Second panel STAMPEDE ADT + abi: n 105/58/30; HR 1.60 (1.16 - 2.21), p 0.004*; at-risk % PCSM GC>0.85 4%/10%/22%, GC0.61-0.85 0%/2%/5%, GC≤0.60 0%/3%/3%. Time points of at-risk columns not printed. [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 data captured as a table — no point estimates read from the curve] Endpoint: Distant metastasis (RWD NCCN high risk panel) GC > 0.85 | 2230 GC 0.61 - 0.85 | 1565 GC ≤ 0.60 | 1402 {'timepoint': 'at-risk column 2 (time not printed)', 'arm': 'GC > 0.85', 'value': '9%', 'ci': None} {'timepoint': 'at-risk column 3 (time not printed)', 'arm': 'GC > 0.85', 'value': '14%', 'ci': None} {'timepoint': 'at-risk column 4 (time not printed)', 'arm': 'GC > 0.85', 'value': '16%', 'ci': None} {'timepoint': 'at-risk column 2 (time not printed)', 'arm': 'GC 0.61 - 0.85', 'value': '4%', 'ci': None} {'timepoint': 'at-risk column 3 (time not printed)', 'arm': 'GC 0.61 - 0.85', 'value': '5%', 'ci': None} {'timepoint': 'at-risk column 4 (time not printed)', 'arm': 'GC 0.61 - 0.85', 'value': '5%', 'ci': None} {'timepoint': 'at-risk column 2 (time not printed)', 'arm': 'GC ≤ 0.60', 'value': '2%', 'ci': None} {'timepoint': 'at-risk column 3 (time not printed)', 'arm': 'GC ≤ 0.60', 'value': '3%', 'ci': None} {'timepoint': 'at-risk column 4 (time not printed)', 'arm': 'GC ≤ 0.60', 'value': '5%', 'ci': None} Note: Second panel RWD NCCN very high risk: n 517/157/85; at-risk % GC>0.85 25%/29%/36%, GC0.61-0.85 7%/10%/10%, GC≤0.60 9%/12%/12%. No HR printed. Time points of at-risk columns not printed.

[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 data captured as a table — no point estimates read from the curve] Endpoint: rPFS (investigator) by PSA response at 6 months 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 Note: Within-arm comparisons of PSA responders vs non-responders, not a randomized between-arm comparison. [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] 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] 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 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 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. [ESMO Congress Madrid 2026 logo] [Survival data captured as a table — no point estimates read from the curve] Endpoint: rPFS by baseline ctDNA status (investigator-assessed) 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 Note: Unstratified HRs; DCO Feb 18, 2026; no landmark rates printed. [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. [ESMO Congress Madrid 2026 logo] [Survival data captured as a table — no point estimates read from the curve] Endpoint: rPFS by ctDNA status at baseline and Week 5, TALA + ENZA arm (investigator-assessed) 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 Note: Unstratified; DCO Feb 18, 2026; no landmark rates printed. [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 [ESMO Congress Madrid 2026 logo]

[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).

[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 [ESMO Congress Madrid 2026 logo] [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 [ESMO Congress Madrid 2026 logo] [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 [ESMO Congress Madrid 2026 logo] [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 [ESMO Congress Madrid 2026 logo; lectern sign: MADRID SPAIN 23-27 OCTOBER 2026]

[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 [ESMO Congress Madrid 2026 logo] [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 [ESMO Congress Madrid 2026 logo] [Survival data captured as a table — no point estimates read from the curve] 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 Note: Plasma-visible PTEN loss: HR 3.50 (unadjusted, no CI printed); HR 2.11 (1.38–3.24) adjusted for ctDNA fraction. 58 of 245 patients tissue-only. Prognostic (non-randomized) comparison. [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 [ESMO Congress Madrid 2026 logo]

[Slide 1] 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

[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]

[Slide 2] 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] 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] 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 data captured as a table — no point estimates read from the curve] 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) Note: HR direction: <95% vs ≥95% reduction (HR 3.7). Exploratory biomarker association, single-arm phase I.

[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 [ESMO Congress Madrid 2026 logo] [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 [ESMO Congress Madrid 2026 logo]

[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 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" [Comprehensive Cancer Center logo] FPG500 program, NCT06020625 [Gemelli logo] CONTACT: daniela.arduini@guest.policlinicogemelli.it [X] DanielaArduini3

[Slide 4] 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. [ESMO Congress Madrid 2026 logo]

[Slide 2] [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]

[Slide 1] 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 [LACOG Latin American Cooperative Oncology Group logo] 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. [Disclosures, funding, affiliations fine print illegible] [Survival data captured as a table — no point estimates read from the curve] Endpoint: rPFS by AR amplification status (Figure 2) 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 {'timepoint': '12 mo rPFS', 'arm': 'AR non-amplified', 'value': '42.3%', 'ci': '26.2-67.6%'} {'timepoint': '12 mo rPFS', 'arm': 'AR amplified', 'value': '0.0%', 'ci': 'NE-NE'} {'timepoint': '12 mo OS', 'arm': 'AR non-amplified', 'value': '88.9%', 'ci': '73.0-95.7%'} {'timepoint': '12 mo OS', 'arm': 'AR amplified', 'value': '50%', 'ci': '15.2-77.5%'} Note: OS (Figure 3): amplified 6/8 events, median 11.6 mo (3.7-NE); non-amplified 9/38 events, median NR (21.6-NE); HR 0.19 (0.07-0.55), p=0.001. Exploratory biomarker analysis of single-arm phase II; small subgroup (n=8 amplified).

[Slide 1] 4737RO [Madrid 2026 ESMO congress logo] 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] 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] 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]]

[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. [ESMO Congress Madrid 2026 logo] [Survival data captured as a table — no point estimates read from the curve] Endpoint: EFS and OS (single-arm), FAS Toripalimab + TIP (EFS) | 25 | Not reached Toripalimab + TIP (OS) | 25 | Not reached {'timepoint': '12 mo', 'arm': 'EFS', 'value': '72.0%', 'ci': '50.1-85.5'} {'timepoint': '24 mo', 'arm': 'EFS', 'value': '67.8%', 'ci': '45.7-82.4'} {'timepoint': '12/24 mo', 'arm': 'OS', 'value': '88.0%', 'ci': '67.3-96.0'} Note: Single-arm phase II, no comparator HR. Median follow-up 21.4 mo. Historical comparators printed: ORR 50%, pCR 10% for TIP chemo. [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. [ESMO Congress Madrid 2026 logo] [Survival data captured as a table — no point estimates read from the curve] 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 {'timepoint': '1 yr', 'arm': 'HPV positive (OS)', 'value': '100%', 'ci': None} {'timepoint': '1 yr', 'arm': 'HPV negative (OS)', 'value': '75%', 'ci': None} Note: Exploratory, not statistically significant. pCR 66.7% vs 33.3% (p=0.22); 1-yr OS p=0.070. Arm-color mapping of at-risk rows inferred from curve labels (blue=HPV negative, red=HPV positive).

[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 data captured as a table — no point estimates read from the curve] Endpoint: rPFS (investigator-assessed), CTC-evaluable population, by baseline CTC 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) Note: Unstratified within-arm prognostic comparisons (not randomized treatment comparisons). No landmark rates printed. At-risk rows too small to read reliably. [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 data captured as a table — no point estimates read from the curve] Endpoint: rPFS (investigator-assessed), CTC-evaluable population, by Week 9 CTC conversion / CTC0 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) Note: Small subgroups; unstratified within-arm comparisons. No landmark rates printed. At-risk rows too small to read reliably.

[Slide 1] 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

[Slide 4] 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 1] 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.

[Slide 1] 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

[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

[Slide 1] [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). [ESMO Congress Madrid 2026 logo] [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 [ESMO Congress Madrid 2026 logo]

[Slide 1] 13:30 - 15:00 Rapid oral: Genitourinary tumours, prostate, penile and testis CHAIRS: THOMAS HWANG, EDMOND KWAN, ALISON REID [ESMO Congress Madrid 2026 logo] 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 [ESMO Congress Madrid 2026 logo] [Speaker card] Arun Azad BILBAO AUDITORIUM - HALL 3 [Survival data captured as a table — no point estimates read from the curve] Endpoint: Time to (definitive) deterioration — GHS/QoL (EORTC QLQ-C30), urinary symptoms (EORTC QLQ-PR25), worst pain (BPI-SF) 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 Note: Three panels on one slide; stratified HR (95% CI). Worst-pain PBO+ENZA lower CI bound partly illegible (23.6 or 23.9). No p-values printed. [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] [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. [ESMO Congress Madrid 2026 logos] [Survival data captured as a table — no point estimates read from the curve] 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) {'timepoint': '1 yr', 'arm': 'HPV positive', 'value': 'OS 100%', 'ci': None} {'timepoint': '1 yr', 'arm': 'HPV negative', 'value': 'OS 75%', 'ci': None} Note: OS by HPV: HR 0.42 (95% CI 0.09–1.96), log-rank p=0.24. 1-yr OS p=0.070. pCR 66.7% vs 33.3% p=0.22. Medians not printed.

[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 [ESMO Congress Madrid 2026 logo] [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 [ESMO Congress Madrid 2026 logo] [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 [ESMO Congress Madrid 2026 logo] [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 [ESMO Congress Madrid 2026 logo]

[Slide 1] [Photo of presenter beside poster; signs: POSTER AREA A; RENAL CANCER; board 3601P] 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, Renee 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 [affiliation superscripts illegible] 1. Dana-Farber Cancer Institute, Boston, MA, USA; 2. Beth Israel Deaconess Medical Center, Boston, MA, USA; 3. Yale Cancer Center, New Haven, CT, USA; 4. Johns Hopkins Hospital, Baltimore, MD, USA; 5. Memorial Sloan Kettering Cancer Center, NYC, NY, USA; 6. Brigham and Women's Hospital, Boston, MA, USA 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 [Figure 1. Study overview; Figure 2. Variant calling pipelines for significantly mutated genes (SMGs) and recurrent copy number variants (CNVs); Figure 3. Clinical characteristics of the aggregated cohort - figure text largely illegible] 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-y 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). [Figure 4. Genomic and survival analyses of NF2 loss-of-function (LOF) tumors - details illegible] - 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-y 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. [Figure 5. Genomic and clinical features of MLH1, B2M, and STK11 - illegible] Email: eddy_saad@dfci.harvard.edu @eddy_saad [Logos: Dana-Farber Cancer Institute, Yale Cancer Center, Beth Israel Deaconess Medical Center] [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 [Predicine logo]

[Slide 1] 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] 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 data captured as a table — no point estimates read from the curve] Endpoint: rPFS (ITT), primary endpoint 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 Note: CIs are 90% CIs as printed. n taken from number at risk at time 0. [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] 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 data captured as a table — no point estimates read from the curve] Endpoint: OS by baseline ctDNA tumor fraction (>2% vs ≤2%) TF >2% (High) | 46 TF ≤2% (Low) | 17 P: <0.0001 Note: No medians printed. Continuous TF Cox HR per unit TF: OS 25.6 (P 2.6e-6), CR-PFS 6.5, PSA-PFS 6.4. [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.

[Slide 1] [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 data captured as a table — no point estimates read from the curve] 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 Note: No medians, HR, or at-risk numbers printed. Interpretation as Alzheimer's/dementia-free survival is inferred from ADGC (Alzheimer's Disease Genetics Consortium) label and caption context.

[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 data captured as a table — no point estimates read from the curve] Endpoint: rPFS by baseline ctDNA status 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 Note: Unstratified HRs: A vs B 1.59 (0.94-2.70) p=0.08; A vs C 0.47 (0.28-0.78) p=0.003; B vs D 0.62 (0.39-0.96) p=0.03; C vs D 2.15 (1.42-3.26) p=0.0002. [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 data captured as a table — no point estimates read from the curve] Endpoint: rPFS by ctDNA status at BL/Week 5 (TALA + ENZA arm) 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 Note: Also C vs A 0.23 (0.09-0.56) p=0.0004; C vs B 0.72 (0.38-1.39) p=0.33. Unstratified. [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 data captured as a table — no point estimates read from the curve] Endpoint: rPFS by ctDNA status at BL/Week 5 (PBO + ENZA arm) 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 Note: Also C vs A 0.23 (0.12-0.43) p<0.0001; C vs B 0.55 (0.33-0.92) p=0.02. Unstratified.

[Slide 2] [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] 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] 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] [Best Poster rosette] BioChiP: A phase II, open-label, biomarker-guided study of carboplatin efficacy in pretreated metastatic castration resistant prostate cancer [author list - illegible] [VHIO logo] [Hospital Universitario 12 de Octubre logo] 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] 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] 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] 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. [ESMO Congress Madrid 2026 logo] [Survival data captured as a table — no point estimates read from the curve] Endpoint: rPFS (investigator) by ctDNA status at baseline and Week 5, PBO + ENZA arm 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 Note: Additional unstratified HRs printed: C vs A 0.23 (0.12–0.43) p<0.0001; C vs B 0.55 (0.33–0.92) p=0.02. Data cutoff Feb 18, 2026. [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. [ESMO Congress Madrid 2026 logo] [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... [Podium: ESMO congress MADRID SPAIN 23-27 OCTOBER 2026] BILBAO AUDITORIUM - HALL 3

[Slide 3] 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] [QR code] himalayasprostatestudy.com

[Slide 1] Prior hypothesis 1. Findings in muscle invasive bladder cancer Ambassador: Bladder primary, node negative, with or without prior cisplatin [Small KM: Disease-free survival; Pembro vs Observ.; embedded table: PEMBROLIZUMAB [illegible] NE (NE-NE); OBSERVATION [illegible] 27.1 (13.6-NE)?; HR (95% CI) [illegible]; at-risk rows illegible] CheckMate 274: Bladder primary, prior cisplatin Treatment | Median DFS (95% CI), months NIVO (n = 142): 19.6 (15.6-48.2) PBO (n = 142): 8.3 (5.6-11.2) HR (95% CI), 0.63 (0.47-0.84) Landmarks: NIVO 46.8% (36 mo), 42.6% (48 mo), 33.4% (60 mo); PBO 29.3% (36 mo), 27.8% (48 mo), 26.2% (60 mo) Disease-free survival probability (%) / Months No. at risk NIVO: 142 105 89 71 64 60 56 52 47 43 9 4 3 2 1 1 0 PBO: 142 77 55 46 43 40 38 38 36 32 3 1 1 1 1 0 0 Apolo et al. ESMO 2024 Galsky et al. Ann Oncol. Volume 37, Issue 1, 2026 [ESMO Congress Madrid 2026 logo] [Survival data captured as a table — no point estimates read from the curve] Endpoint: DFS (CheckMate 274 subgroup: bladder primary, prior cisplatin) NIVO | 142 | 19.6 mo | 15.6-48.2 PBO | 142 | 8.3 mo | 5.6-11.2 HR: 0.63 HR CI: 0.47-0.84 {'timepoint': '36 months', 'arm': 'NIVO', 'value': '46.8%', 'ci': None} {'timepoint': '48 months', 'arm': 'NIVO', 'value': '42.6%', 'ci': None} {'timepoint': '60 months', 'arm': 'NIVO', 'value': '33.4%', 'ci': None} {'timepoint': '36 months', 'arm': 'PBO', 'value': '29.3%', 'ci': None} {'timepoint': '48 months', 'arm': 'PBO', 'value': '27.8%', 'ci': None} {'timepoint': '60 months', 'arm': 'PBO', 'value': '26.2%', 'ci': None} Note: AMBASSADOR panel too small to read reliably; not tabulated. [Slide 2] Prior hypothesis 2. Findings in metastatic disease Concomitant ICPI and chemotherapy may be detrimental JAVELIN 100 bladder trial: overall survival Avelumab Plus BSC (n = 350) | BSC Alone (n = 350) Events, No. (%): 215 (61.4) | 237 (67.7) OS, median (95% CI), months: 23.8 (19.9 to 28.8) | 15.0 (13.5 to 18.2) Stratified HR (95% CI): 0.76 (0.63 to 0.91) 2-sided P: .0036 Landmarks: Avelumab 49.8% (24 mo), 36.0% (36 mo); BSC 38.4% (24 mo), 29.8% (36 mo) OS (%) / Time (months) No. at risk: Avelumab plus BSC: 350 318 274 237 216 183 164 140 99 74 53 31 13 4 1 0 BSC alone: 350 304 243 190 158 131 121 103 82 62 46 27 10 7 0 Powles et al. J Clin Oncol. 2023 Apr 18;41(19):3486–3492 [Survival data captured as a table — no point estimates read from the curve] Endpoint: Overall survival (JAVELIN Bladder 100) Avelumab plus BSC | 350 | 23.8 mo | 19.9 to 28.8 BSC alone | 350 | 15.0 mo | 13.5 to 18.2 HR: 0.76 HR CI: 0.63 to 0.91 P: .0036 (2-sided) {'timepoint': '24 months', 'arm': 'Avelumab plus BSC', 'value': '49.8%', 'ci': None} {'timepoint': '36 months', 'arm': 'Avelumab plus BSC', 'value': '36.0%', 'ci': None} {'timepoint': '24 months', 'arm': 'BSC alone', 'value': '38.4%', 'ci': None} {'timepoint': '36 months', 'arm': 'BSC alone', 'value': '29.8%', 'ci': None} Note: Published long-term follow-up (Powles JCO 2023). [Slide 3] Prior hypothesis 2. Findings in metastatic disease Keynote-361: Cisplatin-allocated subgroup Progression-free survival: Pembrolizumab plus cisplatin-based chemotherapy - events 109 (69.9), PFS median 8.5 (8.2-10.3); Cisplatin-based chemotherapy - 98 (62.8), 7.1 (6.3-8.0); HR 0.67 (0.51-0.89) [small print, read with moderate confidence] Landmarks: 36.8% vs 16.4% (12 mo); 23.6% vs 10.6% (24 mo) No. at risk: 156 128 116 62 47 43 34 29 23 17 11 5 3 0 0 / 156 125 88 24 13 10 7 7 6 4 3 3 1 0 0 [small print] Overall survival: Pembrolizumab plus cisplatin-based chemotherapy - 102 (65.4), OS median 20.1 (14.6-24.1); Cisplatin-based chemotherapy - 108 (69.2), 16.4 (13.6-19.3); HR 0.80 (0.67-1.15) [small print] Landmarks: 67.3% vs 62.8% (12 mo); 42.8% vs 35.9% (24 mo) CheckMate 901: Gem-cis +/- nivolumab Progression-free Survival: Nivolumab+Gemcitabine-Cisplatin 211/304, median PFS 7.9 (7.6-9.5) mo; Gemcitabine-Cisplatin 191/304, 7.6 (6.1-7.8); Hazard ratio for disease progression or death, 0.72 (95% CI, 0.59-0.88); P=0.001 Landmarks: 34.2 (95% CI, 28.6-40.0) and 23.5 (95% CI, 18.3-29.0) nivo arm; 21.8 (95% CI, 16.1-27.9) and 9.6 (95% CI, 5.6-15.0) gem-cis [small print] No. at risk: Nivolumab+gemcitabine-cisplatin 304 179 82 57 41 31 19 11 6 1 0; Gemcitabine-cisplatin 304 119 35 17 10 8 5 1 0 0 0 Overall Survival: Nivolumab+Gemcitabine-Cisplatin 172/304, median OS 21.7 (18.6-26.4) mo; Gemcitabine-Cisplatin 193/304, 18.9 (14.7-22.4); Hazard ratio for death, 0.78 (95% CI, 0.63-0.96); P=0.02 Landmarks: 70.2 (95% CI, 64.6-75.2) and 46.9 (95% CI, 40.7-52.8) nivo arm; 62.7 (95% CI, 56.8-68.1) and 40.7 (95% CI, 34.6-46.7) gem-cis [small print] No. at risk: 304 264 196 142 97 69 48 25 15 7 2 0 / 304 242 166 122 82 49 33 17 13 4 1 0 van der Heijden et al. N Engl J Med 2023;389:1778-89 Powles et al. Clin Genitourin Cancer. 2025 Feb;23(1):102261 [ESMO Congress Madrid 2026 logo] [Survival data captured as a table — no point estimates read from the curve] Endpoint: PFS (CheckMate 901, nivolumab + gem-cis vs gem-cis) Nivolumab+Gemcitabine-Cisplatin | 304 | 7.9 mo | 7.6-9.5 Gemcitabine-Cisplatin | 304 | 7.6 mo | 6.1-7.8 HR: 0.72 HR CI: 0.59-0.88 P: 0.001 Note: CheckMate 901 OS: 172/304 21.7 (18.6-26.4) vs 193/304 18.9 (14.7-22.4), HR 0.78 (0.63-0.96), P=0.02. KEYNOTE-361 cisplatin subgroup PFS HR 0.67 (0.51-0.89), OS HR 0.80 (0.67-1.15) - small print, verify against source. Landmark values small and partially uncertain, omitted here. [Slide 4] Prior hypothesis 3. Findings from NIAGARA Event-free survival HR, 0.68; 95% CI, 0.56–0.82; P<0.0001 12 months: 76.0% (Durvalumab arm) vs 69.9% (Comparator arm) 24 months: 67.8% vs 59.8% Probability of EFS / Time from randomization (months) No. of patients at risk Durvalumab arm: 533 454 386 348 330 312 255 180 115 32 1 Comparator arm: 530 416 343 300 281 259 214 159 94 24 2 Overall survival Durvalumab arm N=533 | Comparator arm N=530 Number of deaths, n (%): 136 (25.5) | 169 (31.9) HR (95% CI): 0.75 (0.59–0.93) Stratified log-rank P value*: 0.0106 12 months: 89.5% vs 86.5%; 24 months: 82.2% vs 75.2% Probability of survival / Time from randomisation (months) [OS number-at-risk rows too small to read reliably] Powles et al. 2024. NEJM:391; 1773 - 1786 [ESMO Congress Madrid 2026 logo] [Survival data captured as a table — no point estimates read from the curve] Endpoint: Event-free survival (NIAGARA) Durvalumab arm | 533 Comparator arm | 530 HR: 0.68 HR CI: 0.56–0.82 P: <0.0001 {'timepoint': '12 months', 'arm': 'Durvalumab arm', 'value': '76.0%', 'ci': None} {'timepoint': '12 months', 'arm': 'Comparator arm', 'value': '69.9%', 'ci': None} {'timepoint': '24 months', 'arm': 'Durvalumab arm', 'value': '67.8%', 'ci': None} {'timepoint': '24 months', 'arm': 'Comparator arm', 'value': '59.8%', 'ci': None} Note: OS panel: deaths 136 (25.5%) vs 169 (31.9%), HR 0.75 (0.59–0.93), stratified log-rank P=0.0106; 12-mo 89.5 vs 86.5%, 24-mo 82.2 vs 75.2%. Medians not printed.

[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 data captured as a table — no point estimates read from the curve] Endpoint: Investigator-assessed rPFS 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 {'timepoint': '36 months (dashed line)', 'arm': 'TALA + ENZA', 'value': '77%', 'ci': '71–81'} {'timepoint': '36 months (dashed line)', 'arm': 'PBO + ENZA', 'value': '56%', 'ci': '50–62'} Note: Landmark % labels printed at the dashed 36-month line; timepoint inferred from dashed line position at x=36 (not printed as text). [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] Patient with metastatic disease • histologically confirmed chRCC • ECOG PS 0-3 • measurable disease per RECIST 1.1 • any IMDC risk • ≥18 years old INCLUSION no previous systemic therapy → PEMBROLIZUMAB + LENVATINIB n=20 progression on systemic therapy, lenvatinib-naïve → LENVATINIB + EVEROLIMUS n=20 Primary endpoint: objective response rate Secondary endpoints: progression-free survival, overall survival, duration of response, safety [Slide 2] Duration of Tumor Control in ChromoKey Cohort A Patient (top to bottom): 15, 4, 13, 2, 10, 17, 6, 16, 5, 1, 18, 11, 7, 3, 20, 9, 19, 8, 12, 14 Legend: Partial response (PR); Stable disease (SD); Progression; Ongoing / censored X-axis: Duration of response or stable disease (months), 0–18

[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 data captured as a table — no point estimates read from the curve] 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* {'timepoint': 'last at-risk column (time not labelled)', 'arm': 'ADT: GC > 0.85', 'value': '26%', 'ci': None} {'timepoint': 'last at-risk column (time not labelled)', 'arm': 'ADT: GC 0.61-0.85', 'value': '12%', 'ci': None} {'timepoint': 'last at-risk column (time not labelled)', 'arm': 'ADT: GC ≤ 0.60', 'value': '7%', 'ci': None} {'timepoint': 'last at-risk column (time not labelled)', 'arm': 'ADT + abi: GC > 0.85', 'value': '22%', 'ci': None} {'timepoint': 'last at-risk column (time not labelled)', 'arm': 'ADT + abi: GC 0.61-0.85', 'value': '5%', 'ci': None} {'timepoint': 'last at-risk column (time not labelled)', 'arm': 'ADT + abi: GC ≤ 0.60', 'value': '3%', 'ci': None} Note: Percentages are cumulative % PCSM printed in the at-risk table; intermediate columns in ocr_text. Column timepoints not labelled. [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] (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] HRR mutations concordant; PTEN loss graded, not absent HRR / BRCA and PI3K-AKT alterations: plasma ctDNA vs matched tissue FRESH (n=102) | ARCHIVAL (n=72) ctDNA fraction % biopsy Gene rows: BRCA2, BRCA1, ATM, CDK12, CHEK2, PALB2, NBN, RAD51B, RAD51C, RAD51D, PIK3CA, PIK3R1, AKT1 PTEN tissue PTEN ctDNA per-gene rows (black = CHIP, excluded on matched leukocyte DNA - a false plasma call) concordant | tissue-only | ctDNA-only | clonal haematopoiesis PTEN block - resolved by depth tissue biallelic loss (called) | ctDNA clonal biallelic | not evaluable (ctDNA fraction < 20%) tissue loss below the biallelic call | ctDNA subclonal / single-copy | excluded, coverage dropout [Slide 3] | 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% [Slide 4] BRCA1/2 + HRRm 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: 7 | 4 4 plasma-only: 2 | 1 | 1 not evaluable - ctDNA fraction below the class floor | technical - coverage / SNP filter or annotation assessable, still discordant - biology | no explanation identified

[Slide 1] [Photo of auditorium screen and speaker at podium; ESMO Congress Madrid Spain 23-27 October 2026] Still Hooked on the Receptor Androgen Receptor (AR) targeted therapies are the mainstay of prostate cancer Resistance to 2nd generation AR pathway inhibitor (ARPI) is still driven by AR What comes after abiraterone, enzalutamide, apalutamide, and darolutamide ? [Diagram: Androgen synthesis: adrenal glands; Androgen activity: e.g. prostate; hypothalamus-pituitary-testes axis; small labels illegible] All "-lutamide" ARPIs bind to the androgen receptor [AR structure: Ligand-binding domain; DNA-binding domain; N-terminal domain] Alice BERNARD-TESSIER, MD Watson et al., Nat Rev Cancer 2015 [ESMO Congress Madrid 2026 logo]

[Slide 1] Figure 1. Proposed mechanism of action of SYNC-T Therapy SV-102 PATIENT-SPECIFIC TUMOR ANTIGENS RELEASED TUMOR + DRUG AND TUMOR ANTIGENS SYNCHRONIZE LOCATION AND ACTIVATE THE IMMUNE SYSTEM TUMOR = Created in BioRender 1 Partial cryolysis of a single target tumor to release a heterogeneous population of patient-specific tumor antigens 2 Through same device, a multi-target drug is infused into the same tumor. Antigens & drug drain into TDLNs, synchronizing location with immune cells 3 Co-localization can create conditions for T cell activation and systemic anti-tumor response while minimizing systemic drug exposure & toxicity [Slide 2] Overall survival probability (%) / Time (months) Number of patients: 15 Number of events: 6 Median (95% CI): NA (27.9, NA) n: 15 13 12 3 0 (at 0, 10, 20, 30, 40 months) [Survival data captured as a table — no point estimates read from the curve] Endpoint: Overall survival All patients | 15 | NA (not reached) | 27.9, NA Note: Single-arm; no landmark rates printed. [Slide 3] Overall survival probability (%) / Time (months) Number of patients: 5 [HS]; 10 [CRPC] Number of events: 0 [HS]; 6 [CRPC] Median (95% CI): NA (NA, NA) [HS]; 27.9 (17.1, NA) [CRPC] Legend: HS | CRPC n (CRPC): 10 8 7 2 0 n (HSPC): 5 5 5 1 0 (at 0, 10, 20, 30, 40 months) [Survival data captured as a table — no point estimates read from the curve] Endpoint: Overall survival by disease state HS (hormone-sensitive) | 5 | NA | NA, NA CRPC | 10 | 27.9 | 17.1, NA Note: Descriptive, no comparison statistics printed. Legend says HS, at-risk row says HSPC. [Slide 4] Grade 1–2 | Grade 3 Hematuria 4 (26.7%) Fever 3 (20.0%) Diarrhea 2 (13.3%) Vomiting 2 (13.3%) Urinary tract infection 2 (13.3%) Diaphoresis / hyperhidrosis 2 (13.3%) Rash 2 (13.3%) Urinary retention 2 (13.3%) Spinal cord compression 1 (6.7%) Patients with event, % (N=15)

[Slide 1] HRD (ProBio definition) Left: No longer clinically benefitting (%) vs Time from randomization, months Physician's choice At Risk 19 5 2 1 1 | Events 0 12 14 14 14 Niraparib At Risk 19 11 4 1 0 | Events 0 6 12 12 12 Right: Survival probability (%) vs Time from randomization, months Physician's choice At Risk 19 12 4 2 1 | Events 0 4 9 9 9 Niraparib At Risk 19 14 10 2 0 | Events 0 3 5 9 10 HRD (MAGNITUDE definition) Left: Physician's choice At Risk 16 4 2 1 1 | Events 0 10 11 11 11 Niraparib At Risk 19 11 4 1 0 | Events 0 6 12 12 12 Right: Physician's choice At Risk 16 11 4 2 1 | Events 0 2 7 7 7 Niraparib At Risk 19 14 10 2 0 | Events 0 3 5 9 10 BRCA complex Left: Physician's choice At Risk 10 1 0 0 | Events 0 7 8 8 Niraparib At Risk 9 4 2 1 | Events 0 3 5 5 Right: Physician's choice At Risk 10 7 3 1 0 | Events 0 2 6 6 6 Niraparib At Risk 9 5 4 1 0 | Events 0 2 3 4 4 Group: Physician's choice | Niraparib [Survival data captured as a table — no point estimates read from the curve] Endpoint: Time to no longer clinically benefitting (left panels) and overall survival (right panels), by biomarker subgroup Niraparib — HRD (ProBio) | 19 Physician's choice — HRD (ProBio) | 19 Niraparib — HRD (MAGNITUDE) | 19 Physician's choice — HRD (MAGNITUDE) | 16 Niraparib — BRCA complex | 9 Physician's choice — BRCA complex | 10 Note: No medians, HRs or p-values printed on this figure. Events = final cumulative events row.

[Slide 1] Figure 2: Response to gem-iDRS monotherapy in SunRISe-1 Cohort 2 among patients who had sequenced urine samples (N=58) Best overall response [Swimmer plot: one bar per patient, treatment time 0–39 months] Best overall response CR ≥12 months CR <12 months Non-CR Not evaluable Response assessment CR Non-CR Not evaluable utDNA MRD status Positive Negative Status On treatment CR Treatment time (months) 0 3 6 9 12 15 18 21 24 27 30 33 36 39

[Slide 3] Figure 3: Daily life activity behavior in the ITP ADT-naïve subgroup. Panels: Low HF severity | High HF severity Top row Y-axis: Mean ± SEM awake sedentary time (min/d); Bottom row Y-axis: Mean ± SEM awake MVPA time (%) X-axis timepoints: Baseline, C4D1, C7D1 [Line plots with error bars; no data values printed] [Slide 4] Figure 4: Daily life activity behavior in the ITP HF-emergent subgroup. Panels: Low HF severity | High HF severity Top row Y-axis: Mean ± SEM awake sedentary time (min/d); Bottom row Y-axis: Mean ± SEM awake MVPA time (%) X-axis timepoints: Baseline, C4D1, C7D1 [Line plots with error bars; no data values printed]

[Slide 3] rPFS per PCWG3-Modified RECIST v1.1 by BICR Total population [KM curve] | | 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) | Landmark labels at dashed line (6 months): 49% (pembro + chemo), 26% (chemo) Y axis: rPFS, %; X axis: Months (0-16) No. at risk 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. [MADRID 2026 ESMO congress logo; presenter video inset] [Survival data captured as a table — no point estimates read from the curve] Endpoint: rPFS per PCWG3-modified RECIST v1.1 by BICR Pembro + chemo (total) | 40 | 5.1 mo | 4.1-6.3 Chemo (total) | 30 | 4.0 mo | 2.1-5.0 Pembro + chemo (CPS >=1) | 15 | 6.0 mo | 3.9-NR Chemo (CPS >=1) | 12 | 3.8 mo | 0.4-5.0 Pembro + chemo (CPS <1) | 23 | 5.1 mo | 4.0-6.3 Chemo (CPS <1) | 16 | 4.1 mo | 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) {'timepoint': '6 mo (dashed line)', 'arm': 'Pembro + chemo (total)', 'value': '49%', 'ci': None} {'timepoint': '6 mo (dashed line)', 'arm': 'Chemo (total)', 'value': '26%', 'ci': None} Note: Landmark percentages printed beside dashed vertical line at 6 months. Data cutoff November 19, 2025. No p-values printed. [Slide 4] Overall Survival Total population [KM curve] | | 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) | Landmark labels at dashed line (6 months): 85% (pembro + chemo), 59% (chemo) Y axis: OS, %; X axis: Months (0-24) No. at risk 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. [MADRID 2026 ESMO congress logo; presenter video inset] [Survival data captured as a table — no point estimates read from the curve] Endpoint: Overall survival Pembro + chemo (total) | 40 | 10.2 mo | 8.0-11.4 Chemo (total) | 30 | 7.1 mo | 4.0-8.2 Pembro + chemo (CPS >=1) | 15 | 11.0 mo | 8.7-11.9 Chemo (CPS >=1) | 12 | 4.8 mo | 1.9-8.6 Pembro + chemo (CPS <1) | 23 | 8.1 mo | 6.6-13.0 Chemo (CPS <1) | 16 | 7.1 mo | 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) {'timepoint': '6 mo (dashed line)', 'arm': 'Pembro + chemo (total)', 'value': '85%', 'ci': None} {'timepoint': '6 mo (dashed line)', 'arm': 'Chemo (total)', 'value': '59%', 'ci': None} Note: OS secondary endpoint; PD-L1 subgroups post hoc (per s1). Data cutoff November 19, 2025. No p-values printed.

[Slide 1] WUTSUP-02: Background • UTUC is frequently invasive at diagnosis with high recurrence risks, and its neoadjuvant chemotherapy pCR rate is generally lower than that of bladder cancer; Furthermore, Renal function impairment and cisplatin ineligibility limit treatment options for many UTUC patients, highlighting an urmet need for novel perioperative strategies • This phase II study aimed to explore the efficacy and safety profile of neoadjuvant Disitamab vedotin (DV, a HER2-targeting ADC) combined with perioperative Tislelizumab (a PD-1 inhibitor) in high risk, non-metastatic UTUC with HER2 expression 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%) [Small KM plots: Overall Survival; Metastasis-free Survival; Survival Probability vs Time (Months) 0-30]

[Slide 4] Safety profile was consistent with prior reports, with no new safety signals observed with long-term follow-up1-3 TRAEs, n (%) | EV+P (pooled DE/A/K) n=121 | EV mono (cohort K) n=73 Any TRAE | 119 (98.3) | 68 (93.2) Any grade ≥3 TRAE | 78 (64.5) | 35 (47.9) Any serious TRAE | 26 (21.5) | 11 (15.1) TRAEs leading to discontinuation of EV only | 34 (28.1) | 17 (23.3) TRAEs leading to discontinuation of P only | 24 (19.8) | – TRAEs leading to discontinuation of both EV and P | 13 (10.7) | – TRAE leading to discontinuation of either EV or P | 61 (50.4) | 17 (23.3) Any TRAE leading to deatha | 4 (3.3) | 2 (2.7) Data cutoff: February 20, 2026. Treatment relatedness was determined by the investigator, including causality to EV and/or P as applicable. aCauses of death in the EV+P pooled DE/A/K cohort were pneumonitis (n=1), multiple organ dysfunction syndrome (n=1), respiratory failure (n=1), and sepsis (n=1). Causes of death in the EV mono cohort were respiratory failure (n=1) and multiple organ dysfunction syndrome (n=1). DE, dose escalation; EV, enfortumab vedotin; P, pembrolizumab; TRAE, treatment-related adverse event. 1. Rosenberg JE, et al. ESMO Open. 2026;11:107700. 2. Friedlander TW, et al. ESMO 2025. Abstract 3074P (Poster). 3. O'Donnell PH, et al. J Clin Oncol. 2023;41:4107-4117. Presented by: Dr. Jose Pablo Maroto Rey

[Slide 3] Event-free survival and bladder preservation At a median follow-up of 15.7 months, no recurrence was observed after pCR Individual patient course · All 31 treated patients N+V (n = 16) | N+P (n = 15) Legend: Major pathological response | pCR: pT0 | Pathological non-response | Response not assessable | Survival follow-Up End of trial (EOT) | Recurrence | Re-TURBT | Radical cystectomy | Treatment for recurrence | Adjuvant treatment Time from initiation of trial treatment [months] 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 Key findings • No recurrence after pCR: N+V 5/5 · N+P 1/1 without recurrence; Median follow-up 15.7 months (IQR 15.1–16.5) • Bladder preservation: 4/5 N+V patients with pCR underwent re-TURBT alone and retained their bladder • EFS in the ITT population: 6 events, 3 per arm; median not reached • Adjuvant therapy: Received by 5 patients, all in the N+P arm. Prof. Andrea Necchi [Slide 4] Adverse event profile No new safety signal was identified with the addition of visugromab Adverse events in all treated patients Patients, n (%) | N+P n = 15 | N+V n = 16 Any TEAE | 15 (100) | 16 (100) Grade 3–4 TEAE | 12 (80) | 8 (50) Any TRAE | 8 (53) | 14 (88) Grade 3–4 TRAE | 3 (20) | 2 (13) Serious TRAE | 1 (7) | 2 (13) Immune-related AE, any grade | 4 (27) | 5 (31) Immune-related AE, grade 3–4 | 3 (20) | 2 (13) TRAE leading to discontinuation | 2 (13) | 0 Treatment-related death | 0 | 0 Treatment-related delay of surgery | 0 | 0 Key findings • Any-grade TRAEs were more frequent with N+V. Grade 3–4 TRAEs and immune-related AE rates were similar between arms • Two Grade 3–4 TRAEs with N+V, each reported in 1 pt: Immune-mediated hepatic disorder with transaminase increase; interstitial lung disease • No treatment-related discontinuation, death or surgical delay with N+V. No postoperative complication was attributed to neoadjuvant treatment • Most frequent TRAEs with N+V, all grade 1–2: asthenia 38%, hypothyroidism 25%, pruritus 19% Prof. Andrea Necchi

[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 [ESMO Congress Madrid 2026 logo] [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 [ESMO Congress Madrid 2026 logo] [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 [ESMO Congress Madrid 2026 logo]

[Slide 1] Study Design & Patients Study Design • Type: Multicenter, single-arm Phase II (Simon's two-stage) • Regimen: Toripalimab (240mg d1) + TIP chemo q3w (up to 4 cycles pre-op) • Primary endpoint: Objective response rate (ORR) • Maintenance: Toripalimab monotherapy (up to 1 year post-op) Patient Characteristics (N=25) • Status: Systematic treatment-naïve • ECOG PS: 0-2 • Stage: cT4 (1pt) or cN2 (6pts) or cN3 (18pts) • HPV positive status: 12pts • PD-L1 positivity (CPS≥1): 23pts • PD-L1 high expression (CPS≥20): 15pts The trial is registered at ClinicalTrials.gov: NCT06415318. [ESMO Congress Madrid 2026 logo]

[Slide 1] 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 Allocation Baseline questionnaires: Baseline 91 received / 94 expected (96.8%) | 91 received / 95 expected (95.8%) On-treatment questionnaire completion rates: Lu-PSMA-617: 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, Numerical Score 0.0–120.0; Lu-P vs Doce; Total, TOI, FACT-G, PWB, FWB, PCS, EWB, SWB] Presenter: Dr. Daniel J. Khalaf [ESMO Congress Madrid 2026 logo] [Slide 3] 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 % patients 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 [ESMO Congress Madrid 2026 logo]

[Slide 1] BACKGROUND: • CH arises from fitness-enhancing hematopoietic mutations, occasionally involving HRR genes, which may lead to false-positive results in liquid biopsies, leading to PARPi treatment in patients who do not harbor tumor-derived HRR mutations. • In addition, CH alterations may be linked to increased risk of hematological toxicities after therapy, such as myelodysplastic syndrome and acute myeloid leukemia (MDS/AML). • An algorithm, called "variant origin prediction" (VOP), leveraging cell-free DNA fragmentomics can help infer CH-derived variants without relying on white blood cell sequencing. • The VOP algorithm achieves 99%, 94%, 94% sensitivity and 100%, 94%, 95% specificity in classifying germline, tumor-somatic and CH-derived short variants, respectively, at the >0.5 threshold. AIMS: (1) Describe the prevalence of mutations predicted of CH origin versus germline/tumor origin in HRR genes (2) Explore the association between CH-predicted vs. germline/tumor HRR genes and outcomes to PARPi (3) Assess changes in CH prevalence following PARPi exposure [Diagram] VARIANT ORIGIN PREDICTION (VOP) Instance-Specific Features: Sample features; Variant features; Fragment features Dataset-Learned Features: Origin propensity -> ML Based Algorithm -> Variant Origin Probabilities: Probability tumor-somatic + Probability germline + Probability clonal hematopoiesis = 100% (Probability tumor) Brown DW, et al. J Liq Biopsy 2025; 9: 100311. Emmanuel S. Antonarakis, MD [ESMO Congress Madrid 2026 logo; inset speaker video with 'MADRID SPAIN 23-27 OCTOBER 2026' podium] [Slide 2] METHODS • We included patients with prostate cancer and FoundationOne® Liquid CDx (F1LCDx) comprehensive genomic profiling. • Clinical data were obtained from the US-wide deidentified Flatiron Health-Foundation Medicine clinico-genomic database, from ~280 cancer clinics (~800 sites of care), between 1/2014 - 9/2025. • Time-to-next treatment (TTNT) and real-world overall survival (rwOS) were compared by Cox models between mCRPC pts receiving PARPi monotherapy with CH-predicted vs. germline/tumor HRR alterations in BRCA1/2, PALB2, RAD51C/D, BARD1, adjusting for clinical covariates. • CH mutation prevalence in common CH-associated genes was compared pre- vs post- PARPi exposure. [Flow diagram] 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 Emmanuel S. Antonarakis, MD [ESMO Congress Madrid 2026 logo; inset speaker video with 'MADRID SPAIN 23-27 OCTOBER 2026' podium] [Slide 3] VARIANT ORIGIN PREDICTION IN HRR GENES [Left bar chart] HRR gene alteration prevalence Total prostate cancer patients, N = 18691 Prevalence among prostate cancer patients: ATM 13.5% (n=2523); CHEK2 10.3% (n=1922); BRCA2 6.4% (n=1193); CDK12 3.7% (n=685); BRCA1 1.6% (n=298); 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) [Right stacked bar chart] Variant origin prediction by gene (mutations only) Patients with an HRR gene mutation, N = 5632 (Germline / Tumor-somatic / CH-predicted only, % as printed) ATM (n=2423): 9% / 15% / 76% CHEK2 (n=1905): 8% / 3% / 87% 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=35): 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 • In prostate cancer patients with detected HRR mutations (n=5,632), 56% had only CH-predicted HRR mutations, including BRCA1 (32%) and PALB2 (24%). In BRCA2, 7% were CH-derived. Emmanuel S. Antonarakis, MD [ESMO Congress Madrid 2026 logo; inset speaker video with 'MADRID SPAIN 23-27 OCTOBER 2026' podium]