ESMO 2026 Breast Cancer Slides

TROPION-Breast02: Study Design Randomised, phase 3, open-label, global study (NCT05374512) Key inclusion criteria: • Patients with histologically or cytologically documented locally recurrent inoperable or metastatic TNBC* • No prior chemotherapy or targeted systemic therapy in the locally recurrent inoperable or metastatic setting • Immunotherapy not an option† • ECOG PS 0 or 1 • No minimum DFI‡ 1:1 Dato-DXd 6 mg/kg IV Day 1 Q3W (n=323) Investigator's choice of chemotherapy (ICC)# Paclitaxel, nab-paclitaxel, capecitabine, eribulin mesylate/eribulin, carboplatin (n=321) Endpoints Dual primary: • OS • PFS by BICR per RECIST v1.1 Secondary included: • PFS (investigator-assessed) • ORR, DoR • Safety Randomisation stratified by: • Geographic region (US/Canada/Europe vs other geographic regions) • PD-L1 status (high [CPS ≥10] vs low [CPS <10])§ • DFI history (de novo vs prior DFI 0–12 months vs prior DFI >12 months)¶ • Treatment continued until investigator-assessed RECIST v1.1 progressive disease, unacceptable toxicity, or another discontinuation criterion was met • Following progression or discontinuation of study treatment, patients could receive subsequent therapies, including approved ADCs or chemotherapy, at the investigator's discretion‖ *According to ASCO/CAP criteria. †Including patients with PD-L1-low tumours, or patients with PD-L1-high tumours with (a) disease relapse after prior PD-(L)1 inhibitor therapy for early-stage breast cancer, (b) comorbidities precluding PD-(L)1 inhibitor therapy, or (c) no regulatory access to PD-(L)1 inhibitor therapy. ‡DFI defined as time between date of completion of treatment with curative intent and date of first documented local or distant disease recurrence. §Recruitment of patients with PD-L1-high tumours who would otherwise be eligible for pembrolizumab if regulatory access was available was capped at ~10% of randomised patients. ¶Recruitment of patients with DFI 0–12 months was capped at ~20% of randomised patients. #If no prior taxane, or prior taxane in the (neo)adjuvant setting and DFI >12 months: paclitaxel 80 mg/m2 IV, D1, 8, 15, Q3W, or nab-paclitaxel 100 mg/m2 IV, D1, 8, 15, Q4W; if prior taxane and DFI 0–12 months: capecitabine 1000 or 1250 mg/m2 orally twice daily, D1–14, Q3W (dose determined by standard institutional practice), or eribulin mesylate 1.4 mg/m2 / eribulin 1.23 mg/m2 IV, Day 1, 8, Q3W, or carboplatin AUC6 IV, D1, Q3W. ‖In the Dato-DXd vs ICC arm, 65% vs 72% of patients received any subsequent therapy in any treatment line; 14% vs 30% received a subsequent ADC (sacituzumab govitecan, sacituzumab tirumotecan, trastuzumab deruxtecan). Rebecca A. Dent

[Slide 1] Proffered paper 2: Breast cancer, early and advanced stage CHAIRS: SARA HURVITZ, SEOCK-AH IM CNS PFS in Patients with Brain Metastases | | Dato-DXd n=35* | ICC n=26* | | CNS PFS events, n (%)† | 18 (51.4) | 16 (61.5) | | Median CNS PFS, months (95% CI) | 9.5 (6.9-19.3) | 2.6 (1.5-5.8) | | HR (95% CI) | 0.27 (0.13-0.56) | Y axis: Probability of PFS; X axis: Time from randomisation (months) Number at risk Dato-DXd: 35, 30, 20, 13, 8, 7, 3, 2, 2, 2, 1, 1, 0 ICC: 26, 6, 2, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0 (at 0, 3, 6, 9, 12, 15, 18, 21, 24, 27, 30, 33, 36 months) Dato-DXd demonstrated CNS PFS benefit compared with ICC, with an HR of 0.27 *Analysis included patients with any brain scan at baseline; †CNS PFS was defined as the time from randomisation to intracranial progression (determined by CNS BICR per CNS-modified RECIST v1.1) or death. Patients without an event were censored at their last evaluable CNS-modified RECIST assessment; events occurring after ≥2 consecutive missed assessments were censored at the last evaluable assessment before the event. CNS PFS events listed only include events that occurred within 2 assessments of the last evaluable assessment (CNS-modified RECIST progression: Dato-DXd: n=15 [42.9%], ICC: n=15 [57.7%] and death in the absence of progression: Dato-DXd: n=3 [8.6%], ICC n=1 [3.8%]). Rebecca Dent - First-line (1L) datopotamab deruxtecan (Dato-DXd) vs chemotherapy in patients (pts) with locally recurrent inoperable or metastatic triple-negative breast cancer (mTNBC) for whom immunotherapy was not an option: Exploratory analysis of pts with brain m... Survival table Endpoint: CNS PFS (CNS BICR per CNS-modified RECIST v1.1) in patients with brain metastases Arm | N | Median | Median CI Dato-DXd | 35 | 9.5 months | 6.9-19.3 ICC | 26 | 2.6 months | 1.5-5.8 HR: 0.27 HR CI: 0.13-0.56 [Slide 2] Systemic Efficacy in Patients With Brain Metastases In patients with or without prior CNS-directed therapy | Dato-DXd | ICC PFS by BICR* (median months): - Patients with BMs (n=61): 7.7 vs 3.5; HR 0.37 (0.20-0.72) - Previously treated BMs (n=31): 6.8 vs 3.5; HR 0.43 (0.17-1.11) - Previously untreated BMs (n=30): 7.7 vs 2.6†; HR 0.29 (0.12-0.76) OS (median months): - Patients with BMs: 13.5† vs 10.7; HR 0.55 (0.29-1.05) - Previously treated BMs: 19.3† vs 12.8†; HR 0.77 (0.28-2.08) - Previously untreated BMs: 11.6† vs 9.1; HR 0.35 (0.14-0.84) ORR by BICR* (%): - Patients with BMs: 62.9 vs 19.2; OR 7.11 (2.16-23.41) - Previously treated BMs: 62.5 vs 33.3; NA‡ - Previously untreated BMs: 63.2 vs 0; NA‡ Dato-DXd demonstrated consistent clinical efficacy benefit vs ICC across PFS, OS and ORR in patients with brain metastases HR 95% CIs are shown in brackets; error bars represent 95% CIs for median values. Data for patients without BMs (n=583): median PFS 11.1 vs 5.6 months (HR 0.56 [0.46-0.69]); median OS 23.7 vs 19.3 months (HR 0.80 [0.64-1.00]); ORR 62.5% vs 30.2% (OR 3.86 [2.73-5.44]). *Per RECIST v1.1. †Upper 95% CI not calculable. ‡ORs are not available for subgroups with less than 20 events across both treatment arms. [Slide 3] Duration of Intracranial Response DoR by CNS BICR per CNS-modified RECIST v1.1 in evaluable disease* Dato-DXd (n=35*) | ICC (n=22*) [swimmer plot; ICC rows labeled N, N, P] Legend: CR; PR; Response start; Response end; Response censored date; Without prior CNS-directed therapy; Type of chemotherapy: P = Paclitaxel, N = Nab-paclitaxel Time from randomisation (weeks) 0-165 Intracranial responses with Dato-DXd were durable, with the longest DoR observed up to 117 weeks (~27 months) *Evaluable disease includes patients who had baseline and post-baseline brain scans. Intracranial response in TLs was classified as CR, PR, SD, PD, or NE, and in NTLs was classified as CR, non-CR/non-PD, PD, or NE. NTLs could not be assessed for PR or SD. Among patients with evaluable disease, intracranial confirmed ORR was 40.0% (5 CR, 9 PR) with Dato-DXd (n=35) and 13.6% (0 CR, 3 PR) with ICC (n=22); intracranial DCR at 12 weeks was 85.7% and 27.3%, respectively. DoR was defined as the time from the first confirmed CNS CR or PR to progression or death. Patients without an event were censored at their last evaluable CNS-modified RECIST v1.1 assessment; patients with events occurring after ≥2 consecutive missed assessments were censored at the last evaluable assessment before the event. Grey lines represent duration of study participation for each patient; arrows represent patients ongoing in the study at the time of DCO. [last line partially cut off] [Slide 4] Intracranial Response in Patients with Measurable Disease | Measurable disease* | Dato-DXd (n=16) | ICC (n=8) | | Intracranial confirmed ORR,† % | 62.5 | 37.5 | | Intracranial DCR at 12 weeks, % | 75.0 | 50.0 | Best percentage change from baseline in measurable disease [waterfall plots: Dato-DXd (n=16); ICC (n=8)] Type of chemotherapy: P = Paclitaxel, N = Nab-paclitaxel, E = Eribulin mesylate Confirmed best objective response: CR, PR, SD, PD; ◆ Patients without prior CNS-directed therapy In patients with measurable disease, Dato-DXd demonstrated a confirmed intracranial ORR of 62.5% vs 37.5% with ICC *Measurable disease defined as patients with TLs at baseline. TLs defined as: contrast enhancing, suitable for accurate repeated measurements, and ≥10 mm in ≥1 dimension with CT or MRI or no less than double the slice thickness, whichever was largest. Intracranial disease that did not qualify as a TL or any surplus in TLs was classified as intracranial NTLs. †Intracranial confirmed ORR defined as the percentage of patients with ≥2 visits with a CNS response of PR or CR by CNS BICR per CNS-modified RECIST v1.1. Dashed lines represent thresholds for PR (-30%) and PD (20%).

CNS PFS in Patients with Brain Metastases Dato-DXd n=35* | ICC n=26* CNS PFS events, n (%)† | 18 (51.4) | 16 (61.5) Median CNS PFS, months (95% CI) | 9.5 (6.9–19.3) | 2.6 (1.5–5.8) HR (95% CI) | 0.27 (0.13–0.56) Number at risk (months 0,3,6,...,36) Dato-DXd 35 30 20 13 8 7 3 2 2 2 1 1 0 ICC 26 6 2 1 1 0 0 0 0 0 0 0 0 Dato-DXd demonstrated CNS PFS benefit compared with ICC, with an HR of 0.27 *Analysis included patients with any brain scan at baseline; †CNS PFS was defined as the time from randomisation to intracranial progression (determined by CNS BICR per CNS-modified RECIST v1.1) or death. Patients without an event were censored at their last evaluable CNS-modified RECIST assessment; events occurring after ≥2 consecutive missed assessments were censored at the last evaluable assessment before the event. CNS PFS events listed only include events that occurred within 2 [remainder of footnote cut off / illegible] Survival table Endpoint: CNS PFS (patients with brain metastases) Arm | N | Median | Median CI Dato-DXd | 35 | 9.5 months | 6.9–19.3 ICC | 26 | 2.6 months | 1.5–5.8 HR: 0.27 HR CI: 0.13–0.56

[Slide 1] 08:30 - 10:00 Proffered paper 2: Breast cancer, early and advanced stage CHAIRS: SARA HURVITZ, SEOCK-AH IM LBA 96: TROPION-Breast02 Results | Outcome | Dato-DXd (n=35) | Chemotherapy (n=26) | Effect | | Systemic ORR | 62.9% | 19.2% | OR 7.11 | | Intracranial ORR (measurable disease) | 62.5% (n=16) | 37.5% (n=8) | | | Intracranial ORR (evaluable disease) | 40.0% | 13.6% | | | Median CNS PFS | 9.5 mos | 2.6 mos | HR 0.27 | - Brain is sanctuary site—up to 6% of pts with early TNBC have brain as first/only site of recurrence - Prior to ADCs, mOS after brain mets 3-7 mos Dawood et al, Annals of Oncology 2009; Dawood et al, Cancer 2012 Rita Nanda, MD Rita Nanda - Invited Discussant LBA26, LBA96 and LBA15 ALICANTE AUDITORIUM - HALL 6 [Slide 3] 08:30 - 10:00 Proffered paper 2: Breast cancer, early and advanced stage CHAIRS: SARA HURVITZ, SEOCK-AH IM HEALER Results - N=885 pts at 76 centers across China - mPFS for HB1801+ anbenitamab vs THP was NR vs 14.3 mos; HR=0.46 (0.37-0.56) - ORR 87.6% vs 73.4% - Increased toxicity with investigational arm (cytopenias) - Grade 3 TEAEs: 45.9% vs 34.4% - Dose reductions: 11.0% vs 3.9% - SAEs: 20.0% vs 16.2% Rita Nanda, MD ALICANTE AUDITORIUM - HALL 6 [Slide 4] 08:30 - 10:00 Proffered paper 2: Breast cancer, early and advanced stage CHAIRS: SARA HURVITZ, SEOCK-AH IM HEALER: Strengths and Caveats - Strengths - Large randomized trial, central HER2 confirmation, BICR primary endpoint, large effect size - Clear PFS advantage over THP (SOC at time of study start) - Caveats - Relevant comparator now T-DXd + P not THP (mPFS 40.7 vs 26.9 mos) - Very few pts had prior H (19%) or P (7%); most HER2-directed tx naive - Two variables changed at once (docetaxel->albumin-bound; HP->bispecific) - Only enrolled in China, agents with limited/no availability in other countries - Role - SOC has evolved since start of study—both in the frontline and maintenance settings - Efficacy after T-DXd+P? CNS activity? - Clearly better than THP; consider if T-DXd+P unavailable/contraindicated Rita Nanda, MD ALICANTE AUDITORIUM - HALL 6

[Slide 1] [Session screen banner] 08:30 - 10:00 Proffered paper 2: Breast cancer, early and advanced stage CHAIRS: SARA HURVITZ, SEOCK-AH IM Intracranial Response in Patients with Measurable Disease Measurable disease* | | Dato-DXd (n=16) | ICC (n=8) | | Intracranial confirmed ORR,† % | 62.5 | 37.5 | | Intracranial DCR at 12 weeks, % | 75.0 | 50.0 | Best percentage change from baseline in measurable disease Dato-DXd (n=16) [waterfall plot; y-axis: Best % change in sum of diameters from baseline] ICC (n=8) [waterfall plot; bars labeled by chemotherapy type: N, P, E, P, E, P, N, N] Type of chemotherapy P = Paclitaxel N = Nab-paclitaxel E = Eribulin mesylate Confirmed best objective response: CR PR SD PD ◆ Patients without prior CNS-directed therapy In patients with measurable disease, Dato-DXd demonstrated a confirmed intracranial ORR of 62.5% vs 37.5% with ICC *Measurable disease defined as patients with TLs at baseline. TLs defined as: contrast enhancing, suitable for accurate repeated measurements, and ≥10 mm in ≥1 dimension with CT or MRI or no less than double the slide thickness, whichever was largest. [I]ntracranial disease that did not qualify as a TL or any surplus in TLs was classified as intracranial NTLs. †Intracranial confirmed ORR defined as the percentage of patients with ≥2 visits with a CNS response of PR or CR by CNS BICR per CNS-modified [R]ECIST v1.1. Dashed lines represent thresholds for PR (-30%) and PD (20%). [R]ebecca Dent CR, complete response; DCR, disease control rate; NTL, non-target lesion; PD, progressive disease; PR, partial response; SD, stable disease; TL, target lesion. [Banner] ALICANTE AUDITORIUM - HALL 6 [Slide 2] [Session screen banner] 08:30 - 10:00 Proffered paper 2: Breast cancer, early and advanced stage CHAIRS: SARA HURVITZ, SEOCK-AH IM Duration of Intracranial Response DoR by CNS BICR per CNS-modified RECIST v1.1 in evaluable disease* [Swimmer plot] Dato-DXd (n=35*) ; ICC (n=22*) [ICC rows labeled N, N, P] Legend: CR; PR; ▲ Response start; ● Response end; ■ Response censored date; ◆ Without prior CNS-directed therapy Type of chemotherapy: P = Paclitaxel; N = Nab-paclitaxel Time from randomisation (weeks) Intracranial responses with Dato-DXd were durable, with the longest DoR observed up to 117 weeks (~27 months) *Evaluable disease includes patients who had baseline and post-baseline brain scans. Intracranial response in TLs was classified as CR, PR, SD, PD, or NE, and in NTLs was classified as CR, non-CR/non-PD, PD, or NE. NTLs could not be assessed for PR or SD. Among patients with evaluable disease, intracranial confirmed ORR was 40.0% (5 CR, 9 PR) with Dato-DXd (n=35) and 13.6% (0 CR, 3 PR) with ICC (n=22); intracranial DCR at 12 weeks was 85.7% and 27.3%, respectively. DoR was defined as the time from the first confirmed CNS CR or PR to progression or death. Patients without an event were censored at their last evaluable CNS-modified RECIST v1.1 assessment; patients with events occurring after ≥2 consecutive missed assessments were censored at the last evaluable assessment before the event. Grey lines represent duration of study participation for each patient; arrows represent patients ongoing in the study at the time of DCO. Rebecca Dent DCO, data cutoff; DoR, duration of response; NE, not evaluable. [Banner] ALICANTE AUDITORIUM - HALL 6 [Slide 4] [Session screen banner] 08:30 - 10:00 Proffered paper 2: Breast cancer, early and advanced stage CHAIRS: SARA HURVITZ, SEOCK-AH IM Conclusions In a subset of patients with metastatic TNBC and asymptomatic brain metastases, with or without prior CNS-directed therapy: • Dato-DXd demonstrated systemic efficacy benefit across PFS, OS and ORR compared with ICC • Dato-DXd demonstrated higher intracranial ORR, greater DCR, longer DoR and CNS PFS benefit vs ICC [Callout box with person icon] TROPION-Breast02 enrolled patients without prior CNS-directed therapy, who are often excluded from clinical trials In TROPION-Breast02, 1L Dato-DXd provided efficacy in patients with metastatic TNBC and brain metastases, regardless of prior CNS-directed therapy Rebecca Dent [Banner] ALICANTE AUDITORIUM - HALL 6

Systemic Efficacy in Patients With Brain Metastases In patients with or without prior CNS-directed therapy Dato-DXd | ICC PFS by BICR* (median, months) Patients with BMs (n=61): 7.7 vs 3.5; HR 0.37 (0.20–0.72) Previously treated BMs (n=31): 6.8 vs 3.5; HR 0.43 (0.17–1.11) Previously untreated BMs (n=30): 7.7 vs 2.6†; HR 0.29 (0.12–0.76) OS (median, months) Patients with BMs: 13.5† vs 10.7; HR 0.55 (0.29–1.05) Previously treated BMs: 19.3† vs 12.8†; HR 0.77 (0.28–2.08) Previously untreated BMs: 11.6† vs 9.1; HR 0.35 (0.14–0.84) ORR by BICR* (%) Patients with BMs: 62.9 vs 19.2; OR 7.11 (2.16–23.41) Previously treated BMs: 62.5 vs 33.3; NA‡ Previously untreated BMs: 63.2 vs 0; NA‡ Dato-DXd demonstrated consistent clinical efficacy benefit vs ICC across PFS, OS and ORR in patients with brain metastases HR 95% CIs are shown in brackets; error bars represent 95% CIs for median values. Data for patients without BMs (n=583): median PFS 11.1 vs 5.6 months (HR 0.56 [0.46–0.69]); median OS 23.7 vs 19.3 months (HR 0.80 [0.64–1.00]); ORR 62.5% vs 30.2% (OR 3.86 [2.73–5.44]). *Per RECIST v1.1; †Upper 95% CI not calculable; ‡ORs are not available for subgroups with less than 20 events across both treatment arms. Rebecca Dent BMs, brain metastases; OR, odds ratio.

TROPION-Breast02 Study Design1,2 Key inclusion criteria: • Patients with previously untreated locally recurrent inoperable or metastatic TNBC • Immunotherapy not an option • No minimum DFI • Patients with asymptomatic treated or untreated brain metastases at baseline were eligible 1:1 Dato-DXd* 6 mg/kg IV Day 1 Q3W (n=323) Investigator's choice of chemotherapy (ICC)† (n=321) Endpoints Dual primary: PFS by BICR per RECIST v1.1, and OS Exploratory analysis in patients with previously treated and untreated brain metastases: • PFS by BICR, OS, confirmed systemic ORR by BICR • Intracranial response and PFS Randomisation stratified by: • Geographic region (US/Canada/Europe vs other geographic regions) • PD-L1 status (high [CPS ≥10] vs low [CPS <10]) • DFI history (de novo disease vs prior DFI 0–12 months vs prior DFI >12 months) TROPION-Breast02 included patients with asymptomatic brain metastases at baseline, including both those with or without prior CNS-directed therapy *Treatment continued until investigator-assessed RECIST v1.1 progressive disease, unacceptable toxicity, or another discontinuation criterion was met. †Paclitaxel, nab-paclitaxel, capecitabine, eribulin mesylate/eribulin, or carboplatin. Rebecca Dent 1. Dent R, et al. Oral presentation at ESMO 2025 (Abstract LBA21). 2. Dent R, et al. Ann Oncol 2026;37:1066–80.

[Slide 3] CNS PFS in Patients with Brain Metastases Dato-DXd n=35* | ICC n=26* CNS PFS events, n (%)† 18 (51.4) | 16 (61.5) Median CNS PFS, months (95% CI) 9.5 (6.9-19.3) | 2.6 (1.5-5.8) HR (95% CI) 0.27 (0.13-0.56) [KM: Probability of PFS vs Time from randomisation (months) 0-36] Number at risk Dato-DXd 35 30 20 13 8 7 3 2 2 2 1 1 0 ICC 26 6 2 1 1 0 0 0 0 0 0 0 0 Dato-DXd demonstrated CNS PFS benefit compared with ICC, with an HR of 0.27 *Analysis included patients with any brain scan at baseline; †CNS PFS was defined as the time from randomisation to intracranial progression (determined by CNS BICR per CNS-modified RECIST v1.1) or death. Patients without an event were censored at their last evaluable CNS-modified RECIST assessment; events occurring after >=2 consecutive missed assessments were censored at the last evaluable assessment before the event. CNS PFS events listed only include events that occurred within 2 assessments of the last evaluable assessment (CNS-modified RECIST progression: Dato-DXd: n=15 [42.9%], ICC: n=15 [57.7%] and death in the absence of progression: Dato-DXd: n=3 [8.6%], ICC n=1 [3.8%]). Rebecca Dent Survival table Endpoint: CNS PFS (patients with brain metastases, any baseline brain scan) Arm | N | Median | Median CI Dato-DXd | 35 | 9.5 months | 6.9-19.3 ICC | 26 | 2.6 months | 1.5-5.8 HR: 0.27 HR CI: 0.13-0.56 (95% CI) [Slide 4] Conclusions In a subset of patients with metastatic TNBC and asymptomatic brain metastases, with or without prior CNS-directed therapy: - Dato-DXd demonstrated systemic efficacy benefit across PFS, OS and ORR compared with ICC - Dato-DXd demonstrated higher intracranial ORR, greater DCR, longer DoR and CNS PFS benefit vs ICC TROPION-Breast02 enrolled patients without prior CNS-directed therapy, who are often excluded from clinical trials [person icon] In TROPION-Breast02, 1L Dato-DXd provided efficacy in patients with metastatic TNBC and brain metastases, regardless of prior CNS-directed therapy Rebecca Dent

Demographics and Baseline Characteristics Columns: Patients with baseline brain metastases: Dato-DXd (n=35) | ICC (n=26) || Patients without baseline brain metastases: Dato-DXd (n=288) | ICC (n=295) Median age (range), years: 62 (37–79) | 55 (26–77) | 55 (27–85) | 57 (23–83) Female, n (%): 35 (100) | 26 (100) | 288 (100) | 293 (99) Race, n (%) Black or African American: 0 | 0 | 13 (5) | 14 (5) Asian: 18 (51) | 11 (42) | 133 (46) | 120 (41) White: 15 (43) | 14 (54) | 116 (40) | 139 (47) Other: 2 (6)* | 1 (4)* | 26 (9)† | 22 (7)‡ Geographic region, n (%) US, Canada, Europe: 12 (34) | 7 (27) | 108 (38) | 113 (38) Other geographic regions: 23 (66) | 19 (73) | 180 (63) | 182 (62) DFI history, n (%) De novo: 8 (23) | 5 (19) | 101 (35) | 105 (36) Prior DFI 0–12 months: 6 (17) | 8 (31) | 61 (21) | 58 (20) Prior DFI 0–6 months: 4 (11) | 6 (23) | 42 (15) | 38 (13) Prior DFI >12 months: 21 (60) | 13 (50) | 126 (44) | 132 (45) Number of metastatic sites, n (%) <3: 8 (23) | 5 (19) | 199 (69) | 210 (71) ≥3: 27 (77) | 21 (81) | 89 (31) | 85 (29) Previous cancer therapy, n (%): 28 (80) | 21 (81) | 187 (65) | 186 (63) Prior CNS-directed therapy, n (%) Radiotherapy alone: 13 (37) | 15 (58) | NA | NA Radiotherapy + surgery: 3 (9) | 0 PD-L1 status, n (%)§ Low (CPS <10): 33 (94) | 22 (85) | 254 (88) | 269 (91) High (CPS ≥10): 2 (6) | 4 (15) | 32 (11) | 25 (9) *Including not reported (Dato-DXd: n=2, ICC: n=1). †Includes Native Hawaiian or Other Pacific Islander and Asian (n=1), White and Brown (n=1), other (n=5), and not reported (n=19). ‡Includes other (n=6) and not reported (n=16). §Based on central laboratory testing, using Agilent PD-L1 IHC 22C3 pharmDx Assay (Agilent Technologies, Santa Clara, CA). Among patients without baseline brain metastases, PD-L1 status was missing/NA in two patients in the Dato-DXd arm and one patient in the ICC arm. Rebecca Dent IHC, immunohistochemistry; NA, not applicable.

[Slide 1] TUXEDO-3 Key Questions 1. What is the clinical need for CNS-active therapies in breast and lung cancer? 2. How can ADCs exert an effect in the CNS? 3. How do the results compare to other ADCs? 4. What is on the horizon? 5. What are implications for clinical trial design? [Slide 2] CNS Outcomes to ADCs in Breast Cancer TDM1 | T-DXd | Dato-DXd | HER3-DXd KAMILLA (HER2+ BC)1: -CNS ORR 32.7% in prev tx'd BM; -CNS ORR 49.3% in untreated BM DESTINY-Breast12 (HER2+ BC)2: -CNS ORR 71% (82% in untreated BM); -12-mo OS 90.3% TUXEDO-1/DEBBRAH/US Case Series – (HER2+ BC) Pooled Analysis3: -CNS ORR 64.9%, active BM; -Median OS 22.5 mo DEBBRAH (HER2-low BC)4: -CNS ORR 33-50% DESTINY-Breast04 (HER2-low BC)5: -CNS ORR 25% TUXEDO-2 (TNBC)6: -CNS ORR 37.5% at interim DATO-BASE (HER2-neg BC)7: -30% LANO response in LMD TROPION-Breast02 (1L TNBC)8: -CNS ORR 62.5%, mCNS-PFS 9.5mo TUXEDO-3 (BC, All subtypes)9: -CNS ORR 23.8%; -Median OS 7.9 mo 1Montemurro et al, Ann Oncol 2020; 2Harbeck et al, Nat Med 2024; 3Bartsch et al, ESMO Open 2025; 4Batista et al, ESMO Open 2024; 5Tsurutani et al, ESMO 2024; 6Bartsch et al, ESMO 2024; 7Tarantino et al, [cut off] [Slide 3] IT cDC1 Therapy Key Questions 1. How does this compare to outcomes with usual care? 2. Is this just patient selection / quality of patient care? 3. What is the best way to integrate into care given rapid trajectory of LMD? 4. How can we get patients to trials, and trials to patients? [Slide 4] Summary TUXEDO-3 • Clinically meaningful intracranial activity of Patritumab-DXd • Breast Cancer: IC-ORR 23.8%, mOS 7.9 mo • NSCLC: IC-ORR 30.0%, mOS 6.7 mo • LMD: mOS 10.5 mo Intrathecal DC1 Therapy • Very novel approach • Evidence of immune modulation in CSF • Intriguing signal of extended OS beyond historical controls • Excited to see the next trials of this approach Management of CNS disease is key component of overall disease control – continued research in this space remains greatly needed.

[Slide 1] 14:45 - 16:15 Key challenges in managing patients with metastatic breast cancer CHAIRS: RUPERT BARTSCH, EVA CIRUELOS TNBC molecular subtypes A PAM50 | B Lehmann et al (2011) | C Curtis et al (2012) | D Lehmann et al (2016) | E Burstein et al (2016) | F Jiang et al (2019) Fig. 1 Triple-negative breast cancer molecular subtypes across studies. Transcriptomic-based and gene expression-based subtypes of triple-negative breast cancer (TNBC) according to PAM50 (a) and defined by Lehmann et al. (b and d), Curtis et al. (c), Burstein et al. (e), and Jiang et al. (f). BLIA basal-like immune-activated, BLIS basal-like immunosuppressed, BL1 basal-like 1, BL2 basal-like 2, IM immunomodulatory, IntClust integrative clusters, LAR luminal androgen receptor, M mesenchymal, MES mesenchymal, MSL mesenchymal stem-like, UNS unstable. Figures was generated by reanalysis of publicly available studies and open-source platforms (cBioPortal: https://www.cbioportal.org). Marra et al, 2020 Biomarker-driven therapeutic approaches - Outline Luminal androgen receptor: • Anti-androgen therapy • Enrichment in PIK3CA mutations • Potential combinations of anti-androgens with CDK 4/6 and PIK3CA inhibitors Immune-enriched: • ↑ TILs • ↑ PD-L1 • Enrichment in immune-response genes • Immune checkpoint inhibitors PI3K/AKT/mTOR activated: • PIK3CA mutations • PTEN loss • Selective PIK3CA and AKT inhibitors • Potential combinations with ICIs DNA repair deficiency: • BRCA 1 and 2 mutations and BRCAness • HRD scores • Platinum compounds • PARP inhibitors Surface antigens: • LIV-1 • Trop-2 • GPNMB • HER2 (HER2-low) MAPK pathway: • BRAF and MEK inhibitors • BET inhibitors • Potential combinations with ICIs Eva M Ciruelos, MD, PhD Eva Ciruelos - Turning the corner: Personalized therapy for triple-negative breast cancer becomes a reality [Slide 2] First-line TROP2 ADCs in mTNBC: ASCENT-03 vs TROPION-Breast02 | Sacituzumab govitecan · ASCENT-03 [1] 10 mg/kg D1,8 q21d vs chemotherapy | Datopotamab deruxtecan · TROPION-Breast02 [2] 6 mg/kg D1 q21d vs chemotherapy Patients | N=558 · PD-(L)1 inhibitor not an option | N=644 · immunotherapy not an option Disease-free interval | ≥6 months required | No minimum: ~15% relapsed within 6 months Control arm | Taxane (55%) or gemcitabine-carboplatin (45%) | Single-agent investigator's choice (taxane 82%) Primary endpoint | PFS (BICR) | Dual: PFS (BICR) and OS Median PFS | 9.7 vs 6.9 mo HR 0.62 (95% CI 0.50–0.77) | 10.8 vs 5.6 mo HR 0.57 (95% CI 0.47–0.69) Median OS | Immature HR 0.98; 82% crossover to SG | 23.7 vs 18.7 mo HR 0.79 (95% CI 0.64–0.98) ORR · DoR | 48% vs 46% DoR 12.2 vs 7.2 mo | 62.5% vs 29.3% DoR 12.3 vs 7.1 mo Signature toxicity | Neutropenia (G≥3 43%), diarrhoea | Stomatitis (57%), ocular events; ILD uncommon Stopped for AEs | 4% vs 12% with chemotherapy | 4% vs 7% with chemotherapy After progression | 82% of next-line therapy in chemo arm was SG | Later ADC: 41% (chemo arm) vs 21% (Dato-DXd arm) BothADCs superior to chemotherapy on PFS; Dato-DXd has shown an OS benefit so far. Cross-trial view: eligibility, controls and post-progression therapy differed. [1] Cortés J et al. N Engl J Med 2025 (ASCENT-03; ESMO 2025 LBA20). [2] TROPION-Breast02: Ann Oncol 2026; Dent R et al. ESMO 2025 LBA21. BICR, blinded independent central review; DFI, disease-free interval; DoR, duration of response; G, grade; IO, immunotherapy. Eva M Ciruelos, MD, PhD [Slide 3] Previously untreated advanced TNBC: ongoing immunotherapy-ADC combination trials Trial | Phase | Population | Design | Endpoint | Results TROPION-Breast05 NCT06103864 | Phase III n 625 | PD-L1>10 inoperable / mTNBC; 1L | Dato-DXd + durvalumab vs Chemo + Pembro | 1º: PFS (BICR) 2º: OS, ORR, safety | Ongoing BEGONIA1 NCT03742102 | Phase Ib/II platform, ongoing | Unresectable LA / mTNBC; 1L cohorts (arms 7 & 8) | Durvalumab + novel partners Dato-DXd, T-DXd, others | ORR, safety | ORR 79% mDOR 17.6m mPFS 14m Saci-IO TNBC NCT04468061 | Phase II randomized | PD-L1–negative 1L mTNBC | Sacituzumab govitecan +/- Pembro | PFS | Ongoing MORPHEUS-panBC2 NCT03424005 | Phase Ib/II umbrella | Metastatic 1L m TNBC TNBC cohorts PDL1 Ventana ≥1% (n31) | Atezolizumab + sacituzumab govitecan ± other novel combinations | ORR in multiple cohort controls | TNBC: ORR 76.7% mDOR 14 m mPFS 12.2m TROFUSE0113 NCT06841354 | Phase III open label | Metastatic 1L m TNBC CPS <10 | SacTMT +/- Pembro vs TPC | 1º: PFS (BICR), OS 2º: ORR, safety | Ongoing 1 Schmid P, Ann Oncol 2026; 2 Schmid P Esmo Open 2024; 3 Rugo H, Future Oncol 2026; ClinicalTrials.gov Eva M Ciruelos, MD, PhD [Slide 4] Future of PARP inhibition in TNBC Functional HRD • Move beyond static genomic scars • RAD51 foci: real-time HR competency • Predict sensitivity and resistance RADIOLA: Study design & translational research (SOLTI) N=65. Key eligibility criteria: • Unresectable locally advanced/ Metastatic HER2-negative BC; • ≤ 2 prior chemotherapy lines in metastatic setting • If HR-positive disease: relapsed/progressed on at least one line of ET±CDK4/6 inhibitor • Measurable disease • No PD to platinum, and if platinum exposure, the biopsy was obtained after treatment Screening: RAD51 test (HRD was defined as ≤10% RAD51 foci-positive cells (RAD51-Low)) N = 55 COHORT 1: Somatic or germline BRCA1/2, PALB2 or RAD51C/D PV and RAD51-evaluable* (*Blinded RAD51 test result) N = 10 COHORT 2: WT or unknown BRCA1/2, PALB2 or RAD51C/D status and RAD51-low tumor Olaparib 300 mg P.O. BID Until PD or toxicity Primary Endpoint: • ORR by RECIST v.1.1 in RAD51-low vs RAD51-high tumors in COHORT 1 Secondary Endpoints: • ORR in COHORT 2 • PFS, CBR, DoR, TtR, in both cohorts • ctDNA drop after 4 weeks of treatment in both cohorts • Clinical benefit of Olaparib according to the different HRR mutated genes in COHORT 1 • Safety Functional RAD51 testing may improve patient selection for PARP inhibitors beyond genomic testing alone Better ORR and PFS in the low-RAD51 subgroup in BRCA1/2 and PALB2 carriers OR (95%CI): 7.90 (2.07 - 39.54) p= 0.002 Observed delta: 46.9% (68.3% - 21.4%); Expected delta: 45.0% (70.0% - 25.0%) Overall response rate: RAD51-LOW HRD (n=41) 68.3% (28/41); RAD51-HIGH HRP (n=14) 21.4% (3/14) PFS | RAD51-low | RAD51-high Median PFS | 7.1 m | 5.6 m HR (95%CI) | 0.51 (0.26 - 0.98) Pimentel I, SABCS 2025 Eva M Ciruelos, MD, PhD

Patient Case • 56-year-old patient with metastatic TNBC with metastases to brain, lung, and distant and regional lymph nodes • Received neoadjuvant chemotherapy with taxane/anthracycline • Previous CNS-directed therapy: No • Onset of intracranial response at 1.5 months, intracranial PD at 19.3 months, DoR 17.8 months • Received 26 cycles of Dato-DXd, discontinued due to objective disease progression, alive at the time of DCO [Brain MRI images, two axial levels] Baseline | 5.5 months Complete intracranial response was observed and sustained for 17.8 months Rebecca Dent

Intracranial Response in Patients with Measurable Disease Measurable disease* Dato-DXd (n=16) ICC (n=8) Intracranial confirmed ORR,† % 62.5 37.5 Intracranial DCR at 12 weeks, % 75.0 50.0 Best percentage change from baseline in measurable disease Dato-DXd (n=16) [waterfall plot] ICC (n=8) [waterfall plot; bars labelled N, P, E, P, E, P, N, N] Type of chemotherapy: P = Paclitaxel; N = Nab-paclitaxel; E = Eribulin mesylate Confirmed best objective response: CR, PR, SD, PD; ◆ Patients without prior CNS-directed therapy In patients with measurable disease, Dato-DXd demonstrated a confirmed intracranial ORR of 62.5% vs 37.5% with ICC *Measurable disease defined as patients with TLs at baseline. TLs defined as: contrast enhancing, suitable for accurate repeated measurements, and ≥10 mm in ≥1 dimension with CT or MRI or no less than double the slide thickness, whichever was largest. Intracranial disease that did not qualify as a TL or any surplus in TLs was classified as intracranial NTLs. †Intracranial confirmed ORR defined as the percentage of patients with ≥2 visits with a CNS response of PR or CR by CNS BICR per CNS-modified RECIST v1.1. Dashed lines represent thresholds for PR (-30%) and PD (20%). Rebecca Dent CR, complete response; DCR, disease control rate; NTL, non-target lesion; PD, progressive disease; PR, partial response; SD, stable disease; TL, target lesion.

Methods and Patient Disposition • For patients with brain metastases documented at baseline, brain MRI or CT was mandated at baseline, Q6W for 48 weeks, then Q9W until RECIST v1.1 progression – Patients with no baseline brain metastases did not require additional brain scans after screening unless clinically indicated • Systemic responses were assessed per RECIST v1.1 • Intracranial responses were assessed per CNS-modified RECIST v1.1 [Flow diagram] ITT population N=644 Patients with baseline brain metastases n=61 Dato-DXd n=35 -> Previously treated brain metastases n=16 ; Previously untreated brain metastases n=19 ICC n=26 -> Previously treated brain metastases n=15 ; Previously untreated brain metastases n=11 Rebecca Dent CT, computed tomography; ITT, intention-to-treat; MRI, magnetic resonance imaging.

08:30 - 10:00 Proffered paper 2: Breast cancer, early and advanced stage CHAIRS: SARA HURVITZ, SEOCK-AH IM Conclusions In a subset of patients with metastatic TNBC and asymptomatic brain metastases, with or without prior CNS-directed therapy: • Dato-DXd demonstrated systemic efficacy benefit across PFS, OS and ORR compared with ICC • Dato-DXd demonstrated higher intracranial ORR, greater DCR, longer DoR and CNS PFS benefit vs ICC TROPION-Breast02 enrolled patients without prior CNS-directed therapy, who are often excluded from clinical trials In TROPION-Breast02, 1L Dato-DXd provided efficacy in patients with metastatic TNBC and brain metastases, regardless of prior CNS-directed therapy Rebecca Dent [Speaker video panel] Rebecca Dent First-line (1L) datopotamab deruxtecan (Dato-DXd) vs chemotherapy in patients (pts) with locally recurrent inoperable or metastatic triple-negative breast cancer (mTNBC) for whom immunotherapy was not an option: Exploratory analysis of pts with brain m... ALICANTE AUDITORIUM - HALL 6

ADCs Have Transformed the Therapeutic Landscape of mTNBC - Dato-DXd and sacituzumab govitecan (SG) both approved in 1L setting for PD-L1-negative/IO ineligible mTNBC - SG plus pembrolizumab approved in 1L setting for PD-L1+ mTNBC - Trastuzumab deruxtecan approved for HR-negative/HER2-low breast cancer in 2L+ | | HER2-low | PD-L1-/IO ineligible | PD-L1-/IO ineligible | PD-L1+ | | Phase III trial | DESTINY-Breast04 | TROPION-Breast02 | ASCENT-03 | ASCENT-04 | | Treatment | T-DXd vs TPC | Dato-DXD vs TPC | SG vs TPC | SG+pembro vs KN-355 | | Prior lines of chemo for MBC | 1-2 | 0 | 0 | 0 | | Median PFS HR (95% CI) | 6.3 vs. 2.9 mo. HR 0.29 (0.15-0.57) | 10.8 vs 5.6 mo. HR 0.57 (0.47-0.69) | 9.7 vs 6.9 mo. HR 0.62 (0.50-0.77) | 11.2 vs 7.8 mo. HR 0.65 (0.51-0.84) | | Median OS HR (95% CI) | 17.1 vs 8.3 mo. HR 0.58 (0.31-1.08) | 23.7 vs 18.7 mo. HR 0.79 (0.64-0.98) | N/A (immature) | N/A (immature) | | ORR | 50.0% vs 16.7% | 62.5% vs 29.3% | 48% vs 46% | 60% vs 53% | Modi et al, NEJM 2022; Dent et al, Annals of Oncology 2026; Cortes et al, NEJM 2025; Tolaney et al, 2026 Rita Nanda, MD

SERENA-4: camizestrant upfront in all-comers SERENA-4 (NCT04711252) ER+/HER2− advanced BC, no prior therapy for advanced disease · all-comers, ESR1 not required · n = 1,371, double-blind · recurrent pts: ≥ 24 mo adjuvant ET, ≥ 12 mo since last AI Camizestrant 75 mg + palbociclib Anastrozole + palbociclib X PRIMARY PFS END POINT NOT MET numerical improvement only · magnitude not disclosed · no new safety signal · full data pending AstraZeneca press release, 11 Sep 2026 — one week after FDA accelerated approval (4 Sep 2026) of camizestrant + CDK4/6i at emergent ESR1-mut (SERENA-6; Guardant360 CDx). THE PATTERN — 1L ORAL SERD + CDK4/6i: WHO IS ENROLLED DECIDES THE RESULT Trial | Regimen | Population | n | PFS result persevERA | giredestrant + palbociclib | all-comers, endocrine-sensitive | 992 | HR 0.89 · NEGATIVE SERENA-4 | camizestrant + palbociclib | all-comers, upfront 1L | 1,371 | not met · NEGATIVE SERENA-6 | camizestrant + same CDK4/6i | ctDNA-emergent ESR1-mut only | 315 | HR 0.45 · POSITIVE 1L backbone stays AI + CDK4/6i. Two all-comer phase 3 trials (2,363 patients) could not show that an oral SERD beats an AI from day 1. Oral SERD enters 1L through ctDNA, not upfront. Camizestrant's 1L role is the SERENA-6 switch at emergent ESR1-mut — ESR1 testing on therapy becomes the actionable step. Still open: size of the SERENA-4 numerical difference · pionERA (giredestrant, ET-resistant relapse) · adjuvant CAMBRIA-1 / -2. "what the CDK4/6 inhibitor does". In persevERA and SERENA-4 it controls the whole luminal population in both arms, so the endocrine partner becomes a minor variable.

HR+/HER2− MBC: Current 1L treatment algorithm Timing of relapse decides endocrine sensitivity → biomarker decides the regimen ENDOCRINE-SENSITIVE De novo · DFI > 12 mo after ET no prior CDK4/6i AI + CDK4/6i MONALEESA-2 / -7 · MONARCH-3: OS benefit PALOMA-2: PFS only PRIOR ADJUVANT CDK4/6i ET-sensitive relapse after adjuvant abemaciclib / ribociclib CDK4/6i rechallenge or clinical trial No dedicated phase III data Judge by endocrine timing, not CDK4/6i exposure ET-RESISTANT · PIK3CA-WT On ET or ≤ 12 mo after ET PIK3CA-WT or not tested Fulvestrant + CDK4/6i MONALEESA-3 · MONARCH 2 · PALOMA-3 Exemestane + everolimus if CDK4/6i unsuitable (BOLERO-2) Post-adj. CDK4/6i rechallenge: FREEASIA (upcoming) ET-RESISTANT · PIK3CA-MUT On ET or ≤ 12 mo after ET Inavolisib + Palbociclib + Fulvestrant INAVO120: PFS 15.0 vs 7.3 mo (HR 0.43) · OS benefit Monitor hyperglycemia, stomatitis SOLAR-1 (alpelisib): pre-CDK4/6i era VISCERAL CRISIS Rapid progression Chemotherapy or T-DXd Taxane / capecitabine T-DXd if HER2-low / ultralow (NCCN) ESMO 2026: ChT only [V, B] 1.5L — on AI + CDK4/6i : ctDNA q2–3 mo → emergent ESR1-mut, no imaging PD → switch AI → camizestrant, same CDK4/6i SERENA-6 · FDA 9/2026 · EU 7/2026 · PADA-1 (fulvestrant) PD → 2L / 3L biomarker-directed algorithm · re-biopsy / ctDNA: ESR1 · PIK3CA · AKT1 · PTEN · gBRCA · HER2 status Before 1L: tissue / ctDNA for ESR1, PIK3CA, AKT1, PTEN · germline BRCA1/2 · HER2 IHC Premenopausal: OFS or oophorectomy with every endocrine regimen (MONALEESA-7) gBRCA1/2: ET + CDK4/6i as other columns, PARPi after ET (OlympiAD / EMBRACA) Endocrine monotherapy (AI or fulvestrant alone, FALCON) for patients unsuitable for CDK4/6i. Navy = current standard · Magenta = 1.5L ctDNA-guided switch · Purple = no phase III data

[search icon] Proffered Paper session LBA25 - Adjuvant ribociclib (RIB) + nonsteroidal aromatase inhibitor (NSAI) in patients (pts) with HR+/HER2− early breast cancer (EBC): NATALEE 6-y overall survival (OS) and efficacy outcomes Speakers Dennis J. Slamon (Los Angeles, United States of America) Lecture Time 14:25 - 14:35 Abstract Background The phase 3 NATALEE trial demonstrated a statistically significant invasive disease–free survival (iDFS) benefit with adjuvant RIB + NSAI in pts with high-risk stage II/III HR+/HER2− EBC. We present a prespecified analysis of OS and efficacy outcomes with 6 y of follow-up. Methods Pts with HR+/HER2− EBC were randomized 1:1 to receive RIB (400 mg/d; 3 wk on/1 wk off for 3 y) + NSAI (letrozole 2.5 mg/d or anastrozole 1 mg/d for 5 y) or NSAI alone. Premenopausal women and men also received goserelin. Pts had stage IIA (if N1 or if N0 with additional high-risk factors), stage IIB, or stage III disease per AJCC, 8th ed. OS, breast cancer–specific survival (BCSS), iDFS, distant disease–free survival (DDFS), and distant recurrence–free survival (DRFS) were evaluated via Kaplan-Meier methods. Statistical comparisons were conducted by stratified log-rank test. Results At the data cutoff (15 May 2026), all pts were off RIB for a median of 38 mo and beyond the predefined NSAI study regimen. With a median OS follow-up of 66.5 mo, RIB + NSAI showed an OS benefit over NSAI alone (HR, 0.809 [95% CI: 0.661-0.989]; nominal 1-sided P=.019), with increasing absolute benefit over time from 1.2% (93.89% vs 92.73%) at 5 y to 2.1% (91.86% vs 89.73%) at 6 y, that was consistent across key subgroups. BCSS (HR, 0.706 [95% CI: 0.559-0.890]) was supportive of OS benefit. With a median iDFS follow-up of 66.3 mo, RIB + NSAI demonstrated a persistent iDFS benefit over NSAI alone (HR, 0.741 [95% CI: 0.646-0.849]; nominal 1-sided P<.0001), with the absolute iDFS benefit increasing from 4.1% (85.32% vs 81.24%) at 5 y to 5.3% (83.23% vs 77.92%) at 6 y. DDFS (HR, 0.744 [95% CI: 0.646-0.858]) and DRFS (HR, 0.738 [95% CI: 0.634-0.858]) benefits were also maintained. No new safety signals were observed. Conclusions With 6 y of follow-up and all pts off RIB (median, ≈3 y), the sustained reduction in recurrence risk and positive OS results with RIB + NSAI over NSAI alone continue to demonstrate a substantial and clinically relevant benefit for pts with high-risk stage II/III HR+/HER2− EBC, supporting the use of RIB + NSAI in the adjuvant setting across the broad NATALEE-eligible population.

[Slide 1] NATALEE: open-label, multicenter, randomized, phase 3 trial1,2 Adult patients with stage II or III HR+/HER2- EBC - Prior ET allowed up to 12 months - Anatomical stage IIAa - N1 - N0 with: - Grade 2 and evidence of high risk: - Ki-67 >=20% - Oncotype DX Breast Recurrence Score >=26 or - High risk via genomic risk profiling - Grade 3 - Anatomical stage IIBa - N0 or N1 - Anatomical stage III - N0, N1, N2, or N3 N = 5101 R 1:1 RIB 400 mg/day 3 weeks on/1 week off for 3 y + NSAI Letrozole or anastrozoleb for >=5 y + goserelin in men and premenopausal women NSAI Letrozole or anastrozoleb for >=5 y + goserelin in men and premenopausal women Primary end point: iDFS using STEEP criteria Secondary end points: OS, DDFS, RFS; PROs; Safety and tolerability; PK Exploratory end points: DRFS, BCSS; Gene expression and alterations in tumor ctDNA/ctRNA samples Efficacy outcomes for the 6-year analysis were estimated by the Kaplan-Meier method, and the results are descriptive. A Cox proportional hazards model was used to estimate the HRs and 95% CIs. Timeline (Median iDFS follow-up): Primary analysis3 27.7 mo | Final iDFS4 33.3 mo | 4-year exploratory5 44.2 mo | 5-year prespecified6 55.4 mo | 6-year prespecified 66.3 mo Current readout | and beyond BCSS, breast cancer-specific survival; ctDNA/RNA, circulating tumor DNA/RNA; DDFS, distant disease-free survival; DRFS, distant recurrence-free survival; EBC, early breast cancer; ET, endocrine therapy; HR+/HER2-, hormone receptor positive, human epidermal growth factor receptor 2 negative; HR, hazard ratio; iDFS, invasive disease-free survival; N, node; NSAI, nonsteroidal aromatase inhibitor; OS, overall survival; PK, pharmacokinetics; PRO, patient-reported outcome; RIB, ribociclib; RFS, recurrence-free survival; STEEP, Standardized Definitions for Efficacy End Points. a Enrollment of patients with stage II disease was capped at 40%. b Per investigator choice. 1. ClinicalTrials.gov. Accessed 26 August 2026. https://clinicaltrials.gov/study/NCT03701334. 2. Slamon D et al. Ther Adv Med Oncol. 2023;15:1-16. 3. Slamon D et al. New Engl J Med. 2024;390(12):1080-1091. 4. Hortobagyi G et al. Ann Oncol. 2025;36(2):149-157. 5. Fasching PA et al. JAMA Oncol. 2025;11(11):1364-1372. 6. Crown J et al. ESMO Open. 2025;10(11):105858. Dennis J. Slamon, MD, PhD [Slide 2] Positive OS results at this prespecified 6-year analysis Landmark labels: 60 mo: 93.9% (Δ=1.2%) RIB + NSAI vs 92.7% NSAI alone; 72 mo: 91.9% (Δ=2.1%a) RIB + NSAI vs 89.7% NSAI alone Median follow-up: 66.5 months RIB + NSAI (n = 2549) | NSAI alone (n = 2552) No. of OS events 178 | 205 HR (95% CI) 0.809 (0.661-0.989) Nominal P value (1-sided)b 0.019 Y axis: Overall survival (%); X axis: Time since randomization (Months) No. of patients at risk (0,6,...,90 mo): RIB + NSAI: 2549 2404 2336 2299 2258 2219 2184 2132 2086 2047 1982 1521 761 85 2 0 NSAI alone: 2552 2301 2254 2208 2159 2117 2064 2012 1981 1936 1866 1426 730 81 3 0 HR, hazard ratio; NSAI, nonsteroidal aromatase inhibitor; RIB, ribociclib. a The difference between percentages does not equal 2.2 due to rounding. b Comparison of survival between treatment arms was generated by stratified log-rank test (1-sided P value, descriptive). Dennis J. Slamon, MD, PhD Survival table Endpoint: OS (prespecified 6-year analysis) RIB + NSAI | 2549 NSAI alone | 2552 HR: 0.809 HR CI: 0.661-0.989 P: 0.019 (nominal, 1-sided, descriptive) Landmark rates: 60 mo: RIB + NSAI 93.9% | NSAI alone 92.7% 72 mo: RIB + NSAI 91.9% | NSAI alone 89.7% [Slide 3] Consistent OS benefit across key subgroups Columns: RIB + NSAI Events/n | 6-y OS rate, % | NSAI alone Events/n | 6-y OS rate, % | 6-y Δ, % | HR (95% CI) | HR | 95% CI ITT 178/2549 | 91.86 | 205/2552 | 89.73 | 2.1 | 0.809 | 0.661-0.989 SUBGROUPS Region North America/Western Europe/Oceania 103/1563 | 91.96 | 113/1565 | 90.63 | 1.3 | 0.859 | 0.657-1.122 Rest of world 75/986 | 91.70 | 92/987 | 88.38 | 3.3 | 0.776 | 0.572-1.053 Menopausal status Premenopausal women and men [boxed] 53/1126 | 94.60 | 71/1132 | 92.17 | 2.4 | 0.705 | 0.494-1.006 Postmenopausal women 125/1423 | 89.73 | 134/1420 | 87.88 | 1.9 | 0.884 | 0.692-1.129 Prior CTa Yes 167/2249 | 91.23 | 187/2245 | 89.20 | 2.0 | 0.834 | 0.677-1.028 No 11/300 | 96.41 | 18/307 | 93.25 | 3.2 | 0.623 | 0.294-1.320 Prior ET Yes 117/1833 | 92.61 | 126/1807 | 91.07 | 1.5 | 0.858 | 0.667-1.105 No 61/716 | 90.09 | 79/745 | 86.65 | 3.4 | 0.767 | 0.549-1.072 AJCC stage Stage II [boxed] 39/1012 | 96.16 | 47/1034 | 94.48 | 1.7 | 0.830 | 0.543-1.269 Stage III 139/1527 | 88.60 | 158/1512 | 85.98 | 2.6 | 0.798 | 0.635-1.003 Nodal statusb N0 [boxed] 15/285 | 94.39 | 21/329 | 92.22 | 2.2 | 0.790 | 0.407-1.533 N1-N3 163/2261 | 91.48 | 184/2218 | 89.31 | 2.2 | 0.817 | 0.661-1.009 Ki-67 statusc Ki-67 <=20% 68/1200 | 93.70 | 81/1241 | 91.58 | 2.1 | 0.827 | 0.598-1.143 Ki-67 >20% 80/919 | 89.57 | 93/932 | 87.27 | 2.3 | 0.812 | 0.602-1.095 Axis 0.0 0.5 1.0 1.5 2.0: Favors RIB + NSAI | Favors NSAI alone AJCC, American Joint Committee on Cancer; CT, chemotherapy; ET, endocrine therapy; HR, hazard ratio; ITT, intent to treat; N, node; NSAI, nonsteroidal aromatase inhibitor; OS, overall survival; RIB, ribociclib. a Includes neoadjuvant and adjuvant chemotherapy. b Nodal status classification according to AJCC staging. Nodal status is from the most advanced stage derived per surgical specimen or at diagnosis. c From archival tumor tissue. Dennis J. Slamon, MD, PhD [Slide 4] BCSS results and survival status support OS benefit BCSS Landmark labels: 60 mo: 95.5% (Δ=1.5%) RIB + NSAI vs 94.0% NSAI alone; 72 mo: 94.2% (Δ=2.6%) RIB + NSAI vs 91.6% NSAI alone Median follow-up: 66.5 months RIB + NSAI (n = 2549) | NSAI alone (n = 2552) No. of BCSS events 126 | 166 HR (95% CI) 0.706 (0.559-0.890) [circled] Nominal P value (1-sided) .002 Y axis: Breast cancer-specific survival (%); X axis: Time since randomization (Months) No. of patients at risk: RIB + NSAI: 2549 2404 2336 2299 2258 2219 2184 2132 2086 2047 1982 1521 761 85 2 0 NSAI alone: 2552 2301 2254 2208 2159 2117 2064 2012 1981 1936 1866 1426 730 81 3 0 Deaths and survival status (stacked bars, Number of patients): RIB + NSAI: Alive with metastatic disease 5.0% (n = 127); Deaths due to BC 4.9%a (n = 126); Deaths not related to BC 2.0% (n = 52); total bracket 9.9% NSAI alone: Alive with metastatic disease 6.8% (n = 173); Deaths due to BC 6.5%a (n = 166); Deaths not related to BC 1.5% (n = 39); total bracket 13.3% Legend: Alive with metastatic disease [boxed]; Deaths due to BC; Deaths not related to BC BC, breast cancer; BCSS, breast cancer-specific survival; HR, hazard ratio; ITT, intention to treat; NSAI, nonsteroidal aromatase inhibitor; OS, overall survival; RIB, ribociclib. a One patient randomized to the RIB + NSAI arm did not receive RIB. This patient is included in the RIB + NSAI arm in the ITT population. In the safety set population, they are included in the NSAI-alone arm. Dennis J. Slamon, MD, PhD Survival table Endpoint: BCSS (breast cancer-specific survival, exploratory) RIB + NSAI | 2549 NSAI alone | 2552 HR: 0.706 HR CI: 0.559-0.890 P: .002 (nominal, 1-sided) Landmark rates: 60 mo: RIB + NSAI 95.5% | NSAI alone 94.0% 72 mo: RIB + NSAI 94.2% | NSAI alone 91.6%

13:30 - 15:00 Proffered paper 1: Breast cancer, early and advanced stage CHAIRS: KOMAL JHAVERI, BARBARA PISTILLI Greater postprogression CDK4/6 inhibitor use in NSAI-alone arm Post–distant recurrence therapy | RIB + NSAI (n = 219) | NSAI alone (n = 320) Received first-line anticancer therapy, n (%) | 192 (87.7) | 285 (89.1) ET only, n (%) | 61 (27.9) | 66 (20.6) ET plus targeted therapy, n (%)a | 57 (26.0) | 144 (45.0) Targeted therapy type CDK4/6 inhibitorb | 46 (21.0) | 142 (44.4) Abemaciclib | 23 (10.5) | 40 (12.5) Palbociclib | 14 (6.4) | 37 (11.6) Ribociclib | 8 (3.7) | 65 (20.3) AKT inhibitor | 5 (2.3) | 0 Checkpoint inhibitor | 2 (0.9) | 0 PI3K inhibitor | 5 (2.3) | 1 (0.3) Otherc | 5 (2.3) | 2 (0.6) Chemotherapy only, n (%) | 48 (21.9) | 46 (14.4) Other, n (%) | 26 (11.9) | 29 (9.1) CDK, cyclin-dependent kinase; ET, endocrine therapy; NSAI, nonsteroidal aromatase inhibitor; PI3K, phosphatidylinositol 3-kinase; RIB, ribociclib. a In the ET plus targeted therapy category, patients may receive more than 1 targeted therapy, so the individual counts may exceed the number of patients in the overall category. b One patient in the RIB + NSAI arm received ET plus dalpiciclib. c In the NSAI-alone arm, 1 patient received the CDK4 inhibitor atirmociclib plus ET, and 1 patient received the CDK2 inhibitor tegtociclib plus ET. Dennis J. Slamon, MD, PhD [Speaker panel] Dennis Slamon — Adjuvant ribociclib (RIB) + nonsteroidal aromatase inhibitor (NSAI) in patients (pts) with HR+/HER2− early breast cancer (EBC): NATALEE 6-y overall survival (OS) and efficacy outcomes ALICANTE AUDITORIUM - HALL 6

[Slide 1] Proffered Paper session LBA25 - Adjuvant ribociclib (RIB) + nonsteroidal aromatase inhibitor (NSAI) in patients (pts) with HR+/HER2− early breast cancer (EBC): NATALEE 6-y overall survival (OS) and efficacy outcomes Speakers Dennis J. Slamon (Los Angeles, United States of America) Lecture Time 14:25 - 14:35 [Abstract] Background The phase 3 NATALEE trial demonstrated a statistically significant invasive disease−free survival (iDFS) benefit with adjuvant RIB + NSAI in pts with high-risk stage II/III HR+/HER2− EBC. We present a prespecified analysis of OS and efficacy outcomes with 6 y of follow-up. [Slide 2] Results At the data cutoff (15 May 2026), all pts were off RIB for a median of 38 mo and beyond the predefined NSAI study regimen. With a median OS follow-up of 66.5 mo, RIB + NSAI showed an OS benefit over NSAI alone (HR, 0.809 [95% CI: 0.661-0.989]; nominal 1-sided P=.019), with increasing absolute benefit over time from 1.2% (93.89% vs 92.73%) at 5 y to 2.1% (91.86% vs 89.73%) at 6 y, that was consistent across key subgroups. BCSS (HR, 0.706 [95% CI: 0.559-0.890]) was supportive of OS benefit. With a median iDFS follow-up of 66.3 mo, RIB + NSAI demonstrated a persistent iDFS benefit over NSAI alone (HR, 0.741 [95% CI: 0.646-0.849]; nominal 1-sided P<.0001), with the absolute iDFS benefit increasing from 4.1% (85.32% vs 81.24%) at 5 y to 5.3% (83.23% vs 77.92%) at 6 y. DDFS (HR, 0.744 [95% CI: 0.646-0.858]) and DRFS (HR, 0.738 [95% CI: 0.634-0.858]) benefits were also maintained. No new safety signals were observed. Conclusions With 6 y of follow-up and all pts off RIB (median, ≈3 y), the sustained reduction in recurrence risk and positive OS results with RIB + NSAI over NSAI alone continue to demonstrate a substantial and clinically relevant benefit for pts with high-risk stage II/III HR+/HER2− EBC, supporting the use of RIB + NSAI in the adjuvant setting across the broad NATALEE-eligible population. [Slide 3] Methods Pts with HR+/HER2− EBC were randomized 1:1 to receive RIB (400 mg/d; 3 wk on/1 wk off for 3 y) + NSAI (letrozole 2.5 mg/d or anastrozole 1 mg/d for 5 y) or NSAI alone. Premenopausal women and men also received goserelin. Pts had stage IIA (if N1 or if N0 with additional high-risk factors), stage IIB, or stage III disease per AJCC, 8th ed. OS, breast cancer−specific survival (BCSS), iDFS, distant disease−free survival (DDFS), and distant recurrence−free survival (DRFS) were evaluated via Kaplan-Meier methods. Statistical comparisons were conducted by stratified log-rank test.

[Slide 1] NATALEE: ADJUVANT RIBOCICLIB NATALEE: A CONSISTENT OS SIGNAL AT 6 YEARS [KM curve: Overall survival (%) vs Time since randomization (Months)] Labels: 93.9% (Δ=1.2%) / 92.7% at 60 months; 91.9% (Δ=2.1%^a) / 89.7% at 72 months Median follow-up: 66.5 months | | RIB + NSAI (n = 2549) | NSAI alone (n = 2552) | No. of OS events | 178 | 205 HR (95% CI) | 0.809 (0.661-0.989) Nominal P value (1-sided)^b | 0.019 No. of patients at risk (0,6,12,...,84+ months): RIB + NSAI: 2549, 2404, 2336, 2299, 2258, 2219, 2184, 2132, 2086, 2047, 1982, 1521, 761, 85, 2, 0 NSAI alone: 2552, 2301, 2254, 2208, 2159, 2117, 2064, 2012, 1981, 1936, 1866, 1426, 730, 81, 3, 0 6-YEAR OVERALL SURVIVAL 91.9% vs 89.7% Δ +2.1% | HR 0.809 (0.661-0.989) BREAST CANCER-SPECIFIC SURVIVAL HR 0.706 (0.559-0.890) 126 vs 166 deaths: 40 fewer breast-cancer deaths Nominal 1-sided P=0.019 - descriptive analysis NOT multiplicity-controlled confirmatory testing Non-breast-cancer deaths: 52 vs 39; specific causes not detailed. Six-year OS and BCSS support ribociclib; the OS inference remains descriptive NATALEE LBA25, ESMO 2026: 6-y OS HR 0.809 (0.661-0.989), BCSS HR 0.706 (0.559-0.890). Nominal 1-sided P=.019. Survival table Endpoint: Overall survival (NATALEE, 6-year) RIB + NSAI | 2549 NSAI alone | 2552 HR: 0.809 HR CI: 0.661-0.989 P: Nominal 1-sided P=0.019 (descriptive) Landmark rates: 60 months: RIB + NSAI 93.9% | NSAI alone 92.7% 72 months (6-year): RIB + NSAI 91.9% | NSAI alone 89.7% [Slide 2] NATALEE: ADJUVANT RIBOCICLIB NATALEE: WHY DOES THE iDFS GAP WIDEN AFTER RIBOCICLIB? [KM curve: Invasive disease-free survival (%) vs Time since randomization (Months)] Labels: 85.3% (Δ=4.1%) / 81.2% at 60 months; 83.2% (Δ=5.3%) / 77.9% at 72 months Median follow-up: 66.3 months | | RIB + NSAI (n = 2549) | NSAI alone (n = 2552) | No. of iDFS events | 374 | 461 HR (95% CI) | 0.741 (0.646-0.849) Nominal P value (1-sided) | <.0001 No. of patients at risk: RIB + NSAI: 2549, 2352, 2277, 2212, 2141, 2086, 2029, 1971, 1916, 1853, 1753, 1300, 629, 64, 0, 0 NSAI alone: 2552, 2240, 2168, 2082, 2009, 1941, 1875, 1805, 1745, 1675, 1590, 1173, 566, 48, 1, 0 iDFS ABSOLUTE DIFFERENCE 5-year Δ 4.1% -> 6-year Δ 5.3% Median time since stopping ribociclib: 38 months All patients had discontinued ribociclib WHY DOES THE GAP WIDEN AFTER TREATMENT? - Eradication? (loss of residual clones) - Prolonged suppression? (senescence / immune effects?) - Delayed clinical recurrence? (later catch-up?) The widening iDFS gap raises biological hypotheses, not proof of a mechanism NATALEE LBA25: both 5-/6-year values from same 2026 cutoff. Here Δ in % denotes an absolute percentage-point difference; mechanisms are speculative. Survival table Endpoint: Invasive disease-free survival (NATALEE) RIB + NSAI | 2549 NSAI alone | 2552 HR: 0.741 HR CI: 0.646-0.849 P: Nominal 1-sided <.0001 Landmark rates: 60 months (5-year): RIB + NSAI 85.3% | NSAI alone 81.2% 72 months (6-year): RIB + NSAI 83.2% | NSAI alone 77.9%

13:30 - 15:00 Proffered paper 1: Breast cancer, early and advanced stage CHAIRS: KOMAL JHAVERI, BARBARA PISTILLI Positive OS results at this prespecified 6-year analysis [KM plot: Overall survival (%) vs Time since randomization (Months) 0-84; dashed lines at 60 and 72 months] RIB + NSAI: 93.9% (Δ=1.2%) at 60 mo line; 91.9% (Δ=2.1%a) at 72 mo line NSAI alone: 92.7% at 60 mo line; 89.7% at 72 mo line Median follow-up: 66.5 months RIB + NSAI (n = 2549) | NSAI alone (n = 2552) No. of OS events 178 | 205 HR (95% CI) 0.809 (0.661-0.989) Nominal P value (1-sided)b 0.019 No. of patients at risk RIB + NSAI 2549 2404 2336 2299 2258 2219 2184 2132 2086 2047 1982 1521 761 85 2 0 NSAI alone 2552 2301 2254 2208 2159 2117 2064 2012 1981 1936 1866 1426 730 81 3 0 HR, hazard ratio; NSAI, nonsteroidal aromatase inhibitor; RIB, ribociclib. a The difference between percentages does not equal 2.2 due to rounding. b Comparison of survival between treatment arms was generated by stratified log-rank test (1-sided P value, descriptive). Dennis J. Slamon, MD, PhD [Speaker card] Dennis Slamon — Adjuvant ribociclib (RIB) + nonsteroidal aromatase inhibitor (NSAI) in patients (pts) with HR+/HER2- early breast cancer (EBC): NATALEE 6-y overall survival (OS) and efficacy outcomes ALICANTE AUDITORIUM - HALL 6 Survival table Endpoint: Overall survival (NATALEE 6-year prespecified analysis) RIB + NSAI | 2549 NSAI alone | 2552 HR: 0.809 HR CI: 0.661-0.989 P: 0.019 (nominal, 1-sided, descriptive) Landmark rates: 60 months (5-yr): RIB + NSAI 93.9% | NSAI alone 92.7% 72 months (6-yr): RIB + NSAI 91.9% | NSAI alone 89.7%

NATALEE: ADJUVANT RIBOCICLIB WHO BENEFITS? CLINICAL AND GENOMIC RISK: NO VALIDATED PREDICTIVE MARKER NATALEE: selected 6-year OS subgroups All patients (100%) 0.81 (0.66–0.99) N0 (~12%) 0.79 (0.41–1.53) N1–3 (~88%) 0.82 (0.66–1.01) Stage II (~40%) 0.83 (0.54–1.27) Stage III (~60%) 0.80 (0.64–1.00) Ki-67 >20% (~36%*) 0.81 (0.60–1.10) [Forest plot axis: 0.5 0.7 1.0 1.5] CLINICAL AND GENOMIC MARKERS OS HR (95% CI) - N0: 0.79 (0.41–1.53) - Stage II: 0.83 (0.54–1.27) GENOMIC SIGNATURES - NATALEE PAM50 (ASCO 2026): benefit across subtypes - Treatment interaction: P = 0.34 (not significant) - monarchE: exploratory 21-gene analyses ctDNA: DYNAMIC MRD RISK - monarchE ctDNA strongly prognostic; - not validated to select, omit or stop CDK4/6i. No validated marker yet selects the relative CDK4/6i benefit Risk can be estimated; no validated marker yet selects the relative CDK4/6i benefit LBA25 OS subgroups (exploratory, low event counts). NATALEE gene-expression: Chia et al. ASCO 2026 abstract 501, n=3,022, PAM50 P-int=.34. monarchE ctDNA: ASCO 2024 LBA507. *Ki-67 high =36% of all randomized; results available only in biomarker-evaluable subset.

[Slide 1] 08:30 - 09:15 Management of early breast cancer in young patients CHAIR: SHANI PALUCH-SHIMON Adjuvant Endocrine Therapy in Premenopausal Women Proper risk stratification is crucial Tamoxifen alone in low-risk SOFT trial No chemotherapy cohort Predominantly pT1pN0 grade 1-2 OFS + (preferably) AI in high-risk SOFT trial Chemotherapy cohort (HER2-negative) SOFT & TEXT trials Chemotherapy cohort (HER2-negative) B DRFI, by cohort D OS, by cohort Francis PA et al, Ann Oncol 2026;37(10):1367-85 Matteo Lambertini, MD PhD @matteolambe matteo.lambertini@unige.it Matteo Lambertini Presentation by Expert [Slide 2] 08:30 - 09:15 Management of early breast cancer in young patients CHAIR: SHANI PALUCH-SHIMON Adjuvant CDK4/6 Inhibitors in Young Women monarchE NATALEE COHORT 1 91% R 1:1 N = 5637* Abemaciclib (150mg twice daily) + Endocrine Therapy Endocrine Therapy Ribociclib 400 mg/day 3 weeks on/1 week off for 3 y NSAI R 1:1 N = 5101 Primary End Point – iDFS using STEEP criteria 44% premenopausal patients Tamoxifen alone allowed OFS + AI mandatory Johnston SRD et al, J Clin Oncol 2020;38(34):3987-98. Harbeck N et al, Ann Oncol 2021;32(12):1571-81 Matteo Lambertini, MD PhD @matteolambe matteo.lambertini@unige.it [Slide 3] 08:30 - 09:15 Management of early breast cancer in young patients CHAIR: SHANI PALUCH-SHIMON Germline BRCA Testing: Therapeutic Implications Adjuvant choices in BRCA carriers with HR+/HER2- breast cancer Results: Overall Survival in ER+/HER2- Breast Cancer Overall Survival (n=9,072) | HR (95%CI) | p-value Non-carriers | Reference ATM | 0.9 (0.5 – 1.8) | 0.91 BRCA1 | 1.8 (0.9 – 3.7) | 0.08 BRCA2 | 1.8 (1.1 – 3.0) | 0.01 CHEK2 | 0.8 (0.5 – 1.4) | 0.44 PALB2 | 1.5 (0.8 – 2.9) | 0.23 Results: Analyses by Tumor Subtype Disease-Free Survival Overall Survival RS<18 Total (95% CI) 0.27 [0.14, 0.51] RS 18-30 Total (95% CI) 1.74 [1.28, 2.37] RS >30 Total (95% CI) 3.71 [2.55, 5.40] Yadav S et al, ASCO 2025. Arecco L et al, Ann Oncol 2024;35(9):792-804. Davey MG et al, Eur J Cancer 2021;154:209-16 Matteo Lambertini, MD PhD @matteolambe matteo.lambertini@unige.it [Slide 4] 08:30 - 09:15 Management of early breast cancer in young patients CHAIR: SHANI PALUCH-SHIMON Pregnancy after Breast Cancer – Is It Safe ? Safety after breast cancer overall Studies Adjusting for the Potential «Guarantee-time Bias» Disease-free Survival Overall Survival HR 0.74 (95% CI 0.58-0.96); p=0.023 HR 0.52 (95% CI 0.42-0.65); p<0.001 Safety after ER+ breast cancer Author | Year | HR (95% CI) Valentini A. et al. | 2013 | 0.39 (0.04, 3.63) Lambertini M. et al. | 2018 | 0.84 (0.60, 1.18) Chuang S. et al. | 2020 | 0.23 (0.07, 0.77) Rauh-Hain A.J. et al. | 2021 | 0.43 (0.20, 0.91) Anderson R.A. et al. | 2022 | 0.66 (0.37, 1.18) Bae S.Y. et al. | 2022 | 0.25 (0.14, 0.43) Random effect (I-squared = 69.6%, p = 0.006) | 0.46 (0.27, 0.77) Safety in BRCA carriers HR 0.63; 95% CI 0.53 – 0.74 HR 0.43; 95% CI 0.30 – 0.61 Lambertini M et al, J Clin Oncol 2021;39(29):3293-305. Arecco L et al, ESMO Open 2023;8(6):102031. Lambertini M et al, J Clin Oncol 2020;38:3012-23. Lambertini M et al, JAMA 2024;331(1):49-59 Matteo Lambertini, MD PhD @matteolambe matteo.lambertini@unige.it

Robust iDFS benefit continues to deepen over time 85.3% (Δ=4.1%) 83.2% (Δ=5.3%) 81.2% 77.9% Median follow-up: 66.3 months RIB + NSAI (n = 2549) | NSAI alone (n = 2552) No. of iDFS eventsᵃ: 374 | 461 HR (95% CI): 0.741 (0.646-0.849) Nominal P value (1-sided): <.0001 Y-axis: Invasive disease-free survival (%); X-axis: Time since randomization (Months) 0-84 No. of patients at risk RIB + NSAI: 2549 2352 2277 2212 2141 2086 2029 1971 1916 1853 1753 1300 629 64 0 0 NSAI alone: 2552 2240 2168 2082 2009 1941 1875 1805 1745 1675 1590 1173 566 48 1 0 HR, hazard ratio; iDFS, invasive disease-free survival; NSAI, nonsteroidal aromatase inhibitor; RIB, ribociclib. ᵃDistant recurrence events accounted for 242 iDFS events in the RIB + NSAI arm and 326 iDFS events in the NSAI-alone arm. Dennis J. Slamon, MD, PhD Survival table Endpoint: Invasive disease-free survival (iDFS), NATALEE 6-year analysis RIB + NSAI | 2549 NSAI alone | 2552 HR: 0.741 HR CI: 0.646-0.849 P: <.0001 (nominal, 1-sided) Landmark rates: 60 mo: RIB + NSAI 85.3% | NSAI alone 81.2% 72 mo: RIB + NSAI 83.2% | NSAI alone 77.9%

DDFS and DRFS results support reduced recurrence risk with RIB + NSAI [Left panel] DDFS 86.5% (Δ=3.7%) 84.5% (Δ=4.9%) 82.8% 79.6% Median follow-up: 66.3 months RIB + NSAI (n = 2549) | NSAI alone (n = 2552) No. of DDFS events: 344 | 423 HR (95% CI): 0.744 (0.646, 0.858) Nominal P value (1-sided): <.001 Y-axis: Distant disease-free survival (%); X-axis: Time since randomization (Months) 0-84 No. of patients at risk RIB + NSAI: 2549 2354 2284 2220 2154 2098 2042 1982 1924 1864 1764 1309 632 65 0 0 NSAI alone: 2552 2244 2172 2094 2025 1956 1892 1820 1760 1693 1609 1187 576 50 1 0 [Right panel] DRFS 88.2% (Δ=3.5%) 86.0% (Δ=4.4%) 84.7% 81.6% Median follow-up: 66.3 months RIB + NSAI (n = 2549) | NSAI alone (n = 2552) No. of DRFS events: 305 | 378 HR (95% CI): 0.738 (0.634-0.858) Nominal P value (1-sided): <.0001 Y-axis: Distant recurrence-free survival (%); X-axis: Time since randomization (Months) 0-84 No. of patients at risk RIB + NSAI: 2549 2362 2292 2238 2171 2119 2066 2007 1955 1892 1794 1327 634 66 0 0 NSAI alone: 2552 2251 2181 2110 2045 1981 1915 1846 1788 1725 1640 1215 589 51 1 0 DDFS, distant disease-free survival; DRFS, distant recurrence-free survival; HR, hazard ratio; NSAI, nonsteroidal aromatase inhibitor; RIB, ribociclib. Dennis J. Slamon, MD, PhD Survival table Endpoint: Distant disease-free survival (DDFS); DRFS on same slide (see notes) RIB + NSAI | 2549 NSAI alone | 2552 HR: 0.744 HR CI: 0.646-0.858 P: <.001 (nominal, 1-sided) Landmark rates: 60 mo: RIB + NSAI 86.5% | NSAI alone 82.8% 72 mo: RIB + NSAI 84.5% | NSAI alone 79.6%

Conclusions - In the NATALEE trial, this 6-year analysis demonstrates a positive OS result, with a 2.1% absolute OS benefit with RIB + NSAI vs NSAI alone across the broad study population - OS benefit with RIB + NSAI has consistently shown a ≈ 20% reduction in the risk of death over NSAI alone and is supported by a BCSS benefit (≈ 30% reduction in risk of breast cancer-related death) - The rate of patients alive with metastatic disease was lower in the RIB + NSAI arm (5.0%) than in the NSAI-alone arm (6.8%) - Patients treated with RIB + NSAI showed ≈ 5% absolute improvements in 6-year iDFS, DDFS, and DRFS rates over those treated with NSAI alone, which continue to deepen - With patients off RIB (median, 38 months) and beyond the protocol-defined NSAI treatment regimen duration, the safety profile of RIB remains unchanged This prespecified 6-year analysis of NATALEE supports the use of RIB + NSAI to reduce recurrence risk and ultimately improve OS in patients with HR+/HER2- EBC at high risk of recurrence BCSS, breast cancer-specific survival; DDFS, distant disease-free survival; DRFS, distant recurrence-free survival; EBC, early breast cancer; HR+/HER2-, hormone receptor positive, human epidermal growth factor receptor 2 negative; iDFS, invasive disease-free survival; NSAI, nonsteroidal aromatase inhibitor; OS, overall survival; RIB, ribociclib. Dennis J. Slamon, MD, PhD

NATALEE: open-label, multicenter, randomized, phase 3 trial Adult patients with stage II or III HR+/HER2− EBC • Prior ET allowed up to 12 months • Anatomical stage IIA(a) • N1 • N0 with: • Grade 2 and evidence of high risk: • Ki-67 ≥20% • Oncotype DX Breast Recurrence [partially obscured] • High risk via genomic risk profil[ing — partially obscured] • Grade 3 • Anatomical stage IIB(a) • N0 or N1 • Anatomical stage III • N0, N1, N2, or N3 N = 5101 Rationale for broad population of patients: Patients with stage II or III HR+/HER2− EBC, including those with no nodal involvement, are at risk of disease recurrence up to decades after initial diagnosis(1,5) RIB 400 mg/day 3 weeks on/1 week off for 3 y + NSAI Letrozole or anastrozole(b) for ≥5 y ± goserelin in men and premenopausal women NSAI Letrozole or anastrozole(b) for ≥5 y ± goserelin in men and premenopausal women Rationale for 400 mg RIB: To improve tolerability while maintaining efficacy(1) Rationale for 3-year treatment duration: Extended duration of treatment is crucial to prolong cell cycle arrest and drive more tumor cells into irreversible senescence(1-4) [Partially obscured endpoints box] ... OS, DDFS, RFS Efficacy outcomes for the 6-year analysis were estimated by the Kapl[an-Meier method — partially obscured]. A Cox proportional hazards model was used to estimate the HRs and 95% CIs. Timeline: Primary analysis 27.7 mo ✓ | Final iDFS 33.3 mo ✓ | 4-year exploratory 44.2 mo ✓ | 5-year prespecified 55.4 mo ✓ | 6-year prespecified 66.3 mo — Current readout — and beyond (Median iDFS follow-up) BCSS, breast cancer–specific survival; ctDNA/RNA, circulating tumor DNA/RNA; DDFS, distant disease–free survival; DRFS, distant recurrence–free survival; EBC, early breast cancer; ET, endocrine therapy; HR+/HER2−, hormone receptor positive, human epidermal growth factor receptor 2 negative; HR, hazard ratio; iDFS, invasive disease–free survival; N, node; NSAI, nonsteroidal aromatase inhibitor; OS, overall survival; PK, pharmacokinetics; PRO, patient-reported outcome; RIB, ribociclib; RFS, recurrence-free survival; STEEP, Standardized Definitions for Efficacy End Points. (a)Enrollment of patients with stage II disease was capped at 40%. (b)Per investigator choice. 1. Slamon D et al. Ther Adv Med Oncol. 2023;15:1-16. 2. Kovatcheva M et al. Oncotarget. 2015;6:8226-8243. 3. Rader J et al. Clin Cancer Res. 2013;19:6173-6182. 4. Klein ME et al. Cancer Cell. 2018;34:9-20. 5. Pan H et al. N Engl J Med. 2017;377(19):1836-1846. Dennis J. Slamon, MD, PhD

Adjuvant CDK4/6 Inhibitors in Young Women monarchE ITT population (includes both Cohort 1 and Cohort 2) First patient randomized July 2017; N=5,120 HR+, HER2-, Node-Positive, High-Risk EBC COHORT 1 91% COHORT 1: High-Risk based on clinical pathological features • ≥4 ALN OR • 1-3 ALN and at least 1 of the below: - Grade 3 disease - Tumor size ≥5 cm Cohort 2 9% Cohort 2: High risk based on Ki-67 First patient randomized one year later; N=517 R 1:1 N = 5637 On-study treatment period 2 years Abemaciclib (150mg twice daily) + Endocrine Therapy Endocrine Therapy Follow-up period Endocrine Therapy 3-8 years as clinically indicated Primary Objective: IDFS Secondary Objectives: IDFS in high Ki-67 populations, DRFS, OS, safety, PK, PROs NATALEE • Adult patients with HR+/HER2− EBC • Prior ET allowed up to 12 mo • Anatomical stage IIA • Anatomical stage IIB • Anatomical stage III N = 5101 R 1:1 Ribociclib 400 mg/day 3 weeks on/1 week off for 3 y NSAI Letrozole or anastrozole for ≥ 5 y + goserelin in men and premenopausal women NSAI Letrozole or anastrozole for ≥ 5 y + goserelin in men and premenopausal women Primary End Point – iDFS using STEEP criteria Secondary End Points – Recurrence-free survival – Distant disease–free survival – OS – PROs – Safety and tolerability – PK Exploratory End Points – Locoregional recurrence–free survival – Gene expression and alterations in tumor ctDNA/ctRNA samples 44% premenopausal patients Tamoxifen alone allowed OFS + AI mandatory Johnston SRD et al, J Clin Oncol 2020;38(34):3987-98. Harbeck N et al, Ann Oncol 2021;32(12):1571-81 Matteo Lambertini, MD PhD @matteolambe matteo.lambertini@unige.it

[Slide 2] In early HR+/HER2- breast cancer… … three layers of information drive the primary treatment decision 1 Anatomy cT cN cM breast, axilla ± systemic imaging 2 Baseline biology Grade | ER/PR | HER2 | Ki67 Genomic score prognostic; partly predicts CT benefit 3 Dynamic biology Ki67 after 2-4 weeks of ET endocrine sensitivity in vivo Upfront surgery vs neoadjuvant therapy Mafalda Oliveira, MD PhD ER: estrogen receptor; PR: progesterone receptor; CT: chemotherapy; ET: endocrine therapy [Slide 3] Practical framework: a proposal Note: this practical framework is a personal opinion of the presenter; it is not intended to be a guideline, nor is it guideline-endorsed as presented All patients: complete staging · ER/PR, grade, Ki67 · gBRCA1/2 if criteria met cT1-2 cN0-1 | Postmenopausal, low/intermediate clinical risk Upfront surgery → genomic score Genomic low or intermediate risk, pN0-1: ET alone Genomic high risk: CT → ET Validated assays: Oncotype DX (TAILORx, RxPONDER) · MammaPrint (MINDACT) · Prosigna (OPTIMA) · EndoPredict (ABCSG, TransATAC) 4-9 nodes: OPTIMA, short follow-up Lobular: surgery first; genomic-high risk is rare cT1-2 cN0 | Premenopausal, low/intermediate clinical risk Upfront surgery → genomic score pN0, low genomic risk: ET alone pN0 intermediate or high genomic risk, or pN1: CT → OFS-based ET OFS + AI alone could be considered, although investigational, in intermediate risk (OPTIMA) Best evidence ≤50 y: Oncotype DX (TAILORx); MammaPrint (MINDACT) if clinically low risk Doubtful CT indication | e.g., cT2 N0-1, intermediate Ki67, pre- or postmenopausal Genomic test on core biopsy + short-course ET 2-4 wk → Ki67 RS ≤25 + Ki67 ≤10%: ET (neoadjuvant or adjuvant) Ki67 >10% or RS >25: CT ET: AI (+ OFS if premenopausal) ADAPT: response spared CT only in pN0–1 with RS 12–25 c/pN2–3: this strategy has not been validated High risk or downstaging needed | e.g. cT3-4 N0-1, cN2-3, G3, high Ki67 Neoadjuvant CT: Anthracycline + taxane based, dose dense considered or NET ≥4-6 months: postmenopausal, ER-rich, low grade (Ki67-guided) Legend: endocrine-based, CT spared | chemotherapy Then escalate by stage and risk OFS-based ET and CDK4/6i also carry burden (toxicity, adherence, QoL) Abemaciclib 2 y | monarchE · ESMO-MCBS B Ribociclib 3 y | NATALEE · ESMO-MCBS A Olaparib 1 y | OlympiA, gBRCA · ESMO-MCBS A Extended ET and/or oral SERDs Clinical trials | e.g., NoLEEta Sparano JA, NEJM 2018 | Kalinsky K, NEJM 2021 | Piccart M, Lancet Oncol 2021 | Stein RC, JCO 2026 (ASCO) | Nitz UA, JCO 2022 | Gluz O, Ann Oncol 2026 | NRG-BR009 | Buus R, JNCI 2016 | Sestak I, BCRT 2019 | monarchE: Johnston SRD, JCO 2020 | NATALEE: Slamon D, NEJM 2024 | OlympiA: Tutt ANJ, NEJM 2021 | Loibl S, Ann Oncol 2024 | Andre F, JCO 2022 | AGO 2026 | NCCN Breast Cancer (25 Sep 2026) | ESMO-MCBS Scorecards (accessed Oct 2026) CTS5: Clinical Treatment Score post-5 years; BCI: Breast Cancer Index; ESMO-MCBS: ESMO Magnitude of Clinical Benefit Scale Mafalda Oliveira, MD PhD [Slide 4] Take-home messages 1 Stage carefully: anatomy sets the starting point Mammography and ultrasound of breast and axilla in everyone; MRI for selected cases. Biopsy and clip suspicious nodes before any systemic therapy. Systemic staging generally from stage IIB or with high-risk features or symptoms. 2 Upfront surgery plus a validated genomic score remains the standard for cT1-2 cN0-1 Postmenopausal pN0-1 with RS ≤25 or genomic low risk can safely omit chemotherapy (TAILORx, RxPONDER, MINDACT, OPTIMA). In premenopausal women the CT benefit persists and may partly reflect ovarian suppression: CT → OFS-based ET stays standard while NRG-BR009 reads out. 3 In vivo endocrine sensitivity provides valuable additional information over baseline biology Ki67 after 2-4 weeks of ET is prognostic (POETIC). Endocrine responders with pN0-1 and RS 12-25 may safely omit chemotherapy, while non-responders do worse despite dose-dense CT. 4 In young women, the endocrine therapy given matters as much as tumor biology ET responses in premenopausal women receiving AI + OFS match those from postmenopausal women on an AI. A short ET window works in routine care but needs reliable Ki67 and logistics. 5 ET-sensitive tumors rarely achieve pCR after chemotherapy, despite clinical high risk Whether CDK4/6i + ET can fill this gap is currently being investigated. Mafalda Oliveira, MD PhD

Disease-Free Survival [KM plot: Disease-Free Survival 0.0-1.0 vs Years from randomization 0-15+; dashed lines near 10 and 15 years] Letrozole 74.9% vs Placebo 72.1% (Δ = 2.8%) at 10-yr line Letrozole 61.6% vs Placebo 56.5% (Δ = 5.1%) at 15-yr line # Events Analysis | P | L | HR (95% CI) | P-value 10-yr | 505 | 435 | 0.85 (0.74-0.96) | 0.01 15-yr | 787 | 691 | 0.85 (0.76-0.94) | 0.002 Patients excluded: • 38 (no follow-up, or not at risk for DFS) • 22 (no clinical follow-up assessment) At risk (0, 5, 10, 15 years) Placebo 1954 1559 1125 651 Letrozole 1952 1601 1171 698 [NRG Oncology logo] Eleftherios P. Mamounas Survival table Endpoint: Disease-free survival (10-yr and 15-yr analyses) Placebo (P) | 1954 Letrozole (L) | 1952 HR: 10-yr: 0.85; 15-yr: 0.85 HR CI: 10-yr: 0.74-0.96; 15-yr: 0.76-0.94 P: 10-yr: 0.01; 15-yr: 0.002 Landmark rates: 10 years: Letrozole 74.9% | Placebo 72.1% 15 years: Letrozole 61.6% | Placebo 56.5%

[Slide 1] MADRID 2026 ESMO congress Target switching with conserved-payload novel ADCs in patients with metastatic breast cancer following prior ADC therapy (SWITCH): a prospective, open-label, Phase 2 platform trial Lei Fan1, Xi-Yu Liu1, Yun-Song Yang1, Fei Liang2, Chao Chen1, Yin Liu1, Fan Yang1, Shen Zhao1, Li Chen1, Wen-Juan Zhang1, Peng Ji1, Guan-Tian Lang1, Qian-Nan Liang5, Yu Shen5, Lei-Ping Wang3, Jin Zhang4, Zhong-Hua Wang1, Ke-Da Yu1, Yi-Zhou Jiang1, Zhi-Ming Shao1 October 26th, 2026 Affiliations: 1. Department of Breast Surgery, Fudan University Shanghai Cancer Center and Key Laboratory of Breast Cancer in Shanghai, Department of Oncology, Shanghai Medical College, Fudan University, Shanghai, China 2. Department of Biostatistics, Zhongshan Hospital, Fudan University, Shanghai, China 3. Department of Medical Oncology, Fudan University Shanghai Cancer Center and Department of Oncology, Shanghai Medical College, Fudan University, Shanghai, China 4. The Third Department of Breast Cancer, Tianjin Medical University Cancer Institute & Hospital, Tianjin, China 5. Department of Clinical Research & Development, Jiangsu Hengrui Pharmaceuticals Co., Ltd., Shanghai, China [Slide 2] Study design Target-switching ADCs after prior ADC in metastatic breast cancer (SWITCH) > SWITCH was a prospective, open-label, Phase 2 platform trial (NCT06649331). > To answer a critical question: is changing the monoclonal antibody target alone sufficient to reverse resistance? > Based on prior-ADC treatment, participants were randomized into one of four cohorts of novel target-switching ADCs: SHR-A1811 (Cohort A, HER2-ADC), SHR-A1921 (Cohort B, TROP-2-ADC), SHR-A2009 (Cohort C, HER3-ADC), or SHR-A2102 (Cohort D, Nectin-4-ADC). Patients with HER2-negative MBC - Prior ADC treatment - Measurable lesions (RECIST v 1.1) - For HR+/HER2-MBC: prior endocrine therapy with CDK4/6i R Cohort A (HER2-ADC) SHR-A1811: 4.8mg/kg, ivgtt, Q3W Cohort B (TROP-2-ADC) SHR-A1921: 3.0mg/kg, ivgtt, Q3W Cohort C (HER3-ADC) SHR-A2009: 8mg/kg, ivgtt, Q3W Cohort D (Nectin-4-ADC) SHR-A2102: 8mg/kg, ivgtt, Q3W Treatment until disease progression, intolerance, or withdrawal for other reasons Primary endpoint - ORR Secondary endpoint - DCR - PFS - OS - Safety Exploratory Endpoint - Biomarker analysis Statistical considerations Each cohort was planned to enroll approximately 10-30 patients and was monitored independently using a Bayesian predictive probability design. A cohort was deemed effective if the posterior probability (ORR > 20%) > 0.80. RECIST, Response Evaluation Criteria in Solid Tumors; ORR: objective response rate; PFS: progression-free survival; DCR: disease control rate; OS: overall survival. [Slide 3] Objective Response Rate (ORR) > Confirmed ORRs were consistently low across all cohorts, ranging from 0% to 8.3%. > Clinical responses only occurred after switching from MMAE-based ADC to TOP1i-based ADCs. | Cohort A (SHR-A1811) (N=10) (%) | Cohort B (SHR-A1921) (N=12) (%) | Cohort C (SHR-A2009) (N=12) (%) | Cohort D (SHR-A2102) (N=20) (%) PR, n (%) | 0 (0.0) | 1 (8.3) | 0 (0.0) | 1 (5.0) SD, n (%) | 8 (80.0) | 2 (16.7) | 5 (41.7) | 10 (50.0) PD, n (%) | 2 (20.0) | 9 (75.0) | 7 (58.3) | 9 (45.0) cORR, % (95% CI) | 0.0 (0.0-30.8) | 8.3 (0.2-38.5) | 0.0 (0.0-26.5) | 5.0 (0.1-24.9) DCR, % (95% CI) | 80.0 (44.4-97.5) | 25.0 (5.5-57.2) | 41.7 (15.2-72.3) | 55.0 (31.5-76.9) Cohort A (SHR-A1811) Posterior mean ORR: 4.5% 95% CrI: 0.0% - 21.7% Pr(ORR >= 20%) = 0.0324 Cohort B (SHR-A1921) Posterior mean ORR: 11.5% 95% CrI: 0.9% - 32.8% Pr(ORR >= 20%) = 0.1551 Cohort C (SHR-A2009) Posterior mean ORR: 3.8% 95% CrI: 0.0% - 18.5% Pr(ORR >= 20%) = 0.0193 Cohort D (SHR-A2102) Posterior mean ORR: 7.1% 95% CrI: 0.5% - 21.1% Pr(ORR >= 20%) = 0.0319 Waterfall legend: PR; SD; PD; TNBC; HR+/HER2-; Progression on prior HER2 ADC; Progression on prior TROP2 ADC; Progression on prior Nectin-4 ADC Best change from baseline (%); Individual patients * Prior ADC: disitamab vedotin (RC48, payload MMAE) Clinical activity was limited across cohorts Abbreviations: PR, partial response; SD, stable disease; PD, progressive disease; CI, confidence interval; cORR, confirmed objective response rate; DCR, disease control rate. [Slide 4] Progression-Free Survival (PFS) > mPFS was uniformly short ranging from 1.4 months to 3.0 months. > Consistent with ORR, patients previously treated with the microtubule inhibitor RC48 exhibited numerically longer PFS on their second ADC. Cohort | mPFS (95% CI) Cohort A (SHR-A1811) 3.0 (1.5 - 5.7) Cohort B (SHR-A1921) 1.4 (1.0 - 2.7) Cohort C (SHR-A2009) 1.5 (1.3 - 2.9) Cohort D (SHR-A2102) 2.7 (1.3 - 3.1) ADC2 PFS (probability) No. at risk Cohort A 10 2 0 0 Cohort B 12 1 1 0 Cohort C 12 0 0 0 Cohort D 20 1 0 0 ADC1 PFS (prior ADC) | ADC2 PFS (current ADC) PFS (months) T-DXd; SG; SHR-A1921; SHR-A2102; x toxicity; RC48; SHR-A1811; SKB264 Patients previously treated with RC48 exhibited numerically longer PFS on the second ADC (ADC2) than those previously exposed to other ADCs. Progression-free survival benefit was limited across cohorts Survival table Endpoint: ADC2 PFS Arm | N | Median | Median CI Cohort A (SHR-A1811) | 10 | 3.0 mo | 1.5-5.7 Cohort B (SHR-A1921) | 12 | 1.4 mo | 1.0-2.7 Cohort C (SHR-A2009) | 12 | 1.5 mo | 1.3-2.9 Cohort D (SHR-A2102) | 20 | 2.7 mo | 1.3-3.1

Conclusions ➤ Target switching with novel TOP1i-based ADCs showed limited efficacy in TOP1i-based ADC-refractory HER2-negative metastatic breast cancer. • Objective response rates ranging from 0% to 8.3% across cohorts • Median progression-free survival ranging from 1.4 to 3.0 months across cohorts ➤ SWITCH is the first prospective study exploring ADC switching in HER2-negative MBC. Payload cross-resistance may be a major determinant of post-ADC treatment failure. ➤ ADCs with novel, non-cross-resistant payloads are urgently needed, alongside the ongoing exploration of additional effective therapeutic strategies. Presenter: Lei Fan, MD

Switching target fails; switching payload works modestly Response rates after progression on a TOP1 ADC, across prospective studies BFv: Nectin-4-MMAE | 30% SWITCH B: TROP2-TOP1i | 8% SWITCH D: Nectin-4-TOP1i | 5% SWITCH C: HER3-TOP1i | 0% SWITCH A: HER2-TOP1i | 0% HER3-DXd Part B | 5% SATEEN: SG + trastuzumab (HER2+) | 4% Same target: Nectin-4, different payload Caveat- different trials, lines HER3-DXd Part B: 422RO and SATEEN: LBA4, ESMO Breast 2026. HER3-DXd ORR in ADC-naive Part A: 39% (HR+), 25% (TNBC). Sherene Loi, Peter MacCallum Cancer Centre

[Slide 1] SWITCH: changing the target, but keeping the payload does not work Four novel TOP1 ADCs after prior ADC (94% TOP1-pretreated): confirmed ORR 0-8% IMPORTANT NEGATIVE STUDY Cohort | Target (all TOP1i) | n | cORR | DCR A SHR-A1811 | HER2 | 10 | 0% | 80% B SHR-A1921 | TROP2 | 12 | 8.3% | 25% C SHR-A2009 | HER3 | 12 | 0% | 42% D SHR-A2102 | Nectin-4 | 20 | 5.0% | 55% Median PFS 1.4-3.0 mo across cohorts. Both responders had switched from an MMAE ADC (RC48; n=3). Caveats • Allocation by prior treatment, not true randomisation • Small cohorts (n=10-20); mixed HR+ and TNBC; 11% ECOG 2 • Payload-switch signal = 2 patients • ATM/ctDNA: requires validation and control for CHIP • ADC1 PFS did not predict ADC2 PFS Sherene Loi, Peter MacCallum Cancer Centre [Slide 2] Four abstracts with one main theme What should follow a TOP1-payload ADC in metastatic breast cancer? Abstract | Target | Payload | Population | Prior TOP1-ADC | Topline Result 4639RO OptiTROP-Breast01 sac-TMT vs chemo (Ph3) | TROP2 | TOP1i | mTNBC, ≥2 prior lines | 0% (ADC-naive) | OS HR 0.60 4RO SWITCH 4 novel ADCs (Ph2 platform) | HER2 / TROP2 / HER3 / Nectin-4 | TOP1i | HER2-neg MBC | 94% | ORR 0-8% 4640RO Bulumtatug fuvedotin (single-arm Ph2) | Nectin-4 | MMAE | mTNBC | 100% | ORR 30% 5RO SYS6043 (Ph1/2) | B7-H3 | TOP1i | All subtypes | Not reported | ORR 54-75% TOP1i topoisomerase-I inhibitor payload MMAE microtubule-inhibitor payload Sherene Loi, Peter MacCallum Cancer Centre [Slide 3] Bulumtatug fuvedotin: activity after a TOP1 ADC Nectin-4-MMAE ADC in mTNBC, 100% prior TOP1 ADC (48% TROP2, 33% HER2), n=60; ORR =30% [Waterfall plot with ORR/DCR table: values [illegible] at this resolution] Med PFS 3.7 mo Med DoR 6.8 mo Comments • Responders appear higher among Nectin-4-positive tumours but number of negative tumors was low (65% v22%) • ORR after prior ICB (35%), TROP2 (28%) vs HER2 ADC (30%): broadly the same • Neuropathy (MMAE class effect) G1/2 ~20%; • 45% treatment interruptions largely due to neutropenia • Adverse event profile NOT similar to Enfortumab Vedotin (no rash) • Phase III study ongoing: Met. TNBC post 1L ADC vs TPC Sherene Loi, Peter MacCallum Cancer Centre

Patient Disposition Between Oct 2024 and Dec 2025, 54 patients with HER2-negative, ADC-refractory MBC were enrolled and randomized into the SWITCH trial 58 HER2-negative patients screened for eligibility 4 excluded 3 did not meet inclusion criteria 1 withdrew consent 54 randomly assigned 10 Cohort A HER2 ADC: SHR-A1811 12 Cohort B Trop2 ADC: SHR-A1921 12 Cohort C HER3 ADC: SHR-A2009 20 Cohort D Nectin-4 ADC: SHR-A2102 10 progressive disease 0 treatment ongoing 12 progressive disease 0 treatment ongoing 11 progressive disease 1 discontinued for vomiting 20 progressive disease 0 treatment ongoing 10 included in FAS and SAS 12 included in FAS and SAS 12 included in FAS and SAS 20 included in FAS and SAS Presenter: Lei Fan, MD • Abstract published at DCO, March 14, 2026 • Updated data presented at DCO, July 31, 2026 • Median follow-up: 8.6 months (range, 4.7–14.0) at data cutoff • With all patients included in full and safety analyses

Post-hoc analysis > No significant correlation between ADC1 PFS and ADC2 PFS, nor between ADC2 PFS and the ADC1-to-ADC2 washout interval. ADC2 PFS (months) vs ADC1 PFS (months): R² = 0.000, P = 0.970 ADC2 PFS (months) vs Washout interval (months): R² = 0.000, P = 0.918 Cohort A; Cohort B; Cohort C; Cohort D > By a post-hoc 5-month ADC1 PFS cutoff ADC1 PFS < 5 months (n = 17) ADC1 PFS ≥ 5 months (n = 37) ADC1 PFS | ADC2 PFS Cohort A (SHR-A1811); Cohort B (SHR-A1921); Cohort C (SHR-A2009); Cohort D (SHR-A2102) PFS (months) ADC1: T-DXd; SG; SHR-A1921; SHR-A2102; RC48; SHR-A1811; SKB264; x toxicity Patients with ADC1 PFS <5 months: ADC2 PFS >3 months rate, numerically higher (35.3% vs 21.6%; Fisher's exact P = 0.33) Patients progressing rapidly on their first ADC (< 5 months) were more likely to achieve a PFS2 >3 months on ADC2, whereas those with PFS1 ≥ 5 months derived less apparent benefit from subsequent ADCs.

Background UNMET NEED > Antibody-drug conjugates (ADCs) have become integral to the treatment of metastatic breast cancer (MBC)1-3. > Resistance to ADCs has emerged as a major therapeutic challenge in MBC4-5. > The optimal sequencing strategy after progression on prior ADC therapy remains unclear6-7. NOVEL ADCs > SHR-A1811 (anti-HER2)8,9, SHR-A1921 (anti-Trop-2)10, SHR-A2009 (anti-HER3)11, and SHR-A2102 (anti-Nectin-4)12, are four novel ADCs with a TOP1-inhibitor payload, all of which have demonstrated considerable efficacy in the treatment of advanced breast Cancer. The SWITCH trial assessed the efficacy and safety of a "target switching" strategy using novel ADCs in patients with ADC-refractory MBC Presenter: Lei Fan, MD

[Slide 1] Do these pharmacological differences translate into clinically meaningful advantages as compared to approved ADCs? We do not know, because we do not have studies with head-to-head ADC comparison in this setting -Pivotal trials should generate evidence for regulatory approval by comparing new ADCs with the current SoC -T-DXd is now a standard treatment option for patients with HR+/HER2-low ABC after ET or following chemotherapy (where approved) -Today, we need head-to-head trials comparing new ADCs against ADCs that represent the current standard of care to: - establish whether pharmacological innovations translate into clinical benefits -identify biomarkers that guide optimal patient–ADC matching Head-to-Head ADC Trials We Need Today Randomized Head-to-Head ADC Trial in Biomarker-Selected HR+/HER2– MBC Study population: HR+/HER2– metastatic breast cancer; Prior endocrine therapy + CDK4/6 inhibitor; Measurable or evaluable disease; Adequate tissue and blood for biomarker analysis Biomarker selection / stratification: Quantitative target expression, internalization / trafficking markers, TOP1 payload sensitivity, resistance signature N=~300 R 1:1 ADC A Biomarker-matched patients ADC B Biomarker-matched patients Stratification: Biomarker-defined subgroup; Prior chemotherapy for MBC (yes/no); Visceral disease (yes/no) Primary endpoint PFS Key secondary endpoints • ORR • OS • Safety / tolerability • Patient-reported outcomes • Biomarker-treatment interaction Figure created using ChatGPT (OpenAI) Barbara Pistilli - Invited Discussant LBA23, LBA13 and LBA27 [Slide 3] CHAIRS: KOMAL JHAVERI, BARBARA PISTILLI Why a quantitative assessment of the target ? Target expression level and ADC activity HER2: [KM curve] HR+ on Primary Tumor (n=126); Median PFS (m): COHORT 1 11.0 [6.9;14.4]; COHORT 2 6.9 [5.5;8.7]; COHORT 3 4.5 [1.5;6.9] Number at risk: COHORT 1 44 37 32 23 12 6 2 0; COHORT 2 57 47 32 17 15 6 2 0; COHORT 3 25 16 7 3 1 0 0 0 Greater efficacy of T-DXd in patients with HER2-high and HER2-low breast cancer TROP2: [KM curve] SG - Trop-2 high / medium / low; TPC - Trop-2 high / medium / low [events/censored values illegible] Greater efficacy (mPFS, ORR) of SG in patients with greater TROP2 expression Key gaps to fill in to support ADCs' development as Precision Medicine [Table: Priority | Challenge/Gap | Why It Matters | Current Limitation | Potential Directions] 1 Antigen expression and intratumoral and intertumoral heterogeneity | a minimum target threshold necessary to have some antitumor benefit is not defined for the majority of ADCs; spatial and temporal heterogeneity reduces effective target engagement | IHC-based assays do not provide a quantitative target assessment and intra- and intertumoral variability | fully quantitative methods for target assessment on histology slides; spatial biology tools (multiplex imaging, digital pathology); immuno-PET imaging to assess whole-body target heterogeneity and longitudinal assessment of antigen expression 2 ADC pharmacology in tumors and TME | ADC activity depends on: 1. the tumor: antigen internalization kinetics, intracellular trafficking, and payload release 2. the TME that influences ADC penetration, payload diffusion, and immune modulation | IHC does not capture spatial heterogeneity or internalization; limited tools to assess ADC distribution, intracellular processing, payload release, and immune effects within tumors | assays capturing cytoplasmic/internalized targets; on-treatment biopsies with spatial transcriptomics/proteomics to map intratumoral distribution; integration of immune profiling 3 Resistance mechanisms | resistance limits the durability of response and long-term clinical benefit | limited clinical data; lack of longitudinal sampling and validated resistance biomarkers | longitudinal tumor and liquid biopsies; identification of resistance pathways; develop strategies to overcome resistance (e.g., novel payload classes and efflux pump inhibitors) 4 Biomarker selection | target expression alone is insufficient to predict ADC response and the durability of benefit | most approved ADCs lack validated predictive biomarkers; reliance on IHC with limited biological resolution | develop composite biomarkers integrating quantitative and spatial target expression, genomics, proteomics, and immune contexture; leverage deep learning to integrate multi-omics data 5 Validation of standardized biomarker assays | reproducible diagnostics are essential for clinical implementation and regulatory approval | lack of harmonized companion diagnostics and standardized thresholds across trials | validate assay-specific CDx platforms with reproducible scoring systems and clinically meaningful cutoffs across clinical trials with ADC 6 Patient selection strategies | optimal benefit requires integrating multiple biological determinants | most trials rely on single biomarkers and do not incorporate multidimensional tumor features | implement in clinics composite selection scores integrating antigen expression, internalization, genomics, and TME features to select optimal ADC for each patient and fulfill a precision medicine development of ADCs 7 Design of new ADCs | improve efficacy and reduce toxicity through optimized selection of the antibody, linker-cleavage technology, and payload | suboptimal patient selection, inconsistent efficacy signals, and avoidable toxicity | design of next-generation ADCs through the selection of antibodies, linker-cleavage technologies, and payloads based on multidimensional biomarkers While target expression may correlate with ADC activity, associations are weak, reflecting the limited analytical precision of semi-quantitative IHC scoring Mosele et al, Nat Medicine 2023; Bardia et al, Ann Oncol 2021; Pistilli et al, Cancer Cell 2026; Liu et al, Nat Rev Clin Oncol 2026 Barbara Pistilli - Invited Discussant LBA23, LBA13 and LBA27

[Slide 1] The relevant unit of decision is the whole treatment journey FIRST-LINE FORK → ACTIVE MAINTENANCE → CNS TRAJECTORY → POST–T-DXd SEQUENCING → LONG RESPONDERS Lifetime patient benefit Disease control across successive therapies + cumulative toxicity + CNS control + access to subsequent treatment + patient-reported outcomes Antonino Musolino, MD, PhD, MSc [Slide 2] First-line therapy is now a choice between two therapeutic strategies ESMO 2026 first-line algorithm: THP [I,A; MCBS 4]; T-DXd + pertuzumab [I,A; MCBS 2] T-DXd + pertuzumab: Continuous HER2 inhibition + continuous cytotoxic payload | Highest first-line efficacy benchmark | No planned maintenance transition VS THP → HP-based maintenance: Finite taxane induction | Then lower-burden HER2 blockade | Maintenance can be biologically intensified Tolaney SM et al. N Engl J Med. 2026; Swain SM et al. Lancet Oncol. 2020; de Azambuja E, et al. Ann Oncol. 2026. Antonino Musolino, MD, PhD, MSc [Slide 4] Clinical framework for management after progression on T-DXd Progression after frontline T-DXd Systemic progression (without CNS progression): - Tucatinib-naïve? consider tucatinib–trastuzumab–capecitabine - If not suitable / previously exposed: consider T-DM1, HER2-directed chemo, alternate ADC, trial - Reassess biology when feasible: HER2/HR status; distinguish toxicity from resistance Isolated CNS progression (extracranial disease controlled): - First isolated CNS event? consider SRS/surgery and continue effective systemic therapy - Later / multifocal CNS event: consider CNS-active systemic switch and clinical trial - MRI surveillance and symptoms: neurocognition, steroid need, WBRT avoidance when possible CNS + extracranial progression (concurrent progression): - Systemic treatment change: prioritise CNS-active therapy if tucatinib-naïve - If prior tucatinib exposure: alternate HER2 strategy, payload/target switch, clinical trial - Manage residual toxicity: ILD, cytopenias, performance status, patient preference Assess and individualise at progression: prior T-DM1/tucatinib exposure • residual T-DXd toxicity • HER2/HR reassessment • performance status • patient goals • trial access Evidence categories used in the framework: direct prospective evidence | indirect / retrospective / extrapolated | investigational / clinical trial preferred Antonino Musolino, MD, PhD, MSc

Exploratory analyses Key findings 52 evaluable patients with baseline ctDNA profilin. Stratified by ADC2 PFS status at 3 months. > SYNE1, CACNA1C, ALK, ATM showed numerical differences; exploratory candidate biomarkers. > ATM mutation was associated with longer ADC2 PFS. > ATM mutation associated with reduced angiogenesis and vascular abnormality scores. PFS < 3 months | PFS ≥ 3 months Gene | Mut/WT | HR (95% CI) | P CACNA1C | 6/46 | 3.76 (1.47-9.56) | 0.006 ALK | 7/45 | 2.71 (1.17-6.26) | 0.020 IGF1R | 6/46 | 1.68 (0.70-4.05) | 0.248 GNAS | 6/46 | 1.67 (0.70-4.02) | 0.250 SYNE1 | 16/36 | 1.65 (0.89-3.04) | 0.112 CBL | 6/46 | 1.60 (0.67-3.82) | 0.288 NOTCH2 | 8/44 | 0.66 (0.31-1.42) | 0.284 EGFR | 7/45 | 0.65 (0.29-1.45) | 0.292 ABL1 | 5/47 | 0.58 (0.23-1.47) | 0.247 ATM | 9/43 | 0.41 (0.18-0.93) | 0.032 Hazard ratio Cohen's d (ATM mut - ATM WT) | Nominal P Immune suppression | 0.431 Endocytosis lysosome | 0.392 ECM remodeling | 0.462 EMT plasticity | 0.518 Angiogenesis and vascular abnormality | 0.034* DDR | 0.802 Efflux detoxification | 0.878 Apoptotic escape | 0.389 ATM mutation and microenvironmental features emerged as candidate response correlates

[Slide 1] 10:15 - 11:45 Rapid oral 2: Breast cancer, advanced stage CHAIRS: ANTONIO LLOMBART CUSSAC, SHERENE LOI, PEDRAM RAZAVI Baseline Characteristics 1.25 mg/kg, D1,D8,Q3W (n=60) Age Median (range) | 50.5 (32,73) ECOG PS n(%) | 0 | 11 (18) | 1 | 49 (82) Metastatic sites n(%) | Visceral metastasis | 39 (65) | Liver metastasis | 20 (33) | Brain metastasis | 4 (7) Nectin-4 expression | Positive (TPS ≥1%) | 39 (65) | Negative (TPS <1%) | 13 (22) | unknown | 8 (13) 1.25 mg/kg, D1,D8,Q3W (n=60) Prior lines of therapy n(%) | 1 line | 10 (17) | 2 lines | 25 (42) | 3 lines | 22 (37) | 4 lines | 3 (5) Prior treatment n(%) | Taxane | 60 (100) | Anthracycline | 46 (77) | Platinum | 21 (35) | Checkpoint in hibitor | 21 (35) | HER-2 ADC* | 20 (33) | TROP2 ADC* | 29 (48) | Other ADC | 12 (20) *One patient received both a HER2-targeted ADC and a TROP2-targeted ADC. ECOG, Eastern Cooperative Oncology Group; PS, performance status; TPS, Tumor Proportion Score Dr Yanchun Meng Yanchun Meng Bulumtatug fuvedotin (BFv, 9MW2821), a nectin-4-directed antibody-drug conjugate (ADC), in patients with locally advanced or metastatic triple negative breast cancer (la/mTNBC) previously treated with topoisomerase I inhibitor (TopI)-based ADCs: Resu... ALICANTE AUDITORIUM - HALL 6 [Slide 2] 10:15 - 11:45 Rapid oral 2: Breast cancer, advanced stage CHAIRS: ANTONIO LLOMBART CUSSAC, SHERENE LOI, PEDRAM RAZAVI Efficacy 1.25 mg/kg, D1,D8,Q3W (N= 60) ORR (95% CI) | 30.0 (18.85,43.21) DCR (95% CI) | 80.0 (67.67, 89.22) Change from Baseline (%) Best Overall Response: Complete Response; Partial Response; Stable Disease; Progressive Disease Nectin-4 Status: Positive; Negative Swimmer plot Individual Patients Time (Months) First radiologic response (CR or PR) Disease progression Dr Yanchun Meng Median follow up time 10.7 months ORR, objective response rate; CI, confidence interval; DCR, disease control rate; CR, complete response; PR, partial response; Yanchun Meng Bulumtatug fuvedotin (BFv, 9MW2821), a nectin-4-directed antibody-drug conjugate (ADC), in patients with locally advanced or metastatic triple negative breast cancer (la/mTNBC) previously treated with topoisomerase I inhibitor (TopI)-based ADCs: Resu... ALICANTE AUDITORIUM - HALL 6

Why a quantitative assessment of the target ? IHC: semi-quantitative assessment of target expression [IHC micrographs] 1+ score | 2+ score Score 1+ Membranous staining not visible at low magnification Identified at Gx10 - Gx20 Clearly visible at high magnification Incomplete weak >10% Courtesy of Dr Magali Lacroix-Triki Quantitative high-sensitivity HER2 and TROP2 multiplex immunofluorescence assay Immunofluorescence Assay: Tyramide signal amplification; HRP-polymer 2' antibodies; Cy5; Biotin/streptavidin-AF750; Anti-HER2; Anti-TROP2; DAPI; Anti-pan-cytokeratin - AF488 Molecular Compartmentalization: Tumor (CK+) | HER2 and TROP2 in CK Qymia score = Σ Intensity / Σ Compartment area (µm2) Cell Line Array Standard: HER2 TROP2 DAPI CK (off) HER2 & TROP2 quantification LC-MS/MS (protein per cell); amol/cell to amol/mm2; Cell segmentation Convert Qymia Score to amol/mm2 HER2, Qymia Score vs HER2 amol/mm2 (cell lines): R2 = 0.906, y = 0.118x TROP2, Qymia Score vs TROP2 amol/mm2 (cell lines): R2 = 0.991, y = 0.008x Report Quantitative Measurements: HER2 1840 amol/mm2, 82nd Percentile; TROP2 890 amol/mm2, 13th Percentile Objective quantification of HER2/TROP2 expression (attomoles/mm2): assessment of both absolute target levels and relative expression within similar BC population Robbins et al, JCO Precsion Oncol 2025

[Slide 1] 08:30 - 10:00 Proffered paper 2: Breast cancer, early and advanced stage CHAIRS: SARA HURVITZ, SEOCK-AH IM KEYNOTE-B49 (NCT04895358) Key Eligibility Criteria • HR+/HER2− locally recurrent inoperable or metastatic breast cancer • Tumor PD-L1 CPS ≥1 • Candidate for chemotherapy (no prior chemotherapy for metastatic breast cancer) • Progression on ≥1 prior line of ET (with a CDK4/6i) • If no CDK4/6i was given, relapse within 24 mo of primary surgery while on adjuvant ET and progression within 6 mo of starting first line of ET R 1:1 N = 342* → Pembrolizumab 200 mg Q3Wa + Chemotherapyb → Placeboa + Chemotherapyb Endpoints Primary • PFS per RECIST v1.1 by BICR in participants with PD-L1 CPS ≥1 Key Secondary • OS in participants with PD-L1 CPS ≥1 Secondary • PFS per RECIST v1.1 by BICR in participants with PD-L1 CPS ≥10 • OS in participants with PD-L1 CPS ≥10 • ORR, DCR, and DOR per RECIST v1.1 by BICR in participants with PD-L1 CPS ≥1 and ≥10 • Safety Stratification Factors • Tumor PD-L1 expression (CPS 1–9 vs CPS ≥10) • Presence of visceral disease (yes vs no) • Planned on-study chemotherapy (taxanes vs liposomal doxorubicin vs capecitabine) *The planned study sample size was reduced from 800 to 340 after a protocol amendment due to low prevalence of PD-L1–positive tumors BICR, blinded independent central review. aParticipants received pembrolizumab or placebo for up to 35 cycles or until PD, unacceptable toxicity, or participant withdrawal. bPhysician's choice of paclitaxel 90 mg/m2 or nab-paclitaxel 100 mg/m2 on days 1, 8, and 15 Q4W; liposomal doxorubicin 50 mg/m2 Q4W; or capecitabine 1000 mg/m2 twice daily on days 1–14 Q3W. ALICANTE AUDITORIUM - HALL 6 [Slide 2] 08:30 - 10:00 Proffered paper 2: Breast cancer, early and advanced stage CHAIRS: SARA HURVITZ, SEOCK-AH IM Baseline Characteristics Columns: Pembro + chemo (n = 171) | Placebo + chemo (n = 171) Age, median (range), y: 58 (27–85) | 57 (21–83) Geographic region North America: 30 (17.5) | 21 (12.3) Europe: 71 (41.5) | 72 (42.1) Asia: 38 (22.2) | 38 (22.2) Rest of world: 32 (18.7) | 40 (23.4) PD-L1 CPS [highlighted] 1–9: 135 (78.9) | 135 (78.9) ≥10: 36 (21.1) | 36 (21.1) ECOG PS 0: 96 (56.1) | 106 (62.0) 1: 75 (43.9) | 65 (38.0) Number of metastatic sites <3: 86 (50.3) | 91 (53.2) ≥3: 85 (49.7) | 80 (46.8) On-study chemotherapy Paclitaxel: 28 (16.4) | 19 (11.1) Nab-paclitaxel: 41 (24.0) | 49 (28.7) Liposomal doxorubicin: 10 (5.8) | 9 (5.3) Capecitabine [highlighted]: 91 (53.2) | 91 (53.2) Not treated: 1 (0.6) | 3 (1.8) Presence of visceral metastases Yes [highlighted]: 145 (84.8) | 146 (85.4) No: 26 (15.2) | 25 (14.6) ER status <1%: 2 (1.2) | 4 (2.3) 1%–10%: 11 (6.4) | 7 (4.1) >10% [highlighted]: 158 (92.4) | 160 (93.6) HER2 status 0 by IHC: 38 (22.2) | 46 (26.9) 1+ by IHC: 84 (49.1) | 67 (39.2) 2+ by IHC (FISH negative): 49 (28.7) | 58 (33.9) Menopausal statusa Premenopausal: 25 (14.6) | 26 (15.2) Post menopausal: 143 (83.6) | 143 (83.6) Prior CDK4/6 therapy [highlighted]: 170 (99.4) | 169 (98.8) Prior ET for metastatic disease 0: 0 | 1 (0.6) 1: 60 (35.1) | 67 (39.2) 2: 94 (55.0) | 79 (46.2) ≥3: 17 (9.9) | 24 (14.0) Prior taxanes in the neoadjuvant/adjuvant setting: 91 (53.2) | 88 (51.5) FISH, fluorescence in situ hybridization; IHC, immunohistochemistry. Data are n (%) unless otherwise specified. aFemale participants only. Data cutoff, Mar 16, 2026. ALICANTE AUDITORIUM - HALL 6 [Slide 3] 08:30 - 10:00 Proffered paper 2: Breast cancer, early and advanced stage CHAIRS: SARA HURVITZ, SEOCK-AH IM PFS by BICR in Participants with PD-L1 CPS ≥1 (Primary Endpoint) IA: Primary PFS Analysis 12-mo rate (95% CI) 26.3% (19.1%–34.1%) 20.3% (14.0%–27.4%) 24-mo rate (95% CI) 15.0% (9.0%–22.5%) 4.3% (1.6%–9.3%) No. at risk (Time, mo: 0 3 6 9 12 15 18 21 24 27 30 33 36 39) Pembro + chemo: 171 131 95 43 30 24 18 11 8 4 2 2 1 0 Placebo + chemo: 171 126 76 42 26 19 8 2 2 1 0 0 0 0 IA | Events | Median (95% CI), mo | HR (95% CI) Pembro + chemo | 72.5% | 6.7 (6.2–8.3) | 0.74 (0.58–0.95) P = 0.0089^b Placebo + chemo | 81.3% | 6.0 (4.2–8.1) FA: Updated Analysis^a 12-mo rate (95% CI) 26.3% (19.1%–34.1%) 20.3% (14.0%–27.4%) 24-mo rate (95% CI) 14.1% (8.5%–21.1%) 3.5% (1.1%–8.2%) No. at risk (Time, mo: 0 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 51 54) Pembro + chemo: 171 131 95 43 30 24 20 16 13 8 4 2 1 1 1 0 0 0 0 Placebo + chemo: 171 126 76 42 26 19 8 4 3 2 0 0 0 0 0 0 0 0 0 FA | Events | Median (95% CI), mo | HR (95% CI) Pembro + chemo | 74.3% | 6.7 (6.2–8.3) | 0.74 (0.58–0.95) Placebo + chemo | 82.5% | 6.0 (4.2–8.1) Median follow-up, 29.1 months (range, 18.1–49.4) for IA (data cutoff, Sep 22, 2025) and 34.9 months (range, 23.9–55.1) for FA (data cutoff, Mar 16, 2026). Median PFS based on the Kaplan-Meier method. HRs (95% CIs) based on a Cox regression model with treatment as a covariate and the Efron method of tie handling stratified by the randomization stratification factors. ^aUpdated analysis was descriptive. ^bMet the prespecified statistical criterion of α = 0.025. ALICANTE AUDITORIUM - HALL 6 Survival table Endpoint: PFS by BICR, PD-L1 CPS ≥1 (primary endpoint) — IA (DCO Sep 22, 2025) and FA updated descriptive analysis (DCO Mar 16, 2026) Arm | N | Median | Median CI Pembro + chemo (IA) | 171 | 6.7 | 6.2–8.3 Placebo + chemo (IA) | 171 | 6.0 | 4.2–8.1 Pembro + chemo (FA) | 171 | 6.7 | 6.2–8.3 Placebo + chemo (FA) | 171 | 6.0 | 4.2–8.1 HR: 0.74 HR CI: 0.58–0.95 P: 0.0089 (IA; met prespecified criterion α=0.025); FA HR 0.74 (0.58–0.95) descriptive Landmark rates: 12 mo (IA): Pembro + chemo 26.3% (CI 19.1%–34.1%) | Placebo + chemo 20.3% (CI 14.0%–27.4%) 24 mo (IA): Pembro + chemo 15.0% (CI 9.0%–22.5%) | Placebo + chemo 4.3% (CI 1.6%–9.3%) 12 mo (FA): Pembro + chemo 26.3% (CI 19.1%–34.1%) | Placebo + chemo 20.3% (CI 14.0%–27.4%) 24 mo (FA): Pembro + chemo 14.1% (CI 8.5%–21.1%) | Placebo + chemo 3.5% (CI 1.1%–8.2%) [Slide 4] 08:30 - 10:00 Proffered paper 2: Breast cancer, early and advanced stage CHAIRS: SARA HURVITZ, SEOCK-AH IM PFS by BICR in Participants with PD-L1 CPS ≥10 (Secondary Endpoint) IA | Events | Median (95% CI), mo | HR (95% CI) Pembro + chemo | 80.6% | 7.5 (4.2–10.3) | 0.59 (0.35–0.98) Placebo + chemo | 91.7% | 4.4 (4.0–8.3) 12-mo rate (95% CI) 23.6% (11.1%–38.7%) 9.3% (2.4%–22.2%) No. at risk (Time, mo: 0 3 6 9 12 15 18 21 24 27 30 33 36 39) Pembro + chemo: 36 28 21 12 8 8 7 5 4 1 1 1 0 0 Placebo + chemo: 36 27 15 7 3 1 1 0 0 0 0 0 0 0 Median follow-up, 29.1 months (range, 18.1–49.4). Median PFS based on the Kaplan-Meier method. HR (95% CI) based on a Cox regression model with treatment as a covariate and the Efron method of tie handling stratified by the randomization stratification factors. Data cutoff, Sep 22, 2025. ALICANTE AUDITORIUM - HALL 6 Survival table Endpoint: PFS by BICR, PD-L1 CPS ≥10 (secondary endpoint), IA Arm | N | Median | Median CI Pembro + chemo | 36 | 7.5 | 4.2–10.3 Placebo + chemo | 36 | 4.4 | 4.0–8.3 HR: 0.59 HR CI: 0.35–0.98 Landmark rates: 12 mo: Pembro + chemo 23.6% (CI 11.1%–38.7%) | Placebo + chemo 9.3% (CI 2.4%–22.2%)

Treatment Disposition of All Randomized Participants at Final Analysis 1767 patients screened 1425 patients did not meet eligibility criteria^a 342 participants randomized (19.4%) Pembro + Chemo • 171 allocated • 170 Started treatment • 3 Completed treatment • 160 Discontinued treatment • 145 Progressive disease^b • 9 AE • 5 Participant withdrawal • 1 Physician decision • 7 Ongoing Analysis Populations • ITT population: n = 171 • Safety-evaluable population: n = 170 Placebo + Chemo • 171 allocated • 168 Started treatment • 1 Completed treatment • 164 Discontinued treatment • 154 Progressive disease^b • 7 AE • 3 Participant withdrawal • 3 Ongoing Analysis Populations • ITT population: n = 171 • Safety-evaluable population: n = 168 • Median follow-up (time from randomization to data cutoff) – IA (data cutoff, Sep 22, 2025): 29.1 months (range, 18.1–49.4) – FA (data cutoff, Mar 16, 2026): 34.9 months (range, 23.9–55.1) ^a59% due to negative PD-L1 expression. ^bIncludes clinical progression and radiographic progression (initial).

Proffered Paper session LBA51 - Results from KEYNOTE-B49: A phase 3, randomized, double-blind study of pembrolizumab (pembro) vs placebo (pbo) plus chemotherapy (CT) in participants (pts) with HR+/HER2− advanced breast cancer (ABC) Speakers Hope S. Rugo (Duarte, United States of America) Lecture Time 08:30 - 08:40 Abstract Background After failure of endocrine-based therapy (ET), patients with HR+/HER2− ABC have poor outcomes. We report results of the phase 3 KEYNOTE-B49 (NCT04895358) study of pembro + CT vs pbo + CT in pts with PD-L1–positive HR+/HER2− ABC. Methods Pts with locally recurrent inoperable or metastatic HR+/HER2− ABC with PD-L1 CPS ≥1 who had prior ET and were candidates for first-line CT were randomized to pembro 200 mg or pbo Q3W for up to 35 cycles + CT (paclitaxel, nab-paclitaxel, liposomal doxorubicin, or capecitabine). The primary endpoint was PFS per RECIST v1.1 by BICR in pts with PD-L1 CPS ≥1. The key secondary endpoint was OS in pts with PD-L1 CPS ≥1. Following a protocol amendment, the sample size was reduced due to low PD-L1 prevalence. Results 342 pts were randomized. Median age was 57 y; 99.4% of pts in the pembro + CT arm and 98.8% in the pbo + CT arm had prior CDK4/6i. At primary analysis (data cutoff, Sep 22, 2025; median follow-up, 29.1 mo), pembro + CT showed statistically significant improvement in PFS vs pbo + CT (HR, 0.74; 95% CI, 0.58–0.95; P = 0.0089; table). In pts with PD-L1 CPS ≥10, the HR for PFS (0.59; 95% CI, 0.35–0.98) favored pembro + CT (median [95% CI] PFS, 7.5 [4.2–10.3] mo in the pembro + CT arm and 4.4 [4.0–8.3] mo in the pbo + CT arm). At final analysis (data cutoff, Mar 16, 2026; median follow-up, 34.9 mo), PFS benefit was maintained (HR, 0.74; 95% CI, 0.58–0.95), but there was no significant difference in OS (HR, 1.04; 95% CI, 0.81–1.33; P = 0.6185; table). Grade ≥3 treatment-related AEs occurred in 37.6% in the pembro + CT arm and 32.7% in the pbo + CT arm; 11.2% and 5.4% discontinued treatment due to treatment-related AEs, respectively. Conclusions In pts with PD-L1 CPS ≥1 HR+/HER2− ABC, pembro + CT demonstrated a statistically significant improvement in PFS, indicating a 26% risk reduction, and a manageable safety profile. A numerical effect favoring pembro + CT for PFS was observed in pts with PD-L1 CPS ≥10. Table: LBA51 | Pembro + CT | Pbo + CT Primary analysis for PFS PD-L1 CPS ≥1a Events, n/N | 124/171 | 139/171 Median (95% CI), mo | 6.7 (6.2–8.3) | 6.0 (4.2–8.1) HR (95% CI) | 0.74 (0.58–0.95); P = 0.0089 Final analysis for OS PD-L1 CPS ≥1a Events, n/N | 126/171 | 124/171 Median (95% CI), mo | 21.0 (18.9–22.9) | 22.2 (17.6–26.3) HR (95% CI) | 1.04 (0.81–1.33); P = 0.6185 aMultiplicity was controlled at α = 0.025 (1-sided)

[Slide 1] [Session banner] 08:30 - 10:00 Proffered paper 2: Breast cancer, early and advanced stage CHAIRS: SARA HURVITZ, SEOCK-AH IM OS in Participants with PD-L1 CPS >=10 (Secondary Endpoint) FA | Events | Median (95% CI), mo | HR (95% CI) Pembro + chemo | 69.4% | 22.9 (18.2-33.4) | 0.82 (0.47-1.43) Placebo + chemo | 77.8% | 22.5 (12.7-30.4) | Y-axis: OS, %; X-axis: Time, mo No. at risk (0/6/12/18/24/30/36/42/48/54): Pembro + chemo: 36 / 32 / 29 / 26 / 17 / 9 / 3 / 2 / 1 / 0 Placebo + chemo: 36 / 32 / 26 / 21 / 15 / 11 / 4 / 3 / 1 / 0 Median follow-up, 34.9 months (range, 23.9-55.1). Median OS based on the Kaplan-Meier method. HRs (95% CIs) based on a Cox regression model with treatment as a covariate and the Efron method of tie handling stratified by the randomization stratification factors. Data cutoff, Mar 16, 2026. [Speaker video] Hope Rugo Results from KEYNOTE-B49: A phase 3, randomized, double-blind study of pembrolizumab (pembro) vs placebo (pbo) plus chemotherapy (CT) in participants (pts) with HR+/HER2- advanced breast cancer (ABC) ALICANTE AUDITORIUM - HALL 6 Survival table Endpoint: OS, PD-L1 CPS >=10 (final analysis) Arm | N | Median | Median CI Pembro + chemo | 36 | 22.9 | 18.2-33.4 Placebo + chemo | 36 | 22.5 | 12.7-30.4 HR: 0.82 HR CI: 0.47-1.43 [Slide 2] [Session banner] 08:30 - 10:00 Proffered paper 2: Breast cancer, early and advanced stage CHAIRS: SARA HURVITZ, SEOCK-AH IM OS in Participants with PD-L1 CPS ≥1 (Key Secondary Endpoint) FA | Events | Median (95% CI), mo | HR (95% CI) Pembro + chemo | 73.7% | 21.0 (18.9–22.9) | 1.04 (0.81–1.33) P = 0.6185^a Placebo + chemo | 72.5% | 22.2 (17.6–26.3) | OS, % vs Time, mo (0-54) No. at risk Pembro + chemo: 171 154 131 103 69 33 12 5 2 0 Placebo + chemo: 171 151 124 96 76 44 16 9 2 0 Median follow-up, 34.9 months (range, 23.9–55.1). Median OS based on the Kaplan-Meier method. HRs (95% CIs) based on a Cox regression model with treatment as a covariate and the Efron method of tie handling stratified by the randomization stratification factors. ^aDid not meet the prespecified statistical criterion of α = 0.025. Data cutoff, Mar 16, 2026. [Speaker panel] Hope Rugo Results from KEYNOTE-B49: A phase 3, randomized, double-blind study of pembrolizumab (pembro) vs placebo (pbo) plus chemotherapy (CT) in participants (pts) with HR+/HER2- advanced breast cancer (ABC) ALICANTE AUDITORIUM - HALL 6 Survival table Endpoint: OS (key secondary), PD-L1 CPS >=1, final analysis Arm | N | Median | Median CI Pembro + chemo | 171 | 21.0 mo | 18.9–22.9 Placebo + chemo | 171 | 22.2 mo | 17.6–26.3 HR: 1.04 HR CI: 0.81–1.33 P: 0.6185 (did not meet prespecified alpha 0.025) [Slide 3] [Session banner] 08:30 - 10:00 Proffered paper 2: Breast cancer, early and advanced stage CHAIRS: SARA HURVITZ, SEOCK-AH IM Adjusted Analysis of OS^a (Post Hoc Exploratory Analysis) • 73.1% of participants in the pembrolizumab plus chemotherapy arm and 86.0% in the placebo plus chemotherapy arm received subsequent therapy (difference of 12.9%) – 50 (29.2%) and 65 (38.0%) participants received ADCs, respectively | Pembro + chemo (n = 171) | Placebo + chemo (n = 171) Subsequent therapy use, n (%) | 125 (73.1) | 147 (86.0) OS Unadjusted HR (95% CI) | 1.04 (0.81–1.33) Adjusted^a HR (95% CI) | 0.94 (0.74–1.21) • The adjusted HR for OS suggests the observed OS treatment effect may have been influenced by the confounding effect of subsequent therapy Median follow-up, 34.9 months (range, 23.9–55.1). ^aA 2-stage model was used to adjust the estimated effect on OS for the potential confounding effect of the difference in subsequent therapy use between arms. Data cutoff, Mar 16, 2026. [Speaker panel] Hope Rugo Results from KEYNOTE-B49: A phase 3, randomized, double-blind study of pembrolizumab (pembro) vs placebo (pbo) plus chemotherapy (CT) in participants (pts) with HR+/HER2- advanced breast cancer (ABC) ALICANTE AUDITORIUM - HALL 6 [Slide 4] [Session banner] 08:30 - 10:00 Proffered paper 2: Breast cancer, early and advanced stage CHAIRS: SARA HURVITZ, SEOCK-AH IM Conclusions • The phase 3 KEYNOTE-B49 study (NCT04895358) is the largest study evaluating immunotherapy plus chemotherapy in endocrine-resistant HR+/HER2−, PD-L1–positive (CPS ≥1) advanced breast cancer • Pembrolizumab plus chemotherapy significantly improved PFS vs placebo plus chemotherapy in participants with PD-L1 CPS ≥1 tumors, with a numerically greater benefit observed in those with PD-L1 CPS ≥10 tumors and a generally consistent treatment effect across other prespecified subgroups • No significant improvement in OS was observed in participants with PD-L1 CPS ≥1 tumors. Interpretation of OS results is limited, as a protocol amendment reduced enrollment from 800 to 340 participants due to low prevalence of PD-L1–positive (CPS ≥1) disease; reducing statistical power for OS to ~65% • Safety was consistent with the known profiles of each regimen with no new safety signals • KEYNOTE-B49 met its primary endpoint and demonstrated improved PFS in a PD-L1–defined subset of participants with HR+/HER2− advanced breast cancer. The ongoing TroFuse-010 study (NCT06312176) is further investigating the efficacy of pembrolizumab in combination with sacituzumab tirumotecan (sac-TMT) in both the overall population and a PD-L1–defined subgroup [Speaker panel] Hope Rugo Results from KEYNOTE-B49: A phase 3, randomized, double-blind study of pembrolizumab (pembro) vs placebo (pbo) plus chemotherapy (CT) in participants (pts) with HR+/HER2- advanced breast cancer (ABC) ALICANTE AUDITORIUM - HALL 6

OS in Participants with PD-L1 CPS ≥1 (Key Secondary Endpoint) FA | Events | Median (95% CI), mo | HR (95% CI) Pembro + chemo | 73.7% | 21.0 (18.9–22.9) | 1.04 (0.81–1.33), P = 0.6185(a) Placebo + chemo | 72.5% | 22.2 (17.6–26.3) No. at risk (0–54 mo, every 6 mo) Pembro + chemo: 171 154 131 103 69 33 12 5 2 0 Placebo + chemo: 171 151 124 96 76 44 16 9 2 0 Median follow-up, 34.9 months (range, 23.9–55.1). Median OS based on the Kaplan-Meier method. HRs (95% CIs) based on a Cox regression model with treatment as a covariate and the Efron method of tie handling stratified by the randomization stratification factors. (a)Did not meet the prespecified statistical criterion of α = 0.025. Data cutoff, Mar 16, 2026. Survival table Endpoint: OS, PD-L1 CPS ≥1 (key secondary) Arm | N | Median | Median CI Pembro + chemo | 171 | 21.0 mo | 18.9–22.9 Placebo + chemo | 171 | 22.2 mo | 17.6–26.3 HR: 1.04 HR CI: 0.81–1.33 P: 0.6185 (did not meet α=0.025)

Summary of Treatment-Related^a AEs | Pembro + chemo (n = 170) | Placebo + chemo (n = 168) Treatment-related AEs | 155 (91.2) | 152 (90.5) Grade ≥3 | 64 (37.6) | 55 (32.7) Serious | 23 (13.5) | 13 (7.7) Led to any treatment discontinuation | 19 (11.2) | 9 (5.4) Led to death | 3 (1.8)^b | 0 Treatment-Related AEs with Incidence ≥10% in Either Treatment Arm (Incidence, %; Pembro + chemo / Placebo + chemo; bars split Grade 1/2 and 3–5) PPE: 30.6 / 33.3 Anemia: 29.4 / 24.4 Nausea: 25.9 / 20.8 Diarrhea: 21.2 / 22.6 Alopecia: 18.2 / 20.8 Fatigue: 21.2 / 17.9 Neutropenia: 17.6 / 13.1 Neutrophil count ↓: 17.1 / 12.5 Asthenia: 13.5 / 15.5 WBC count ↓: 11.8 / 16.1 AST ↑: 14.7 / 11.9 ALT ↑: 11.8 / 9.5 Hypothyroidism: 18.2 / 3.0 Vomiting: 10.0 / 8.9 Rash: 10.6 / 5.4 PPE, palmar-plantar erythrodysesthesia syndrome. Data in the table are n (%). ^aDetermined by the investigator to be related to any study treatment. ^bGrade 5 events were autoimmune hepatitis, gastrointestinal inflammation, and pneumonitis. Data cutoff, Mar 16, 2026.

Annals of Oncology Available online 24 October 2026 In Press, Journal Pre-proof What's this? Original Article Pembrolizumab Plus Chemotherapy for HR+/HER2– PD-L1–Positive Advanced Breast Cancer: Final Analysis of the Phase 3 KEYNOTE-B49 Study H.S. Rugo 1, M. Kok 2, A. Goncalves 3, D. Loirat 4, G. Bianchini 5 6, P.A. Fasching 7, D.W. Cescon 8, R. Romero Díaz 9, F. Zagouri 10, E. Korbenfeld 11, F. Dalenc 12, Y. Jerez Gilarranz 13, J. Hashimoto 14, Y.H. Park 15, S.-A. Im 16, X. Hu 17, K.M. Hirshfield 18, L. Jia 18, K. Tryfonidis 18, P. Schmid 19

PFS by BICR in Participants with PD-L1 CPS ≥10 (Secondary Endpoint) IA | Events | Median (95% CI), mo | HR (95% CI) Pembro + chemo Placebo + chemo 12-mo rate (95% CI) PFS, % Time, mo No. at risk Median follow-up, 29.1 months (range, 18.1–49.4). Median PFS based on the Kaplan-Meier method. HR (95% CI) based on a Cox regression model with treatment as a covariate and the Efron method of tie handling stratified by the randomization stratification factors. Data cutoff, Sep 22, 2025. Survival table Endpoint: PFS by BICR, PD-L1 CPS ≥10 (secondary endpoint) Arm | N | Median | Median CI Pembro + chemo | 36 | 7.5 mo | 4.2–10.3 Placebo + chemo | 36 | 4.4 mo | 4.0–8.3 HR: 0.59 HR CI: 0.35–0.98 12-mo PFS: 23.6% (11.1%–38.7%) vs 9.3% (2.4%–22.2%)

Tumor Response Outcomes at Final Analysis (Secondary Endpoints) PD-L1 CPS ≥1 ORR, % (95% CI) Pembro + chemo: 31.6 (54/171) Placebo + chemo: 28.1 (48/171) PD-L1 CPS ≥10 Pembro + chemo: 44.4 (16/36) Placebo + chemo: 30.6 (11/36) PD-L1 CPS ≥1 | Pembro + chemo | Placebo + chemo DOR,^a median (range), mo | 8.3 (1.0+ to 40.0+) | 8.4 (2.2+ to 27.3) PD-L1 CPS ≥10 | Pembro + chemo | Placebo + chemo DOR,^a median (range), mo | 8.3 (4.3–29.7+) | 6.2 (4.2–12.5) Median follow-up, 34.9 months (range, 23.9–55.1). "+" indicates no PD by the time of the participant's last disease assessment. ^aIn participants with a confirmed objective response. Data cutoff, Mar 16, 2026.

[Slide 1] PFS by BICR in Participants with PD-L1 CPS >=1 (Primary Endpoint) IA: Primary PFS Analysis 12-mo rate (95% CI): 26.3% (19.1%-34.1%) [pembro + chemo]; 20.3% (14.0%-27.4%) [placebo + chemo] 24-mo rate (95% CI): 15.0% (9.0%-22.5%) [pembro + chemo]; 4.3% (1.6%-9.3%) [placebo + chemo] No. at risk (IA), Time 0-39 mo: Pembro + chemo: 171 131 95 43 30 24 18 11 8 4 2 2 1 0 Placebo + chemo: 171 126 76 42 26 19 8 2 2 1 0 0 0 0 IA | Events | Median (95% CI), mo | HR (95% CI) Pembro + chemo | 72.5% | 6.7 (6.2-8.3) | 0.74 (0.58-0.95) P = 0.0089(b) Placebo + chemo | 81.3% | 6.0 (4.2-8.1) FA: Updated Analysis(a) 12-mo rate (95% CI): 26.3% (19.1%-34.1%); 20.3% (14.0%-27.4%) 24-mo rate (95% CI): 14.1% (8.5%-21.1%); 3.5% (1.1%-8.2%) No. at risk (FA), Time 0-54 mo: Pembro + chemo: 171 131 95 43 30 24 20 16 13 8 4 2 1 1 1 0 0 0 0 Placebo + chemo: 171 126 76 42 26 19 8 4 3 2 0 0 0 0 0 0 0 0 0 FA | Events | Median (95% CI), mo | HR (95% CI) Pembro + chemo | 74.3% | 6.7 (6.2-8.3) | 0.74 (0.58-0.95) Placebo + chemo | 82.5% | 6.0 (4.2-8.1) Median follow-up, 29.1 months (range, 18.1-49.4) for IA (data cutoff, Sep 22, 2025) and 34.9 months (range, 23.9-55.1) for FA (data cutoff, Mar 16, 2026). Median PFS based on the Kaplan-Meier method. HRs (95% CIs) based on a Cox regression model with treatment as a covariate and the Efron method of tie handling stratified by the randomization stratification factors. (a)Updated analysis was descriptive. (b)Met the prespecified statistical criterion of alpha = 0.025. Survival table Endpoint: PFS by BICR, PD-L1 CPS>=1 (primary endpoint) Arm | N | Median | Median CI Pembro + chemo (IA) | 171 | 6.7 | 6.2-8.3 Placebo + chemo (IA) | 171 | 6.0 | 4.2-8.1 Pembro + chemo (FA) | 171 | 6.7 | 6.2-8.3 Placebo + chemo (FA) | 171 | 6.0 | 4.2-8.1 HR: 0.74 HR CI: 0.58-0.95 P: 0.0089 (IA) Landmark rates: 12 mo (IA): Pembro + chemo 26.3% (CI 19.1%-34.1%) | Placebo + chemo 20.3% (CI 14.0%-27.4%) 24 mo (IA): Pembro + chemo 15.0% (CI 9.0%-22.5%) | Placebo + chemo 4.3% (CI 1.6%-9.3%) 12 mo (FA): Pembro + chemo 26.3% (CI 19.1%-34.1%) | Placebo + chemo 20.3% (CI 14.0%-27.4%) 24 mo (FA): Pembro + chemo 14.1% (CI 8.5%-21.1%) | Placebo + chemo 3.5% (CI 1.1%-8.2%) [Slide 2] PFS by BICR in Participants with PD-L1 CPS >=10 (Secondary Endpoint[)] IA | Events | Median (95% CI), mo | HR (95% CI) Pembro + chemo | 80.6% | 7.5 (4.2-10.3) | 0.59 (0.35-0.98) Placebo + chemo | 91.7% | 4.4 (4.0-8.3) 12-mo rate (95% CI): 23.6% (11.1%-38.7%) [pembro + chemo]; 9.3% (2.4%-22.2%) [placebo + chemo] Y axis: PFS, %; X axis: Time, mo (0-39) No. at risk Pembro + chemo: 36 28 21 12 8 8 7 5 4 1 1 1 0 0 Placebo + chemo: 36 27 15 7 3 1 1 0 0 0 0 0 0 0 Median follow-up, 29.1 months (range, 18.1-49.4). Median PFS based on the Kaplan-Meier method. HR (95% CI) based on a Cox regression model with treatment as a covariate and the Efron method of tie handling stratified by the randomization stratification factors. Data cutoff, Sep 22, 2025. Survival table Endpoint: PFS by BICR, PD-L1 CPS>=10 (secondary endpoint) Arm | N | Median | Median CI Pembro + chemo | 36 | 7.5 | 4.2-10.3 Placebo + chemo | 36 | 4.4 | 4.0-8.3 HR: 0.59 HR CI: 0.35-0.98 Landmark rates: 12 mo: Pembro + chemo 23.6% (CI 11.1%-38.7%) | Placebo + chemo 9.3% (CI 2.4%-22.2%) [Slide 3] Subgroup Analyses of PFS by BICR in Participants with PD-L1 CPS >=1 Subgroup | Events/participants | HR (95% CI) Overall | 263/342 | 0.74 (0.58-0.95) Age <65 y | 184/244 | 0.77 (0.57-1.02) Age >=65 y | 79/98 | 0.69 (0.44-1.09) Geographic region: North America | 40/51 | 0.68 (0.36-1.29) Europe | 100/143 | 0.85 (0.57-1.26) Asia | 63/76 | 0.75 (0.45-1.24) Rest of the world | 60/72 | 0.66 (0.39-1.12) PD-L1 CPS 1-9 | 201/270 | 0.80 (0.60-1.05) PD-L1 CPS >=10 | 62/72 | 0.59 (0.35-0.98) ECOG PS 0 | 148/202 | 0.68 (0.49-0.94) ECOG PS 1 | 115/140 | 0.85 (0.59-1.23) Number of metastatic sites <3 | 138/179 | 0.58 (0.41-0.83) Number of metastatic sites >=3 | 125/163 | 0.92 (0.64-1.32) On-study chemotherapy: Paclitaxel | 32/47 | 0.87 (0.43-1.76) Nab-paclitaxel | 72/90 | 0.64 (0.40-1.03) Liposomal doxorubicin | 17/19 | 1.18 (0.45-3.08) Capecitabine | 140/182 | 0.75 (0.54-1.05) Presence of visceral metastases: Yes | 230/291 | 0.76 (0.58-0.98) No | 33/51 | 0.70 (0.34-1.45) ER status: ER 1%-10% | 15/18 | 0.42 (0.15-1.21) ER >10% | 244/318 | 0.76 (0.59-0.98) HER2 status: 0 by IHC | 68/84 | 0.68 (0.42-1.12) 1+ by IHC | 117/151 | 0.69 (0.48-1.00) 2+ by IHC (FISH negative) | 78/107 | 0.88 (0.56-1.37) Menopausal status(a): Premenopausal | 37/51 | 1.08 (0.56-2.07) Postmenopausal | 221/286 | 0.69 (0.53-0.91) Prior ET for metastatic disease: 1 | 97/128 | 0.53 (0.35-0.81) >=2 | 165/213 | 0.90 (0.67-1.23) Prior taxanes in the neoadjuvant/adjuvant setting: Yes | 146/179 | 0.55 (0.40-0.77) No | 117/163 | 0.98 (0.68-1.42) Axis 0.1-5; Favors pembro + chemo <- -> Favors placebo + chemo [Median] follow-up, 29.1 months (range, 18.1-49.4). HRs (95% CIs) based on Cox regression models with treatment as a covariate and the Efron method of tie handling; analyses in [ov]erall and PD-L1 subgroups were stratified; analyses in other subgroups were unstratified. (a)Female participants only. Data cutoff, Sep 22, 2025.

[Slide 1] [Session banner] 08:30 - 10:00 Proffered paper 2: Breast cancer, early and advanced stage CHAIRS: SARA HURVITZ, SEOCK-AH IM Tumor Response Outcomes at Final Analysis (Secondary Endpoints) PD-L1 CPS >=1: ORR, % (95% CI) Pembro + chemo 31.6 (54/171); Placebo + chemo 28.1 (48/171) PD-L1 CPS >=10: ORR, % (95% CI) Pembro + chemo 44.4 (16/36); Placebo + chemo 30.6 (11/36) PD-L1 CPS >=1 | Pembro + chemo | Placebo + chemo DOR,(a) median (range), mo | 8.3 (1.0+ to 40.0+) | 8.4 (2.2+ to 27.3) PD-L1 CPS >=10 | Pembro + chemo | Placebo + chemo DOR,(a) median (range), mo | 8.3 (4.3-29.7+) | 6.2 (4.2-12.5) Median follow-up, 34.9 months (range, 23.9-55.1). "+" indicates no PD by the time of the participant's last disease assessment. (a)In participants with a confirmed objective response. Data cutoff, Mar 16, 2026. [Speaker video] Hope Rugo Results from KEYNOTE-B49: A phase 3, randomized, double-blind study of pembrolizumab (pembro) vs placebo (pbo) plus chemotherapy (CT) in participants (pts) with HR+/HER2- advanced breast cancer (ABC) ALICANTE AUDITORIUM - HALL 6 [Slide 2] [Session banner] 08:30 - 10:00 Proffered paper 2: Breast cancer, early and advanced stage CHAIRS: SARA HURVITZ, SEOCK-AH IM Summary of Treatment-Related(a) AEs | Pembro + chemo (n = 170) | Placebo + chemo (n = 168) Treatment-related AEs | 155 (91.2) | 152 (90.5) Grade >=3 | 64 (37.6) | 55 (32.7) Serious | 23 (13.5) | 13 (7.7) Led to any treatment discontinuation | 19 (11.2) | 9 (5.4) Led to death | 3 (1.8)(b) | 0 Treatment-Related AEs with Incidence >=10% in Either Treatment Arm (Incidence, %; Pembro + chemo / Placebo + chemo; Grade 1/2 and 3-5 stacked): PPE 30.6 / 33.3; Anemia 29.4 / 24.4; Nausea 25.9 / 20.8; Diarrhea 21.2 / 22.6; Alopecia 18.2 / 20.8; Fatigue 21.2 / 17.9; Neutropenia 17.6 / 13.1; Neutrophil count decreased 17.1 / 12.5; Asthenia 13.5 / 15.5; WBC count decreased 11.8 / 16.1; AST increased 14.7 / 11.9; ALT increased 11.8 / 9.5; Hypothyroidism 18.2 / 3.0; Vomiting 10.0 / 8.9; Rash 10.6 / 5.4 PPE, palmar-plantar erythrodysesthesia syndrome. Data in the table are n (%). (a)Determined by the investigator to be related to any study treatment. (b)Grade 5 events were autoimmune hepatitis, gastrointestinal inflammation, and pneumonitis. Data cutoff, Mar 16, 2026. [Speaker video] Hope Rugo Results from KEYNOTE-B49: A phase 3, randomized, double-blind study of pembrolizumab (pembro) vs placebo (pbo) plus chemotherapy (CT) in participants (pts) with HR+/HER2- advanced breast cancer (ABC) ALICANTE AUDITORIUM - HALL 6

Progression-Free Survival, % / Time, mo Events, n (%) | Median (95% CI), mo | HR (95% CI) Pembrolizumab + chemotherapy: 124 (73) | 6.7 (6.2–8.3) | 0.74 (0.58–0.95), P = 0.0089 Placebo + chemotherapy: 139 (81) | 6.0 (4.2–8.1) 12-mo rate (95% CI): 26% (19%–34%) [pembro + chemo]; 20% (14%–27%) [placebo + chemo] 24-mo rate (95% CI): 15% (9%–23%) [pembro + chemo]; 4% (2%–9%) [placebo + chemo] No. at risk (months 0,3,6,9,12,15,18,21,24,27,30,33,36,39) Pembrolizumab + chemotherapy: 171 131 95 43 30 24 18 11 8 4 2 2 1 0 Placebo + chemotherapy: 171 126 76 42 26 19 8 2 2 1 0 0 0 0 No. censored Pembrolizumab + chemotherapy: 0 6 14 31 32 33 37 41 43 46 46 46 46 47 Placebo + chemotherapy: 0 10 19 25 26 26 28 31 31 32 32 32 32 32 Survival table Endpoint: PFS Arm | N | Median | Median CI Pembrolizumab + chemotherapy | 171 | 6.7 mo | 6.2–8.3 Placebo + chemotherapy | 171 | 6.0 mo | 4.2–8.1 HR: 0.74 HR CI: 0.58–0.95 P: 0.0089 Landmark rates: 12 mo: Pembrolizumab + chemotherapy 26% (CI 19%–34%) | Placebo + chemotherapy 20% (CI 14%–27%) 24 mo: Pembrolizumab + chemotherapy 15% (CI 9%–23%) | Placebo + chemotherapy 4% (CI 2%–9%)

[Slide 1] FourLight-1 (NCT06105632) Study Design Key Eligibility Criteria • Aged ≥18 years • HR+/HER2− a/mBC with either measurable disease or non-measurable bone-only disease by RECIST v1.1 • Previously received a CDK4/6 inhibitor either as adjuvant therapy, with metastatic recurrence during treatment or within 12 months after completion, or as 1L treatment for a/mBC, followed by disease progression • 2 prior lines of systemic therapy for a/mBC were permitted, one must have been a CDK4/6 inhibitor plus NSAI and the other an approved treatment targeting ESR1, PIK3CA/AKT1/PTEN, or BRCA mutations • ECOG performance status of ≤2 1:1 Treatment Atirmociclib: 300 mg PO BID, continuously throughout 28-day cycles + Fulvestrant: 500 mg IM on day 1 and 15 of cycle 1 and then on day 1 of each cycle thereafter Investigator's Choice of Therapy (ICT)a: Fulvestrant: 500 mg IM on day 1 and 15 of cycle 1 and then on day 1 of each cycle thereafter OR Everolimus: 10 mg PO QD, continuously + Exemestane: 25 mg PO QD, continuously Endpoints Primary • Investigator-assessed PFS Secondary • OS • Antitumor activity (investigator-assessed) - OR - DoR - CBR • Safety and tolerability • At randomization, patients were stratified by visceral disease (yes vs no) and prior CDK4/6 inhibitor treatment duration (<12 vs ≥12 months). aICT regimen was declared for all patients prior to randomization and could not be changed during a patient's course of treatment. 1L, first-line; a/mBC, advanced or metastatic breast cancer; BID, twice daily; CBR, clinical benefit rate; CDK4/6, cyclin-dependent kinase 4/6; DoR, duration of response; ECOG, Eastern Cooperative Oncology Group; HER2−, human epidermal growth factor receptor 2-negative; HR+, hormone receptor-positive; ICT, investigator's choice of therapy; IM, intramuscularly; NSAI, non-steroidal aromatase inhibitor; OR, objective response; OS, overall survival; PFS, progression-free survival; PO, orally; QD, once daily; RECIST, Response Evaluation Criteria in Solid Tumors. Presenter: Antonio Giordano [Slide 3] Safety and Tolerability • Atirmociclib demonstrated a manageable safety and tolerability profile with low rates of serious treatment-emergent adverse events (TEAEs) and treatment discontinuations due to TEAEs Safety and tolerability summary | Atirmociclib arm (n = 125) | ICT arm (n = 136) Any TEAE | 121 (96.8) | 120 (88.2) Serious TEAE | 14 (11.2) | 27 (19.9) Maximum grade 3 or 4 TEAE | 47 (37.6) | 33 (24.3) Maximum grade 5 TEAE* | 4 (3.2) | 5 (3.7) Discontinuation due to TEAEs ≥1 study treatment | 8 (6.4) | 9 (6.6) All study treatment | 4 (3.2) | 6 (4.4) Atirmociclib | 8 (6.4) | NA Dose reduction due to TEAEs ≥1 study treatment | 17 (13.6) | 24 (17.6) Atirmociclib | 17 (13.6) | NA Dose interruption due to TEAEs ≥1 study treatment | 55 (44.0) | 32 (23.5) Atirmociclib | 55 (44.0) | NA * One death in the atirmociclib arm was considered related to treatment by the investigator (Pneumonitis) TEAEs ≥10% in Either Treatment Arm (Proportion of patients, %) | Atirmociclib | ICT Any | 97 (grade 3: 35) | 88 (grade 3: 21) Neutropenia | 47 (grade 3: 22) | 8 Leukopenia | 34 | 10 Diarrhea | 28 | 13 Fatigue | 21 | 9 Nausea | 18 | 13 Rash | 18 | 13 AST increased | 14 | 16 Headache | 14 | 4 Hyperglycemia | 14 | 12 Lymphopenia | 14 | 4 Anemia | 13 | 16 Arthralgia | 12 | 10 Vomiting | 12 | 5 Cough | 11 | 10 Thrombocytopenia | 11 | 10 ALT increased | 9 | 14 Back pain | 6 | 13 Stomatitis | 5 | 30 Legend: Atirmociclib Grade 1/2, Grade 3, Grade 4; ICT Grade 1/2, Grade 3, Grade 4 ALT, alanine aminotransferase; AST, aspartate aminotransferase; ICT, investigator's choice of therapy; TEAE, treatment-emergent adverse event. Presenter: Antonio Giordano [Slide 4] Conclusions • Atirmociclib, a potential first-in-class, highly selective CDK4 inhibitor, in combination with fulvestrant demonstrated statistically significant and clinically meaningful PFS improvement in patients whose disease progressed on a prior CDK4/6 inhibitor, while having a manageable safety and tolerability profile. • Among patients with HR+/HER2− mBC, those who received atirmociclib plus fulvestrant had a 40% relative risk reduction in disease progression or death compared with those who received ICT. • These findings support the continued development of atirmociclib in earlier lines of therapy, including an ongoing phase 3 study of atirmociclib versus CDK4/6 inhibitors, in combination with letrozole, as 1L treatment of HR+/HER2− advanced or mBC (FourLight-3; NCT06760637; accrual completed). 1L, first-line; CDK, cyclin-dependent kinase; HER2−, human epidermal growth factor receptor 2-negative; HR+, hormone receptor-positive; ICT, investigator's choice of therapy; mBC, metastatic breast cancer; PFS, progression-free survival. Presenter: Antonio Giordano

[Slide 2] Patient Demographic and Clinical Characteristics Characteristic | Atirmociclib arm (n = 126) | ICT arm (n = 137) Age Median, years (range) | 57 (28–85) | 58 (28–82) <65 years, n (%) | 93 (74) | 94 (69) ≥65 years, n (%) | 33 (26) | 43 (31) Female sex, n (%) | 125 (99) | 136 (99) Postmenopausal | 92 (73) | 103 (75) Race, n (%) Asian | 65 (52) | 71 (52) White | 54 (43) | 57 (42) American Indian or Alaska Native | 5 (4) | 7 (5) Black or African American | 1 (1) | 1 (1) Unknown | 1 (1) | 1 (1) Geographic region, n (%) Asia/Pacific | 66 (52) | 73 (53) US/Canada | 6 (5) | 5 (4) Europe | 6 (5) | 7 (5) Rest of the world | 48 (38) | 52 (38) ECOG performance status 0 or 1, n (%) | 126 (100) | 135 (99) Stage IV at initial diagnosis (de novo mBC), n (%) | 57 (45) | 55 (40) Stage IV at current diagnosis, n (%) | 126 (100) | 137 (100) Measurable disease, n (%) | 109 (87) | 115 (84) No. of disease sites, n (%) 1 | 30 (24) | 38 (28) 2 | 33 (26) | 42 (31) ≥3 | 63 (50) | 57 (42) Characteristic | Atirmociclib arm (n = 126) | ICT arm (n = 137) Visceral metastases (any),ᵃ n (%) | 82 (65) | 87 (64) Liver metastases, n (%) | 55 (44) | 40 (29) Bone-only disease, n (%) | 18 (14) | 20 (15) ESR1 status, n (%) Altered | 52 (41) | 48 (35) Wildtype | 48 (38) | 61 (45) Unknown | 26 (21) | 28 (20) PIK3CA/AKT1/PTEN status, n (%) Altered | 47 (37) | 61 (45) Wildtype | 53 (42) | 48 (35) Unknown | 26 (21) | 28 (20) Prior anticancer surgery, n (%) | 76 (60) | 85 (62) Prior anticancer radiotherapy, n (%) | 62 (49) | 79 (58) Prior systemic anticancer therapy, n (%) (Neo)adjuvant setting only | 2 (2) | 8 (6) 1 regimen in mBC setting | 122 (97) | 126 (92) 2 regimens in mBC setting | 2 (2) | 3 (2) Prior CDK4/6 inhibitor use, n (%) (Neo)adjuvant setting | 2 (2) | 8 (6) mBC setting | 124 (98) | 129 (94) Duration of prior CDK4/6 inhibitor treatment, n (%) <12 months | 33 (26) | 40 (29) ≥12 months | 93 (74) | 97 (71) <3 months from last CDK4/6 inhibitor dose to randomization, n (%) | 116 (92) | 117 (85) ᵃVisceral refers to ≥1 lesion in liver, lung, pleura, peritoneum, brain, other visceral organs. CDK4/6, cyclin-dependent kinase 4/6; ECOG, Eastern Cooperative Oncology Group; ICT, investigator's choice of therapy; mBC, metastatic breast cancer. Presenter: Antonio Giordano [Slide 4] Background • CDK4/6 inhibitors plus ET are the standard-of-care for HR+/HER2− mBC but resistance ultimately limits durable disease control and marks a setting of substantial unmet need.¹,² • Antitumor activity in ER-positive disease is driven primarily by CDK4, while neutropenia is linked to CDK6 inhibition.³ • Atirmociclib (PF-07220060) is a potential first-in-class, highly selective and potent CDK4 inhibitor with up to 33-fold greater selectivity than approved CDK4/6 inhibitors.³ ATIRMOCICLIB | Proposed Mechanism of Action [Diagram labels: Estrogen receptor; Estrogen; Cytoplasm; Atirmociclib inhibits the CDK4/cyclin D1 complex and blocks Rb phosphorylation; Cyclin D1; CDK4; Rb; E2F; Cell cycle progression; G1, S, G2, M; Nuclear membrane] 1L, first-line; CDK, cyclin-dependent kinase; ET, endocrine therapy; HER2−, human epidermal growth factor receptor 2-negative; HR+, hormone receptor-positive; mBC, metastatic breast cancer. 1. McAndrew NP, et al. JCO Oncol Pract. 2022;18:319-327. 2. Teysir J, et al. Am Soc Clin Oncol Educ Book. 2025;45:e473372. 3. Palmer CL, et al. Cancer Cell. 2025;43:464-481.e14. Presenter: Antonio Giordano

Investigator-Assessed PFS | Atirmociclib arm (n = 126) | ICT arm (n = 137) Median follow-up, months | 13.7 | 13.6 Events | 72 | 93 Median PFS, months (95% CI) | 9.4 (7.3–13.6) | 6.6 (3.8–7.8) Stratified HR (95% CI) | 0.60 (0.440–0.825) one-sided P value | 0.0007 Curve annotations: 3 mo: Atirmociclib 82.5%, ICT 68.9%; 6 mo: Atirmociclib 62.3%, ICT 51.9%; 12 mo: Atirmociclib 44.8%, ICT 22.5% Time (months): 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 No. at risk Atirmociclib arm: 126 120 100 98 83 83 72 71 64 63 55 49 34 32 13 12 12 3 3 2 0 ICT arm: 137 127 90 84 66 65 58 54 45 45 31 25 17 16 7 7 6 1 0 0 0 CI, confidence interval; HR, hazard ratio; ICT, investigator's choice of therapy; PFS, progression-free survival. Presenter: Antonio Giordano

Tumor Response Characteristic | Atirmociclib arm (n = 126) | ICT arm (n = 137) Measurable disease, n | 109 | 115 Confirmed best overall response, n (%) Complete response | 1 (0.9) | 1 (0.9) Partial response | 20 (18.3) | 12 (10.4) Stable disease | 64 (58.7) | 56 (48.7) Progressive disease | 20 (18.3) | 35 (30.4) Not evaluable | 4 (3.7) | 11 (9.6) Objective response rate^a, n (%) | 21 (19.3) | 13 (11.3) 95% CI | 13.0–27.7 | 6.7–18.4 Nominal 2-sided P value | 0.0901 Median duration of response^b (95% CI), months | 14.8 (7.4–NE) | 9.3 (5.5–NE) Clinical benefit rate^c, n (%) | 78 (61.9) | 63 (46.0) 95% CI | 53.2–69.9 | 37.9–54.3 Nominal 2-sided P value | 0.0083 ^aDefined as confirmed complete or partial response. ^bAmong patients with confirmed complete or partial response. ^cFull analysis set population, defined as confirmed complete response, partial response, or stable disease/non-complete response/non-progressive disease (for patients with nonmeasurable disease) ≥24 weeks. CI, confidence interval; ICT, investigator's choice of therapy; NE, not estimable. Presenter: Antonio Giordano

Fourlight-3 Study design [Overlay sticker, red: The actual question] Screening: Adults with HR+/HER2– a/mBC who have not received prior systemic anticancer treatment for advanced disease Randomization^1 1:1 N = 1020 Arm A: Investigational arm, n = 510, atirmociclib 300 mg (twice daily) + letrozole 2.5 mg (once daily), 28-day cycle Arm B: Comparator arm, n = 510, CDK4/6 inhibitor^a + letrozole 2.5 mg (once daily), 28-day cycle Tumor assessments every 8±1 weeks for the first 48 weeks and every 12±1 weeks thereafter Safety follow-up 28–35 days after last dose Survival follow-up every 3 months after last dose Active, not recruiting ⓘ — No longer looking for participants Updated on May 18, 2026 Study of PF-07220060 With Letrozole in Adults With HR-positive HER2-negative Breast Cancer Who Have Not Received Anticancer Treatment for Advanced/Metastatic Disease (FourLight-3) ClinicalTrials.gov ID NCT06760637 Peter A. Fasching (peter.fasching@uk-erlangen.de) Madrid ESMO Conference 2026 E47P https://clinicaltrials.gov/study/NCT06760637

Investigator-Assessed PFS [Kaplan-Meier curves: PFS probability (%) vs Time (months) 0-20] Landmark labels: 82.5%, 62.3%, 44.8% (atirmociclib); 68.9%, 51.9%, 22.5% (ICT) at months 3, 6, 12 Atirmociclib arm (n = 126) ICT arm (n = 137) Median follow-up, months 13.7 13.6 Events 72 93 [partially legible in this photo] Median PFS, months (95% CI) 9.4 (7.3-13.6) 6.6 (3.8-7.8) Stratified HR (95% CI) 0.60 (0.440-0.825) one-sided P value 0.0007 No. at risk Atirmociclib arm: 126 120 100 98 83 83 72 71 64 63 55 49 34 32 13 12 12 3 3 2 0 ICT arm: 137 127 90 84 66 65 58 [remaining values illegible in this photo] CI, confidence interval; HR, hazard ratio; ICT, investigator's choice of therapy; PFS, progression-free survival Presenter: Antonio Giordano Survival table Endpoint: Investigator-assessed PFS Arm | N | Median | Median CI Atirmociclib | 126 | 9.4 mo | 7.3-13.6 Investigator's choice (ICT) | 137 | 6.6 mo | 3.8-7.8 HR: 0.60 HR CI: 0.440-0.825 P: 0.0007 (one-sided) Landmark rates: 3 mo: Atirmociclib 82.5% | ICT 68.9% 6 mo: Atirmociclib 62.3% | ICT 51.9% 12 mo: Atirmociclib 44.8% | ICT 22.5%

Investigator-Assessed PFS Table: Atirmociclib arm (n = 126) | ICT arm (n = 137) Median follow-up, months: 13.7 | 13.6 Events: 72 | 93 Median PFS, months (95% CI): 9.4 (7.3-13.6) | 6.6 (3.8-7.8) Stratified HR (95% CI): 0.60 (0.440-0.825) one-sided P value: 0.0007 Curve annotations: 3 mo 82.5% (atirmociclib) vs 68.9% (ICT); 6 mo 62.3% vs 51.9%; 12 mo 44.8% vs 22.5% Y axis: PFS probability (%); X axis: Time (months) 0-20 No. at risk Atirmociclib arm: 126 120 100 98 83 83 72 71 64 63 55 49 34 32 13 12 12 3 3 2 0 ICT arm: 137 127 90 84 66 65 58 54 45 45 31 25 17 16 7 7 6 1 0 0 0 CI, confidence interval; HR, hazard ratio; ICT, investigator's choice of therapy; PFS, progression-free survival. Presenter: Antonio Giordano Survival table Endpoint: Investigator-assessed PFS Arm | N | Median | Median CI Atirmociclib arm | 126 | 9.4 | 7.3-13.6 ICT arm | 137 | 6.6 | 3.8-7.8 HR: 0.60 (stratified) HR CI: 0.440-0.825 P: 0.0007 (one-sided) Landmark rates: 3 mo: Atirmociclib arm 82.5% | ICT arm 68.9% 6 mo: Atirmociclib arm 62.3% | ICT arm 51.9% 12 mo: Atirmociclib arm 44.8% | ICT arm 22.5%

[Slide 1] CDK4 vs. CDK6 knock out by CRISPR A [Box plots of CDK4 CERES (left panel) and CDK6 CERES (right panel) by Tumor type; x-axis 0, -1, -2] Tumor type (rows, top to bottom): Chronic Myeloid Leukemia B-Lymphoblastic Leukemia/Lymphoma Ampullary Carcinoma Lung Neuroendocrine Tumor Cutaneous Squamous Cell Carcinoma Prostate Adenocarcinoma_Primary Cervical Squamous Cell Carcinoma Esophageal Squamous Cell Carcinoma Bladder Urothelial Carcinoma Diffuse large B-cell lymphoma Multiple Myeloma Head and Neck Squamous Cell Carcinoma Synovial Sarcoma Esophagogastric Adenocarcinoma Acute Myeloid Leukemia [boxed in red] Diffuse Glioma Ovarian Epithelial Tumor Pleural Mesothelioma T-Lymphoblastic Leukemia/Lymphoma [boxed in red] Intraductal Papillary Neoplasm of the Bile Duct Endometrial Carcinoma Rhabdomyosarcoma Neuroblastoma Renal Cell Carcinoma Anaplastic large cell lymphoma Pancreatic Adenocarcinoma Osteosarcoma Non-Small Cell Lung Cancer Anaplastic Thyroid Cancer Intracholecystic Papillary Neoplasm ER_Negative_Breast Hepatocellular Carcinoma Colorectal Adenocarcinoma Burkitt Lymphoma Melanoma Liposarcoma Ocular Melanoma Nerve Sheath Tumor PRAD_Metastatic [red, boxed] Ewing Sarcoma [red, boxed] ER_Positive_Breast [red, boxed] CDK4 CERES CDK6 CERES ER positive breast cancer relies on CDK4 for cell growth, but not on CDK6 Peter A. Fasching (peter.fasching@uk-erlangen.de) Palmer et al. 2025 Cancer Cell with permission [Slide 2] Investigator-Assessed PFS Atirmociclib arm (n = 126) ICT arm (n = 137) Events 72 93 Median PFS, months 9.4 6.6 (95% CI) (7.3-13.6) (3.8-7.8) Stratified HR (95% CI) 0.60 (0.440-0.825) one-sided P value 0.0007 [KM curve, PFS probability (%) y-axis; landmark labels on curve: atirmociclib 82.5%, 62.3%, 44.8%; ICT 68.9%, 51.9%, 22.5%] [Overlaid inset panels boxed in red:] A: Tumor volume (mm3) vs Days - Veh, Fulv, At, At+Fulv B: Spheroid diameter (mm) vs Days - Veh, Fulv, At, At+Fulv Participants With CCCA at Day 14*: Atirmociclib + Letrozole 88% (45/51); Letrozole 18% (10/56) Absolute Ki-67 Value at Baseline and Day 14*: Atirmociclib + Letrozole n = 49, n = 51, Median change = -30.1%; Letrozole n = 54, n = 56, Median change = -17.3%; Screening / Day 14; Visit Presenter: Antonio Giordano Survival table Endpoint: Investigator-assessed PFS Arm | N | Median | Median CI Atirmociclib | 126 | 9.4 months | 7.3-13.6 ICT (investigator's choice of therapy) | 137 | 6.6 months | 3.8-7.8 HR: 0.60 HR CI: 0.440-0.825 P: 0.0007 (one-sided) Landmark rates: [timepoint not legible; first marker]: Atirmociclib 82.5% | ICT 68.9% [timepoint not legible; second marker]: Atirmociclib 62.3% | ICT 51.9% [timepoint not legible; third marker]: Atirmociclib 44.8% | ICT 22.5% [Slide 3] Safety Atirmociclib vs. historic Palbociclib data TEAEs >=10% in Either Treatment Arm Atirmociclib ICT Any 97 (35 grade 3 segment) 21 / 88 Neutropenia 47 (22) 8 Leukopenia 34 10 Diarrhea 28 13 Fatigue 21 9 Nausea 18 13 Rash 18 13 AST increased 14 16 Headache 14 4 Hyperglycemia 14 12 Lymphopenia 14 4 Anemia 13 16 Arthralgia 12 10 Vomiting 12 5 Cough 11 10 Thrombocytopenia 11 10 ALT increased 9 14 Back pain 6 13 Stomatitis 5 30 Legend: Atirmociclib Grade 1/2; Atirmociclib Grade 3; Atirmociclib Grade 4; ICT Grade 1/2; ICT Grade 3; ICT Grade 4 Proportion of patients (%) [100 80 60 40 20 0 20 40 60 80 100] Neutropenia in Paloma-3 (Turner et al. NEJM 2015): - All grade: 78.8% - Grade 3 and Grade 4: 62.0% Neutropenia in Fourlight-1 (Giordano ESMO 2026): - All grade: 47% - Grade 3 and Grade 4: 22% [Diagram: Therapeutic index; Neutrophil; Hematologic effect (D3, CDK6); Tumor effect (D1, CDK4); Dual CDK4/6 inhibitor vs CDK4 selective inhibitor (Atirmociclib); Exposure*; *In vivo site of action concentration] Peter A. Fasching (peter.fasching@uk-erlangen.de) Palmer et al. 2025 Cancer Cell with permission [Slide 4] Is there a hook for omitting CDK6 Inhibition The effect of CDK6 vs. CDK4 inhibition on T-Cell activation [Kinome trees: Palbociclib, Ribociclib, Abemaciclib, Atirmociclib; Node Color 0 50 100; Node Size 0 100] Kinase (Ki, nM) Palbociclib Ribociclib Abemaciclib Atirmociclib CDK4 1.4 6.0 0.3 0.7 CDK6 2.2 26 4.3 23 CDK4 selectivity (CDK4:CDK6) 1.5 Fold 4 Fold 14 Fold 33 Fold E [Bar chart: [IL2] (pg/mL), 0-200; siCTRL, siCDK4, siCDK6; groups UT, TCR/IgG, TCR/PD-1; * significance markers] UT: Unstimulated control - no activating stimulation. TCR/IgG: Anti-CD3 + anti-CD28 activate the cells; control IgG provides no additional PD-1 signal TCR/PD-1: The same activating stimulation plus PD-1 engagement, which applies an inhibitory signal and reduces activation Peter A. Fasching (peter.fasching@uk-erlangen.de) Palmer et al. 2025 Cancer Cell with permission https://s206.q4cdn.com/795948973/files/doc_downloads/2024/02/29/Pfizer-Oncology-Innovation-Day-Presentation_FINAL.pdf

[Slide 1] 08:30 - 10:00 Proffered paper 2: Breast cancer, early and advanced stage CHAIRS: SARA HURVITZ, SEOCK-AH IM Investigator-Assessed PFS | Atirmociclib arm (n = 126) | ICT arm (n = 137) Median follow-up, months | 13.7 | 13.6 Events | 72 | 93 Median PFS, months (95% CI) | 9.4 (7.3–13.6) | 6.6 (3.8–7.8) Stratified HR (95% CI) | 0.60 (0.440–0.825) one-sided P value | 0.0007 Curve annotations: Atirmociclib 82.5%, 62.3%, 44.8%; ICT 68.9%, 51.9%, 22.5% (at dotted reference lines at 3, 6, 12 months) Y axis: PFS probability (%); X axis: Time (months) 0-20 No. at risk (months 0,1,...,20) Atirmociclib arm: 126 120 100 98 83 83 72 71 64 63 55 49 34 32 13 12 12 3 3 2 0 ICT arm: 137 127 90 84 66 65 58 54 45 45 31 25 17 16 7 7 6 1 0 0 0 CI, confidence interval; HR, hazard ratio; ICT, investigator's choice of therapy; PFS, progression-free survival. Presenter: Antonio Giordano [Speaker video panel] Antonio Giordano Efficacy and safety of atirmociclib (ATI) + fulvestrant (FUL) in patients (pts) with HR+/HER2– advanced/metastatic breast cancer (mBC) who progressed on prior CDK4/6 inhibitor (CDK4/6i) in a randomized phase 2 study (FOURLIGHT-1) ALICANTE AUDITORIUM - HALL 6 Survival table Endpoint: Investigator-assessed PFS Arm | N | Median | Median CI Atirmociclib + fulvestrant | 126 | 9.4 months | 7.3–13.6 ICT | 137 | 6.6 months | 3.8–7.8 HR: 0.60 HR CI: 0.440–0.825 P: 0.0007 (one-sided) Landmark rates: 3 months (dotted line): Atirmociclib 82.5% | ICT 68.9% 6 months (dotted line): Atirmociclib 62.3% | ICT 51.9% 12 months (dotted line): Atirmociclib 44.8% | ICT 22.5% [Slide 2] 08:30 - 10:00 Proffered paper 2: Breast cancer, early and advanced stage CHAIRS: SARA HURVITZ, SEOCK-AH IM Tumor Response Characteristic | Atirmociclib arm (n = 126) | ICT arm (n = 137) Measurable disease, n | 109 | 115 Confirmed best overall response, n (%) Complete response | 1 (0.9) | 1 (0.9) Partial response | 20 (18.3) | 12 (10.4) Stable disease | 64 (58.7) | 56 (48.7) Progressive disease | 20 (18.3) | 35 (30.4) Not evaluable | 4 (3.7) | 11 (9.6) Objective response ratea, n (%) | 21 (19.3) | 13 (11.3) 95% CI | 13.0–27.7 | 6.7–18.4 Nominal 2-sided P value | 0.0901 Median duration of responseb (95% CI), months | 14.8 (7.4–NE) | 9.3 (5.5–NE) Clinical benefit ratec, n (%) | 78 (61.9) | 63 (46.0) 95% CI | 53.2–69.9 | 37.9–54.3 Nominal 2-sided P value | 0.0083 aDefined as confirmed complete or partial response. bAmong patients with confirmed complete or partial response. cFull analysis set population, defined as confirmed complete response, partial response, or stable disease/non-complete response/non-progressive disease (for patients with nonmeasurable disease) ≥24 weeks. CI, confidence interval; ICT, investigator's choice of therapy; NE, not estimable. Presenter: Antonio Giordano [Speaker video panel] Antonio Giordano Efficacy and safety of atirmociclib (ATI) + fulvestrant (FUL) in patients (pts) with HR+/HER2– advanced/metastatic breast cancer (mBC) who progressed on prior CDK4/6 inhibitor (CDK4/6i) in a randomized phase 2 study (FOURLIGHT-1) ALICANTE AUDITORIUM - HALL 6 [Slide 3] 08:30 - 10:00 Proffered paper 2: Breast cancer, early and advanced stage CHAIRS: SARA HURVITZ, SEOCK-AH IM Safety and Tolerability • Atirmociclib demonstrated a manageable safety and tolerability profile with low rates of serious treatment-emergent adverse events (TEAEs) and treatment discontinuations due to TEAEs Safety and tolerability summary | Atirmociclib arm (n = 125) | ICT arm (n = 136) Any TEAE | 121 (96.8) | 120 (88.2) Serious TEAE | 14 (11.2) | 27 (19.9) Maximum grade 3 or 4 TEAE | 47 (37.6) | 33 (24.3) Maximum grade 5 TEAE* | 4 (3.2) | 5 (3.7) Discontinuation due to TEAEs ≥1 study treatment | 8 (6.4) | 9 (6.6) All study treatment | 4 (3.2) | 6 (4.4) Atirmociclib | 8 (6.4) | NA Dose reduction due to TEAEs ≥1 study treatment | 17 (13.6) | 24 (17.6) Atirmociclib | 17 (13.6) | NA Dose interruption due to TEAEs ≥1 study treatment | 55 (44.0) | 32 (23.5) Atirmociclib | 55 (44.0) | NA * One death in the atirmociclib arm was considered related to treatment by the investigator (Pneumonitis) TEAEs ≥10% in Either Treatment Arm (Atirmociclib | ICT, %) Any | 97 (grade 3 bar label 35) | 88 (grade 3 bar label 21) Neutropenia | 47 (grade 3 bar label 22) | 8 Leukopenia | 34 | 10 Diarrhea | 28 | 13 Fatigue | 21 | 9 Nausea | 18 | 13 Rash | 18 | 13 AST increased | 14 | 16 Headache | 14 | 4 Hyperglycemia | 14 | 12 Lymphopenia | 14 | 4 Anemia | 13 | 16 Arthralgia | 12 | 10 Vomiting | 12 | 5 Cough | 11 | 10 Thrombocytopenia | 11 | 10 ALT increased | 9 | 14 Back pain | 6 | 13 Stomatitis | 5 | 30 Legend: Atirmociclib Grade 1/2; Atirmociclib Grade 3; Atirmociclib Grade 4; ICT Grade 1/2; ICT Grade 3; ICT Grade 4 Proportion of patients (%) ALT, alanine aminotransferase; AST, aspartate aminotransferase; ICT, investigator's choice of therapy; TEAE, treatment-emergent adverse event. Presenter: Antonio Giordano [Speaker video panel] Antonio Giordano Efficacy and safety of atirmociclib (ATI) + fulvestrant (FUL) in patients (pts) with HR+/HER2– advanced/metastatic breast cancer (mBC) who progressed on prior CDK4/6 inhibitor (CDK4/6i) in a randomized phase 2 study (FOURLIGHT-1) ALICANTE AUDITORIUM - HALL 6

FourLight-1 is a randomized phase 2 trial evaluating the efficacy and safety of atirmociclib + fulvestrant vs investigator's choice of therapy (ICT) in patients with HR+/HER2− mBC whose disease progressed on a CDK4/6 inhibitor Atirmociclib is a potential first-in-class, highly selective CDK4 inhibitor with improved CDK4 target coverage, which may offer a wider therapeutic index and greater antitumor response than currently approved CDK4/6 inhibitors. [Diagram] Atirmociclib Proposed Mechanism of Action — Cytoplasm; Estrogen; Estrogen receptor; Cyclin D1; CDK4; Rb; E2F; Cell cycle progression (G1, S, G2, M). Atirmociclib inhibits the CDK4/cyclin D1 complex and blocks Rb phosphorylation Patients with HR+/HER2− mBC Prior treatment with CDK4/6 inhibitor with recurrence/progression Randomized 1:1 Atirmociclib arm: Atirmociclib + Fulvestrant, n = 126 ICT arm: Fulvestrant alone OR Everolimus + Exemestane, n = 123 Atirmociclib reduced the risk of disease progression or death by 40% vs ICT Hazard ratio = 0.60; 95% CI: 0.440–0.825, P = 0.0007 PFS probability (%) vs Time (months) 0–19 Atirmociclib arm: 82.5% (3 mo), 62.3% (6 mo), 44.8% (12 mo) ICT arm: 68.9% (3 mo), 51.9% (6 mo), 22.5% (12 mo) A consistent PFS benefit with atirmociclib was observed across most prespecified subgroups. Secondary efficacy outcomes (Atirmociclib arm / ICT arm) Objective response rate: 19% / 11% Clinical benefit rate: 62% / 46% Duration of response: 14.8 months / 9.3 months TEAEs in ≥10% of patients in either treatment arm (Proportion of patients, %; Atirmociclib / ICT) Any 97 [grade 3 segment 35] / 88 [grade 3 segment 21] Neutropenia 47 [22] / 8 Leukopenia 34 / 10 Diarrhea 28 / 13 Fatigue 21 / 9 Nausea 18 / 13 Rash 18 / 13 AST increased 14 / 16 Headache 14 / 4 Hyperglycemia 14 / 12 Lymphopenia 14 / 4 Anemia 13 / 16 Arthralgia 12 / 10 Vomiting 12 / 5 Cough 11 / 10 Thrombocytopenia 11 / 10 ALT increased 9 / 14 Back pain 6 / 13 Stomatitis 5 / 30 Legend: Atirmociclib Grade 1/2, Grade 3, Grade 4; ICT Grade 1/2, Grade 3, Grade 4 Atirmociclib demonstrated a manageable safety and tolerability profile with low rates of serious TEAEs (11.2%) and treatment discontinuations due to TEAEs (6.4%). Relevance • Atirmociclib in combination with fulvestrant showed significant and clinically meaningful PFS improvement in patients who progressed on a prior CDK4/6 inhibitor, with a manageable safety and tolerability profile. • These findings support the continued development of atirmociclib in earlier lines of therapy, including an ongoing phase 3 study of atirmociclib versus CDK4/6 inhibitors, in combination with letrozole, as first-line treatment of HR+/HER2− mBC. ALT, alanine aminotransferase; AST, aspartate aminotransferase; CDK, cyclin-dependent kinase; ECOG, Eastern Cooperative Oncology Group; HER2−, human epidermal growth factor receptor 2-negative; HR+, hormone receptor-positive; ICT, investigator's choice of therapy; mBC, advanced/metastatic breast cancer; PFS, progression-free survival; TEAE, treatment-emergent adverse event. Survival table Endpoint: PFS (FourLight-1 primary analysis) Atirmociclib + fulvestrant | 126 ICT (fulvestrant alone or everolimus + exemestane) | 123 HR: 0.60 HR CI: 0.440–0.825 P: 0.0007 Landmark rates: 3 mo: Atirmociclib 82.5% | ICT 68.9% 6 mo: Atirmociclib 62.3% | ICT 51.9% 12 mo: Atirmociclib 44.8% | ICT 22.5%

[Slide 1] ESMO / BREASTCANCER·ORG New CDK4 inhibitor slows progression of HR+ metastatic breast cancer FOURLIGHT-1 IS A RANDOMIZED PHASE 2 TRIAL EVALUATING THE EFFICACY AND SAFETY OF ATIRMOCICLIB + FULVESTRANT VS INVESTIGATOR'S CHOICE OF THERAPY (ICT) IN PATIENTS WITH HR+/HER2- MBC WHOSE DISEASE PROGRESSED ON A CDK4/6 INHIBITOR Atirmociclib is a potential first-in-class, highly selective CDK4 inhibitor with improved CDK4 target coverage, which may offer a wider therapeutic index and greater antitumor response than currently approved CDK4/6 Atirmociclib Proposed Mechanism of Action Atirmociclib inhibits the CDK4/cyclin D1 complex and blocks Rb phosphorylation Cytoplasm Estrogen Estrogen receptor Cyclin D1 CDK4 Rb E2F Cell cycle progression Patients with HR+/HER2− mBC Prior treatment with CDK4/6 inhibitor with recurrence/progression Randomized 1:1 Atirmociclib arm Atirmociclib + Fulvestrant n = 126 ICT arm Fulvestrant alone OR Everolimus + Exemestane n = 137 [Slide 2] ESMO / BREASTCANCER·ORG New CDK4 inhibitor slows progression of HR+ metastatic breast cancer FOURLIGHT-1 IS A RANDOMIZED PHASE 2 TRIAL EVALUATING THE EFFICACY AND SAFETY OF ATIRMOCICLIB + FULVESTRANT VS INVESTIGATOR'S CHOICE OF THERAPY (ICT) IN PATIENTS WITH HR+/HER2- MBC WHOSE DISEASE PROGRESSED ON A CDK4/6 INHIBITOR Atirmociclib reduced the risk of disease progression or death by 40% vs ICT Hazard ratio = 0.60;95% CI: 0.440–0.825, P = 0.0007 Atirmociclib arm ICT arm PFS probability (%) Time (months) A consistent PFS benefit with atirmociclib was observed across most prespecified subgroups. Secondary efficacy outcomes Atirmociclib arm | ICT arm Objective response rate | 19% | 11% Clinical benefit rate | 62% | 46% Duration of response | 14.8 months | 9.3 months Survival table Endpoint: Progression-free survival Atirmociclib + fulvestrant | 126 Investigator's choice therapy (ICT) | 137 HR: 0.60 HR CI: 0.440–0.825 P: 0.0007 3 months (printed label): 82.5% vs 68.9% 6 months (printed label): 62.3% vs 51.9% 12 months (printed label): 44.8% vs 22.5% [Slide 3] ESMO / BREASTCANCER·ORG New CDK4 inhibitor slows progression of HR+ metastatic breast cancer FOURLIGHT-1 IS A RANDOMIZED PHASE 2 TRIAL EVALUATING THE EFFICACY AND SAFETY OF ATIRMOCICLIB + FULVESTRANT VS INVESTIGATOR'S CHOICE OF THERAPY (ICT) IN PATIENTS WITH HR+/HER2- MBC WHOSE DISEASE PROGRESSED ON A CDK4/6 INHIBITOR TEAEs in ≥10% of patients in either treatment arm (Atirmociclib | ICT) Any | 97 (35) | 88 (21) Neutropenia | 47 (22) | 8 Leukopenia | 34 | 10 Diarrhea | 28 | 13 Fatigue | 21 | 9 Nausea | 18 | 13 Rash | 18 | 13 AST increased | 14 | 16 Headache | 14 | 4 Hyperglycemia | 14 | 12 Lymphopenia | 14 | 4 Anemia | 13 | 16 Arthralgia | 12 | 10 Vomiting | 12 | 5 Cough | 11 | 10 Thrombocytopenia | 11 | 10 ALT increased | 9 | 14 Back pain | 6 | 13 Stomatitis | 5 | 30 Proportion of patients (%) Atirmociclib Grade 1/2 Atirmociclib Grade 3 Atirmociclib Grade 4 ICT Grade 1/2 ICT Grade 3 ICT Grade 4 Atirmociclib demonstrated a manageable safety and tolerability profile with low rates of serious TEAEs (11.2%) and treatment discontinuations due to TEAEs (6.4%). Relevance Atirmociclib in combination with fulvestrant showed significant and clinically meaningful PFS improvement in patients who progressed on a prior CDK4/6 inhibitor, with a manageable safety and tolerability profile. These findings support the continued development of atirmociclib in earlier lines of therapy, including an ongoing phase 3 study of atirmociclib versus CDK4/6 inhibitors, in combination with letrozole, as first-line treatment of HR+/HER2− mBC.

[Slide 1] MADRID 2026 ESMO congress Efficacy and safety of atirmociclib + fulvestrant in patients with HR+/HER2– advanced/metastatic breast cancer who progressed on prior CDK4/6 inhibitor in a randomized phase 2 study (FOURLIGHT-1) Antonio Giordano, Binghe Xu, Yeon Hee Park, Chiun-Sheng Huang, Guillermo Lela Larsa[?], Huiping Li, Ayelet Shai, Georgina Garnica Jaliffe, Mitsuya Itoh, Anja Williams, Shom Goel, Seock-Ah Im, Patrick Cobb, Marzieh Golmakani[?], Feng Liu, Jason Gong, Alicia Vana, Michail Ignatiadis, Tetsuhiro Yoshinami, Hope S. Rugo Slide deck Plain language summary Copies of the slide deck and plain language summary obtained through the Quick Response (QR) codes are for personal use only and may not be reproduced without permission from ESMO or the authors [Slide 2] Investigator-Assessed PFS | Atirmociclib arm (n = 128) | ICT arm (n = 127) Median follow-up, months | 13.7 | 13.0[?] Events | 72 | 93 Median PFS, months (95% CI) | 9.4 (7.3–13.6) | 6.6 (3.8–7.8) Stratified HR (95% CI) | 0.60 (0.440–0.825) one-sided P value | 0.0007 Curve labels: 82.5% / 68.9% (3 months); 62.3% / 51.9% (6 months); 44.8% / 22.5% (12 months) Y axis: PFS probability (%); X axis: Time (months) 0–20 No. at risk Atirmociclib arm: 128 120 109[?] 98 83 82 72 71 64 63 55 48 34 32 13 12 12 3 3 3 0 [many values blurry] ICT arm: 127 127 90[?] 84[?] 66[?] 65[?] 56[?] 54[?] 40[?] 40[?] 31[?] 26[?] 17 14[?] 7 7 4 1 3[?] 0[?] 0[?] [blurry] CI, confidence interval; HR, hazard ratio; ICT, investigator's choice of therapy; PFS, progression-free survival Presenter: Antonio Giordano Survival table Endpoint: Investigator-assessed PFS Arm | N | Median | Median CI Atirmociclib + fulvestrant | 128 | 9.4 months | 7.3–13.6 ICT (investigator's choice of therapy) | 127 | 6.6 months | 3.8–7.8 HR: 0.60 HR CI: 0.440–0.825 P: 0.0007 (one-sided) Landmark rates: 3 months: Atirmociclib 82.5% | ICT 68.9% 6 months: Atirmociclib 62.3% | ICT 51.9% 12 months: Atirmociclib 44.8% | ICT 22.5%

[search icon] Proffered Paper session LBA23 - Final results of TALENT: A neoadjuvant trial of T-DXd with or without anastrozole, or in sequence with chemotherapy for HER2-low, HR+ early-stage breast cancer Speakers Sara A. Hurvitz (Seattle, United States of America) Lecture Time 13:30 - 13:40 Abstract Background Pathologic complete response (pCR) in HR+/HER2-negative breast cancer (BC) is uncommon. T-DXd, a HER2-targeted antibody drug conjugate, improves progression-free and overall survival vs chemotherapy in metastatic HER2-low BC. TALENT (NCT04553770) is a phase II, 2-stage neoadjuvant study evaluating T-DXd + endocrine therapy or in sequence with chemotherapy in pts with HR+/HER2-low early BC. Methods Pts with untreated, operable invasive early HR+/HER2-low BC were randomized 1:1 to receive 6 cy of T-DXd (5.4 mg/kg q21d) alone (Arm A) or with anastrozole (1 mg PO QD) (Arm B). An arm was expanded if ≥2 pCR (RCB0) were observed. After 29 pts enrolled, subsequent pts received 8 cy. After Arms A/B, Arm C was added to evaluate 4 cy T-DXd → 4 cy docetaxel 75 mg/m2 + cyclophosphamide 600 mg/m2 (TC) in <58 pts. Biopsies were taken at baseline, pre-cy2 & surgery. Imaging/biomarkers were assessed at baseline, cy 2 & surgery. Endpoints: Residual Cancer Burden (RCB), safety, overall response rate (ORR) & biomarkers. Ki67, HER2, ER & PR were evaluated centrally, and TROP2 & HER2 levels were quantified (quant) by TROPLEX for Arms A & B. Results From 11/2020 - 02/2026, 132 pts were enrolled. Pts in Arm C were more likely to have node+ disease (81%) vs. Arms A/B (53%, 55%). Overall rates of pCR/near-pCR (RCB 0/1) were 16% (Arm A), 17% (Arm B), and 17% (10/58 Arm C). HER2 levels were higher in pts with RCB0 (2691 & 1221 amol/mm2) vs RCB1/2/3 (715-947 & 440-608 amol/mm2) at Baseline & cy 2, respectively. Table: LBA23 [Table] | Arm A (N=45) n (%) | Arm B (N=29) n (%) | Arm C (N=58) n (%) Baseline characteristics Stage I/II/III | 0/32 (71)/13 (29) | 0/26 (90)/3 (10) | 2 (3)/43 (74)/13 (22) Node+ | 24 (53) | 16 (55) | 47 (81) Grade III | 19 (42) | 11 (38) | 29 (50) Ki67>20% (central) | 33 (73) | 24 (83) | 43 (74) Antigen quant (med amol/mm2) Baseline→Cy2→Surgery HER2 | 707→334→150 | 778→331→83 | - TROP2 | 1097→1371→10011001→952→925- [as printed; columns run together] Results: Evaluable RCB0 | 3/42 (7) | 1/29 (3) | 5/55 (9) ORR | 28/42 (67) | 19/29 (66) | 37/57 (65) Results ITT | 6 cycles (N=15) | 8 cycles (N=30) | 6 cycles (N=14) | 8 cycles (N=15) | 8 cycles (N=58) RCB0 | 0 | 3 (10) | 0 | 1 (7) | 5 (9) RCB1 | 2 (13) | 2 (7) | 1 (7) | 3 (20) | 5 (9) Gr >3 Treatment Emergent AE | 3 (20) | 8 (27) | 3 (21) | 2 (13) | 16 (28) Interstitial Lung Disease (all Gr 2) | 0 | 0 | 1 (7) | 0 | 2 (3) Conclusions In this first study to report outcomes with neoadjuvant T-DXd in HR+/HER2-low BC, T-DXd–based therapy was associated with low pCR rates. ILD was rare. HER2 was higher in tumors with RCB0 versus RCB1/2/3, suggesting HER2 level may correlate with treatment outcome. HER2 levels declined significantly on treatment, whereas TROP2 expression remained stable. ER, PR, & Ki67 analyses will be presented.

[Slide 1] 13:30 - 15:00 Proffered paper 1: Breast cancer, early and advanced stage CHAIRS: KOMAL JHAVERI, BARBARA PISTILLI Conclusions • Neoadjuvant T-DXd demonstrated modest clinical activity in HER2 low, HR+ localized breast cancer • Addition of endocrine therapy to T-DXd did not improve pCR rate • pCR rate T-DXd: 7%, T-DXd + anastrozole 3% • Sequencing T-DXd with multiagent chemotherapy only marginally improved pCR rate • pCR rate T-DXd: 7%, T-DXd->TC 9% • Safety: Nausea and fatigue were the most common AE; 3 cases of grade 2 ILD; 1 patient death • HER2 protein levels decreased in Arms A&B after treatment with T-DXd • Higher baseline HER2 protein level may be associated with pCR in Arms A&B but sample size is small and data from Arm C are pending • Ongoing analyses include: • Event free survival • Breast conserving surgery rates • Biomarkers: ctDNA, whole genome sequencing, intrinsic subtyping, • multiplex IHC, RNA seq transcriptomic analyses • Patient-reported outcomes • Ovarian function Sara Hurvitz Final results of TALENT: A neoadjuvant trial of T-DXd with or without anastrozole, or in sequence with chemotherapy for HER2-low, HR+ early-stage breast cancer ALICANTE AUDITORIUM - HALL 6 [Slide 2] 13:30 - 15:00 Proffered paper 1: Breast cancer, early and advanced stage CHAIRS: KOMAL JHAVERI, BARBARA PISTILLI TALENT: Biomarker Collection & Imaging Assessments Study Population: •Hormone receptor-positive •HER2-low (by local and/or central review) •Stage II-III •Operable •Men or Pre-/Post-menopausal women N=58 (1:1) Arm A T-DXd 5.4mg/kg Arm B T-DXd 5.4mg/kg + anastrozole (+GnRH analog for men/premenopausal women) Arm C T-DXd x 4 -> TC x 4 -> Surgery** • Tumor sample from archival or fresh biopsy at baseline, biopsy between C1D17-C1D21, and tumor sample at surgery. • Circulating tumor DNA analysis collected at baseline, C1D1, C2D1 and end of treatment • Tumor imaging performed (ultrasound and mammogram) at baseline and EOT. Ultrasound or mammogram at Cycle 2 (modality at the discretion of the investigator) All tissue collected for study: pathology centrally reviewed Sara A. Hurvitz, MD Sara Hurvitz Final results of TALENT: A neoadjuvant trial of T-DXd with or without anastrozole, or in sequence with chemotherapy for HER2-low, HR+ early-stage breast cancer ALICANTE AUDITORIUM - HALL 6

[Slide 1] TALENT: Addition of Third Arm Study Population: •Hormone receptor-positive •HER2-low (by local and/or central review) •Stage II-III •Operable •Men or Pre-/Post-menopausal women N=58 (1:1) -> Arm A T-DXd 5.4mg/kg ; Arm B T-DXd 5.4mg/kg + anastrozole (+GnRH analog for men/premenopausal women) N=29 -> Arm C T-DXd x 4 -> TC x 4 Stage I: Any arm with ≥2 pCR pass to Stage II expansion Stage II: Arm A [check] -> Enroll additional patients; Arm B [X] -> STOP; Arm C [check] -> Enroll additional patients* *Arm C started enrollment after completion of Arm A/B [GnR]H: gonadotropin releasing hormone; Hormone receptor: estrogen receptor and/or progesterone receptor ≥1% by immunohistochemistry on local testing; pCR: [path]ologic complete response (ypT0/is, ypN0) Sara A. Hurvitz, MD [Slide 2] Baseline Characteristics (continued) PATIENT CHARACTERISTICS | Arm A (T-DXd) N=45 | Arm B (T-DXd + Anastrozole) N = 29 | Arm C (T-DXd x4 -> TC x4) N = 58 Baseline Ki67 Scores (Central Review) <10%: 2 (4%) | 0 | 8 (14%) ≥10% and <30%: 20 (44%) | 10 (34%) | 19 (33%) ≥30%: 23 (51%) | 19 (66%) | 31 (53%) Tumor Grade (local) Grade 1: 4 (9%) | 1 (3%) | 0 Grade 2: 22 (49%) | 16 (55%) | 29 (50%) Grade 3: 19* (42%) | 11** (38%) | 29 (50%) Not reported: 0 | 1 (3%) | 0 * 1 patient with single lesion and multiple grades (2 & 3) included in grade 3 only ** 2 patients with 2 masses and multiple grades (2 & 3) included in grade 3 only Sara A. Hurvitz, MD [Slide 3] Objective Response Rate Arm A: ORR 28/45 = 62%; CR 5/45 = 11%; PR 23/45 = 51% Arm B: ORR 19/29 = 66%; CR 2/29 = 7%; PR 17/29 = 59% Arm C: ORR 42/58 = 72%; CR 9/58 = 16%; PR 33/58 = 57% [Waterfall plots: Arm A (T-DXd) N=45; Arm B (T-DXd + Anastrozole) N=29; Arm C (T-DXd x 4 -> TC x4) N=58; legend CR / PR / SD / PD] [...]patients did not have EOT imaging, but included [in] intention to treat (ITT) denominator for ORR [ana]lysis per protocol 3 patients did not have EOT imaging, but included in ITT denominator for ORR analysis per protocol [Wat]erfall plot with bars representing change in tumor size after treatment with T-DXd, compared to baseline, as per RECIST v1.1. Intention to treat population for ORR includes all who received at least 1 cycle of protocol therapy.*patients who received 6 cycles [Slide 4] Pathologic Response (by arm and cycles) | Arm A T-DXd | Arm B T-DXd/anastrozole | Arm C T-DXd->TC pCR Rate: 3/45 = 7% | 1/29 = 3% | 5/58 = 9% RCB0/1: 7/45 = 16% | 5/29 = 17% | 10/58 = 17% [Stacked bar chart, Percent; RCB1 / RCB0] Arm A (N=45)*: 6 cycles (N=15): RCB1 13%; 8 cycles (N=30): RCB0 10%, RCB1 7% Arm B (N=29): 6 cycles (N=14): RCB1 7%; 8 cycles (N=15): RCB0 7%, RCB1 20% Arm C (N=58)**: 8 cycles (N=58): RCB0 9%, RCB1 9% *3 pts with no surgery results enrolled to 8 cycles of treatment (1 – Pt. withdrew consent; 1 – Pt. expired; 1 – Pt. had PD); **2 pts with no surgery results (withdrew consent/loss to follow-up); RCBi = Residual cancer burden index; RCB 0 = pCR; Histology or IHC Status did not appear to be associated with RCB response Sara A. Hurvitz, MD

[Slide 1] CHAIRS: KOMAL JHAVERI, BARBARA PISTILLI Pathologic Response (by arm and cycles) | Arm A T-DXd | Arm B T-DXd/anastrozole | Arm C T-DXd->TC pCR Rate | 3/45 = 7% | 1/29 = 3% | 5/58 = 9% RCB0/1 | 7/45 = 16% | 5/29 = 17% | 10/58 = 17% [Stacked bar chart, Percent 0-30; RCB1 (orange), RCB0 (blue)] Arm A (N=45)*: 6 cycles (N=15): RCB1 13%; 8 cycles (N=30): RCB0 10%, RCB1 7% Arm B (N=29): 6 cycles (N=14): RCB1 7%; 8 cycles (N=15): RCB0 7%, RCB1 20% Arm C (N=58)**: 8 cycles (N=58): RCB0 9%, RCB1 9% *3 pts with no surgery results enrolled to 8 cycles of treatment (1 - Pt. withdrew consent; 1 - Pt. expired; 1 - Pt. had PD); **2 pts with no surgery results (withdrew consent/loss to follow-up); RCBi = Residual cancer burden index; RCB 0 = pCR; Histology or IHC Status did not appear to be associated with RCB response Sara A. Hurvitz, MD [Slide 3] HER2 Protein Quantification (ammol/mm2) by TROPLEX(TM) [Box plot: HER2 levels at Baseline, Cycle 2, Surgery; legend and axis labels illegible] - HER2 protein levels were similar for Arms A and B - HER2 levels significantly decreased over time (p <0.05) after treatment with T-DXd - Arm A: 707 -> 334 -> 150 - Arm B: 778 -> 331 -> 83 Arm A: T-DXd (n/N [illegible]); Arm B: T-DXd + anastrozole ([illegible]) Sample size with Tissue: Baseline (n=[illegible]) Cycle 2 (n=47) Surgery (n=44) [partially legible] [Footer references in small blurred text - illegible] Hurvitz et al ALICANTE AUDITORIUM - HALL 6 [partially visible] [Slide 4] 13:30 - 15:00 Proffered paper 1: Breast cancer, early and advanced stage CHAIRS: KOMAL JHAVERI, BARBARA PISTILLI Expression of Ki-67 (IHC) by arm and timepoint [Box plot: Ki-67 (% of positive cells), 0-80, at Baseline, Cycle 2, Surgery; legend Arm A / Arm B / Arm C] - Ki-67 decreased from baseline to cycle 2 in all arms - Arm A: p=0.015 - Arm B: p<0.001 - Arm C: p<0.004 - Ki67 decreased from baseline to surgery in all arms - Arm A, B, C: p<0.001 Samples available: Arm A: T-DXd - Baseline (45/45), Cycle 2 (41/45), Surgery (36/45); Arm B: T-DXd + anastrozole - Baseline (29/29), Cycle 2 (27/29), Surgery (28/29); Arm C: T-DXd x4 -> TC x4 - Baseline (58/58), Cycle 2 (56/58), Surgery (46/58) Breast cancer biomarker status was evaluated in a CLIA-certified and CAP-accredited central laboratory (Breast Cancer Analysis Laboratory, USC / Norris Comprehensive Cancer Center, LA, CA) using IHC. Immunohistochemistry (IHC) methods were performed using formalin-fixed, paraffin-embedded tissue sections with validated(1-4) primary anti-ER (6F11), anti-PR (PGR636), anti-Ki-67 (MIB-1), and anti- HER2 (SK001, HercepTest) antibodies. Antigen retrieval techniques and secondary antibodies with conjugated horseradish peroxidase-labelled-dextran polymer were used in all IHC assays. Interpretation involved an assessment of percent positively IHC-stained carcinoma cells for each biomarker according to ASCO-CAP guidelines. 1. Steroids 67: 795-813, 2002; 2. Breast Cancer Research and Treatment 172 (2), 327-338, 2018; 3. Journal of Clinical Oncology 20: 3095-3105, 2002; 4. Journal of Clinical Oncology 34 (29): 3518-3528, 2016.

[Slide 1] Proffered Paper session LBA23 - Final results of TALENT: A neoadjuvant trial of T-DXd with or without anastrozole, or in sequence with chemotherapy for HER2-low, HR+ early-stage breast cancer Speakers Sara A. Hurvitz (Seattle, United States of America) Lecture Time 13:30 - 13:40 [Abstract] Background Pathologic complete response (pCR) in HR+/HER2-negative breast cancer (BC) is uncommon. T-DXd, a HER2-targeted antibody drug conjugate, improves progression-free and overall survival vs chemotherapy in metastatic HER2-low BC. TALENT (NCT04553770) is a phase II, 2-stage neoadjuvant study evaluating T-DXd + endocrine therapy or in sequence with chemotherapy in pts with HR+/HER2-low early BC. [Slide 2] Conclusions In this first study to report outcomes with neoadjuvant T-DXd in HR+/HER2-low BC, T-DXd−based therapy was associated with low pCR rates. ILD was rare. HER2 was higher in tumors with RCB0 versus RCB1/2/3, suggesting HER2 level may correlate with treatment outcome. HER2 levels declined significantly on treatment, whereas TROP2 expression remained stable. ER, PR, & Ki67 analyses will be presented. Clinical trial identification NCT04553770. Legal entity responsible for the study Translational Research In Oncology - US (TRIO-US).

[Slide 1] TALENT: Biomarker Collection & Imaging Assessments Study Population: - Hormone receptor-positive - HER2-low (by local and/or central review) - Stage II-III - Operable - Men or Pre-/Post-menopausal women N=58 (1:1) Arm A: T-DXd 5.4mg/kg Arm B: T-DXd 5.4mg/kg + anastrozole (+GnRH analog for men/premenopausal women) Arm C: T-DXd x 4 -> TC x 4 -> Surgery** - Tumor sample from archival or fresh biopsy at baseline, biopsy between C1D17-C1D21, and tumor sample at surgery. - Circulating tumor DNA analysis collected at baseline, C1D1, C2D1 and end of treatment - Tumor imaging performed (ultrasound and mammogram) at baseline and EOT. Ultrasound or mammogram at Cycle 2 (modality at the discretion of the investigator) All tissue collected for study: pathology centrally reviewed Sara A. Hurvitz, MD [Slide 2] Baseline Characteristics | Arm A (N = 45) | Arm B (N = 29) | Arm C (N = 58) Age, years, median (range) | 57 (33 - 87) | 55 (32 - 83) | 52 (29 - 79) Race American Indian/Alaska Native | 0 | 0 | 1 (2%) Asian | 9 (20%) | 3 (10%) | 12 (21%) Black or African American | 2 (4%) | 1 (3%) | 1 (2%) Decline to answer/Other/Unknown | 5 (11%) | 1 (3%) | 16 (28%) White | 29 (64%) | 24 (83%) | 28 (48%) Menopausal Status Post- (2 men in Arm A) | 28 (62%) | 17 (59%) | 26 (45%) [boxed in red] Baseline HER2 Expression (Central Review)* 0** | 3 (7%) | 2 (7%) | 0 1+ | 32 (71%) | 25 (86%) | 36 (62%) 2+ | 9 (20%) | 2 (7%) | 22 (38%) Histology Invasive Ductal | 40 (89%) | 26 (90%) | 53 (91%) Invasive Lobular | 5 (11%) | 3 (10%)*** | 5 (9%) Baseline Hormone Receptor Status (Local Review) ER+ and PR+ | 39 (87%) | 26 (90%) | 48 (83%) Lymph Node Status Node positive | 24 (53%) | 16 (55%) | 47 (81%) [boxed in red] Stage IB | 0 | 0 | 2 (3%) IIA/IIB | 12 (27%) / 20 (44%) | 11 (38%) / 15 (52%) | 18 (31%) / 25 (43%) IIIA/IIIB/IIIC | 10 (22%) / 2 (4%) / 1 (2%) | 3 (10%) / 0 / 0 | 8 (14%) / 4 (7%) / 1 (2%) *1 Arm A patient with 1+ and 2+ (multicentric lesion); **Baseline HER2 expression was 1+ or 2+ based on local review; ***1 patient with mixed ductal/lobular

HER2 Protein Quantification (ammol/mm²) by TROPLEX™ [Box plot: HER2 (amol/mm2), log scale 20-20000, Arm A (blue) vs Arm B (pink) at Baseline, Cycle 2, Surgery; Compared to baseline * p<0.05, ** p<0.01, *** p<0.001; * at Cycle 2, ** and *** at Surgery] • HER2 protein levels were similar for Arms A and B • HER2 levels significantly decreased over time (p <0.05) after treatment with T-DXd • Arm A: 707→334→150 • Arm B: 778→331→83 Arm A: T-DXd (29/45); Arm B: T-DXd + anastrozole (21/29) Sample size with Tissue: Baseline (n=50) Cycle 2 (n=47) Surgery (n =44) • Measured in the Quantitative Diagnostics in Anatomic Pathology (QDAP) CLIA lab at Yale Univ Dept of Pathology; Lines on box plots indicate median • The Troplex assay is an analytic assay that uses quantitative immunofluorescence, standardized to cell lines with mass spec measured targets with a CV<10. * Robbins CJ, He M, Chan N, Khaimova R, Bates K, Trontzas IP, Scott L, Moutafi M, Coleman CB, Hill S, Liebler DC, Fulton R, Rimm DL. Quantitative Multiplex Immunofluorescence Assay for Trophoblast Cell-Surface Antigen 2 and Human Epidermal Growth Factor Receptor 2 Expression in Breast Cancer: Toward Guiding Patient Selection for Antibody-Drug Conjugate Therapies. JCO Precis Oncol. 2025 Jul;9:e2500128. doi: 10.1200/PO-25-00128. Epub 2025 Jul 16. PMID: 40669019; PMCID: PMC12490811. Hurvitz et al.

Pathologic Response (by arm and cycles) | | Arm A | Arm B | Arm C | | pCR Rate | 3/45 = 7% | 1/29 = 3% | 5/58 = 9% | | RCB0/1 | 7/45 = 16% | 5/29 = 17% | 10/58 = 17% | Percent (y-axis 0-30) Arm A (N=45)*: 6 cycles (N=15): RCB1 13% | 8 cycles (N=30): RCB0 10%, RCB1 7% Arm B (N=29): 6 cycles (N=14): RCB1 7% | 8 cycles (N=15): RCB0 7%, RCB1 20% Arm C (N=58)**: 8 cycles (N=58): RCB0 9%, RCB1 9% Legend: RCB1, RCB0 *3 pts with no surgery results enrolled to 8 cycles of treatment (1 – Pt. withdrew consent; 1 – Pt. expired; 1 – Pt. had PD); **2 pts with no surgery results (withdrew consent/loss to follow-up); RCBi = Residual cancer burden index; RCB 0 = pCR; Histology or IHC Status did not appear to be associated with RCB response Figure. Pathological response to T-DXd alone (Arm A), T-DXd plus anastrozole (Arm B) and T-DXd followed by chemotherapy in patients with HR-positive, HER2-low, stage II–III breast cancer (ESMO Congress 2026, LBA23)

Safety Summary | | Arm A (N = 45) T-DXd n (%) | Arm B (N = 29) T-DXd/anastrozole n (%) | Arm C (N = 58) T-DXd→TC n (%) | Treatment Related Adverse Events (TRAEs) Grade ≥ 3 | 11 (24) | 5 (17) | 16 (28) Treatment Related SAEs | 2 (4) | 1 (3) | 10 (17) TRAEs leading to any dose reduction | 4 (9) | 0 | 9 (16) TRAEs leading to any drug interruption | 5 (11) | 3 (10) | 5 (9) TRAEs leading to any treatment discontinuation | 2 (4) | 3 (10) | 5 (9) Any TRAEs with outcome of death | 1 (2) | 0 | 0 Adverse Event of Special Interest (Drug-related ILD/pneumonitis) Grade 2 (all were grade 2) | 0 | 1 (3) | 2 (3) Sara A. Hurvitz, MD

Final results of TALENT: A neoadjuvant trial of T-DXd with or without anastrozole, or in sequence with chemotherapy for HER2-low, HR+ early-stage breast cancer Sara A. Hurvitz, Nicholas P. McAndrew, David Chan, Lisa S. Wang, Vu Phan, Thomas Lomis, Nina M. Lamble, Justin Tiulim, Brian DiCarlo, Deborah Villa, Nihal Abdulla, Giribala Patel, Aashini K. Master, Swati Sikaria, Shaker Dakhil, Vidhya Nair, Michael F. Press, David L. Rimm, Chi-Hong Tseng, Aditya Bardia Presenting Author: Sara A. Hurvitz, MD Fred Hutch Cancer Center University of Washington Seattle, WA, USA 23 October 2026

[Session banner, partially cut: CHAIRS: KOMAL JHAVERI, BARBARA PISTILLI] TALENT: Study Design Study Population: •Hormone receptor-positive •HER2-low (by local and/or central review) •Stage II-III •Operable •Men or Pre-/Post-menopausal women N=58 (1:1) Stratified by HER2 expression level (1+ vs 2+) and menopausal status (pre vs post) Arm A T-DXd 5.4mg/kg Treatment: 6 or 8 cycles* + EOT Arm B T-DXd 5.4mg/kg + anastrozole (+GnRH analog for men/premenopausal women) -> Surgery** *Originally, 6 cycles of treatment were given but after 29 patients were enrolled, an amendment increased the number of treatment cycles to 8 cycles **After surgery, adjuvant therapy was per the discretion of the treating provider

Study Objectives & Statistical Considerations • Primary Objective: • Evaluate pathologic complete response (pCR) rate in breast and lymph nodes (ypT0/is ypTN0; residual cancer burden index, RCBi = 0) • Secondary Objectives: • Objective response rate (ORR) • Pathological assessment according to RCBi at surgery • Tumor biomarkers: HER2, ER, PR, Ki-67, Trop2 • Safety • Assess the impact of T-DXd or T-DXd followed by chemotherapy on ovarian function • Exploratory Objectives: • Serum and tumor predictive biomarkers • Event free survival • Patient reported outcomes • Simon's minimax two-stage study design: • 58 participants randomly assigned 1:1 to one of two treatment arms • No formal statistical comparison between the arms • pCR ≤ 5% vs > 15% as statistical benchmark for Simon 2 stage design [presenter name cut off at bottom edge]

Proffered Paper session LBA13 - Trastuzumab rezetecan (SHR-A1811) in HER2-low advanced breast cancer (BC): A randomized, phase 3 study (HORIZON-Breast06) Speakers Ying Fan (Beijing, China) Lecture Time 13:40 - 13:50 Abstract Background HORIZON-Breast06 is a pivotal, randomized, multicenter, open-label, phase 3 study comparing trastuzumab rezetecan (a HER2 ADC) with physician's choice of chemotherapy in patients (pts) with HER2-low advanced BC who had received ≤1 prior line of chemotherapy for metastatic disease, with prior endocrine therapy required for hormone receptor (HR)-positive disease. Methods Pts were randomized (1:1) to receive trastuzumab rezetecan (6.4 mg/kg IV, Q3W) or physician's choice of chemotherapy (taxanes, eribulin, capecitabine, or gemcitabine). The co-primary endpoints were PFS by blinded independent central review (BICR) in chemo-naïve (0 prior line of chemotherapy for metastatic disease) and overall pts. Results As of Mar 20, 2026, 551 pts were randomized (trastuzumab rezetecan, n=277; chemotherapy, n=274; chemo-naïve: 81.5%; HR-positive: 73.0%; prior CDK4/6 inhibitor: 64.8%). Median follow-up was 18.0 months (95% CI 16.8–19.3) for all patients. Trastuzumab rezetecan significantly prolonged PFS by BICR in both chemo-naïve (median 16.6 vs. 6.8 months; HR 0.25, 95% CI 0.19–0.33; P<0.0001; table) and overall pts (median 16.4 vs. 6.8 months; HR 0.27, 95% CI 0.21–0.34; P<0.0001). The PFS benefit was consistent across all prespecified subgroups, including HR-negative disease (16.5 vs. 6.8 months; HR 0.34) and HR-positive disease (16.4 vs. 6.7 months; HR 0.24). OS data remain immature, with a trend favoring trastuzumab rezetecan. Grade ≥3 treatment-related adverse events occurred in 84.1% (trastuzumab rezetecan) and 67.9% (chemotherapy) of patients, respectively. Interstitial lung disease (ILD) was reported in only 7 pts (2.5%) with trastuzumab rezetecan. Table: LBA13 [Table] Chemo-naïve patients: Trastuzumab rezetecan (N=225) | Chemotherapy (N=224) | Overall patients: Trastuzumab rezetecan (N=277) | Chemotherapy (N=274) PFS Events, n (%) | 91 (40.4) | 146 (65.2) | 130 (46.9) | 185 (67.5) Median (95% CI), months | 16.6 (14.0–19.4) | 6.8 (5.7–7.0) | 16.4 (14.0–16.8) | 6.8 (5.7–6.9) 12-mo rate, % (95% CI) | 65.2 (57.6–71.8) | 18.9 (12.8–25.8) | 63.1 (56.4–69.1) | 18.0 (12.7–24.1) HR (95% CI); 1-sided P | 0.25 (0.19–0.33); P<0.0001 | 0.27 (0.21–0.34); P<0.0001 OS Events, n (%) | 43 (19.1) | 55 (24.6) | 66 (23.8) | 77 (28.1) 12-mo rate, % (95% CI) | 92.0 (87.3–95.1) | 83.0 (76.9–87.6) | 90.6 (86.2–93.6) | 81.4 (75.9–85.8) ORR, % (95% CI) | 76.0 (69.9–81.4) | 43.3 (36.7–50.1) | 75.5 (69.9–80.4) | 41.2 (35.4–47.3) Conclusions Trastuzumab rezetecan significantly improved PFS over physician's choice of chemotherapy in pts with HER2-low advanced BC regardless of prior lines of chemotherapy or HR status, with a trend toward OS benefit and a favorable safety profile including a low incidence of ILD.

Primary endpoint: PFS in chemo-naïve patients (per BICR) [Kaplan-Meier plot; y-axis: Progression-free survival (%); x-axis: Time (months) 0-20] Legend: — Trastuzumab rezetecan; — Physician's choice; + Censor Table: | Trastuzumab rezetecan | Physician's choice Median PFS (95% CI), months | 16.6 (14.0-19.4) | 6.8 (5.7-7.0) Harzard ratio (95% CI) | 0.25 (0.19-0.33) P<0.001 No. at Risk Trastuzumab rezetecan: 225 216 207 197 193 182 170 155 147 130 114 103 84 83 67 62 59 34 33 32 18 Physician's choice: 224 214 167 145 138 113 88 67 59 42 31 27 19 18 14 12 11 7 7 7 4 Presenter: Ying Fan MD PFS, progression-free survival, BICR, blinded independent central review, Survival table Endpoint: PFS per BICR, chemo-naïve patients (primary endpoint) Arm | N | Median | Median CI Trastuzumab rezetecan | 225 | 16.6 months | 14.0-19.4 Physician's choice | 224 | 6.8 months | 5.7-7.0 HR: 0.25 HR CI: 0.19-0.33 P: P<0.001

[Slide 1] HORIZON-Breast06 study design Randomized, open-label, multicenter, phase 3 trial (NCT05814354) Eligibility criteria • Centrally confirmed HER2-low (IHC 1+ or IHC 2+/ISH–) metastatic breast cancer • ≤1 prior line of chemotherapy for metastatic disease • Prior endocrine therapy required for HR-positive disease. R 1:1 Trastuzumab rezetecan (6.4 mg/kg, IV, Q3W) Physician's choice of chemotherapy (taxanes, eribulin, capecitabine, or gemcitabine) Primary endpoint: PFS (per BICR) • in chemo-naïve patients • in overall patients Key secondary endpoints: • OS in overall patients • OS in chemo-naïve patients • PFS (per BICR) in chemo-treated patients Stratification factors: ✓ Prior lines of chemotherapy for metastatic disease (0 vs. 1) ✓ HER2 status (IHC 1+ vs. IHC 2+/ISH–) ✓ HR status (HR+ with prior CDK4/6 inhibitor vs. HR+ without prior CDK4/6 inhibitor vs. HR–) Presenter: Ying Fan MD HER2, human epidermal growth factor receptor 2, IHC, immunohistochemistry, ISH, in situ hybridization, IV, Intravenous, PFS, progression-free survival, BICR, blinded independent central review, OS, overall survival, HR, hormone receptor, CDK4/6, cyclin-dependent kinase 4 and 6 [Slide 2] PFS in HR-positive and HR-negative patients (per BICR) Fig.1 PFS in HR-positive patients [Kaplan-Meier plot; y-axis Progression-free survival (%); x-axis Time (months) 0-20] Median PFS (95% CI), months: Trastuzumab rezetecan 16.4 (13.9-16.8) | Physician's choice 6.7 (5.7-7.0) Harzard ratio (95% CI): 0.24 (0.18-0.33) No. at Risk Trastuzumab rezetecan: 201 191 183 177 173 164 155 140 133 119 105 96 80 79 65 60 57 33 32 31 16 Physician's choice: 201 191 153 130 125 100 75 58 50 34 25 21 14 13 8 7 6 3 3 3 2 Fig.2 PFS in HR-negative patients Median PFS (95% CI), months: Trastuzumab rezetecan 16.5 (11.5-27.7) | Physician's choice 6.8 (4.1-8.5) Harzard ratio (95% CI): 0.34 (0.22-0.54) No. at Risk Trastuzumab rezetecan: 76 74 71 66 65 59 53 51 47 42 38 35 30 30 25 24 23 14 14 14 10 Physician's choice: 73 71 51 44 41 35 30 24 22 17 15 14 10 10 9 7 6 5 5 4 2 Legend: Trastuzumab rezetecan; Physician's choice; + Censor Presenter: Ying Fan MD PFS, progression-free survival, BICR, blinded independent central review; HR, hormone receptor Survival table Endpoint: PFS per BICR by HR status (overall population subgroups) HR+ Trastuzumab rezetecan | 201 | 16.4 months | 13.9-16.8 HR+ Physician's choice | 201 | 6.7 months | 5.7-7.0 HR− Trastuzumab rezetecan | 76 | 16.5 months | 11.5-27.7 HR− Physician's choice | 73 | 6.8 months | 4.1-8.5 HR: HR+ 0.24; HR− 0.34 HR CI: HR+ 0.18-0.33; HR− 0.22-0.54 [Slide 3] Overall survival Fig.1 Overall survival in overall patients | Trastuzumab rezetecan | Physician's choice Events, n (%): 66 (23.8) | 77 (28.1) 12-month OS rate % (95% CI): 90.6 (86.2-93.6) | 81.4 (75.9-85.8) Hazard ratio (95% CI): 0.67 (0.48-0.94) No. at Risk Trastuzumab rezetecan: 277 268 257 233 193 153 121 85 48 15 3 Physician's choice: 274 258 240 195 163 137 92 61 28 11 2 0 Fig.2 Overall survival in chemo-naïve patients Events, n (%): 43 (19.1) | 55 (24.6) 12-month OS rate % (95% CI): 92.0 (87.3-95.1) | 83.0 (76.9-87.6) Hazard ratio (95% CI): 0.62 (0.41-0.93) No. at Risk Trastuzumab rezetecan: 225 218 209 187 150 113 86 52 27 8 0 Physician's choice: 224 210 196 157 127 104 63 39 14 4 2 0 [Kaplan-Meier plots; y-axis Overall survival (%); x-axis Time (months) 0-33] Overall survival data remained immature. Further analysis of overall survival, including an interim analysis at 274 OS events (maturity: ~52%) and a final analysis at 392 OS events (maturity: ~74%) are planned. Presenter: Ying Fan MD Survival table Endpoint: Overall survival (immature) Overall: Trastuzumab rezetecan | 277 Overall: Physician's choice | 274 Chemo-naïve: Trastuzumab rezetecan | 225 Chemo-naïve: Physician's choice | 224 HR: Overall 0.67; chemo-naïve 0.62 HR CI: Overall 0.48-0.94; chemo-naïve 0.41-0.93 Landmark rates: 12 months: Overall: Trastuzumab rezetecan 90.6% (CI 86.2-93.6) | Overall: Physician's choice 81.4% (CI 75.9-85.8) | Chemo-naïve: Trastuzumab rezetecan 92.0% (CI 87.3-95.1) | Chemo-naïve: Physician's choice 83.0% (CI 76.9-87.6) [Slide 4] Conclusion • Compared physician's choice of chemotherapy, trastuzumab rezetecan demonstrated: – Superiority in PFS: ✓ Chemo-naïve patients: 16.6 vs. 6.8 months (HR 0.25 [95% CI 0.19–0.33]; P<0.001) assessed by BICR ✓ Overall patients: 16.4 vs. 6.8 months (HR 0.27 [95% CI 0.21–0.34]; P<0.001) assessed by BICR – Consistent PFS benefit across all prespecified subgroups: ✓ HR-negative cohort: 16.5 vs. 6.8 months (HR 0.34 [95% CI 0.22–0.54]) ✓ HR-positive cohort: 16.4 vs. 6.7 months (HR 0.24 [95% CI 0.18–0.33]) • Safety was manageable including a low incidence of ILD. Presenter: Ying Fan MD

Primary endpoint: PFS in chemo-naïve patients (per BICR) [KM curve: Progression-free survival (%) vs Time (months) 0-20] Legend: Trastuzumab rezetecan / Physician's choice / + Censor | Trastuzumab rezetecan | Physician's choice Median PFS (95% CI), months | 16.6 (14.0-19.4) | 6.8 (5.7-7.0) Harzard ratio (95% CI) | 0.25 (0.19-0.33) P<0.001 No. at Risk (months 0-20): Trastuzumab rezetecan: 225 216 207 197 193 182 170 155 147 130 114 103 84 83 67 62 59 34 33 32 18 Physician's choice: 224 214 167 145 138 113 88 67 59 42 31 27 19 18 14 12 11 7 7 7 4 Presenter: Ying Fan MD Survival table Endpoint: PFS per BICR, chemo-naive patients (primary endpoint) Arm | N | Median | Median CI Trastuzumab rezetecan | 225 | 16.6 months | 14.0-19.4 Physician's choice | 224 | 6.8 months | 5.7-7.0 HR: 0.25 HR CI: 0.19-0.33 P: P<0.001

Baseline characteristics Columns: Chemotherapy-naïve patients [Trastuzumab rezetecan (N=225) | Physician's choice (N=224)] | Chemotherapy-treated patients [Trastuzumab rezetecan (N=52) | Physician's choice (N=50)] | Overall patients [Trastuzumab rezetecan (N=277) | Physician's choice (N=274)] Median age, years (IQR): 56.0 (50.0-62.0) | 56.0 (49.0-62.5) | 53.5 (45.5-59.0) | 54.0 (47.0-60.0) | 55.0 (49.0-62.0) | 55.0 (49.0-61.0) ≥65 years, n (%): 48 (21.3) | 45 (20.1) | 6 (11.5) | 5 (10.0) | 54 (19.5) | 50 (18.2) Female, n (%): 225 (100.0) | 224 (100.0) | 52 (100.0) | 50 (100.0) | 277 (100.0) | 274 (100.0) ECOG PS score of 1, n (%): 108 (48.0) | 111 (49.6) | 32 (61.5) | 31 (62.0) | 140 (50.5) | 142 (51.8) HER2 status, n (%) IHC 1+ and ISH unknown: 137 (60.9) | 134 (59.8) | 33 (63.5) | 34 (68.0) | 170 (61.4) | 168 (61.3) IHC 2+ and ISH–: 78 (34.7) | 79 (35.3) | 18 (34.6) | 16 (32.0) | 96 (34.7) | 95 (34.7) IHC 1+ and ISH–: 10 (4.4) | 11 (4.9) | 1 (1.9) | 0 | 11 (4.0) | 11 (4.0) Hormone receptor positive, n (%): 166 (73.8) | 166 (74.1) | 35 (67.3) | 35 (70.0) | 201 (72.6) | 201 (73.4) Visceral metastases, n (%): 170 (75.6) | 180 (80.4) | 41 (78.8) | 38 (76.0) | 211 (76.2) | 218 (79.6) Liver metastases, n (%): 121 (53.8) | 124 (55.4) | 28 (53.8) | 26 (52.0) | 149 (53.8) | 150 (54.7) Lung metastases, n (%): 80 (35.6) | 85 (37.9) | 18 (34.6) | 17 (34.0) | 98 (35.4) | 102 (37.2) Brain metastases, n (%): 1 (0.4) | 5 (2.2) | 0 | 3 (6.0) | 1 (0.4) | 8 (2.9) Treatment in the recurrence and metastasis stage, n (%) Chemotherapy: 0 | 0 | 52 (100.0) | 50 (100.0) | 52 (18.8) | 50 (18.2) Endocrine therapy: 188 (83.6) | 188 (83.9) | 42 (80.8) | 36 (72.0) | 230 (83.0) | 224 (81.8) Targeted therapy: 144 (64.0) | 139 (62.1) | 33 (63.5) | 29 (58.0) | 177 (63.9) | 168 (61.3) CDK4/6 inhibitor: 141 (62.7) | 135 (60.3) | 28 (53.8) | 25 (50.0) | 169 (61.0) | 160 (58.4) ECOG PS, Eastern Cooperative Oncology Group performance status, HER2, human epidermal growth factor receptor 2, IHC, immunohistochemistry, ISH, in situ hybridization, HR, hormone receptor, CDK4/6, cyclin-dependent kinase 4 and 6 Presenter: Ying Fan MD

[Slide 2] Primary endpoint: PFS in chemo-naïve patients (per BICR) | | Trastuzumab rezetecan | Physician's choice | Median PFS (95% CI), months | 16.6 (14.0-19.4) | 6.8 (5.7-7.0) Harzard ratio (95% CI) | 0.25 (0.19-0.33) P<0.001 Legend: Trastuzumab rezetecan / Physician's choice / + Censor Y-axis: Progression-free survival (%); X-axis: Time (months) 0-20 No. at Risk (months 0-20) Trastuzumab rezetecan: 225 216 207 197 193 182 170 155 147 130 114 103 84 83 67 62 59 34 33 32 18 Physician's choice: 224 214 167 145 138 113 88 67 59 42 31 27 19 18 14 12 11 7 7 7 4 Presenter: Ying Fan MD PFS, progression-free survival, BICR, blinded independent central review Survival table Endpoint: PFS per BICR, chemo-naïve patients (primary) Arm | N | Median | Median CI Trastuzumab rezetecan | 225 | 16.6 mo | 14.0-19.4 Physician's choice | 224 | 6.8 mo | 5.7-7.0 HR: 0.25 HR CI: 95% CI 0.19-0.33 P: <0.001 [Slide 3] Baseline characteristics HR-positive patients | Trastuzumab rezetecan (N=201) | Physician's choice (N=201) Median age, years (IQR) | 56.0 (48.0–62.0) | 55.0 (47.0–61.0) ≥65 years, n (%) | 44 (21.9) | 28 (13.9) Female, n (%) | 201 (100.0) | 201 (100.0) ECOG PS score of 1, n (%) | 103 (51.2) | 102 (50.7) HER2 status, n (%) IHC 1+ and ISH unknown | 122 (60.7) | 122 (60.7) IHC 2+ and ISH– | 71 (35.3) | 72 (35.8) IHC 1+ and ISH– | 8 (4.0) | 7 (3.5) Visceral metastases, n (%) | 162 (80.6) | 166 (82.6) Liver metastases, n (%) | 117 (58.2) | 123 (61.2) Lung metastases, n (%) | 76 (37.8) | 70 (34.8) Brain metastases, n (%) | 0 | 6 (3.0) Treatment in the recurrence and metastasis stage, n (%) Chemotherapy | 35 (17.4) | 35 (17.4) Endocrine therapy | 201 (100.0) | 200 (99.5) Targeted therapy | 157 (78.1) | 149 (74.1) CDK4/6 inhibitor | 152 (75.6) | 145 (72.1) HR-negative patients | Trastuzumab rezetecan (N=76) | Physician's choice (N=73) Median age, years (IQR) | 54.5 (50.0–60.5) | 56.0 (50.0–66.0) ≥65 years, n (%) | 10 (13.2) | 22 (30.1) Female, n (%) | 76 (100.0) | 73 (100.0) ECOG PS score of 1, n (%) | 37 (48.7) | 40 (54.8) HER2 status, n (%) IHC 1+ and ISH unknown | 48 (63.2) | 46 (63.0) IHC 2+ and ISH– | 25 (32.9) | 23 (31.5) IHC 1+ and ISH– | 3 (3.9) | 4 (5.5) Visceral metastases, n (%) | 49 (64.5) | 52 (71.2) Liver metastases, n (%) | 32 (42.1) | 27 (37.0) Lung metastases, n (%) | 22 (28.9) | 32 (43.8) Brain metastases, n (%) | 1 (1.3) | 2 (2.7) PD-L1 positive, n (%) | 21 (27.6) | 21 (28.8) PD-L1 negative/unknown, n (%) | 55 (72.4) | 52 (71.2) Treatment in the recurrence and metastasis stage, n (%) Chemotherapy | 17 (22.4) | 15 (20.5) Targeted therapy | 20 (26.3) | 19 (26.0) [Abbreviations footnote: illegible] Presenter: Ying Fan MD

ORR and DoR (per BICR) Y-axis: Objective response rate (%) Chemotherapy-naïve patients: Trastuzumab rezetecan (N=225) 76.0 | Physician's choice (N=224) 43.3 Chemotherapy-treated patients: Trastuzumab rezetecan (N=52) 73.1 | Physician's choice (N=50) 32.0 Overall patients: Trastuzumab rezetecan (N=277) 75.5 | Physician's choice (N=274) 41.2 Clinical benefit rate, n (%): 187 (83.1) | 115 (51.3) | 42 (80.8) | 22 (44.0) | 229 (82.7) | 137 (50.0) Median DoR (95% CI), months: 15.3 (12.7-20.6) | 6.9 (5.5-7.2) | 10.4 (7.0-13.8) | 6.1 (3.9-8.1) | 15.0 (12.6-17.6) | 6.8 (5.6-7.2) ORR, objective response rate, DoR, duration of response Presenter: Ying Fan MD

Background HER2-low breast cancer > defined as IHC 1+ or IHC 2+/ ISH-negative1,2 > Historically limited benefit from traditional HER2-targeted therapies and poor prognosis.3 Distribution of breast cancer by HER2 status HER2+ 15% | HER2-low 45%-55% | HER2-negative 30%-40% Trastuzumab rezetecan (SHR-A1811) > HER2-targeted ADC > Cleavable tetrapeptide linker > Highly potent topoisomerase I inhibitor [ADC structure diagram: SHR-A1811 - Cys - Linker - Enzyme cleavable - SHR169265] Current study Phase 3 HORIZON-Breast06 Final analysis for progression-free survival (PFS) in patients with HER2-low breast cancer Presenter: Ying Fan MD 1. J Clin Oncol. Jun 10 2020;38(17):1951-1962. doi:10.1200/jco.19.02488 2. J Clin Oncol. Aug 1 2023;41(22):3867-3872. doi:10.1200/jco.22.02864 3. JAMA. Jan 22 2019;321(3):288-300. doi:10.1001/jama.2018.19323

Efficacy in HR-negative patients [Bar chart: Objective response rate (%)] PD-L1 negative / unknown patients: Trastuzumab rezetecan 70.9; Physician's choice 42.3 PD-L1 positive patients: Trastuzumab rezetecan 95.2; Physician's choice 42.9 | Progression-free survival | PD-L1 neg/unknown: Trastuzumab rezetecan (N=55) | Physician's choice (N=52) | PD-L1 positive: Trastuzumab rezetecan (N=21) | Physician's choice (N=21) | | Events, n (%) | 30 (54.5) | 33 (63.5) | 7 (33.3) | 18 (85.7) | | Median (95% CI), months | 14.0 (9.6-27.7) | 7.1 (2.7-11.1) | 16.6 (11.2-NR) | 6.7 (3.9-8.5) | | Hazard ratio (95% CI) | 0.43 (0.26-0.73) | | 0.25 (0.10-0.59) | | PD-L1 positive is combined positive score ≥1. PD-L1, Programmed Cell Death Ligand 1; HR-negative, hormone receptor-negative; NR, not reached Presenter: Ying Fan MD

Chemo-naïve patients | Overall patients Trastuzumab rezetecan (N=225) | Chemotherapy (N=224) | Trastuzumab rezetecan (N=277) | Chemotherapy (N=274) PFS Events, n (%) 91 (40.4) | 146 (65.2) | 130 (46.9) | 185 (67.5) Median (95% CI), months 16.6 (14.0–19.4) | 6.8 (5.7–7.0) | 16.4 (14.0–16.8) | 6.8 (5.7–6.9) 12-mo rate, % (95% CI) 65.2 (57.6–71.8) | 18.9 (12.8–25.8) | 63.1 (56.4–69.1) | 18.0 (12.7–24.1) HR (95% CI); 1-sided P 0.25 (0.19–0.33); P<0.0001 | 0.27 (0.21–0.34); P<0.0001 OS Events, n (%) 43 (19.1) | 55 (24.6) | 66 (23.8) | 77 (28.1) 12-mo rate, % (95% CI) 92.0 (87.3–95.1) | 83.0 (76.9–87.6) | 90.6 (86.2–93.6) | 81.4 (75.9–85.8) ORR, % (95% CI) 76.0 (69.9–81.4) | 43.3 (36.7–50.1) | 75.5 (69.9–80.4) | 41.2 (35.4–47.3)

JAMA | Research Summary Trastuzumab Rezetecan in ERBB2-Low Advanced Breast Cancer This trial compared trastuzumab rezetecan vs physician's choice of chemotherapy in patients with erythroblastic oncogene B (ERBB2)–low metastatic breast cancer. Why Does This Clinical Trial Matter? Patients with ERBB2-low breast cancer have limited targeted treatment options, especially those with hormone receptor–negative disease. This trial tested whether trastuzumab rezetecan could improve outcomes compared with standard chemotherapy in patients who had received ≤1 prior line of chemotherapy treatment for metastatic disease. Trastuzumab rezetecan 277 patients Physician's choice chemotherapy 274 patients Who Was Involved? • Participants Analyzed: 551 Females were randomized and included in the primary analysis • Type: Females with unresectable locally recurrent or metastatic breast cancer • Age (mean): 55 Years • Locations: Multisite trial at 73 hospitals in China • Eligibility Criteria: Centrally confirmed ERBB2-low disease, age 18-75 years, ≤1 prior chemotherapy line for metastatic disease, and prior endocrine therapy required for hormone receptor–positive disease How Was the Trial Done? • Trial Type: Open-label, randomized, multicenter phase 3 clinical trial • Randomization Details: Individual participants were randomized 1:1 • Interventions: Trastuzumab rezetecan vs physician's choice of chemotherapy • Intervention Details: Trastuzumab rezetecan 6.4 mg/kg every 3 weeks; chemotherapy options were taxanes, eribulin, capecitabine, or gemcitabine • Mode of Delivery: Intravenous trastuzumab rezetecan; chemotherapy per local label or guidelines • Timing of Outcome Assessment: Tumors were assessed every 6 weeks for 48 weeks, then every 12 weeks What Was Found? Primary Outcomes • Progression-free survival was longer with trastuzumab rezetecan than chemotherapy in chemotherapy-naive patients (median, 16.6 [95% CI, 14.0-19.4] vs 6.8 [95% CI, 5.7-7.0] months; hazard ratio, 0.25 [95% CI, 0.19-0.33]; P < .001) and in all randomized patients (median, 16.4 [95% CI, 14.0-16.8] vs 6.8 [95% CI, 5.7-6.9] months; hazard ratio, 0.27 [95% CI, 0.21-0.34]; P < .001) Secondary Outcomes • Benefit was also seen in hormone receptor–negative disease (16.5 vs 6.8 months) and hormone receptor–positive disease (16.4 vs 6.7 months) Adverse Events • Grade ≥3 treatment-related adverse events occurred in 232 of 276 patients (84.1%) with trastuzumab rezetecan vs 184 of 271 (67.9%) with chemotherapy; decreased neutrophil count was the most common Progression-free survival for all patients Trastuzumab rezetecan 16.4 mo (95% CI, 14.0 to 16.8) Physician's choice chemotherapy 6.8 mo (95% CI, 5.7 to 6.9) What This Means This study suggests trastuzumab rezetecan may be more effective than standard chemotherapy for ERBB2-low advanced breast cancer, including in hormone receptor–positive and hormone receptor–negative disease. However, it caused more severe treatment-related adverse events, so close monitoring and dose adjustment may be needed in practice. Limitations and Knowledge Gaps Overall survival data were immature. The trial included only patients in China, which may limit generalizability. The chemotherapy control group was heterogeneous, and treatment standards are evolving. Further studies should confirm longer-term benefits and compare this drug with other current options. Citation: Fan Y, Liu Y, Wang Y, et al. Trastuzumab rezetecan in ERBB2-low advanced breast cancer: the HORIZON-Breast06 randomized clinical trial. JAMA. Published online October 23, 2026. doi:10.1001/jama.2026.21372 Clinical Trial Registration: NCT05814354 Funding: Jiangsu Hengrui Pharmaceuticals, Chinese Academy of Medical Sciences Innovation Fund for Medical Sciences, Noncommunicable Chronic Diseases-National Science and Technology Major Project, the National Natural Science Foundation of China, Beijing Natural Science Foundation Additional Information: GPT 4.1 (OpenAI) generated a draft of the manuscript text on October 6, 2026, following an editor-developed prompt. The final manuscript was prepared and edited by editorial staff, who take full responsibility for the content. © 2026 American Medical Association. All rights reserved, including those for text and data mining, AI training, and similar technologies.

13:30 - 15:00 Proffered paper 1: Breast cancer, early and advanced stage CHAIRS: KOMAL JHAVERI, BARBARA PISTILLI HORIZON-Breast06 study design Randomized, open-label, multicenter, phase 3 trial (NCT05814354) Eligibility criteria • Centrally confirmed HER2-low (IHC 1+ or IHC 2+/ISH–) metastatic breast cancer • ≤1 prior line of chemotherapy for metastatic disease • Prior endocrine therapy required for HR-positive disease. R 1:1 Trastuzumab rezetecan (6.4 mg/kg, IV, Q3W) Physician's choice of chemotherapy (taxanes, eribulin, capecitabine, or gemcitabine) Primary endpoint: PFS (per BICR) • in chemo-naive patients • in overall patients Key secondary endpoints: • OS in overall patients • OS in chemo-naive patients • PFS (per BICR) in chemo-treated patients Stratification factors: ✓ Prior lines of chemotherapy for metastatic disease (0 vs. 1) ✓ HER2 status (IHC 1+ vs. IHC 2+/ISH–) ✓ HR status (HR+ with prior CDK4/6 inhibitor vs. HR+ without prior CDK4/6 inhibitor vs. HR–) Presenter: Ying Fan MD Ying Fan Trastuzumab rezetecan (SHR-A1811) in HER2-low advanced breast cancer (BC): A randomized, phase 3 study (HORIZON-Breast06) ALICANTE AUDITORIUM - HALL 6 Experimental ADC could be option for HER2-low Metastatic Breast Cancer ESMO / BREASTCANCER•ORG

[Slide 1] PANKU-Breast01 Study Design An open-label, multicenter, randomized, phase III study conducted at 72 study centers across China Key eligibility criteria • Unresectable LA/M HR+/HER2- BC (HR+, ER≥1% and/or PR≥1%; HER2-, IHC 0, IHC 1+, or IHC 2+/ISH-) • Progressed after at least 1 line of endocrine therapy • 1 or 2 lines of prior chemotherapy • Previously received CDK4/6 inhibitors and taxanes • Measurable lesion per RECIST v1.1 • ECOG PS 0-1 R 1:1 N=383 Iza-bren 2.5 mg/kg D1D8 Q3W TPC • Eribulin, D1D8 Q3W • Capecitabine, BID, Q3W • Gemcitabine, D1D8 Q3W • Vinorelbine, D1D8 Q3W Stratified by • Previous lines of chemotherapy (1 line vs 2 lines) • CNS metastases (Yes vs No) • HER2 expression (IHC 0 vs IHC 1+ vs IHC 2+/ISH-) Treatment until Disease progression per RECIST v1.1 or intolerable toxicity Primary endpoint: PFS (BICR) Key secondary endpoint: OS H_PFS α=0.025 → H_OS α=0 Other secondary endpoints • PFS (INV) • ORR (BICR & INV) • DCR (BICR & INV) • DOR (BICR & INV) • Safety, PK, Immunogenicity Note: Iza-bren was dosed with 10% compensation per protocol; eribulin 1.4 mg/m2 on days 1 and 8 Q3W; capecitabine 1000-1250 mg/m2, BID from days 1 to 14 Q3W; gemcitabine 1000 mg/m2 on days 1 and 8 Q3W; vinorelbine 25 mg/m2 on days 1 and 8 Q3W. ECOG PS, Eastern Cooperative Oncology Group performance status; Q3W, every 3 weeks; D, day; PFS, progression-free survival; BICR, blinded independent central review; OS, overall survival; INV, investigator assessed; ORR, objective response rate (confirmed); DCR, disease control rate; DOR, duration of response. Bo Lan, MD [Slide 2] PFS by BICR: Primary Endpoint [Table] Iza-bren (N = 190) | TPC (N = 193) Number of events: 130 | 159 Median, mo (95% CI): 8.2 (6.9, 8.6) | 4.0 (2.9, 4.3) HR (95% CI): 0.41 (0.33, 0.53)† P-value: <0.0001* [Kaplan-Meier plot: Progression-free survival probability (%) vs Time (Months), 0-24] Curve labels: Iza-bren 62.8% (6 mo), 41.8% (9 mo), 26.5% (12 mo); TPC 24.2% (6 mo), 14.4% (9 mo), 9.2% (12 mo) Legend: Iza-bren; TPC; + Censored Subjects at risk (months 0/3/6/9/12/15/18/21/24): Iza-bren: 190, 138, 97, 63, 38, 22, 12, 4, 0 TPC: 193, 96, 34, 17, 10, 4, 2, 1, 0 Iza-bren demonstrated a statistically significant and clinically meaningful improvement in PFS vs TPC with doubling of mPFS and a HR of 0.41 PFS (BICR) was evaluated in the mITT population. The median duration of follow-up for PFS was 15.2 months. ITT population is identical to the mITT population. †Based on a stratified Cox regression model. *One-sided, based on a stratified log-rank test. Stratification factor CNS metastases (yes vs no) was not included in the stratified analysis due to the small number of patients with CNS metastases in both arms (n=4 or 3). Data cut off: July 21, 2026 Bo Lan, MD Survival table Endpoint: PFS by BICR (primary endpoint), mITT Arm | N | Median | Median CI Iza-bren | 190 | 8.2 mo | 6.9, 8.6 TPC | 193 | 4.0 mo | 2.9, 4.3 HR: 0.41 HR CI: 0.33, 0.53 P: <0.0001 (one-sided, stratified log-rank) Landmark rates: 6 mo: Iza-bren 62.8% | TPC 24.2% 9 mo: Iza-bren 41.8% | TPC 14.4% 12 mo: Iza-bren 26.5% | TPC 9.2% [Slide 3] TRAEs in ≥ 30% of Patients TRAEs with Frequency ≥ 30% Iza-bren (N=190) vs TPC (N=193) [Butterfly bar chart, Incidence (%), Preferred Term] Preferred terms (top to bottom): Anaemia; Leukopenia; Thrombocytopenia; Neutropenia; Alopecia; Asthenia; Nausea; AST increased; Decreased appetite; Constipation; Hypoalbuminaemia; ALT increased; Lymphocyte count decreased; Diarrhoea; Hypokalaemia; Stomatitis; Hypertriglyceridaemia [Bar value labels, read top-to-bottom in photo order; small, perspective-skewed, row alignment uncertain] Iza-bren (Grade ≥3 | Grade 1-2): 49.5 | 44.8; 48.4 | 38.4; 46.3 | 38.4; 53.7 | 25.8; 0.5 | 52.7; 13.2 | 36.3; 2.6 | 43.7; 0.5 | 37.9; 2.1 | 34.7; — | 35.8; — | 35.8; 1.1 | 34.2; 13.7 | 17.9; 3.2 | 27.9; 7.4 | 23.7; 2.1 | 28.4; 1.6 | 28.4 TPC (Grade 1-2 | Grade ≥3): 56.5 | 6.2; 44.6 | 40.4; 31.1 | 5.7; 32.1 | 54.9; 43.0; 19.7 | 2.6; 24.4 | 0.5; 34.7 | 2.1; 13.5; 21.8; 11.4; 36.2 | 1.6; 10.9 | 5.7; 7.3; 6.2; 5.7 | 0.5; 19.6 | 3.6 Legend: Iza-bren | Grade 1-2; Iza-bren | Grade ≥3; TPC | Grade 1-2; TPC | Grade ≥3 Most common TRAEs were hematological, including anemia, thrombocytopenia and neutropenia, which were effectively managed through standard supportive measures including growth factor support, dose delay and dose reductions, which allowed majority of patients to continue iza-bren treatment with low discontinuation rate. Primary G-CSF prophylaxis is not mandatory. Grade ≥3 anemia resolved in a median of 6 days in the iza-bren arm. Grade ≥3 neutropenia and thrombocytopenia generally resolved within 2 and 3 days in the iza-bren arm. Neutropenic fever (7.9%) and grade 3 or higher infections (4.7%) were low in the iza-bren arm. Non-hematologic AEs were mostly low grade; alopecia was the most common in both groups (53.2% with iza-bren and 43.0% with TPC), followed by asthenia (49.5% and 22.3%, respectively). Adjudicated treatment-related ILD occurred in two patients (1.1%) in the iza-bren arm (one case of grade 1 and one case of grade 2). None in the TPC arm. Data cut off: July 21, 2026 Bo Lan, MD [Slide 4] Safety Summary [Table] Iza-bren (N = 190) | TPC (N = 193) TRAEs, n (%): 189 (99.5) | 188 (97.4) ≥Grade 3 TRAEs, n (%): 162 (85.3) | 127 (65.8) Treatment-related SAEs, n (%): 79 (41.6) | 22 (11.4) TRAEs leading to death, n (%): 2 (1.1) | 1 (0.5) TRAEs leading to treatment discontinuation, n (%): 3 (1.6) | 2 (1.0) TRAEs leading to dose reduction, n (%): 123 (64.7) | 55 (28.5) TRAEs leading to dose interruption, n (%): 107 (56.3) | 73 (37.8) • Median (range) duration of treatment, mo: – Iza-bren: 7.3 (0.3, 22.5) – TPC: 4.3 (0.3, 26.0) • Median (range) number of treatment cycles: – Iza-bren: 10 (1, 30) – TPC: 6 (1, 37) • Iza-bren had a manageable safety profile. The rate of treatment discontinuation due to TRAEs was low (1.6%) vs TPC (1.0%). • TRAEs leading to death occurred in 2 patients (1.1%) receiving iza-bren and in 1 patient (0.5%) receiving TPC. TRAE, treatment related adverse events. Safety data based on the safety analysis set. TRAEs leading to death occurred in 2 patients (1.1%) receiving iza-bren: Death (unknown cause) and Pneumonia. TRAEs leading to death occurred in 1 patient (0.5%) receiving TPC: Sepsis. Data cut off: July 21, 2026 Bo Lan, MD

Antitumor Responses Iza-bren (N=190) | TPC (N=193) Best overall response, n (%) PR: 77 (40.5) | 39 (20.2) Non-CR/Non-PD: 3 (1.6) | 3 (1.6) SD: 84 (44.2) | 81 (42.0) PD: 16 (8.4) | 61 (31.6) NE: 10 (5.3) | 9 (4.7) ORR, % (95% CI): 40.5 (33.5, 47.9) | 20.2 (14.8, 26.6) Difference, % (95% CI): 20.2 (11.2, 29.0) Odds ratio (95% CI): 2.7 (1.7, 4.3) DCR, % (95% CI): 86.3 (80.6, 90.9) | 63.7 (56.5, 70.5) CR, complete response; PR, partial response; SD, stable disease; PD, progressive disease; NE, not evaluable [Waterfall plot] Iza-bren mDOR = 6.9 mo — 77.9% of patients (148/190) had tumor shrinkage and the median (range) shrinkage (%) was -37.4 (-89.9, -0.1). [Waterfall plot] TPC mDOR = 4.2 mo — 60.6% of patients (117/193) had tumor shrinkage and the median (range) shrinkage (%) was -24.5 (-90.7, -0.1). Y-axis: Best Change from Baseline (%) • OS data not mature*: Median follow-up 17.8 months, and no detrimental effect was observed • The study is continuing to the next planned analysis for OS *: Information fraction: 62.8% BICR-assessed confirmed ORR was doubled in iza-bren versus TPC with longer durability of response. Data cut off: July 21, 2026 Bo Lan, MD

[search icon] Proffered Paper session LBA50 - PANKU-Breast01: Randomized phase III study of izalontamab brengitecan (iza-bren) in previously treated unresectable locally advanced or metastatic (LA/M) hormone receptor-positive/HER2-negative (HR+/HER2-) breast cancer (BC) Speakers Bo Lan (Beijing, China) Lecture Time 14:35 - 14:45 Abstract Background Although first-line endocrine therapy plus CDK4/6 inhibitors has improved clinical outcome of HR+/HER2- BC, patients (pts) inevitably develop resistance and survival outcomes remain poor with subsequent therapy. Iza-bren is a first-in-class bispecific antibody-drug conjugate (ADC) against EGFR and HER3, containing a potent topoisomerase I inhibitor payload (Ed-04). Methods Eligible pts had unresectable LA/M HR+/HER2- BC (HR+, ER≥1% and/or PR≥1%; HER2-, IHC 0, 1+, or 2+/ISH-), progressed after at least 1L of endocrine therapy and 1 or 2L prior chemotherapy, and had previously received CDK4/6 inhibitors and taxanes. Pts were randomized 1:1 to receive iza-bren (2.5 mg/kg D1D8 Q3W) or TPC (eribulin, capecitabine, gemcitabine, or vinorelbine). The primary endpoint was PFS by blinded independent central review (BICR). Secondary endpoint included OS. The pre-planned final analysis for PFS serves as the interim analysis for OS. Results As of Jul 21, 2026, a total of 383 pts were randomized and treated (iza-bren/TPC, 190/193). Median follow-up for PFS was 15.24 vs 13.63 mo (iza-bren/TPC). Iza-bren significantly improved PFS by BICR vs TPC; mPFS was 8.18 mo (95% CI, 6.93-8.58) vs 4.01 mo (95% CI, 2.92-4.30) (HR, 0.41; 95% CI, 0.33-0.53; P<0.0001). OS was immature at the interim analysis, and no detrimental effect was observed. The cORR by BICR was 40.5% with iza-bren vs 20.2% with TPC, and mDOR was 6.93 vs 4.21 mo, respectively. Grade ≥3 TRAEs were predominantly hematologic in nature. TRAEs were manageable with standard supportive care, and only 1.6% of pts discontinued iza-bren treatment due to TRAEs. Adjudicated ILD occurred in 2 (1.1%) pts with iza-bren (grade 1 or 2). No new safety signals were identified. Conclusions The study met its primary endpoint at the prespecified analysis. Iza-bren demonstrated a statistically significant and clinically meaningful improvement in PFS vs TPC in previously treated HR+/HER2- BC. The safety profile was manageable and consistent with previous studies. The results support iza-bren as a new standard of care option in this population in China.

[Slide 1] Proffered Paper session LBA50 - PANKU-Breast01: Randomized phase III study of izalontamab brengitecan (iza-bren) in previously treated unresectable locally advanced or metastatic (LA/M) hormone receptor-positive/HER2-negative (HR+/HER2-) breast cancer (BC) Speakers Bo Lan (Beijing, China) Lecture Time 14:35 - 14:45 [Abstract] Background Although first-line endocrine therapy plus CDK4/6 inhibitors has improved clinical outcome of HR+/HER2- BC, patients (pts) inevitably develop resistance and survival outcomes remain poor with subsequent therapy. Iza-bren is a first-in-class bispecific antibody-drug conjugate (ADC) against EGFR and HER3, containing a potent topoisomerase I inhibitor payload (Ed-04). [Slide 2] Methods Eligible pts had unresectable LA/M HR+/HER2- BC (HR+, ER≥1% and/or PR≥1%; HER2-, IHC 0, 1+, or 2+/ISH-), progressed after at least 1L of endocrine therapy and 1 or 2L prior chemotherapy, and had previously received CDK4/6 inhibitors and taxanes. Pts were randomized 1:1 to receive iza-bren (2.5 mg/kg D1D8 Q3W) or TPC (eribulin, capecitabine, gemcitabine, or vinorelbine). The primary endpoint was PFS by blinded independent central review (BICR). Secondary endpoint included OS. The pre-planned final analysis for PFS serves as the interim analysis for OS. [Slide 3] Results As of Jul 21, 2026, a total of 383 pts were randomized and treated (iza-bren/TPC, 190/193). Median follow-up for PFS was 15.24 vs 13.63 mo (iza-bren/TPC). Iza-bren significantly improved PFS by BICR vs TPC; mPFS was 8.18 mo (95% CI, 6.93-8.58) vs 4.01 mo (95% CI, 2.92-4.30) (HR, 0.41; 95% CI, 0.33-0.53; P<0.0001). OS was immature at the interim analysis, and no detrimental effect was observed. The cORR by BICR was 40.5% with iza-bren vs 20.2% with TPC, and mDOR was 6.93 vs 4.21 mo, respectively. Grade ≥3 TRAEs were predominantly hematologic in nature. TRAEs were manageable with standard supportive care, and only 1.6% of pts discontinued iza-bren treatment due to TRAEs. Adjudicated ILD occurred in 2 (1.1%) pts with iza-bren (grade 1 or 2). No new safety signals were identified. Conclusions The study met its primary endpoint at the prespecified analysis. Iza-bren demonstrated a statistically significant and clinically meaningful improvement in PFS vs TPC in previously treated HR+/HER2- BC. The safety profile was manageable and consistent with previous studies. The results support iza-bren as a new standard of care option in this population in China.

Conclusions • In this pre-specified analysis, iza-bren demonstrated a statistically significant and clinically meaningful improvement in PFS vs TPC in previously treated HR+/HER2- BC, meeting the primary endpoint. • Iza-bren demonstrated an improved confirmed ORR and more durable response than TPC. • OS is immature at this interim analysis, no detrimental effect was observed. • The safety profile of iza-bren was manageable and consistent with previous studies. – TRAEs leading to discontinuation (1.6%) were low. – The incidence of adjudicated ILD was 1.1% PFS (Median): Iza-bren 8.2 mo; TPC 4.0 mo; HR: 0.41, 95% CI: 0.33 - 0.53 cORR: Iza-bren 40.5%; TPC 20.2%; Difference: 20.2%, 95% CI: 11.2 - 29.0 DOR (Median): Iza-bren 6.9 mo; TPC 4.2 mo The results support iza-bren as a potential new standard of care option in this population. Bo Lan, MD

[Slide 2] 13:30 - 15:00 Proffered paper 1: Breast cancer, early and advanced stage CHAIRS: KOMAL JHAVERI, BARBARA PISTILLI PFS by BICR: Primary Endpoint | Iza-bren (N = 190) | TPC (N = 193) Number of events | 130 | 159 Median, mo (95% CI) | 8.2 (6.9, 8.6) | 4.0 (2.9, 4.3) HR (95% CI) | 0.41 (0.33, 0.53)† P-value | <0.0001* Curve labels: Iza-bren 62.8% (6 mo), 41.8% (9 mo), 26.5% (12 mo); TPC 24.2% (6 mo), 14.4% (9 mo), 9.2% (12 mo) Legend: Iza-bren; TPC; + Censored Y-axis: Progression-free survival probability (%) X-axis: Time (Months) Subjects at risk (months 0 / 3 / 6 / 9 / 12 / 15 / 18 / 21 / 24): Iza-bren: 190 / 138 / 97 / 63 / 38 / 22 / 12 / 4 / 0 TPC: 193 / 96 / [illegible]4 / 17 / 10 / 4 / 2 / 1 / 0 Iza-bren demonstrated a statistically significant and clinically meaningful improvement in PFS vs TPC with doubling of mPFS and a HR of 0.41 PFS (BICR) was evaluated in the mITT population. The median duration of follow-up for PFS was 15.2 months. ITT population is identical to the mITT population. †Based on a stratified Cox regression model. *One-sided, based on a stratified log-rank test. Stratification factor CNS metastases (yes vs no) was not included in the stratified analysis due to the small number of patients with CNS metastases in both arms (n=4 or 3). Data cut off: July 21, 2026 Bo Lan, MD [Speaker panel] Bo Lan — PANKU-Breast01: Randomized phase III study of izalontamab brengitecan (iza-bren) in previously treated unresectable locally advanced or metastatic (LA/M) hormone receptor-positive/HER2-negative (HR+/HER2-) breast cancer (BC) ALICANTE AUDITORIUM - HALL 6 Survival table Endpoint: PFS by BICR (primary endpoint), mITT Arm | N | Median | Median CI Iza-bren | 190 | 8.2 mo | 6.9, 8.6 TPC | 193 | 4.0 mo | 2.9, 4.3 HR: 0.41 HR CI: 0.33, 0.53 P: <0.0001 (one-sided, stratified log-rank) Landmark rates: 6 mo: Iza-bren 62.8% | TPC 24.2% 9 mo: Iza-bren 41.8% | TPC 14.4% 12 mo: Iza-bren 26.5% | TPC 9.2% [Slide 3] 13:30 - 15:00 Proffered paper 1: Breast cancer, early and advanced stage CHAIRS: KOMAL JHAVERI, BARBARA PISTILLI Safety Summary | Iza-bren (N = 190) | TPC (N = 193) TRAEs, n (%) | 189 (99.5) | 188 (97.4) ≥Grade 3 TRAEs, n (%) | 162 (85.3) | 127 (65.8) Treatment-related SAEs, n (%) | 79 (41.6) | 22 (11.4) TRAEs leading to death, n (%) | 2 (1.1) | 1 (0.5) TRAEs leading to treatment discontinuation, n (%) | 3 (1.6) | 2 (1.0) TRAEs leading to dose reduction, n (%) | 123 (64.7) | 55 (28.5) TRAEs leading to dose interruption, n (%) | 107 (56.3) | 73 (37.8) • Median (range) duration of treatment, mo: – Iza-bren: 7.3 (0.3, 22.5) – TPC: 4.3 (0.3, 26.0) • Median (range) number of treatment cycles: – Iza-bren: 10 (1, 30) – TPC: 6 (1, 37) • Iza-bren had a manageable safety profile. The rate of treatment discontinuation due to TRAEs was low (1.6%) vs TPC (1.0%). • TRAEs leading to death occurred in 2 patients (1.1%) receiving iza-bren and in 1 patient (0.5%) receiving TPC. TRAE, treatment related adverse events. Safety data based on the safety analysis set. TRAEs leading to death occurred in 2 patients (1.1%) receiving iza-bren: Death (unknown cause) and Pneumonia. TRAEs leading to death occurred in 1 patient (0.5%) receiving TPC: Sepsis. Data cut off: July 21, 2026 Bo Lan, MD [Speaker panel] Bo Lan — PANKU-Breast01: Randomized phase III study of izalontamab brengitecan (iza-bren) in previously treated unresectable locally advanced or metastatic (LA/M) hormone receptor-positive/HER2-negative (HR+/HER2-) breast cancer (BC) ALICANTE AUDITORIUM - HALL 6 [Slide 4] 13:30 - 15:00 Proffered paper 1: Breast cancer, early and advanced stage CHAIRS: KOMAL JHAVERI, BARBARA PISTILLI Conclusions • In this pre-specified analysis, iza-bren demonstrated a statistically significant and clinically meaningful improvement in PFS vs TPC in previously treated HR+/HER2- BC, meeting the primary endpoint. • Iza-bren demonstrated an improved confirmed ORR and more durable response than TPC. • OS is immature at this interim analysis, no detrimental effect was observed. • The safety profile of iza-bren was manageable and consistent with previous studies. – TRAEs leading to discontinuation (1.6%) were low. – The incidence of adjudicated ILD was 1.1% PFS (Median): Iza-bren 8.2 mo; TPC 4.0 mo HR: 0.41, 95% CI: 0.33 - 0.53 cORR: Iza-bren 40.5%; TPC 20.2% Difference: 20.2%, 95% CI: 11.2 - 29.0 DOR (Median): Iza-bren 6.9 mo; TPC 4.2 mo The results support iza-bren as a potential new standard of care option in this population. Bo Lan, MD [Speaker panel] Bo Lan — PANKU-Breast01: Randomized phase III study of izalontamab brengitecan (iza-bren) in previously treated unresectable locally advanced or metastatic (LA/M) hormone receptor-positive/HER2-negative (HR+/HER2-) breast cancer (BC) ALICANTE AUDITORIUM - HALL 6

Baseline Characteristics | | Iza-bren (N = 190) | TPC (N = 193) | Median (range) age, years | 54.0 (33.0, 71.0) | 54.0 (27.0, 70.0) Age group (years), n (%) < 65 | 166 (87.4) | 172 (89.1) ≥ 65 | 24 (12.6) | 21 (10.9) Previous chemotherapy lines, n (%) 1 line | 135 (71.1) | 139 (72.0) 2 lines | 55 (28.9) | 54 (28.0) CNS metastases, n (%) Yes | 4 (2.1) | 3 (1.6) No | 186 (97.9) | 190 (98.4) HER2 expression, n (%) IHC 0 | 64 (33.7) | 66 (34.2) IHC 1+ | 73 (38.4) | 74 (38.3) IHC 2+/ISH- | 53 (27.9) | 53 (27.5) | | Iza-bren (N = 190) | TPC (N = 193) | Duration of prior CDK-inhibitor use ≥ 12 months, n (%) | 93 (48.9) | 103 (53.4) Prior (neo)adjuvant therapy, n (%) | 152 (80.0) | 149 (77.2) Prior anthracycline use, n (%) | 152 (80.0) | 160 (82.9) Duration of endocrine therapy in advanced disease ≥ 6 months, n (%) | 162 (85.3) | 164 (85.0) Progression within 6 months after (neo)adjuvant chemotherapy, n (%) | 16 (8.4) | 18 (9.3) Baseline liver metastasis, n (%) | 136 (71.6) | 137 (71.0) ECOG PS, n (%) 0 | 68 (35.8) | 69 (35.8) 1 | 122 (64.2) | 124 (64.2) De novo metastasis at initial diagnosis, n (%) | 33 (17.4) | 42 (21.8) Previous radiotherapy, n (%) | 106 (55.8) | 103 (53.4) Data cut off: July 21, 2026 Bo Lan, MD

PANKU-Breast01: Randomized Phase III Study of Izalontamab Brengitecan (iza-bren) in Previously Treated Unresectable Locally Advanced or Metastatic (LA/M) Hormone Receptor-Positive/HER2-negative (HR+/HER2-) Breast Cancer (BC) Bo Lan1, Yehui. Shi2, Xiying Shao3, Fan Wu4, Shusen Wang5, Weimin Xie6, Qingyuan Zhang7, Quchang Ouyang8, Jiuwei Cui9, Xiaoling Ling10, Haijun Yu11, Huihui Li12, Xianjun Tang13, Sijuan Ding14, Chunfang Hao15, Bing Zhao16, Sa Xiao17, Hai Zhu18, Yi Zhu19, Binghe Xu1* On behalf of the PANKU-Breast01 investigators *Leading PI 1Department of Medical Oncology, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China; 2Department of Breast Medical Oncology, Tianjin Medical University Cancer Institute and Hospital, Tianjin, China; 3Department of Medical Oncology for Breast Diseases, Zhejiang Cancer Hospital, Hangzhou, China; 4Department of Breast Medical Oncology, Fujian Cancer Hospital, Fuzhou, China; 5Department of Medical Oncology, Sun Yat-sen University Cancer Center, Guangzhou, China; 6Department of Breast, Bone & Soft Tissue Oncology, Guangxi Medical University Cancer Hospital, Nanning, China; 7Department of Breast Medical Oncology, Harbin Medical University Cancer Hospital, Harbin, China; 8Department of Breast Medical Oncology, Hunan Cancer Hospital, Changsha, China; 9Department of Cancer Center, The First Hospital of Jilin University, Changchun, China; 10Department of Medical Oncology, The First Hospital of Lanzhou University, Lanzhou, China; 11Department of Tumor chemoradiotherapy, Zhongnan Hospital of Wuhan University, Wuhan, China; 12Department of Breast Medical Oncology, Cancer Hospital of Shandong First Medical University, Jinan, China; 13Department of Breast Oncology Center, Chongqing University Cancer Hospital, Chongqing, China; 14Department of Oncology, The Central Hospital of Yongzhou, Yongzhou, China; 15Department of Breast Tumor, Tianjin Cancer Hospital Airport Hospital, Tianjin, China; 16Department of Breast Medical Oncology, Affiliated Cancer Hospital of Xinjiang Medical University, Urumqi, China; 17Medical and Pharmacological Research Department, Baili-Bio (Chengdu) Pharmaceutical Co., Ltd., Chengdu, China; 18Department of Biometrics, SystImmune Inc., Redmond, United States of America; 19President and CSO, Sichuan Biokin Pharmaceutical Co., Ltd, Chengdu, China 23 October 2026

EUROPEAN SOCIETY FOR MEDICAL ONCOLOGY (ESMO) CONGRESS 2026 Articles Izalontamab brengitecan for HR+/HER2– breast cancer in China (PANKU-Breast01): a randomised, phase 3 trial Bo Lan*, Yehui Shi*, Xiying Shao*, Fan Wu*, Shusen Wang*, Weimin Xie, Qingyuan Zhang, Quchang Ouyang, Jiuwei Cui, Xiaoling Ling, Haijun Yu, Huihui Li, Xianjun Tang, Sijuan Ding, Chunfang Hao, Bing Zhao, Yang Li, Hua Yang, Min Yan, Guohui Han, Xujuan Wang, Fanfan Li, Peijian Peng, Chang Gong, Yunjiang Liu, Junlan Guo, Yongmei Yin, Man Li, Yue'e Teng, Jiuda Zhao, Yongli Gao, Jingfen Wang, Jing Luo, Xinghua Han, Xiaoping Li, Bin Liu, Shuqun Zhang, Bihui Li, Lei Yang, Sa Xiao, Hai Zhu, Yi Zhu, Binghe Xu, for the PANKU-Breast01 Investigators† Summary Background Patients with heavily pretreated hormone receptor-positive and human epidermal growth factor receptor 2-negative (HR+/HER2–) breast cancer have few treatment options. Izalontamab brengitecan (iza-bren) is a first-in-class, epidermal growth factor receptor and HER3 bispecific antibody–drug conjugate that contains a potent topoisomerase I inhibitor payload (Ed-04). The aim of PANKU-Breast01 was to compare the efficacy and safety of iza-bren versus chemotherapy in patients with pretreated HR+/HER2– breast cancer. Published Online October 23, 2026 https://doi.org/10.1016/S0140-6736(26)02096-9 *Co-first authors †Investigators are listed in the appendix (pp 2–5) [The Lancet crossmark/access icons]

PANKU-Breast01 Study Design An open-label, multicenter, randomized, phase III study conducted at 72 study centers across China Key eligibility criteria • Unresectable LA/M HR+/HER2- BC (HR+, ER≥1% and/or PR≥1%; HER2-, IHC 0, IHC 1+, or IHC 2+/ISH-) • Progressed after at least 1 line of endocrine therapy • 1 or 2 lines of prior chemotherapy • Previously received CDK4/6 inhibitors and taxanes • Measurable lesion per RECIST v1.1 • ECOG PS 0-1 R 1:1 N=383 Iza-bren 2.5 mg/kg D1D8 Q3W TPC: Eribulin, D1D8 Q3W; Capecitabine, BID, Q3W; Gemcitabine, D1D8 Q3W; Vinorelbine, D1D8 Q3W Stratified by • Previous lines of chemotherapy (1 line vs 2 lines) • CNS metastases (Yes vs No) • HER2 expression (IHC 0 vs IHC 1+ vs IHC 2+/ISH-) Treatment until Disease progression per RECIST v1.1 or intolerable toxicity Primary endpoint: PFS (BICR) Key secondary endpoint: OS HPFS α=0.025 | HOS α=0 Other secondary endpoints • PFS (INV) • ORR (BICR & INV) • DCR (BICR & INV) • DOR (BICR & INV) • Safety, PK, Immunogenicity Bo Lan, MD Novel ADC slows growth of metastatic HR-positive breast cancer ESMO / BREASTCANCER.ORG

[Slide 1] VIKTORIA-1 Study Design HR+/HER2- Advanced Breast Cancer Eligibility Criteria • Pre- & postmenopausal women & men • Progression on/after CDK4/6i + NSAI • ≤2 lines of prior ET for ABC • Measurable disease, RECIST v1.1 • Screening result for PIK3CA status • No T2DM with HbA1c >6.4% or T1DM • No prior mTORi, PI3Ki, or AKTi • No prior chemotherapy for ABC Stratification Factors • Lung/liver metastases (yes/no) • TTP on immediate prior therapy (≤ or >6 months) • Region (US/Canada or ROW) STUDY 2 PIK3CA-MT — R 3:3:1 (N=362) Arm D^a: Gedatolisib | 180 mg IV once weekly, 3 weeks on, 1 week off Palbociclib | 125 mg daily, 21 days on, 7 days off Fulvestrant | 500 mg; cycle 1, days 1 & 15, then every 4 weeks Arm E: Alpelisib | 300 mg orally daily for 4 weeks Fulvestrant | 500 mg; cycle 1, days 1 & 15, then every 4 weeks Arm F^a: Gedatolisib | 180 mg IV once weekly, 3 weeks on, 1 week off Fulvestrant | 500 mg; cycle 1, days 1 & 15, then every 4 weeks Primary Endpoint: PFS (BICR) Arm D vs Arm E Secondary Endpoints: OS; PFS (BICR F vs E); Response; Safety; QoL Assign STUDY 1 PIK3CA-WT — R 1:1:1 (N=392) Arm A^a: Gedatolisib + Palbociclib + Fulvestrant Arm B^a: Gedatolisib^a + Fulvestrant Arm C: Fulvestrant ^aProphylactic use of a steroid-containing "swish and spit" regimen was protocol-mandated; oral non-sedating antihistamine therapy was recommended [abbreviations footnote small, partly illegible] Study 2 Data cut-off: March 9, 2026 Study 1 Data cut-off: May 30, 2025 Presenter: Barbara Pistilli, MD [Slide 2] Conclusions • Gedatolisib + fulvestrant ± palbociclib significantly improved median PFS compared with alpelisib + fulvestrant in patients with and without visceral or liver metastases - Liver metastases: mPFS 9.1–11.1 vs 4.6 months (present); 11.5–21.5 vs 9.1 months (absent) - Visceral metastases: mPFS 9.8–11.3 vs 5.6 months (present); 21.5 months–NR vs 9.2 months (absent) • Responses (confirmed and unconfirmed) were more frequent and durable with gedatolisib-based regimens - ORR: 43–59% vs 33% - mDOR: 19.4–27.8 vs 7.5 months • Gedatolisib discontinuations due to treatment-emergent adverse events were infrequent compared with alpelisib - Geda + Palbo + Fulv: 5% (1% for stomatitis, 0% for hyperglycemia) - Geda + Fulv: 4% (0% due to stomatitis or hyperglycemia) - Alpe + Fulv: 19% (3% for diarrhea, 1% for hyperglycemia) • Grade 3 stomatitis with gedatolisib was transient: <3% by cycle 2; ~89% of grade 3 events improved to a lower grade within 1-2 weeks These analyses support gedatolisib + fulvestrant ± palbociclib as a potential new standard of care for HR+/HER2-/PIK3CA-MT ABC after progression on a CDK4/6 inhibitor [abbreviations footnote blurred] Presenter: Barbara Pistilli, MD

[Slide 1] VIKTORIA-1 Study Design HR+/HER2- Advanced Breast Cancer Eligibility Criteria • Pre- & postmenopausal women & men • Progression on/after CDK4/6i + NSAI • ≤2 lines of prior ET for ABC • Measurable disease, RECIST v1.1 • Screening result for PIK3CA status • No T2DM with HbA1c >6.4% or T1DM • No prior mTORi, PI3Ki, or AKTi • No prior chemotherapy for ABC Stratification Factors • Lung/liver metastases (yes/no) • TTP on immediate prior therapy (≤ or >6 months) • Region (US/Canada or ROW) Assign → STUDY 2 PIK3CA-MT → R 3:3:1 (N=362) Arm D^a: Gedatolisib 180 mg IV once weekly, 3 weeks on, 1 week off; Palbociclib 125 mg daily, 21 days on, 7 days off; Fulvestrant 500 mg; cycle 1, days 1 & 15, then every 4 weeks Arm E: Alpelisib 300 mg orally daily for 4 weeks; Fulvestrant 500 mg; cycle 1, days 1 & 15, then every 4 weeks Arm F^a: Gedatolisib 180 mg IV once weekly, 3 weeks on, 1 week off; Fulvestrant 500 mg; cycle 1, days 1 & 15, then every 4 weeks Primary Endpoint: PFS (BICR) Arm D vs Arm E Secondary Endpoints: OS; PFS (BICR F vs E); Response; Safety; QoL Assign → STUDY 1 PIK3CA-WT → R 1:1:1 (N=392) Arm A^a: Gedatolisib + Palbociclib + Fulvestrant Arm B^a: Gedatolisib^a + Fulvestrant Arm C: Fulvestrant ^a Prophylactic use of a steroid-containing "swish and spit" regimen was protocol-mandated; oral non-sedating antihistamine therapy was recommended ABC, advanced breast cancer; AKTi, protein kinase B inhibitor; BICR, blinded independent central review; CDK4/6i, cyclin-dependent kinase 4 and 6 inhibitor; ET, endocrine therapy; HbA1c, hemoglobin A1c; HER2-, human epidermal growth factor receptor 2-negative; HR+, hormone receptor-positive; IV, intravenous; MT, mutant; mTORi, mechanistic target of rapamycin inhibitor; NSAI, non-steroidal aromatase inhibitor; OS, overall survival; PFS, progression-free survival; PI3Ki, phosphatidylinositol 3-kinase inhibitor; QoL, quality of life; R, randomization; RECIST v1.1, Response Evaluation Criteria in Solid Tumors, version 1.1; ROW, rest of world; T1DM, type 1 diabetes mellitus; T2DM, type 2 diabetes mellitus; TTP, time to progression; US, United States; WT, wild-type. Study 2 Data cut-off: March 9, 2026 Study 1 Data cut-off: May 30, 2025 Presenter: Barbara Pistilli, MD [Slide 2] PFS in Key Subgroup: Liver Metastases VIKTORIA-1 Study 2 No Liver Metastases — Gedatolisib Triplet G + P + F (n=77) | Alpe + Fulv (n=76) Median PFS, mos (95% CI): 21.5 (11.4-NR) | 9.1 (5.6-16.9) HR (90% CI): 0.46 (0.30-0.72); P = 0.0016 No. at Risk: G + P + F 77 72 66 65 56 55 44 43 34 29 25 23 15 15 13 11 11 9 9 9 9 9 7 5 5 5 2 2 0 Alpe + Fulv 76 70 63 61 45 43 34 33 25 25 23 22 18 16 11 11 9 7 7 7 4 3 2 1 1 1 1 1 1 1 1 0 Liver Metastases — Gedatolisib Triplet G + P + F (n=78) | Alpe + Fulv (n=79) Median PFS, mos (95% CI): 9.1 (5.7-11.1) | 4.6 (3.7-5.6) HR (90% CI): 0.47 (0.34-0.65); P < 0.0001 No. at Risk: G + P + F 78 78 68 66 54 52 33 33 28 28 17 15 12 10 7 5 5 3 3 3 0 Alpe + Fulv 79 78 53 46 33 31 19 18 10 9 7 5 4 4 2 2 2 1 1 1 1 1 1 1 0 No Liver Metastases — Gedatolisib Doublet Geda + Fulv (n=28) | Alpe + Fulv (n=76) Median PFS, mos (95% CI): 11.5 (9.2-NR) | 9.1 (5.6-16.9) HR (90% CI): 0.71 (0.41-1.25); P = 0.1627 No. at Risk: Geda + Fulv 28 25 21 20 20 20 17 17 12 12 6 6 5 5 4 3 3 3 3 3 2 2 2 1 1 1 1 1 1 0 Alpel + Fulv 76 70 63 61 45 43 34 33 25 25 23 22 18 16 11 11 9 7 7 7 4 3 2 1 1 1 1 1 1 1 1 0 Liver Metastases — Gedatolisib Doublet Geda + Fulv (n=24) | Alpe + Fulv (n=79) Median PFS, mos (95% CI): 11.1 (3.7-NR) | 4.6 (3.7-5.6) HR (90% CI): 0.38 (0.23-0.62); P = 0.0006 No. at Risk: Geda + Fulv 24 23 21 21 16 16 14 13 11 11 9 9 5 5 4 4 4 4 3 2 2 1 1 1 1 1 1 1 1 0 Alpe + Fulv 79 78 53 46 33 31 19 18 10 9 7 5 4 4 2 2 2 1 1 1 1 1 1 1 0 Alpe, alpelisib; CI, confidence interval; F, fulvestrant; Fulv, fulvestrant; G, gedatolisib; Geda, gedatolisib; HR, hazard ratio; NR, not reached; P, palbociclib; PFS, progression-free survival. Study 2 Data cut-off: March 9, 2026 Presenter: Barbara Pistilli, MD Survival table Endpoint: PFS by liver metastases, VIKTORIA-1 Study 2 Arm | N | Median | Median CI G+P+F, no liver mets | 77 | 21.5 mo | 11.4-NR Alpe+Fulv, no liver mets | 76 | 9.1 mo | 5.6-16.9 G+P+F, liver mets | 78 | 9.1 mo | 5.7-11.1 Alpe+Fulv, liver mets | 79 | 4.6 mo | 3.7-5.6 Geda+Fulv, no liver mets | 28 | 11.5 mo | 9.2-NR Geda+Fulv, liver mets | 24 | 11.1 mo | 3.7-NR HR: Liver mets triplet 0.47; no liver mets triplet 0.46; doublet no liver 0.71; doublet liver 0.38 HR CI: 90% CIs: 0.34-0.65; 0.30-0.72; 0.41-1.25; 0.23-0.62 P: <0.0001; 0.0016; 0.1627; 0.0006 [Slide 3] PFS in Key Subgroup: Visceral Metastases (Lung/Liver) VIKTORIA-1 Study 2 No Visceral Metastases — Gedatolisib Triplet G + P + F (n=38) | Alpe + Fulv (n=37) Median PFS, mos (95% CI): 21.5 (9.6-NR) | 9.2 (3.9-16.9) HR (90% CI): 0.38 (0.21-0.73); P = 0.0059 No. at Risk: G + P + F 38 37 33 32 31 31 23 23 20 17 14 12 10 10 9 7 7 6 6 6 5 5 3 2 2 2 2 2 0 Alpe + Fulv 37 35 33 32 19 18 14 14 11 11 10 10 7 5 4 4 3 1 1 1 0 Visceral Metastases — Gedatolisib Triplet G + P + F (n=117) | Alpe + Fulv (n=118) Median PFS, mos (95% CI): 9.8 (7.5-12.9) | 5.6 (4.6-7.2) HR (90% CI): 0.52 (0.39-0.70); P < 0.0001 No. at Risk: G + P + F 117 113 101 99 79 76 54 53 42 40 28 26 17 15 11 9 9 6 6 4 4 4 3 3 3 0 Alpe + Fulv 118 113 83 75 59 56 39 37 24 23 20 17 15 15 9 9 8 7 7 7 5 4 3 1 1 1 1 1 1 1 1 0 No Visceral Metastases — Gedatolisib Doublet Geda + Fulv (n=13) | Alpe + Fulv (n=37) Median PFS, mos (95% CI): NR (1.6-NR) | 9.2 (3.9-16.9) HR (90% CI): 0.40 (0.15-1.1); P = 0.0703 No. at Risk: Geda + Fulv 13 11 9 9 9 9 8 8 4 4 1 1 1 1 1 1 1 1 1 1 0 Alpe + Fulv 37 35 33 32 19 18 14 14 11 11 10 10 7 5 4 4 3 1 1 1 0 Visceral Metastases — Gedatolisib Doublet Geda + Fulv (n=39) | Alpe + Fulv (n=118) Median PFS, mos (95% CI): 11.3 (7.4-22.1) | 5.6 (4.6-7.2) HR (90% CI): 0.53 (0.35-0.79); P = 0.0042 No. at Risk: Geda + Fulv 39 37 33 32 27 27 23 22 19 19 14 14 9 9 7 6 6 6 6 5 4 4 3 2 2 2 2 1 1 1 0 Alpe + Fulv 118 113 83 75 59 56 39 37 24 23 20 17 15 15 9 9 8 7 7 7 5 4 3 1 1 1 1 1 1 1 1 0 Alpe, alpelisib; CI, confidence interval; F, fulvestrant; Fulv, fulvestrant; G, gedatolisib; Geda, gedatolisib; HR, hazard ratio; NR, not reached; P, palbociclib; PFS, progression-free survival. Study 2 Data cut-off: March 9, 2026 Presenter: Barbara Pistilli, MD Survival table Endpoint: PFS by visceral (lung/liver) metastases, VIKTORIA-1 Study 2 Arm | N | Median | Median CI G+P+F, no visceral mets | 38 | 21.5 mo | 9.6-NR Alpe+Fulv, no visceral mets | 37 | 9.2 mo | 3.9-16.9 G+P+F, visceral mets | 117 | 9.8 mo | 7.5-12.9 Alpe+Fulv, visceral mets | 118 | 5.6 mo | 4.6-7.2 Geda+Fulv, no visceral mets | 13 | NR | 1.6-NR Geda+Fulv, visceral mets | 39 | 11.3 mo | 7.4-22.1 HR: Visceral triplet 0.52; no visceral triplet 0.38; doublet no visceral 0.40; doublet visceral 0.53 HR CI: 90% CIs: 0.39-0.70; 0.21-0.73; 0.15-1.1; 0.35-0.79 P: <0.0001; 0.0059; 0.0703; 0.0042

[Slide 1] Efficacy Summary VIKTORIA-1 Study 2 Endpoint | Geda + Palbo/Fulv (n=155) | Geda + Fulvestrant (n=52) | Alpelisib + Fulvestrant (n=155) mPFS, no liver metastases | 21.5 months (n=77) | 11.5 months (n=28) | 9.1 months (n=76) mPFS, liver metastases | 9.1 months (n=78) | 11.1 months (n=24) | 4.6 months (n=79) mPFS, no visceral metastases | 21.5 months (n=38) | Not reached (n=13) | 9.2 months (n=37) mPFS, visceral metastases | 9.8 months (n=117) | 11.3 months (n=39) | 5.6 months (n=118) ORR^a, response evaluable | 59% (n=133) | 43% (n=42) | 33% (n=127) mDOR^a | 19.4 months (n=78) | 27.8 months (n=18) | 7.5 months (n=42) ^a Includes patients with unconfirmed partial responses, confirmed partial responses, and complete responses. Partial responders included uPRs in Arm D (n=13), Arm E (n=9), and Arm F (n=3). DOR, duration of response; Fulv, fulvestrant; Geda, gedatolisib; m, median; ORR, objective response rate; Palbo, palbociclib; PFS, progression-free survival; uPR, unconfirmed partial response. Study 2 Data cut-off: March 9, 2026 Presenter: Barbara Pistilli, MD [Slide 2] Conclusions • Gedatolisib + fulvestrant ± palbociclib significantly improved median PFS compared with alpelisib + fulvestrant in patients with and without visceral or liver metastases – Liver metastases: mPFS 9.1–11.1 vs 4.6 months (present); 11.5–21.5 vs 9.1 months (absent) – Visceral metastases: mPFS 9.8–11.3 vs 5.6 months (present); 21.5 months–NR vs 9.2 months (absent) • Responses (confirmed and unconfirmed) were more frequent and durable with gedatolisib-based regimens – ORR: 43–59% vs 33% – mDOR: 19.4–27.8 vs 7.5 months • Gedatolisib discontinuations due to treatment-emergent adverse events were infrequent compared with alpelisib – Geda + Palbo + Fulv: 5% (1% for stomatitis, 0% for hyperglycemia) – Geda + Fulv: 4% (0% due to stomatitis or hyperglycemia) – Alpe + Fulv: 19% (3% for diarrhea, 1% for hyperglycemia) • Grade 3 stomatitis with gedatolisib was transient: <3% by cycle 2; ~89% of grade 3 events improved to a lower grade within 1-2 weeks These analyses support gedatolisib + fulvestrant ± palbociclib as a potential new standard of care for HR+/HER2-/PIK3CA-MT ABC after progression on a CDK4/6 inhibitor ABC, advanced breast cancer; Alpe, alpelisib; CDK4/6, cyclin-dependent kinase 4 and 6; DOR, duration of response; Fulv, fulvestrant; Geda, gedatolisib; HER2-, human epidermal growth factor receptor 2-negative; HR+, hormone receptor-positive; m, median; MT, muta[tion]; NR, not reached; ORR, objective response rate; Palbo, palbociclib; PFS, progression-free survival Presenter: Barbara Pistilli, MD

Open questions - Do you need a triplet?....Would a doublet accomplish similar benefit and spare some toxicity/cost? - IV/IM/+/- PO.....will patients and physicians accept an endocrine strategy that requires near weekly visits? - Would there be any different results if co-mutations were present?....samples not collected as part of the study - If both options available, flexibility to escalate or de-escalate depending on the patient. Bill Gradishar MD

Grade 3 Stomatitis: Prevalence by Treatment Cycle VIKTORIA-1 Study 2 (Safety Analysis Set, Months 1-12) Prevalence of Stomatitis Arm D: Gedatolisib + Palbociclib + Fulvestrant [Bar chart: Patients (%) 0-20 by Treatment Month M1-M12] Prevalence of Stomatitis Arm F: Gedatolisib + Fulvestrant [Bar chart: Patients (%) 0-20 by Treatment Month M1-M12] By cycle 2, fewer than 3% of patients in either gedatolisib arm had grade 3 stomatitis, and about 89% of grade 3 events improved to a lower grade within 1 to 2 weeks. [Study] 2 Data cut-off: March 9, 2026 [Pres]enter: Barbara Pistilli, MD

VIKTORIA-1 Study 2: Efficacy Summary Endpoint | Gedatolisib + Palbociclib + Fulvestrant | Gedatolisib + Fulvestrant | Alpelisib + Fulvestrant mPFS, no visceral metastases (n=38/13/37) | 21.5 months | Not reached | 9.2 months mPFS, visceral metastases (n=117/39/118) | 9.8 months | 11.3 months | 5.6 months mPFS, no liver metastases (n= 77/28/78) | 21.5 months | 11.5 months | 9.1 months mPFS, liver metastases (n=78/24/79) | 9.1 months | 11.1 months | 4.6 months ORR (n=133/42/127) | 59% | 43% | 33% mDOR (n=133/42/127) | 19.4 months | 27.8 months | 7.5 months DOR, duration of response; m, median; ORR, objective response rate; PFS, progression-free survival Study 2 Data cut off: March 9, 2026 Presenter: Barbara Pistilli, MD

APPROVED PI3K/AKT/mTOR-PATHWAY REGIMENS MOVING TO NEW POPULATIONS Where the triplets are headed: expansion in motion All three approved PI3K/AKT/mTOR-pathway regimens are now being tested in an earlier or different 1L population — the same agents, a new setting. The strategy: extend proven regimens along the endocrine-sensitivity axis. APPROVED (today) Inavolisib · PI3Kα inhibitor INAVO120 Endocrine-RESISTANT, PIK3CA-mut + Palbociclib + Fulvestrant — PFS 15.0 vs 7.3 mo (HR 0.43) · FDA approved 10/2024 Capivasertib · AKT inhibitor CAPItello-291 Post-ET 2L, AKT-pathway altered + Fulvestrant — PFS 7.3 vs 3.1 mo (HR 0.50) · FDA approved 11/2023 Gedatolisib · pan-PI3K / mTORC1/2 inhibitor VIKTORIA-1 Post-CDK4/6i 2L, PIK3CA-WT (mut sNDA submitted 8/2026) + Fulv ± Palbo — PFS 9.3 / 7.4 vs 2.0 mo (HR 0.24 / 0.33) · FDA approved 7/2026 population shift ONGOING (expansion) ONGOING INAVO123 Endocrine-SENSITIVE, PIK3CA-mut + CDK4/6i + Letrozole — phase III, ~450 pts · primary EP PFS (NCT06790693) ONGOING CAPItello-292 1L, ET-resistant (relapse on / ≤ 12 mo after adjuvant ET) + CDK4/6i + Fulvestrant — phase Ib/III triplet · all-comers (NCT04862663) ONGOING VIKTORIA-2 1L, ET-resistant (Study 1) + ET-sensitive (Study 2) cohorts + Palbo + Fulvestrant (Study 1) or + Palbo + Letrozole (Study 2) vs Ribociclib + ET — phase III (NCT06757634) The pattern: prove a regimen in one population, then move it along the axis — inavolisib toward endocrine-sensitive, capivasertib toward CDK4/6i-combined 1L, gedatolisib toward treatment-naïve 1L. If positive, the PI3K-pathway branch shifts even earlier, reshaping the whole sequence. INAVO123 (NCT06790693), CAPItello-292 (NCT04862663), VIKTORIA-2 (NCT06757634) — ongoing, no efficacy readout yet. Approved-regimen figures per reported readouts; status as of Oct 2026.

PFS in Key Subgroup: Liver Metastases VIKTORIA-1 Study 2 No Liver Metastases: G + P + F (n=77) 21.5 (11.4-NR) | Alpe + Fulv (n=76) 9.1 (5.6-16.9); HR (90% CI) 0.4[6] (0.30-0.72), P = 0.0016 Liver Metastases: G + P + F (n=78) 9.1 (5.7-11.1) | Alpe + Fulv (n=79) 4.6 (3.7-5.6); HR (90% CI) 0.47 (0.34-0.65), P < 0.0001 Gedatolisib Doublet, No Liver Metastases: Geda + Fulv (n=28) 11.5 (9.2-NR) | Alpe + Fulv (n=76) 9.1 (5.6-16.9); HR 0.71 (0.41-1.25), P = 0.1627 Gedatolisib Doublet, Liver Metastases: Geda + Fulv (n=24) 11.1 (3.7-NR) | Alpe + Fulv (n=79) 4.6 (3.7-5.6); HR 0.38 (0.23-0.62), P = 0.0006 No. at risk rows: [illegible at this resolution] Presenter: Barbara Pistilli, MD Study 2 Data cut-off: March 9, 2026 Survival table Endpoint: PFS by liver metastases, VIKTORIA-1 Study 2 Arm | N | Median | Median CI G+P+F, no liver mets | 77 | 21.5 mo | 11.4-NR Alpe+Fulv, no liver mets | 76 | 9.1 mo | 5.6-16.9 G+P+F, liver mets | 78 | 9.1 mo | 5.7-11.1 Alpe+Fulv, liver mets | 79 | 4.6 mo | 3.7-5.6 Geda+Fulv, no liver mets | 28 | 11.5 mo | 9.2-NR Geda+Fulv, liver mets | 24 | 11.1 mo | 3.7-NR HR: Liver mets triplet 0.47; no liver mets triplet 0.46; doublet no liver 0.71; doublet liver 0.38 HR CI: 90% CIs: 0.34-0.65; 0.30-0.72; 0.41-1.25; 0.23-0.62 P: <0.0001; 0.0016; 0.1627; 0.0006

PRO-CTCAE joint pain: More patients on giredestrant vs SOC ET reported "never" experiencing joint pain, with minimal changes in severity and interference from BL Frequency (legend: Never; Rarely; Occasionally; Frequently; Almost constantly) Severity (legend: None; Mild; Moderate; Severe; Very severe) Interference (legend: Not at all; A little bit; Somewhat; Quite a bit; Very much) Rows: Giredestrant Patients; SOC ET Patients; stacked bars 0-100% at BL, C2, C3, C6, C9, C12, C15 with 50% reference line Patients (n) - Frequency: Giredestrant 1579 1558 1543 1544 1571 1574 1572; SOC ET 1566 1545 1537 1517 1540 1509 1487 Patients (n) - Severity: Giredestrant 964 1015 1085 1183 1219 1215 1227; SOC ET 1004 1108 1188 1227 1247 1179 1192 Patients (n) - Interference: Giredestrant 912 970 1040 1142 1173 1173 1200; SOC ET 943 1046 1142 1185 1194 1132 1146 - Increased frequency in both arms with longer treatment duration - Fewer giredestrant-treated patients reporting "frequent/almost constant" joint pain vs SOC ET Comparable severity and interference between giredestrant and SOC ET Data cutoff: 8 August 2025. Denominators are numbers of patients who completed PRO-CTCAE responses at the key timepoints shown. BL, baseline (C1); C, cycle; PRO-CTCAE, Patient-Reported Outcomes version of the Common Terminology Criteria for Adverse Events; SOC ET, standard-of-care endocrine therapy. Presented by Gustavo Werutsky, MD, PhD.

[Slide 1] The unmet need in ER+ early breast cancer Distant recurrence over 20 years, by nodal status — Risk of Distant Recurrence N4–9: 22, 36, 45, 52 | N1–3: 10, 19, 25, 31 | N0: 6, 11, 16, 22 (at years 5, 10, 15, 20) No. at Risk (0, 5, 10, 15, 20 yrs): N4–9 12,333 8,116 2165 259 52 | N1–3 31,936 23,576 7250 949 183 | N0 29,925 24,081 8571 1982 414 No. of Events — annual rate (%): N4–9 2568 (4.8) 969 (4.0) 121 (3.1) 13 (2.2) | N1–3 3126 (2.2) 1421 (1.9) 241 (1.7) 39 (1.8) | N0 1646 (1.2) 835 (1.1) 272 (1.3) 68 (1.4) ~70% of breast cancers are ER+/HER2− 20–40% distant recurrence despite ET, by risk features ~1 in 3 patients on adjuvant ET eventually recur ~50% stop ET before 5 years, mainly due to toxicity The Unmet Need is to Identify: Less for some. More for others. Tailored to clinical + molecular risk. Pan H et al. N Engl J Med 2017;377:1836-46 (figure); Hershman DL et al. J Clin Oncol 2010;28:4120-8 Komal Jhaveri Refining endocrine therapy with CDK4/6 inhibitors in early breast cancer [Slide 2] Why a Better Endocrine Backbone in Early Disease? Preclinical: deeper ER blockade | AI | TAM | Giredestrant Antagonism | None | Partial | Full Degrades ER | No | No | Yes ESR1-mutant | Escapes | Reduced | Active • More potent antiproliferation than earlier SERDs across ER+ cell lines • Regression of ESR1 Y537S PDX at low oral doses, alone or with a CDK4/6i Liang et al. J Med Chem 2021;64:11841–56 coopERA: Ki-67 suppression Giredestrant vs anastrozole × 2wk (n = 221) Geometric mean Ki-67 change at week 2 (%): Giredestrant −75%; Anastrozole −67%; p = 0.043 Hurvitz et al. Lancet Oncol 2023;24:1029–41 Real-world adherence to ET Kaiser Permanente cohort (n = 8,769) Discontinued by 4.5 years 32% Non-adherent among those continuing 28% Full duration, on schedule 49% Only about half of patients finish 5 years of adjuvant ET as prescribed Hershman et al. J Clin Oncol 2010;28:4120–8 Adherence over 5+ years matters as much as potency AI, aromatase inhibitor; ET, endocrine therapy; PDX, patient-derived xenograft; TAM, tamoxifen. Non-adherence = taking ≤80% of prescribed doses. Komal Jhaveri Refining endocrine therapy with CDK4/6 inhibitors in early breast cancer [Slide 3] Putting lidERA into Context for the Clinic: Open Questions No SOC CDK4/6i use in lidERA: Could SERD + CDK4/6i additive? Can SERD replace CDK4/6i? Who? Early readout: will benefit persist? Late events? Luminal A vs B? Fewer events in stage I and medium-risk groups: needs maturity Premenopausal (40%): heterogeneous treatment Few ESR1m tumors: why SERD over AI? Biomarkers to predict benefit? Cost when giredestrant is approved [Slide 4] Key Take Home Messages: From one size fits many to the right therapy at the right time for the right patient 1. CDK4/6 inhibitors: proven, durable carryover benefit, and established OS with abemaciclib and OS trend with ribociclib 2. Oral SERDs: lidERA is the first positive adjuvant SERD trial, with more phase III readouts ahead as early as 2027 3. ctDNA/MRD: the next frontier is adapting therapy to molecular risk over time, currently being tested in ongoing trials Komal Jhaveri Refining endocrine therapy with CDK4/6 inhibitors in early breast cancer

Conclusions Sustained long-term tolerability with giredestrant: - Over 50% of giredestrant-treated patients reported zero side-effect bother across more than 3 years of treatment Reduced joint pain frequency with giredestrant: - More patients with giredestrant vs AIs reported "never" experiencing any joint pain over the first 15 cycles of treatment, mitigating a major cause of early treatment discontinuation Comparable hot flashes and fatigue profiles to SOC ET: - The majority of patients in both arms reported low levels of these symptoms Preserved day-to-day function and HRQoL with giredestrant: - Patients maintained physical and role functioning, and HRQoL - a critical requirement for patients receiving curative treatment Giredestrant demonstrated superior efficacy vs SOC ET in early breast cancer, with favourable tolerability and preserved HRQoL from the patient perspective AI, aromatase inhibitor; HRQoL, health-related quality of life; SOC ET, standard-of-care endocrine therapy. Presented by Gustavo Werutsky, MD, PhD.

PRO-CTCAE joint pain: More patients on giredestrant vs AIs reported "never" experiencing any joint pain within the first 15 cycles Giredestrant vs AIs: Proportion of patients who reported "never" experiencing joint pain Critical 6-month window of peak ET discontinuation [box around C2-C6] Patients reporting "never" (Giredestrant | AIs | difference): BL: 38.9% | 35.9% C2: 34.9% | 28.2% | +6.7% C3: 29.7% | 22.2% | +7.5% C6: 23.4% | 17.4% | +6.0% C9: 22.4% | 17.3% | +5.1% C12: 22.8% | 20.5% | +2.3% C15: 21.9% | 18.6% | +3.3% Chi-square analysis at C6 (absolute difference: 6%) - Odds of reporting "never" having joint pain were 44% higher with giredestrant vs AIs (OR: 1.44 [95% CI: 1.20, 1.73]; descriptive p = 0.0001) Whilst clinician-reported AE rates for arthralgia were similar between giredestrant and the AI subgroup, PRO data demonstrated a meaningful difference Data cutoff: 8 August 2025. Denominators are numbers of patients who completed PRO-CTCAE responses at the key timepoints shown. AE, adverse event; AI, aromatase inhibitor; BL, baseline (C1); C, cycle; CI, confidence interval; ET, endocrine therapy; OR, odds ratio; (PRO)-CTCAE, (Patient-Reported Outcomes) version of the Common Terminology Criteria for Adverse Events. Presented by Gustavo Werutsky, MD, PhD.

Day-to-day functioning and HRQoL were preserved with giredestrant; no clinically meaningful deterioration was observed in either treatment arm [Line charts: Mean change from BL, -10 to 10; Giredestrant vs SOC ET; Improving / Worsening] Physical functioning - High BL physical functioning fully maintained through 3+ years of therapy BL scores: Giredestrant: 86.8; SOC ET: 86.6 Patients (n) Giredestrant/SOC ET: BL 1601/1586; C3 1537/1519; C6 1511/1452; C9 1467/1414; C12 1422/1365; C15 1401/1322; C18 1369/1293; C21 1349/1253; C24 1332/1227; C30 1057/966 Role functioning - High BL role functioning fully maintained through 3+ years of therapy BL scores: Giredestrant: 83.9; SOC ET: 84.0 Patients (n): BL 1601/1586; C3 1537/1519; C6 1511/1452; C9 1467/1414; C12 1422/1365; C15 1400/1322; C18 1369/1293; C21 1349/1253; C24 1332/1227; C30 1057/966 HRQoL - HRQoL stable across treatment cycles BL scores: Giredestrant: 74.6; SOC ET: 73.7 Patients (n): BL 1601/1586; C3 1536/1516; C6 1511/1449; C9 1467/1414; C12 1421/1365; C15 1401/1321; C18 1369/1293; C21 1349/1253; C24 1332/1227; C30 1057/966 Data cutoff: 8 August 2025. EORTC QLQ-C30 responses up to Cycle 30 shown, where > 50% of patients remained on-treatment. BL, baseline; C, cycle; EORTC QLQ-C30, European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire C-30; HRQoL, health-related quality of life; SOC ET, standard-of-care endocrine therapy. Presented by Gustavo Werutsky, MD, PhD.

[Slide 2] A Lesson from Moneyball: Adjuvant ER+ Therapy has made the same shift Scouting → Clinical risk Treat by appearance: stage, nodes, grade Analytics → Molecular risk Treat by biology, measured over time [Baseball field diagram] [Slide 3] Formula 1: The Telemetry Works. Does an Early Pit Stop Win the Race? LIGHTS OUT THRESHOLD: MRD POSITIVE PIT STOP? under study Radiologic recurrence Better finish? LAPS → serial ctDNA/MRD monitoring over time Lead time: months before imaging Live telemetry = serial ctDNA/MRD: prognostic, with lead time Steady laps? = stay the course (de-escalation also unproven) Pit stop? = escalate or switch early: being tested in trials The signal is real and early; whether acting on it changes the finish is what trials must prove.

16:15 - 17:45 Rapid oral: Breast cancer, early stage CHAIRS: MARIA VITTORIA DIECI, SANDRA XIMENA FRANCO MILLAN, LILIT HARUTYUNYAN FACT-G item GP5 overall bother: Sustained tolerability with giredestrant, with low overall bother from side effects [Line chart; y-axis: Patients reporting being "not at all" bothered, 0%-100%; 50% reference] "Not at all" bothered Giredestrant: C3 53%, C12 51%, C24 54%, C42 51% SOC ET: C3 45%, C12 47%, C24 50%, C42 46% Patients (n/N) Giredestrant: C3 812/1543; C12 799/1572; C24 820/1519; C42 130/256 SOC ET: C3 692/1533; C12 714/1508; C24 701/1416; C42 100/219 Across more than 3 years of treatment: - > 50% of giredestrant-treated patients reported being "not at all" bothered by treatment-related side-effects; a consistently higher proportion vs SOC ET Data cutoff: 8 August 2025. Numerators are numbers of patients who answered "not at all" and denominators are numbers of patients who completed FACT-G item GP5 at the key timepoints shown. C, cycle; FACT-G, Functional Assessment of Cancer Therapy - General; SOC ET, standard-of-care endocrine therapy. Presented by Gustavo Werutsky, MD, PhD. Gustavo Werutsky Patient (pt)-reported outcomes (PROs) from the Phase III lidERA Breast Cancer (BC) trial of giredestrant (GIRE) in pts with oestrogen receptor-positive, HER2-negative early BC (ER+, HER2- eBC) ALICANTE AUDITORIUM - HALL 6

ESMO CONGRESS 2026 PRO-CTCAE joint pain: More patients on giredestrant vs SOC E[T] experiencing joint pain, with minimal changes in severity and i[nterference] Frequency Severity Never | Rarely | Occasionally | Frequently | Almost constantly None | Mild | Moderate | Severe | Very severe Giredestrant Patients SOC ET Patients BL C2 C3 C6 C9 C12 C18 Patients (n) • Increased frequency in both arms with longer treatment duration • Fewer giredestrant-treated patients reporting "frequent/almost constant" joint pain vs SOC ET Comparable severity and interference between gi[redestrant and SOC ET] Presented by Gustavo Werutsky, MD, PhD. RAPID ORAL SESSION 538RO: Giredestrant in Early Breast Cancer: Insights into Patients Quality of Life

PRO-CTCAE joint pain: More patients on giredestrant vs SOC ET reported "never" experiencing joint pain, with minimal changes in severity and interference from BL lidERA Breast Cancer Frequency | Severity | Interference Never / Rarely / Occasionally / Frequently / Almost constantly None / Mild / Moderate / Severe / Very severe Not at all / A little bit / Somewhat / Quite a bit / Very much Giredestrant Patients SOC ET Patients BL C2 C3 C6 C9 C12 C15 Patients (n) Giredestrant 1579 1558 1543 1544 1571 1574 1572 SOC ET [illegible] • Increased frequency in both arms with longer treatment duration • Fewer giredestrant-treated patients reporting "frequent/almost constant" joint pain vs SOC ET Comparable severity and interference between giredestrant and SOC ET Data cutoff: 8 August 2025. Denominators are numbers of patients who completed PRO-CTCAE responses at the key timepoints shown. BL, baseline (C1); C, cycle; PRO-CTCAE, Patient-Reported Outcomes version of the Common Terminology Criteria for Adverse Events; SOC ET, standard-of-care endocrine therapy Presented by Gustavo Werutsky, MD, PhD. People Tolerated Giredestrant Better Than Tamoxifen, Aromatase Inhibitors ESMO Congress 2026 / BREASTCANCER.ORG

Proffered Paper session LBA24 - Final analysis of NRG oncology/NSABP B-42, a randomized, double-blinded, placebo-controlled trial of extended letrozole in postmenopausal women with hormone-receptor positive breast cancer who have completed previous adjuvant therapy with an aromatase inhibitor Speakers Eleftherios P. Mamounas (Orlando, United States of America) Lecture Time 14:15 - 14:25 Abstract Background NSABP B-42 evaluated whether 5 yrs of letrozole (L) vs placebo (P) improves disease-free survival (DFS) in patients who received 5 yrs of endocrine therapy. Previous 10-yr results showed that L significantly improved the primary endpoint of DFS and secondary endpoints of breast cancer-free interval (BCFI) and distant recurrence (DR) with no significant improvement in overall survival (OS). Here we report the final trial analysis of 15-yr outcome data. Methods A total of 3,966 postmenopausal patients with stage I-III, hormone-receptor positive breast cancer, disease-free after 5 yrs of an aromatase inhibitor (AI) or ≤3 yrs of tamoxifen à AI for total of 5 yrs, were randomized 1:1 to 5 yrs of L 2.5 mg PO daily or P. Stratification was by pathologic nodal status, prior tamoxifen or not, and baseline DEXA T scores (>-2.0, ≤-2.0 SD). Primary endpoint was DFS. Secondary endpoints included OS, BCFI and DR. Differences in outcomes between L and P were assessed by stratified log-rank tests and Cox proportional hazards models. Results Follow-up of trial patients was completed in 08/26, with median follow-up of 15.9 yrs. Since the 10-yr update, 538 additional DFS events were recorded. The 15-yr results continue to show a modest but significant improvement in DFS (HR=0.85, 95%CI=0.77 to 0.94,15-yr DFS: P=56.5%, L=61.6%), BCFI (HR=0.78, 95%CI=0.67 to 0.92,15-yr BCFI risk: P=18.3%, L=14.7%) and DR (HR=0.72, 95%CI=0.57 to 0.91,15-yr DR risk: P=9.7%, L=6.9%). The L effect on DFS was consistent across subgroups by stratification factors. There was no significant improvement in OS (HR=0.92, 95%CI=0.82 to 1.05, 15-yr OS: P=72.3%, L=74.0%). Conclusions With over 15 yrs of follow-up, the B-42 trial continues to demonstrate a sustained and clinically meaningful benefit with extended letrozole therapy on DFS and key secondary endpoints consistent with the original and 10-yr updated results. No attenuation of treatment effect was observed from year 10 to year 15.

Distant Recurrence by Nodal Status Node-Negative Analysis 15-yr | # Events P 54, L 42 | HR (95% CI) 0.78 (0.52-1.16) | P-value 0.22 [Cumulative incidence curve; y-axis Cumulative Incidence of DR 0.00-0.20; x-axis Years after Random Assignment 0-15] Curve labels: 10 yr 4.0% (placebo) / 3.3% (letrozole); 15 yr 5.6% (placebo) / 4.0% (letrozole); Delta = 1.6% Number at risk (0, 5, 10, 15 yr): Placebo 1116, 975, 762, 475; Letrozole 1126, 979, 776, 491 Node-Positive Analysis 15-yr | # Events P 115, L 81 | HR (95% CI) 0.69 (0.52-0.92) | P-value 0.011 Curve labels: 10 yr 12.1% (placebo) / 8.9% (letrozole); 15 yr 15.1% (placebo) / 10.7% (letrozole); Delta = 4.4% Number at risk (0, 5, 10, 15 yr): Placebo 838, 694, 524, 314; Letrozole 826, 699, 531, 329 Patients excluded: - 38 (no follow-up, or not at risk for DFS) - 22 (no clinical follow-up assessment) Eleftherios P. Mamounas Survival table Endpoint: Cumulative incidence of distant recurrence (DR) by nodal status, 15-yr analysis Node-negative - Placebo | 1116 Node-negative - Letrozole | 1126 Node-positive - Placebo | 838 Node-positive - Letrozole | 826 Landmark rates: 10 yr: Node-negative - Placebo 4.0% | Node-negative - Letrozole 3.3% | Node-positive - Placebo 12.1% | Node-positive - Letrozole 8.9% 15 yr: Node-negative - Placebo 5.6% | Node-negative - Letrozole 4.0% | Node-positive - Placebo 15.1% | Node-positive - Letrozole 10.7%

[Slide 1] NSABP B-42 Conclusions - 15-year results of B-42 continue to demonstrate that extended L results in: - Statistically significant improvement in DFS: 15% reduction in DFS event; 5.1% absolute improvement - Statistically significant improvement in BCFI: 22% reduction in BCFI event; 3.6% absolute improvement - Statistically significant reduction in DR: 28% reduction in DR; 2.8% absolute improvement - No significant difference in overall survival: Trend in reduction of breast cancer deaths - No significant increase in osteoporotic fractures or arterial thrombotic events Eleftherios P. Mamounas [Slide 2] NRG Oncology NSABP B-42 - Postmenopausal pts with ER+ or PR+ breast cancer - Stage I, II, or IIIa invasive BC at diagnosis - Disease-free after 5 years of endocrine therapy AI x 5 yrs or TAM x <= 3 yrs -> AI to 5 yrs Stratification: Pathological nodal status (Negative, Positive) Prior adjuvant TAM (Yes, No) Lowest BMD T score: spine, hip, femur (>-2.0, <=-2.0 SD) R -> Letrozole x 5 yrs | Placebo x 5 yrs Endpoints Primary: - Disease-Free Survival (DFS): - Local, regional, distant recurrence, contralateral BC, 2nd non-breast primary cancer, and death from any cause as first events Secondary: - Overall Survival (OS) - Breast Cancer-Free Interval (BCFI): - Recurrence or contralateral BC as first event - Distant Recurrence (DR) - Osteoporotic Fractures (OF) - Arterial Thrombotic events (AT) Eleftherios P. Mamounas [Slide 3] Disease-Free Survival [KM curve; y-axis Disease-Free Survival 0.0-1.0; x-axis Years from randomization 0-15] Curve labels: 10 yr 74.9% (letrozole) / 72.1% (placebo), Delta = 2.8%; 15 yr 61.6% (letrozole) / 56.5% (placebo), Delta = 5.1% Analysis | # Events P | L | HR (95% CI) | P-value 10-yr | 505 | 435 | 0.85 (0.74-0.96) | 0.01 15-yr | 787 | 691 | 0.85 (0.76-0.94) | 0.002 Number at risk (0, 5, 10, 15 yr): Placebo 1954, 1559, 1125, 651; Letrozole 1952, 1601, 1171, 698 Patients excluded: - 38 (no follow-up, or not at risk for DFS) - 22 (no clinical follow-up assessment) Eleftherios P. Mamounas Survival table Endpoint: Disease-free survival (primary endpoint) Placebo | 1954 Letrozole | 1952 HR: 0.85 HR CI: 0.76-0.94 P: 0.002 Landmark rates: 10 yr: Letrozole 74.9% | Placebo 72.1% 15 yr: Letrozole 61.6% | Placebo 56.5%

[Slide 2] Cohort Characteristics - From SEP 2006 to JAN 2010, 3966 pts were randomized - Follow-up of trial patients was completed in AUG 2026 - Median follow-up of 15.9 yrs - Since the 10-yr update, 538 DFS events were recorded. - No significant differences in the distribution of patient, tumor, and prior treatment characteristics between P and L: - 34% <60 years of age; 93% white; 4% black - 57% node-negative; 78% HER-2 negative (8% unknown) - 25% had lowest BMD score <= -2.0 - 39% received prior tamoxifen - 61% had breast-conserving surgery Eleftherios P. Mamounas [Slide 3] Previously Reported 10-Year Results - Extended letrozole significantly improved: - Disease-free Survival: HR: 0.85 (0.74-0.96), p=0.01, Abs. Diff: 3.3% - Breast Cancer-free Interval: HR: 0.75 (0.62-0.91), p=0.003, Abs. Diff: 2.7% - Distant recurrence: HR: 0.72 (0.55-0.92), p=0.01, Abs. Diff: 1.8% - There was no significant improvement in overall survival. - Extended letrozole did not significantly increase risk of osteoporotic fractures or arterial thrombotic events. - Here we report the final trial analysis of 15-yr outcome data. Mamounas, EP et al. J Natl Cancer Inst. 2023;115(11):1302-1309. Eleftherios P. Mamounas

[Slide 2] Distant Recurrence by Nodal Status Node-Negative # Events: Analysis 15-yr | P 54 | L 42 | HR (95% CI) 0.78 (0.52-1.16) | P-value 0.22 Curve labels: Placebo 4.0% (10 yr), 5.6% (15 yr); Letrozole 3.3% (10 yr), 4.0% (15 yr); Δ = 1.6% At risk (years 0, 5, 10, 15): Placebo 1116, 975, 762, 475; Letrozole 1126, 979, 776, 491 Node-Positive # Events: Analysis 15-yr | P 115 | L 81 | HR (95% CI) 0.69 (0.52-0.92) | P-value 0.011 Curve labels: Placebo 12.1% (10 yr), 15.1% (15 yr); Letrozole 8.9% (10 yr), 10.7% (15 yr); Δ = 4.4% At risk (years 0, 5, 10, 15): Placebo 838, 694, 524, 314; Letrozole 826, 699, 531, 329 [Cumulative incidence plots; y-axis: Cumulative Incidence of DR 0.00-0.20; x-axis: Years after Random Assignment] Patients excluded: • 38 (no follow-up, or not at risk for DFS) • 22 (no clinical follow-up assessment) [NRG Oncology logo] Eleftherios P. Mamounas Survival table Endpoint: Distant recurrence (cumulative incidence) by nodal status, 15-year analysis Node-negative: Placebo | 1116 Node-negative: Letrozole | 1126 Node-positive: Placebo | 838 Node-positive: Letrozole | 826 HR: Node-negative 0.78; Node-positive 0.69 HR CI: Node-negative 0.52-1.16; Node-positive 0.52-0.92 P: Node-negative 0.22; Node-positive 0.011 Landmark rates: 10 yr: Node-negative: Placebo 4.0% | Node-negative: Letrozole 3.3% | Node-positive: Placebo 12.1% | Node-positive: Letrozole 8.9% 15 yr: Node-negative: Placebo 5.6% | Node-negative: Letrozole 4.0% | Node-positive: Placebo 15.1% | Node-positive: Letrozole 10.7% [Slide 3] Cumulative Incidence of Death by Cause Breast Cancer # Events Analysis | P | L | HR (95% CI) | P-value 15-yr | 114 | 91 | 0.78 (0.59-1.03) | 0.08 Curve labels at 15 yr: 6.5% [placebo, blue]; 5.0% [letrozole, red] Y-axis: Cumulative Incidence of breast cancer death X-axis: Years after Random Assignment Non-Breast Cancer/Unknown # Events Analysis | P | L | HR (95% CI) | P-value 15-yr | 411 | 401 | 0.96 (0.84-1.11) | 0.60 Curve labels at 15 yr: 21.2% [placebo, blue]; 21.0% [letrozole, red] Y-axis: Cumulative Incidence of death from other causes X-axis: Years after Random Assignment Number at risk (both panels; years 0 / 5 / 10 / 15): Placebo: 1965 / 1736 / 1405 / 924 Letrozole: 1963 / 1725 / 1399 / 980 Patients excluded: • 38 (no follow-up, or not at risk for DFS) [NRG Oncology logo] Eleftherios P. Mamounas Survival table Endpoint: Cumulative incidence of death by cause (breast cancer death; non-breast cancer/unknown death), 15-yr analysis Placebo | 1965 Letrozole | 1963 HR: 0.78 (breast cancer death); 0.96 (non-breast cancer/unknown death) HR CI: 0.59-1.03 (breast cancer); 0.84-1.11 (non-breast cancer/unknown) P: 0.08 (breast cancer); 0.60 (non-breast cancer/unknown) Landmark rates: 15 yr: Placebo 6.5% breast cancer death | Letrozole 5.0% breast cancer death | Placebo 21.2% non-breast cancer/unknown death | Letrozole 21.0% non-breast cancer/unknown death

Overall Survival [KM curves; y-axis: Overall Survival; x-axis: Years after Random Assignment] Curve labels: 85.4% / 85.2% (10 yr); 74.0% / 72.3% (15 yr) # Events Analysis | P | L | HR (95% CI) | P-value 10-yr | 272 | 262 | 0.97 (0.82-1.15) | 0.74 15-yr | 525 | 492 | 0.92 (0.82-1.05) | 0.21 Patients excluded: • 38 (no follow-up, or not at risk for DFS) No. at risk (0, 5, 10, 15 yr): Placebo: 1965, 1736, 1405, 924 Letrozole: 1963, 1725, 1399, 980 [NRG Oncology logo] Eleftherios P. Mamounas Survival table Endpoint: Overall survival Placebo | 1965 Letrozole | 1963 HR: 10-yr 0.97; 15-yr 0.92 HR CI: 10-yr 0.82-1.15; 15-yr 0.82-1.05 P: 10-yr 0.74; 15-yr 0.21 Landmark rates: 10 yr: Placebo 85.4% | Letrozole 85.2% 15 yr: Letrozole 74.0% | Placebo 72.3%

[Slide 3] Disease-Free Survival Landmarks: 10 yr — Letrozole 74.9%, Placebo 72.1%, Δ = 2.8%; 15 yr — Letrozole 61.6%, Placebo 56.5%, Δ = 5.1% Analysis | # Events P | L | HR (95% CI) | P-value 10-yr | 505 | 435 | 0.85 (0.74-0.96) | 0.01 15-yr | 787 | 691 | 0.85 (0.76-0.94) | 0.002 Y: Disease-Free Survival; X: Years from randomization 0,5,10,15 Number at risk: Placebo 1954 1559 1125 651; Letrozole 1952 1601 1171 698 Patients excluded: • 38 (no follow-up, or not at risk for DFS) • 22 (no clinical follow-up assessment) [NRG Oncology logo] Eleftherios P. Mamounas Survival table Endpoint: Disease-free survival (NSABP B-42, letrozole vs placebo) Placebo | 1954 Letrozole | 1952 HR: 0.85 (15-yr analysis) HR CI: 0.76-0.94 P: 0.002 Landmark rates: 10 yr: Letrozole 74.9% | Placebo 72.1% 15 yr: Letrozole 61.6% | Placebo 56.5% [Slide 4] DFS First Events by Treatment First Event | Placebo (n=1954) # | % | Letrozole (n=1952) # | % Distant Recurrence | 137 | 7.0 | 100 | 5.1 Local Recurrence | 70 | 3.6 | 63 | 3.2 Second Primary Cancer | 332 | 17.0 | 282 | 14.4 Breast | 117 | 6.0 | 98 | 5.0 Non-Breast | 215 | 11.0 | 184 | 9.4 Death | 248 | 12.7 | 246 | 12.6 Total First Event | 787 | 40.3 | 691 | 35.4 [NRG Oncology logo] Eleftherios P. Mamounas

[Slide 2] Breast Cancer-Free Interval [Cumulative incidence curves; y-axis: Cumulative Incidence of BCFI; x-axis: Years after Random Assignment] Curve labels: 13.5% / 10.9%, Δ = 2.6% (10 yr); 18.3% / 14.7%, Δ = 3.6% (15 yr) # Events Analysis | P | L | HR (95% CI) | P-value 10-yr | 246 | 189 | 0.75 (0.62-0.91) | 0.003 15-yr | 326 | 264 | 0.78 (0.67-0.92) | 0.003 Patients excluded: • 38 (no follow-up, or not at risk for DFS) • 22 (no clinical follow-up assessment) No. at risk (0, 5, 10, 15 yr): Placebo: 1954, 1559, 1125, 651 Letrozole: 1952, 1601, 1171, 698 [NRG Oncology logo] Eleftherios P. Mamounas Survival table Endpoint: Breast cancer-free interval (cumulative incidence of BCFI events) Placebo | 1954 Letrozole | 1952 HR: 10-yr 0.75; 15-yr 0.78 HR CI: 10-yr 0.62-0.91; 15-yr 0.67-0.92 P: 10-yr 0.003; 15-yr 0.003 Landmark rates: 10 yr: Placebo 13.5% | Letrozole 10.9% 15 yr: Placebo 18.3% | Letrozole 14.7% [Slide 3] Distant Recurrence [Cumulative incidence curves; y-axis: Cumulative Incidence of DR; x-axis: Years after Random Assignment] # Events Analysis | P | L | HR (95% CI) | P-value 10-yr | 139 | 101 | 0.72 (0.55-0.92) | 0.01 15-yr | 169 | 123 | 0.72 (0.57-0.91) | 0.005 Curve labels: 7.4% / 5.7%, Δ = 1.7% (10 yr); 9.7% / 6.9%, Δ = 2.8% (15 yr) Patients excluded: • 38 (no follow-up, or not at risk for DFS) • 22 (no clinical follow-up assessment) No. at risk (0, 5, 10, 15 yr): Placebo: 1954, 1669, 1286, 789 Letrozole: 1952, 1678, 1307, 820 [NRG Oncology logo] Eleftherios P. Mamounas Survival table Endpoint: Distant recurrence (cumulative incidence) Placebo | 1954 Letrozole | 1952 HR: 10-yr 0.72; 15-yr 0.72 HR CI: 10-yr 0.55-0.92; 15-yr 0.57-0.91 P: 10-yr 0.01; 15-yr 0.005 Landmark rates: 10 yr: Placebo 7.4% | Letrozole 5.7% 15 yr: Placebo 9.7% | Letrozole 6.9%

[Slide 3] A | Abemaciclib + Fulvestrant (n=182) | Placebo + Fulvestrant (n=186) Events | 90 | 110 Median (95% CI), months | 32.1 (27.2-38.1) | 24.6 (22.5-28.8) HR (95% CI); log-rank P | 0.80 (0.60-1.06); 0.11 OS (%) / Time (months) Number at Risk Abemaciclib + Fulvestrant | 182 | 168 | 151 | 139 | 124 | 114 | 107 | 99 | 91 | 85 | 79 | 67 | 42 | 30 | 17 | 1 Placebo + Fulvestrant | 186 | 176 | 162 | 153 | 143 | 128 | 109 | 102 | 86 | 74 | 66 | 56 | 37 | 21 | 9 | 1 B | Abemaciclib + Fulvestrant (n=182) | Placebo + Fulvestrant (n=186) Events | 145 | 156 Median (95% CI), months | 5.9 (5.5-8.6) | 5.3 (3.7-5.6) HR (95% CI); log-rank P | 0.78 (0.62-0.99); 0.039 Curve annotations at 6 months: 49% (abemaciclib) / 37% (placebo) PFS (%) / Time (months) Number at Risk Abemaciclib + Fulvestrant | 182 | 124 | 79 | 64 | 40 | 35 | 28 | 26 | 19 | 16 | 15 | 13 | 6 | 2 | 2 | 0 Placebo + Fulvestrant | 186 | 114 | 61 | 49 | 33 | 23 | 18 | 12 | 10 | 9 | 8 | 7 | 3 | 1 | 0 | 0 Survival table Endpoint: A: Overall survival; B: Progression-free survival (updated) Arm | N | Median | Median CI Abemaciclib + Fulvestrant (OS) | 182 | 32.1 months | 27.2-38.1 Placebo + Fulvestrant (OS) | 186 | 24.6 months | 22.5-28.8 Abemaciclib + Fulvestrant (PFS) | 182 | 5.9 months | 5.5-8.6 Placebo + Fulvestrant (PFS) | 186 | 5.3 months | 3.7-5.6 HR: OS 0.80; PFS 0.78 HR CI: OS 0.60-1.06; PFS 0.62-0.99 P: OS 0.11; PFS 0.039 (log-rank) Landmark rates: 6 months: Abemaciclib + Fulvestrant (PFS) 49% | Placebo + Fulvestrant (PFS) 37% [Slide 4] Phase III postMONARCH: Conclusions • Overall survival numerically favored abemaciclib plus fulvestrant following disease progression on prior AI and a CDK4/6i; median 32.1 vs 24.6 months (Δ 7.5 months; HR 0.80, 95% CI 0.60-1.06); the difference did not reach statistical significance • OS was consistent across prespecified subgroups, including prior CDK4/6i and duration of prior CDK4/6i • Subsequent anti-cancer therapy was balanced across arms • The OS trend is directionally consistent with the PFS outcomes; no new safety signals were identified • Together these data continue to support the consideration of abemaciclib-based treatments after progression on a prior CDK4/6i Kevin Kalinsky, MD, MS, FASCO

[Slide 1] First Subsequent Post-Discontinuation Therapy Characteristic, n (%) | Abemaciclib + Fulvestrant n = 182 | Placebo + Fulvestrant n = 186 Discontinueda | 171 (94) | 178 (96) Any therapy | 132 (77) | 147 (83) Chemotherapyb | 88 (51) | 96 (54) ADC | 1 (1) | 6 (3) Endocrine therapyb | 40 (23) | 51 (29) Targeted agent therapyb | 36 (21) | 41 (23) CDK4/6i | 1 (1) | 6 (3) PI3K/AKT/mTOR inhibitor | 30 (18) | 31 (17) Immunotherapyb,c | 3 (2) | 2 (1) aPercentages out of randomized participants. All other percentages are out of participants who discontinued study treatment; bSome participants were counted more than once due to receiving combinations of listed agents; cIncludes 3 participants who received blinded pembrolizumab or placebo (treatment assignment not disclosed). ADC, antibody-drug conjugate; CDK4/6i, CDK4/6 inhibitor; PI3K, phosphoinositide 3-kinase; mTOR, mammalian target of rapamycin Data cut: 06 March 2026 No meaningful imbalance in first subsequent therapy Kevin Kalinsky, MD, MS, FASCO [Slide 2] Updated PFS was consistent with the primary analysis Investigator- PFS (Primary endpoint) | Abemaciclib + Fulvestrant (n=182) | Placebo + Fulvestrant (n=186) Events | 145 | 156 Median (95% CI), months | 5.9 (5.5-8.6) | 5.3 (3.7-5.6) HR (95% CI); log-rank P | 0.78 (0.62-0.99); 0.039 Curve labels: 6 mo 49% vs 37%; 12 mo 28% vs 22%; 24 mo 14% vs 8% Number at Risk | 0 | 3 | 6 | 9 | 12 | 15 | 18 | 21 | 24 | 27 | 30 | 33 | 36 | 39 | 42 | 45 Abemaciclib + Fulvestrant | 182 | 124 | 79 | 64 | 40 | 35 | 28 | 26 | 19 | 16 | 15 | 13 | 6 | 2 | 2 | 0 Placebo + Fulvestrant | 186 | 114 | 61 | 49 | 33 | 23 | 18 | 12 | 10 | 9 | 8 | 7 | 3 | 1 | 0 | 0 BICR-PFS (Secondary endpoint) | Abemaciclib + Fulvestrant (n=182) | Placebo + Fulvestrant (n=186) Events | 69 | 100 Median (95% CI), months | 13.8 (9.5-NR) | 5.6 (4.6-9.2) HR (95% CI); log-rank P | 0.60 (0.44-0.83); 0.0019 Estimates impacted by informative censoring Curve labels: 6 mo 69% vs 46%; 12 mo 53% vs 39%; 24 mo 40% vs 26% Number at Risk | 0 | 3 | 6 | 9 | 12 | 15 | 18 | 21 | 24 | 27 | 30 | 33 | 36 | 39 | 42 | 45 Abemaciclib + Fulvestrant | 182 | 118 | 75 | 62 | 39 | 30 | 24 | 22 | 17 | 15 | 13 | 10 | 3 | 1 | 1 | 0 Placebo + Fulvestrant | 186 | 105 | 53 | 46 | 32 | 22 | 16 | 10 | 9 | 7 | 6 | 6 | 2 | 1 | 0 | 0 Data cut: 06 March 2026 Concordant treatment effect across Investigator-PFS and BICR-PFS Kevin Kalinsky, MD, MS, FASCO Survival table Endpoint: Progression-free survival (Investigator-assessed, primary; BICR-assessed, secondary) - updated analysis Arm | N | Median | Median CI Abemaciclib + Fulvestrant (INV-PFS) | 182 | 5.9 months | 5.5-8.6 Placebo + Fulvestrant (INV-PFS) | 186 | 5.3 months | 3.7-5.6 Abemaciclib + Fulvestrant (BICR-PFS) | 182 | 13.8 months | 9.5-NR Placebo + Fulvestrant (BICR-PFS) | 186 | 5.6 months | 4.6-9.2 HR: INV-PFS 0.78; BICR-PFS 0.60 HR CI: INV-PFS 0.62-0.99; BICR-PFS 0.44-0.83 P: INV-PFS log-rank P 0.039; BICR-PFS log-rank P 0.0019 Landmark rates: 6 months: Abemaciclib + Fulvestrant (INV-PFS) 49% | Placebo + Fulvestrant (INV-PFS) 37% | Abemaciclib + Fulvestrant (BICR-PFS) 69% | Placebo + Fulvestrant (BICR-PFS) 46% 12 months: Abemaciclib + Fulvestrant (INV-PFS) 28% | Placebo + Fulvestrant (INV-PFS) 22% | Abemaciclib + Fulvestrant (BICR-PFS) 53% | Placebo + Fulvestrant (BICR-PFS) 39% 24 months: Abemaciclib + Fulvestrant (INV-PFS) 14% | Placebo + Fulvestrant (INV-PFS) 8% | Abemaciclib + Fulvestrant (BICR-PFS) 40% | Placebo + Fulvestrant (BICR-PFS) 26% [Slide 3] Consistent Overall Survival Effect Across Subgroups Subgroup | Abemaciclib Events/N | Placebo Events/N | HR (95% CI) | Interaction p-value Overall | 90/182 | 110/186 | 0.796 (0.600, 1.057) | Age, years | | | | 0.163 <65 | 64/126 | 65/118 | 0.936 (0.662, 1.322) | ≥65 | 26/56 | 45/68 | 0.613 (0.378, 0.995) | Region | | | | 0.416 Other | 67/133 | 84/134 | 0.719 (0.521, 0.993) | United States | 14/28 | 14/28 | 1.198 (0.571, 2.513) | East Asia | 9/21 | 12/24 | 0.973 (0.410, 2.310) | Measurable disease | | | | 0.494 Yes | 66/131 | 75/127 | 0.858 (0.616, 1.195) | No | 24/51 | 35/59 | 0.692 (0.411, 1.163) | Visceral metastasis | | | | 0.993 Yes | 60/112 | 70/109 | 0.797 (0.564, 1.126) | No | 30/70 | 40/77 | 0.799 (0.497, 1.283) | Liver metastasis | | | | 0.963 Yes | 39/68 | 49/71 | 0.803 (0.527, 1.223) | No | 51/114 | 61/115 | 0.814 (0.561, 1.181) | Bone-only disease | | | | 0.550 Yes | 14/32 | 23/42 | 0.667 (0.343, 1.297) | No | 76/150 | 87/144 | 0.834 (0.613, 1.136) | PR status | | | | 0.753 Positive | 74/144 | 90/150 | 0.841 (0.618, 1.145) | Negative | 16/37 | 17/32 | 0.746 (0.377, 1.478) | Previous CDK4/6 inhibitor duration | | | | 0.820 ABC ≥12 months or after adjuvant CDK4/6i | 58/129 | 84/144 | 0.751 (0.537, 1.050) | ABC <12 months or during adjuvant CDK4/6i | 32/53 | 26/40 | 0.807 (0.480, 1.355) | Previous CDK4/6 inhibitor | | | | 0.196 Palbociclib | 52/107 | 64/110 | 0.809 (0.561, 1.167) | Ribociclib | 34/61 | 38/61 | 0.985 (0.620, 1.565) | Abemaciclib | 4/14 | 8/14 | 0.302 (0.091, 1.004) | Favors Abemaciclib <-> Favors Placebo (axis 0.25, 0.5, 1, 2) Kevin Kalinsky, MD, MS, FASCO ABC, advanced breast cancer; CDK4/6i, cyclin-dependent kinase 4/6 inhibitor Data cut: 06 March 2026 [Slide 4] Final Overall Survival Numerically Favored Abemaciclib | Abemaciclib + Fulvestrant (n=182) | Placebo + Fulvestrant (n=186) Events | 90 | 110 Median (95% CI), months | 32.1 (27.2-38.1) | 24.6 (22.5-28.8) HR (95% CI); log-rank P | 0.80 (0.60-1.06); 0.11 Curve labels: 32.1 months (Δ = 7.5); 24.6 months Overall Survival (%) vs Time (months) 0-45 Number at Risk | 0 | 3 | 6 | 9 | 12 | 15 | 18 | 21 | 24 | 27 | 30 | 33 | 36 | 39 | 42 | 45 Abemaciclib + Fulvestrant | 182 | 168 | 151 | 139 | 124 | 114 | 107 | 99 | 91 | 85 | 79 | 67 | 42 | 30 | 17 | 1 Placebo + Fulvestrant | 186 | 176 | 162 | 153 | 143 | 128 | 109 | 102 | 86 | 74 | 66 | 56 | 37 | 21 | 9 | 1 Data cut: 06 March 2026 Abemaciclib + fulvestrant led to numerical 20% reduction in the risk of death Kevin Kalinsky, MD, MS, FASCO Survival table Endpoint: Overall survival (final analysis) Arm | N | Median | Median CI Abemaciclib + Fulvestrant | 182 | 32.1 months | 27.2-38.1 Placebo + Fulvestrant | 186 | 24.6 months | 22.5-28.8 HR: 0.80 HR CI: 0.60-1.06 P: 0.11 (log-rank)

Phase III postMONARCH: Study Design and Background Phase III postMONARCH Study (NCT05169567) Eligibility: HR+, HER2- ABC; Men & Pre/postmenopausal women; Prior Therapy: • ABC: Disease progression on CDK4/6i + AI as initial therapy • Adjuvant: Disease recurrence on/after CDK4/6i + ET • No other therapy for ABC Randomization 1:1, N = 368 -> Abemaciclib + Fulvestrant | Placebo + Fulvestrant Endpoints — Primary: Investigator-assessed PFS; Secondary: OS, BICR-assessed PFS; Exploratory: PFS2, TTC Primary analysis previously reported(1): OS data were immature Prespecified final OS analysis: 200 OS events; Median follow-up time: 37 months • CDK4/6i + ET are standard first-line treatment of HR+, HER2- ABC(2) • After first-line treatment, safe and effective therapies that delay infusional treatment are an important need • The postMONARCH trial previously demonstrated that abemaciclib + fulvestrant significantly improved PFS vs placebo + fulvestrant after progression on CDK4/6i + ET (HR: 0.73 [95% CI 0.57-0.95], P = 0.017)(1) • We now report results of the final OS analysis in addition to secondary and exploratory endpoints HR, hormone receptor; HER2, human epidermal growth factor receptor 2; ABC, advanced breast cancer; CDK4/6i, cyclin-dependent kinase 4/6 inhibitor; AI, aromatase inhibitor; ET, endocrine therapy; OS, overall survival; PFS, progression-free survival 1. Kalinsky et al. 2024 JCO 2. Gradishar et al. 2025 J Natl Compr Canc Netw Kevin Kalinsky, MD, MS, FASCO

[Slide 3] Updated PFS was consistent with the primary analysis Investigator- PFS (Primary endpoint) | Abemaciclib + Fulvestrant (n=182) | Placebo + Fulvestrant (n=186) Events | 145 | 156 Median (95% CI), months | 5.9 (5.5-8.6) | 5.3 (3.7-5.6) HR (95% CI); | 0.78 (0.62-0.99); log-rank P | 0.039 Curve labels: 49% / 37%; 28% / 22%; 14% / 8% Number at Risk Abemaciclib + Fulvestrant: 182 124 79 64 40 35 28 26 19 16 15 13 6 2 2 0 Placebo + Fulvestrant: 186 114 61 49 33 23 18 12 10 9 8 7 3 1 0 0 BICR-PFS (Secondary endpoint) | Abemaciclib + Fulvestrant (n=182) | Placebo + Fulvestrant (n=186) Events | 69 | 100 Median (95% CI), months | 13.8 (9.5-NR) | 5.6 (4.6-9.2) HR (95% CI); | 0.60 (0.44-0.83); log-rank P | 0.0019 Estimates impacted by informative censoring Curve labels: 69% / 46%; 53% / 39%; 40% / 26% Number at Risk Abemaciclib + Fulvestrant: 182 118 75 62 39 30 24 22 17 15 13 10 3 1 1 0 Placebo + Fulvestrant: 186 105 53 46 32 22 16 10 9 7 6 6 2 1 0 0 Time (months) 0-45 Data cut: 06 March 2026 Concordant treatment effect across Investigator-PFS and BICR-PFS Kevin Kalinsky, MD, MS, FASCO Survival table Endpoint: Investigator-assessed PFS (primary); BICR-PFS (secondary) reported separately in landmarks/notes Arm | N | Median | Median CI Abemaciclib + Fulvestrant (INV-PFS) | 182 | 5.9 | 5.5-8.6 Placebo + Fulvestrant (INV-PFS) | 186 | 5.3 | 3.7-5.6 Abemaciclib + Fulvestrant (BICR-PFS) | 182 | 13.8 | 9.5-NR Placebo + Fulvestrant (BICR-PFS) | 186 | 5.6 | 4.6-9.2 HR: 0.78 (INV-PFS); 0.60 (BICR-PFS) HR CI: 0.62-0.99 (INV-PFS); 0.44-0.83 (BICR-PFS) P: 0.039 (INV-PFS, log-rank); 0.0019 (BICR-PFS, log-rank) Landmark rates: 12 months (INV-PFS): Abemaciclib + Fulvestrant 28% | Placebo + Fulvestrant 22% 24 months (INV-PFS): Abemaciclib + Fulvestrant 14% | Placebo + Fulvestrant 8% 6 months (BICR-PFS): Abemaciclib + Fulvestrant 69% | Placebo + Fulvestrant 46% 12 months (BICR-PFS): Abemaciclib + Fulvestrant 53% | Placebo + Fulvestrant 39% 24 months (BICR-PFS): Abemaciclib + Fulvestrant 40% | Placebo + Fulvestrant 26% [Slide 4] PFS2 and Time to Chemotherapy (TTC) PFS2 (Exploratory endpoint) | Abemaciclib + Fulvestrant (n=182) | Placebo + Fulvestrant (n=186) Events | 135 | 151 Median (95% CI), months | 13.9 (11.8-17.2) | 12.3 (10.7-13.8) HR (95% CI); | 0.76 (0.59-0.97); log-rank P | 0.029 Time to Second Progression (%) Number at Risk Abemaciclib + Fulvestrant: 182 164 137 117 86 72 56 44 36 30 30 19 9 2 2 0 Placebo + Fulvestrant: 186 167 146 114 84 59 43 29 21 15 11 8 4 1 0 0 TTC (Exploratory endpoint) | Abemaciclib + Fulvestrant (n=182) | Placebo + Fulvestrant (n=186) Events | 111 | 132 Median (95% CI), months | 13.6 (9.7-17.2) | 9.7 (7.7-11.5) HR (95% CI); | 0.76 (0.58-0.99); log-rank P | 0.039 Time to Chemotherapy (%) Number at Risk Abemaciclib + Fulvestrant: 182 149 113 89 72 61 46 42 38 31 29 26 14 10 5 0 Placebo + Fulvestrant: 186 146 110 80 62 48 37 30 26 21 17 15 9 5 3 0 Time (months) 0-45 Data cut: 06 March 2026 PFS2 and TTC endpoints favored abemaciclib + fulvestrant PFS2, Time from randomization to the discontinuation date of next-line (first line of post-discontinuation treatment), or starting date of the 2nd line of post-discontinuation treatment, or death from any cause, whichever is earliest; TTC, Time from randomization to the initiation of first-post-discontinuation chemotherapy, including ADC. Kevin Kalinsky, MD, MS, FASCO Survival table Endpoint: PFS2 (exploratory); TTC (exploratory) on same slide Arm | N | Median | Median CI Abemaciclib + Fulvestrant (PFS2) | 182 | 13.9 | 11.8-17.2 Placebo + Fulvestrant (PFS2) | 186 | 12.3 | 10.7-13.8 Abemaciclib + Fulvestrant (TTC) | 182 | 13.6 | 9.7-17.2 Placebo + Fulvestrant (TTC) | 186 | 9.7 | 7.7-11.5 HR: 0.76 (PFS2); 0.76 (TTC) HR CI: 0.59-0.97 (PFS2); 0.58-0.99 (TTC) P: 0.029 (PFS2, log-rank); 0.039 (TTC, log-rank)

A OS (%) Abemaciclib + Fulvestrant (n=182) | Placebo + Fulvestrant (n=186) Events | 90 | 110 Median (95% CI), months | 32.1 (27.2-38.1) | 24.6 (22.5-28.8) HR (95% CI); log-rank P | 0.80 (0.60-1.06); 0.11 Time (months) 0 | 3 | 6 | 9 | 12 | 15 | 18 | 21 | 24 | 27 | 30 | 33 | 36 | 39 | 42 | 45 Number at Risk Abemaciclib + Fulvestrant | 182 | 168 | 151 | 139 | 124 | 114 | 107 | 99 | 91 | 85 | 79 | 67 | 42 | 30 | 17 | 1 Placebo + Fulvestrant | 186 | 176 | 162 | 153 | 143 | 128 | 109 | 102 | 86 | 74 | 66 | 56 | 37 | 21 | 9 | 1 B PFS (%) Abemaciclib + Fulvestrant (n=182) | Placebo + Fulvestrant (n=186) Events | 145 | 156 Median (95% CI), months | 5.9 (5.5-8.6) | 5.3 (3.7-5.6) HR (95% CI); log-rank P | 0.78 (0.62-0.99); 0.039 49% 37% Time (months) 0 | 3 | 6 | 9 | 12 | 15 | 18 | 21 | 24 | 27 | 30 | 33 | 36 | 39 | 42 | 45 Number at Risk Abemaciclib + Fulvestrant | 182 | 124 | 79 | 64 | 40 | 35 | 28 | 26 | 19 | 16 | 15 | 13 | 6 | 2 | 2 | 0 Placebo + Fulvestrant | 186 | 114 | 61 | 49 | 33 | 23 | 18 | 12 | 10 | 9 | 8 | 7 | 3 | 1 | 0 | 0 Survival table Endpoint: Overall survival (panel A); PFS (panel B) Arm | N | Median | Median CI Abemaciclib + Fulvestrant (OS) | 182 | 32.1 months | 27.2-38.1 Placebo + Fulvestrant (OS) | 186 | 24.6 months | 22.5-28.8 Abemaciclib + Fulvestrant (PFS) | 182 | 5.9 months | 5.5-8.6 Placebo + Fulvestrant (PFS) | 186 | 5.3 months | 3.7-5.6 HR: OS 0.80; PFS 0.78 HR CI: OS 0.60-1.06; PFS 0.62-0.99 P: OS log-rank P 0.11; PFS log-rank P 0.039 Landmark rates: 6 months (PFS): Abemaciclib + Fulvestrant 49% | Placebo + Fulvestrant 37%

[Slide 1] 10:15 - 11:45 Rapid oral 1: Breast cancer, advanced stage CHAIRS: VESNA BJELIC-RADISIC, GÉRALDINE GEBHART, WILLIAM GRADISHAR Final Overall Survival Numerically Favored Abemaciclib | Abemaciclib + Fulvestrant (n=182) | Placebo + Fulvestrant (n=186) Events | 90 | 110 Median (95% CI), months | 32.1 (27.2-38.1) | 24.6 (22.5-28.8) HR (95% CI); log-rank P | 0.80 (0.60-1.06); 0.11 Curve labels: 32.1 months (Δ = 7.5); 24.6 months Overall Survival (%) vs Time (months) 0-45 Number at Risk | 0 | 3 | 6 | 9 | 12 | 15 | 18 | 21 | 24 | 27 | 30 | 33 | 36 | 39 | 42 | 45 Abemaciclib + Fulvestrant | 182 | 168 | 151 | 139 | 124 | 114 | 107 | 99 | 91 | 85 | 79 | 67 | 42 | 30 | 17 | 1 Placebo + Fulvestrant | 186 | 176 | 162 | 153 | 143 | 128 | 109 | 102 | 86 | 74 | 66 | 56 | 37 | 21 | 9 | 1 Data cut: 06 March 2026 Abemaciclib + fulvestrant led to numerical 20% reduction in the risk of death Kevin Kalinsky, MD, MS, FASCO Kevin Kalinsky Overall survival from the phase 3 postMONARCH trial of abemaciclib + fulvestrant vs placebo + fulvestrant for HR+, HER2- advanced breast cancer following progression on a prior CDK4/6 inhibitor + endocrine therapy Survival table Endpoint: Overall survival (final analysis) Arm | N | Median | Median CI Abemaciclib + Fulvestrant | 182 | 32.1 months | 27.2-38.1 Placebo + Fulvestrant | 186 | 24.6 months | 22.5-28.8 HR: 0.80 HR CI: 0.60-1.06 P: 0.11 (log-rank) [Slide 2] 10:15 - 11:45 Rapid oral 1: Breast cancer, advanced stage CHAIRS: VESNA BJELIC-RADISIC, GÉRALDINE GEBHART, WILLIAM GRADISHAR Updated PFS was consistent with the primary analysis Investigator- PFS (Primary endpoint) | Abemaciclib + Fulvestrant (n=182) | Placebo + Fulvestrant (n=186) Events | 145 | 156 Median (95% CI), months | 5.9 (5.5-8.6) | 5.3 (3.7-5.6) HR (95% CI); log-rank P | 0.78 (0.62-0.99); 0.039 Curve labels: 49% / 37% (6 mo); 28% / 22% (12 mo); 14% / 8% (24 mo) Progression-Free Survival (%) ; Time (months) 0-45 Number at Risk Abema: 182 124 79 64 40 35 28 26 19 16 15 13 6 2 2 0 Placebo: 186 114 61 49 33 23 18 12 10 9 8 7 3 1 0 0 BICR-PFS (Secondary endpoint) | Abemaciclib + Fulvestrant (n=182) | Placebo + Fulvestrant (n=186) Events | 69 | 100 Median (95% CI), months | 13.8 (9.5-NR) | 5.6 (4.6-9.2) HR (95% CI); log-rank P | 0.60 (0.44-0.83); 0.0019 Estimates impacted by informative censoring Curve labels: 69% / 46% (6 mo); 53% / 39% (12 mo); 40% / 26% (24 mo) Number at Risk Abema: 182 118 75 62 39 30 24 22 17 15 13 10 3 1 1 0 Placebo: 186 105 53 46 32 22 16 10 9 7 6 6 2 1 0 0 Data cut: 06 March 2026 Concordant treatment effect across Investigator-PFS and BICR-PFS Kevin Kalinsky, MD, MS, FASCO Survival table Endpoint: Investigator-assessed PFS (primary); BICR-PFS (secondary) also shown Arm | N | Median | Median CI Abemaciclib + Fulvestrant | 182 | 5.9 | 5.5-8.6 Placebo + Fulvestrant | 186 | 5.3 | 3.7-5.6 HR: 0.78 HR CI: 0.62-0.99 P: 0.039 (log-rank) Landmark rates: 6 mo: Abemaciclib + Fulvestrant 49% | Placebo + Fulvestrant 37% 12 mo: Abemaciclib + Fulvestrant 28% | Placebo + Fulvestrant 22% 24 mo: Abemaciclib + Fulvestrant 14% | Placebo + Fulvestrant 8%

ESMO 2026 ORAL PRESENTATIONS FPN: 1RO RAPID ORAL Overall survival from the Phase 3 postMONARCH trial of abemaciclib + fulvestrant vs placebo + fulvestrant for HR+, HER2- advanced breast cancer following progression on a prior CDK4/6 inhibitor + endocrine therapy Rapid oral 1: Breast cancer, advanced stage Presenter: Kevin Kalinsky October 24 Almeria Auditorium - Hall 9 10:59 – 11:04 AM Intended for HCPs only [Lilly logo] A MEDICINE COMPANY

10:15 - 11:45 Rapid oral 1: Breast cancer, advanced stage CHAIRS: VESNA BJELIC-RADISIC, GÉRALDINE GEBHART, WILLIAM GRADISHAR Phase III postMONARCH: Conclusions Overall survival numerically favored abemaciclib plus fulvestrant following disease progression on prior AI and a CDK4/6i; median 32.1 vs 24.6 months (Δ 7.5 months; HR 0.80, 95% CI 0.60-1.06); the difference did not reach statistical significance OS was consistent across prespecified subgroups, including prior CDK4/6i and duration of prior CDK4/6i Subsequent anti-cancer therapy was balanced across arms The OS trend is directionally consistent with the PFS outcomes; no new safety signals were identified Together these data continue to support the consideration of abemaciclib-based treatments after progression on a prior CDK4/6i Kevin Kalinsky, MD, MS, FASCO Kevin Kalinsky Overall survival from the phase 3 postMONARCH trial of abemaciclib + fulvestrant vs placebo + fulvestrant for HR+, HER2- advanced breast cancer following progression on a prior CDK4/6 inhibitor + endocrine therapy ALMERIA AUDITORIUM - HALL 9 Verzenio Combination Linked to Longer Overall Survival in Advanced Breast Cancer ESMO / BREASTCANCER·ORG

[Slide 1] Background [Schematic of biparatopic antibody: Anti-Her2 ECD2 (VH/VL), Anti-Her2 ECD4 (VL/VH), Cleavable linker, Topo1 inhibitor] DAR 5.8 ~ 6.0 - HER2-negative breast cancers account for approximately 80% of breast cancers, with HER2-low tumors (IHC 1+ or IHC 2+/ISH-) comprising nearly 60% and occurring predominantly in HR-positive disease.1 - Rolditamig Deuderuxtecan (TQB2102) is a first-in-class biparatopic ADC targeting both ECD2 and ECD4 of HER2, with bispecific binding that enhances receptor cross-linking and internalization, indicating its superior potential against HER2-low malignancies.2 - TQB2102 has shown promising efficacy in heavily treated patients with HER2-low mBC in phase 1 and phase 1b trials, with an ORR of 58.3% at the 7.5 mg/kg dose.3 Here, we present the primary results from APEX-Breast01, a randomized phase 3 study evaluating the efficacy and safety of TQB2102 versus chemotherapy in patients with HER2-low mBC who had not received prior chemotherapy for metastatic disease. Abbreviations: HER2, human epidermal growth factor receptor 2; IHC, immunohistochemistry; ISH, in situ hybridization; HR, hormone receptor; ADC, antibody-drug conjugate; mBC, metastatic breast cancer; ORR, objective response rate. 1. Tarantino P, et al. J Clin Oncol, 2020. 2. D. Ruan, et al. Ann Oncol, 2026. 3. S. Wang, et al. 2025 ASCO Abstract #1090. Presenter: Shusen Wang, MD [Slide 2] APEX-Breast01 Study Design Randomized, open-label, multi-center, active-control, phase 3 trial (NCT06561607) Key Eligibility - HER2-low (IHC 1+ or IHC 2+/ISH-) r/m BC - No prior chemotherapy in r/m BC setting - HR+/HER2-low: ≥1 line of ET for r/m BC or progression within 24 months after initiation of (neo)adjuvant ET - HR-/HER2-low: Patients with known PD-L1 CPS ≥1 were excluded R 1:1 TQB2102 (N=269) 7.5 mg/kg IV, Q3W HR+/HER2-low = 461 HR-/HER2-low = 82 Chemotherapy (N=274) (capecitabine, nab-paclitaxel, paclitaxel) Primary endpoint - PFS by BIRC in HR+/HER2-low pts Key Secondary endpoint - PFS by BIRC in all HER2-low pts Secondary and other endpoints - PFS (INV), OS, ORR, DoR, CBR - Safety - efficacy in HR-/HER2-low pts - Treatment continued until disease progression/death, intolerable toxicity, patient withdrawal, or investigator decision. - Tumor response was assessed per RECIST v1.1. Stratification Factors: - HR+ (Prior CDK4/6i use) vs HR+ (No prior CDK4/6i use) vs HR- - HER2 expression (IHC 1+ vs IHC 2+/ISH-) - Prior taxane in non-metastatic setting (yes vs no) Statistical Considerations - Hierarchical testing was performed first for PFS in HR+/HER2-low pts, followed by PFS in all HER2-low pts, both at a two-sided α level of 0.05. One interim PFS analysis was planned. - Prespecified IA for PFS in HR+/HER2-low: approximately 216 (75%) PFS events. - Prespecified IA for PFS in all HER2-low: approximately 255 (75%) PFS events. Abbreviations: r/m BC, recurrent or metastatic breast cancer; ET, endocrine therapy; Q3W, every 3 weeks; HR+, Hormone Receptor-Positive; CDK4/6i, cyclin dependent kinase 4/6 inhibitor; BIRC, blinded independent review committee; INV, investigator; pts, patients; PFS, progression-free survival; OS, overall survival; ORR, objective response rate; DoR, duration of response; CBR, clinical benefit rate; IA, interim analysis. Presenter: Shusen Wang, MD [Slide 3] PFS (BIRC) in HR+/HER2-low: primary endpoint [KM curve: Progression-Free survival (%) vs Time (months), 0-18] Landmark annotations at 6 months: 81.4% (TQB2102); 45.9% (Chemotherapy) Legend: TQB2102 / Chemotherapy / + Censored | | TQB2102 (n=228) | Chemotherapy (n=233) | | PFS events, n (%) | 66 (29.0) | 148 (63.5) | | Median PFS (95% CI), mo | 16.4 (11.9, NE) | 5.5 (4.3, 6.9) | HR=0.29 (95% CI: 0.22, 0.39) p<0.0001* No. at risk (months 0 / 3 / 6 / 9 / 12 / 15 / 18): TQB2102: 228 / 191 / 123 / 75 / 25 / 11 / 0 Chemotherapy: 233 / 146 / 72 / 29 / 5 / 2 / 0 - The investigator-assessed PFS was consistent with BIRC: HR 0.31 (95% CI: 0.23, 0.42). *Based on prespecified PFS IA, P value was below the prespecified efficacy boundary (two-sided α = 0.0186; Lan-DeMets O'Brien-Fleming α-spending function). Abbreviations: PFS, progression-free survival; BIRC, blinded independent review committee; CI, confidence interval; HR, hazard ratio (stratified); NE, not estimable. Presenter: Shusen Wang, MD Survival table Endpoint: PFS by BIRC in HR+/HER2-low (primary endpoint) Arm | N | Median | Median CI TQB2102 | 228 | 16.4 mo | 11.9, NE Chemotherapy | 233 | 5.5 mo | 4.3, 6.9 HR: 0.29 HR CI: 0.22, 0.39 P: <0.0001 Landmark rates: 6 months: TQB2102 81.4% | Chemotherapy 45.9% [Slide 4] ORR, CBR and DoR by BIRC [Waterfall plots: Best Percent Change from Baseline (%) — left panel TQB2102, right panel Chemotherapy; bars colored HR+ / HR-] | | HR+/HER2-low TQB2102 (n=228) | HR+/HER2-low Chemo (n=233) | All HER2-low TQB2102 (n=269) | All HER2-low Chemo (n=274) | HR-/HER2-low TQB2102 (n=41) | HR-/HER2-low Chemo (n=41) | | ORR, n (%) | 152 (66.7) | 73 (31.3) | 185 (68.8) | 85 (31.0) | 33 (80.5) | 12 (29.3) | | CBR, n (%) | 177 (77.6) | 105 (45.1) | 212 (78.8) | 121 (44.2) | 35 (85.4) | 16 (39.0) | | Median DoR, mo (95% CI) | 15.0 (12.7, NE) | 7.1 (5.5, 9.8) | 15.0 (11.3, NE) | 7.1 (5.5, 9.8) | 11.3 (6.8, NE) | 11.1 (1.4, NE) | Abbreviations: Chemo, Chemotherapy; ORR, objective response rate; DoR, duration of response; CBR, clinical benefit rate; NE, not evaluable; CI, confidence interval. Presenter: Shusen Wang, MD

[Slide 3] OS in HR+ and all HER2-low: secondary endpoints • TQB2102 showed a clear OS benefit trend in both HR+ and all HER2-low patients. [Left panel: KM curve, Overall survival (%) vs Time (months)] HR+/HER2-low | TQB2102 (n=228) | Chemotherapy (n=233) OS events, n (%) | 16 (7.0) | 31 (13.3) HR (95% CI) | 0.47 (0.26, 0.87) Legend: TQB2102 / Chemotherapy / + Censored No. at risk (months 0, 3, 6, 9, 12, 15, 18, 21): TQB2102: 228, 221, 170, 112, 61, 24, 2, 0 Chemotherapy: 233, 218, 153, 97, 48, 18, 5, 0 [Right panel: KM curve, Overall survival (%) vs Time (months)] All HER2-low | TQB2102 (n=269) | Chemotherapy (n=274) OS events, n (%) | 17 (6.3) | 36 (13.1) HR (95% CI) | 0.44 (0.25, 0.79) Legend: TQB2102 / Chemotherapy / + Censored No. at risk (months 0, 3, 6, 9, 12, 15, 18, 21): TQB2102: 269, 261, 196, 127, 70, 25, 2, 0 Chemotherapy: 274, 256, 175, 110, 57, 19, 5, 0 Abbreviations: OS, overall survival; CI, confidence interval; HR, hazard ratio (stratified). Presenter: Shusen Wang, MD Survival table Endpoint: OS (secondary endpoints) - HR+/HER2-low and All HER2-low TQB2102 (HR+/HER2-low) | 228 Chemotherapy (HR+/HER2-low) | 233 TQB2102 (All HER2-low) | 269 Chemotherapy (All HER2-low) | 274 HR: HR+/HER2-low 0.47; All HER2-low 0.44 HR CI: HR+/HER2-low 0.26-0.87; All HER2-low 0.25-0.79 [Slide 4] PFS and OS in HR−/HER2-low: prespecified exploratory analyses • Consistent PFS (HR=0.27) and OS (HR=0.21) benefit trend of TQB2102 in HR−/HER2-low patients. [Left panel: PFS (BIRC) KM curve, Progression-Free survival (%) vs Time (months); annotations at 6 months: 82.8% and 41.2%] Legend: TQB2102 / Chemotherapy / + Censored No. at risk (months 0, 3, 6, 9, 12, 15): TQB2102: 41, 29, 18, 7, 3, 0 Chemotherapy: 41, 23, 10, 4, 3, 0 PFS | TQB2102 (n=41) | Chemotherapy (n=41) PFS events, n (%) | 12 (29.3) | 26 (63.4) Median PFS (95% CI), mo | 8.5 (8.2, 14.0) | 5.4 (2.9, 6.9) HR (95% CI) | 0.27 (0.13, 0.56) [Right panel: OS KM curve, Overall survival (%) vs Time (months)] Legend: TQB2102 / Chemotherapy / + Censored No. at risk (months 0, 3, 6, 9, 12, 15, 18): TQB2102: 41, 40, 26, 15, 9, 1, 0 Chemotherapy: 41, 38, 22, 13, 9, 1, 0 OS | TQB2102 (n=41) | Chemotherapy (n=41) OS events, n (%) | 1 (2.44) | 5 (12.2) HR (95% CI) | 0.21 (0.02, 1.85) Abbreviations: PFS, progression-free survival; BIRC, blinded independent review committee; OS, overall survival; CI, confidence interval; HR, hazard ratio (stratified). Presenter: Shusen Wang, MD Survival table Endpoint: PFS (BIRC) and OS - HR-/HER2-low, prespecified exploratory Arm | N | Median | Median CI TQB2102 | 41 | PFS 8.5 mo | 8.2-14.0 Chemotherapy | 41 | PFS 5.4 mo | 2.9-6.9 HR: PFS 0.27; OS 0.21 HR CI: PFS 0.13-0.56; OS 0.02-1.85 Landmark rates: 6 months: TQB2102 82.8% | Chemotherapy 41.2%

PFS (BIRC) in all HER2-low: key secondary endpoint [Kaplan-Meier curve: Progression-Free survival (%) vs Time (months), 0-18] Legend: TQB2102 / Chemotherapy / + Censored Landmark labels at 6 months on curve: 81.7% (TQB2102); 45.3% (Chemotherapy) TQB2102 (n=269) Chemotherapy (n=274) PFS events, n (%) 78 (29.0) 174 (63.5) Median PFS (95% CI), mo 16.4 (11.9, NE) 5.5 (4.3, 6.9) HR=0.29 (95% CI: 0.22, 0.38) p<0.0001* No. at risk (months 0, 3, 6, 9, 12, 15, 18) TQB2102: 269, 220, 141, 82, 28, 11, 0 Chemotherapy: 274, 169, 82, 33, 8, 2, 0 • The investigator-assessed PFS was consistent with BIRC: HR 0.29 (95% CI: 0.22, 0.39). *Based on prespecified PFS IA, P value was below the prespecified efficacy boundary (two-sided α = 0.0180; Lan-DeMets O'Brien-Fleming α-spending function). [footnote small/blurry] Abbreviations: PFS, progression-free survival; BIRC, blinded independent review committee; CI, confidence interval; HR, hazard ratio (stratified); NE, not estimable. Presenter: Shusen Wang, MD ALICANTE AUDITORIUM - HALL 6 Survival table Endpoint: PFS by BIRC, all HER2-low (key secondary endpoint) Arm | N | Median | Median CI TQB2102 | 269 | 16.4 mo | 11.9, NE Chemotherapy | 274 | 5.5 mo | 4.3, 6.9 HR: 0.29 HR CI: 0.22, 0.38 P: <0.0001 Landmark rates: 6 months: TQB2102 81.7% | Chemotherapy 45.3%

Demographics and Baseline Characteristics | Baseline characteristics | HR+/HER2-low TQB2102 (n=228) | HR+/HER2-low Chemotherapy (n=233) | All HER2-low TQB2102 (n=269) | All HER2-low Chemotherapy (n=274) | | Female, n (%) | 225 (98.7) | 232 (99.6) | 266 (98.9) | 272 (99.3) | | ECOG performance status, n (%) | | | | | | 0 | 77 (33.8) | 85 (36.5) | 95 (35.3) | 104 (38.0) | | 1 | 151 (66.2) | 147 (63.1) | 174 (64.7) | 169 (61.7) | | HER2 status*, n (%) | | | | | | IHC 1+ | 133 (58.3) | 135 (57.9) | 157 (58.4) | 158 (57.7) | | IHC2+/ISH- | 95 (41.7) | 98 (42.1) | 112 (41.6) | 116 (42.3) | | HR status*, n (%) | | | | | | ER+/PR+ | 132 (57.9) | 132 (56.7) | 132 (49.1) | 132 (48.2) | | ER+/PR- | 94 (41.2) | 96 (41.2) | 94 (34.9) | 96 (35.0) | | ER-/PR+ | 2 (0.9) | 5 (2.1) | 2 (0.7) | 5 (1.8) | | ER-/PR- | 0 | 0 | 41 (15.2) | 41 (15.0) | | De-novo disease at diagnosis, n (%) | 52 (22.8) | 50 (21.5) | 65 (24.2) | 60 (21.9) | | Visceral metastasis, n (%) | 190 (83.3) | 184 (79.0) | 219 (81.4) | 213 (77.7) | | Liver metastasis, n (%) | 144 (63.2) | 125 (53.7) | 162 (60.2) | 139 (50.7) | | Brain metastasis, n (%) | 10 (4.4) | 13 (5.6) | 12 (4.5) | 13 (4.7) | *HER2-low status defined by central laboratory testing data; HR status was evaluated at the pathology department of the participating study site. Presenter: Shusen Wang, MD

Conclusions - TQB2102 demonstrated a statistically significant and clinically meaningful improvement in PFS and a clear trend toward OS benefit versus chemotherapy in patients with HER2-low recurrent/metastatic breast cancer. - PFS by BIRC in all HER2-low patients: median 16.4 vs 5.5 mo (HR=0.29; 95% CI, 0.22-0.38; p<0.0001). - PFS benefit was observed across all predefined subgroups. - OS in all HER2-low patients: HR=0.44 (95% CI, 0.25-0.79). - ORR by BIRC in all HER2-low patients: 68.8% (95% CI, 62.9-74.3) vs 31.0% (95% CI, 25.6-36.9). - The overall safety profile of TQB2102 was manageable and generally consistent with previous experience, with a low incidence of ILD. The results of APEX-Breast01 support TQB2102, a first-in-class HER2 biparatopic ADC, as a promising treatment option for patients with HER2-low breast cancer who have not previously received chemotherapy in the metastatic setting. Presenter: Shusen Wang, MD

[Slide 2] 13:30 - 15:00 Proffered paper 1: Breast cancer, early and advanced stage CHAIRS: KOMAL JHAVERI, BARBARA PISTILLI PFS (BIRC) in HR+/HER2-low: primary endpoint TQB2102 (n=228) | Chemotherapy (n=233) PFS events, n (%) | 66 (29.0) | 148 (63.5) Median PFS (95% CI), mo | 16.4 (11.9, NE) | 5.5 (4.3, 6.9) HR=0.29 (95% CI: 0.22, 0.39) p<0.0001* [curve annotations at 6 months: 81.4% (TQB2102), 45.9% (Chemotherapy)] Legend: TQB2102 | Chemotherapy | + Censored No. at risk (Time, months: 0, 3, 6, 9, 12, 15, 18) TQB2102: 228, 191, 123, 75, 25, 11, 0 Chemotherapy: 233, 146, 72, 29, 5, 2, 0 • The investigator-assessed PFS was consistent with BIRC: HR 0.31 (95% CI: 0.23, 0.42). *Based on prespecified PFS IA, P value was below the prespecified efficacy boundary (two-sided α = 0.0186; Lan-DeMets O'Brien-Fleming α-spending function). Abbreviations: PFS, progression-free survival; BIRC, blinded independent review committee; CI, confidence interval; HR, hazard ratio (stratified); NE, not estimable. Presenter: Shusen Wang, MD Shusen Wang APEX-Breast01: TQB2102 versus chemotherapy in HER2-low recurrent or metastatic breast cancer (r/m BC) - Results from a randomized phase 3 study ALICANTE AUDITORIUM - HALL 6 Survival table Endpoint: PFS by BIRC, HR+/HER2-low (primary endpoint; prespecified interim analysis) Arm | N | Median | Median CI TQB2102 | 228 | 16.4 months | 11.9, NE Chemotherapy | 233 | 5.5 months | 4.3, 6.9 HR: 0.29 HR CI: 0.22, 0.39 P: <0.0001 Landmark rates: 6 months: TQB2102 81.4% | Chemotherapy 45.9% [Slide 3] 13:30 - 15:00 Proffered paper 1: Breast cancer, early and advanced stage CHAIRS: KOMAL JHAVERI, BARBARA PISTILLI OS in HR+ and all HER2-low: secondary endpoints • TQB2102 showed a clear OS benefit trend in both HR+ and all HER2-low patients. [Left KM: Overall survival (%) vs Time (months) 0-21] HR+/HER2-low TQB2102 (n=228) Chemotherapy (n=233) OS events, n (%) 16 (7.0) 31 (13.3) HR (95% CI) 0.47 (0.26, 0.87) No. at risk (0, 3, 6, 9, 12, 15, 18, 21) TQB2102: 228, 221, 170, 112, 61, 24, 2, 0 Chemotherapy: 233, 218, 153, 97, 48, 18, 5, 0 [Right KM: Overall survival (%) vs Time (months) 0-21] All HER2-low TQB2102 (n=269) Chemotherapy (n=274) OS events, n (%) 17 (6.3) 36 (13.1) HR (95% CI) 0.44 (0.25, 0.79) No. at risk (0, 3, 6, 9, 12, 15, 18, 21) TQB2102: 269, 261, 196, 127, 70, 25, 2, 0 Chemotherapy: 274, 256, 175, 110, 57, 19, 5, 0 Legend: TQB2102 / Chemotherapy / + Censored Abbreviations: OS, overall survival; CI, confidence interval; HR, hazard ratio (stratified). Presenter: Shusen Wang, MD [Speaker video: Shusen Wang] Shusen Wang APEX-Breast01: TQB2102 versus chemotherapy in HER2-low recurrent or metastatic breast cancer (r/m BC) - Results from a randomized phase 3 study ALICANTE AUDITORIUM - HALL 6 Survival table Endpoint: OS (secondary endpoint), HR+/HER2-low and all HER2-low TQB2102 (HR+/HER2-low) | 228 Chemotherapy (HR+/HER2-low) | 233 TQB2102 (all HER2-low) | 269 Chemotherapy (all HER2-low) | 274 HR: 0.47 (HR+/HER2-low); 0.44 (all HER2-low) HR CI: 0.26, 0.87 (HR+/HER2-low); 0.25, 0.79 (all HER2-low)

TQB2102 HR+ (n=228) | TPC HR+ (n=233) | TQB2102 HER2-low (n=269) | TPC HER2-low (n=274) | TQB2102 HR- (n=41) | TPC HR- (n=41) mPFS^a, mo (95% CI) 16.4 (11.9, NE) | 5.5 (4.3, 6.9) | 16.4 (11.9, NE) | 5.5 (4.3, 6.9) | 8.5 (8.2, 14.0) | 5.4 (2.9, 6.9) PFS HR (95% CI), P value^b 0.29 (0.22, 0.39) p<0.0001 | 0.29 (0.22, 0.38) p<0.0001 | 0.27 (0.13, 0.56) OS HR (95% CI) 0.47 (0.26, 0.87) | 0.44 (0.25, 0.79) | 0.21 (0.02, 1.85) ORR^a, % (95% CI) 66.7(60.1, 72.8) | 31.3 (25.4, 37.7) | 68.8 (62.9, 74.3) | 31.0 (25.6, 36.9) | 80.5 (65.1, 91.2) | 29.3 (16.1, 45.5) NE, non-estimable. ^a by BIRC. ^b 2-sided log-rank test.

[Slide 1] Background and Study Design • In OptiTROP-Breast01 (NCT05347134), sac-TMT significantly improved PFS versus physician's choice of chemotherapy (HR 0.32) and OS at the prespecified interim analysis (HR 0.53; P = 0.0005) in patients with previously treated TNBC.1 Key Eligibility Criteria • ECOG PS 0 or 1 • Histologically confirmed unresectable locally advanced or metastatic TNBC • ≥2 prior systemic therapies, including ≥1 for metastatic disease • Prior taxane therapy Stratification factors • Number of prior lines of therapy (2–3 vs >3) • Presence of liver metastases (yes vs no) R 1:1 N = 263 Sac-TMT 5 mg/kg IV, Q2W (n = 130) Physician's choice of chemotherapy: eribulin, capecitabine, gemcitabine, or vinorelbine (n = 133) Primary endpoint • PFS assessed by BICR Secondary endpoints • OS (key secondary endpoint) • Investigator-assessed PFS • ORR, DCR, DOR • Safety Patients received sac-TMT or chemotherapy until disease progression, unacceptable toxicity or any other reason for discontinuation Here, we report the final OS analysis of OptiTROP-Breast01 after extended follow-up. 1. Yin Y et al. Nat Med. 2025 Jun;31(6):1969-1975. BICR, blinded independent central review; DCR, disease control rate; DOR, duration of response; ECOG PS, Eastern Cooperative Oncology Group performance status; HR, hazard ratio; ORR, objective response rate; OS, overall survival, PFS, progression-free survival; R, randomization; sac-TMT, sacituzumab tirumotecan; TNBC, triple-negative breast cancer. Professor Man Li [Slide 2] Overall Survival Sac-TMT demonstrated a sustained OS benefit over chemotherapy, with a 40% reduction in the risk of death. Overall Survival (%) mOS: 14.7 mo mOS: 9.9 mo 26.8% 13.2% sac-TMT Chemotherapy Time in Months | Sac-TMT (n = 130) | Chemo (n = 133) OS events, n (%) | 88 (67.7) | 109 (82.0) Median OS, mo (95% CI) | 14.7 (12.1, 20.1) | 9.9 (8.6, 11.8) 30-mo OS rate, % (95% CI) | 26.8 (19.0, 35.2) | 13.2 (7.8, 20.0) HR 0.60 (95% CI: 0.45–0.80) Nominal P = 0.0004 At Risk sac-TMT 130 120 111 93 73 58 52 45 38 30 17 5 0 Chemotherapy 133 119 95 70 49 42 32 30 24 13 7 3 1 0 Median follow-up: 29.4 months. Final OS analyses were descriptive. OS met the prespecified significance criterion at the interim analysis. Data cutoff: 17 September 2025. CI, confidence interval; mo, month. Professor Man Li Survival table Endpoint: Overall survival (final analysis, descriptive) Arm | N | Median | Median CI Sac-TMT | 130 | 14.7 mo | 12.1, 20.1 Chemotherapy | 133 | 9.9 mo | 8.6, 11.8 HR: 0.60 HR CI: 0.45–0.80 P: Nominal P = 0.0004 30-mo OS rate: Sac-TMT 26.8% (19.0, 35.2) vs Chemo 13.2% (7.8, 20.0) [Slide 3] Subsequent Anticancer Therapy and Adjusted OS In the chemotherapy group, 27.8% of patients received subsequent ADC therapy. RPSFT adjustment supported the ITT OS findings (adjusted HR 0.55). | Sac-TMT (n = 130) | Chemo (n = 133) Subsequent anticancer therapy, n (%) | 88 (67.7) | 98 (73.7) Chemotherapy | 75 (57.7) | 82 (61.7) Targeted therapy | 55 (42.3) | 65 (48.9) ADC | 24 (18.5) | 37 (27.8) TROP2-directed ADCs | 6 (4.6) | 29 (21.8) SG/Dato-DXd/others | 3 (2.3) / 0 / 3 (2.3) | 17 (12.8) / 5 (3.8) / 7 (5.3) HER2-directed ADCs | 16 (12.3) | 11 (8.3) T-DXd/others | 15 (11.5) / 2 (1.5) | 8 (6.0) / 3 (2.3) ADCs with other targets | 5 (3.8) | 4 (3.0) Immunotherapy | 18 (13.8) | 20 (15.0) Adjusted overall survival* | Sac-TMT (n = 130) | Chemo (n = 133) OS events, n (%) | 88 (67.7) | 96 (72.2) Median OS, mo (95% CI) | 14.7 (12.1, 20.1) | 9.4 (8.4, 11.2) HR 0.55 (95% CI: 0.39–0.73) Overall Survival (%) sac-TMT Chemotherapy Time in Months At Risk sac-TMT 130 120 111 93 73 58 52 45 38 30 17 5 0 Chemotherapy 133 119 92 67 43 34 22 4 0 Data cutoff: 17 September 2025. *An RPSFT model was used to account for the potential impact of subsequent ADC therapy on OS among patients in the chemotherapy group who received such treatment. ADC, antibody-drug conjugate; Dato-DXd, datopotamab deruxtecan; HER2, Human epidermal growth factor receptor 2; ITT, intention-to-treat; RPSFT, rank-preserving structural failure time model; SG, sacituzumab govitecan; T-DXd, trastuzumab deruxtecan. Professor Man Li Survival table Endpoint: RPSFT-adjusted overall survival Arm | N | Median | Median CI Sac-TMT | 130 | 14.7 mo | 12.1, 20.1 Chemo | 133 | 9.4 mo | 8.4, 11.2 HR: 0.55 HR CI: 0.39–0.73 [Slide 4] OS by Prior PD-1/PD-L1 Inhibitor Exposure and HER2 Expression OS favored sac-TMT in subgroups defined by prior PD-1/PD-L1 inhibitor exposure and HER2 expression. A. Prior treatment with PD-1/PD-L1 inhibitor | Sac-TMT (n = 32) | Chemo (n = 36) Median OS, mo | 17.4 | 11.2 HR (95% CI) | 0.54 (0.32, 0.93) B. Prior treatment without PD-1/PD-L1 inhibitor | Sac-TMT (n = 98) | Chemo (n = 97) Median OS, mo | 14.6 | 9.8 HR (95% CI) | 0.64 (0.46, 0.90) C. HER2 Expression: 0 | Sac-TMT (n = 74) | Chemo (n = 77) Median OS, mo | 17.4 | 10.9 HR (95% CI) | 0.64 (0.43, 0.93) D. HER2 Expression: Low | Sac-TMT (n = 56) | Chemo (n = 56) Median OS, mo | 14.0 | 8.6 HR (95% CI) | 0.60 (0.39, 0.91) Data cutoff: 17 September 2025. Subgroup analyses were descriptive. HRs and 95% CIs were estimated using unstratified Cox models. Professor Man Li Survival table Endpoint: OS by subgroup (descriptive) Prior PD-1/PD-L1: Sac-TMT | 32 | 17.4 Prior PD-1/PD-L1: Chemo | 36 | 11.2 No prior PD-1/PD-L1: Sac-TMT | 98 | 14.6 No prior PD-1/PD-L1: Chemo | 97 | 9.8 HER2 0: Sac-TMT | 74 | 17.4 HER2 0: Chemo | 77 | 10.9 HER2-low: Sac-TMT | 56 | 14.0 HER2-low: Chemo | 56 | 8.6 HR: A 0.54; B 0.64; C 0.64; D 0.60 HR CI: A 0.32, 0.93; B 0.46, 0.90; C 0.43, 0.93; D 0.39, 0.91

10:15 - 11:45 Rapid oral 2: Breast cancer, advanced stage CHAIRS: ANTONIO LLOMBART CUSSAC, SHERENE LOI, PEDRAM RAZAVI Conclusions • After a median follow-up of 29.4 months, sac-TMT demonstrated a clinically meaningful and sustained OS benefit versus chemotherapy (HR 0.60, 95% CI 0.45–0.80). • RPSFT adjustment for subsequent ADC therapy supported the ITT OS findings (adjusted HR 0.55, 95% CI 0.39–0.73). • OS findings generally favored sac-TMT across prespecified subgroups, including those defined by prior PD-1/PD-L1 inhibitor exposure and HER2 expression. • The safety profile remained manageable, with no new safety signals identified during extended follow-up. Final OS Median OS, months Sac-TMT 14.7 Chemo 9.9 HR 0.60 (95% CI 0.45–0.80); Nominal P = 0.0004 RPSFT-adjusted OS Median OS, months Sac-TMT 14.7 Chemo 9.4 HR 0.55 (95% CI 0.39–0.73) These final OS results further support sac-TMT as an effective treatment option for patients with previously treated locally recurrent or metastatic TNBC. Professor Man Li

[Slide 1] Why the payload resistance dominates Most known resistant mechanisms acts downstream of the antigen Binding & internalisation -> Lysosomal processing -> Payload in cytoplasm -> TOP1-DNA cleavage complex -> DNA damage -> cell death Antigen loss; TACSTD2 mutation | Trafficking / lysosomal defects | Efflux pumps (e.g. ABCG2: SN-38) | TOP1 mutation / loss; SLFN11 loss | Enhanced DDR/HR repair Fixed by switching target Implication: need biomarkers that measure payload sensitivity (TOP1, SLFN11, efflux; rebiopsy or ctDNA) and not antigen alone Sherene Loi, Peter MacCallum Cancer Centre [Slide 2] Antigen expression does not predict for responding patients • B7-H3 responses did not differ by B7H3 responses • NECTIN-4 responses recorded in negative patients • TROP2- no benefit recorded across TROP2 levels Caveat: mainly archival IHC, often not from the progressing lesion & IHC is a blunt tool But the bystander effect and payload sensitivity seem to matter more than antigen density: probably likely explanation for T-Dxd and HER2 low activity Sherene Loi MD PhD [Slide 4] Conclusions 1 OptiTROP-Breast01: durable OS benefit (HR 0.60). Confirms the TROP2-TOP1 class effect seen in ASCENT; 2 After a TOP1 ADC, switching the target alone gives ~5% ORR (consistent with other datasets/cohorts) 3 Switching the payload (Nectin-4 MMAE) restores responses in ~30%, but PFS is 3.7 mo: needs a prospective randomized study 4 Antigen expression does not select patients: we need biomarkers of payload sensitivity 5 Rationale combinations needed with ADCs in 1L and 2L post settings Take home messages Every ADC trial should report prior ADC target, payload and previous PFS Randomise payload-defined sequences against active chemotherapy Biopsy or ctDNA at ADC progression Novel payloads and dual-payload ADCs needed Sherene Loi, Peter MacCallum Cancer Centre

[Slide 1] Safety Summary and Common TRAEs Safety-related variables | Sac-TMT (n = 130) n (%) | Chemo (n = 132) n (%) TRAEs | 130 (100) | 127 (96.2) Grade ≥3 TRAEs | 84 (64.6) | 76 (57.6) Serious TRAEs | 28 (21.5) | 18 (13.6) TRAEs leading to discontinuation | 2 (1.5) | 2 (1.5) TRAEs leading to death | 1 (0.8)* | 0 Selected common TRAEs† | All grade | Grade ≥3 | All grade | Grade ≥3 Anemia | 108 (83.1) | 38 (29.2) | 72 (54.5) | 8 (6.1) NEUT count decreased | 102 (78.5) | 47 (36.2) | 98 (74.2) | 62 (47.0) WBC count decreased | 100 (76.9) | 38 (29.2) | 101 (76.5) | 48 (36.4) Stomatitis | 65 (50.0) | 13 (10.0) | 6 (4.5) | 1 (0.8) PLT count decreased | 54 (41.5) | 17 (13.1) | 34 (25.8) | 5 (3.8) Nausea | 52 (40.0) | 0 | 26 (19.7) | 0 Vomiting | 46 (35.4) | 0 | 18 (13.6) | 0 Rash | 44 (33.8) | 5 (3.8) | 1 (0.8) | 0 ALT increased | 37 (28.5) | 1 (0.8) | 47 (35.6) | 0 AST increased | 32 (24.6) | 3 (2.3) | 63 (47.7) | 1 (0.8) Data cutoff: 17 September 2025. †Summary of any grade TRAEs with incidence ≥30% in either treatment group. *One treatment-related death due to multiple organ dysfunction syndrome was previously reported.¹ 1. Yin Y et al. Nat Med. 2025 Jun,31(6):1969-1975. ALT, alanine aminotransferase; AST, aspartate aminotransferase; NEUT, neutrophil count; PLT, platelet; TRAE, treatment-related adverse event; WBC, white blood cell. • Hematologic toxicities were the most common TRAEs in both groups. Stomatitis occurred in 50.0% of patients receiving sac-TMT (grade 3, 10.0%). • TRAEs led to treatment discontinuation in 1.5% of patients in each group. • Ocular events with sac-TMT included dry eye (3.8%) and blurred vision (1.5%), all grade 1–2. • The safety profile remained manageable, with no new safety signals identified during extended follow-up. Professor Man Li [Slide 4] Overall Survival: Subgroup Analyses A consistent OS benefit of sac-TMT over chemotherapy was observed across all prespecified subgroups. Subgroup | Number of events/patients sac-TMT | Chemotherapy | Hazard ratio (95%CI) All Patients | 88/130 | 109/133 | 0.60 (0.45-0.80) Age <65 yr | 77/119 | 96/119 | 0.58 (0.43-0.78) ECOG Performance Status 0 | 26/40 | 31/39 | 0.59 (0.35-1.00) ECOG Performance Status 1 | 62/90 | 78/94 | 0.62 (0.44-0.87) Prior Line of Therapy 2~3 | 75/114 | 93/113 | 0.58 (0.43-0.79) Prior Line of Therapy >3 | 13/16 | 16/20 | 0.84 (0.40-1.75) Liver Metastases Yes | 33/45 | 38/45 | 0.67 (0.42-1.07) Liver Metastases No | 55/85 | 71/88 | 0.58 (0.41-0.83) Initial Diagnosis of TNBC Yes | 62/92 | 69/88 | 0.62 (0.44-0.88) Initial Diagnosis of TNBC No | 26/38 | 40/45 | 0.61 (0.37-1.00) Prior Treatment with PD-1 or PD-L1 Inhibitors Yes | 23/32 | 31/36 | 0.54 (0.32-0.93) Prior Treatment with PD-1 or PD-L1 Inhibitors No | 65/98 | 78/97 | 0.64 (0.46-0.90) Lymph Node Metastases Yes | 59/81 | 64/75 | 0.59 (0.42-0.85) Lymph Node Metastases No | 29/49 | 45/58 | 0.59 (0.37-0.95) Visceral Metastases Yes | 82/115 | 92/113 | 0.65 (0.48-0.88) Visceral Metastases No | 6/15 | 17/20 | 0.34 (0.13-0.86) HER2 Expression 0 | 48/74 | 60/77 | 0.64 (0.43-0.93) HER2 Expression Low Expression | 40/56 | 49/56 | 0.60 (0.39-0.91) <---sac-TMT Better Chemotherapy Better---> Data cutoff: 17 September 2025. Statistical analysis was not performed when the number of patients in the subgroup was less than 10% of the total patients. Professor Man Li

A Final OS No. of Patients / No. of Events / Median OS (Months) (95% CI) sac-TMT 130 88 14.7 (12.1, 20.1) Chemotherapy 133 109 9.9 (8.6, 11.8) HR (95% CI) 0.60 (0.45, 0.80) P Value 0.0004 26.8% 13.2% sac-TMT Chemotherapy Time in Months At Risk (censored) sac-TMT 130 (0) 120 (3) 111 (3) 93 (5) 73 (6) 58 (8) 52 (9) 45 (10) 38 (11) 30 (15) 17 (25) 5 (37) 0 (42) Chemotherapy 133 (0) 119 (3) 95 (4) 70 (4) 49 (6) 42 (8) 32 (9) 30 (9) 24 (10) 13 (12) 7 (18) 3 (21) 1 (23) 0 (24) B RPSFT-adjusted OS sac-TMT 130 88 14.7 (12.1, 20.1) Chemotherapy 133 96 9.4 (8.4, 11.2) HR (95% CI) 0.55 (0.39, 0.73) P Value 0.0004 At Risk (censored) sac-TMT 130 (0) 120 (3) 111 (3) 93 (5) 73 (6) 58 (8) 52 (9) 45 (10) 38 (11) 30 (15) 17 (25) 5 (37) 0 (42) Chemotherapy 133 (0) 119 (3) 92 (4) 67 (4) 43 (8) 34 (8) 22 (16) 4 (33) 0 (37) C Censoring-adjusted OS sac-TMT 130 75 14.7 (12.1, 20.4) Chemotherapy 133 81 9.4 (8.4, 11.8) HR (95% CI) 0.59 (0.43, 0.82) P Value 0.0013 At Risk (censored) sac-TMT 130 (0) 119 (4) 107 (7) 86 (12) 68 (14) 52 (19) 45 (22) 36 (26) 29 (29) 21 (36) 12 (43) 3 (52) 0 (55) Chemotherapy 133 (0) 113 (9) 82 (20) 49 (31) 31 (38) 26 (40) 19 (44) 18 (44) 15 (45) 7 (46) 5 (48) 3 (49) 1 (51) 0 (52) Survival table Endpoint: Overall survival (A final; B RPSFT-adjusted; C censoring-adjusted) Arm | N | Median | Median CI sac-TMT (final OS) | 130 | 14.7 mo | 12.1, 20.1 Chemotherapy (final OS) | 133 | 9.9 mo | 8.6, 11.8 sac-TMT (RPSFT-adjusted) | 130 | 14.7 mo | 12.1, 20.1 Chemotherapy (RPSFT-adjusted) | 133 | 9.4 mo | 8.4, 11.2 sac-TMT (censoring-adjusted) | 130 | 14.7 mo | 12.1, 20.4 Chemotherapy (censoring-adjusted) | 133 | 9.4 mo | 8.4, 11.8 HR: Final 0.60; RPSFT 0.55; censoring-adjusted 0.59 HR CI: 0.45, 0.80; 0.39, 0.73; 0.43, 0.82 P: Final 0.0004; RPSFT 0.0004; censoring-adjusted 0.0013 Final OS at 30 months: 26.8% vs 13.2% (printed labels)

[Slide 2] 10:15 - 11:45 Rapid oral 2: Breast cancer, advanced stage CHAIRS: ANTONIO LLOMBART CUSSAC, SHERENE LOI, PEDRAM RAZAVI Overall Survival Sac-TMT demonstrated a sustained OS benefit over chemotherapy, with a 40% reduction in the risk of death. Overall Survival (%) mOS: 14.7 mo mOS: 9.9 mo 26.8% 13.2% sac-TMT Chemotherapy Time in Months | Sac-TMT (n = 130) | Chemo (n = 133) OS events, n (%) | 88 (67.7) | 109 (82.0) Median OS, mo (95% CI) | 14.7 (12.1, 20.1) | 9.9 (8.6, 11.8) 30-mo OS rate, % (95% CI) | 26.8 (19.0, 35.2) | 13.2 (7.8, 20.0) HR 0.60 (95% CI: 0.45–0.80) Nominal P = 0.0004 At Risk sac-TMT 130 120 111 93 73 58 52 45 38 30 17 5 0 Chemotherapy 133 119 95 70 49 42 32 30 24 13 7 3 1 0 Median follow-up: 29.4 months. Final OS analyses were descriptive. OS met the prespecified significance criterion at the interim analysis. Data cutoff: 17 September 2025. CI, confidence interval; mo, month. Professor Man Li Man Li Final OS analysis of sacituzumab tirumotecan versus chemotherapy in previously treated locally recurrent or metastatic triple-negative breast cancer: Phase III OptiTROP-Breast01 ALICANTE AUDITORIUM - HALL 6 Survival table Endpoint: Overall survival (final analysis, descriptive) Arm | N | Median | Median CI Sac-TMT | 130 | 14.7 mo | 12.1, 20.1 Chemotherapy | 133 | 9.9 mo | 8.6, 11.8 HR: 0.60 HR CI: 0.45–0.80 P: Nominal P = 0.0004 30-mo OS rate: Sac-TMT 26.8% (19.0, 35.2) vs Chemo 13.2% (7.8, 20.0) [Slide 3] 10:15 - 11:45 Rapid oral 2: Breast cancer, advanced stage CHAIRS: ANTONIO LLOMBART CUSSAC, SHERENE LOI, PEDRAM RAZAVI OS by Prior PD-1/PD-L1 Inhibitor Exposure and HER2 Expression OS favored sac-TMT in subgroups defined by prior PD-1/PD-L1 inhibitor exposure and HER2 expression. A. Prior treatment with PD-1/PD-L1 inhibitor | Sac-TMT (n = 32) | Chemo (n = 36) Median OS, mo | 17.4 | 11.2 HR (95% CI) | 0.54 (0.32, 0.93) B. Prior treatment without PD-1/PD-L1 inhibitor | Sac-TMT (n = 98) | Chemo (n = 97) Median OS, mo | 14.6 | 9.8 HR (95% CI) | 0.64 (0.46, 0.90) C. HER2 Expression: 0 | Sac-TMT (n = 74) | Chemo (n = 77) Median OS, mo | 17.4 | 10.9 HR (95% CI) | 0.64 (0.43, 0.93) D. HER2 Expression: Low | Sac-TMT (n = 56) | Chemo (n = 56) Median OS, mo | 14.0 | 8.6 HR (95% CI) | 0.60 (0.39, 0.91) Data cutoff: 17 September 2025. Subgroup analyses were descriptive. HRs and 95% CIs were estimated using unstratified Cox models. Professor Man Li Man Li Final OS analysis of sacituzumab tirumotecan versus chemotherapy in previously treated locally recurrent or metastatic triple-negative breast cancer: Phase III OptiTROP-Breast01 ALICANTE AUDITORIUM - HALL 6 Survival table Endpoint: OS by subgroup (descriptive) Prior PD-1/PD-L1: Sac-TMT | 32 | 17.4 Prior PD-1/PD-L1: Chemo | 36 | 11.2 No prior PD-1/PD-L1: Sac-TMT | 98 | 14.6 No prior PD-1/PD-L1: Chemo | 97 | 9.8 HER2 0: Sac-TMT | 74 | 17.4 HER2 0: Chemo | 77 | 10.9 HER2-low: Sac-TMT | 56 | 14.0 HER2-low: Chemo | 56 | 8.6 HR: A 0.54; B 0.64; C 0.64; D 0.60 HR CI: A 0.32, 0.93; B 0.46, 0.90; C 0.43, 0.93; D 0.39, 0.91

A Final OS No. of Patients | No. of Events | Median OS (Months) (95% CI) sac-TMT | 130 | 88 | 14.7 (12.1, 20.1) Chemotherapy | 133 | 109 | 9.9 (8.6, 11.8) HR (95% CI) 0.60 (0.45, 0.80) P Value 0.0004 26.8% 13.2% sac-TMT — Chemotherapy Overall Survival (%) Time in Months At Risk (censored) B RPSFT-adjusted OS sac-TMT | 130 | 88 | 14.7 (12.1, 20.1) Chemotherapy | 133 | 96 | 9.4 (8.4, 11.2) HR (95% CI) 0.55 (0.39, 0.73) P Value 0.0004 C Censoring-adjusted OS sac-TMT | 130 | 75 | 14.7 (12.1, 20.4) Chemotherapy | 133 | 81 | 9.4 (8.4, 11.8) HR (95% CI) 0.59 (0.43, 0.82) P Value 0.0013 Survival table Endpoint: Final OS Arm | N | Median | Median CI sac-TMT | 130 | 14.7 | 12.1, 20.1 Chemotherapy | 133 | 9.9 | 8.6, 11.8 HR: 0.60 HR CI: 0.45, 0.80 P: 0.0004 30 mo (marked on curve): sac-TMT 26.8%, Chemotherapy 13.2%

[Slide 1] DESTINY-Breast09 study design Randomized, multicenter, Phase III trial (NCT04784715)1 Brain MRI/CT imaging schedule All patients at BLb Only patients with INV-assessed BMs at BL were mandated regular on-treatment brain MRI/CT imagingc All patients at EoTb Patient population Key eligibility criteria: • HER2+ a/mBC • Asymptomatic/inactive BMs allowed • DFI >6 mo from last chemotherapy or HER2-targeted therapy in neoadjuvant/adjuvant setting Prior lines of therapy: • One prior line of ET for mBC permitted • No other prior systemic treatment for a/mBCa Stratification factors • De novo vs recurrent disease • HR+ or HR− • PIK3CAm (detected vs non-detected) Rd 1:1:1 n=387 T-DXde + placebo Blinded until final PFS analysis n=383 T-DXde + Pf n=387 THP Taxane (paclitaxel or docetaxel)g + trastuzumabh + Pf ITT Primary endpoint • PFS (BICR per RECIST 1.1) Key secondary endpoint • OS Other secondary endpoints • PFS (INV) • ORR (BICR/INV) • DOR (BICR/INV) • PFS2 (INV) • Safety/tolerability Exploratory efficacy analysis • All brain MRI/CT scans were assessed retrospectively for BMs at BL by BICR after EoT • Systemic and CNS responses were assessed by BICRi [small-print footnotes largely illegible] 1. Tolaney SM, et al. N Engl J Med. 2026;394:551–562. 2. Eisenhauer EA, et al. Eur J Cancer. 2009;45:228–247. Cristina Saura DESTINY-Breast09 [Slide 2] Exploratory efficacy analysis R 1:1:1 ITT population (n=770)a Patient population • BICR-assessed BMs at BL, which must be: – Clinically inactive OR – Treated and asymptomatic, requiring no further treatment with corticosteroids or anticonvulsants (≥2 weeks between the end of whole brain radiotherapy and trial enrolment)b • ≥1 follow-up brain scan after BL Patients with BICR-assessed BMs at BL (n=61)c T-DXd + P (n=31) THP (n=30) Exploratory outcomes by BICRd • Systemic PFS, CNS-PFS • Systemic cORR, CNS-cORR • DOR, CNS-DOR Data cutoff: February 26, 2025 aT-DXd + P and THP arms only; the T-DXd + placebo arm will remain blinded until the final PFS analysis. bPatients with treated, no longer symptomatic BMs could be included in the study if patients had recovered from the acute toxic effect of radiotherapy. cA total of 47 patients (T-DXd + P, n=25; THP, n=22) had INV-assessed BMs at BL before retrospective BICR assessment at EoT. Retrospective BICR assessment confirmed a total of 61 patients with BMs at BL: 32 patients had BMs at BL as assessed both by BICR and INV (16 in each arm), while 29 patients had BMs at BL as assessed by BICR but not by INV (T-DXd + P, n=15; THP, n=14). Discordance levels across BICR- and INV-assessed BMs at BL (i.e., number of patients whose BMs were not identified during either BICR or INV assessments divided by the overall number of assessments) were comparable across treatment arms (T-DXd + P vs THP: 0.063 vs 0.052). dEndpoints not prespecified in the clinical study protocol. CNS-cORR, confirmed objective response rate in the CNS; CNS-DOR, duration of response in the CNS; CNS-PFS, progression-free-survival in the CNS; cORR, confirmed objective response rate. Cristina Saura DESTINY-Breast09 [Slide 3] Systemic PFS and CNS-PFS in patients with BMs at BL BMs at BL assessed by BICR Systemic PFS T-DXd + P | THP PFS | T-DXd + P (n=31) | THP (n=30) Events | 13 | 22 Median, mo (95% CI) | 36.5 (19.1–NE) | 11.1 (7.5–15.4) HR (95% CI) | 0.30 (0.14–0.60) 12-month PFS 86.5% 39.4% Δ47.1% HR 0.30 (95% CI, 0.14–0.60) mPFS PFS, % Time from randomization (mo) No. at risk CNS-PFS CNS-PFS | T-DXd + P (n=31) | THP (n=30) Events | 9 | 17 Median, mo (95% CI) | NE (19.1–NE) | 9.7 (4.2–19.2) HR (95% CI) | 0.26 (0.11–0.57) 12-month CNS-PFS 88.9% 42.5% Δ46.4% HR 0.26 (95% CI, 0.11–0.57) mCNS-PFS CNS-PFS, % Patients with BMs at BL in the T-DXd + P arm had substantially longer systemic PFS and CNS-PFS than those in the THP arm Responses by BICR. Data cutoff: February 26, 2025 (median [range] follow-up: 23.6 mo [2.8–41.4]). Results should be interpreted with caution due to the low number of patients in the subgroup. Due to the low number of events recorded, PFS could not be calculated for the 32 patients with BMs at BL assessed by both BICR and INV. CI, confidence interval; m(CNS)-PFS, median (central nervous system) progression-free survival; NE, not evaluable. Cristina Saura DESTINY-Breast09 Survival table Endpoint: Systemic PFS and CNS-PFS (BICR), patients with brain metastases at baseline Arm | N | Median | Median CI T-DXd + P - systemic PFS | 31 | 36.5 mo | 19.1-NE THP - systemic PFS | 30 | 11.1 mo | 7.5-15.4 T-DXd + P - CNS-PFS | 31 | NE | 19.1-NE THP - CNS-PFS | 30 | 9.7 mo | 4.2-19.2 HR: Systemic PFS 0.30; CNS-PFS 0.26 HR CI: Systemic PFS 0.14-0.60; CNS-PFS 0.11-0.57 12-month PFS 86.5% vs 39.4% (delta 47.1%) 12-month CNS-PFS 88.9% vs 42.5% (delta 46.4%) [Slide 4] Systemic cORR, CNS-cORR, and DORs Systemic cORRa Δ13.9% 83.9 (66.3–94.5) | 70.0 (50.6–85.3) T-DXd + P (n=31): CR 9.7, PR 74.2 THP (n=30): PR 70.0 CNS-cORRb Δ24.9% 51.6 (33.1–69.8) | 26.7 (12.3–45.9) T-DXd + P (n=31): CR 35.5, PR 16.1 THP (n=30): CR 23.3, PR 3.3 PR | CR Patients with response, % (95% CI) Responders with BICR-assessed BMs at BLc Response duration | T-DXd + P (n=31) | THP (n=30) Patients with objective response, n (%) | 26 (83.9) | 21 (70.0) Median DOR, mo (95% CI) | 35.1 (17.1–NE) | 10.5 (6.7–30.1) Patients with objective CNS response, n (%) | 16 (51.6) | 8 (26.7) Median CNS-DOR, mo (95% CI) | NE (14.3–NE) | 18.5 (8.3–NE) In patients with BICR-confirmed BMs at BL, T-DXd + P demonstrated deeper and more durable systemic responses versus THP; CNS-cORR was ~2-fold higher (51.6% vs 26.7%), and CNS-CR rates were 35.5% versus 23.3%, respectively Responses by BICR. Data cutoff: February 26, 2025 (median [range] follow-up: 23.6 mo [2.8–41.4]). Results should be interpreted with caution due to the low number of patients in the subgroup. [remaining small-print footnotes partly illegible] Cristina Saura DESTINY-Breast09

[Slide 3] Systemic cORR, CNS-cORR, and DORs Systemic cORRa Δ13.9% 83.9 (66.3–94.5) | 70.0 (50.6–85.3) T-DXd + P (n=31): CR 9.7, PR 74.2 THP (n=30): PR 70.0 CNS-cORRb Δ24.9% 51.6 (33.1–69.8) | 26.7 (12.3–45.9) T-DXd + P (n=31): CR 35.5, PR 16.1 THP (n=30): CR 23.3, PR 3.3 PR | CR Patients with response, % (95% CI) Responders with BICR-assessed BMs at BLc Response duration | T-DXd + P (n=31) | THP (n=30) Patients with objective response, n (%) | 26 (83.9) | 21 (70.0) Median DOR, mo (95% CI) | 35.1 (17.1–NE) | 10.5 (6.7–30.1) Patients with objective CNS response, n (%) | 16 (51.6) | 8 (26.7) Median CNS-DOR, mo (95% CI) | NE (14.3–NE) | 18.5 (8.3–NE) In patients with BICR-confirmed BMs at BL, T-DXd + P demonstrated deeper and more durable systemic responses versus THP; CNS-cORR was ~2-fold higher (51.6% vs 26.7%), and CNS-CR rates were 35.5% versus 23.3%, respectively Responses by BICR. Data cutoff: February 26, 2025 (median [range] follow-up: 23.6 mo [2.8–41.4]). Results should be interpreted with caution due to the low number of patients in the subgroup. aConfirmed systemic responses not captured in the ORR: T-DXd + P, SD (n=5, 16.1%), PD (n=0), NE (n=0); THP, SD (n=5, 16.7%), PD (n=3, 10.0%), NE (n=1, 3.3%). bConfirmed CNS responses not captured in the ORR: T-DXd + P, SD (n=12, 38.7%), PD (n=3, 9.7%), NE (n=0); THP, SD (n=13, 43.3%), PD (n=6, 20.0%), NE (n=3, 10.0%). cFor patients with both BICR- and INV-assessed BMs and BL, systemic cORR (95% CI) was 87.5% (61.7–98.4) with T-DXd + P versus 62.5% (35.4–84.8) with THP; CNS-cORR (95% CI) was 75.0% (47.6–92.7) with T-DXd + P versus 31.3% (11.0–58.7) with THP. (CNS-)CR, complete response (in CNS); PR, partial response; SD, stable disease. Cristina Saura DESTINY-Breast09 [Slide 4] Conclusions • In this exploratory analysis of DESTINY-Breast09, T-DXd + P demonstrated robust systemic and intracranial disease control versus standard-of-care THP, supporting a meaningful clinical benefit in patients with BMs at BL – Longer systemic PFS (HR 0.30) with T-DXd + P versus THP – Deeper systemic and CNS responses with T-DXd + P versus THP (systemic cORR: 83.9% vs 70.0%, respectively), with CNS-cORR ~2-fold higher with T-DXd + P versus THP (CNS-cORR: 51.6% vs 26.7%) – More durable systemic responses with T-DXd + P versus THP (median DOR: 35.1 vs 10.5 mo) • These findings further strengthen the body of evidence which supports the CNS activity of T-DXd in 1L HER2+ a/mBC, building on the efficacy observed in later lines of therapy from DESTINY-Breast12 and pooled analyses of DESTINY-Breast01/02/031,2 T-DXd + P demonstrated substantial and durable systemic and intracranial disease control versus THP, supporting its use as 1L therapy in HER2+ a/mBC 1. Harbeck N, et al. Nat Med. 2024;30:3717–3727. 2. André F, et al. Ann Oncol. 2024;35:1169–1180. Cristina Saura DESTINY-Breast09

1st HER2+ MBC – DB-09 for all? Ongoing studies Post - Cleopatra INAVO122/WO44263 trial | TT | ET | Anti HER2 Untreated HER2+ ABC and completed Induction therapy (4-8 cy taxane Pz/Tz) HER2+ and PIK3CA mut 1:1 Maintenance Therapy: Phesgo + placebo | Phesgo + inavolisib Long term Follow up: PD; Limiting toxicity heredERA trial | TT | ET | Anti HER2 Patients with ER+, HER2+ LA/mBC, previously untreated in the mBC setting N = 812 Induction: PH FDC SC + taxane 4-6 cycles R 1:1 N = 730 Maintenance: giredestrant 30 mg PO QD + PH FDC SC Q3W | PH FDC SC Q3W Until disease progression or unacceptable toxicity or withdrawal Stratification factors: Site of disease (visceral vs nonvisceral); Type of stage IV presentation (de novo metastatic vs recurrent metastatic disease); [illegible]; OR during the induction therapy phase (PR/CR vs SD [or non-CR/non-PD for participants with nonmeasurable disease]) Primary endpoint: Progression-free survival (as determined by the investigator according to RECIST v1.1, time from randomization to the first occurrence of disease progression or death) Secondary endpoints: Overall survival; Objective response rate; Duration of response; Clinical benefit rate; Patient-reported outcomes; Safety Post – DB-09 DEMETHER Induction therapy: T-DXd 6 cycles# | Maintenance therapy: PHESGO FDC SC* | Post-treatment FU: EoT# | Physician's choice# | EoS$ Baseline C2D1 | C1D1 C2D1 #PD, death, 3 years after T-DXd initiation, other anti-cancer treatment, or discontinuation for any other reason Destiny-Breast GUIDE T-DXd + pertuzumab (+/- ET) 18-24 cycles N = 200 → Trastuzumab + pertuzumab + ET + palbociclib until PD or unacceptable toxicity → PD → Retreatment with T-DXd post-progression ET: aromatase inhibitor, fulvestrant or tamoxifen. TO OPEN IN LATE 2026 Antonio Llombart-Cussac, MD, PhD

[Slide 3] Baseline demographics and disease characteristics Characteristic, n (%) unless otherwise specified | Patients with BICR-confirmed BMs at BL: T-DXd + P (n=31) | THP (n=30) | ITT population: T-DXd + P (n=383) | THP (n=387) Age, median (range), years | 56 (27–79) | 54 (26–71) | 54 (27–85) | 54 (20–81) Geographical region Asia | 17 (54.8) | 11 (36.7) | 188 (49.1) | 191 (49.4) Western Europe and North America | 5 (16.1) | 5 (16.7) | 87 (22.7) | 78 (20.2) Other geographic regions | 9 (29.0) | 14 (46.7) | 108 (28.2) | 118 (30.5) ECOG performance status 0 (normal activity) / 1 (restricted activity) | 20 (64.5) / 11 (35.5) | 18 (60.0) / 12 (40.0) | 256 (66.8) / 127 (33.2) | 246 (63.6) / 141 (36.4) HER2 score by central testa IHC 3+ | 28 (90.3) | 24 (80.0) | 318 (83.0) | 315 (81.4) IHC <3 / ISH+ | 3 (9.7) | 6 (20.0) | 62 (16.2) | 71 (18.3) HR status Positiveb / Negative | 16 (51.6) / 15 (48.4) | 14 (46.7) / 16 (53.3) | 207 (54.0) / 176 (46.0) | 209 (54.0) / 178 (46.0) De novo disease at diagnosis | 11 (35.5) | 7 (23.3) | 200 (52.2) | 200 (51.7) PIK3CAm detected | NA | NA | 116 (30.3) | 121 (31.3) Visceral metastases | 30 (96.8) | 28 (93.3) | 281 (73.4) | 268 (69.3) Prior radiotherapy to the brain | 7 (22.6)c | 6 (20.0)c | NA | NA Baseline demographics and disease characteristics of patients with BMs at BL were generally consistent with the ITT population, although rates of de novo disease were higher in the T-DXd + P arm versus the THP arm in the BM subgroup Data cutoff: February 26, 2025 (median [range] follow-up: 23.6 mo [2.8–41.4]). aIHC NR / ISH+ status: subgroup with BMs at BL, NA (both treatment arms); ITT population, 3 (0.8%) and 1 (0.3%) patients in the T-DXd + P and THP treatment arms, respectively. bDefined as ER-positive and/or PR-positive. cMedian (range) time from last radiotherapy to study treatment start date was 26.0 (20–43) days and 29.0 (15–90) days in the T-DXd + P and THP arms, respectively. ECOG, Eastern Cooperative Oncology Group; ER, estrogen receptor; IHC, immunohistochemistry; ISH+, in situ hybridization positive; NA, not available; NR, not recorded; PR, progesterone receptor. Cristina Saura DESTINY-Breast09

1st HER2+ MBC – Do all patients require DB-09? DB-09 Cleopatra "plus" PATINA – HER2CLIMB-05 Yes, DB-09 in all cases It will depend on the scenario LBA049: It is clinically reasonable to consider brain metastases as a factor favoring a DB09-based strategy LBA049: Reopen the question of routine baseline brain MRI to inform first-line treatment selection in HER2-positive MBC. Antonio Llombart-Cussac, MD, PhD

A DESTINY-Breast09 analysis of trastuzumab deruxtecan (T-DXd) + pertuzumab (P) vs taxane + trastuzumab + pertuzumab (THP) for patients with HER2+ advanced/metastatic breast cancer (a/mBC) and baseline (BL) brain metastases (BMs) Cristina Saura, Sibylle Loibl, ZeFei Jiang, Romualdo Barroso-Sousa, Yeon Hee Park, Mothaffar F Rimawi, Andreas Schneeweiss, Masakazu Toi, Keun Seok Lee, Yongkui Lu, Mehmet AN Sendur, Toshimi Takano, Quchung Ouyang, Annie Darilay, Jagdish Shetty, Angela Zeng, Sara M Tolaney October 26, 2026 Presentation LBA49 DESTINY Breast09 MADRID 2026 ESMO congress

Overall HRQoL Outcomes Were Comparable, With Domain-Specific Differences Between T-DXd and T-DM1 TDD,a median (95% CI), months: T-DXd (n = 818) | T-DM1 (n = 817) | Hazard Ratio (95% CI) EORTC QLQ-C30 GHS/QoL: 24.02 (22.64-NE) | 23.66 (22.64-25.26) | 1.15 (0.98-1.34) Pain symptoms: 24.11 (23.75-31.70) | 22.74 (18.07-24.48) | 0.77 (0.65-0.90) Physical functioning: 30.09 (30.09-NE) | 33.91 (NE-NE) | 1.17 (0.95-1.43) Emotional functioning: 26.25 (24.02-NE) | 27.93 (27.93-NE) | 1.02 (0.85-1.22) Social functioning: 24.15 (23.66-NE) | NE (26.58-NE) | 1.29 (1.08-1.54) EORTC QLQ-BR23 Arm symptoms: 23.13 (18.04-24.15) | 17.54 (12.22-21.98) | 0.82 (0.70-0.95) Breast symptoms: NE (26.48-NE) | NE (27.20-NE) | 0.83 (0.66-1.03) EQ-5D-5L VAS: 24.15 (23.36-29.80) | 23.85 (23.36-26.58) | 1.05 (0.89-1.24) [Forest plot axis 0.4-1.8; Favors T-DXd / Favors T-DM1] - There was no difference observed in TDD between T-DXd and T-DM1 across overall HRQoL, physical functioning, emotional functioning, breast symptoms, and EQ-5D-5L VAS - T-DXd was associated with a longer time to deterioration in pain and arm symptoms compared with T-DM1, while T-DM1 was associated with a longer time to deterioration in social function compared with T-DXd during the study period [Abbreviation footnotes - fine print, partly illegible] Dr. Carlos Barrios

[Slide 1] Conclusions - Overall, key HRQoL domains were maintained with both T-DXd and T-DM1 in patients with high-risk HER2-positive eBC who had residual invasive disease after neoadjuvant therapy - PRO assessments were comparable between treatments, with some differences in selected domains: o T-DXd was associated with a longer time to deterioration in pain and arm symptoms compared with T-DM1 during the study period o T-DM1 was associated with a longer time to deterioration in social function compared with T-DXd during the study period DESTINY-Breast05 HRQoL findings demonstrate preservation of patient-reported quality of life and, together with previously reported efficacy data, support T-DXd as a new standard of care in the post-neoadjuvant setting eBC, early breast cancer; HER2, human epidermal growth factor receptor 2; HRQoL, health-related quality of life; PRO, patient-reported outcome; T-DM1, trastuzumab emtansine; T-DXd, trastuzumab deruxtecan. Dr. Carlos Barrios [Slide 2] Overall HRQoL Outcomes Were Comparable, With Domain-Specific Differences Between T-DXd and T-DM1 TDD,a median (95% CI), months: T-DXd (n = 818) | T-DM1 (n = 817) | Hazard Ratio (95% CI) EORTC QLQ-C30 GHS/QoL: 24.02 (22.64-NE) | 23.66 (22.64-25.26) | 1.15 (0.98-1.34) Pain symptoms: 24.11 (23.75-31.70) | 22.74 (18.07-24.48) | 0.77 (0.65-0.90) Physical functioning: 30.09 (30.09-NE) | 33.91 (NE-NE) | 1.17 (0.95-1.43) Emotional functioning: 26.25 (24.02-NE) | 27.93 (27.93-NE) | 1.02 (0.85-1.22) Social functioning: 24.15 (23.66-NE) | NE (26.58-NE) | 1.29 (1.08-1.54) EORTC QLQ-BR23 Arm symptoms: 23.13 (18.04-24.15) | 17.54 (12.22-21.98) | 0.82 (0.70-0.95) Breast symptoms: NE (26.48-NE) | NE (27.20-NE) | 0.83 (0.66-1.03) EQ-5D-5L VAS: 24.15 (23.36-29.80) | 23.85 (23.36-26.58) | 1.05 (0.89-1.24) [Forest plot axis 0.4-1.8; Favors T-DXd <- -> Favors T-DM1] - There was no difference observed in TDD between T-DXd and T-DM1 across overall HRQoL, physical functioning, emotional functioning, breast symptoms, and EQ-5D-5L VAS - T-DXd was associated with a longer time to deterioration in pain and arm symptoms compared with T-DM1, while T-DM1 was associated with a longer time to deterioration in social function compared with T-DXd during the study period CI, confidence interval; EORTC QLQ-BR23, European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Breast Cancer Module 23; EORTC QLQ-C30, European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; GHS, global health status; HRQoL, health-related quality of life; QoL, quality of life; NE, not estimable; TDD, time to definitive deterioration; T-DM1, trastuzumab emtansine; T-DXd, trastuzumab deruxtecan; VAS, visual analog scale. TDD was analyzed using Kaplan-Meier methods and Cox proportional hazards models. aClinically meaningful definitive deterioration during the study period, defined as a change of >=10 points from baseline at either 2 or more consecutive time points, last PRO assessment, or death by the first survival follow-up visit. Dr. Carlos Barrios

[Slide 1] 16:15 - 17:45 Rapid oral: Breast cancer, early stage CHAIRS: MARIA VITTORIA DIECI, SANDRA XIMENA FRANCO MILLAN, LILIT HARUTYUNYAN DESTINY-Breast05 Study Design and Interim Results A global, multicenter, randomized, open-label phase 3 trial (NCT04622319) Population • Residual invasive disease in the breast and/or axillary lymph nodes after neoadjuvant chemotherapy with HER2-directed therapy (NAT)(a) • High risk of recurrence(b) • Centrally confirmed HER2+ (IHC 3+ and/or ISH+) eBC • ECOG PS 0 or 1 N = 1635 R 1:1 T-DXd 5.4 mg/kg IV Q3W for 14 cycles N = 818 T-DM1 3.6 mg/kg IV Q3W for 14 cycles N = 817 40-day safety follow-up Primary endpoint • IDFS by INV Key secondary endpoint • DFS by INV Other secondary endpoints • OS, DRFI by INV, BMFI by INV, safety Exploratory endpoints • HRQoL/PRO Interim Efficacy and Safety Results (DCO: July 02, 2025)(1,2,c) Post-neoadjuvant T-DXd demonstrated superior efficacy with no new safety signals in patients with high-risk HER2+ eBC and residual invasive disease after NAT,(1,2) and is now FDA approved(3) • IDFS events (T-DXd, 51 [6.2%] vs T-DM1, 102 [12.5%]) hazard ratio, 0.47 [95% CI, 0.34-0.66]; P <0.001) • 53% reduction in the risk of invasive disease recurrence or death • Estimated 3-year IDFS rates were 92.4% for T-DXd and 83.7% for T-DM1 The overall safety profile of T-DXd showed no new safety signals • In both treatment arms, >72% of patients had completed 14 treatment cycles(c); treatment had concluded for all patients in the study at DCO • Adjudicated drug-related ILD cases were mostly low grade and reversible with treatment management guidelines We report exploratory PRO results at interim analysis (DCO: July 02, 2025) BMFI, brain metastasis-free interval; CI, confidence interval; DCO, data cutoff; DFS, disease free-survival; DRFI, distant recurrence-free interval; eBC, early breast cancer; ECOG PS, Eastern Cooperative Oncology Group performance status; HER2, human epidermal growth factor receptor 2; HRQoL, health-related quality of life; IDFS, invasive disease-free survival; IHC, immunohistochemistry; ILD, interstitial lung disease; INV, investigator; ISH, in situ hybridization; IV, intravenous; NAT, neoadjuvant therapy; OS, overall survival; PRO, patient-reported outcome; Q3W, every 3 weeks; R, randomization; T-DM1, trastuzumab emtansine; T-DXd, trastuzumab deruxtecan. (a)NAT is defined as ≥16 weeks NAT with ≥9 weeks of trastuzumab ± pertuzumab and ≥9 weeks of taxane-based chemotherapy. (b)High-risk defined as Inoperable (cT4, N0-3, M0 or cT1-3, N2-3, M0) before NAT or Operable (cT1-3, N0-1, M0) before NAT with axillary node-positive disease (ypN1-3) after NAT. (c)Median study treatment duration: T-DXd, 9.8 months; T-DM1, 9.7 months. 1. Loibl S et al. N Engl J Med. 2026;394(9):845-857. 2. Geyer CE Jr et al. Presented at: ESMO Annual Congress; October 17-21, 2025; Berlin, Germany. 3. FDA approves two separate indications for fam-trastuzumab deruxtecan-nxki in HER2-positive early-stage breast cancer. Press release. May 15, 2026. Accessed July 10, 2026. Dr. Carlos Barrios [Slide 3] Overall HRQoL Was Preserved From Baseline With Both T-DXd and T-DM1 (EORTC QLQ-C30(a)) [three line charts: Mean (SD) CFB by cycle C2, C6, C10, C14; Better/Worse; ± 10-point change is considered clinically meaningful] GHS/QoL Physical Functioning Emotional Functioning Patients with CFB data,(c) n | C2 | C6 | C10 | C14 T-DXd (n = 818) | 715 | 591 | 545 | 456 T-DM1 (n = 817) | 701 | 619 | 562 | 480 (same patient counts shown under each of the three panels) Legend: T-DXd (solid) / T-DM1 (dashed) Mean change from baseline remained stable (within ± 10 points) over time across GHS/QoL, physical functioning, and emotional functioning in both treatment groups, supporting preservation of HRQoL during treatment C, cycle; CFB, change from baseline; EORTC QLQ-BR23, European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Breast Cancer Module 23; EORTC QLQ-C30, European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; EQ-5D-5L, EuroQol 5-Dimension 5-Level questionnaire; GHS, global health status; PRO, patient-reported outcome; QoL, quality of life; T-DM1, trastuzumab emtansine; T-DXd, trastuzumab deruxtecan; VAS, visual analog scale. (a)Higher scores in the EORTC QLQ-C30 GHS/QoL and functioning scales correspond to a higher/healthier level of functioning or QoL. (b)Day 1 of each cycle. (c)PRO data collected after new anti-cancer therapy, completion of 14 treatment cycles, or discontinuation of study drug are excluded. Dr. Carlos Barrios [Slide 4] Exploratory PRO Assessments Questionnaire | Description | Scales | Analyses EORTC QLQ-C30 | Oncology-specific questionnaire | • Global health status (GHS)/QoL • Functioning scales: physical, emotional, cognitive, and social • Symptom scales: pain, fatigue, nausea/vomiting • Single-item scales: dyspnea, sleep disturbance, appetite loss, constipation, diarrhea, and financial impact | • Change from baseline • Time to definitive deterioration (clinically meaningful definitive deterioration during the study period, defined as a change of ≥10 points from baseline at either 2 or more consecutive time points, last PRO assessment, or death by the first survival follow-up visit) EORTC QLQ-BR23 | Breast cancer-specific questionnaire | • Subscales: breast symptoms, arm symptoms, body image, sexual functioning, and systemic therapy side effects | • Change from baseline • Time to definitive deterioration EQ-5D-5L | Generic questionnaire | • Self-rated health status (VAS) | • Time to definitive deterioration PRO endpoints are reported for the following study time points(a): Cycle 1 (baseline) | Cycle 2 | Every 4 cycles (cycle 6, 10, 14) | End of treatment Completion(b) was comparable throughout treatment in both treatment groups, supporting interpretability of the longitudinal PRO analyses [abbreviations footnote] (a)PRO assessments began before infusion on day 1 of cycle 1; 1 cycle = 21 days. (b)The number of patients who completed at least one question from a questionnaire divided by the number of patients randomized within each treatment arm. Dr. Carlos Barrios

Breast Symptoms (EORTC QLQ-BR23) and Health Status (EQ-5D-5L VAS) Were Preserved in Both Treatment Groups EORTC QLQ-BR23 Breast symptoms^a (± 10-point change is considered clinically meaningful; Worse ↑ / Better ↓) Patients with CFB data,^d n T-DXd (n = 818): 699, 574, 533, 448 T-DM1 (n = 817): 692, 609, 557, 476 EQ-5D-5L VAS^a,b (Better ↑ / Worse ↓) Patients with CFB data,^d n T-DXd (n = 818): 698, 574, 533, 447 T-DM1 (n = 817): 686, 605, 556, 475 Cycles^c C2, C6, C10, C14; Mean (SD) CFB Legend: T-DXd, T-DM1 Mean change from baseline remained stable (within ± 10 points) over time across breast symptoms and EQ-5D-5L VAS scale, consistent with preservation of patient-reported health status, with both T-DXd and T-DM1 C, cycle; CFB, change from baseline; EORTC QLQ-BR23, ...; EQ-5D-5L, EuroQol 5-Dimension 5-Level questionnaire; GHS, global health status; PRO, patient-reported outcome; QoL, quality of life; T-DM1, trastuzumab emtansine; T-DXd, trastuzumab deruxtecan; VAS, visual analog scale. ^aHigher scores on the EORTC QLQ-BR23 symptom scales indicate a higher level of symptomatology (ie, a worse state of health for the patient); higher scores on the EQ 5D-5L VAS correspond to better health states. ^bMeasured on a 20 cm vertical VAS with endpoints labelled "the best health you can imagine" and "the worst health you can imagine." ^cDay 1 of each cycle. ^dPRO data collected after new anti-cancer therapy, completion of 14 treatment cycles, or discontinuation of study drug are excluded. Dr. Carlos Barrios

16:15 - 17:45 Rapid oral: Breast cancer, early stage CHAIRS: MARIA VITTORIA DIECI, SANDRA XIMENA FRANCO MILLAN, LILIT HARUTYUNYAN Exploratory PRO Assessments Questionnaire | Description | Scales | Analyses EORTC QLQ-C30 | Oncology-specific questionnaire | • Global health status (GHS)/QoL • Functioning scales: physical, emotional, cognitive, and social • Symptom scales: pain, fatigue, nausea/vomiting • Single-item scales: dyspnea, sleep disturbance, appetite loss, constipation, diarrhea, and financial impact | • Change from baseline • Time to definitive deterioration (clinically meaningful definitive deterioration during the study period, defined as a change of ≥10 points from baseline at either 2 or more consecutive time points, last PRO assessment, or death by the first survival follow-up visit) EORTC QLQ-BR23 | Breast cancer–specific questionnaire | • Subscales: breast symptoms, arm symptoms, body image, sexual functioning, and systemic therapy side effects | • Change from baseline • Time to definitive deterioration EQ-5D-5L | Generic questionnaire | • Self-rated health status (VAS) | • Time to definitive deterioration PRO endpoints are reported for the following study time points(a): Cycle 1 (baseline) | Cycle 2 | Every 4 cycles (cycle 6, 10, 14) | End of treatment Completion(b) was comparable throughout treatment in both treatment groups, supporting interpretability of the longitudinal PRO analyses EORTC QLQ-BR23, European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Breast Cancer Module 23; EORTC QLQ-C30, European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; EQ-5D-5L, EuroQol 5-Dimension 5-Level questionnaire; GHS, global health status; HR, hormone receptor; HRQoL, health-related quality of life; PRO, patient-reported outcome; QoL, quality of life; TDD, time to definitive deterioration; VAS, visual analog scale. (a)PRO assessments began before infusion on day 1 of cycle 1; 1 cycle = 21 days. (b)The number of patients who completed at least one question from a questionnaire divided by the number of patients randomized within each treatment arm. Dr. Carlos Barrios Speaker panel: Carlos Barrios — Patient-reported outcomes (PROs) from DESTINY-Breast05: Health-related quality of life (HRQoL) with trastuzumab Deruxtecan (T-DXd) versus trastuzumab emtansine (T-DM1) in patients (Pts) with high-risk HER2-positive early breast cancer (eBC) ALICANTE AUDITORIUM - HALL 6

[Slide 1] CCTG MA.40/FINER Trial Design Study population: • Stage IV ER+/HER-2 negative breast cancer • Prior line of systemic treatment with CDK 4/6i and AI • ECOG PS 0–1 • Measurable or evaluable disease CDK 4/6i and AI progression → randomization ≦ 4 weeks cfDNA for Stratification (FoundationOne®Liquid CDx) N=250 R 1:1 Ipatasertib + Fulvestrant 400 mg PO day 1-21 Stratification • PIK3CA/PTEN/AKT altered vs wild type/unknown • Primary versus secondary endocrine resistance Placebo + Fulvestrant Placebo PO day 1-21 Primary endpoint • PFS by RECIST v1.1 in intent to treat (ITT) population Key secondary endpoints • PFS - by BICR in ITT - in AKT pathway altered cohort • OS • AEs • QoL - Health Economics - Correlative Today we will present the updated OS analysis results from the ITT and biomarker altered cohort Andrew Redfern [Slide 2] Results: Progression-free Survival in ITT Cohort Outcome | Ipatasertib + fulvestrant | placebo + fulvestrant Median PFS months, (95% CI) | 5.36 (3.71–6.37) | 1.94 (1.84–3.48) Hazard Ratio (95% CI), p | 0.68 (0.52–0.88), p=0.0036 232 events Legend: Ipatasertib (blue), Placebo (red); y-axis Probability of Progression-free Survival (%); x-axis Time (months) 0-48 At risk (Ipatasertib): 125 51 35 16 9 4 3 2 0 At risk (Placebo): 125 32 13 10 6 3 1 0 0 Andrew Redfern Survival table Endpoint: PFS, ITT cohort (primary) Arm | N | Median | Median CI Ipatasertib + fulvestrant | 125 | 5.36 mo | 3.71–6.37 Placebo + fulvestrant | 125 | 1.94 mo | 1.84–3.48 HR: 0.68 HR CI: 0.52–0.88 P: 0.0036 [Slide 3] Results: PFS in AKT Pathway-Altered Cohort Outcome | Ipatasertib + fulvestrant | placebo + fulvestrant Median PFS months, (95% CI) | 5.45 (3.55–9.13) | 1.92 (1.81–3.48) Hazard Ratio (95% CI), p | 0.55 (0.37–0.81), p=0.0023 Legend: Ipatasertib (blue), Placebo (red); y-axis Probability of Progression-free Survival (%); x-axis Time (months) 0-48 At risk (Ipatasertib): 57 26 18 8 5 1 0 0 0 At risk (Placebo): 54 11 4 3 1 0 0 0 0 Andrew Redfern Survival table Endpoint: PFS, AKT pathway-altered cohort Arm | N | Median | Median CI Ipatasertib + fulvestrant | 57 | 5.45 mo | 3.55–9.13 Placebo + fulvestrant | 54 | 1.92 mo | 1.81–3.48 HR: 0.55 HR CI: 0.37–0.81 P: 0.0023 [Slide 4] Adverse events possibly attributed to ipatasertib with > 10% frequency Placebo + fulvestrant (n=124) | Ipatasertib + fulvestrant (n=124) Total (%) / Grade 3 (%) [placebo | ipatasertib] Diarrhea: 37.9/0.0 | 92.7/16.9 Nausea: 47.6/0.8 | 79.8/0.8 Fatigue: 63.7/0.8 | 67.7/4.0 Vomiting: 21.8/0.8 | 50.0/1.6 Hyperglycemia: 33.1/0.0 | 45.2/0.0 Headache: 33.9/0.0 | 40.3/0.8 Anorexia: 20.2/0.0 | 32.3/0.0 Rash maculo-papular: 12.1/0.0 | 29.8/1.6 Mucositis oral: 9.7/0.0 | 21.0/0.0 Pruritus: 9.7/0.0 | 16.1/0.0 Incidence axis: 80% 60% 40% 20% 0% 20% 40% 60% 80% 100% Legend: Total (%) / Grade 3 (%) Placebo: grade 1, grade 2, grade 3; Ipatasertib: grade 1 (green), grade 2 (blue), grade 3 (yellow) Andrew Redfern

LBA14 - A double-blind placebo controlled randomized phase III trial of fulvestrant and ipatasertib as treatment for HER-2 negative and estrogen receptor positive (ER+) metastatic breast cancer (MBC): Overall survival results of the CCTG/BCT MA.40/FINER study Speakers Andrew D. Redfern (Perth, Australia) Lecture Time 08:40 - 08:50 Abstract Background The MA.40/FINER Study compared the efficacy and safety of the pan-AKT inhibitor ipatasertib versus (vs) placebo plus fulvestrant in patients with ER+/HER2-negative MBC immediately following progression on a CDK4/6 inhibitor (CDK4/6i) plus aromatase inhibitor (AI). Both the primary objective comparing investigator-assessed progression-free survival (PFS) between treatment arms in the ITT population and key secondary objective comparing PFS in the AKT pathway-altered cohort demonstrated statistically and clinically significant improvements. Here we report on the overall survival (OS) analysis. Methods Pre-, peri-, and postmenopausal women and men with ER+/HER-2-negative MBC were randomly assigned 1:1 to receive ipatasertib (400mg PO daily, days 1-21) plus fulvestrant (500mg IM, days 1 and 15, cycle 1 then day 1 of subsequent cycles) vs placebo plus fulvestrant. Randomization was stratified by AKT pathway altered (PIK3CA/AKT1/PTEN genomic alteration determined by cfDNA NGS assay) vs wild type/unknown and primary (<6 months (mo) CDK4/6i) vs secondary (>=6 mo CDK4/6i) endocrine resistance. OS was a secondary endpoint. Results 250 participants were randomized to ipatasertib (125) or placebo (125). At a median follow-up of 35.0 mo, 3.2% ipatasertib vs 2.4% placebo arm participants remained on protocol treatment. Median OS for ITT populations were 26.68 mo (95% CI: 21.42 to 30.78 mo) vs 25.76 mo (95% CI: from 19.98 to 28.88 mo) for ipatasertib vs placebo arms (HR = 0.97, 95% C.I. 0.70 to 1.33, p = 0.83). In an exploratory analysis, median OS for AKT pathway-altered cohorts were 23.66 mo (95% CI: 18.53 to 28.65 mo) vs 25.23 mo (95% CI: 16.99 to 28.88 mo) [HR = 1.14 95% CI: 0.72 to 1.80; p=0.58]. No new safety signals were observed with longer duration follow-up. Conclusions Despite meaningful PFS improvements, the addition of ipatasertib to fulvestrant for the second-line treatment of ER+/HER-2-negative MBC after CDK4/6i plus AI did not impact OS for either total or AKT pathway-altered patient cohorts.

[Slide 2] Results: Overall Survival in ITT Cohort Outcome | Ipatasertib + fulvestrant | placebo + fulvestrant Median OS months, (95% CI) | 26.68 (21.42-30.78) | 25.76 (19.598-28.88) Hazard Ratio (95% CI) | 0.97 (0.70-1.33), p=0.83 156 events Legend: Ipatasertib (blue); Placebo (red) Y-axis: Probability of Overall Survival (%); X-axis: Time (months) 0-60 At risk (Ipatasertib): 125 | 115 | 99 | 81 | 60 | 30 | 20 | 15 | 7 | 1 | 0 (months 0,6,12,18,24,30,36,42,48,54,60) At risk (Placebo): 125 | 106 | 95 | 79 | 57 | 31 | 20 | 10 | 4 | 2 | 0 Andrew Redfern Survival table Endpoint: OS (ITT) Arm | N | Median | Median CI Ipatasertib + fulvestrant | 125 | 26.68 mo | 21.42-30.78 Placebo + fulvestrant | 125 | 25.76 mo | 19.598-28.88 HR: 0.97 HR CI: 0.70-1.33 P: 0.83 [Slide 3] Results: Overall Survival in AKT Pathway-Altered Cohort Outcome | Ipatasertib + fulvestrant | placebo + fulvestrant Median OS months, (95% CI) | 23.66 (18.53-28.65) | 25.23 (16.99-28.88) Hazard Ratio (95% CI), p | 0.88 (0.56-1.39) Legend: Ipatasertib (blue); Placebo (red) Y-axis: Probability of Overall Survival (%); X-axis: Time (months) 0-60 At risk (Ipatasertib): 57 | 53 | 47 | 37 | 25 | 12 | 7 | 4 | 1 | 0 | 0 (months 0,6,12,18,24,30,36,42,48,54,60) At risk (Placebo): 54 | 47 | 43 | 34 | 27 | 12 | 8 | 4 | 1 | 1 | 0 Andrew Redfern Survival table Endpoint: OS (AKT pathway-altered cohort) Arm | N | Median | Median CI Ipatasertib + fulvestrant | 57 | 23.66 mo | 18.53-28.65 Placebo + fulvestrant | 54 | 25.23 mo | 16.99-28.88 HR: 0.88 HR CI: 0.56-1.39 [Slide 4] Conclusions The ipatasertib and fulvestrant combination is generally tolerable with a low rate of discontinuation due to toxicity. The combination provides statistically significant and clinically meaningful improvements in PFS for both ITT and AKT pathway altered cohorts. No improvement in overall survival was observed in either cohort although power was low to detect modest differences. Further research looking at mechanisms of resistance to AKT inhibitors may allow disease control benefits to be translated into longer overall survival. Andrew Redfern

Results: Overall Survival in ITT Cohort Outcome | Ipatasertib + fulvestrant | placebo + fulvestrant Median OS months, (95% CI) | 26.68 (21.42–30.78) | 25.76 (19.598–28.88) Hazard Ratio (95% CI) | 0.97 (0.70–1.33), p=0.83 156 events Legend: Ipatasertib (blue), Placebo (red); y-axis Probability of Overall Survival (%); x-axis Time (months) 0-60 At risk (Ipatasertib): 125 115 99 81 60 30 20 15 7 1 0 At risk (Placebo): 125 106 95 79 57 31 20 10 4 2 0 Andrew Redfern Survival table Endpoint: OS, ITT cohort Arm | N | Median | Median CI Ipatasertib + fulvestrant | 125 | 26.68 mo | 21.42–30.78 Placebo + fulvestrant | 125 | 25.76 mo | 19.598–28.88 HR: 0.97 HR CI: 0.70–1.33 P: 0.83

[Slide 1] 08:30 - 10:00 Proffered paper 2: Breast cancer, early and advanced stage CHAIRS: SARA HURVITZ, SEOCK-AH IM Key Takeaway Points • At a median follow-up of 35.0 months, the FINER trial did not demonstrate a prolongation of overall survival in either the ITT or AKT pathway-altered cohorts with the addition of ipatasertib to fulvestrant (secondary endpoint). • In contrast, the previously observed prolongation of progression free-survival persisted, remaining significant in both overall ITT and AKT pathway-altered cohorts. • No new safety signals were observed with longer duration follow-up, with <10% of participants discontinuing therapy due to adverse events. • Targeting the AKT pathway with ipatasertib in advanced, ER+/HER2-breast cancer is feasible and associated with meaningful disease control. Andrew Redfern [Speaker video panel] Andrew Redfern A double-blind placebo controlled randomized phase III trial of fulvestrant and ipatasertib as treatment for HER-2 negative and estrogen receptor positive (ER+) metastatic breast cancer (MBC): Overall survival results of the CCTG/BCT MA.40/FINER study ALICANTE AUDITORIUM - HALL 6 [Slide 2] 08:30 - 10:00 Proffered paper 2: Breast cancer, early and advanced stage CHAIRS: SARA HURVITZ, SEOCK-AH IM Results: Overall Survival in ITT Cohort Outcome | Ipatasertib + fulvestrant | placebo + fulvestrant Median OS months, (95% CI) | 26.68 (21.42-30.78) | 25.76 (19.598-28.88) Hazard Ratio (95% CI) | 0.97 (0.70-1.33), p=0.83 156 events Legend: Ipatasertib (blue); Placebo (red) Y-axis: Probability of Overall Survival (%); X-axis: Time (months) 0-60 At risk (Ipatasertib): 125 | 115 | 99 | 81 | 60 | 30 | 20 | 15 | 7 | 1 | 0 (months 0,6,12,18,24,30,36,42,48,54,60) At risk (Placebo): 125 | 106 | 95 | 79 | 57 | 31 | 20 | 10 | 4 | 2 | 0 Andrew Redfern [Speaker video panel] Andrew Redfern - A double-blind placebo controlled randomized phase III trial of fulvestrant and ipatasertib as treatment for HER-2 negative and estrogen receptor positive (ER+) metastatic breast cancer (MBC): Overall survival results of the CCTG/BCT MA.40/FINER study ALICANTE AUDITORIUM - HALL 6 Survival table Endpoint: OS (ITT) Arm | N | Median | Median CI Ipatasertib + fulvestrant | 125 | 26.68 mo | 21.42-30.78 Placebo + fulvestrant | 125 | 25.76 mo | 19.598-28.88 HR: 0.97 HR CI: 0.70-1.33 P: 0.83 [Slide 3] 08:30 - 10:00 Proffered paper 2: Breast cancer, early and advanced stage CHAIRS: SARA HURVITZ, SEOCK-AH IM Results: Overall Survival in AKT Pathway-Altered Cohort Outcome | Ipatasertib + fulvestrant | placebo + fulvestrant Median OS months, (95% CI) | 23.66 (18.53-28.65) | 25.23 (16.99-28.88) Hazard Ratio (95% CI), p | 0.88 (0.56-1.39) Legend: Ipatasertib (blue); Placebo (red) Y-axis: Probability of Overall Survival (%); X-axis: Time (months) 0-60 At risk (Ipatasertib): 57 | 53 | 47 | 37 | 25 | 12 | 7 | 4 | 1 | 0 | 0 (months 0,6,12,18,24,30,36,42,48,54,60) At risk (Placebo): 54 | 47 | 43 | 34 | 27 | 12 | 8 | 4 | 1 | 1 | 0 Andrew Redfern [Speaker video panel] Andrew Redfern - A double-blind placebo controlled randomized phase III trial of fulvestrant and ipatasertib as treatment for HER-2 negative and estrogen receptor positive (ER+) metastatic breast cancer (MBC): Overall survival results of the CCTG/BCT MA.40/FINER study ALICANTE AUDITORIUM - HALL 6 Survival table Endpoint: OS (AKT pathway-altered cohort) Arm | N | Median | Median CI Ipatasertib + fulvestrant | 57 | 23.66 mo | 18.53-28.65 Placebo + fulvestrant | 54 | 25.23 mo | 16.99-28.88 HR: 0.88 HR CI: 0.56-1.39 [Slide 4] 08:30 - 10:00 Proffered paper 2: Breast cancer, early and advanced stage CHAIRS: SARA HURVITZ, SEOCK-AH IM Results: Time to Commencement of Subsequent Line of Systemic Therapy (TSST) or Death in ITT Cohort Outcome | Ipatasertib + fulvestrant | placebo + fulvestrant Median TSST months, (95% CI) | 6.24 (4.80-8.11) | 3.75 (2.83-5.03) Hazard Ratio (95% CI), p | 1.46 (1.13-1.88), p=0.0037 Legend: Ipatasertib (blue); Placebo (red) Y-axis: Probability of New Therapy Start or Death (%); X-axis: Time (months) 0-48 At risk (Ipatasertib): 125 | 65 | 37 | 20 | 9 | 4 | 2 | 1 | 0 (months 0,6,12,18,24,30,36,42,48) At risk (Placebo): 125 | 44 | 14 | 8 | 6 | 3 | 3 | 0 | 0 Andrew Redfern [Speaker video panel] Andrew Redfern - A double-blind placebo controlled randomized phase III trial of fulvestrant and ipatasertib as treatment for HER-2 negative and estrogen receptor positive (ER+) metastatic breast cancer (MBC): Overall survival results of the CCTG/BCT MA.40/FINER study ALICANTE AUDITORIUM - HALL 6 Survival table Endpoint: Time to subsequent systemic therapy or death (TSST), ITT Arm | N | Median | Median CI Ipatasertib + fulvestrant | 125 | 6.24 mo | 4.80-8.11 Placebo + fulvestrant | 125 | 3.75 mo | 2.83-5.03 HR: 1.46 HR CI: 1.13-1.88 P: 0.0037

[Slide 1] evERA BC exploratory analysis: Assessing the relationship between the genomic landscape at baseline and clinical outcomes Giredestrant + everolimus demonstrated statistically significant improvements in median INV-PFS vs SOC ET + everolimus in both the overall study and ESR1m populations:1 • Overall study population: HR 0.56; 8.77 vs 5.49 months • ESR1m population: HR 0.38; 9.99 vs 5.45 months Consistent benefit of giredestrant + everolimus in patients with PIK3CAm and PI3K pathway alterations has been reported previously1,2 Exploratory biomarker analysis: Methodology Enrolled patients1: • ER+, HER2- aBC (1L-3L of therapy; enriched to 55% ESR1m*) • ≤2 prior lines of ET in the aBC setting • PD or relapse during/post-CDK4/6i + ET • No prior chemotherapy in the aBC setting N = 373 Plasma collection: Baseline ctDNA collected for F1LCDx to determine ESR1m status (evaluable for 365 enrolled patients) Biomarker assessments: cTF† - Overall tumour burden ESR1 landscape - Assessment of variant prevalence, multiplicity and clonality Broad genomic landscape - Identification of key oncogenic mutations PFS analyses‡ -> Evaluation of giredestrant + everolimus benefit consistency across biomarkers ClinicalTrials.gov number, NCT05306340. * Centrally tested at baseline via ctDNA. † Proportion of ctDNA to total cell-free DNA in the blood. ‡ Data shown are INV-PFS; labelled as PFS for simplicity. 1L-3L, first to third line; (a)BC, (advanced) breast cancer; CDK4/6i, cyclin-dependent kinase 4/6 inhibitor; ctDNA, circulating tumour DNA; cTF, circulating tumour fraction; ER+, oestrogen receptor-positive; F1LCDx, FoundationOne®Liquid CDx; HER2-, HER2-negative; HR, hazard ratio; (INV)-PFS, (investigator-assessed) progression-free survival; m, mutation; PD, progressive disease; (SOC) ET, (standard-of-care) endocrine therapy. 1. Mayer EL, et al. N Engl J Med 2026; 395:1285-1298; 2. Rugo HS, et al. SABCS 2025 (oral GS3-09). Presented by Sara M Tolaney, MD, MPH. [Slide 2] PFS benefit was greater with giredestrant + everolimus vs SOC ET + everolimus in the overall study and ESR1m populations, regardless of cTF status Overall study population* [KM curves: Giredestrant + everolimus, cTF <1%; Giredestrant + everolimus, cTF ≥1%; SOC ET + everolimus, cTF <1%; SOC ET + everolimus, cTF ≥1%] PFS (%) / Months No. at risk (0 3 6 9 12 15 18 21 24 27): Giredestrant + everolimus, cTF <1%: 60 51 34 31 22 15 7 2 2 1 Giredestrant + everolimus, cTF ≥1%: 116 92 58 46 23 13 7 1 1 0 SOC ET + everolimus, cTF <1%: 61 45 26 20 16 12 7 3 3 1 SOC ET + everolimus, cTF ≥1%: 126 83 40 29 10 2 0 0 0 0 cTF | Aggregated mPFS (months) | Giredestrant + everolimus mPFS (months) | SOC ET + everolimus mPFS (months) | HR (95% CI) <1% | 8.08 | 11.24 | 5.59 | 0.61 (0.40, 0.94) ≥1% | 5.59 | 7.39 | 4.60 | 0.52 (0.39, 0.70) ESR1m population† [KM curves, same 4 groups] No. at risk (0 3 6 9 12 15 18 21 24 27): Giredestrant + everolimus, cTF <1%: 21 19 15 14 8 4 2 1 1 0 Giredestrant + everolimus, cTF ≥1%: 80 65 45 37 19 9 4 1 1 0 SOC ET + everolimus, cTF <1%: 25 16 9 7 4 1 1 1 1 0 SOC ET + everolimus, cTF ≥1%: 80 51 26 18 6 1 0 0 0 0 cTF | Aggregated mPFS (months) | Giredestrant + everolimus mPFS (months) | SOC ET + everolimus mPFS (months) | HR (95% CI) <1% | 8.08 | 12.94 | 5.52 | 0.32 (0.16, 0.68) ≥1% | 6.05 | 9.26 | 5.45 | 0.45 (0.31, 0.65) In evaluable patients (n = 363), cTF was well balanced between arms; overall, 67% had high (≥1%) cTF *Among patients with evaluable ctDNA in the overall study population, 60 of 176 patients (34.1%) in the giredestrant + everolimus arm and 61 of 187 patients (32.6%) in the SOC ET + everolimus arm had a cTF <1%. †Among patients with evaluable ctDNA in the ESR1m population, 21 of 101 patients (20.8%) in the giredestrant + everolimus arm and 25 of 105 patients (23.8%) in the SOC ET + everolimus arm had a cTF <1%. HR estimates are unstratified. CI, confidence interval; ctDNA, circulating tumour DNA; cTF, circulating tumour fraction; ESR1m, ESR1 mutation; HR, hazard ratio; (m)PFS, (median) progression-free survival; SOC ET, standard-of-care endocrine therapy. Presented by Sara M Tolaney, MD, MPH. Survival table Endpoint: INV-PFS by circulating tumour fraction (cTF) Overall - Giredestrant + everolimus, cTF <1% | 60 | 11.24 mo Overall - SOC ET + everolimus, cTF <1% | 61 | 5.59 mo Overall - Giredestrant + everolimus, cTF ≥1% | 116 | 7.39 mo Overall - SOC ET + everolimus, cTF ≥1% | 126 | 4.60 mo ESR1m - Giredestrant + everolimus, cTF <1% | 21 | 12.94 mo ESR1m - SOC ET + everolimus, cTF <1% | 25 | 5.52 mo ESR1m - Giredestrant + everolimus, cTF ≥1% | 80 | 9.26 mo ESR1m - SOC ET + everolimus, cTF ≥1% | 80 | 5.45 mo HR: Overall cTF<1% 0.61; Overall cTF≥1% 0.52; ESR1m cTF<1% 0.32; ESR1m cTF≥1% 0.45 HR CI: 0.40-0.94; 0.39-0.70; 0.16-0.68; 0.31-0.65 [Slide 3] PFS benefit with giredestrant + everolimus was observed independent of ESR1m clonality, and across ESR1 variants ESR1 heterogeneity was common: 34% of patients harboured >1 ESR1 variant, and most had ≥1 clonal variant despite subclonal heterogeneity [Bar chart: Frequency (n) vs No. of ESR1 variants per patient (1-9); All variants subclonal / ≥1 clonal variant] The most common ESR1 variants were at Y537 (in 60% of patients) and D538 (in 57% of patients) [Lollipop plot: Observed no. per variant; ERα domains with specific residues of alterations: AF-1 domain, DBD, LBD/AF-2; residues 1, 185, 250, 312, 546, 595; E380, S463, L536, Y537, D538] Patients (n): Total | Giredestrant + everolimus | SOC ET + everolimus | PFS HR (95% CI) ESR1m ESR1m, multiple | 71 | 35 | 36 | 0.48 (0.28, 0.84) ESR1m, single | 136 | 67 | 69 | 0.38 (0.25, 0.58) ESR1m by clonality ≥1 clonal variant | 119 | 51 | 68 | 0.39 (0.25, 0.61) All variants subclonal | 87 | 51 | 36 | 0.44 (0.26, 0.74) Y537m | 124 | 60 | 64 | 0.35 (0.22, 0.54) D538m | 117 | 56 | 61 | 0.58 (0.38, 0.89) L536m | 26 | 14 | 12 | 0.67 (0.28, 1.63) E380m | 26 | 16 | 10 | 0.30 (0.11, 0.83) [Forest plot axis 0.1 - 1.0 - 3.0; Giredestrant + everolimus better | SOC ET + everolimus better] The PFS benefit was consistent across ESR1 variants, and irrespective of ESR1m clonality status and residue mutated Patients with "single" ESR1m were those where only one residue was mutated, regardless of the number of variants at that residue. Patients with "multiple" ESR1m were those with mutations at more than one residue. Some variants at relevant residues (e.g., Y537D) are known pathogenic variants but were not included in the biomarker definition for evERA BC; these have been included in the analysis. HR estimates are unstratified. AF-1, activation function-1; AF-2, activation function-2; BC, breast cancer; CI, confidence interval; DBD, DNA-binding domain; ERα, oestrogen receptor alpha; HR, hazard ratio; LBD, ligand-binding domain; m, mutation; PFS, progression-free survival; SOC ET, standard-of-care endocrine therapy. Presented by Sara M Tolaney, MD, MPH. [Slide 4] PFS benefit was observed across key poor-prognostic genomic subgroups in the overall study and ESR1m populations Overall study population - Patients (n): Total | Giredestrant + everolimus | SOC ET + everolimus | PFS HR (95% CI) Gene alterations: TP53m | 152 | 81 | 71 | 0.65 (0.45, 0.93) TP53nmd | 213 | 97 | 116 | 0.49 (0.36, 0.68) GATA3m | 65 | 31 | 34 | 0.80 (0.48, 1.35) GATA3nmd | 300 | 147 | 153 | 0.53 (0.41, 0.70) RB1m | 64 | 36 | 28 | 0.65 (0.38, 1.12) RB1nmd | 301 | 142 | 159 | 0.54 (0.41, 0.70) CDH1m | 45 | 20 | 25 | 0.81 (0.41, 1.57) CDH1nmd | 320 | 158 | 162 | 0.54 (0.42, 0.69) Pathway alterations: MAPKm | 80 | 36 | 44 | 0.98 (0.62, 1.57) MAPKnmd | 285 | 142 | 143 | 0.50 (0.38, 0.66) ESR1m population - Patients (n): Total | Giredestrant + everolimus | SOC ET + everolimus | PFS HR (95% CI) TP53m | 77 | 42 | 35 | 0.54 (0.31, 0.92) TP53nmd | 130 | 60 | 70 | 0.34 (0.22, 0.52) GATA3m | 41 | 20 | 21 | 0.58 (0.29, 1.14) GATA3nmd | 166 | 82 | 84 | 0.40 (0.28, 0.58) RB1m | 38 | 24 | 14 | 0.56 (0.27, 1.17) RB1nmd | 169 | 78 | 91 | 0.39 (0.27, 0.56) CDH1m | 26 | 12 | 14 | 0.98 (0.39, 2.46) CDH1nmd | 181 | 90 | 91 | 0.35 (0.24, 0.50) MAPKm | 49 | 21 | 28 | 0.55 (0.29, 1.03) MAPKnmd | 158 | 81 | 77 | 0.40 (0.27, 0.59) [Forest plot axes 0.1 - 1.0 - 3.0; Giredestrant + everolimus better | SOC ET + everolimus better] Genes were considered mutated if a known or likely pathogenic variant was detected. The MAPK pathway was defined by the following genes: BRAF, KRAS, NF1, MAP2K4, MAP3K13, RAF1, MAP2K2, HRAS, MYC, NFKBIA, MAPK1, MAP2K1, JUN and RAC1. The pathway was considered mutated if a patient tested positive for one or more mutations in the pathway gene list. HR estimates are unstratified. PIK3CA and PI3K pathway subgroup analyses have been previously described and also showed consistent PFS benefit.1,2 CI, confidence interval; HR, hazard ratio; m, mutation; nmd, no mutation detected; PFS, progression-free survival; SOC ET, standard-of-care endocrine therapy. 1. Mayer EL, et al. N Engl J Med 2026; 395:1285-1298; 2. Rugo HS, et al. SABCS 2025 (oral GS3-09). Presented by Sara M Tolaney, MD, MPH.

Key Observations/Questions ESR1m, post CDK4/6i Endocrine Therapies Trial | Regimen | ESR1m mPFS (mo) | HR | Comparator evERA | Giredestrant + everolimus | 10.0 vs 5.5 | 0.38 | SOC ET + everolimus EMBER-3 | Imlunestrant + abemaciclib | ~9.4 vs 5.5 (all-comers) | 0.57 | Imlunestrant alone EMBER-3 | Imlunestrant alone | 5.5 vs 3.8 | 0.62 | SOC ET VERITAC-2 | Vepdegestrant | 5.0 vs 2.1 | 0.57 | Fulvestrant EMERALD | Elacestrant | 3.8 vs 1.9 | 0.55 | SOC ET - Different populations! - Focus on control arm - HR for evERA 0.38 earned against a real combination; other trials with ESR1m tumors largely resistant to control arm [arrow annotation to evERA 10.0 vs 5.5] evERA enrichced for RB1 and TP53 so poor prognosis group [arrow annotation to evERA 10.0 vs 5.5] There is no Giredestrant only arm so difficult to determine if "Better SERD" from SERD plus mTOR synergy Bill Gradishar MD

The baseline genomic landscape of patients in evERA BC was characteristic of a CDK4/6i-pretreated population1,2 evERA Breast Cancer Mutation prevalence (%) [bar chart, y-axis 0-60] ESR1 57% | TP53 42% | DNMT3A* 32% | PIK3CA 32% | GATA3 18% | RB1 18% | FGF3 15% | FGF4 15% | CCND1 15% | FGF19 15% | CDH1 12% | ATM* 12% | FGFR1 12% | PTEN 11% | TET2* 11% | WHSC1L1 11% | ZNF703 10% | CHEK2* 10% | ARID1A 8% | MYC 8% | AKT1 7% | NF1 7% | BRCA2 6% | GNAS 6% | ASXL1* 5% | EMSY 5% | MAP3K1 5% | SF3B1 5% | ERBB2 4% The most common non-ESR1 genomic alterations were in TP53, PIK3CA, GATA3 and RB1 Per protocol, the evERA BC patient population was enriched for ESR1m†,3 ESR1m was defined as the presence of >= 1 of the following variants: E380Q, S463P, P535H, L536H/P/Q/R, Y537C/N/S or D538G. PIK3CA mutations included activating mutations. AKT1 mutations were limited to E17K. All other genes were detected via any known/likely mutation including short variants, copy number alterations or rearrangements. * Common Clonal Hematopoiesis of Indeterminate Potential (CHIP) mutation. † Centrally tested at baseline via ctDNA. BC, breast cancer; CDK4/6i, cyclin-dependent kinase 4/6 inhibitor; ctDNA, circulating tumour DNA. 1. Chaudhary N, et al. NPJ Breast Cancer 2024; 10:15; 2. Pontolillo L, et al. NPJ Breast Cancer 2025; 11:93; 3. Mayer EL, et al. N Engl J Med 2026; 395:1285-1298. Presented by Sara M Tolaney, MD, MPH.

Conclusions: Exploratory analysis of evERA BC Giredestrant + everolimus Genomic alterations: Generally balanced between arms. Consistent with previously described datasets from CDK4/6i-pretreated populations1,2 Circulating tumour fraction: cTF ≥ 1% negative prognostic feature. 67% of patients had a cTF ≥ 1% ESR1 alterations: High prevalence of subclonal ESR1 variants Poor-prognosis subgroups: TP53 and PIK3CA were the most prevalent alterations Benefit of giredestrant + everolimus observed in key subgroups: • cTF status (≥ 1% vs < 1%) • ESR1 variants, clonality and specific residue mutated • Poor-prognostic genomic subgroups (including those with ESR1, TP53, RB1 and PI3K pathway alterations) Assessment of on-treatment and end-of-treatment ctDNA dynamics is ongoing BC, breast cancer; CDK4/6i, cyclin-dependent kinase 4/6 inhibitor; ctDNA, circulating tumour DNA; cTF, circulating tumour fraction. 1. Chaudhary N, et al. NPJ Breast Cancer 2024; 10:15; 2. Pontolillo L, et al. NPJ Breast Cancer 2025; 11:93. Presented by Sara M Tolaney, MD, MPH.

Key Observations/Questions ESR1m, post CDK4/6i Endocrine Therapies [left column partially cropped: 'rial', '-ERA', 'MBER-3', 'IBER-3', 'RITAC-2', 'ERALD'] Trial | Regimen | ESR1m mPFS (mo) | HR | Comparator evERA | Giredestrant + everolimus | 10.0 vs 5.5 | 0.38 | SOC ET + everolimus EMBER-3 | Imlunestrant + abemaciclib | ~9.4 vs 5.5 (all-comers) | 0.57 | Imlunestrant alone EMBER-3 | Imlunestrant alone | 5.5 vs 3.8 | 0.62 | SOC ET VERITAC-2 | Vepdegestrant | 5.0 vs 2.1 | 0.57 | Fulvestrant EMERALD | Elacestrant | 3.8 vs 1.9 | 0.55 | SOC ET - Different populations! - Focus on control arm - HR for evERA 0.38 earned against a real combination; other trials with ESR1m tumors largely resistant to control arm [...]shar MD

Conclusions • In DESTINY-Breast11, rapid and deep antitumor activity was seen with T-DXd treatment, and responses continued to deepen with subsequent THP, highlighting the importance of completing the full neoadjuvant regimen to achieve maximal response – MRI-based ORR was 77.3% versus 50.8% after 4 cycles of T-DXd versus ddAC, increasing to 92.9% versus 87.2% with subsequent THP. The higher ORR with T-DXd versus ddAC was irrespective of HR status – Mid-treatment median tumor size reduction was 59.0% with T-DXd versus 34.8% with ddAC, increasing by EoT to 80.8% with T-DXd-THP versus 74.4% with ddAC-THP • pCR rates were higher among patients with an MRI-based response at mid-treatment and EoT. However, some responders did not have a pCR, and a substantial number of patients who did not respond had a pCR – The predictive value of MRI differed according to HR status, with greater association between MRI response and pCR in patients with HR-negative disease – MRI-based response may not be a sensitive predictor of pCR T-DXd demonstrated rapid and deep antitumor activity and responses deepened over time with subsequent THP. These findings support T-DXd-THP as an effective neoadjuvant regimen for patients with HER2-positive eBC Jean-François Boileau

[Slide 1] [Livestream screen capture] 16:15 - 17:45 Rapid oral: Breast cancer, early stage CHAIRS: MARIA VITTORIA DIECI, SANDRA XIMENA FRANCO MILLAN, LILIT HARUTYUNYAN MRI-based analysis of response DESTINY-Breast11: A randomized, global, multicenter, open-label, Phase III study (NCT05113251)1 Breast MRI scans | Baseline | Mid-treatment (at C5D1, after T-DXd/ddAC) | EoT (after T-DXd/ddAC and THP) Patient population: • ≥18 years of age • Previously untreated HER2-positive eBC • Confirmed high-risk status, defined as: – ≥cT3 and N0–3 or cT0–4 and N1–3 – Inflammatory breast cancer Stratification factors • HR status: ER and/or PR-positive or -negative • HER2 status: IHC 3+ or ISH+ in the absence of IHC 3+ status R 1:1:1 n=321 → T-DXd(a) 4 cycles → THP(b) 4 cycles → MRI-evaluable patients 309/321 (96.3%)(f) n=320 → ddAC(c) 4 cycles → THP(d) 4 cycles → 313/320 (97.8%)(f) [greyed] n=286 → T-DXd(a,e) 8 cycles → Surgery(g) / Post-neoadjuvant treatment(h) Data cut-off: March 12, 2025 Primary endpoint: • pCR (ypT0/is ypN0) Secondary endpoints: • pCR (ypT0 ypN0) • EFS • IDFS • OS • HRQoL • Safety • Pharmacokinetics and immunogenicity Prespecified exploratory endpoint: • ORR (per RECIST v1.1) – Investigator-assessed in patients with baseline MRI scans(i) – Association with pCR (ypT0/is ypN0) (post hoc) [Footnotes, small print, partially legible:] High-resolution computed tomography chest scans were performed every 6 weeks during treatment; if ILD/pneumonitis was suspected while receiving T-DXd, treatment was interrupted and a full investigation completed. Echocardiograms or multigated acquisition scans were performed during screening (<28 days prior to randomization), during treatment (<3 days before Cycle 5), and at the end of treatment to assess left ventricular ejection fraction. ... aThe T-DXd alone arm closed on March 13, 2024, following an Independent Data Monitoring Committee recommendation ... 12 patients in the T-DXd-THP arm and 7 patients in the ddAC-THP arm did not receive baseline MRI scans ... MRI evaluations were investigator or local radiologist assessed and not centrally reviewed. [remaining footnote text illegible] 1. Harbeck N, et al. Ann Oncol. 2026;37:166–79. Jean-François Boileau [Speaker video panel] Jean-Francois Boileau DESTINY-Breast11: MRI-assessed objective responses during neoadjuvant treatment (NAT) with trastuzumab deruxtecan (T-DXd) followed by paclitaxel + trastuzumab + pertuzumab (THP) vs dose-dense doxorubicin + cyclophosphamide followed by THP (ddAC-THP) in... ALICANTE AUDITORIUM - HALL 6 [Slide 2] [Livestream screen capture] 16:15 - 17:45 Rapid oral: Breast cancer, early stage CHAIRS: MARIA VITTORIA DIECI, SANDRA XIMENA FRANCO MILLAN, LILIT HARUTYUNYAN Mid-treatment: association of MRI-based response with pCR Bar chart - Patients who had pCR, %: T-DXd-THP: CR 90.2 (95% CI 76.9–97.3) 37/41; PR 66.7 (95% CI 59.6–73.2) 132/198; Non-CR/PR 54.3 (95% CI 41.9–66.3) 38/70 ddAC-THP: CR 87.5 (95% CI 47.3–99.7) 7/8; PR 71.5 (95% CI 63.6–78.6) 108/151; Non-CR/PR 41.6 (95% CI 33.7–49.8) 64/154 Predictive values for pCR (ypT0/is ypN0), % Mid-treatment response | T-DXd-THP PPV | NPV | ddAC-THP PPV | NPV CR | 90.2 | 36.6 | 87.5 | 43.6 OR | 70.7 | 45.7 | 72.3 | 58.4 | Sensitivity | Specificity | Sensitivity | Specificity CR | 17.9 | 96.1 | 3.9 | 99.3 OR | 81.6 | 31.4 | 64.2 | 67.2 PPV = [(patients with CR or OR who had pCR)/(all patients with CR or OR)]*100 NPV = [(patients without CR or OR who did not have pCR)/(patients without CR or OR)]*100 Sensitivity = [(patients with CR or OR who had pCR)/(all patients who had pCR)]*100 Specificity = [(patients without CR or OR who did not have pCR)/(all patients who did not have pCR)]*100 CR at mid-treatment was associated with pCR, particularly in patients with HR-negative disease(a); however, lack of MRI response did not exclude pCR NPV, negative predictive value; OR, objective response; PPV, positive predictive value. MRI evaluations were investigator or local radiologist assessed and not centrally reviewed. aSupplementary data Jean-François Boileau [Speaker video panel] Jean-Francois Boileau DESTINY-Breast11: MRI-assessed objective responses during neoadjuvant treatment (NAT) with trastuzumab deruxtecan (T-DXd) followed by paclitaxel + trastuzumab + pertuzumab (THP) vs dose-dense doxorubicin + cyclophosphamide followed by THP (ddAC-THP) in... ALICANTE AUDITORIUM - HALL 6 [Slide 3] [Livestream screen capture] 16:15 - 17:45 Rapid oral: Breast cancer, early stage CHAIRS: MARIA VITTORIA DIECI, SANDRA XIMENA FRANCO MILLAN, LILIT HARUTYUNYAN EoT: association of MRI-based response with pCR Bar chart - Patients who had pCR, %: T-DXd-THP: CR 83.7 (95% CI 74.2–90.8) 72/86; PR 64.7 (95% CI 57.6–71.3) 130/201; Non-CR/PR 22.7 (95% CI 7.8–45.4) 5/22 ddAC-THP: CR 75.3 (95% CI 64.2–84.4) 58/77; PR 58.2 (95% CI 50.9–65.2) 114/196; Non-CR/PR 17.5 (95% CI 7.3–32.8) 7/40 Predictive values for pCR (ypT0/is ypN0), % EoT response | T-DXd-THP PPV | NPV | ddAC-THP PPV | NPV CR | 83.7 | 39.5 | 75.3 | 48.7 OR | 70.4 | 77.3 | 63.0 | 82.5 | Sensitivity | Specificity | Sensitivity | Specificity CR | 34.8 | 86.3 | 32.4 | 85.8 OR | 97.6 | 16.7 | 96.1 | 24.6 PPV = [(patients with CR or OR who had pCR)/(all patients with CR or OR)]*100 NPV = [(patients without CR or OR who did not have pCR)/(patients without CR or OR)]*100 Sensitivity = [(patients with CR or OR who had pCR)/(all patients who had pCR)]*100 Specificity = [(patients without CR or OR who did not have pCR)/(all patients who did not have pCR)]*100 Across arms, most patients with CR at EoT had a pCR. However, 22.7% and 17.5% of patients with non-CR/PR in the T-DXd-THP and ddAC-THP arms, respectively, had a pCR NPV, negative predictive value; OR, objective response; PPV, positive predictive value. MRI evaluations were investigator or local radiologist assessed and not centrally reviewed. Jean-François Boileau [Speaker video panel] Jean-Francois Boileau DESTINY-Breast11: MRI-assessed objective responses during neoadjuvant treatment (NAT) with trastuzumab deruxtecan (T-DXd) followed by paclitaxel + trastuzumab + pertuzumab (THP) vs dose-dense doxorubicin + cyclophosphamide followed by THP (ddAC-THP) in... ALICANTE AUDITORIUM - HALL 6 [Slide 4] [Livestream screen capture] 16:15 - 17:45 Rapid oral: Breast cancer, early stage CHAIRS: MARIA VITTORIA DIECI, SANDRA XIMENA FRANCO MILLAN, LILIT HARUTYUNYAN Conclusions • In DESTINY-Breast11, rapid and deep antitumor activity was seen with T-DXd treatment, and responses continued to deepen with subsequent THP, highlighting the importance of completing the full neoadjuvant regimen to achieve maximal response – MRI-based ORR was 77.3% versus 50.8% after 4 cycles of T-DXd versus ddAC, increasing to 92.9% versus 87.2% with subsequent THP. The higher ORR with T-DXd versus ddAC was irrespective of HR status – Mid-treatment median tumor size reduction was 59.0% with T-DXd versus 34.8% with ddAC, increasing by EoT to 80.8% with T-DXd-THP versus 74.4% with ddAC-THP • pCR rates were higher among patients with an MRI-based response at mid-treatment and EoT. However, some responders did not have a pCR, and a substantial number of patients who did not respond had a pCR – The predictive value of MRI differed according to HR status, with greater association between MRI response and pCR in patients with HR-negative disease – MRI-based response may not be a sensitive predictor of pCR T-DXd demonstrated rapid and deep antitumor activity and responses deepened over time with subsequent THP. These findings support T-DXd-THP as an effective neoadjuvant regimen for patients with HER2-positive eBC Jean-François Boileau [Speaker video panel] Jean-Francois Boileau DESTINY-Breast11: MRI-assessed objective responses during neoadjuvant treatment (NAT) with trastuzumab deruxtecan (T-DXd) followed by paclitaxel + trastuzumab + pertuzumab (THP) vs dose-dense doxorubicin + cyclophosphamide followed by THP (ddAC-THP) in... ALICANTE AUDITORIUM - HALL 6

Mid-treatment: association of MRI-based response with pCR Bar chart - Patients who had pCR, %: T-DXd-THP: CR 90.2 (95% CI 76.9–97.3) 37/41; PR 66.7 (95% CI 59.6–73.2) 132/198; Non-CR/PR 54.3 (95% CI 41.9–66.3) 38/70 ddAC-THP: CR 87.5 (95% CI 47.3–99.7) 7/8; PR 71.5 (95% CI 63.6–78.6) 108/151; Non-CR/PR 41.6 (95% CI 33.7–49.8) 64/154 Predictive values for pCR (ypT0/is ypN0), % Mid-treatment response | T-DXd-THP PPV | NPV | ddAC-THP PPV | NPV CR | 90.2 | 36.6 | 87.5 | 43.6 OR | 70.7 | 45.7 | 72.3 | 58.4 | Sensitivity | Specificity | Sensitivity | Specificity CR | 17.9 | 96.1 | 3.9 | 99.3 OR | 81.6 | 31.4 | 64.2 | 67.2 PPV = [(patients with CR or OR who had pCR)/(all patients with CR or OR)]*100 NPV = [(patients without CR or OR who did not have pCR)/(patients without CR or OR)]*100 Sensitivity = [(patients with CR or OR who had pCR)/(all patients who had pCR)]*100 Specificity = [(patients without CR or OR who did not have pCR)/(all patients who did not have pCR)]*100 CR at mid-treatment was associated with pCR, particularly in patients with HR-negative disease(a); however, lack of MRI response did not exclude pCR NPV, negative predictive value; OR, objective response; PPV, positive predictive value. MRI evaluations were investigator or local radiologist assessed and not centrally reviewed. aSupplementary data Jean-François Boileau

[Slide 2] • The DESTINY[-Breast11 ...] (T-DXd-THP) versus ddAC-THP in pati[ents ...] • T-DXd-THP led to a statistically significant and clinically m[eaningful ...] (67.3% vs [illegible]), an early trend for favorable event-free [...] and an improved safety profile versus ddAC-THP1,2 – pCR benefit was irrespective of disease stage and H[R ...] – Among patients with baseline positive regional lymph[ ...] nodal pCR rate (62.2% vs 73.3%) versus ddAC-THP[ ...] • T-DXd-THP is approved by the US FDA and recommend[ed ...] the treatment of HER2-positive eBC3,4 • There is clinical interest in mid-treatment MRI-based ass[essment ...] provide early prognostic information5-7 Here, we present the exploratory analyses [...] from DESTINY-Breast11 and their [...] Jean-François Boileau Dr. Jean-François Boileau on the DESTINY-Breast11 clinical trial Breast Cancer Canada

[Slide 1] Study design HEALER is a multicenter, randomized, controlled, open-label, phase 3 study conducted at 76 sites in China (NCT05838066). Key eligibility • Centrally confirmed HER2 positive breast cancer • No prior chemotherapy or HER2-targeted therapy for recurrent or metastatic disease • At least one assessable lesion • ECOG PS of 0 or 1 R 1:1 Anbenitamab 30 mg/kg and HB1801* 100 mg/m2 Q3W (n=445) Trastuzumab 6 mg/kg†, pertuzumab 420 mg†, and docetaxel* 75 mg/m2 Q3W (n=440) Primary endpoints: ➤ PFS assessed by BIRC per RECIST v1.1 Secondary endpoints: ➤ PFS assessed by investigator ➤ ORR, DoR, DCR assessed by BIRC and investigator ➤ OS ➤ Safety Stratification factors: – Prior neoadjuvant or adjuvant therapy (yes vs. no) – Hormone receptor status (negative vs. positive) – Visceral metastasis (yes vs. no) * HB1801 or docetaxel was given for a minimum of 6 cycles, unless intolerable toxicity or disease progression; After 6 cycles, continuation of treatment was determined at the discretion of the investigator in consultation with the patient. †Trastuzumab: loading dose 8 mg/kg i.v., then 6 mg/kg Q3W; Pertuzumab: loading dose 840 mg i.v., then 420 mg Q3W. BIRC, blinded independent review committee; DCR, disease control rate; DoR, duration of response; ECOG PS, Eastern Cooperative Oncology Group performance status; HER2, human epidermal growth factor receptor 2; IHC, immunohistochemistry; ORR, objective response, rate; OS, overall survival; PFS, progression-free survival; Q3W, every 3 weeks. Binghe Xu, MD, PhD [Slide 2] BIRC-assessed PFS: primary endpoint Anbenitamab plus HB1801 significantly reduced the risk of progression or death by 54% versus THP. Landmark labels on curve: Anbenitamab plus HB1801 80.2% (12 mo), 66.9% (18 mo), 58.5% (24 mo); THP 56.9% (12 mo), 42.4% (18 mo), 34.4% (24 mo) Anbenitamab THP Median PFS, mo NR 14.3 95% CI (24.9 to NE) (12.4 to 16.4) HR, 95% CI 0.46 (0.37 to 0.56) P value 6.37*10^-15 Number at risk (months 0/3/6/9/12/15/18/21/24/27/30/33): Anbenitamab plus HB1801: 445 430 407 345 324 217 173 113 86 38 7 0 THP: 440 410 368 274 232 144 110 75 50 28 5 0 At cutoff date of April 30, 2026, 401 PFS events occurred; The median follow-up duration was 21.4 months (range, 0.0 to 33.1) in the anbenitamab group and 20.8 months (range, 0.0 to 32.5) in the THP group. BIRC, blinded independent review committee; CI, confidence interval; HR, hazard ratio; PFS, progression-free survival. Binghe Xu, MD, PhD Survival table Endpoint: BIRC-assessed PFS (primary endpoint) Arm | N | Median | Median CI Anbenitamab plus HB1801 | 445 | NR | 24.9 to NE THP | 440 | 14.3 mo | 12.4 to 16.4 HR: 0.46 HR CI: 0.37 to 0.56 P: 6.37*10^-15 Landmark rates: 12 mo: Anbenitamab plus HB1801 80.2% | THP 56.9% 18 mo: Anbenitamab plus HB1801 66.9% | THP 42.4% 24 mo: Anbenitamab plus HB1801 58.5% | THP 34.4% [Slide 3] BIRC and investigator-assessed ORR and DoR Anbenitamab + HB1801 (n=445) THP (n=440) BIRC-assessed Investigator-assessed BIRC-assessed Investigator-assessed BOR, n (%) CR 17 (3.8) 13 (2.9) 10 (2.3) 12 (2.7) PR 373 (83.8) 342 (76.9) 313 (71.1) 298 (67.7) SD 29 (6.5) 49 (11.0) 71 (16.1) 86 (19.5) Non-CR/Non-PD 17 (3.8) 28 (6.3) 21 (4.8) 22 (5.0) PD 5 (1.1) 8 (1.8) 21 (4.8) 18 (4.1) NE 4 (0.9) 5 (1.1) 4 (0.9) 4 (0.9) ORR, %, 95% CI 87.6 (84.2–90.6) 79.8 (75.7–83.4) 73.4 (69.0–77.5) 70.5 (66.0–74.7) DCR, %, 95% CI 98.0 (96.2–99.1) 97.1 (95.1–98.4) 94.3 (91.7–96.3) 95.0 (92.5–96.8) Median DoR, mo, 95% CI NE (27.0–NE) NE (24.8–NE) 16.6 (12.5–20.8) 15.3 (12.5–18.7) BIRC, blinded independent review committee, BOR, best overall response; CI, confidence interval, CR, complete response, DCR, disease control rate, DoR, duration of response, NE, not estimable, ORR, objective response rate, PD, progressive disease. [DoR Kaplan-Meier curve, right panel] Anbenitamab THP Median DoR, mo (95% CI) not reached (27.0, not estimatable) 16.6 (12.5, 20.8) Number at risk (months 0/3/6/9/12/15/18/21/24/27/30): Anbenitamab plus HB1801: 390 381 342 301 244 169 138 84 53 16 0 THP: 323 319 274 204 154 100 86 46 36 7 0 Binghe Xu, MD, PhD Survival table Endpoint: Duration of response (BIRC) Arm | N | Median | Median CI Anbenitamab plus HB1801 | 390 | not reached | 27.0, not estimatable THP | 323 | 16.6 mo | 12.5, 20.8 [Slide 4] Overall survival (8.2% maturity) Early overall survival data suggest a positive trend favoring anbenitamab plus HB1801. Anbenitamab THP Median OS, mo Not reached Not reached 95% CI (not estimable to not estimable) (not estimable to not estimable) HR, 95% CI 0.71 (0.44 to 1.12) Number at risk (months 0/3/6/9/12/15/18/21/24/27/30/33/36): Anbenitamab plus HB1801: 445 439 436 434 432 356 285 226 161 103 27 1 0 THP: 440 436 434 431 421 349 277 215 147 88 22 0 Binghe Xu, MD, PhD CI, confidence interval; HR, hazard ratio; OS, overall survival. Survival table Endpoint: Overall survival (8.2% maturity) Arm | N | Median | Median CI Anbenitamab plus HB1801 | 445 | Not reached | not estimable to not estimable THP | 440 | Not reached | not estimable to not estimable HR: 0.71 HR CI: 0.44 to 1.12

[Slide 3] Investigator-assessed PFS Consistent improvement in investigator-assessed PFS was also observed with anbenitamab plus HB1801. Landmark labels on curve: Anbenitamab plus HB1801 79.0% (12 mo), 64.4% (18 mo), 54.0% (24 mo); THP 61.3% (12 mo), 42.9% (18 mo), 35.3% (24 mo) Anbenitamab THP Median PFS, mo NR 14.6 95% CI (22.8 to NE) (13.0 to 16.6) HR, 95% CI 0.52 (0.43 to 0.64) P value 3.05*10^-11 Number at risk (months 0/3/6/9/12/15/18/21/24/27/30/33): Anbenitamab plus HB1801: 445 427 405 354 328 223 178 113 86 40 9 0 THP: 440 414 380 299 252 154 116 76 53 26 5 0 At cutoff date of April 30, 2026, 401 PFS events occurred; The median follow-up duration was 21.4 months (range, 0.0 to 33.1) in the anbenitamab group and 20.8 months (range, 0.0 to 32.5) in the THP group. CI, confidence interval; HR, hazard ratio; PFS, progression-free survival. Binghe Xu, MD, PhD Survival table Endpoint: Investigator-assessed PFS (secondary endpoint) Arm | N | Median | Median CI Anbenitamab plus HB1801 | 445 | NR | 22.8 to NE THP | 440 | 14.6 mo | 13.0 to 16.6 HR: 0.52 HR CI: 0.43 to 0.64 P: 3.05*10^-11 Landmark rates: 12 mo: Anbenitamab plus HB1801 79.0% | THP 61.3% 18 mo: Anbenitamab plus HB1801 64.4% | THP 42.9% 24 mo: Anbenitamab plus HB1801 54.0% | THP 35.3% [Slide 4] Conclusion - Anbenitamab plus HB1801 achieved clinical meaningful and statistically significant PFS compared with THP in patients with HER2-positive breast cancer. - Median PFS assessed by BIRC not reached vs. 14.3 months, with 24-month PFS rate of 58.5% vs. 34.4% (HR=0.46; 95% CI, 0.37 to 0.56; P<0.0001). - Consistent results were observed in investigator-assessed PFS and across all prespecified subgroups. - Similar improvement were observed in ORR, DCR and DoR. - Anbenitamab plus HB1801 had a favorable safety profile. - The safety profiles were consistent with known profiles of this class of combination therapies. - The majority of these events were CTCAE grade 1-2 and manageable. - HERLER demostrated that anbenitamab plus HB1801 offer a promising treatment option for these patients. Binghe Xu, MD, PhD

HEALER: anbenitamab + HB1801 substantially improves PFS vs THP Phase III • prespecified interim analysis • N=885 • 76 sites in China PRIMARY ENDPOINT: BICR-ASSESSED PFS A Progression-free Survival as Assessed by Blinded Independent Central Review Percentage of Patients Months since Randomization Anbenitamab plus HB1801 THP Hazard ratio for disease progression or death, 0.46 (95% CI, 0.37 to 0.56) P<0.001 Prespecified interim analysis: 401 PFS events; median follow-up 21.0 months KEY RESULTS HR 0.46 95% CI 0.37-0.56 24-mo PFS 58.5% vs 34.4% ORR (BICR) 87.6% vs 73.4% OS immature SAFETY • Grade ≥3 AEs: 45.9% vs 34.4% • Anemia 74%; diarrhea 53%; alopecia 52%; neutropenia 41% • Grade ≥3 neutropenia 13.1%; leukopenia 9.7% • ILD 0.9% (1 grade ≥3 event) CAVEATS / TRANSFERABILITY • Anbenitamab currently approved in China for 2L+ HER2+ gastric/GEJ cancer • Comparator was THP - not the current T-DXd + pertuzumab efficacy benchmark • Limited prior anti-HER2 exposure: trastuzumab 172/885 (19.4%); pertuzumab 61/885 (6.9%) • Two variables changed simultaneously: HP → Anbenitamab and Docetaxel → Albumin-bound Docetaxel (HB1801) WHY IT IS INTERESTING • Finite chemotherapy exposure; antibody-only maintenance is possible • No continuous ADC cytotoxic payload exposure • Median exposure: HB1801 4.8 mo vs anbenitamab 16.6 mo Wang J et al. ESMO 2026; N Engl J Med. Published online October 24, 2026. DOI: 10.1056/NEJMoa2610345. Antonino Musolino, MD, PhD, MSc Survival table Endpoint: PFS by BICR (prespecified interim analysis) Anbenitamab plus HB1801 THP HR: 0.46 HR CI: 95% CI 0.37-0.56 P: <0.001 24-mo PFS 58.5% vs 34.4%

DEMETHER: Optimizing the 1L treatment in HER2+ BC Phase II trial Open-label • single-arm • multicentre N = 165 KEY SELECTION CRITERIA: • Centrally confirmed HER2-positive breast cancer • Unresectable locally advanced or metastatic disease • Stable brain metastases allowed • No prior systemic treatment for advanced disease; one prior line of ET for MBC allowed • Prior neo/adjuvant chemotherapy or HER2-targeted therapy allowed if DFI ≥ 12 months • No prior T-DXd in the neo/adjuvant setting Induction therapy Maintenance therapy Post-treatment FU T-DXd 6 cycles PHESGO FDC SC until PD / toxicity / 3 years EoT* Physician's choice# EoS Baseline C2D1 C1D1 C2D1 *PD, death, 3 years after T-DXd initiation, other anti-cancer treatment, or discontinuation for any other reason #Subsequent anti-cancer treatment at physician discretion CO-PRIMARY ENDPOINTS: • 1-year PFS • 3-year OS SECONDARY OBJECTIVES: • PFS • Other efficacy endpoints • QoL (intra-patient comparative) • Safety and toxicity EXPLORATORY OBJECTIVES: • Biomarkers (TBD) • ctDNA monitoring Adapted from NCT06172127 and Cortés et al. Clin Cancer Res. 2024. Antonino Musolino, MD, PhD, MSc

Subgroup Analysis of Progression-free Survival Subgroup | Anbenitamab plus HB1801 | THP | Hazard Ratio for Disease Progression or Death (95% CI) (no. of patients with event/total no. of patients) All patients | 149/445 | 252/440 | 0.46 (0.38 to 0.57) Age <65 yr | 135/389 | 218/362 | 0.45 (0.37 to 0.56) ≥65 yr | 14/56 | 34/78 | 0.47 (0.25 to 0.88) ECOG performance-status score at baseline 0 | 71/251 | 130/236 | 0.38 (0.28 to 0.51) 1 | 78/194 | 120/202 | 0.59 (0.44 to 0.79) Previous neoadjuvant or adjuvant therapy Yes | 60/165 | 95/161 | 0.49 (0.36 to 0.68) No | 89/280 | 157/279 | 0.45 (0.35 to 0.58) Previous treatment with trastuzumab Yes | 30/81 | 56/91 | 0.47 (0.30 to 0.74) No | 119/364 | 196/349 | 0.46 (0.37 to 0.58) Previous treatment with pertuzumab Yes | 14/29 | 23/32 | 0.46 (0.23 to 0.90) No | 135/416 | 229/408 | 0.46 (0.38 to 0.58) Hormone-receptor status Positive | 95/244 | 150/243 | 0.48 (0.37 to 0.62) Negative | 54/201 | 102/197 | 0.44 (0.31 to 0.61) Visceral metastasis Yes | 114/316 | 190/316 | 0.47 (0.37 to 0.59) No | 35/129 | 62/124 | 0.46 (0.30 to 0.69) Liver metastasis Yes | 57/156 | 111/181 | 0.46 (0.33 to 0.63) No | 92/289 | 141/259 | 0.47 (0.36 to 0.62) Bone metastasis Yes | 85/206 | 119/193 | 0.55 (0.42 to 0.73) No | 64/239 | 133/247 | 0.38 (0.29 to 0.52) No. of distant metastatic sites 1 or 2 | 92/310 | 148/288 | 0.46 (0.36 to 0.60) ≥3 | 56/125 | 98/141 | 0.50 (0.36 to 0.70) HER2 status IHC score of 3+ | 108/374 | 207/385 | 0.43 (0.34 to 0.54) IHC score of 2+ and FISH-positive | 41/71 | 45/55 | 0.44 (0.29 to 0.69) Stage IV disease at initial diagnosis Yes | 85/253 | 139/246 | 0.48 (0.37 to 0.63) No | 64/192 | 113/194 | 0.45 (0.33 to 0.61) [x-axis 0.125 0.25 0.50 1.00 2.00] Anbenitamab plus HB1801 Better ← → THP Better

ESMO OPEN SCIENCE FOR OPTIMAL CANCER CARE ORIGINAL ARTICLE · Articles in Press, 108641, October 26, 2026 Open Access Addition of Tucatinib to Trastuzumab/Pertuzumab After Local Therapy for Isolated Intracranial Progression in HER2-Positive Metastatic Breast Cancer: Primary Results of the Phase II BRIDGET Trial S.L. Sammons 1 · A.E.D. Van Swearingen 2 · N.A. Bagegni 3 · ... · R. Murthy 9 · N.U. Lin 10 · C.K. Anders 2 ... Show more esmoopen.com

[Slide 1] Background Tucatinib is a CNS-penetrant, HER2-selective TKI that improved intracranial PFS (iPFS) when added to HER2-directed backbones (HER2CLIMB: 9.9 vs 4.2 mo; HER2CLIMB-05 on TP maintenance: 8.5 vs 4.3 mo, exploratory)1,2 BRIDGET tests whether adding tucatinib to TP or T-DM1 after local therapy prevents further intracranial progression in patients with isolated brain progression Screening | Within 12 weeks | 21-day cycles until progression or toxicity HER2+ MBC On TP or T-DM1 1st or 2nd isolated intracranial progression Isolated brain relapse off therapy* Extracranial disease stable or absent No WBRT, LMD, or prior HER2 TKI Local therapy SRS or resection Tucatinib 300 mg BID + continue TP or T-DM1 Registration Staging scans including brain MRI every 9 weeks PRIMARY ENDPOINT iPFS (RANO-BM) NULL HYPOTHESIS Historical median iPFS 5.0 mo STATISTICAL TEST One-sample log-rank; 1-sided α 0.05 POWER / SAMPLE SIZE 90% power; n = 43 Secondary endpoints: Extracranial PFS (ePFS) and Overall PFS (oPFS) (RECIST 1.1; PFS = progression at any site or death); site of first progression (intracranial v extracranial); OS; toxicity; PROs, MRD *TP initiated with tucatinib as 1L therapy. 1. Lin NU, et al. J Clin Oncol 2020;38:2610-9. 2. HER2CLIMB-05. J Clin Oncol 2026, doi:10.1200/JCO-25-02600. Sarah Sammons MD [Slide 3] CNS First and Only Site Population Survival in CNS Only as Site of First Metastasis (n=25) Legend: ePFS, iPFS, oPFS, OS Survival Probability Months from Treatment Start Kaplan-Meier estimates, patients whose first and only site of metastatic relapse was the CNS (n = 25) Consistent with the full cohort: 12-mo iPFS 62%, 18-mo iPFS 53%; no extracranial progression through 18 mo. Endpoint | Events, n (%) | Median | 6 mo | 12 mo | 18 mo CNS-only subset (n = 25) Intracranial PFS | 10 (40.0) | NR | 88% (67–96) | 62% (39–78) | 53% (30–73) Extracranial PFS | 1 (4.0) | NR | 100% | 100% | 100% Overall survival | 1 (4.0) | NR | 100% | 100% | 100% Sarah Sammons MD Survival table Endpoint: Intracranial PFS, CNS-only first-relapse subset Intracranial PFS | 25 | NR Extracranial PFS | 25 | NR Overall survival | 25 | NR iPFS 6 mo 88% (67–96); 12 mo 62% (39–78); 18 mo 53% (30–73) ePFS 6/12/18 mo 100% OS 6/12/18 mo 100% [Slide 4] Conclusions BRIDGET met its primary endpoint: adding tucatinib to TP after local therapy for isolated intracranial progression improved iPFS vs a 5.0-mo historical control (p < 0.0001) Durable CNS control: 12- and 18-mo iPFS 64% and 59% Extracranial control and survival preserved: 18-mo extracranial PFS 98%; 46/47 alive at median 19.6-mo follow-up Consistent when the brain was the first and only relapse site (n = 25; 12-mo iPFS 62%) Manageable toxicity: no grade 4–5 treatment-related AEs; 6.4% discontinued for AEs These results support TP + tucatinib as an effective strategy for isolated or brain-only relapse in HER2+ metastatic breast cancer. Sarah Sammons MD

ESMO OPEN SCIENCE FOR OPTIMAL CANCER CARE ORIGINAL ARTICLE · Articles in Press, 108641, October 26, 2026 · Open Access Addition of Tucatinib to Trastuzumab/Pertuzumab After Local Therapy for Isolated Intracranial Progression in HER2-Positive Metastatic Breast Cancer: Primary Results of the Phase II BRIDGET Trial S.L. Sammons · A.E.D. Van Swearingen · N.A. Bagegni · L.A. Huppert · S-H. Jung · G. Broadwater · K. Strickland · S. Beyer · A. Conlin · A. Morikawa · R. Murphy · N.U. Lin · C.K. Anders Endpoint | Events, n (%) | Median (95% CI), mo | 6-mo Est (95% CI) | 12-mo Est (95% CI) | 18-mo Est (95% CI) Full cohort (n = 47) Intracranial PFS | 17 (36.2) | NR | 0.91 (0.79–0.97) | 0.64 (0.48–0.76) | 0.59 (0.42–0.73) Extracranial PFS | 2 (4.3) | NR | 1.00 (NE) | 0.98 (0.84–1.00) | 0.98 (0.84–1.00) Overall PFS | 17 (36.2) | NR | 0.89 (0.76–0.95) | 0.65 (0.49–0.77) | 0.62 (0.47–0.75) Overall survival | 1 (2.1) | NR | 1.0 (NE) | 1.0 (NE) | 1.0 (NE) CNS-only subset (n = 25) Intracranial PFS | 10 (40.0) | NR | 0.88 (0.67–0.96) | 0.62 (0.39–0.78) | 0.53 (0.30–0.73) Extracranial PFS | 1 (4.0) | NR | 1.00 (NE) | 1.00 (NE) | 1.00 (NE) Overall PFS | 10 (40.0) | NR | 0.84 (0.63–0.94) | 0.63 (0.41–0.79) | 0.58 (0.36–0.75) Overall survival | 1 (4.0) | NR | 1.0 (NE) | 1.0 (NE) | 1.0 (NE)

LBA26 - Short-term preoperative durvalumab for stage I TNBC: First results from the phase II UNLOCK 14 - POP-DURVA trial Speakers Joana M. Ribeiro (Villejuif, France) Lecture Time 09:15 - 09:25 Abstract Background Short-course checkpoint blockade can induce immune activation and pathological responses without chemotherapy in early triple-negative breast cancer (TNBC). POP-DURVA assessed clinical and immune effects of anti-PD-L1 monotherapy in cT1N0 TNBC. Methods Multicentre, single-arm phase II, preoperatory trial in untreated cT1N0, TNBC with locally assessed stromal tumour-infiltrating lymphocytes (sTILs)≥5%. Durvalumab 10 mg/kg was given on days 1 and 15, followed by primary surgery (S) or end-of-treatment biopsy (EOT-B) before standard therapy; invasive residual disease at EOT was counted as non-pCR. Primary endpoint was pCR; RECIST 1.1 response and safety were secondary. Paired samples underwent central assessment of sTILs (International TILs Working Group), PD-L1 combined positive score (CPS; SP263/Roche Ventana), tertiary lymphoid structures (TLS) and spatial immune phenotype (inflamed/excluded/desert). Results The efficacy population included 71 patients (46 S; 25 EOT-B. Median age was 55 years; 45% were premenopausal, 79% had cT1c and 68% grade 3 tumours; median baseline sTILs was 20% (IQR 10–30). pCR occurred in 1/71 (1.4%; 95% CI 0.1–7.2); objective response rate was 11.6%. Among 46 patients who underwent S,19 (41.3%) achieved a tumor downstaging to pT0/T1a/T1b. In 63 paired samples, treatment promoted substantial immune remodeling of the tumor microenvironment, with median increases of 10% points for sTILs (IQR 0–30) and 9 points for PD-L1 CPS (IQR 1–20; p<0.0001). Notably, TLS emerged in 16/49 (33%) of initially negative tumours and 14/32(44%) baseline excluded tumours transitioned to an inflamed phenotype. Treatment-related adverse events occurred in 33/71 (46%); four patients (6%) had grade 2 events and none grade ≥3. Thyroid immune-mediated events occurred in 15%, all grade 1–2. Conclusions Short-course durvalumab monotherapy was well tolerated and induced rapid and frequent immune remodelling, with higher sTILs and PD-L1 expression, emergence of TLS and conversion to an inflamed phenotype despite low pCR and radiological response rates. Baseline sTILs ≥5% alone did not enrich for clinical response, supporting the need for integrated biomarkers. Integrated multi-omic profiling will be presented.

[Slide 1] Results - Immune transcriptional programmes increased after durvalumab Bulk RNA-seq · 62 patients with paired samples Biopsy–surgery: n = 44 · Biopsy–biopsy: n = 18 Individual interferon gamma (IFNG) changes IFNG log2 fold change (EOT / baseline) Increase ≥2-fold | Increase <2-fold | No increase [waterfall plot; y-axis 5.0, 2.5, 0.0, -2.5; dashed line labeled 2-fold] IFNG induction 52% 32 of 62 patients met criterion for immune activation (at least a twofold increase in expression). Immune programmes and cytotoxic gene expression Within-patient changes in immune readouts CYT (Rooney) Effector CD8 (Bagaev) IFNG IFN-γ 6-gene (Ayers) TLS 9-gene TRM (Savas) EOT − baseline (SD units, 95% CI) All six immune programmes increased overall — FDR < 0.001 PRF1 FDR <0.001 | GZMB FDR <0.001 [paired boxplots Baseline vs EOT] Increased expression of cytolytic effector genes Increased gene-expression signals associated with immune activation and cytotoxicity after durvalumab EOT core biopsies were obtained before subsequent NAST. Upper-quartile normalised bulk RNA-seq. Changes in SD units of the baseline distribution. Wilcoxon signed-rank test with Benjamini–Hochberg FDR correction across six readouts. ALICANTE AUDITORIUM - HALL 6 [Slide 2] Results - Immune remodelling after durvalumab Translational analysis set (n=76) Change in sTILs, baseline to EOT Paired-evaluable patients: n= 63 [waterfall: Change in sTILs from baseline to EOT (% points); Patients, ranked by change; dashed line Median +10] Median sTILs 10% → 22% Baseline → EOT (p<0.0001) sTIL category transitions: baseline to EOT Paired-evaluable patients: n= 63 Baseline: ≥50% n=5 (8%); 30–<50% n=10 (16%); 5–<30% n=37 (59%); <5% n=11 (17%) EOT: ≥50% n=19 (30%); 30–<50% n=11 (17%); 5–<30% n=28 (44%); <5% n=5 (8%) Higher category | Same category | Lower category sTILs <30% → ≥30% 18 of 48 tumours (38%) Baseline → EOT sTILs <50% → ≥50% 14 of 58 tumours (24%) Baseline → EOT PD-L1 CPS (SP263) Median 10 → 20 Baseline → EOT (p<0.0001; n= 57) Concordant with CD274 mRNA (ρ = 0.36) and IFN-γ (ρ = 0.38)† Short-course durvalumab was followed by increased sTILs and PD-L1 CPS EOT: D22 core biopsy or surgical specimen. sTILs centrally assessed, PD-L1 by SP263 (combined positive score). p-values: Wilcoxon signed-rank test. † Spearman ρ of changes, RNA-seq subset (n = 52). Joana M Ribeiro [Slide 3] Results - T-cell states and clonal remodelling Single-cell RNA/TCR-seq · exploratory analyses T-cell states and effector gene expression [UMAP plots: CD4 T CXCL13+, CD4 T naive/memory, CD4 T stem-like, CD4 Treg FOXP3+, CD8 GZMH+, CD8 IL7R+ KLRB1+, CD8 T GZMK+, CD8 T ZEB2+, CD8 Tex, CD8 ZNF683+ KLRC2+, ILC-like IL7R+, MAIT, NK, NK CD103+ KRT86+, NK CD56bright-like, T cycling, T naive/memory; GZMB, PRF1, IFNG feature plots] Exhausted CD8 T cells show high effector-gene expression and may contribute to the bulk signal Clone-size distribution by T-cell state Baseline and EOT pooled Clone size (within sample): 1 | 2–10 | 11–100 | >100 CD8 Tex (n=4,140) T cycling (n=1,482) CD8 T ZEB2+ (n=3,389) CD4 T CXCL13+ (n=4,823) CD8 IL7R+ KLRB1+ (n=474) CD8 T GZMK+ (n=4,539) CD8 ZNF683+ KLRC2+ (n=1,908) T naive/memory (n=3,382) CD8 GZMH+ (n=739) CD4 Treg FOXP3+ (n=5,018) CD4 T naive/memory (n=2,403) CD4 T stem-like (n=2,878) Fraction of cells Large clones among CD8 Tex cells are consistent with previous antigen stimulation. Paired changes during treatment 11 paired patients, evaluable n varies by state Fraction of cells in expanded clones (clone size ≥2) Each line represents one patient CD8 Tex (n=7) | CD4 CXCL13+ (n=9) [baseline vs EOT paired plots] Clonal remodelling may reflect expansion or accumulation of antigen-experienced T cells scRNA/TCR-seq (10x GEM-X 5'): 75,671 T/NK cells, 42 samples, 26 patients, 16 paired baseline–EOT; TCR dynamics in 11 patients with paired TCR (EOT: 7 surgical specimens, 4 EOT biopsies). Clone = cells sharing a TCR within a sample, expanded = ≥2 cells. state = majority state per clone. Texh, exhausted CD8 T cells. Joana M Ribeiro [Slide 4] Conclusions ▪ In stage I TNBC with biomarker enrichment limited to sTILs ≥5%, pCR was rare after two doses of durvalumab alone; MPR was observed in 11% of patients. ▪ No grade ≥3 treatment-related adverse events were reported. Thyroid dysfunction occurred in 12% of patients. ▪ How does the tumour immune microenvironment change after two doses of durvalumab? – Immune remodeling: Greater lymphocytic infiltration, more inflamed tumours and more frequent TLS detection. – Stronger immune effector gene-expression signals ▪ Are these changes associated with pathological response? – Greater sTIL gains associated with less residual viable tumour Perspectives: Identify the minnority of patients sensitive to short-course PD-L1 blockade. Assess the prognostic value of induced high TILs to inform future personalisation and de-escalation studies.

[Slide 2] Results - Pathological response Main Efficacy analysis set (n=71) Primary endpoint pCR after durvalumab alone 1/71 1.4% (95% CI, 0.1–7.2) Assessed at Day-29 surgery NAST counted as non-pCR Surgery cohort (n = 46) Exploratory endpoint: Pathological response by % residual viable tumour * Breast-conserving surgery: 33 (72%) Mastectomy: 13 (28%) MPR† ≤10% RVT n = 5 (11%) | PPR >10% to ≤50% RVT n = 30 (65%) | PNR >50% RVT n = 11 (24%) Neoadjuvant cohort (n = 25) Secondary endpoint: pCR after NAST 14/23 (61%) 2 patients NAST ongoing Neoadjuvant systemic therapy after durvalumab window Regimen | n (%) EC -> paclitaxel | 14 (56%) EC and paclitaxel–carboplatin (either sequence)(a) | 5 (20%) Paclitaxel–carboplatin | 3 (12%) Pembrolizumab-containing(b) | 3 (12%) Protocol version 1 – separate efficacy analysis ( N= 5): No pCR was observed after 2 doses of durvalumab (0/5). Follow-up ongoing: overall study population (n = 78) Median 12.5 months (95% CI, 8.6–17.9) No EFS events or deaths reported Data cut-off: 10 July 2026 MPR, major pathological response; PPR, partial pathological response; PNR, pathological non-response. † MPR includes pCR. * Exploratory categories adapted from INMC (Tetzlaff et al., Ann Oncol 2018). EFS includes recurrence, invasive contralateral breast cancer, second cancer or death. Radiological progression during the window is excluded from EFS. Median follow-up by reverse Kaplan–Meier. Joana M Ribeiro [Slide 3] Results - Treatment-related adverse events¥ Safety analysis set (n=78) - No grade ≥3 treatment-related adverse events were reported ¥ - Thyroid dysfunction was the most common immune-related AE, affecting 12% of patients. Adverse event | Any grade n (%) | Grade 1 n (%) | Grade 2 n (%) Patients with ≥1 treatment-related AE* | 36 (46%) | 34 (44%) | 5 (6%) Treatment-related AEs by grade£ Fatigue/Asthenia | 20 (26%) | 19 (24%) | 1 (1%) Thyroid dysfunction | 9 (12%) | 7 (9%) | 2 (3%) Diarrhoea | 6 (8%) | 6 (8%) | 0 Nausea | 8 (10%) | 8 (10%) | 0 Pruritus | 1 (1%) | 1 (1%) | 0 Skin rash | 1 (1%) | 0 | 1 (1%) Neutropenia | 1(1%) | 0 | 1(1%) Other grade 1 adverse events considered possibly related to treatment included myalgia (n=1), abdominal pain (n=1), constipation (n=3), headache (n=1), dizziness (n=2), palpitations (n=1), hypertension (n=1), and dyspnoea (n=1). ¥AE collection window: up to 30 days after the last durvalumab dose (End-of-Study visit). Late-onset irAEs reported to pharmacovigilance per protocol will be included in the final analysis and publication. * AEs - Overall grade-specific counts are non-exclusive; three patients experienced both grade 1 and grade 2 treatment-related AEs. £Individual AE categories are reported by maximum grade per patient. Joana M Ribeiro [Slide 4] Study design Multicenter, single-arm, phase II, window of opportunity study Protocol version 2 Protocol version 1: three doses of durvalumab followed by mandatory surgery. Key eligibility criteria • TNM Stage I (cT1N0) • TNBC • ER and PR <10%;HER2-negative • ECOG PS 0–1 • sTILs ≥5% (Locally assessed) Durvalumab 10 mg/kg IV, Q2W × 2 cycles 4-week window Response assessment Investigator choice* Primary endpoint¹ pCR after Durvalumab 2 cycles Surgery | pCR assessed on the surgical specimen NAST then later surgery | NAST counted as non-pCR in the primary analysis Other endpoints: Safety, radiological response, % RVT, immune dynamics and response biomarkers Screening (Day −30) | Day 1 | Day 15 | Day 22 (EOT biopsy if NAST planned) | Day 29 (+7 days) (Surgical specimen) | After NAST (months later): pCR after NAST (secondary endpoint) | Follow-up (all patients): EFS and OS Legend: Breast ultrasound | Breast MRI | Tumour biopsy | Blood sample | Clinical visit | Stool sample Statistical plan: pCR rate with exact 95% CI (Blaker) Planned n = 195, assuming pCR 20% and a 95% CI precision of ±5.7 percentage points Study on hold from January 2023 to May 2024 for a protocol amendment. Recruitment resumed in May 2024 and continued until May 2026. * Investigator choice - EOT biopsy required only if subsequent NAST was planned. ¹ pCR: ypT0/is ypN0. Subsequent NAST counted as non-pCR for the primary endpoint. NAST: neoadjuvant systemic therapy Joana M Ribeiro

Results - Immune state and residual viable tumour Surgery cohort · exploratory analyses [Scatter plot] y-axis: Residual viable tumour (% of tumour bed); x-axis: Change in sTILs, baseline to surgery (% points) Greater sTIL increases were associated with lower residual viable tumour ΔsTILs vs RVT (n = 41) Spearman ρ = −0.39 (95% CI −0.66 to −0.07); p = 0.01 Baseline sTILs vs RVT: ρ = 0.02 (n = 40) Baseline phenotype: Desert; Excluded; Inflamed; No phenotype data End of treatment: Inflamed at EoT (became or remained inflamed) [ΔsT]ILs = surgery − baseline (percentage points); complete-case analysis. Fill: baseline immune phenotype; ring: inflamed at end of treatment (became or remained inflamed). Joana M Ribeiro

[Slide 1] Study Design EXEMESTANE VS ANASTROZOLE VS LETROZOLE Patient population - Stage I-III breast cancer - Postmenopausal - HR+ / HER2 any 2 X 3 FACTORIAL DESIGN N = 3697 R 1:1:1:1:1:1 ANASTROZOLE 5 YRS TAMOXIFEN 2 YRS -> ANASTROZOLE 3 YRS EXEMESTANE 5 YRS TAMOXIFEN 2 YRS -> EXEMESTANE 3 YRS LETROZOLE 5 YRS TAMOXIFEN 2 YRS -> LETROZOLE 3 YRS Primary Endpoint - Disease-Free Survival (DFS-DCIS) Secondary Endpoints - Overall Survival (OS) - Toxicity according to CTCAE v.3.0 Methods In parallel, we were able to detect a 2% absolute difference in DFS in the 3 AI head-to-head comparisons, requiring a total of 792 events. If the result of the primary global comparison was statistically significant, prespecified pairwise comparisons, with Holm-Bonferroni correction, of AIs were performed. Stratification: - ER/PgR status (+/+, +/- or -/+) - Previous chemotherapy (no/adjuvant/neoadjuvant or both) - pN (pN0/pN1/pN2 or pN3) - HER2 (positive/negative/unknown) ClinicalTrials.gov ID NCT00541086 EUDRACT-2006-004018-42 Pietro De Placido, MD, PhD e-mail: pietro.deplacido@unina.it AI: aromatase inhibitors; HR+, hormone receptor-positive; DFS: disease-free survival; STEEP: standard definitions for efficacy end points; OS, overall survival; CTCAE: common terminology criteria for adverse events [Slide 2] Schedule Comparison Switch vs Upfront DFS and OS comparison Legend: Switch / Upfront Disease-free survival At 120 months: Upfront 77.5%; Switch 75.7% HR 0.95, 95% CI 0.82-1.09; p=0.45 Overall survival At 120 months: Upfront 86.7%; Switch 85.6% HR 0.88, 95% CI 0.74-1.06; Y-axis: Probability of DFS / Probability of OS; X-axis: Months (0-216) N* at risk (censored): Switch 1850 ...; Upfront 1847 ... [remaining values too small to transcribe reliably] No difference in terms of DFS or OS in the Switch versus Upfront schedule comparison Pietro De Placido, MD, PhD e-mail: pietro.deplacido@unina.it Enrolment: 2007-2012. Data cutoff: February 1, 2026. Median follow-up: 88.9 months; Survival table Endpoint: DFS and OS (switch vs upfront) Switch | 1850 Upfront | 1847 HR: 0.95 (DFS) HR CI: 0.82-1.09 P: 0.45 (DFS) Landmark rates: 120 mo DFS: Upfront 77.5% | Switch 75.7% 120 mo OS: Upfront 86.7% | Switch 85.6% [Slide 3] Conclusions - In GIM3 FATA, upfront aromatase inhibitor and switching from tamoxifen showed no statistically significant differences in DFS or OS, supporting either schedule as an acceptable option in the overall population. - Exemestane was associated with significantly worse DFS than anastrozole, with an absolute 10-year difference of 4.9%, with OS being non-significant. - Study limitations are that late recurrences in HR-positive disease continue beyond 10 years, and use of extended endocrine therapy was not standard at the time of the trial. - Evidence from clinical trials, together with tolerability, comorbidities, patient preference, and the differential acquisition cost, should help inform backbone selection as adjuvant therapy evolves towards combinations and further risk stratification. Pietro De Placido, MD, PhD e-mail: pietro.deplacido@unina.it [Slide 4] Aromatase Inhibitors Comparison 10-year head-to-head DFS and OS comparison Legend: Anastrozole / Exemestane / Letrozole Disease-free survival At 120 months: Anastrozole 79.1%; Letrozole 76.6%; Exemestane 74.2% Pairwise comparisons HR 1.22 (95% CI 1.03-1.45); E vs A p=0.01* HR 1.14 (95% CI 0.96-1.34); E vs L p=0.06 HR 1.04 (95% CI 0.95-1.13); L vs A p=0.51 Global p-value: 0.028* Overall survival At 120 months: Anastrozole 87.9%; Exemestane 85.7%; Letrozole 84.9% Pairwise comparisons HR 1.13 (95% CI 0.91-1.41); E vs A HR 1.01 (95% CI 0.82-1.25); E vs L HR 1.06 (95% CI 0.95-1.18); L vs A N* at risk (censored): A 1226 ...; E 1238 ...; L 1233 ... [remaining values too small to transcribe reliably] DFS was worse with exemestane than with anastrozole, with a 4.9% delta at 10 years favoring anastrozole Pietro De Placido, MD, PhD e-mail: pietro.deplacido@unina.it Enrolment: 2007-2012. Data cutoff: February 1, 2026. Median follow-up: 88.9 months; Survival table Endpoint: DFS and OS (10-year, pairwise AI comparison) Anastrozole | 1226 Exemestane | 1238 Letrozole | 1233 HR: 1.22 (DFS, E vs A) HR CI: 1.03-1.45 P: 0.01 (DFS E vs A); global DFS p=0.028 Landmark rates: 10 yr (120 mo) DFS: Anastrozole 79.1% | Letrozole 76.6% | Exemestane 74.2% 10 yr (120 mo) OS: Anastrozole 87.9% | Exemestane 85.7% | Letrozole 84.9%

Aromatase Inhibitors Comparison 10-year head-to-head DFS and OS comparison Legend: Anastrozole, Exemestane, Letrozole Disease-free survival (Probability of DFS) Labels at 120-month marker: 79.1% (anastrozole), 76.6% (letrozole), 74.2% (exemestane) Pairwise comparisons HR 1.22 (95% CI 1.03–1.45); E vs A p=0.01* HR 1.14 (95% CI 0.96–1.34); E vs L p=0.06 HR 1.04 (95% CI 0.95–1.13); L vs A p=0.51 Global p-value: 0.028* Overall survival (Probability of OS) Labels at 120-month marker: 87.9% (anastrozole), 85.7% (exemestane), 84.9% (letrozole) Pairwise comparisons HR 1.13 (95% CI 0.91–1.41); E vs A HR 1.01 (95% CI 0.82–1.25); E vs L HR 1.06 (95% CI 0.95–1.18); L vs A X-axis: Months (0–216) N* at risk (censored) rows A, E, L: [small print, largely illegible]; first values A 1226, E 1238, L 1233 DFS was worse with exemestane than with anastrozole, with a 4.9% delta at 10 years favoring anastrozole Pietro De Placido, MD, PhD e-mail: pietro.deplacido@unina.it Enrolment: 2007-2012. Data cutoff: February 1, 2026. Median follow-up: 88.9 months; Survival table Endpoint: DFS and OS, aromatase inhibitor comparison (10-year) Anastrozole (A) | 1226 Exemestane (E) | 1238 Letrozole (L) | 1233 HR: DFS: E vs A 1.22, E vs L 1.14, L vs A 1.04; OS: E vs A 1.13, E vs L 1.01, L vs A 1.06 HR CI: DFS: 1.03–1.45, 0.96–1.34, 0.95–1.13; OS: 0.91–1.41, 0.82–1.25, 0.95–1.18 P: DFS: E vs A p=0.01, E vs L p=0.06, L vs A p=0.51; global p=0.028 Landmark rates: 120 mo (DFS): Anastrozole 79.1% | Letrozole 76.6% | Exemestane 74.2% 120 mo (OS): Anastrozole 87.9% | Exemestane 85.7% | Letrozole 84.9%

[Slide 1] MADRID 2026 ESMO congress Within-trial economic analysis of chemotherapy for high-risk early breast cancer informed by a 50-gene assay 2-year results from the OPTIMA randomised trial Peter S Hall, Alison Smith, Christopher McCabe and the OPTIMA Trial Management Group 25 October 2026 [Slide 2] 2 year hospital costs Chemotherapy savings outweighed the cost of testing Net difference between arms −£654 per patient (95% CI −£1,525 to £218) Legend: Adds cost | Saves cost | Net 50-gene test +1,955 Chemotherapy −3,611 Ovarian suppression +487 Abemaciclib +382 Radiotherapy +137 Other −4 Net difference −654 Patients assigned chemotherapy, UK Standard care 100% (1,697) Test-directed 32% (546 of 1,728) 68% fewer Chemotherapy cost per patient, £ | Control | Test-directed | Difference Drugs | 124 | 39 | −82 Delivery | 3,621 | 1,162 | −2,370 Supportive medicines | 1,207 | 388 | −792 Related admissions | 588 | 207 | −367 Total | 5,540 | 1,796 | −3,611 ROR ≤60 only: chemotherapy cost ≈£5,580 control v £0 test-directed £ per patient; adjusted mean differences. ROR ≤60 control cost back-calculated (unadjusted) from the whole control arm (£5,540, n = 1,697) and the ROR >60 group (£5,461, n = 510), assuming patients without an ROR group cost the arm average. [Slide 3] Conclusions • Test-directed treatment dominated standard care over 2 years and can be considered cost-effective: £654 less per patient and 0.049 more QALYs. • A test-directed strategy saved £3,611 per patient in chemotherapy costs. • The test stays cost-saving up to a list price of about £2,550, and cost-effective up to £3,520 at £20,000 per QALY. • With non-inferior invasive breast cancer-free survival, these results support the adoption of test-directed chemotherapy in the UK National Health Service, subject to confirmation in the lifetime model.

Our Institutional approach to genomic assays in HR+/HER2- EBC Stage I-III HR+/HER2- Neoadjuvant Systemic Therapy Adjuvant Systemic Therapy N2 N1 N0 Postmenopausal Premenopausal Premenopausal Postmenopausal Age < 40 Age ≥ 40 Mammaprint (MINDACT + FLEX) Prosigna (OPTIMA) Not enough data Prosigna (OPTIMA) Oncotype (RxPONDER) Oncotype (TAILORx) Kevin Kainsky, MD, MS, FASCO

JAMA QUESTION Is trastuzumab rezetecan an effective treatment for advanced erythroblastic oncogene B (ERBB2)–low breast cancer? CONCLUSION Trastuzumab rezetecan showed significant progression-free survival improvement over physician's choice of chemotherapy in patients with ERBB2-low advanced breast cancer, with benefit observed in the both overall population and chemotherapy-naive patients. POPULATION 551 Females Females aged 18 to 75 years with centrally confirmed ERBB2-low unresectable locally recurrent or metastatic breast cancer Mean age: 55 years LOCATION 73 Hospitals in China INTERVENTION 551 Patients randomized 277 Trastuzumab rezetecan — Patients received 6.4 mg/kg intravenously every 3 weeks 274 Physician's choice chemotherapy — Patients received the physician's choice of chemotherapy (taxanes, eribulin, capecitabine, or gemcitabine) PRIMARY OUTCOME Progression-free survival FINDINGS Progression-free survival for all patients Trastuzumab rezetecan 16.4 mo (95% CI, 14.0 to 16.8) Physician's choice chemotherapy 6.8 mo (95% CI, 5.7 to 6.9) Trastuzumab rezetecan significantly prolonged progression-free survival compared with chemotherapy: Hazard ratio, 0.27 (95% CI, 0.21 to 0.34) P < .001 © AMA Fan Y, Liu Y, Wang Y, et al. Trastuzumab rezetecan in ERBB2-low advanced breast cancer: the HORIZON-Breast06 randomized clinical trial. JAMA. Published online October 23, 2026. doi:10.1001/jama.2026.21372

[Slide 1] Changing the Standard: CDK4/6 Inhibitors in HR+ mBC show a significant improvement in PFS +/- OS Trial | Population | Randomization | Arms | PFS (months) | OS (months) MONARCH-3 | Endocrine-sensitive | R 2:1 | Abemaciclib + NSAI / Placebo + NSAI | 29 / 14.8 | 67.1 / 54.5 MONARCH-2 | Endocrine-resistance (ACQUIRED) Endocrine-resistance (PRIMARY) | R 2:1 | Abemaciclib + Fulvestrant / Placebo + Fulvestrant | 16.44 / 9.3 | 46.7 / 37.3 PALOMA-2 | Endocrine-sensitive | R 2:1 | Palbociclib + Letrozole / Placebo + Letrozole | 24.8 / 14.5 | 53.9 / 51.2 PALOMA-3 | Endocrine-resistance (ACQUIRED) Endocrine-resistance (PRIMARY) | R 1:1 | Palbociclib + Fulvestrant / Placebo + Fulvestrant | 9.5 / 4.6 | 34.9 / 28 MONALEESA-2 | Endocrine-sensitive | R 1:1 | Ribociclib + Letrozole / Placebo + Letrozole | 25.3 / 16 | 63.9 / 51.4 MONALEESA-3 | Endocrine-sensitive Endocrine-resistance (ACQUIRED) Endocrine-resistance (PRIMARY) | R 2:1 | Ribociclib + Fulvestrant / Placebo + Fulvestrant | 1L: 33.6 / 19.2; 2L: 14.6 / 9.1 | 1L: 67.6 / 51.8; 2L: 39.7 / 33.7 YoungPEARL | Tamoxifen pre-Tx, 1L CTx (De novo & Sensitive relapse) | R 1:1 | Palbociclib + Exemestane + GnRHa / Capecitabine | 19.5 / 14.0 | 54.8 / 57.8 MONALEESA-7 | Endocrine-sensitive Endocrine-resistance (ACQUIRED) Endocrine-resistance (PRIMARY) | R 1:1 | Ribociclib + NSAI/TAM + GnRHa / Placebo + NSAI/TAM + GnRHa | 23.8 / 13 | 58.7 / 48 Courtesy of prof. SA Im [Slide 3] PIPELINE — KEY PHASE III TRIALS IN HR+/HER2− MBC · STATUS AS OF OCT 2026 HR+ MBC: a crowded, fast-moving trial landscape Today · Oct 2026 1L mBC: persevERA 1L giredestrant + palbociclib, n=992 — NEGATIVE (2025) pionERA 1L giredestrant + CDKi of choice, n=1050 (ETR) SERENA-4 1L camizestrant + palbociclib, n=1,371 — NEGATIVE (2026) SERENA-6 1L cami + CDKi ctDNA switch ESR1mt, n=315 — POSITIVE (HR 0.45) OPERA-2 1L palazestrant + ribociclib, n=1000 INAVO120 1L inavo + palbo + fulv, PIK3CAmt — approved 10/2024 INAVO123 1L inavolisib + palbociclib + AI, PIK3CAmt, n=450 (ETS) CAPItello-292 1L capivasertib + palbo/ribo + fulvestrant, n=895 (Ph1b/3) (ETR) VIKTORIA-2 1L gedatolisib + palbo + fulv/letrozole, n=1,180 (ETR + ETS cohorts) PIKALO-2 1L LY4064809 + CDKi + ET, PIK3CAmt, n=920 (ETS/ETR) Zovegalisib triplet 1L zovegalisib + atirmociclib + AI vs CDK4/6i + AI, PIK3CAmt (ETS) — ph3 planned KANDELA-302 1L BGB-43395 + letrozole vs CDK4/6i + letrozole, n=1,056 FOURLIGHT-3 1L atirmociclib + letrozole, n=1020 — ongoing · FOURLIGHT-1 ph2 HR 0.60 1-2L mBC: evERA 2-3L giredestrant + everolimus, post-CDKi, n=373 — POSITIVE · PDUFA 12/2026 CAPItello-291 2L (69% post-CDKi) capi + fulv, PIK3CA/AKT1/PTEN, n=818 — approved 11/2023 ReDiscover-2 2L zovegalisib + fulv vs capivasertib + fulv, PIK3CAmt post-CDKi, n=540 FINER 2L (ph2) ipatasertib + fulv, post-CDKi, n=250 — POSITIVE (ASCO 2025), not filed INAVO121 1-3L inavolisib + fulvestrant, PIK3CAmt post-CDKi, n=400 EPIK-B5 1-3L alpelisib + fulv, PIK3CAmt post-CDKi, n=188 — DISCONTINUED BGB-43395 2L BGB-43395 + fulvestrant KATSIS-1 2-3L (pivotal) prifetrastat + fulv vs everolimus + ET, post-CDKi, n=400 2L+: EMERALD 2-3L elacestrant, n=466 — approved 1/2023 EMBER-3 2-3L imlun +/- abema, n=860 — imlunestrant approved 9/2025 VERITAC-2 2-3L Vepdeg, n=624 — approved 5/2026 (ESR1m) OPERA-1 2-3L palazestrant, n=510 ELAINE III 2-3L Lasofoxifene + abema, n=500 ADELA 2L elacestrant + everolimus, post-CDKi with ESR1mt, n=240 VIKTORIA-1 2-3L Gedatolisib + fulv +/- palbo, n=701 — WT approved 7/2026 · mut sNDA 8/2026 LITESPARK-029 2L (ph2) belzutifan + fulv vs everolimus + ET, post-CDKi — HIF-2α, first in HR+ MBC Legend: Oral SERD / PROTAC (SERD · PROTAC · CERAN · SERM): giredestrant, inavolisib (Roche); elacestrant (Menarini, SERD); camizestrant (AZ, SERD); imlunestrant (Lilly, SERD); vepdegestrant (Arvinas, PROTAC); palazestrant (Olema, CERAN); lasofoxifene (Sermonix, SERM) PI3K-pathway inhibitor (PI3Ki · AKTi · PI3K/mTOR): alpelisib (NVS, PI3Ki); gedatolisib (Celcuity, PI3K/mTOR); capivasertib (AZ, AKTi); zovegalisib (Relay, PI3Ki); tersolisib (Lilly, PI3Ki) CDK4 inhibitor (CDK4-selective): atirmociclib (Pfizer, CDK4i); BGB-43395 (BeiGene, CDK4i) Others (KAT6i · HIF-2αi): prifetrastat (Pfizer, KAT6i); belzutifan (Merck, HIF-2αi) Note: key Ph3 trials unless otherwise noted · dashed outline = planned / not yet enrolling. [Slide 4] KEY EVIDENCE — RANDOMIZED TRIALS AFTER CDK4/6i · PI3K/AKT/mTOR, ORAL SERD AND COMBINATIONS Post-CDK4/6i randomized trials: median PFS by regimen Median PFS by arm, months · one colour per trial, lighter shade = control arm · PI3K/AKT/mTOR trials above, oral SERD trials below · ITT unless stated INAVO120 (1L) · fulv + palbo, PIK3CAm | 7.3 mo INAVO120 (1L) · inavolisib + fulv + palbo, PIK3CAm | 17.2 mo VIKTORIA-1 (PIK3CA-MT) · alpelisib + fulv | 5.6 mo VIKTORIA-1 (PIK3CA-MT) · gedatolisib + fulv | 11.3 mo VIKTORIA-1 (PIK3CA-MT) · gedatolisib + fulv + palbo | 11.1 mo VIKTORIA-1 (PIK3CA-WT) · fulvestrant | 2.0 mo VIKTORIA-1 (PIK3CA-WT) · gedatolisib + fulv | 7.4 mo VIKTORIA-1 (PIK3CA-WT) · gedatolisib + fulv + palbo | 9.3 mo CAPItello-291 · fulvestrant | 3.6 mo CAPItello-291 · capivasertib + fulv | 7.2 mo CAPItello-291 · capivasertib + fulv (AKT-path) | 7.3 mo FINER · fulvestrant | 1.9 mo FINER · ipatasertib + fulv | 5.3 mo FINER · ipatasertib + fulv (AKT altered) | 5.5 mo evERA · ET + everolimus | 5.5 mo evERA · giredestrant + everolimus | 8.8 mo evERA · giredestrant + everolimus (ESR1m) | 10.0 mo EMERALD · SoC ET | 1.9 mo EMERALD · elacestrant (ESR1m) | 3.8 mo SERENA-2 · fulvestrant | 3.7 mo SERENA-2 · camizestrant 75 mg | 7.2 mo SERENA-2 · camizestrant 150 mg | 7.7 mo EMBER-3 · SoC ET | 3.8 mo EMBER-3 · imlunestrant | 5.5 mo EMBER-3 · imlunestrant + abemaciclib | 9.4 mo VERITAC-2 · fulvestrant | 2.1 mo VERITAC-2 · vepdegestrant | 5.0 mo Approval status FDA | EU as of Oct 2026 (approved · review / other setting · not approved) INAVO120: FDA Oct 2024 · EU Jul 2025 — inavolisib · PIK3CAm (1L) VIKTORIA-1: FDA Jul 2026 · EU not yet — gedatolisib · WT approved, MT cohort filed CAPItello-291: FDA Nov 2023 · EU Jun 2024 — capivasertib · PIK3CA/AKT1/PTEN-alt FINER: Not approved (FDA · EU) — ipatasertib · not filed evERA: FDA review · EU not yet — giredestrant · ESR1m · PDUFA 18 Dec 2026 EMERALD: FDA Jan 2023 · EU Sep 2023 — elacestrant · ESR1m SERENA-2: Not approved in this setting — camizestrant (SERENA-6) · FDA/EU 2026 EMBER-3: FDA Sep 2025 · EU Jan 2026 — FDA mono + abema (Sep 2026) · EU mono only VERITAC-2: FDA May 2026 · EU not yet — vepdegestrant · ESR1m Post-CDK4/6i except INAVO120 (1L, endocrine-resistant). SERENA-2 from primary publication; EMBER-3 ESR1m population (imlunestrant + abemaciclib = all-comers); FINER ITT and AKT-altered cohorts (ASCO 2025); VERITAC-2 ESR1m population (ASCO 2025; ITT 3.8 vs 3.6 mo); evERA ESR1m from NEJM 2026; VIKTORIA-1 MT cohort vs alpelisib + fulv (ESMO 2026 update); FOURLIGHT-1 control = pooled fulv / exe+eve (ESMO 2026). Cross-trial comparison only — populations differ.

KEY TAKEAWAYS Five things to carry into clinic 1 Test early and repeatedly — ctDNA/NGS for ESR1, PIK3CA/AKT/PTEN, HER2 status. Horizon: ctDNA-triggered switching (SERENA-6, FDA 9/2026) — intervene when the target appears; upfront oral SERD in all-comers failed (persevERA, SERENA-4). 2 Prior adjuvant CDK4/6i is now common — judge by relapse timing and biology, not the line label; switch the agent. Horizon: CDK4/6i rechallenge in ET-resistant relapse ≥ 12 mo after adjuvant CDK4/6i (FREEASIA, upcoming); CDK4-selective atirmociclib / BGB-43395 (FOURLIGHT-3, KANDELA-302); CDK2i for CCNE1-amp. 3 After progression on CDK4/6i, change the mechanism — biomarker-directed targeted agent or new ET partner Horizon: mutant-selective PI3Kαi (zovegalisib, tersolisib) and triplets moving to 1L (INAVO123, CAPItello-292, VIKTORIA-2, PIKALO-2); KAT6i prifetrastat, HIF-2αi belzutifan post-CDK4/6i. 4 Move ADCs earlier where the evidence supports it; re-check HER2-low / ultralow status at each PD. Horizon: ADC-after-ADC order and sac-TMT chemo-naive (TroFuse-010); PARP1-selective saruparib + camizestrant in gBRCA 1L (EvoPAR-Breast01). 5 Consider clinical-trial enrollment as a priority option at every decision point. Most of the above is unresolved by randomized data — trials now switch on molecular, not imaging, signals.

New ADCs in an Already Crowded Treatment Landscape Phase 3 studies with approved or in late-stage development ADCs in HR+/HER2-neg/low ABC no longer eligible to ET [Treatment-line arrow:] ET+CDK4/6i > ET+/- TT > 1L CHT > 2L CHT > ≥3L CHT MAIN STUDY READOUTS APPROVED ADCs DESTINY-Breast06 T-DXd vs chemotherapy HER2-low / ultralow; ET-pretreated; 0 prior ABC CHT PFS 13.2 vs 8.1 mo | HR 0.62* OS Immature; HR 0.83* AEs Nausea, fatigue, ILD Positive PFS - OS immature APEX-Breast01 TQB2102 vs chemotherapy HER2-low; 0 prior ABC CHT PFS 16.4 vs 5.5 mo | HR 0.29 OS immature AEs Neutropenia; anemia; nausea Positive PFS - OS Immature DESTINY-Breast04 T-DXd vs chemotherapy HER2-low only; 1-2 prior ABC CHT; HR+ subgroup PFS 10.1 vs 5.4 mo | HR 0.51 OS 23.9 vs 17.6 mo | HR 0.69† AEs Nausea, fatigue, ILD Positive PFS + OS OptiTROP-Breast02 Sac-TMT vs chemotherapy Any HER2-negative; ≥1 prior ABC CHT (57% ≥2) PFS 8.3 vs 4.1 mo | HR 0.35 OS Not reached; early HR 0.33‡ AEs Anemia, neutropenia, stomatitis Positive PFS • OS immature ASCENT-07 SG vs chemotherapy Any HER2-negative; 0 prior ABC CHT PFS 8.3 vs 8.3 mo | HR 0.85 OS Immature AEs Neutropenia; diarrhea; nausea Primary PFS endpoint NOT met HORIZON-Breast06 SHR-A1811 vs chemotherapy HER2-low; 0-1 prior ABC CHT PFS 16.6 vs 6.8 mo | HR 0.25 OS immature AEs Anemia, leucopenia, thrombocytopenia Positive PFS - OS immature TROPION-Breast01 Dato-DXd vs chemotherapy Any HER2-negative; 1-2 prior ABC CHT PFS 6.9 vs 4.9 mo | HR 0.63 OS 18.6 vs 18.3 mo | HR 1.01 AEs Stomatitis, nausea, ocular surface events Positive PFS - Final OS negative TROPiCS-02 SG vs chemotherapy Any HER2-negative; 2-4 prior ABC CHT PFS 5.5 vs 4.0 mo | HR 0.66 OS 14.4 vs 11.2 mo | HR 0.79 AEs Neutropenia; diarrhea; nausea Positive PFS + OS ONGOING STUDIES TroFuse-010 Sac-TMT ± pembrolizumab vs CHT Any HER2-negative; 0 prior ABC CHT PFS Pending OS Pending NCT06312176 HERTHENA-Breast04 HER3-DXd vs physician's choice (including T-DXd) Any HER2-negative; after CDK4/6i; chemo-naive PFS Pending OS Pending (co-primary) NCT07060807 PANKU-Breast01 Iza-bren (EGFR×HER3) vs CHT Any HER2-negative; ≥1 prior ABC CHT PFS Presented in few minutes OS Not maturely reported ESMO 2026 Differences in prior treatments, patient populations and safety profile Access to the ADCs vary across countries

[Slide 1] What changes after ESMO 2026? 1. CNS status is becoming a relevant first-line treatment-selection variable: baseline brain metastases may favor a DB09-based strategy and reopen the question of routine baseline brain MRI. 2. Isolated CNS progression should not automatically be considered systemic treatment failure: BRIDGET supports preserving an effective extracranial backbone while intensifying CNS control. 3. For HER2+ LMD, systemic therapies already show meaningful activity; INSPIRE-LM adds a compelling signal for intrathecal HER2 targeting within a multimodal strategy, although its independent contribution remains uncertain. 4. The emerging paradigm is compartment-informed treatment: target the compartment that is failing while preserving control of those that are not. Antonio Llombart-Cussac, MD, PhD [Slide 2] INSPIRE-LM – Intra-Tecal HP for LMD a Overall Survival b Leptomeningeal PFS Despite these limitations, the remarkably prolonged survival outcomes are difficult to ignore and provide a compelling rationale for further prospective evaluation of intrathecal HER2-directed therapy within a multimodal strategy for HER2+ leptomeningeal disease (95% CI: 19.4 – NR) 12-month OS: 79%. (95% CI 7.2 – NR) 12-month LM-PFS: 61%. Antonio Llombart-Cussac, MD, PhD Kamran Ahmed, et al. 1767RO – ESMO 2026 Survival table Endpoint: OS and leptomeningeal PFS Overall survival (single arm) | [illegible] | 19.4–NR Leptomeningeal PFS (single arm) | [illegible] | 7.2–NR 12-month OS: 79% 12-month LM-PFS: 61% [Slide 4] Level of Evidence LBA49 - DESTINY-Breast09 Best first-line regimen when BM are present? Exploratory non-prespecified CNS analysis from DB09 Ph-III 3rd arm (T-DXd) remains blinded T-DXd + P vs THP Exploratory CNS subgroup 61 patients BICR-confirmed 1767RO - BRIDGET Systemic switch for exclusive CNS escape? Single-arm phase II Historical iPFS control Local Therapy + ongoing ST backbone HP for 44/47 patients Evaluating the addition of Tucatinib 1768RO - INSPIRE-LM Local directed therapy for LMD? Phase II multimodal strategy Historical OS control Local and Systemic therapy 30 patients Evaluating the addition of IT-HP [illegible] et al. [illegible] – ESMO 2026 [illegible] et al. 1767RO – ESMO 2026 Kamran Ahmed, et al. 1768RO – ESMO 2026 Antonio Llombart-Cussac, MD, PhD

[Slide 1] 08:30 - 10:00 Proffered paper 2: Breast cancer, early and advanced stage CHAIRS: SARA HURVITZ, SEOCK-AH IM Evolving Management of HER2-Positive MBC Frontline Treatment • DESTINY-Breast09: T-DXd + P Maintenance • PATINA (HR+): palbociclib + H +/- P + ET • HER2CLIMB-05: tucatinib + HP [KM: Progression-free Survival — Trastuzumab deruxtecan + pertuzumab vs THP] No. of Patients / Median Progression-free Survival (95% CI), mo: Trastuzumab Deruxtecan + Pertuzumab 383, 40.7 (36.5–NC); THP 387, 26.9 (21.8–NC) Hazard ratio for disease progression or death, 0.56 (95% CI, 0.44–0.71) P<0.00001 (prespecified P-value boundary for superiority, 0.00043) [KM: TUC + TRAS + PERT vs PBO + TRAS + PERT — small table: Events 141 | 191; Median (months) (95% CI) 24.9 (21.3 to –) | 16.3 (12.6 to 18.7); HR (95% CI) two-sided P value 0.641 (0.514 to 0.799); <.0001 — partially legible] [KM: Progression-free Survival — Palbociclib + HER2 + ET vs HER2 + ET] No. of Events/Total No. of Patients / Median PFS (95% CI), mo: Palbociclib + HER2 + ET 127/261, 44.3 (32.4–56.8); HER2 + ET 135/257, 29.1 (23.3–38.6) Hazard ratio for disease progression or death, 0.75 (95% CI, 0.59–0.96) Two-sided unstratified P=0.02 by log-rank test Rita Nanda, MD Tolaney et al, NEJM 2026; Metzger et al, NEJM 2026; Dieras et al, JCO 2026 Rita Nanda — Invited Discussant LBA26, LBA96 and LBA15 ALICANTE AUDITORIUM - HALL 6 [Slide 2] 08:30 - 10:00 Proffered paper 2: Breast cancer, early and advanced stage CHAIRS: SARA HURVITZ, SEOCK-AH IM TROPION-Breast02 Brain Metastases Data: The Caveats • Small, exploratory analysis, not stratified (n=61) • Imbalances between arms • Median age 62 vs 55 yrs, prior CNS tx 46 vs 58% • DFI < 12 mos 17 vs 31% (ICC more aggressive biology) • Only pts with asymptomatic brain mets were eligible • Cannot extrapolate to symptomatic disease or leptomeningeal metastases • Brain imaging was only mandatory for pts with mets at baseline • Can dato-DXd delay/prevent new brain metastases→early-stage setting • Nonetheless, impressive intracranial RR and PFS with dato-DXd Rita Nanda, MD Rita Nanda — Invited Discussant LBA26, LBA96 and LBA15 ALICANTE AUDITORIUM - HALL 6 [Slide 3] 08:30 - 10:00 Proffered paper 2: Breast cancer, early and advanced stage CHAIRS: SARA HURVITZ, SEOCK-AH IM POP-DURVA: Results and Relevance Clinically Negative Study • Did not meet primary endpoint (pCR 1.4%) • Only 2 doses over 4-week period • Enrichment for TIL was minimal (>5%); median level across 10% Biologically rich dataset→inform future studies • "Clean" dataset—most studies of de-escalation also include chemo/ADC • Immune activation clearly demonstrated—increase in sTILs, PD-L1, IFN-gamma, inflamed phenotype • Change in TIL, but not baseline TIL, correlated with tumor response Relevance • Predictive biomarker of IO benefit may be dynamic, and in vivo tumor response may better guide tx selection • Future of de-escalation strategies→response-adapted designs Rita Nanda, MD Rita Nanda — Invited Discussant LBA26, LBA96 and LBA15 ALICANTE AUDITORIUM - HALL 6

[Slide 3] What do these two ADCs add to the treatment landscape? What's new in these ADC features and mechanism of action as compared to T-DXd? Trastuzumab rezetecan (SHR-A1811) | Rolditamig Deuderuxtecan (TQB2102) TARGET/EPITOPE HER2 extracellular domain IV (trastuzumab) | HER2, extracellular domain II and IV (biparatopic) CONJUGATION Maleimide–thiol coupling to cystein residues | Maleimide–thiol coupling to cystein residues LINKER GGFG, enzymatically cleavable tetrapeptide | GGFG, enzymatically cleavable tetrapeptide PAYLOAD SHR169265, TOP1 inhibitor, greater membrane permeability and potency | TQ22723 (deuterated DXd), TOP1 inhibitor DAR ≈ 6 | ≈ 6 HALF-LIFE 5.1–7.5 days; fast systemic clearance of the payload potentially reducing off-target toxicities | ≈ 6-8 days, selective replacement of one or more hydrogen atoms with deuterium, potentially increased metabolic stabiliy and free-payload half-life SCHEDULE Q3W | Q3W MECHANISM OF ACTION Single-epitope HER2 binding -> HER2-mediated internalization and by-stander effect | Dual-epitope HER2 binding -> binds to adjacent HER2 molecules in a multivalent trans model, leading to large clusters, increased internalization and payload delivery [Mechanism diagrams: SHR-A1811 binding HER2 ECD IV (DAR ~6, rezetecan SHR169265 TOP1 inhibitor); TQB2102 biparatopic "trans" binding across HER2 receptor A and B] Yao et al., JCO 2024; Ma et al, Cancer Cell 2025; Dan-yun et al, ASCO 2025; Ruan et al, Ann Oncol 2026 [Slide 4] Future study design with ADCs to maximize their efficacy EARLY PHASE TRIALS Early-phase translational studies: • Longitudinal biopsies → Translational insights: tumor eco-system, ADC PD Studies integrating multidimensional biomarkers: • Clinical data • Multiomic data • Imaging → Machine Learning Models: Composite Predictive Signatures Platform studies: • Patients allocation according to composite biomarkers → Precision enrollment Integrated predictive tools Outcomes for next-generation ADCs: Personalized therapy | Rational combinations | Enhanced efficacy and safety LATE PHASE TRIALS Randomized ADC Sequential Strategy Trial in HR+/HER2− MBC Study population: HR+/HER2− metastatic breast cancer; Eligible for 2 ADC options with similar TOP1 payload class; Measurable or evaluable disease; Mandatory biopsy / ctDNA at baseline and progression Biomarker selection / stratification: Target expression, prior ADC exposure, payload sensitivity, resistance mechanisms, ctDNA dynamics N=~300, R 1:1 Strategy A: ADC 1 → ADC 2 (Switch at progression) Strategy B: ADC 2 → ADC 1 (Switch at progression) Stratification: Prior chemotherapy for MBC (yes/no); Biomarker-defined subgroup; Visceral disease (yes/no) Primary endpoint: PFS2 / strategy-level outcome Key secondary endpoints: • PFS with first ADC • OS • Safety / tolerability • Cross-resistance patterns • Patient-reported outcomes Longitudinal biopsies / ctDNA at progression to inform resistance and sequencing Figure created using ChatGPT (OpenAI) Integrating multidimensional biomarkers (including quantitative target expression) and adaptive trial designs for a biologically informed patient–ADC matching Define optimal ADC sequencing by using multidimensional biomarkers Pistilli et al, Cancer Cell 2026; Liu et al, Nat Rev Clin Oncol 2026

[Slide 1] Background Three clinical scenarios across the HER2+ disease continuum BRAIN METASTASES (BM) • Up to 50% of HER2+MBC will develop BM during their disease course • CNS progression may occur despite extracranial control • Local therapy remains central, but systemic CNS activity may matter from first line. "CNS SANCTUARY" as a target • HER2 therapies achieve meaningful intracranial disease control • Must Brain Imaging become standard in asymptomatic HER2+ patients? • May CNS detection change treatment choice? LEPTOMENINGEAL DISEASE (LMD) • Less common than BM, but major morbidity and poor survival • Diagnosis relies on clinical suspicion, MRI and CSF evaluation • Management requires individualized approach; no well established standard of care Initial BM involvement LBA49 - DESTINY-Breast09 Best first-line regimen when BM are present? Exclusive CNS progression 1767RO - BRIDGET Systemic switch for exclusive CNS escape? LMD 1768RO - INSPIRE-LM Local directed therapy for LMD? Antonio Llombart-Cussac, MD, PhD [Slide 2] BRIDGET – Adding Tucatinib for iCNS progression Screening HER2+ MBC ▪ On TP or T-DM1 ▪ 1st or 2nd isolated intracranial progression ▪ Isolated brain relapse off therapy* ▪ Extracranial disease stable or absent ▪ No WBRT, LMD, or prior HER2 TKI Within 12 weeks Local therapy SRS or resection Staging scans including brain MRI every 9 weeks Primary Objective: intracranial PFS (RANO-BM) Historical median iPFS: 5,0 months Patients characteristics: • 94% were receiving HP • 53% had isolated CNS disease as the first metastatic relapse • Patients relapsing in the brain off therapy, initiated TP together with tucatinib. Survival for all eligible patients (n=47) Survival Probability Months from Treatment Start ePFS iPFS oPFS OS median iPFS not reached vs 5.0-mo historical control, p < 0.0001 (Median FU 19.6 mo.) Sammons S, et al. Antonio Llombart-Cussac, MD, PhD

Do these change practice? Study | Design | Key result | Verdict OptiTROP-Breast01 | Ph3 RCT, n=263 | OS HR 0.60 | Confirms: TROP2-TOP1 class effect; option where available SWITCH (+ HER3-DXd B, SATEEN) | Ph2 platform, n=54 | ORR 0-8% | Practice-informing: no responses from TOP1-TOP1 payload switch Bulumtatug fuvedotin | Single-arm Ph2, n=60 | ORR 30%; PFS 3.7 mo | Hypothesis-generating on payload switch; RCT vs chemo or other SOC needed SYS6043 | Ph1/2, n≈139 | ORR 54-75% | Early: unclear post-ADC activity; RCT vs chemo or other SOC needed Sherene Loi, Peter MacCallum Cancer Centre

HR+/HER2− MBC: Proposed 1L treatment algorithm The only timing split is endocrine sensitivity — prior adjuvant CDK4/6i no longer has its own column ET-SENSITIVE DFI > 12 mo · de novo any prior adjuvant CDK4/6i AI + CDK4/6i (or CDK4 inhibitor) FOURLIGHT-3 · KANDELA-302 PIK3CA-mut: INAVO123 · PIKALO-2 · VIKTORIA-2 ET-RESISTANT · PI3K / AKT-ALTERED On ET or ≤ 12 mo after ET PIK3CA · AKT1 · PTEN PI3K / AKT inhibitor + Fulvestrant ± CDK4/6i INAVO120 (HR 0.43, OS) · CAPItello-291 (~13 % 1L) CAPItello-292 · VIKTORIA-2 · PIKALO-2 CDK4/6i-free < 12 mo: no 1L data ET-RESISTANT · PI3K PATHWAY-WT On ET or ≤ 12 mo after ET no actionable PI3K alteration Fulvestrant / oral SERD + CDK4/6i ESR1-mut: giredestrant (pionERA) · WT: fulvestrant Ribociclib + fulvestrant rechallenge vs ET ± everolimus: FREEASIA GERMLINE BRCA1/2 Any timing ET + PARP1-selective inhibitor (saruparib) Saruparib + camizestrant vs CDK4/6i + ET: EvoPAR-Breast01 Olaparib / talazoparib as fallback 2L post-PARPi → ET + CDK4/6i (no data) 1.5L — on AI + CDK4/6 inhibitor only: ESR1-mut emergence in ctDNA → switch AI → camizestrant, CDK4/6 inhibitor continued (SERENA-6, reported) Other ctDNA-guided switches: pending analysis PD → Proposed 2L / 3L algorithm · re-biopsy / ctDNA: PIK3CA · AKT1 · PTEN · gBRCA · HER2 status Assumed: CDK4i 1L approval (FOURLIGHT-3 / KANDELA-302) · oral SERD adjuvant backbone (CAMBRIA, EMBER-4, lidERA) · saruparib 1L in gBRCA. Not assumed: upfront oral SERD in unselected 1L (persevERA, SERENA-4 negative) — oral SERD only for ESR1-mut, baseline or ctDNA-emergent. Orange = ongoing / upcoming · Grey = reported · Red = no data · Baseline testing: ESR1, PIK3CA, AKT1, PTEN, gBRCA1/2, HER2 IHC (ESR1 decides AI vs SERD backbone) · OFS for premenopausal patients.

[Slide 2] Treatment strategy for HR+HER2- EBC Neo-Adjuvant: Diagnosis -> Neo-adjuvant -> Surgery -> Adjuvant Adjuvant: Diagnosis -> Surgery -> Adjuvant Recurrence risk prediction Chemo benefit prediction ET, RT + escalation of adjuvant therapy Risk prediction is a key to choose treatment. Shigehira Saji, MD, PhD [Slide 3] Summary - In phase III HR+/HER2- EBC trials, ICIs provided a ~10-percentage-point pCR gain with an irAE burden similar to that seen in TNBC - Potentially benefiting subgroups include patients with low ER, PD-L1-positive, high TIL levels, MammaPrint High2, or ImPrint-positive tumors - KEYNOTE-756 and CheckMate 7FL continued ICI after surgery and did not incorporate adjuvant CDK4/6 inhibitors, limiting both attribution of EFS benefit and applicability to current practice - SWOG S2206 focuses on MP2 tumors, omits adjuvant ICI, allows adjuvant CDK4/6 inhibitors, and uses EFS as its primary endpoint, making it a pivotal validation trial in this setting - At present, pCR improvement alone does not justify ICI use - Tumor SBRT with preserving lymph nodes may modulate immunity - Future directions: better patient selection or biological sensitization or both and of course newer IO agents such as bispecifics Shigehira Saji, MD, PhD

[Slide 1] 1873P Gedatolisib (Ged) vs Other Regimens Post-CDK4/6i and Endocrine Therapy (ET) in PIK3CA Wild-Type (WT) HR+/ HER2- Advanced Breast Cancer (aBC): A Network Meta-Analysis (NMA) Sara M. Tolaney1, VK Gadi2, Rena D. Callahan3, Barbara Pistilli4, Evan Popoff5, Michael J. Zoratti5, Alana M. Stilla5, Thelma Ubani5, Anik R. Patel5, Sidonie Niba6, Chiemeka Ike6, Seth A. Wander7 1Dana-Farber Cancer Institute, Boston, MA, USA; 2University of Illinois Cancer Center, Chicago, IL, USA; 3UCLA Health Jonsson Comprehensive Cancer Center, Los Angeles, CA, USA; 4Gustave Roussy, Villejuif, France; 5RainCity Analytics, Vancouver, BC, Canada; 6Celcuity, Minneapolis, MN, USA; 7Massachusetts General Hospital, Boston, MA, USA FPN #1873P Presented at the 2026 ESMO Annual Congress BACKGROUND - The treatment landscape of advanced breast cancer (aBC) is evolving rapidly. Notably, the introduction of cyclin-dependent kinase 4/6 inhibitors (CDK4/6i) has substantially improved clinical outcomes for patients with HR+/HER2- aBC. - However, treatment options are limited for patients with PIK3CA wild-type (WT) HR+/HER2- aBC post-progression on CDK4/6i and endocrine therapy (ET). - Gedatolisib, a kinase inhibitor of class I PI3K isoforms (α, β, δ, γ) and mTOR complexes mTORC1 and mTORC2, has emerged as an important investigational therapy with promising activity in combination with ET and CDK4/6i.1 - On July 14, 2026, gedatolisib received US FDA approval for use in combination with fulvestrant, with or without palbociclib, for the treatment of adult patients with HR+, HER2-negative locally advanced or metastatic breast cancer without a PIK3CA mutation detected following progression on or after treatment with at least one line of ET in the metastatic setting. - In the absence of head-to-head trials to inform clinical decision-making, we conducted a systematic literature review (SLR) and NMA to assess the relative efficacy of gedatolisib vs alternate regimens. OBJECTIVE - The objective of this study was to estimate the relative efficacy of gedatolisib-based regimens vs approved/emerging treatments for patients with HR+/HER2- aBC and PIK3CA WT tumors who were previously treated with CDK4/6i. METHODS - Systematic searches were conducted in MEDLINE, Embase, and CENTRAL, supplemented by hand searches of pertinent conferences and US FDA CDER reports, for randomized controlled trials of systemic treatments evaluated among patients with HR+/HER2- aBC with CDK4/6i experience and PIK3CA WT tumors. - Hazard ratios (HRs) for PFS were extracted or estimated from reconstructed Kaplan-Meier data. - A Bayesian NMA was performed for time-to-event outcomes on the log-HR scale with a normal likelihood and vague priors. Fixed-effects models were selected. - A sensitivity analysis was conducted where the assumption of equivalent imlunestrant and fulvestrant efficacy was relaxed by including the EMBER-3 fulvestrant arm, using the broader CDK4/6-experienced PIK3CA-pooled population due to data availability limitations. - All analyses were performed using R (version 4.3.1) and JAGS. METHODS (cont'd.) Figure 1. PICOS and Study Selection [funnel; numbers partly legible: 6,809 records identified; 741 full texts screened; 8 studies were included in the NMA (Figure 2)] Figure 2. Network of Evidence [nodes: Palbociclib + Avelumab + Fulvestrant; Abemaciclib + Fulvestrant; Palbociclib + Fulvestrant; Abemaciclib + Imlunestrant; SERD Monotherapy*; Capivasertib + Fulvestrant; Ipatasertib + Fulvestrant; edges: PACE, postMONARCH, VIKTORIA-1, EMBER-3, CAPItello-291, FINER, MAINTAIN, VERONICA] RESULTS - Database searches were conducted on February 13, 2026, and 13 studies were identified (Figure 1) - Five were not eligible for inclusion due to the absence of a shared comparator (DESTINY-Breast06, TROPiCS-02 and [occluded]) patients with an ESR1 mutation (EMERALD and VERITAC-2). - Finally, eight studies were eligible for inclusion in the analysis set. - Progression-free survival (PFS) was selected as the primary endpoint for the NMA because response and safety outcomes were not co[occluded] studies. - In EMBER-3, PIK3CA WT subgroup data were reported only for imlunestrant plus abemaciclib versus imlunestrant monotherapy. - Imlunestrant was assumed to have comparable efficacy as fulvestrant for the purposes of this analysis. - Study and patient characteristics of included studies are presented in Table 1 and Table 2. Key differences in inclusion criteria across t[occluded] status and prior treatment with chemotherapy. - The gedatolisib triplet demonstrated a statistically significant improvement in PFS relative to all non-gedatolisib regimens, except ri[occluded] favor of the gedatolisib triplet (Figure 3). - The percentage of patients progression-free at 6 months for gedatolisib triplet was 20% higher compared to imlunestrant+ abemac[occluded] - Similarly, statistically significant improvement in PFS was observed with gedatolisib doublet regimen compared to other therapies e[occluded] (Figure 4). - Gedatolisib doublet also demonstrated a significant 31% PFS gain at 6 months compared to capivasertib + fulvestrant. PFS gain [occluded] however, it was not statistically significant. - Results of the sensitivity analysis were consistent with the base-case analysis for both the gedatolisib triplet and doublet vs abemaci[occluded] fulvestrant arm using the broader CDK4/6-experienced, PIK3CA-pooled population did not materially alter the comparative findings. RESULTS (cont'd.) Table 1: Study Characteristics of Trials in the PIK3CA WT Network [illegible at this resolution] Figure 3: Relative Efficacy (PFS) of Ged Triplet vs Other Therapies [forest plot; values illegible] Figure 4: Relative Efficacy (PFS) of Ged Doublet vs Other Therapies [forest plot; values illegible] CONCLUSION [occluded]sed regimens compared to several comparator treatments, supporting their potential role in patients with PIK3CA-WT progression [...] PIK3CA-mutant cohort, which were not included in the present analysis, provide important new evidence for future comparative [...] comparisons, although findings should be interpreted in the context of cross-trial differences and assumptions inherent to NMA. DISCLOSURES: [small text] Study sponsored by Celcuity [Slide 3] 594P Real-world eligibility for DESTINY-Breast11 (DB-11) neoadjuvant escalation criteria and outcomes of DB-11-eligible vs DB-11-ineligible patients with HER2-positive early breast cancer (HER2+ eBC) Paolo Tarantino1, Elsa García Hernández2, Erika Schirghuber2, Ibrahim M. Abbass3, Alisan Zirtiloglu2, Thibaut Sanglier2 1Department of Medical Oncology, Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA, USA; 2F. Hoffmann-La Roche Ltd, Basel, Switzerland; 3Genentech, Inc., South San Francisco, CA, USA Background - Neoadjuvant therapy (NAT) featuring dual HER2-targeted blockade and chemotherapy is the standard of care for HER2+ eBC.1 - The DB-11 trial (NCT05113251) introduced trastuzumab deruxtecan (T-DXd) as a neoadjuvant escalation strategy for a selected high-risk population (≥ cT3,cN0 or cT0-4,cN1-3) with HER2+ eBC.2 - However, the real-world prevalence, characteristics and outcomes of patients meeting vs not meeting high-risk criteria remain incompletely characterised. - In this analysis, we aimed to quantify the distribution and real-world clinical outcomes of patients with HER2+ eBC receiving pertuzumab-based NAT, stratified by their eligibility for the DB-11 escalation strategy. Methods - This retrospective cohort study utilised the Flatiron Health (US) database, which collects longitudinal electronic health record data from academic and community cancer centres. - Patients diagnosed with HER2+ eBC who initiated neoadjuvant pertuzumab between 1 January 2013 and 31 December 2024 were included. Patients were stratified as DB-11-eligible (higher-risk; cT3-cT4 and/or cN > 0) or DB-11-ineligible (lower-risk; cT2 and cN0). No patients received T-DXd, as it was not commercially available in the US during this period. - Endpoints included pathological complete response (pCR; ypT0/is ypN0), 3-year distant recurrence-free survival (DRFS) and overall survival (OS). Results Patients - A total of 14954 patients were included, identified using a data cutoff date of 28 February 2026. - Patient demographics and baseline characteristics are summarised in Table 1. - The DB-11-ineligible lower-risk group comprised 34% (n = 5091) of the cohort, and this proportion showed a significant and progressive increase over time (from 28% in 2013 to 38% in 2024; trend p < 0.001; Figure 1). Table 1. Patient demographics and baseline characteristics | DB-11-eligible higher-risk (n = 9863) | DB-11-ineligible lower-risk (n = 5091) Median age, years (range) | 54.0 (20.0-85.0) | 55.0 (22.0-85.0) HER2 status, n (%) | | IHC 3+ | 8256 (84) | 4016 (79) IHC 2+/FISH-positive | 1557 (16) | 1048 (21) Positive NOS | 50 (1) | 27 (1) HR status, n (%) | | Positive | 6376 (65) | 3625 (71) Negative | 3460 (35) | 1451 (29) Menopausal status, n (%) | | Postmenopausal | 5769 (58) | 3098 (61) Premenopausal | 3733 (38) | 1842 (36) Unknown/not documented | 361 (4) | 151 (3) Neoadjuvant treatment, n (%) | | Chemotherapy | 9647 (98) | 4976 (98) Endocrine therapy | 233 (2) | 89 (2) Taxanes | 9619 (98) | 4964 (98) Anthracyclines | 713 (7) | 197 (4) Platinum salts | 8102 (82) | 4287 (84) Figure 1. Proportion of DB-11-ineligible lower-risk patients over time 2013 27.9% | 2014 28.8% | 2015 32.0% | 2016 33.0% | 2017 32.1% | 2018 32.9% | 2019 33.9% | 2020 34.1% | 2021 34.2% | 2022 34.5% | 2023 35.2% | 2024 37.6% Clinical outcomes - In the DB-11-eligible higher-risk group, the pCR rate was 47% overall and showed a significant upward trend over time, increasing from 32% in 2013 to 49% in 2024 (trend p < 0.001; Figure 2A). - In the DB-11-ineligible lower-risk group, the pCR rate was 48% overall and remained stable over the study period (p = 0.5797; Figure 2B). - The 3-year DRFS and OS were higher in the DB-11-ineligible lower-risk group vs the DB-11-eligible higher-risk group (by 7% and 4%, respectively; Figure 3). Figure 2. pCR over time in A) DB-11-eligible higher-risk and B) DB-11-ineligible lower-risk patients A) 2013 31.9% | 41.2% | 42.3% | 45.4% | 43.8% | 42.7% | 49.9% | 48.9% | 46.6% | 49.1% | 47.0% | 2024 49.2% B) 2013 53.6% | 37.6% | 48.1% | 44.4% | 49.5% | 48.9% | 52.3% | 50.5% | 49.5% | 45.4% | 48.3% | 2024 47.6% Figure 3. A) DRFS* and B) OS† outcomes per eligibility group 3-year DRFS, % (95% CI): DB-11-eligible higher-risk 89 (88, 89); DB-11-ineligible lower-risk 96 (95, 96) 3-year OS, % (95% CI): [higher-risk] 94 (94, 95); [lower-risk] 98 (97, 98) Conclusions - Approximately one [occluded] patients with HER[...] receiving standar[d...] ineligible for T-D[Xd...] per DB-11, with [...] increasing over [...] - DB-11-ineligible [lower-risk] patients (cT2 a[nd cN0]) exhibited favou[rable] 3-year survival [...] suggesting that [...] pertuzumab-bas[ed...] remains an effe[ctive] option for a sig[nificant] proportion of pa[tients] Please contact the lead author at paolo_tarantino@dfci.harvard.edu for permission to reprint and/or distribute. Presented at the ESMO Congress 2026, Madrid, Spain [Slide 4] 2114P Dana-Farber Cancer Institute | Harvard Medical School #7225: Recurrence of Checkpoint Inhibitor-Induced Immune-Related Adverse Events After Tumor Infiltrating Lymphocyte Therapy in Metastatic Melanoma Karam Khaddour, Ann W. Silk, Patrick A. Ott, Justine V. Cohen, David Liu, Edward Brierchek, Megan Insco, Linda Ramsdell, Michael Manos, Charles Yoon, Elizabeth Lilley, Raphael Bueno, Jordan Weiss, Anay Pachori, William Gibson, Rizwan Haq, Elad Sharon, Caron Jacobson, Stephen F. Hodi, Sarah Nikiforow, Elizabeth I. Buchbinder Background - Tumor Infiltrating Lymphocyte (TIL) therapy is a novel treatment for patients with melanoma whose disease progresses after prior checkpoint inhibitor (CPI) therapy, including programmed death-1 (PD-1) blockade. - Limited data exist regarding the risk of flare of prior CPI-induced immune-related adverse events (irAEs) following TIL therapy. - This study describes the risk of irAEs flare in patients with metastatic melanoma treated with TIL therapy from a single institution cohort at the Dana-Farber Cancer Institute, Boston MA, USA. Methods - Patients with metastatic melanoma screened for TIL therapy were eligible if they ultimately received TIL infusion and have been treated previously with [a]nti-PD-1 therapy. - [irAEs were] captured if they developed during any time prior to TIL [...] graded using the CTCAE version 5.0 - [...]stly descriptive given the small cohort size. - [...]een irAEs flare and response to TILs were analyzed. [Coh]ort diagram and irAEs [..] patients screened for TILs at DFCI -> 35 patients excluded due to not receiving TIL infusion -> 25 patients underwent TIL Therapy -> [..] Patients [experi]enced irAEs -> Endocrinopathies (n=8), Hepatitis (n=6), Dermatitis (n=1), Uveitis (n=1) Key message: In a single-center study of patients with metastatic melanoma receiving TIL therapy, prior CPI-induced irAEs were not associated with a confirmed flare after TIL therapy although 2 patients developed severe AE in organs that were previously involved with irAEs. Figure 2: Frequency of irAEs and grade in patients who received TIL therapy Most common prior CPI-induced irAEs Colitis | 8 (44%) Endocrinopathy | 8 (44%) Hepatitis | 6 (33%) Pneumonitis | 2 (11%) Dermatitis | 1 (6%) Uveitis | 1 (6%) x-axis: Patients (n); % of the 18 with prior irAE Pie: Grade ≥3 prior irAE 28%; Prior irAE < grade 3 44%; No prior irAE 28% Figure 2: Frequencies and grade of irAEs in patients who previously received anti-PD-1 therapy and eventually were infused with TILs. Figure 3: Post-TIL flare of prior irAEs Post-TIL outcome per patient with a prior irAE (n = 18) No flare of prior irAE (n = 16) | Indeterminate post-TIL event (n = 2) | Confirmed flare of same irAE (n = 0) Annotations: Prior CPI pneumonitis: acute hypoxic respiratory failure, death; flare vs. TIL toxicity not established. Prior CPI uveitis: bilateral retinal detachment (surgery, vision restored); flare not established Figure 3: Post-TIL flare of prior irAEs. Two patients experienced severe adverse events in organs which were previously involved with irAEs but attribution to an irAE flare was not confirmed. Figure 4: Case study of respiratory failure after TIL therapy and history of prior pneumonitis A 65-year-old man with metastatic melanoma (M1d) started ipi+nivo complicated by G3 pneumonitis after one infusion and required steroids. Month | Clinical Course 0 | Ipi+Nivo (1 infusion) | IV steroids followed by a taper with subsequent resolution of respiratory symptoms and radiological findings of pneumonitis within 1 month 2 | Craniotomy and SRS to brain metastases 3 | Chemotherapy 6 | TIL Infusion with IL2 x2 complicated by CLS and A. Fib 8 | Radiologic response after TILs; Admission due to hypoxia 9 | readmission due to hypoxia course complicated asystole and death [CT images labelled After CPI / During TIL Therapy / After TIL Therapy] Conclusions and future directions - While no confirmed irAE flares were observed after TIL therapy in patients with prior irAE history, two patients developed severe adverse events in organs that were involved by irAEs previously. - No association was observed between immunosuppression, or irAEs with response to TIL therapy. - Further larger studies are essential to evaluate the safety of TIL Funding and Information - Supported in part by the MO Grant at the Dana- Farber Cancer Institute. - Contact: karam_Khaddour@dfci.Harvard.edu

[Slide 1] CHAIRS: KOMAL JHAVERI, BARBARA PISTILLI HORIZON-Breast06, APEX-Breast01: HER2-targeting ADCs in HR+/HER2-low ABC Phase-3 randomized controlled trials STUDY DESIGN HORIZON-Breast06 study design Randomized, open-label, multicenter, phase 3 trial (NCT05814354) Eligibility criteria: • Centrally confirmed HER2-low (IHC 1+ or IHC 2+/ISH−) metastatic breast cancer • ≤1 prior line of chemotherapy for metastatic disease • Prior endocrine therapy required for HR-positive disease R 1:1 → Trastuzumab rezetecan (6.4 mg/kg, IV, Q3W) | Physician's choice of chemotherapy (taxane, eribulin, capecitabine, or gemcitabine) Primary endpoint: PFS (per BICR) • in chemo-naive patients • in overall patients Key secondary endpoints: • OS in overall patients • OS in chemo-naive patients • PFS (per BICR) in chemo-treated patients Stratification factors: ✓ Prior lines of chemotherapy for metastatic disease (0 vs 1) ✓ HER2 status (IHC 1+ vs IHC 2+/ISH−) ✓ HR status (HR+ with prior CDK4/6 inhibitor vs HR+ without prior CDK4/6 inhibitor vs HR−) APEX-Breast01 Study Design Randomized, open-label, multi-center, active-control, phase 3 trial (NCT06561607) Key Eligibility: • HER2-low (IHC 1+ or IHC 2+/ISH−) r/m BC • No prior chemotherapy in r/m BC setting • HR+/HER2-low: ≥1 line of ET for r/m BC or progression within 24 months after initiation of (neo)adjuvant ET • HR−/HER2-low: Patients with known PD-L1 CPS ≥1 were excluded TQB2102 (N=269) 7.5 mg/kg IV, Q3W | Chemotherapy (N=274) (capecitabine, nab-paclitaxel, paclitaxel) HR+/HER2-low = 461; HR−/HER2-low = 82 Primary endpoint: • PFS by BIRC in HR+/HER2-low pts Key Secondary endpoint: • PFS by BIRC in all HER2-low pts Secondary and other endpoints: • PFS (INV), OS, ORR, DoR, CBR • Safety • efficacy in HR−/HER2-low pts • Treatment continued until disease progression/death, intolerable toxicity, patient withdrawal, or investigator decision. • Tumor response was assessed per RECIST v1.1. Stratification Factors: • HR+ (Prior CDK4/6i use) vs HR+ (No prior CDK4/6i use) vs HR− • HER2 expression (IHC 1+ vs IHC 2+/ISH−) • Prior taxane in non-metastatic setting (yes vs no) Statistical Considerations: • Hierarchical testing was performed first for PFS in HR+/HER2-low pts, followed by PFS in all HER2-low pts, both at a two-sided α level of 0.05. One interim PFS analysis was planned. • Prespecified IA for PFS in HR+/HER2-low: approximately 216 (75%) PFS events. • Prespecified IA for PFS in all HER2-low: approximately 253 (75%) PFS events. HORIZON-Breast06 | APEX-Breast01 HR- 27% (149/551) | 15% (82/543) PRIOR CHT FOR ABC 18.5% (102/551) | 0 CHT ARM Taxane, eribuline, capecitabine, gemcitabine | Taxane, capecitabine PRIMARY ENDPOINT Co-primary: PFS in chemo-naïve, PFS in overall patients (BICR) | PFS in HR+ (BICR) Fan et al, ESMO 2026; Wang et al, ESMO 2026 Barbara Pistilli Invited Discussant LBA23, LBA13 and LBA27 [Slide 2] HORIZON-Breast06, APEX-Breast01: HER2-targeting ADCs in HR+/HER2-low ABC Primary endpoint HORIZON-Breast06 Primary endpoint: PFS in chemo-naïve patients (per BICR) [KM curve; y-axis: Progression-free survival (%); x-axis: Time (months) 0-20] Trastuzumab rezetecan | Physician's choice Median PFS (95% CI), months 16.6 (14.0-19.4) | 6.8 (5.7-7.0) Hazard ratio (95% CI) 0.25 (0.19-0.33), P<0.001 No. at Risk (months 0-20) Trastuzumab rezetecan: 225 216 207 197 193 182 170 155 147 130 114 103 84 83 67 62 59 34 33 32 18 Physician's choice: 224 214 167 145 138 113 88 67 59 42 31 27 19 18 14 12 11 7 7 7 4 APEX-Breast01 PFS (BIRC) in HR+/HER2-low: primary endpoint [KM curve; y-axis: Progression-free survival (%); x-axis: Time (months) 0-18; landmark labels 81.4% (TQB2102) and 45.9% (Chemotherapy) at dashed line at 6 months] TQB2102 (n=228) | Chemotherapy (n=233) PFS events, n (%) 66 (29.0) | 148 (63.5) Median PFS (95% CI), mo 16.4 (11.9, NE) | 5.5 (4.3, 6.9) HR=0.29 (95% CI: 0.22, 0.39) p<0.0001* No. at risk (months 0, 3, 6, 9, 12, 15, 18) TQB2102: 228 191 123 75 25 11 0 Chemotherapy: 233 146 72 29 5 2 0 Co-primary: PFS in chemo-treated patients (BICR): HR 0.32 (0.19-0.54) Both studies met their primary PFS endpoints, while OS data are still immature Fan et al, ESMO 2026; Wang et al, ESMO 2026 Barbara Pistilli Invited Discussant LBA23, LBA13 and LBA27 Survival table Endpoint: PFS per BICR/BIRC (primary) — HORIZON-Breast06 chemo-naive patients; APEX-Breast01 HR+/HER2-low Arm | N | Median | Median CI HORIZON-Breast06: Trastuzumab rezetecan (chemo-naive) | 225 | 16.6 mo | 14.0-19.4 HORIZON-Breast06: Physician's choice (chemo-naive) | 224 | 6.8 mo | 5.7-7.0 APEX-Breast01: TQB2102 (HR+/HER2-low) | 228 | 16.4 mo | 11.9-NE APEX-Breast01: Chemotherapy (HR+/HER2-low) | 233 | 5.5 mo | 4.3-6.9 HR: HORIZON-Breast06 0.25; APEX-Breast01 0.29 HR CI: HORIZON 95% CI 0.19-0.33; APEX 95% CI 0.22-0.39 P: HORIZON P<0.001; APEX p<0.0001 Landmark rates: 6 mo: APEX-Breast01: TQB2102 81.4% | APEX-Breast01: Chemotherapy 45.9% [Slide 4] ADCs with same target and same payload class are increasingly available Trastuzumab deruxtecan (T-DXd) • Target: HER2, domain IV (trastuzumab) • Payload: TOP1 inhibitor (DXd) • Linker: Peptide-cleavable (GGFG) • Bystander effect: yes [Diagram: Anti-HER2 ECD4 antibody; GGFG linker; DXd payload; DAR ≈ ~8.0] Trastuzumab rezetecan (SHR-A1811) • Target: HER2, domain IV (trastuzumab) • Payload: TOP1 inhibitor (SHR169265) • Linker: Peptide-cleavable (GGFG) • Bystander effect: yes [Diagram: Anti-HER2 ECD4 antibody; GGFG linker; SHR169265 payload; DAR ≈ ~6] Rolditamig deuderuxtecan (TQB2102) • Target: HER2, domain IV and domain II (biparatopic) • Payload: TOP1 inhibitor (TQ22723) • Linker: Peptide-cleavable (GGFG) • Bystander effect: yes [Diagram: Anti-HER2 ECD2 / Anti-HER2 ECD4 biparatopic antibody; GGFG linker; TQ22723 payload; DAR ≈ 5.8~6.0] Barbara Pistilli Invited Discussant LBA23, LBA13 and LBA27

Guidelines for stage I TNBC | T1a ≤ 5 mm | T1b > 5–10 mm | T1c > 10–20 mm NCCN v8.2026 | In certain cases with high-risk features (young, high grade) chemo can be considered | Consider adjuvant chemotherapy | (Neo)adjuvant chemotherapy ESMO 2024 | Consider no adjuvant therapy | Adjuvant chemotherapy | Neoadjuvant chemotherapy ASCO neoadjuvant 2021/22 | | | (Neo)adjuvant chemotherapy St Gallen 2025 panel | No chemotherapy (threshold > 0.5 cm: 76%) | Adjuvant chemotherapy | (Neo)adjuvant chemotherapy Legend: No chemotherapy | Consider | Chemotherapy | Neoadjuvant

[Slide 1] Proteomic Changes Support Multiple Co-Existing Mechanisms of T-DXd Resistance • ABCB1 and ABCC1 protein expression increased in 5/7 (71%) and 4/7 (57%) T-DXd-treated tumors, reinforcing the NGS findings. • MAN2A2 expression decreased in 5/7 (71%) tumors, suggesting altered glycosylation as an additional potential resistance mechanism. • Recurrent changes in DNA damage response and ferroptosis markers were also observed. By contrast, no recurrent ADC-related proteomic change was observed after SG. Post vs Pre Protein Fold Change of candidate proteins (T-DXd) Pt1 Pt3 Pt6 Pt7 Pt8 Pt18 Pt23 ADC target | Efflux Pumps | Internalization/Lysosome | ADC cleavage | Glycosylation | DNA Damage Response | Apoptosis | Ferroptosis | Cell division | ERBB family | TGFb/EMT | Metabolism | Immunosuppression Receptor Profile: HER2+/HR+ | HER2+/HR- | HER2-/HR+ Fold change Post/Pre ADC: ≥8, 4, 2, 0, -2, -4, ≤-8; |FC| < 2 Boundary Events: Detectable/above LLOQ Post only; Detectable/above LLOQ Pre only Paolo Tarantino, MD, PhD [Slide 2] Preclinical Models Validate Efflux and Impaired Internalization as Mechanisms of T-DXd Resistance • ABCB1 or ABCC1 overexpression in NCI-N87 cells significantly reduced intracellular payload exposure and T-DXd sensitivity. • siRNA-mediated MAN2A2 silencing in KPL4 cells decreased T-DXd binding and internalization, reducing intracellular payload exposure and T-DXd sensitivity. Efflux pump overexpression impacts payload exposure in vitro Parental | ABCC1 OE | ABCB1 OE | ABCG2 OE ABCC1 ABCB1 ABCG2 b-Actin HER2 b-Actin N87 Payload exposure: <0.0001; 0.0755 Sensitivity MAN2A2 downregulation reduces T-DXd internalization in vitro Binding: Trastuzumab binding (% ± SD); Ctrl siMAN2A2; 0.0475 Internalization: Control siMAN2A2; KPL4; Time (h) Payload exposure: Intracellular DXd (% ± SD); Ctrl siMAN2A2; <0.0001 Sensitivity: %Viability (mean ± SD); Control siMAN2A2; T-DXd concentration (µg/mL) Paolo Tarantino, MD, PhD

[Slide 1] Study Design KEY ELIGIBILITY CRITERIA ECOG 0-1 Locally advanced or metastatic TNBC Received taxane or anthycycline based chemotherapy Received Topoisomerase 1 Inhibitor (Top1i)-Based ADCs Prior 1 - 4 lines of therapy* BFv 1.25mg/kg on days 1,8 of 21-day cycles Treatment until disease progression, intolerable toxicity, patient request to discontinue treatment, initiation of a new anticancer therapy, death or others. KEY ENDPOINTS • Primary endpoint Objective response rate (ORR) • Secondary endpoints Duration of response (DOR) Disease control rate (DCR) Progression-free survival (PFS) Overall survival (OS) Safety Open-label, single-arm, multicentre phase 2 study | Tumor response was assessed per RECIST 1.1 Data cut-off for analysis: 19 July 2026 *The eligibility criterion for prior lines of anticancer therapy was revised from 1–4 lines to 1–3 lines Dr Yanchun Meng [Slide 2] Baseline Characteristics 1.25 mg/kg, D1,D8,Q3W (n=60) Age Median (range) 50.5 (32,73) ECOG PS n(%) 0: 11 (18); 1: 49 (82) Metastatic sites n(%) Visceral metastasis 39 (65); Liver metastasis 20 (33); Brain metastasis 4 (7) Nectin-4 expression Positive (TPS ≥1%) 39 (65); Negative (TPS <1%) 13 (22); unknown 8 (13) Prior lines of therapy n(%) 1 line 10 (17); 2 lines 25 (42); 3 lines 22 (37); 4 lines 3 (5) Prior treatment n(%) Taxane 60 (100); Anthracycline 46 (77); Platinum 21 (35); Checkpoint in hibitor 21 (35); HER-2 ADC* 20 (33); TROP2 ADC* 29 (48); Other ADC 12 (20) *One patient received both a HER2-targeted ADC and a TROP2-targeted ADC. ECOG, Eastern Cooperative Oncology Group; PS, performance status; TPS, Tumor Proportion Score Dr Yanchun Meng [Slide 3] Efficacy | 1.25 mg/kg, D1,D8,Q3W (N= 60) | | ORR (95% CI) | 30.0 (18.85,43.21) | | DCR (95% CI) | 80.0 (67.67, 89.22) | Waterfall: Change from Baseline (%); Best Overall Response: Complete Response, Partial Response, Stable Disease, Progressive Disease; Nectin-4 Status: Positive, Negative Swimmer plot: Individual Patients; Time (Months); First radiologic response (CR or PR); Disease progression Dr Yanchun Meng Median follow up time 10.7 months ORR, objective response rate; CI, confidence interval; DCR, disease control rate; CR, complete response; PR, partial response; [Slide 4] Efficacy Duration of Response (DOR) — BFv 1.25 mg/kg (N=12); Probability of Remaining in Response (%); Time (months) Progression-Free Survival (PFS) — BFv 1.25 mg/kg (N=60); Progression-Free Survival Probability(%); Time (months) Overall Survival (OS) — BFv 1.25 mg/kg (N=60); Overall Survival Probability(%); Time (months) | 1.25 mg/kg, D1,D8,Q3W (N= 60) | DOR | | Number of events,n (%) | 5 | | MedianDOR, months (95% CI) | 8.8 (8.51, NR) | | 6-month DOR rate, % (95% CI) | 81.5 (61.14, 100.00) | | 1.25 mg/kg, D1,D8,Q3W (N= 60) | PFS | | Number of events,n (%) | 47 | | MedianPFS, months (95% CI) | 3.7 (2.96, 4.40) | | 6-month PFS rate, % (95% CI) | 30.2 (20.00, 45.49) | | 1.25 mg/kg, D1,D8,Q3W (N= 60) | OS | | Number of events,n (%) | 22 | | Median OS, months (95% CI) | NR (12.71, NR) | | 12-month OS rate, % (95% CI) | 63.5 (51.53, 78.35) | Median follow up time 10.7 months NR, not reached; CI, confidence interval; Dr Yanchun Meng Survival table Endpoint: DOR / PFS / OS (single arm) Arm | N | Median | Median CI BFv 1.25 mg/kg D1,D8 Q3W — DOR (responders, N=12 per curve legend) | 12 | 8.8 | 8.51, NR BFv 1.25 mg/kg D1,D8 Q3W — PFS | 60 | 3.7 | 2.96, 4.40 BFv 1.25 mg/kg D1,D8 Q3W — OS | 60 | NR | 12.71, NR 6-month DOR rate 81.5% (61.14, 100.00) 6-month PFS rate 30.2% (20.00, 45.49) 12-month OS rate 63.5% (51.53, 78.35)

Overall Survival and Leptomeningeal PFS a Cumulative Survival Months b Leptomeningeal PFS Months No at Risk (censored) Endpoint met, Median survival 36.2 months (19.4 – NR) with a 12-month OS of 79%, historical control 4.4 months Median LM-PFS 19.3 months (7.2 – NR) with a 12-month LM-PFS of 61% Kamran Ahmed Survival table Endpoint: Overall survival (panel a) and leptomeningeal PFS (panel b), single arm Arm | N | Median | Median CI RT followed by IT trastuzumab/pertuzumab - OS | 30 | 36.2 months | 19.4 - NR RT followed by IT trastuzumab/pertuzumab - LM-PFS | 30 | 19.3 months | 7.2 - NR 12-month OS 79% 12-month LM-PFS 61% Historical control median OS 4.4 months

[Slide 1] Background and Methods • Mechanisms of resistance to TOPO1i ADCs remain poorly understood. • We conducted multiomic profiling of paired biopsies from 26 patients with metastatic breast cancer, collected before and after T-DXd (n=11), SG (n=12), or sequential T-DXd followed by SG (n=3). Cohort (n=26) T-DXd only (n=11) SG only (n=12) T-DXd then SG (n=3) HR+/HER2+ (n=6) HR-/HER2+ (n=3) HR+/HER2- (n=12) TNBC (n=5) N=26 PRE ADC POST-PROGRESSION NGS IHC MASS SPECTROMETRY PROTEOMIC PROFILING Real-world cohort: biopsy sites and intervening therapies varied. Number of samples with data available varied across different techniques. No samples were excluded due to non-QC reasons Paolo Tarantino, MD, PhD ADC (antibody-drug conjugate); T-DXd (trastuzumab deruxtecan); SG (sacituzumab govitecan), TOP1i (topoisomerase 1 inhibitors) [Slide 2] TP53 Mutations are Acquired in a Subset of Tumors at ADC Progression • Acquired TP53 mutations were detected after T-DXd and/or SG (3/15 patients, 20%) • These mutations included two hotspot mutations and one loss-of-function alteration T-DXd Subtype Timepoint Location 67% TP53 28% PIK3CA 17% GATA3 11% ARID1B 11% NCOR1 11% TSC2 6% PTEN Pt1 Pt10 Pt18 Pt23 Pt49 Pt51 Pt6 Pt61 Pt9 SG 38% TP53 19% ARID1B 6% BRCA2 6% NF1 6% RB1 6% RUNX1 6% TSC1 Pt11 Pt12 Pt13 Pt14 Pt16 Pt17 Pt18 Pt61 Freq (%) Mutation Type: Missense, Nonsense, Frameshift, Splice, Inframe_indel Subtype: HER2-/HR+, HER2+/HR-, HER2+/HR+, TNBC Timepoint: T1, T2 Location: Breast, Liver, LN, Other Paolo Tarantino, MD, PhD [Slide 3] Acquired ABCB1/ABCC1 Copy-Number Gains Emerge at ADC Progression • ABCB1 and ABCC1 showed acquired copy-number gains at ADC progression in 3/8 (37.5%), supporting drug efflux as a potential resistance mechanism. • One tumor exhibited marked co-amplification of both efflux pumps (ABCB1: 4→13; ABCC1: 3→34). ΔCopy Number (Post – Pre) T-DXd: Pt6 Pt10 Pt18 Pt23 Pt49 Pt51 SG: Pt11 Pt16 Pt18 Pt51 ABCB1 (chr7) | -1 | +9 | +0 | +4 | +1 | +1 | -1 | +0 | +0 | +1 ABCC1 (chr16) | -1 | +31 | -1 | -1 | +1 | +2 | -2 | +4 | -1 | +2 ABCG2 (chr4) | -1 | -1 | +0 | +0 | -1 | +0 | +0 | +3 | +0 | +0 MYC (chr8) | +0 | -2 | +1 | -12 | +1 | +1 | +2 | +9 | +1 | +1 TOP1 (chr20) | -1 | -2 | -3 | +0 | +0 | +0 | +0 | +5 | -3 | +0 ADC target ERBB2 (chr17)/TROP2 (chr1) | -5 | -17 | -15 | -2 | +0 | +2 | +0 | +3 | +0 | +1 CN change Post-Pre Copy Number changes <4 are shown in grey Paolo Tarantino, MD, PhD [Slide 4] Conclusions • ADC resistance is multifactorial, with complete target loss being uncommon and multiple resistance mechanisms co-existing within individual tumors. • Functional studies implicate ABCB1/ABCC1-mediated drug efflux and MAN2A2-associated internalization defects as mechanisms of T-DXd resistance • Payload-related mechanisms may be critical determinants of ADC resistance, highlighting the limitations of relying on target expression alone. LIMITATIONS: These results are hypothesis-generating, warranting confirmation in larger, prospective studies.

[Slide 1] Methods Main Study: JCOG1204 Key eligibility criteria: • ≥4 lymph node metastases (LNM) for estrogen receptor (ER)+ breast cancer without neoadjuvant therapy • ≥1 LNM for ER- breast cancer without neoadjuvant therapy • ≥1 LNM after neoadjuvant therapy for both ER+/- breast cancer Parent: JCOG1204 INSPIRE Trial (N = 1034) Primary endpoint: Overall Survival All breast cancer subtypes | High risk of recurrence | After curative surgery R -> Standard Surveillance: Ancillary: JCOG1204A1 ctDNA monitoring (N = 56) ··· ctDNA testing starts ≥1 years after surgery Intensive Surveillance: Ancillary: JCOG1204A1 ctDNA monitoring (N = 189) ··· ctDNA testing starts ≥1 years after surgery Ancillary ctDNA Study: JCOG1204A1 Key inclusion criteria: • Patients enrolled in JCOG1204 • No more than 4 years after surgery • Relapse-free at time of enrollment • Exploratory analysis for early detection of recurrence by surveillance ctDNA began ≥1 years after surgery* *Beginning ctDNA testing >1 year from surgery potentially reducing representation of early relapsed patients with HER2+ or TNBC Yukinori Ozaki, M.D., Ph.D. [Slide 2] Results: Cumulative ctDNA detection across all surveillance arms [Step chart] Legend: Cumulative incidence of ctDNA detection (patient level) Y-axis: Cumulative ctDNA detection (%) 0%-20%; X-axis: Months from surgery 0-96 Box: ctDNA testing began ≥1 years after surgery • Among 210 evaluable patients, 28 (13.3%) became ctDNA positive during a median clinical follow-up period of 72.3 months. • Median time to first ctDNA test: 35.3 (12-51.2) months from surgery • Median duration of ctDNA surveillance: 36.1 (0.0-60) months Yukinori Ozaki, M.D., Ph.D. [Slide 3] Results: RFS in ctDNA-negative vs. ctDNA-positive patients (N = 210, ctDNA status available) Curve labels: 5 yr RFS: 99.5% (ctDNA-neg); 5 yr RFS: 53.6% (ctDNA-pos) Box: ctDNA testing began ≥1 years after surgery HR = 60.3, 95% CI: (20.5–177.5); p < 0.0001 Y-axis: Recurrence-free survival; X-axis: Years after the date of surgery; + Censored Numbers at Risk (0-9 years): ctDNA-neg: 182 182 181 181 181 175 99 51 5 0 ctDNA-pos: 28 28 28 24 22 15 6 3 1 0 • Unstratified Cox proportional hazards model showed HR: 60.3, 95% CI: 20.5–177.5. • Cox proportional hazards model with ctDNA positive/negative as a time-dependent covariate, HR: 27.7, 95% CI: 10.1–75.4. • In the ER+/HER2- subgroup, the 7-year RFS was 94.1% in ctDNA-negative patients, versus 15.9% in ctDNA-positive patients with HR 32.0, 95% CI: 10.8–95.3. Yukinori Ozaki, M.D., Ph.D. Survival table Endpoint: RFS by ctDNA status ctDNA-negative | 182 ctDNA-positive | 28 HR: 60.3 HR CI: 20.5–177.5 P: <0.0001 Landmark rates: 5 yr: ctDNA-negative 99.5% | ctDNA-positive 53.6% 7 yr (ER+/HER2- subgroup): ctDNA-negative 94.1% | ctDNA-positive 15.9% [Slide 4] Results: Lead time interval among patients who converted from ctDNA-negative to ctDNA-positive during surveillance and experienced disease recurrence (N = 16) [Swimmer plot] Standard Surveillance Arm (4 patients); Intensive Surveillance Arm (12 patients) Legend: First ctDNA (–) test; First ctDNA (+) test; MRD-negative period; MRD-positive (lead-time to event); Recurrence or death X-axis: Months from surgery 0-84 Overall lead time: • Median lead time 7.5 months (range: 0.3-18.9). Lead time by surveillance arm: • Standard surveillance arm (n=4): Median lead time of 7.4 months (range: 5.1-8.3) • Intensive surveillance arm (n=12): Median lead time 9.5 months (range: 0.3-18.9) Contextual note: recurrence rate was similar in both arms in the parent trial. Participation in the ctDNA substudy was optional, and was skewed toward the intensive surveillance arm (N=189) vs. standard surveillance arm (N=56). Yukinori Ozaki, M.D., Ph.D.

ADC Target Expression can Decrease at Progression, yet Usually Remains Detectable • HER2 decreased in 8/14 (57%) after T-DXd by IHC and/or MS. • However, some degree of HER2 expression remained detectable in all tumors by IHC and/or MS. • After SG, there was no consistent TROP2 downregulation (5/14 with TROP2 increase, 3/14 with decrease). HER2 IHC HER2 Score by Patient (T1 vs T2) Baseline (T1) Post-treatment (T2) T-DXd HER2 Score 0 0+ 1+ 2+ 3+ HER2 Peptide Levels TDXD Wilcoxon p = 0.156 Paired N = 7 amol/µg LLOQ TROP2 IHC SG TROP2 H-Score Timepoint Paired Wilcoxon: p = 0.307 | N = 11 TROP2 Peptide Levels Saci Wilcoxon p = 0.625 Paired N = 5 Paolo Tarantino, MD, PhD IHC, immunohistochemistry; MS, mass spectrometry.

[Slide 1] DISCUSSION ❖ FDP-M arm demonstrated a significantly better control of nausea translating into higher total control rates in all 3 phases as compared to FDP-O arm (78.5% vs 56.4%; 52.3% vs 25.5% and 46.2% vs 21.8% respectively). ❖ In contrast, previous studies have demonstrated a better control of nausea with the olanzapine- based quadruplet regimen compared with a mirtazapine-based regimen, potentially attributable to the higher doses of dexamethasone (12mg) and olanzapine (10mg) 5. ❖ No significant differences in the control of vomiting in the acute, delayed and overall phases. ❖ Olanzapine- acts on D2, 5-HT, alpha adrenergic, histamine and muscarinic receptors- effective in suppressing vomiting pathways ❖ Control of nausea more complex- involves regulation of cortical and limbic pathways and hence mirtazapine appears to be more effective. 5. Maleki A, Ghadiyani M, Salamzadeh J, Salari S, Banihashem S, Tavakoli-Ardakani M. Comparison of Mirtazapine and Olanzapine on Nausea and Vomiting following Anthracycline-cyclophosphamide Chemotherapy Regimen in Patients with Breast Cancer. Iran J Pharm Res. 2020;19(3):451-64. Anusha Mruthyunjaya Swamy The efficacy and safety of mirtazapine vs olanzapine in preventing chemotherapy induced nausea and vomiting in combination with the triplet regimen in patients receiving highly emetogenic chemotherapy- an open label, prospective, randomized controlled ... [Slide 2] MIRTAZAPINE 'Serotonin by night, appetite by morning- sleep better and eat better ' COMPLEX PHARMACOLOGY ▪ Indirect agonist on CNS 5-HT1A receptors ▪ Inverse agonist on 5-HT2C and H1 receptors ▪ Antagonist on 5-HT2, 5-HT3 and α2 receptors 3 AUXILLARY BENEFITS ✓ Accelerates gastric emptying ✓ Stimulates appetite ✓ Anxiolytic ✓ Improves sleep quality DATA ON CINV ▪ Kim SW et al. evaluated mirtazapine as a rescue antiemetic ▪ It was associated with improvements in pain, depression, anorexia and QoL 4 AS A PART OF THE QUADRUPLET REGIMEN ▪ Maleki et al. compared mirtazapine and olanzapine as part of a quadruplet antiemetic regimen in patients receiving HEC (AC#) for EBC * ▪ Mirtazapine-minimal or no impact of CINV on QoL 5 #AC- Adriamycin Cyclophosphamide *EBC- Early Breast Cancer [reference lines small, partially illegible]

CHAIRS: ANTONIO LLOMBART CUSSAC, SHERENE LOI, PEDRAM RAZAVI Ultra-sensitive MRD Assessment in Long-term Complete Responders to Anti-HER2 Therapy in HER2-positive Metastatic Breast Cancer: Initial Results from HER2-CR Study Kazuki Nozawa, Yukinori Ozaki, Haruru Kotani, Yumiko Akita, Ayumi Kataoka, Akiyo Yoshimura, Kana Yukino, Kiyomi Nonogaki, Toshimi Takano, Takayuki Ueno, Yumi Wanifuchi-Endo, Tomoko Asano, Nanae Horisawa, Mitsuo Terada, Kaori Terata, Derrick Renner, Angel Rodriguez, Minetta C. Liu, Fumikata Hara, Tatsuya Toyama, Hiroji Iwata DOI: 10.1016/j.esmoop.2026.108639 URL: https://www.esmoopen.com/article/S2059-7029(26)02585-8/fulltext QR Code: ESMO OPEN SCIENCE FOR OPTIMAL CANCER CARE ALICANTE AUDITORIUM - HALL 6

Fields of Development - HRpos/HER2neg aBC Endocrine-Based Treatments | Other Pathways | CT/ADCs/Radioligands/IO CDK-targeted developments: CDK4/6i: CDK4/6i + fulvestrant/AI1 CDK4-Specific Inhibitor: Atirmociclib + fulvestrant2 Next-generation CDK inhibitors (CDK2i, CDK7i)3 CDK4-Specific Inhibitor: Atirmociclib + AI [Cell Cycle Control] ESR1mut Associated Resistance: ESR1mut: Elacestrant1 ESR1mut: Imlunestrant ± abemaciclib4,5 ESR1mut: Vepdegestrant6,7 ESR1mut: Lasofoxifene + abemaciclib8 ESR1mut: Palazestrant9 ESR1mut: Giredestrant + everolimus10 ESR1mut: Camizestrant+CDK4/6i27 ESR1mut: Novel SERD, SERM, CERAN, SERCA, PROTAC11 PI3Kα / AKT / mTOR: PIK3CAmut: Alpelisib + fulvestrant1,12 PIK3CAmut, AKTmut, PTENalt: Capivasertib + fulvestrant13-15 PIK3CAmut: Inavolisib±palbociclib+fulvestrant16,17 Everolimus + ET1 PIK3CAmut: Gedatolisib ± palbociclib + fulvestrant18 [Pi3K Pathway] Novel MoAs: Epigenetic modulators (KAT6i)19,20 MAPKmut: Fulvestrant + MEKi1 FGFRalt: Fulvestrant + FGFRi1; Fulvestrant + palbociclib + FGFRi21 Other Targeted Therapy: germline BRCA/PALB2mut: PARPi1 HRD+: PARPi/DDRi/combo1 RB1mut: AURKAi or TTKi1 AR+: Enobosarm ± abemaciclib1 HER2mut: Fulvestrant + neratinib + trastuzumab1 Additional Mechanisms: Chemotherapy1 Sacituzumab govitecan1,22 HER2 low/HER2 ultralow-targeted ADC T-DXd1,23 HER2 low-targeted ADC Trastuzumab pamirtecan24 HER3-targeted ADCs25 Bispecific ADCs25 Novel ADC or immunotherapy combinations1 Radioligand therapy26 Immune-Checkpoint-Inhibitors [Immune Oncology] [Footnotes and references in tiny print - illegible in this photo] Peter A. Fasching (peter.fasching@uk-erlangen.de) [Timer display: 08:56]

Three different ADCs targeting HER2 in HER2-low BC Same target, same linker, same payload class Trastuzumab deruxtecan (T-DXd) • Target: HER2, domain IV (trastuzumab) • Payload: TOP1 inhibitor (DXd) • Linker: Peptide-cleavable (GGFG) • Bystander effect: yes [structure diagram] Anti-HER2 ECD4 | GGFG linker | DAR = ~8.0 | DXd payload Trastuzumab rezetecan (SHR-A1811) • Target: HER2, domain IV (trastuzumab) • Payload: TOP1 inhibitor (SHR169265) • Linker: Peptide-cleavable (GGFG) • Bystander effect: yes [structure diagram] Anti-HER2 ECD4 | GGFG linker | DAR = ~6 | SHR169265 payload Rolditamig deuderuxtecan (TQB2102) • Target: HER2, domain IV and domain II (biparatopic) • Payload: TOP1 inhibitor (TQ22723) • Linker: Peptide-cleavable (GGFG) • Bystander effect: yes [structure diagram] Anti-HER2 ECD2 | Anti-HER2 ECD4 | GGFG linker | DAR = 5.8~6.0 | TQ22723 payload

[Slide 1] A global, interventional, open-label trial of trastuzumab deruxtecan (T-DXd)–based therapy followed by palbociclib, endocrine therapy (ET), and dual HER2 blockade in first-line (1L) human epidermal growth factor receptor 2–positive (HER2+), hormone receptor–positive (HR+) advanced or metastatic breast cancer (mBC) and retreatment with T-DXd (DESTINY-Breast Guide) Sarah Sammons, Giuseppe Curigliano, David W. Cescon, Erika Hamilton, Wolfgang Janni, Yeon Hee Park, Jian Zhang, Viktor Komlósi, Daniel Rodriguez Duque, Alan Zeng, Adrien Decque, Waltraud Pasteiner, Sara M. Tolaney 1795TiP Sections: Plain Language Summary | Background | Study Design | Key Inclusion Criteria | Key Exclusion Criteria | Key Study Endpoints | Abbreviations | Acknowledgments | Disclosures | References Study start: September 4, 2026; recruiting patients Key Eligibility Criteria • Patients aged ≥18 years • HER2+ (IHC 3+ or IHC 2+/ISH+)/HR+, de novo or recurrent mBC • No prior chemotherapy or HER2-targeted therapy for advanced or mBC • ECOG PS 0 or 1 • DFI >6 months • Asymptomatic brain metastases allowed 1. Upfront Treatment Phase: T-DXd + pertuzumab 18-24 cycles N = 200 2. Maintenance Treatment Phase: Trastuzumab + pertuzumab + ET + palbociclib until PD or unacceptable toxicity PD 3. T-DXd Retreatment Phase: Retreatment with T-DXd post-progression 1L Treatment | 2L Treatment Countries with participating study sites Australia, Canada, China, France, Germany, Italy, Republic of Korea, and US Key Study Endpoints Primary endpoint: PFS rate at 24 months [remaining fine print illegible at this resolution] [Body text of Plain Language Summary, Background, Inclusion/Exclusion criteria, endpoints and references too small to transcribe reliably] [Slide 2] Study Design • DB-Guide (NCT07683754) is a global, single-arm, open-label trial that aims to investigate the efficacy and safety of sequencing 1L T-DXd + pertuzumab followed by a palbociclib-containing maintenance regimen and T-DXd retreatment after disease progression in patients with HER2+, HR+ advanced or mBC • In the upfront treatment phase, approximately 200 patients will receive 18-24 cycles (≈12-17 months) of T-DXd 5.4 mg/kg intravenously (IV) every 3 weeks (Q3W) + pertuzumab 840 mg IV loading dose, then 420 mg IV Q3W thereafter – Patients who achieve a confirmed CR per RECIST v1.1 can switch to maintenance therapy earlier after receiving at least 8 cycles of T-DXd + pertuzumab. Confirmation of CR should be obtained during the next protocol-defined scheduled tumor assessment. • In the maintenance phase, patients with at least stable disease (SD) after upfront treatment will transition to trastuzumab 6 mg/kg IV Q3W + pertuzumab 420 mg IV Q3W + ET + palbociclib oral 125 mg daily (3 out of 4 weeks, every 4 weeks) • In the retreatment phase, eligible patients with PD on maintenance therapy will receive 2L T-DXd 5.4 mg/kg IV Q3W Study start: September 4, 2026; recruiting patients Key Eligibility Criteria • Patients aged ≥18 years • HER2+ (IHC 3+ or IHC 2+/ISH+)a/HR+, de novo or recurrent mBC • No prior chemotherapy or HER2-targeted therapy for advanced or mBCb • ECOG PS 0 or 1 • DFI >6 monthsc • Asymptomatic brain metastases allowed 1. Upfront Treatment Phase | 2. Maintenance Treatment Phase | 3. T-DXd Retreatment Phase T-DXd + pertuzumab 18-24 cyclesd N = 200 → Trastuzumab + pertuzumab + ETe + palbociclib until PD or unacceptable toxicity → PD → Retreatment with T-DXd post-progression 1L Treatment | 2L Treatment aAs classified by 2020 ASCO-CAP guidelines. b1 prior line of ET is allowed for mBC. cPatients who have received systemic therapy in the neoadjuvant or adjuvant setting are eligible, with a DFI >6 months from completion of systemic chemotherapy or HER2-targeted therapy (except T-DXd) to advanced or metastatic diagnosis of >6 months and at least >12 months since completion of prior T-DXd treatment in the neoadjuvant or post-neoadjuvant setting. dPatients who experience protocol-defined unacceptable toxicity during upfront therapy can transition to maintenance at any time. eAromatase inhibitor or fulvestrant. Countries with participating study sites Australia, Canada, China, France, Germany, Italy, Republic of Korea, and US [Slide 3] 1795TiP A global, interventional, open-label trial of trastuzumab deruxtecan (T-DXd)–based therapy followed by palbociclib, endocrine therapy (ET), and dual HER2 blockade in first-line (1L) human epidermal growth factor receptor 2–positive (HER2+), hormone receptor–positive (HR+) advanced or metastatic breast cancer (mBC) and retreatment with T-DXd (DESTINY-Breast Guide) Study start: September 4, 2026; recruiting patients T-DXd + pertuzumab 18-24 cycles N = 200 DESTINY Breast-Guide [Presenter standing in front of poster; body text not legible]

[Slide 1] ORIGINAL ARTICLE Anbenitamab plus HB1801 for HER2-Positive Advanced Breast Cancer Authors: Jiayu Wang, M.D., Huiping Li, M.D., Ph.D., Fei Xu, M.D., Tao Wu, M.D., Zhongsheng Tong, M.M., Jiuwei Cui, M.D., Xujuan Wang, M.M., +34, and Binghe Xu, M.D., Ph.D. Author Info & Affiliations Published October 23, 2026 DOI: 10.1056/NEJMoa2610345 Copyright © 2026 [Slide 2] RANDOMIZATION AND MASKING Eligible patients were randomly assigned in a 1:1 ratio to receive anbenitamab plus HB1801 (anbenitamab group) or docetaxel, trastuzumab, and pertuzumab (THP group). Randomization was performed centrally with the use of an interactive Web-response system and stratified according to previous neoadjuvant or adjuvant therapy (yes vs. no), hormone-receptor status (negative vs. positive), and visceral metastasis (yes vs. no). This was an open-label trial; the investigators, patients, trial staff, and primary sponsor were aware of the trial-group assignments. Patients with tumors expressing hormone receptors did not receive endocrine therapy or inhibitors of cyclin-dependent kinase 4 and 6 during this trial.

[Slide 1] Baseline Patient and Tumor Characteristics Characteristic | Total (n = 47) Age at registration, mean (SD), y | 53.5 (12.09); range 32–78 Female sex, n (%) | 46 (97.9) Race, n (%) White | 38 (80.9) Black or African American | 5 (10.6) Asian | 3 (6.4) Unknown | 1 (2.1) ER positive, n (%) | 31 (66.0) ER negative, n (%) | 16 (34.0) Enrolled after 1st intracranial progression, n (%) | 38 (80.9) Enrolled after 2nd intracranial progression, n (%) | 9 (19.1) Sites of Disease at Initial Metastatic Relapse, n (%) Intracranial only | 25 (53.2) Extracranial only | 17 (36.2) Both intracranial and extracranial | 5 (10.6) Backbone therapy Trastuzumab/pertuzumab (TP), n (%) | 44 (93.6) T-DM1 therapy, n (%) | 3 (6.4) Isolated CNS as site of first metastatic relapse, n (%) | 25 (53.2) SRS-to-registration interval, median (IQR), d | 21.5 (15.0–37.0); range 7–110 47 eligible and evaluable patients enrolled across 9 sites in the US Sarah Sammons MD [Slide 2] Primary Endpoint: Intracranial PFS Survival for all eligible patients (n=47) Survival Probability Months from Treatment Start ePFS iPFS oPFS OS Kaplan-Meier estimates, full analysis population (n = 47) Primary endpoint met: median iPFS not reached vs 5.0-mo historical control, one-sample log-rank p < 0.0001. Median follow-up 19.6 mo. Endpoint | Events, n (%) | Median | 6 mo | 12 mo | 18 mo Full cohort (n = 47) Intracranial PFS | 17 (36.2) | NR | 91% (79–97) | 64% (48–76) | 59% (42–73) Extracranial PFS | 2 (4.3) | NR | 100% | 98% (84–100) | 98% (84–100) Overall survival | 1 (2.1) | NR | 100% | 100% | 100% 95% CI in parentheses. NE, not estimable; NR, not reached. Sarah Sammons MD Survival table Endpoint: Intracranial PFS (primary), extracranial PFS, OS - single arm, full analysis population Full cohort - Intracranial PFS | 47 | NR Full cohort - Extracranial PFS | 47 | NR Full cohort - Overall survival | 47 | NR P: one-sample log-rank p < 0.0001 (vs 5.0-mo historical control) iPFS 6 mo 91% (79-97) iPFS 12 mo 64% (48-76) iPFS 18 mo 59% (42-73) ePFS 6 mo 100% ePFS 12 mo 98% (84-100) ePFS 18 mo 98% (84-100) OS 6/12/18 mo 100% [Slide 3] Patient Characteristics Variable | n | % No. of Patients | 30 Age, median (range) | 48 (33-65) Karnofsky Performance Status 90 | 8 | 27% 80 | 7 | 23% 70 | 11 | 37% 60 | 4 | 13% Receptors HR-/HER2+ | 14 | 46% HR+/HER2+ | 16 | 53% Radiation Therapy Administered During Screening Window WBRT | 9 | 30% fSRS/SRS | 3 | 10% CSI | 4 | 13% Spine RT | 3 | 10% None | 13 | 43% Radiation Therapy Administered Prior to Study WBRT | 13 | 43% fSRS/SRS | 13 | 43% CSI | 3 | 10% Spine RT | 4 | 13% None | 7 | 23% LMD Diagnosis at Baseline Radiographic Evidence | 21 | 70% Radiographic Evidence and Positive Cytology | 9 | 30% Variable | n | % Previous Lines of Systemic Therapy Trastuzumab/Pertuzumab | 27 | 90% Tucatinib/Trastuzumab/Xeloda | 10 | 33% Trastuzumab Deruxtecan | 11 | 37% Median Lines of Prior Systemic Therapy (range) | 3 (1-9) Concurrent Systemic Therapy at IT Treatment Start Trastuzumab Deruxtecan | 8 | 27% Trastuzumab/Pertuzumab | 5 | 17% Trastuzumab | 5 | 17% Tucatinib/Trastuzumab/Xeloda | 8 | 27% Prior Intrathecal Trastuzumab Y | 4 | 13% N | 26 | 87% Kamran Ahmed, MD

[Slide 1] Primary endpoint: PFS in overall patients (per BICR) [KM curve; y-axis Progression-free survival (%); x-axis Time (months) 0-20] Legend: Trastuzumab rezetecan; Physician's choice; + Censor | Trastuzumab rezetecan | Physician's choice Median PFS (95% CI), months | 16.4 (14.0-16.8) | 6.8 (5.7-6.9) Harzard ratio (95% CI) 0.27 (0.21-0.34) P<0.001 No. at Risk (months 0-20) Trastuzumab rezetecan: 277 265 254 243 238 223 208 191 180 161 143 131 110 109 90 84 80 47 46 45 26 Physician's choice: 274 262 204 174 166 135 105 82 72 51 40 35 24 23 17 14 12 8 8 7 4 Presenter: Ying Fan MD PFS, progression-free survival, BICR, blinded independent central review. Survival table Endpoint: PFS per BICR (primary endpoint), overall patients Arm | N | Median | Median CI Trastuzumab rezetecan | 277 | 16.4 months | 14.0-16.8 Physician's choice | 274 | 6.8 months | 5.7-6.9 HR: 0.27 HR CI: 0.21-0.34 P: P<0.001 [Slide 3] TRAEs Trastuzumab rezetecan N = 276 [partially legible] | Physician's choice of chemotherapy N = 271 [partially legible] AE | T rezetecan any grade | T rezetecan grade >=3 | Chemo grade >=3 | Chemo any grade Anaemia | 92.8 | 42.8 | 4.8 | 59.0 White blood cell count decreased | 88.4 | 56.5 | 35.4 | 84.5 Neutrophil count decreased | 85.9 | 65.2 | 54.2 | 84.9 Platelet count decreased | 77.5 | 35.9 | 1.8 | 27.7 Nausea | 62.7 | 1.8 | 0.0 | 10.8 Aspartate aminotransferase increased | 61.6 | 0.7 | 4.8 | 59.0 Alanine aminotransferase increased | 58.7 | 1.4 | 3.0 | 55.0 Alopecia | 50.7 | - | - | 42.1 Decreased appetite | 50.0 | 1.4 | 0.7 | 15.9 Weight decreased | 47.5 | 1.1 | 0.4 | 15.5 Vomiting | 40.9 | 2.5 | - | 8.9 Hypertriglyceridaemia | 33.0 | 2.5 | 3.7 | 22.5 Hypoalbuminaemia | 30.8 | - | - | 12.5 Diarrhoea | 27.9 | 1.4 | - | 4.8 Urinary tract infection | 26.1 | 0.4 | 0.4 | 19.2 Gamma-glutamyltransferase increased | 25.4 | 2.5 | 3.7 | 24.0 Hypokalaemia | 25.4 | 7.6 | 2.6 | 13.3 Malaise | 23.2 | 4.7 | 0.4 | 13.3 Fatigue | 23.2 | 1.8 | 0.4 | 8.9 Lymphocyte count decreased | 21.7 | 11.2 | 4.4 | 17.0 Constipation | 20.7 | 0.4 | - | 9.2 Hypoaesthesia | 4.3 | - | 2.6 | 24.4 Interstitial lung disease | 2.5/1.1 | | | Legend: Trastuzumab rezetecan - any grade; Trastuzumab rezetecan - Grade >=3; Physician's choice of chemotherapy - any grade; Physician's choice of chemotherapy - Grade >=3 Incidence (%) [axis 100-0-100] TRAEs occurred in >20% of patients in either group and Interstitial lung disease are presented. Presenter: Ying Fan MD [Slide 4] TRAEs in >=25% of patients (either group) TQB2102 | Chemotherapy AE | TQB2102 any grade | TQB2102 grade >=3 | Chemo grade >=3 | Chemo any grade Leukopenia | 76.6 | 21.6 | 14.0 | 59.8 Neutropenia | 74.3 | 30.1 | 23.2 | 59.8 Anemia | 71.0 | 12.6 | 1.8 | 50.2 Nausea | 65.1 | 3.3 | - | 15.5 AST increased | 63.2 | 1.9 | 2.2 | 38.7 ALT increased | 60.6 | 2.2 | 1.1 | 42.1 Fatigue | 55.0 | 5.6 | 2.2 | 26.6 Decreased appetite | 54.3 | 2.6 | 0.4 | 10.0 Thrombocytopenia | 50.9 | 14.9 | 0.7 | 13.3 Hypokalemia | 47.2 | 17.8 | 1.5 | 5.2 Diarrhea | 46.5 | 3.3 | 0.4 | 3.7 gamma-GGT increased | 42.0 | 6.3 | 2.6 | 14.8 Weight decreased | 42.0 | 1.5 | - | 4.4 Vomiting | 40.9 | 3.3 | - | 6.3 Hypoalbuminemia | 38.3 | 0.4 | - | 7.4 Lymphopenia | 36.1 | 14.1 | 1.5 | 13.7 Alopecia | 30.9 | - | - | 36.5 Peripheral hypesthesia | 6.3 | - | 3.0 | 32.8 Legend: TQB2102, Any Grade; Chemo, Any Grade; TQB2102, Grade>=3; Chemo, Grade>=3 - The most common grade >=3 TRAEs for both TQB2102 and chemotherapy were hematologic toxicities. - ILD occurred in 2.2% of patients (grade 3, 1.1%) in the TQB2102 group, and no ILD-related deaths occurred. Abbreviations: TRAEs, treatment-related adverse events; ILD, interstitial lung disease. Presenter: Shusen Wang, MD

Key take aways: stage I TNBC Outcomes for stage I TNBC are good — and extremely good for TIL-high TNBC No proper RCTs showing the benefit of chemotherapy If chemotherapy is used, guidelines provide several options AC-T (± carboplatin) · TC · carboplatin/taxane Limited consensus on neoadjuvant vs adjuvant size threshold for a neoadjuvant approach varies between 1 and 2 cm Academic trials ongoing to tailor treatment based on TILs, ctDNA and imaging

ESMO OPEN ORIGINAL ARTICLE · Articles in Press, 108639, October 26, 2026 · Open Access Ultra-sensitive MRD Assessment in Long-term Complete Responders to Anti-HER2 Therapy in HER2-positive Metastatic Breast Cancer: Initial Results from HER2-CR Study Kazuki Nozawa · Yukinori Ozaki · Haruru Kotani · Yumiko Akita · Ayumi Kataoka · Akiyo Yoshimura · Kana Yukino · Kiyomi Nonogaki · Toshimi Takano · Takayuki Ueno · Yumi Wanifuchi-Endo · Tomoko Asano · Nanae Horisawa · Mitsuo Terada · Kaori Terata · Derrick Renner · Angel Rodriguez · Minetta C. Liu · Fumikata Hara · Tatsuya Toyama · Hiroji Iwata Subgroup | n | ctDNA trajectory Overall 37: 30 | 2 | 1 | 4 Best overall response Complete response (CR) 27: 24 | 2 | 1 Partial response (PR) 10: 6 | 1 | 3 Anti-HER2 therapy status Ongoing 28: 22 | 2 | 1 | 3 Discontinued 9: 8 | 1 Persistent ctDNA-negative; ctDNA-clearance (pos→neg); Late conversion (neg→pos); Persistent ctDNA-positive Radiographic residual disease at baseline. Raw serial ctDNA trajectories; separate two-sided Fisher's exact tests (2×2), unadjusted for multiplicity; *P < 0.05 Persistent negative: 22/23 vs 8/14, *P = 0.007 Clearance (pos→neg): 1/23 vs 1/14, P = 1.00 Late conversion (neg→pos): 0/23 vs 1/14, P = 0.38 Persistent positive: 0/23 vs 4/14, *P = 0.015 No residual disease (n = 23); With residual disease (n = 14) A Clinical outcome at data cut-off Maintained CR/PR: 8/9 | 2/5 Progressive disease: 1/9 | 3/5 Patients (%)

[Slide 2] HER2-CR Study Design Multicenter, prospective, observational study in 3 Japanese institutions; enrolment Oct 2023 - Mar 2025; data cut-off 1 Apr 2026 First study to evaluate ctDNA in long-term responders to anti-HER2 therapy using a WGS-based ultrasensitive MRD test Key eligibility HER2-positive metastatic / unresectable breast cancer CR or PR maintained ≥ 2 years on anti-HER2 therapy Enrolled whether therapy was ongoing or discontinued 48 patients enrolled Tumor-informed ctDNA assay Signatera™ Genome (Natera, Inc): WGS-based multiplex PCR assay Detection at parts-per-million level Personalized variant panel obtained for 77% (37/48) – evaluable cohort Exclusions: Pathology/concordance fail: 8 Low tumor DNA quality: 3 Serial plasma sampling Baseline, then every 3 months (maximum 5 draws) 142 samples analyzed; median 4 per patient Median 10.6 (range 2.4-24.6) years on anti-HER2 therapy; median follow-up 17.8 (0-23.2) months Primary endpoint: ctDNA-negativity rate, at sample level and at patient level Patient-level trajectories: persistent-negative / ctDNA-clearance / ctDNA-emergence / persistent-positive Secondary: association with best overall response, radiographic residual disease (RECIST v1.1) and treatment status ctDNA, circulating tumor DNA; WGS, whole-genome sequencing Kazuki Nozawa, MD [Slide 3] ctDNA Status in Each Patient Swimmer plot of the 37 analysable patients, sorted by total time on anti-HER2 therapy Years since initiation of anti-HER2 therapy (dashed = off Tx) | HR | ADC use | Imaging | Tx status (off duration) Months from first ctDNA assessment Legend: ctDNA negative; ctDNA positive; PD on imaging; Off-Tx event; On Tx; Off Tx Key observations 129/142 assessments (90.8%) ctDNA-negative; median 4 draws per patient Persistent ctDNA negativity associated with lack of PD both on and off therapy (97%, 29/30) All 5 patients who were ctDNA-positive at the final draw had RRD at baseline ctDNA positivity preceded or coincided with PD in 100% (3/3) patients who had RRD at baseline and progressed 1 patient with PD and no prior ctDNA positivity had no ctDNA testing within 6 months of event One patient had only 1 draw due to transfer to another hospital. ADC, antibody-drug conjugate; ER, estrogen receptor; No res., no radiographic residual disease; PD, progressive disease; T-DM1, trastuzumab emtansine; T-DXd, trastuzumab deruxtecan. Kazuki Nozawa, MD [Slide 4] ctDNA Results: Primary Endpoint N = 37; 142 plasma samples; median follow-up 17.8 months 90.8% (95% CI 84.9-95.0) 129 / 142 plasma samples ctDNA-negative 83.8% (95%CI 68.0-93.8) 31 / 37 patients ctDNA-negative at initial draw 81% (95% CI 65-92) 30 / 37 patients persistently ctDNA-negative 86% (95% CI 71-96) 32 / 37 ctDNA-negative at the final assessment Subgroup | n | ctDNA trajectory Overall | 37 | 30 persistent-negative, 2 clearance, 1 emergence, 4 persistent-positive Best overall response: Complete response (CR) | 27 | 24, 2, 1 Partial response (PR) | 10 | 6, 1, 3 Anti-HER2 therapy status: Ongoing | 28 | 22, 2, 1, 3 Discontinued | 9 | 8, 1 Legend: Persistent ctDNA-negative; ctDNA-clearance (pos→neg); ctDNA emergence (neg→pos); Persistent ctDNA-positive ctDNA further stratified radiographic residual disease (RRD) ctDNA-negative at final assessment: No RRD: 100% (23/23); with RRD: 64% (9/14); P = 0.005 Patients with RRD Maintained CR/PR: always ctDNA-negative: 89% (8/9); ever ctDNA-positive: 40% (2/5) 75% (3/4) persistently ctDNA-positive patients with RRD progressed Two-sided Fisher exact test, unadjusted for multiplicity. CI, confidence interval; CR, complete response; PR, partial response. Kazuki Nozawa, MD

Conclusions First prospective, multicenter evaluation of MRD monitoring in long-term responders to anti-HER2 therapy Sustained mCR • Most patients (86%, 32/37) with sustained response to anti-HER2 therapy were ctDNA-negative at the last available assessment • median duration of anti-HER2 therapy: 10.6 years Beyond imaging • ctDNA restratified radiographic residual disease (RRD): ○ 100% of patients with no RRD had a ctDNA-negative result at final assessment, vs. 64% of patients with RRD (P = 0.005) • ctDNA-negative residual lesions behaved indolently • Persistent ctDNA positivity was associated with clinical progression Next step • First molecular rationale supporting prospective trials of ctDNA-guided de-escalation or discontinuation of anti-HER2 therapy in durable responders • Follow-up study is ongoing ○ (NCT06663787) In long-term responders to anti-HER2 therapy, molecular complete response refines prognosis beyond durable clinical response

Rapid Oral session LBA17 - HER2-PET imaging to predict trastuzumab deruxtecan (T-DXd) efficacy in advanced breast cancer: The phase II PREDICT-HER trial Speakers Juan Jin (Shanghai, China) Lecture Time 10:37 - 10:42 Abstract Background T-DXd has shown remarkable efficacy in HER2-positive, HER2-low, and HER2-ultralow breast cancer (BC). However, single-site biopsy IHC fails to represent whole-body lesion status or in vivo antibody-target binding. HER2-PET imaging provides a non-invasive and whole-body visualization of HER2 target availability. This technology holds significant potential to refine patient selection and more accurately predict the efficacy of T-DXd. Methods This single-arm phase II trial enrolled advanced BC pts scheduled to receive T-DXd. HR+HER2-low/zero pts must have progressed on prior CDK4/6 inhibitors. Participants underwent baseline 18F-FDG PET/CT and HER2-PET/CT using 18F- or 68Ga-labeled affibody molecules within one month prior to T-DXd initiation. Based on HER2-PET visual assessment of active lesions, pts were categorized into "HER2-PET-positive" (presence of HER2-avid lesions) and "HER2-PET-negative" (mild or no HER2 radiotracer uptake). The primary objective was to evaluate whether the negative predictive value (NPV) of HER2-PET-negative status for objective response was ≥85%. Correlations between normalized HER2 SUVmax and T-DXd response were also analyzed. Results By April 2026, 89 pts were enrolled (HER2-positive: 27; HER2-low: 54; IHC 0: 8) with a median of 1 prior line of anti-HER2 therapy or chemotherapy. mPFS was significantly longer in the HER2-PET-positive group vs. the HER2-PET-negative group (n=72 vs. n=17; 13.6 vs. 5.1 mo; multivariable-adjusted HR: 0.11, 95% CI: 0.05–0.26). Among 79 pts with measurable target lesions, the NPV of HER2-PET-negative status for T-DXd response was 88.2% (15/17), reaching 100% (10/10) in HER2-low/zero pts. The positive predictive value of HER2-PET-positive status in HER2-positive pts was 93.8% (15/16). Normalized HER2 SUVmax was higher in responders vs. non-responders (8.1 vs. 3.9, P < 0.001). Conclusions The initial results met the primary endpoint. HER2-PET redefines whole-body HER2 status in advanced BC. High NPV and striking PFS differences indicate that HER2-PET accurately identifies non-responders and could spare PET-negative pts from futile T-DXd. Integrating HER2-PET with IHC will optimize patient selection, particularly in HER2-low/zero BC.

Proteomic Changes Support Multiple Co-Existing Mechanisms of T-DXd Resistance • ABCB1 and ABCC1 protein expression increased in 5/7 (71%) and 4/7 (57%) T-DXd-treated tumors, reinforcing the NGS findings. Post vs Pre Protein Fold Change of candidate proteins (T-DXd) Pt1 Pt3 Pt6 Pt7 Pt8 Pt18 Pt23 Receptor Profile: HER2+/HR+ HER2+/HR- HER2-/HR+ Paolo Tarantino, MD, PhD ALICANTE AUDITORIUM - HALL 6

# 1879P – Real-World Treatment Patterns and Outcomes of Inavolisib in Breast Cancer Patients Maryam Mooghali, MD, MSc; Nicole Casasanta, MD; Mathew Andersen, BA; Ian Krop, MD, PhD; Eric Winer, MD; Joseph S. Ross, MD, MHS; Maryam Lustberg, MD, MPH Department of Internal Medicine, Yale School of Medicine, New Haven, CT, USA; Yale Cancer Center, Yale School of Medicine, New Haven, CT, USA YaleNewHavenHealth Smilow Cancer Hospital | Yale Cancer Center | Yale School of Medicine Internal Medicine INTRODUCTION • HR-positive/HER2-negative breast cancer is the most common advanced breast cancer subtype; endocrine therapy plus a CDK4/6 inhibitor is standard first-line therapy. • Activating PIK3CA mutations occur in 35%-40% of tumors and are linked [to] pathway activation, endocrine resistance, and poorer outcomes. • [partially obscured] ...elective inhibitor of the PI3K p110α catalytic subunit that ...proved PFS and OS in patients with PIK3CA-mutated, HR-...ive advanced breast cancer recurring during or within ...ing adjuvant endocrine therapy in the INAVO120 trial. ...lisib with palbociclib and fulvestrant in October 2024. METHODS [partially obscured by presenter]: Source: Retrospective observational cohort study ...base (October 10, 2024 - March 31, 2026). ...tients with breast cancer who initiated inavolisib ≥6 ...cutoff date. ...es: monthly inavolisib initiations, 6-month overall ...d laboratory-based adverse events; dose reduction, ...ation; and acute health care utilization. Table 1 [partially obscured] ...nts with Breast Cancer Initiating Inavolisib | Overall Cohort, N = 436: 64 (55–72); 427 (97.9%); 345 (79.1%); 40 (9.2%); 21 (4.8%); 33 (7.6%); 388 (89.0%); 27.0 (23.5–31.3); 282 (64.7%); 76 (17.4%); 387 (88.8%) [row labels obscured] RESULTS KEY FINDINGS Inavolisib was rapidly adopted following approval, with real-world overall survival consistent with that observed in INAVO120. Substantial toxicity, particularly metabolic adverse events, and frequent acute health care utilization were observed. 49 | Median monthly initiation (IQR 43–54) 90.3% | 6-month overall survival (95% CI 87.4%–93.2%) 82.6% | Acute healthcare utilization within 30 days after an AE Figure – Six-month Overall Survival of Breast Cancer Patients on Inavolisib [KM curve; x axis Time Since Inavolisib Initiation, months 0-6; numbers at risk rows: Overall cohort, Triplet therapy: Inavolisib + ET + CDK4/6i, Doublet therapy: Inavolisib + ET, Doublet therapy: Inavolisib + CDK4/6i, Monotherapy: Inavolisib — small-print values not reliably legible] TREATMENT MODIFICATION Dose reduction | 123 (28.2%) Interruption (>14-day gap between prescriptions) | 84 (19.3%) Discontinuation (No subsequent prescription within 30 days) | 253 (58.0%) Time to first event in days, median (IQR): dose reduction: 62 (39–131); interruption: 70 (44–144); discontinuation: 76 (30–192) ACUTE CARE UTILIZATION (Within 30 days after an adverse event) ED visit | 307 (70.4%) Hospitalization | 193 (44.3%) ICU admission | 52 (11.9%) Length of stay in days, median (IQR): hospital admission: 4 (1–8); ICU admission: 3 (2–4). Table 2 – Adverse Events After Initiation of Inavolisib Among Patients with Breast Cancer Adverse Event (AE) | Any-grade AE, n (%) | Median (IQR) time to first any-grade AE, days | Grade 3-4 AE, n (%) | Median (IQR) time to first grade 3-4 AE, days Diarrhea | 62 (14.2) | 29 (14-86) | - | - Fatigue | 90 (20.6) | 21 (5-76) | - | - Nausea/vomiting | 90 (20.6) | 21 (8-40) | - | - Ocular toxicity | 31 (7.1) | 29 (12-126) | - | - Pneumonitis | 6 (1.4) | 12 (8-48) | - | - Rash | 23 (5.3) | 28 (19-81) | - | - Stomatitis/mucositis | 42 (9.6) | 41 (13-99) | - | - Anemia | 334 (84.6) | 29 (13-70) | 93 (23.5) | 73 (26-140) Hyperglycemia | 258 (61.1) | 42 (22-89) | 120 (28.4) | 61 (32-127) Neutropenia | 270 (72.8) | 37 (21-77) | 153 (41.2) | 49 (27-93) Thrombocytopenia | 258 (63.9) | 34 (23-62) | 45 (11.1) | 61 (28-117) Transaminitis | 224 (57.3) | 30 (13-118) | 55 (14.0) | 96 (31-161) CONCLUSION • This real-world study of inavolisib in patients with breast cancer showed rapid and consistent uptake after FDA approval. • Early overall survival was consistent with that reported in INAVO120. • Metabolic and hematologic toxicities were frequent, and more than half of patients discontinued treatment during follow-up. Acute health care utilization was common after adverse events. • Proactive monitoring, early toxicity management, and supportive care are necessary to ensure safe use of inavolisib in routine practice. • Mutant-selective PI3Kα inhibitors are under investigation as potentially less toxic therapeutics in PIK3CA-mutated advanced breast cancer. Funding: No external funding was used to support this project. Potential Competing Interests: Dr. Mooghali has no competing interests to disclose. Contact: maryam.mooghali@yale.edu

[Slide 2] Adjuvant or neoadjuvant? No RCTs addressing chemotherapy benefit in this population! Advantages of adjuvant approach • Exact tumor diameter • Accurate grade and histology • Confirmation of nodal status → Less risk of overtreatment Advantages of neoadjuvant approach • Prognostic information on pCR • Response-guided post-neoadjuvant treatment (capecitabine, olaparib) → Risk of overtreatment little evidence on capecitabine/olaparib for stage I St Gallen 2025 panel³ Threshold for neoadjuvant chemo in stage I TNBC [line chart: % favoring treatment vs Tumor size (cm) 1.0–3.0; HER2+ and TNBC] ~10% positive sentinel node in cN0 stage I TNBC at upfront surgery 1,2 Risk rises with size: T1a/b 5–9% · T1c 12–15% Mittendorf SSO abstract 2022, 2. Cortina Ann Surg Oncol 2024, 3. Burstein et al. Ann Oncol 2025 Speaker panel: Marleen Kok — Optimising systemic therapy and sequence for stage I TNBC [Slide 3] Prognostic factors in stage I TNBC Tumor size Stromal TILs Histology: Favorable: adenoid cystic, secretory, low-grade special types. pCR after NACT: 5-yr OS 97% vs 90% with residual disease (aHR 0.23)¹ Germline BRCA1: Worse OS mainly due to second primaries 2,3,4 Less established: grade, Ki-67, LVI, age, gene signatures such as TNBC-DX. Investigational: ctDNA De Graaf et al. ESMO Open 2025, 2. Zhong et al. CCR 2014, 3. Wang et al. BMC Cancer 2025, 4.Lynce et al. ASCO 2025 poster

Key takeaways and potential clinical implications • NeXT Personal MRD test detected ctDNA at pretreatment baseline in 95.1% of patients with stage I–III TNBC • Neoadjuvant therapy produced a rapid and deep molecular response: 80.8% of patients cleared ctDNA from Baseline-to-C3D1 • Residual ctDNA at C3D1 was predominantly detectable only at ultrasensitive levels: 85.7% of detections were below 100 PPM, including 57.1% below 10 PPM • ctDNA clearance at C3D1 was enriched in patients achieving pCR compared to those with non-pCR • Early molecular response differed significantly by tumor stage, occurring in 100% of stage I, 85.5% of stage II, and 44.4% of stage III patients, where 55.6% of stage III patients remained ctDNA-positive at C3D1 • These results suggest a potential clinical opportunity for de-escalation of standard systemic therapy without compromising outcomes. Carlos H Barcenas, MD, MSc

[Slide 1] ET-FES STUDY ER+/HER2- MBC pts Ist line therapy N = 147 → 18F-FES PET/CT + 18F-FDG-PET/CT → 18F-FES SUV≥2 → Endocrine therapy → 18F-FES SUV<2 → R 1:1 → Endocrine therapy / CT or ET + biological agents 124 pts → ET-FES HETEROGENEITY STUDY GROUP A | 100% Concordant: all lesions FDG & FES positive GROUP B | Partially concordant: < 100% FES positive GROUP C | 100% Discordant: 0% FES positive PRIMARY OBJECTIVE To compare the efficacy of first-line ET vs CT in patients with ER+/HER2- MBC and 18F-FES SUV<2 at baseline CT/PET scan. SECONDARY OBJECTIVES • to evaluate DCR/PFS/OS with ET in patients with 18F-FES SUV≥2; • to compare the DCR/PFS/OS with ET in patients with 18F-FES SUV≥2 with 18F-FES SUV <2; • to correlate ER expression in the primary tumor and overall 18F-FES-uptake in metastases; • to assess PFS and OS in all patients and by 18F-FES SUV value. Gennari A. [Slide 3] PROGRESSION-FREE AND OVERALL SURVIVAL Median Progression-free survival Group A | 18.1 months (95%CI 11.3-25.3) Group B/C | 12.0 months (95%CI 6.6-22.9) HR 1.36 (95%CI 0.88-2.10); p=0.162 Progression-free survival (%) | Years (0-11) No. at Risk (No. Cumulative Censors) Group A: 92 (0) | 53 (0) | 39 (0) | 29 (0) | 25 (0) | 21 (1) | 11 (6) | 9 (7) | 7 (9) | 4 (12) | 2 (14) | 0 (16) Group B+C: 31 (0) | 16 (0) | 9 (0) | 7 (0) | 6 (0) | 3 (1) | 2 (2) | 1 (2) | 0 (3) Median Overall survival Group A | 69.6 months (95% CI, 48.3-85.1) Group B/C | 42.5 months (95% CI, 27.0-70.4) HR = 1.87 (95% CI 1.16–3.01); p = 0.010 Overall survival (%) | Years (0-11) No. at Risk (No. Cumulative Censors) Group A: 92 (0) | 90 (0) | 80 (1) | 70 (1) | 57 (2) | 48 (5) | 33 (12) | 26 (15) | 21 (17) | 11 (27) | 5 (32) | 0 (36) Group B+C: 32 (0) | 28 (1) | 23 (1) | 17 (1) | 14 (1) | 11 (2) | 8 (3) | 3 (5) | 2 (6) | 0 (7) Gennari A. Survival table Endpoint: PFS and OS by FES/FDG concordance group (Group A vs Group B/C) Arm | N | Median | Median CI Group A (PFS) | 92 | 18.1 months | 11.3-25.3 Group B/C (PFS) | 31 | 12.0 months | 6.6-22.9 Group A (OS) | 92 | 69.6 months | 48.3-85.1 Group B/C (OS) | 32 | 42.5 months | 27.0-70.4 HR: PFS HR 1.36; OS HR 1.87 HR CI: PFS 0.88-2.10; OS 1.16–3.01 P: PFS p=0.162; OS p=0.010 [Slide 4] MULTIVARIATE ANALYSIS Prognostic Factor | PFS analysis (N=121) HR [95% CI] | P value | OS analysis (N=121) HR [95% CI] | P value Eterogeneity group: Group A | 1 (ref.) | 0.180 | 1 (ref.) | 0.005 Group B-C | 1.40 [0.86-2.27] | | 2.21 [1.30-3.75] | Age at study entry: <60 yrs | 1 (ref.) | 0.706 | 1 (ref.) | 0.337 ≥60 yrs | 0.89 [0.50-1.61] | | 1.39 [0.70-2.77] | Menopausal status: Pre/peri-menopausal | 1 (ref.) | 0.708 | 1 (ref.) | 0.519 Post-menopausal | 1.17 [0.52-2.61] | | 1.36 [0.53-3.48] | ER expression level: ≤50% | 1 (ref.) | 0.253 | 1 (ref.) | 0.885 >50% | 1.52 [0.72-3.25] | | 0.94 [0.42-2.13] | Site of metastases: Other | 1 (ref.) | 0.537 | 1 (ref.) | 0.310 Bone only | 1.15 [0.74-1.78] | | 1.30 [0.78-2.17] | No. of mts lesions | 1.15 [0.99-1.32] | 0.069 | 1.07 [0.91-1.26] | 0.393 Disease-free interval: >24 mos | 1 (ref.) | 0.875 | 1 (ref.) | 0.531 ≤24 mos | 0.96 [0.48-1.93] | | 1.25 [0.56-2.79] | Not applicable | 0.88 [0.54-1.44] | | 1.36 [0.79-2.35] | Gennari A.

[Slide 1] B-STRONGER-I prospective multisite observational trial (NCT06230185) Objective: Relate serial ctDNA dynamics to pathological response Neoadjuvant Treatment Enrollment 1: stage I-III TNBC (n= ~422) Baseline Pretreatment n=142 T0 Cycle 3 Day 1 (+21 d) n=78 T1 End of treatment ongoing T2 Total Baseline/C3D1 samples analyzed n=220 Surgery Pathological Complete Response Report n=44 pCR n=34 residual Primary endpoint 1 (N= ~360): Correlation between Molecular Residue Disease (test) and Pathological Complete Response Endpoint: PPA, NPA Estimated: High-risk (Non-pCR patients, n= ~179) Low-risk (pCR patients n= ~179) Carlos H Barcenas, MD, MSc. [Slide 2] High baseline ctDNA detection rate and significant stage-burden relationship - Baseline ctDNA detection rate: 95.1% (95% CI, 90.1-98.0%) - Baseline ctDNA burden increased across tumor stage (Kruskal-Wallis p<0.0001): - Median 13.83 PPM in stage I - Median 828.21 PPM in stage II - Median 22,732.23 PPM in stage III - 29.7% (19/64) of stage I/II baseline detections were in the ultrasensitive range <100 PPM [Box plot: Baseline ctDNA (PPM; log scale) by AJCC stage I, II, III; p<0.0001] Carlos H Barcenas, MD, MSc [Slide 3] Neoadjuvant therapy produces a rapid and deep molecular response - 98.6% (72/73) of patients had an early molecular response to neoadjuvant therapy from baseline to C3D1 (Fig 1 and 2) - Across all patients, median Baseline-to-C3D1 reduction in ctDNA levels was 99.8% (paired Wilcoxon p<0.001) (Fig 1) - 80.8% of patients cleared ctDNA from Baseline-to-C3D1, during neoadjuvant therapy (59/73; 95% CI, 69.9-89.1%) (Fig 2) - 85.7% (12/14) of residual C3D1 detections were below 100 PPM and 57.1% (8/14) were below 10 PPM (Fig 3) Fig1: ctDNA detection rate - T0 Baseline 135/142 95.1%; T1 C3D1 14/78 17.9% Fig2: ctDNA level (PPM; pseudo-log scale) T0 Baseline to T1 C3D1 [spaghetti plot; small legend text illegible] Fig3: ctDNA level (PPM; log scale), 10 PPM and 100 PPM thresholds; T0 Baseline: <10 PPM: 4.1% (3/73) | <100 PPM: 2[illegible]% (19/73); T1 C3D1: <10 PPM: 57.1% (8/14) | <100 PPM: 85.7% (12/14) Carlos H Barcenas, MD, MSc [Slide 4] C3D1 negativity is enriched in earlier-stage disease and correlates with pCR - ctDNA clearance or persistent negativity at C3D1 was observed in 7/7 (100%) stage I, 53/62 (85.5%) stage II, and 4/9 (44.4%) stage III patients (global exact p=0.0102; Stage I vs III (Fisher p=0.0337) - ctDNA negativity at C3D1 was less frequent in patients with non-pCR (21/30, 70.0%) than in patients achieving pCR (38/43, 88.4%] (one-sided Fisher exact p=0.049) [Bar chart: T1 clearance or persistent negativity (%) by AJCC stage: I 7/7 100.0%; II 53/62 85.5%; III 4/9 44.4%; Global p=0.010, Stage I vs III p=0.034] pCR status | C3D1 ctDNA detected | C3D1 ctDNA clearance pCR (n= 43) | 5 | 38 Non-pCR (n= 30) | 9 | 21 Carlos H Barcenas, MD, MSc

Efficacy | | HR+/HER2- BC (n = 82) | TNBC (n = 32) | HER2-BC (n = 23) | BOR, n (%) | Complete response | 1 (1.2) | 0 | 0 | | Partial response | 43 (52.4) | 24 (75.0) | 16 (69.6) | | Stable disease | 32 (39.0) | 6 (18.8) | 6 (26.1) | | Progressive disease | 6 (7.3) | 2 (6.3) | 1 (4.3) | | Not evaluable* | 0 | 0 | 0 | | ORR, n (%) | 53.7 (42.3, 64.8) | 75.0 (56.6, 88.5) | 69.6 (47.1, 86.8) | | DCR, n (%) | 92.7 (84.8, 97.3) | 93.8 (79.2, 99.2) | 95.7 (78.1, 99.9) | | Median DoR, months (95% CI) | NE (3.9, NE) | 5.5 (4.1, NE) | 7.0 (5.1, NE) | | Median PFS, months (95% CI) | NE (8.4, NE) | 7.2 (5.3, NE) | NE (6.5, NE) | Waterfall plots: HR+/HER2- BC (6.0 mg/kg Q3W, 8.0 mg/kg Q3W, 4.0 mg/kg Q2W, 6.0 mg/kg Q2W); TNBC (6.0 mg/kg Q3W, 6.0 mg/kg Q2W); HER2+ BC (6.0 mg/kg Q3W, 8.0 mg/kg Q3W) Best change from basline in target lesions (%) *The reason for being classified as not evaluable was that stable disease lasted for less than 42 days. BC, breast cancer; BOR, best overall response; CI, confidence interval; DCR, disease control rate, DoR, duration of response, HER2, human epidermal growth factor receptor 2; HR, hormone receptor; NE, not estimable; ORR, objective response rate; TNBC, triple negative breast cancer. Hongxia Wang, MD

[Slide 1] The Design of the PREDICT-HER trial Key Eligibility • Locally advanced or metastatic breast cancer • Scheduled for T-DXd monotherapy • Measurable disease (per RECIST 1.1) or non-measurable bone-only disease • Baseline HER2-PET/CT and FDG-PET/CT Primary Endpoint Negative predictive value (NPV) of HER2-PET-negative status for objective response (RECIST 1.1). Eligible patients treated with T-DXd (n = 89) 18F-HER2 (n = 50) / 68Ga-HER2 (n = 39) HER2-positive* (n = 28) / HER2-low/zero (n = 61) Patients with measurable target lesions (n = 80) — Evaluated for Primary Endpoint (ORR / NPV) Patients with non-measurable bone-only lesions (n = 9) — Included in PFS & Safety Analyses Data cut-off date: August 2026 Median follow-up duration: 13.1 months *: HER2 status was determined by the most recent pathology results. 18F-HER2: 18F-NOTA-HER2-BCH. 68Ga-HER2: 68Ga-NOTA-MAL-MZHER2. Juan Jin [Slide 2] Patterns of HER2–PET/CT Confronted with FDG–PET/CT HER2-PET positive (HER2-PET P): presence of ≥1 HER2-avid lesions (TBD ≥ 4) HER2-PET negative (HER2-PET N): no HER2-avid lesions (all lesions TBD < 4) [PET MIP images: Patient #82 FDG / 18F-HER2; Patient #34 FDG / 68Ga-HER2; Patient #68 FDG / 18F-HER2; Patient #22 FDG / 68Ga-HER2] Maximum intensity projection: Lesion uptake was considered positive when visually distinct from the surrounding background. Tumor-to-background difference (TBD) was defined as the lesion SUVmax minus the surrounding background SUVmean. Methodological note: TBD was selected based on its superior concordance with expert visual consensus over conventional ratio metrics (SUV/TBR/TMR). Juan Jin [Slide 3] HER2-PET Independently Predicts PFS Beyond Pathological Status Multivariable Cox Proportional Hazards Regression Analysis of PFS Variable | | HR (95% CI) | P HER2-PET status | HER2-PET P | Reference | | HER2-PET N | 5.46 (2.68–11.12) | <0.001 HER2 status | HER2-positive | Reference | | HER2-low/zero | 4.90 (2.02–11.88) | <0.001 Hormone receptor status, tracer type, visceral metastases, and prior lines of therapy were evaluated in univariable analysis but excluded from the multivariable model due to lack of statistical significance. PFS according to HER2 Status and HER2–PET/CT Group | mPFS (months) 95% CI HER2-positive+ HER2-PET P | NR (NR – NR) HER2-low/zero+ HER2-PET P | 11.9 (10.3 – NR) HER2-positive+ HER2-PET N | 9.0 (3.6 – NR) HER2-low/zero+ HER2-PET N | 4.8 (2.9 – NR) Y-axis: Progression-free survival (%); X-axis: Time (months) 0–20 Patients at risk (0, 5, 10, 15, 20) | HR (95%CI) | P value HER2-positive +HER2-PET P | 21 20 17 7 1 | Reference | - HER2-low/zero +HER2-PET P | 51 41 21 6 1 | 4.2 (1.0 - 17.8) | 0.05 HER2-positive +HER2-PET N | 7 5 1 0 0 | 7.6 (1.4 - 41.7) | 0.02 HER2-low/zero +HER2-PET N | 10 5 1 0 0 | 16.6 (3.5 - 77.7) | < 0.001 Juan Jin Survival table Endpoint: PFS according to HER2 status and HER2-PET/CT Arm | N | Median | Median CI HER2-positive + HER2-PET P | 21 | NR | NR – NR HER2-low/zero + HER2-PET P | 51 | 11.9 months | 10.3 – NR HER2-positive + HER2-PET N | 7 | 9.0 months | 3.6 – NR HER2-low/zero + HER2-PET N | 10 | 4.8 months | 2.9 – NR HR: vs HER2-positive+HER2-PET P (reference): HER2-low/zero+PET P 4.2; HER2-positive+PET N 7.6; HER2-low/zero+PET N 16.6 HR CI: 1.0 - 17.8; 1.4 - 41.7; 3.5 - 77.7 P: 0.05; 0.02; < 0.001 [Slide 4] Conclusion ➢ Overcoming Spatial Heterogeneity: Whole-body HER2-PET captures the systemic target landscape and in vivo target binding, far beyond the limits of localized tissue biopsies. ➢ Clinical Implications: Baseline HER2-PET serves as a powerful rule-out biomarker for T-DXd in advanced breast cancer, reliably identifying target-deficient patients with primary resistance, particularly in HER2-low/zero populations. Acknowledgements: We would like to thank the patients, their families, our study team. Clinical trial identification: NCT07586215 Juan Jin

[Slide 1] NeoTRIP: IMC Myeloid TME study NeoTRIP trial High-risk (T1c-T2N1, T3N0) or locally advanced eTNBC Randomised 1:1 Intention-to-treat n=280 Samples received n=266 C&I n = 132 C n = 134 Atezolizumab + Carboplatin + nab-paclitaxel Carboplatin + nab-paclitaxel Surgery Assess pCR Baseline 74/33 78/31 On-treatment 92/45 83/41 Post-treatment 108/58 112/58 NT-M: Myeloid Study (n. Patients/Patients with pCR at surgery) Longitudinal Image Mass Cytometry: • Spatial proteomics • 1µm resolution (standard) • FFPE tissue compatible Myeloid-focused 45 marker panel Myeloid: CD14 CD68 CD163 CD204 CD209 CD11b CD40 iNOS CD11c DC-LAMP CLEC9A MPO CD66b MCT Carboplatin Lymphoid: CD20 CD4 CD8 GZMB FOXP3 Checkpoint: IDO VISTA PD-L1 CD86 TIM4 PD-1 DNA and Proliferation: H3 Iridium Lamin B1 Ki67 Stroma: CD31 SMA Vimentin Calponin PDGFRβ PDPN Caveolin-1 Epithelial: panCK CK5/14 TROP2 GATA3 PanTME: HLA-ABC HLA-DR CD45 β-tubulin TME / Epithelial / Both [Slide 2] Baseline myeloid TME predicts response to C&I • CD204+M2-like TAMs predict lower pCR under C&I (OR 0.85, p=0.002) but not C alone (pinteraction =0.007) • Activation (M1-like TAMs) matters: • Combined score achieves high accuracy C&I response predictions: pinteraction = 3.7×10-4 pCR probability Combined score AUC Combined score: C 0.43, C&I 0.85 √%Ki67+ TME: C 0.40, C&I 0.75 √CD204+M2 density: C 0.49, C&I 0.71 C (72/31) C&I (68/32) Odds ratio Pinteraction (qBH) CD204+M2 0.007 (0.11) • Resolution matters: Myeloid pinteraction = 0.6 CD204+M2 pinteraction = 0.007 Density (cells/mm2) [Slide 3] Carboplatin accumulates in myeloid cells • Responders (pCR) have higher tissue Carboplatin, no difference between treatment arms (C&I vs C) • Intra-cellular Carboplatin accumulates preferentially in Myeloid cells (M2-like TAMs) – (Cytoplasm > Nucleus) log(Carboplatin) Arm Response TIM4+M2 Scaled mean Cavity-resident TIM4+ macrophages sequester effector T cells and impair anti-tumour CD8 immunity Cell phenotype at interface log2OR 5.0 ↑enriched 0.0 -5.0 ↓depleted -log10(padj) 0 2 4+ Carboplatin (Pt194) Andrew Chow, et al. Tim-4+ cavity-resident macrophages impair anti-tumor CD8+ T cell immunity. Cancer Cell, 2021, [Slide 4] Take-home messages 1 High-resolution Myeloid Atlas in eTNBC • Large randomized Ph.III trial • Longitudinal assessments (Baseline, On-treatment, Post-treatment) • 16 spatially organized myeloid phenotypes • 15 recurrent multicellular niches 2 De novo baseline biomarker for poor response to C&I • Baseline CD204+M2 density associates with lower pCR upon C&I and not C alone (pint=0.007) • Adds information beyond previously described predictors (i.e. TME proliferation) 3 Mechanistic insights • CD204+M2 cells are enriched in the interface zone and are checkpoint low. • CD204+M2 interactions within the Immunoreg. TAM niche associates with lower pCR upon C&I: • CLEC9A+DC, fibroblasts, Tregs and other TAMs 4 In situ Carboplatin mapping • pCR patients show higher tissue Carboplatin (independently from treatment arm) • Intracellular Carboplatin accumulates mostly in M2-like TAMs. • Carbolatin+TIM4+M2 emerge post-treatment and contact T cells further away from tumour

Conclusion • BFv demonstrated promising antitumor activity with a manageable safety profile in heavily pretreated patients with la/mTNBC who had all received prior Top1i-based ADCs. • The first pivotal trial for this indication worldwide is underway. Dr Yanchun Meng

681eP - Effect of Adding Carboplatin to Taxane Plus Dual HER2 Blockade on Pathologic Complete Response in Neoadjuvant HER2-Positive Breast Cancer: A Meta-Analysis Yakup Ergun, Hayati ARVAS Department of Medical Oncology, Bower Hospital, Diyarbakır, Türkiye, Medical oncology, Mardin Education and Research Hospital, Mardin, Türkiye BACKGROUND To evaluate whether the addition of carboplatin to taxane plus trastuzumab and pertuzumab (THP) in the neoadjuvant treatment of HER2-positive early breast cancer provides an incremental benefit in terms of pathologic complete response (pCR). METHODS A systematic literature search was conducted from inception in PubMed, Embase, and the Cochrane Library. Studies comparing neoadjuvant TCHP versus THP were included. The primary endpoint was pCR in the intention-to-treat (ITT) population. Prespecified subgroup analyses were performed according to hormone receptor (HR) status, lymph node status, clinical stage, and tumor grade. Effect sizes were calculated as odds ratios (ORs) with 95% confidence intervals (CIs) using a random-effects model. Heterogeneity was assessed using the I² statistic. RESULTS A total of 8 studies (2 prospective and 6 retrospective), comprising 2,529 patients, were included. In the ITT population, pCR rates were 62.2% in the TCHP arm and 61.7% in the THP arm. The addition of carboplatin did not result in a statistically significant improvement in pCR (OR 1.15; 95% CI 0.82–1.60; p=0.42; I²=68%). In the HR-negative subgroup, pCR rates were 74.5% versus 78.6% (OR 0.98; 95% CI 0.35–2.76), while in the HR-positive subgroup they were 54.4% versus 54.2% (OR 1.01; 95% CI 0.78–1.30), with no significant differences observed. Among node-negative patients, pCR rates were 67.8% and 67.3% (OR 1.09; 95% CI 0.60–1.98), and among node-positive patients 59.9% and 63.4% (OR 0.87; 95% CI 0.64–1.19). In clinical stage II disease, pCR rates were 63.0% and 65.7% (OR 0.88; 95% CI 0.60–1.29), and in stage III disease 59.8% and 63.6% (OR 0.87; 95% CI 0.48–1.55). Study or Subgroup | TCbHP Events | TCbHP Total | THP Events | THP Total | Weight | Odds Ratio M-H, Random, 95% CI Binwei Lin et al. | 17 | 38 | 22 | 43 | 8.5% | 0.77 [0.32, 1.85] HELEN-006 | 194 | 337 | 220 | 332 | 17.5% | 0.69 [0.50, 0.95] Meirong Zhou et al. | 36 | 76 | 20 | 38 | 9.7% | 0.81 [0.37, 1.77] Minhao Lv et al. | 35 | 46 | 16 | 26 | 6.8% | 1.99 [0.70, 5.63] neoCARHP | 253 | 384 | 245 | 382 | 17.8% | 1.08 [0.80, 1.45] Song Wu et al. | 68 | 103 | 44 | 83 | 12.5% | 1.72 [0.95, 3.12] Xiangmin Ma | 106 | 145 | 49 | 94 | 13.3% | 2.50 [1.45, 4.31] Xiaowei Qi et al. | 197 | 327 | 46 | 75 | 13.9% | 0.96 [0.57, 1.60] Total (95% CI) | | 1456 | | 1073 | 100.0% | 1.15 [0.82, 1.60] Total events | 906 | | 662 Heterogeneity: Tau² = 0.14; Chi² = 21.67, df = 7 (P = 0.003); I² = 68% Test for overall effect: Z = 0.81 (P = 0.42) Favours: THP | TCbHP CONCLUSIONS In neoadjuvant HER2-positive breast cancer, the addition of carboplatin to taxane plus dual HER2 blockade does not significantly improve pCR rates in the overall population or in any of the evaluated clinical subgroups. Based on the available evidence, a patient-tailored, selective approach appears more rational than the routine use of carboplatin REFERENCES 1. Lin B, et al. Technol Cancer Res Treat. 2023;22. 2. Chen XC, et al. Lancet Oncol. 2025;26(1). 3. Zhou M, et al. Ann Transl Med. 2022;10(24). 4. Lv M, et al. Gland Surg. 2020;9(6). 5. Gao HF, et al. J Clin Oncol. 2026;44(17). 6. Wu S, et al. World J Surg Oncol. 2024;22(1). 7. Ma X, et al. Medicine (Baltimore). 2022;101(40). 8. Qi X, et al. Int J Surg. 2026;112(1). CONTACT Yakup ERGUN, MD E-mail: [email address omitted] X account: @dr_yakupergun No conflicts of interest

[Slide 1] Patients and Methods • Key Inclusion Criteria – HER2+ breast cancer patients with leptomeningeal metastases identified by MRI and/or CSF – Life expectancy > 8 weeks – Left ventricular ejection fraction (LVEF) >50% – Karnofsky Performance Status (KPS) ≥ 60 – Age ≥ 18 • Key Exclusion Criteria – History of serious adverse event related to any of the study drugs or study drug components Patients could continue treatment with HER2-directed, hormonal, or chemotherapeutic agents if controlling systemic disease and LM developed on these therapies Trial Schema HER 2+ breast leptomeningeal disease Radiation therapy IT Trastuzumab and Pertuzumab twice weekly X 4 weeks at assigned dose Follow-up response assessment IT Trastuzumab and Pertuzumab once per week X 4 weeks IT Trastuzumab and Pertuzumab every two weeks X 4 weeks Follow-up response assessments q8 weeks Kamran Ahmed, MD [Slide 2] 10:15 - 11:45 Rapid oral 2: Breast cancer, advanced stage CHAIRS: ANTONIO LLOMBART CUSSAC, SHERENE LOI, PEDRAM RAZAVI Overall Survival and Leptomeningeal PFS a Cumulative Survival Months b Leptomeningeal PFS Months No at Risk (censored) Endpoint met, Median survival 36.2 months (19.4 – NR) with a 12-month OS of 79%, historical control 4.4 months Median LM-PFS 19.3 months (7.2 – NR) with a 12-month LM-PFS of 61% Kamran Ahmed Survival table Endpoint: OS and leptomeningeal PFS (single arm) Arm | N | Median | Median CI OS (single arm) | 30 | 36.2 months | 19.4 – NR LM-PFS (single arm) | 30 | 19.3 months | 7.2 – NR 12-month OS 79% 12-month LM-PFS 61% [Slide 3] 10:15 - 11:45 Rapid oral 2: Breast cancer, advanced stage CHAIRS: ANTONIO LLOMBART CUSSAC, SHERENE LOI, PEDRAM RAZAVI Response Rate Individual Patients Time from Treatment Initiation (Months) Response LM Progression CR PR SD Ongoing follow-up Best LM Response: CR 5 (17%) PR 5 (17%) and SD in 17 (57%) Kamran Ahmed Kamran Ahmed Radiation therapy followed by intrathecal (IT) trastuzumab/pertuzumab in the management of HER2+ breast leptomeningeal metastasis: Results from a phase II multi-institutional study, (INSPIRE-LM) ALICANTE AUDITORIUM - HALL 6 [Slide 4] 10:15 - 11:45 Rapid oral 2: Breast cancer, advanced stage CHAIRS: ANTONIO LLOMBART CUSSAC, SHERENE LOI, PEDRAM RAZAVI Conclusions The study met its predetermined endpoint for OS with the longest median OS reported to date in HER2+ LM. Complete responses were noted in a subset of patients. IT trastuzumab/pertuzumab should be considered in the management of HER2+ LM. Kamran Ahmed Kamran Ahmed Radiation therapy followed by intrathecal (IT) trastuzumab/pertuzumab in the management of HER2+ breast leptomeningeal metastasis: Results from a phase II multi-institutional study, (INSPIRE-LM) ALICANTE AUDITORIUM - HALL 6

[Slide 1] [Poster photo with first author standing beside it] 36P - Concomitant radiation therapy and sacituzumab govitecan in metastatic breast cancer patients: feasibility and outcomes from a real-life multicentre international cohort study Luca Visani, Marcin Kubeczko, Ivica Ratosa, Barbro K. Linderholm, Miroslawa Puskulluoglu, Malgorzata Pieniazek, Justyna Zubrowska, Karolina Winsko-Szczesnowicz, Anna Polakiewicz-Gilowska, Michal Jarzab, Jan Kindl, Renata Soumarova, Hana Studentova, Zuzana Bielcikova, Jan Sustr, Francesco Pantano, Marianna Valzano, Carlotta Becherini, Lorenzo Livi, Icro Meattini [affiliations: small print, not transcribed] BACKGROUND Sacituzumab Govitecan (SG) has demonstrated significantly increased efficacy in terms of progression free survival (PFS) and overall survival (OS) in patients with metastatic triple negative breast cancer (mTNBC) and hormone receptors (HR)-positive HER2-negative metastatic breast cancer (mBC). In the advanced disease setting, radiation therapy (RT) is commonly used as part of palliative therapy for pain relief. Moreover, in oligometastatic and oligoprogressive patients, stereotactic body radiotherapy (SBRT) may significantly improve local disease control and OS. We conducted a retrospective multicentre study to assess the safety of the concomitant administration of SG and RT in mBC patients. METHODS A retrospective analysis was conducted on patients with mBC treated at 13 leading European institutions across 5 different countries. Ablative RT was defined based on a biological dose threshold of a minimum of 50Gy EQD2(10) delivered in no more than 12 fractions, as per the European Society for Radiotherapy and Oncology (ESTRO) OligoCare study. The primary objective was to assess the association between RT and adverse events (AEs) greater than grade (G) 2. Acute toxicity | RT (n=76)* | Non-RT (n=222)* | p-value** Hematologic, n (%) Neutropenia | 38 (61.3) | 121 (61.1) | 1.0000 Thrombocytopenia | 8 (14.3) | 22 (11.5) | 0.6419 Anemia | 25 (46.3) | 72 (37.7) | 0.2727 Non-hematologic, n (%) Fatigue | 29 (49.2) | 70 (35.2) | 0.0670 Elevated transaminases | 4 (7.3) | 36 (19.0) | 0.0393 Edema | 3 (5.5) | 1 (0.5) | 0.0364 Nausea | 14 (24.1) | 30 (15.7) | 0.1685 Diarrhea | 11 (19.3) | 39 (19.9) | 1.0000 Radiation dermatitis | 10 (18.5) | 0 (0) | <0.0001 Table 1. Acute toxicity outcomes. Late toxicity | RT (n=76)* | Non-RT (n=222)* | p-value** Hematologic, n (%) Leukopenia | 6 (12.0) | 5 (3.4) | 0.0342 Thrombocytopenia | 4 (8.3) | 0 (0) | 0.0035 Anemia | 10 (20.8) | 6 (4.1) | 0.0009 Non-hematologic, n (%) Fatigue | 10 (19.6) | 7 (4.7) | 0.0024 Elevated transaminases | 3 (6.2) | 6 (4.1) | 0.6913 Edema | 2 (4.2) | 0 (0) | 0.0615 Skin fibrosis and atrophy | 1 (2.1) | 0 (0) | 0.2500 Pulmonary and cardiac toxicity | 2 (4.2) | 0 (0) | 0.0615 Radionecrosis | 1 (2.1) | 1 (0.7) | 0.4385 Table 2. Late toxicity outcomes. RESULTS Data from 298 consecutive patients treated with SG, with or without RT, were retrospectively evaluated. Seventy-six patients received RT immediately before (within 14 days) or during SG without interruptions, amounting to 79 concurrent RT treatments, while 222 patients did not receive RT. The median age at SG initiation was 53 years in the overall population (range, 27-86), and the median follow-up from SG initiation was 12 months (range 1–50). Most patients received SG as second line of systemic therapy for metastatic disease (42.0% of the overall population). The median total dose was 24 Gy (range, 6-60), delivered at a median dose per fraction of 4 Gy (range, 2-24). The median biologically effective dose (BED; alpha/beta=10) was 37.5 Gy (range, 9.6-112.5), and the median equivalent dose in 2-Gy fractions (EQD2) was 31.2 Gy (range, 8.0-93.8). Among the 79 RT courses, bone was the most frequently treated site (n=27, 34.2%). The relationship between RT administration and the development of ≥G3 toxicity was not statistically significant (p=0.6001) at chi-square test. The only acute toxicities significantly associated with RT were radiation dermatitis and edema of any grade (Table 1). Among late toxicities, fatigue, thrombocytopenia, anemia, and leukopenia of any grade were significantly more frequent in the RT group (Table 2). In the adjusted analysis for potential confounding factors, the association between RT and PFS reached nominal statistical significance (HR, 0.707; p=0.0233), with a median of 6.7 months vs 4.8 months in the no RT cohort. CONCLUSIONS Our real-world findings suggest that the combination of SG and concurrent RT does not increase the risk of severe acute and late toxicity. First author (contact at luca.visani@unifi.it) has no conflict of interest to declare related to the presented work. Università degli Studi Firenze | Azienda Ospedaliero Universitaria Careggi | FRO [Slide 2] [Close-up of poster; 'BEST POSTER' rosette attached top-right] 36P - Concomitant radiation therapy and sacituzumab govitecan in metastatic breast cancer patients: feasibility and outcomes from a real-life multicentre international cohort study Luca Visani, Marcin Kubeczko, Ivica Ratosa, Barbro K. Linderholm, Miroslawa Puskulluoglu, Malgorzata Pieniazek, Justyna Zubrowska, Karolina Winsko-Szczesnowicz, Anna Polakiewicz-Gilowska, Michal Jarzab, Jan Kindl, Renata Soumarova, Hana Studentova, Zuzana Bielcikova, Jan Sustr, Francesco Pantano, Marianna Valzano, Carlotta Becherini, Lorenzo Livi, Icro Meattini [affiliations: small print, not transcribed] BACKGROUND Sacituzumab Govitecan (SG) has demonstrated significantly increased efficacy in terms of progression free survival (PFS) and overall survival (OS) in patients with metastatic triple negative breast cancer (mTNBC) and hormone receptors (HR)-positive HER2-negative metastatic breast cancer (mBC). In the advanced disease setting, radiation therapy (RT) is commonly used as part of palliative therapy for pain relief. Moreover, in oligometastatic and oligoprogressive patients, stereotactic body radiotherapy (SBRT) may significantly improve local disease control and OS. We conducted a retrospective multicentre study to assess the safety of the concomitant administration of SG and RT in mBC patients. METHODS A retrospective analysis was conducted on patients with mBC treated at 13 leading European institutions across 5 different countries. Ablative RT was defined based on a biological dose threshold of a minimum of 50Gy EQD2(10) delivered in no more than 12 fractions, as per the European Society for Radiotherapy and Oncology (ESTRO) OligoCare study. The primary objective was to assess the association between RT and adverse events (AEs) greater than grade (G) 2. Acute toxicity | RT (n=76)* | Non-RT (n=222)* | p-value** Hematologic, n (%) Neutropenia | 38 (61.3) | 121 (61.1) | 1.0000 Thrombocytopenia | 8 (14.3) | 22 (11.5) | 0.6419 Anemia | 25 (46.3) | 72 (37.7) | 0.2727 Non-hematologic, n (%) Fatigue | 29 (49.2) | 70 (35.2) | 0.0670 Elevated transaminases | 4 (7.3) | 36 (19.0) | 0.0393 Edema | 3 (5.5) | 1 (0.5) | 0.0364 Nausea | 14 (24.1) | 30 (15.7) | 0.1685 Diarrhea | 11 (19.3) | 39 (19.9) | 1.0000 Radiation dermatitis | 10 (18.5) | 0 (0) | <0.0001 Table 1. Acute toxicity outcomes. Late toxicity | RT (n=76)* | Non-RT (n=222)* | p-value** Hematologic, n (%) Leukopenia | 6 (12.0) | 5 (3.4) | 0.0342 Thrombocytopenia | 4 (8.3) | 0 (0) | 0.0035 Anemia | 10 (20.8) | 6 (4.1) | 0.0009 Non-hematologic, n (%) Fatigue | 10 (19.6) | 7 (4.7) | 0.0024 Elevated transaminases | 3 (6.2) | 6 (4.1) | 0.6913 Edema | 2 (4.2) | 0 (0) | 0.0615 Skin fibrosis and atrophy | 1 (2.1) | 0 (0) | 0.2500 Pulmonary and cardiac toxicity | 2 (4.2) | 0 (0) | 0.0615 Radionecrosis | 1 (2.1) | 1 (0.7) | 0.4385 Table 2. Late toxicity outcomes. RESULTS Data from 298 consecutive patients treated with SG, with or without RT, were retrospectively evaluated. Seventy-six patients received RT immediately before (within 14 days) or during SG without interruptions, amounting to 79 concurrent RT treatments, while 222 patients did not receive RT. The median age at SG initiation was 53 years in the overall population (range, 27-86), and the median follow-up from SG initiation was 12 months (range 1–50). Most patients received SG as second line of systemic therapy for metastatic disease (42.0% of the overall population). The median total dose was 24 Gy (range, 6-60), delivered at a median dose per fraction of 4 Gy (range, 2-24). The median biologically effective dose (BED; alpha/beta=10) was 37.5 Gy (range, 9.6-112.5), and the median equivalent dose in 2-Gy fractions (EQD2) was 31.2 Gy (range, 8.0-93.8). Among the 79 RT courses, bone was the most frequently treated site (n=27, 34.2%). The relationship between RT administration and the development of ≥G3 toxicity was not statistically significant (p=0.6001) at chi-square test. The only acute toxicities significantly associated with RT were radiation dermatitis and edema of any grade (Table 1). Among late toxicities, fatigue, thrombocytopenia, anemia, and leukopenia of any grade were significantly more frequent in the RT group (Table 2). In the adjusted analysis for potential confounding factors, the association between RT and PFS reached nominal statistical significance (HR, 0.707; p=0.0233), with a median of 6.7 months vs 4.8 months in the no RT cohort. CONCLUSIONS Our real-world findings suggest that the combination of SG and concurrent RT does not increase the risk of severe acute and late toxicity. First author (contact at luca.visani@unifi.it) has no conflict of interest to declare related to the presented work. Università degli Studi Firenze | Azienda Ospedaliero Universitaria Careggi | FRO

[Slide 1] neoHIP: INTEGRATION OF ER AND IMMUNE BIOMARKERS HR status | Immune status | Arm A: THP | Arm B: THP + pembro | Δ pCR (B - A) HR- | PD-L1+ or sTILs+ | 64% (7/11) | 86% (6/7) | +22% HR- | PD-L1- and sTILs- | 71% (5/7) | 88% (7/8) | +17% HR+ | PD-L1+ or sTILs+ | 38% (5/13) | 80% (4/5) | +42% HR+ | PD-L1- and sTILs- | 36% (5/14) | 25% (3/12) | -11% Page D, ESMO 2026 3-biomarker rule for "pembro eligibility" Pembro eligible: +30% pCR with pembro Pembro not eligible: -11% pCR with pembro - Exploratory analysis - Very small subgroups - Data-driven biomarker integration - Unbalanced biomarker availability between arms: 55% of pembro arm / 78% of control arm - No test for interaction shown - However, nicely hypothesis-generating Maria Vittoria Dieci [Slide 2] NEOADJUVANT IMMUNOTHERAPY TRIALS IN HER2+ BC NEOADJUVANT | DESIGN · N | REGIMEN · SETTING | PCR RATE (ICI arm / Control (no ICI)) | Δ pCR | p IMpassion050* | Ph3 · randomized n=454 (226 vs 228) | ddAC->THP + atezolizumab; Neoadj -> adj | 62.4% / 62.7% | -0.3 | 0.96 APTneo | Ph3 · randomized n=661 (438 vs 223) | THCP / AC->TCHP + atezolizumab; Neoadj -> adj | 57.8% / 52.0% | +5.8 | 0.09 neoPATH | Ph2 · single-arm n=67 | HP + docetaxel + atezolizumab; Neoadj -> adj | 61% single arm | - | - neoHIP | Ph2 · randomized n=116 (58 vs 58) | THP x4 + pembrolizumab; Neoadjuvant only | 67.2% / 48.3% | +18.9 | 0.03 * Prematurely closed for unfavourable benefit/risk profile (Jan 2021) neoHIP: UNIQUE SIGNAL - Only randomized neoadjuvant ICI trial with a significant pCR gain Neoadj backbone: THP x 4. pCR rates in the control arm in line with CompassHER2, remarkable improvement with pembrolizumab. Small, phase II study. Population: 87% stage II, 84% cT1-2, 60% cN0 Additional chemotherapy before surgery: 4 pts (pembro arm) and 3 pts (control arm) Huober J, J Clin Oncol 2022; Gianni L, SABCS 2023; Ahn HK, JAMA Oncol 2022; Page D, ESMO 2026; McArthur HL, ASCO 2026; Tung N, J Clin Oncol 2026 Maria Vittoria Dieci [Slide 3] IS THERE A ROLE FOR NEOADJUVANT IMMUNOTHERAPY IN HER2+ EBC? IMpassion050: 3-YEAR DFS atezolizumab vs placebo + PH + chemo Subgroup | n | Atezo | Placebo | HR (95% CI) Overall | 434 | 92.9% | 88.5% | 0.71 (0.38-1.32) HR+ | 223 | 94.4% | 87.3% | 0.43 (0.17-1.13) HR- | 211 | 91.6% | 89.7% | 1.07 (0.46-2.53) PD-L1- | 224 | 94.6% | 84.1% | 0.44 (0.19-1.02) PD-L1+ | 210 | 91.1% | 92.9% | 1.38 (0.51-3.69) [Forest plot 0.1-10; <- Atezo better / Placebo ->] Huober J, ESMO Open 2026 SAFETY AEs (%; TRAEs in IMpassion050)) ICI arm / Control (no ICI) IMpassion050 · atezolizumab ~1 year (neoadj + adj): G3-4 47.3 / 42.2; Serious 15.0 / 10.7; AESI G3-4 10.6 / 6.7; G5 0.9 / 0.0 neoHIP · pembrolizumab 4 cycles: G3-4 12.1 / 10.3; Serious 3.4 / 1.7; G5 0.0 / 0.0 Huober J, J Clin Oncol 2022; McArthur HL, ASCO 2026 IMpassion050 - Early closure (2021): ~1/3 stopped treatment - Numerically better DFS (secondary and unpowered endpoint) with atezolizumab Against the neoHIP signal? - Benefit in PD-L1- / HR+: opposite of the neoHIP biomarker hypothesis -> Different population, chemotherapy, ICI type and setting The bar keeps rising - Very high cure rates - T-DXd is further reshaping neoadjuvant Tx Benefit vs risk - ICIs add irAEs - Low absolute rates of severe AEs in neoHIP Maria Vittoria Dieci [Slide 4] CONCLUSIONS 541RO · neoHIP Immunotherapy in HER2+ EBC - Interesting and biologically plausible biomarker signal from a small trial requiring validation. - As the role of neoadjuvant immunotherapy in HER2+ scenario remains to be clarified, these results at least provide the rationale to further test the neoHIP regimen in patients selected based on integrated biomarkers. Maria Vittoria Dieci

[Slide 1] Background • BFv is an ADC delivering monomethyl auristatin E (MMAE) to Nectin-4-expressing cells, a target highly expressed in TNBC. • For patients with prior exposure to Top1i-based ADCs, switching to a Nectin-4-targeted ADC with a distinct payload (MMAE) represents a rational therapeutic strategy. • We report efficacy and safety data from the cohort of la/mTNBC patients previously treated with Top1i-based ADCs in the phase 2 study of BFv (NCT06492005). Linker-Payload IDM | Valine-citrulline | PABC | MMAE A Circulation in Blood B Antigen binding and Internalization C Drug Release D Apoptosis Tumor Cell ADC Treatment Landscape in Late-Stage TNBC TROP2-Targeted ADCs: Sacituzumab Govitecan SN-38 (Top1i); Datopotamab Deruxtecan DXd (Top1i); Sacituzumab Tirumotecan KL610023 (Top1i) HER2-Targeted ADCs: Trastuzumab Deruxtecan DXd (Top1i) Nectin-4-Targeted ADC: Bulumtatug Fuvedotin MMAE Differentiated mechanism for post-Top1i ADC settings Dr Yanchun Meng [Slide 4] Safety TRAEs | N=62 n (%) TRAEs | 62 (100) Grade ≥3 TRAEs | 25(40.3) Serious TRAEs | 8(12.9) TRAEs leading to treatment interruption | 28(45.2) TRAEs leading to dose reduction | 5(8.1) TRAEs leading to discontinuation | 0 TRAEs leading to death | 0 TRAEs, % AST increased Neutrophil count decreased WBC decreased ALT increased Hypertriglyceridemia Hypercholesterolemia Anemia Hyperglycemia Asthenia PLT decreased GGT increased Grade ≥3 · Grade 1-2 TRAE, treatment-related adverse event; WBC, white blood cell; AST, aspartate aminotransferase; ALT, alanine aminotransferase; GGT, gamma-glutamyltransferase; PLT, platelet. Dr Yanchun Meng

[Slide 1] 10:15 - 11:45 Rapid oral 2: Breast cancer, advanced stage CHAIRS: ANTONIO LLOMBART CUSSAC, SHERENE LOI, PEDRAM RAZAVI HER2-CR: 81% persistently ctDNA-negative 90.8% (95% CI 84.9-95.0) 129 / 142 plasma samples ctDNA-negative 83.8% (95%CI 68.0-93.8) 31 / 37 patients ctDNA-negative at initial draw 81% (95% CI 65-92) 30 / 37 patients persistently ctDNA-negative 86% (95% CI 71-96) 32 / 37 ctDNA-negative at the final assessment • 37 of 48 assay success • Median 10.6 years on anti-HER2 therapy, 76% still on treatment • Persistently ctDNA-negative: 30/37 (81%) • CR on imaging: 100% ctDNA-; PR on imaging: 64% ctDNA- • Persistently negative: 29/30 progression-free, on or off therapy • Persistently positive: 3/4 progressed; • 1/4 recurrence with negative serial ctDNA (was it the same tumor or a new primary? • Off therapy (9 patients, up to 14.9 years): 8/9 persistently negative Design: prospective, 3 Japanese centers; HER2+ MBC with CR or PR ≥ 2 years on anti-HER2 therapy, on or off treatment; Signatera Genome (WGS-informed, LOD95 ~10ppm) at baseline and every 3 months, ≤ 5 draws; 142 samples; median follow-up 17.8 months. Discussant: Pedram Razavi Pedram Razavi Invited Discussant 1769RO, 1770RO and 4449RO ALICANTE AUDITORIUM - HALL 6 [Slide 2] 10:15 - 11:45 Rapid oral 2: Breast cancer, advanced stage CHAIRS: ANTONIO LLOMBART CUSSAC, SHERENE LOI, PEDRAM RAZAVI Why it matters, and why not to stop on a negative test yet Why it matters • One of the first studies to assess a molecular definition of “complete response” in outstanding responders • Consistent with an independent tumor-informed series (0/27 exceptional responders MRD-positive at year 1; Morganti et al, medRxiv 2026). • ctDNA refines disease status below the threshold of imaging: ctDNA-negative residual lesions behaved indolently; persistent positivity resulted in progression • A tumor-informed ctDNA assay is specific to tumor and robust to clonal hematopoiesis, reaches ultrasensitive ranges, and suits serial sampling on and off therapy with outcomes linked Some thoughts • Small study with a relatively short follow-up time to inform long-term outcomes • Discordance between imaging and ctDNA need to be clarified: 9/14 patients with residual disease were ctDNA negative; are these scars or imaging misclassification, or true residual disease missed by ctDNA: low shedders or indolent disease? • 23% could not be tested: a tumor-informed assay needs tissue, and a decade-old primary may not represent the metastatic clone • Mostly still on therapy (28/37): negativity under suppression is not negativity after withdrawal. Pedram Razavi Invited Discussant 1769RO, 1770RO and 4449RO ALICANTE AUDITORIUM - HALL 6 [Slide 3] Open questions and future directions Open questions • What defines a molecular CR fit for a stopping rule: how many negative draws, over how long, at what sensitivity, with which imaging, what if discordant with imaging? • Is negativity on therapy informative, or only after withdrawal? Should the first test be a planned interruption under surveillance? • A positive test without progression: continue, intensify, or treat locally? And what is the re-response rate after stopping? Future Directions • Stop in outstanding responders: HEROES (UNICANCER; ≥ 2 y CR/PR + ctDNA-negative → stop; n ≈ 170), STOP-HER2 (TBCRC 062; ≥ 3 y; stop vs continue cohorts; serial ctDNA), Free-HER • Curative intent trials in de novo HER2+ disease: SAPPHO (planned cessation; DFS by MRD), ARCHER (local consolidation) • Can ctDNA-guided adaptive escalation and de-escalation achieve molecular CR, with persistent molecular CR as the primary endpoint for potential cure? HERizon-Breast (ctDNA-guided curative-intent escalation / de-escalation study aiming to eradicate disease) HEROES NCT06450314 · STOP-HER2 NCT05721248 · Free-HER NCT05959291 · SAPPHO NCT06439693 · HERizon-Breast NCT07459673 · ARCHER NCT07053085 Discussant: Pedram Razavi

Conclusions The study met its predetermined endpoint for OS with the longest median OS reported to date in HER2+ LM. Complete responses were noted in a subset of patients. IT trastuzumab/pertuzumab should be considered in the management of HER2+ LM. Kamran Ahmed Radiation therapy followed by intrathecal (IT) trastuzumab/pertuzumab in the management of HER2+ breast leptomeningeal metastasis: Results from a phase II multi-institutional study, (INSPIRE-LM)

[Slide 2] HER2-PET Redefines the Tumor Landscape and Stratifies PFS Redefining HER2 Landscape: Correspondence Between Pathological HER2, HER2-PET, and HR Status (N=89) HER2 Status: HER2-IHC 0 7 | HER2-low 54 | HER2-positive 28 HER2-PET Status: HER2-PET-N 17 | HER2-PET-P 72 HR Status: HR+ 49 | HR- 40 Progression-Free Survival (PFS) according to HER2–PET Status | HER2-PET N | HER2-PET P Events | 14/17 | 27/72 Median PFS (95% CI), months | 5.1 (3.6–NR) | 13.6 (11.9–NR) Hazard ratio (95% CI): 4.09 (2.07-8.11); P<0.001 Progression-free survival (%) ; Time (months) 0 5 10 15 20 Patients at risk HER2-PET N | 17 | 10 | 2 | 0 | 0 HER2-PET P | 72 | 61 | 38 | 13 | 2 Survival table Endpoint: PFS Arm | N | Median | Median CI HER2-PET N | 17 | 5.1 | 3.6–NR HER2-PET P | 72 | 13.6 | 11.9–NR HR: 4.09 HR CI: 2.07-8.11 P: <0.001 [Slide 4] HER2-PET Accurately Predicts RECIST 1.1 Objective Response Confirmed Best Overall Response (N = 80) Best Response | HER2-positive (n = 24) | HER2-low/zero (n = 56) ORR: CR + PR | 18 (75.0%) | 25 (44.6%) CR | 4 (16.7%) | 2 (3.6%) PR | 14 (58.3%) | 23 (41.1%) SD | 3 (12.5%) | 23 (41.1%) PD | 3 (12.5%) | 8 (14.3%) DCR: CR + PR + SD | 21 (87.5%) | 48 (85.7%) ORR: Objective Response Rate. CR: Complete Response. PR: Partial Response. SD: Stable Disease. DCR: Disease Control Rate. Relation between patient RECIST 1.1 response and results of HER2 PET/CT | n | R | NR | NPV | PPV All patients HER2-PET P | 63 | 41 | 22 | | 65.1^b HER2-PET N | 17 | 2 | 15 | 88.2^a | HER2-low/zero patients HER2-PET P | 46 | 25 | 21 | | 54.3^d HER2-PET N | 10 | 0 | 10 | 100^c | Exact 95% confidence intervals as follows: a: 64% - 99%, b: 52% - 77%, c: 69% - 100%, d: 39% - 69%. n number, R responders, NR non-responders, NPV Negative Predictive Value, PPV Positive Predictive Value.

PATIENT-REPORTED OUTCOMES Are we measuring what patients actually value? Validated scales are necessary — but the domains reported determine what we can conclude. WHAT WE MEASURE WELL Global health status Physical and role functioning Fatigue and pain Social functioning Longitudinal group differences WHAT MAY BE MISSING Sexual desire and intimacy Dyspareunia and vaginal symptoms Cognitive fatigue and return to work Financial and caregiver burden Recovery after therapy HOW TO IMPROVE BR45 / PROMIS sexual function Targeted PRO-CTCAE Individual deterioration trajectories Patient-prioritized domains Post-treatment follow-up Sandra Franco, MD

Oncofertility Counseling Is Mandatory after Diagnosis Are healthcare providers compliant with this recommendation? PROJECT528 4,000 survey responses 67 countries 11% living with MBC ESMO BREAST CANCER Jennifer Merschdorf, Young Survival Coalition (YSC) https://project528.youngsurvival.org/ Did a healthcare provider discuss fertility preservation (such as egg freezing) with you? No | Yes [illegible regional bar values] Merschdorf J, ESMO Breast 2025 Matteo Lambertini, MD PhD @matteolambe matteo.lambertini@unige.it MADRID 2026 ESMO congress

[Slide 1] 16:15 - 17:45 Rapid oral 1: Gastrointestinal tumour, upper digestive CHAIRS: MARIA IGNEZ BRAGHIROLI, FENG WANG, LUCJAN WYRWICZ TQB2102-II-04 Study Design Exploratory Phase II study:NCT06767800 Key Eligibility Criteria - Unresectable locally advanced or metastatic GC/GEJC - HER2-positive status - No prior systemic therapy - ECOG performance status 0-1 Key Exclusion Criteria - Other active malignancies - Active autoimmune disease or interstitial lung disease - Symptomatic or untreated CNS metastases - Uncontrolled serious infection requiring systemic therapy Treatment A: TQB2102 7.5 mg/kg + TQB2450, IV Q3W (N=25) B: TQB2102 7.5 mg/kg + TQB2450, IV Q3W + capecitabine (N=37) C: TQB2102 6.0 mg/kg + TQB2450, IV Q3W + capecitabine (N=18) TQB2450 (benmelstobart, humanized IgG1 anti-PD-L1 monoclonal antibody): 1200mg, IV, Q3W Capecitabine: 1000 mg/m2, PO, BID, D1-14, Q3W Follow-up - Treatment continued until: Disease progression; Unaccepted toxicity; Withdrawal of consent; Reaching maximum treatment duration (24 months) per protocol Primary Endpoint: ORR Secondary Endpoints: DCR; TTR; DoR; PFS; OS; Safety Abbreviations: GC, gastric cancer; GEJC, gastro-esophageal junction cancer; HER2, human epidermal growth factor receptor 2; ECOG, Eastern Cooperative Oncology Group; CNS, central nervous system; Q3W, every three weeks; ORR, objective response rate; DCR, disease control rate; TTR, time to response; DoR, duration of response; PFS, progression-free survival; OS, overall survival. Presenter: Miao-Zhen Qiu Miaozhen Qiu TQB2102, a novel HER2 bispecific antibody-drug conjugate, combined with bemotrizumab with or without capecitabine as first-line therapy for HER2-positive advanced gastric/gastroesophageal junction cancer: A phase II study ALMERIA AUDITORIUM - HALL 9 [Slide 2] 16:15 - 17:45 Rapid oral 1: Gastrointestinal tumour, upper digestive CHAIRS: MARIA IGNEZ BRAGHIROLI, FENG WANG, LUCJAN WYRWICZ Efficacy- ORR and DCR - The doublet arm showed an ORR of 69.6% and a DCR of 95.7%. - For the triplet cohorts, the overall ORR was 84.2% with a DCR of 100%, wherein the 6.0 mg triplet subgroup achieved an ORR of 95.8% and a DCR of 100%. [Waterfall plot: Best Reduction From Baseline in Target Lesion (%); legend TQB2102 7.5mg+TQB2450; TQB2102 7.5mg+TQB2450+Cape; TQB2102 6mg+TQB2450+Cape] Treatment response | Doublet arm: TQB2102 7.5mg/kg +TQB2450 (N=25) | Triplet arm: TQB2102 7.5 mg/kg +TQB2450+Cape (N=37) | TQB2102 6.0 mg/kg +TQB2450+Cape (N=24) | Total (N=61) BOR CR, n (%): 0 | 1 (3.0) | 2 (8.3) | 3 (5.3) PR, n (%): 16 (69.6) | 24 (72.7) | 21 (87.5) | 45 (78.9) SD, n (%): 6 (26.1) | 8 (24.2) | 1 (4.2) | 9 (15.8) PD, n (%): 1 (4.3) | 0 | 0 | 0 ORR, n (%) 95% CI: 16 (69.6) (47.1%, 86.8%) | 25 (75.8) (57.7%, 88.9%) | 23 (95.8) (78.9%, 99.9%) | 48 (84.2) (72.1%, 92.5%) cORR, n (%) 95% CI: 14 (60.9) (38.5%, 80.3%) | 23 (69.7) (51.3%, 84.4%) | 22 (91.7) (73.0%, 99.0%) | 45 (78.9) (66.1%, 88.6%) DCR, n (%) 95% CI: 22 (95.7) (78.1%, 99.9%) | 33 (100) (89.4%, 100.0%) | 24 (100) (85.8%, 100.0%) | 57 (100) (93.7%, 100.0%) mTTR, months (min, max): 2.60 (1.18, 2.96) | 1.38 (1.25, 2.92) | 1.45 (1.22, 4.14) | 1.41 (1.22, 4.14) Abbreviations: BOR, best of response; CR, complete response; PR, partial response; SD, stable disease; PD, progressive disease; NE, not evaluable; ORR, objective response rate; cORR, confirmed objective response rate; DCR, disease control rate; CI, confidence interval. Presenter: Miao-Zhen Qiu Miaozhen Qiu TQB2102, a novel HER2 bispecific antibody-drug conjugate, combined with bemotrizumab with or without capecitabine as first-line therapy for HER2-positive advanced gastric/gastroesophageal junction cancer: A phase II study ALMERIA AUDITORIUM - HALL 9 [Slide 3] 16:15 - 17:45 Rapid oral 1: Gastrointestinal tumour, upper digestive CHAIRS: MARIA IGNEZ BRAGHIROLI, FENG WANG, LUCJAN WYRWICZ Efficacy- PFS - As of 03 Sep 2026, the median PFS was 12.68 months(95% CI, 5.78-NR), for the doublet arm and 11.04 months (95% CI, 9.07-NR) for the triplet arm; median PFS for the triplet arm was immature. - In the triplet arm, the 6-month PFS rate was 79.9%, with 73.1% for the 7.5 mg/kg dose group versus 90.2% for the 6 mg/kg dose group. - The OS data was not mature. Doublet arm (TQB2102 7.5 mg/kg + TQB2450): Events 11; Censor (%) 14 (56.0); Median (95% CI) 12.68 (5.78, NR); mPFS: 12.68 mo; mPFS follow-up: 13.6 mo; No. at risk 25, 18, 14, 11, 8, 0 Triplet arm (TQB2102 7.5 mg/kg & 6.0 mg/kg + TQB2450 + Cape): Events 25; Censor (%) 36 (59.0); Median (95% CI) 11.04 (9.07, NR); mPFS: 11.04 mo; mPFS follow-up: 8.51 mo; No. at risk 61, 51, 37, 22, 11, 0 [X-axis Time since randomization (months) 0-15] Presenter: Miao-Zhen Qiu Miaozhen Qiu TQB2102, a novel HER2 bispecific antibody-drug conjugate, combined with bemotrizumab with or without capecitabine as first-line therapy for HER2-positive advanced gastric/gastroesophageal junction cancer: A phase II study ALMERIA AUDITORIUM - HALL 9 Survival table Endpoint: PFS Arm | N | Median | Median CI Doublet: TQB2102 7.5 mg/kg + TQB2450 | 25 | 12.68 mo | 5.78-NR Triplet: TQB2102 7.5 & 6.0 mg/kg + TQB2450 + Cape | 61 | 11.04 mo | 9.07-NR Landmark rates: 6 mo: Triplet arm (overall) 79.9% | Triplet 7.5 mg/kg 73.1% | Triplet 6 mg/kg 90.2%

[Slide 1] Considerations for 1L Therapy for HER2-Positive MBC PATINA THP → ET + HP + Palbociclib ▪ ER+, HER2+ ▪ De novo ▪ Non-visceral ▪ HER2 3+ ▪ PI3KCA wildtype HER2CLIMB-05 THP → ET + HP + Tucatinib ▪ ER+, HER2+ ▪ De novo ▪ Non-visceral ▪ HER2 3+ ▪ PI3KCA wildtype DESTINY-Breast09 T-DXd + Pertuzumab ▪ Recurrent disease within 12 mo adjuvant therapy ▪ HER2 3+ or 2+ ▪ PIK3CA mutant ▪ Significant BMs ▪ Significant visceral disease Can we merge these concepts and substitute taxane induction with T-DXd + P and follow with maintenance strategies? Expert perspectives Sara Tolaney, MD, PhD [Slide 2] Conclusion ▪ Treatment for HER2+ BC continues to evolve, with T-DXd now having a role in 1L MBC and in the preoperative and residual disease settings ▪ Figuring out how to tailor therapy in both the early and metastatic settings will be critical [puzzle-piece graphic]

LMD is an Inflammatory Disease that Activates Innate, But Not Adaptive, Immunity - LMD IFN-γ produced by T & NK cells, supports maturation of cDC2 into migratory DCs (express CCR7 in mice; LAMP3 in humans) - Non-antigen dependent production of ILs that support survival & proliferation of NK cells that are tumor cell cytotoxic. Dr. A Boire MSKCC Nature 2025 - Hypothesis: Intrathecal (IT) DCs in Breast Cancer LMD remodels the CSF and elicits adaptive immune responses Dr. Peter Forsyth

Low pCR rate replacing entire/part of conventional NACT with ADCs in HR+ EBC What's the optimal strategy/combination ? SOLTI-VALENTINE: HER3-DXd [VALENTINE / SOLTI logo] Parallel, randomized, non-comparative, open-label, phase II trial (NCT05569811) N=120 Key eligibility criteria: [small text, partly illegible] Pre- and post-menopausal women, or men; Primary operable breast cancer >=1 cm by MRI; HR+/HER2-negative; Ki67 >=20% and/or high genomic risk (gene signature); No prior treatment for the current breast cancer; Available pre-treatment FFPE core-needle biopsy Baseline | Cycle 2 Day 1 | Surgery R 2:2:1 HER3-DXd 5.6 mg/kg q21d x 6 cycles HER3-DXd + Letrozole 5.6 mg/kg q21d x 6 cycles + letrozole 2.5mg QD +/- OFS Standard Multi-agent Chemotherapy: EC or AC q14d or q21d x 4 cycles followed by weekly paclitaxel 80mg/m2 x 12 weeks pCR | Follow up 5 years Primary Endpoint: Rate of pCR (ypT0/is ypN0) at surgery Secondary Endpoints: ORR; CelTIL score change, Ki67 drop, PAM50, ROR; Safety; [illegible] [Results table] | | HER3-DXd N=50 | HER3-DXd + LET N=48 | Chemotherapy N=24 | Overall N=122 | pCR rate N | 2 | 1 | 1 | 4 % (95%CI) | 4.0% (0.5-13.7) | 2.1% (0.1-11.1) | 4.2% (0.1-21.1) | 3.3% (0.9-8.2) ORR N | 35 | 39 | 17 | 91 % (95%CI) | 70.0% (55.4-82.1) | 81.3% (67.4-91.1) | 70.8% (48.9-87.4) | 74.6% (65.9-82.0) PD N (%) | 0 | 1 (2.1%) | 1 (4.2%) | 2 (1.6%) [waterfall plot of best response by arm; axis text illegible] I-SPY 2.2: DATO-DXd +/-Durvalumab; T-DXd+Rilvegostomig [Schema: Screen -> Randomize -> BLOCK A (RILVEGOSTOMIG + T-DXd) -> BLOCK B (BEST BY RPS) -> BLOCK C (RESCUE CHEMO) -> Surgery; EARLY ESCALATION / DE-ESCALATION arrows; CONTROL ARM -> BLOCK B -> BLOCK C -> Surgery; Block B and C Treatment Assignments table (small, illegible)] Dato-DXd Signature | N | N pCR after Block A | Est. pCR Rate (s.d.) HR+HER2- | 53 | 3 | 0.08 (0.04) HR-HER2- | 50 | 15 | 0.25 (0.06) HR+HER2-Immune-DRD- | 36 | 1 | 0.03 (0.02) HR-HER2-Immune-DRD- | 11 | 3 | 0.15 (0.06) HER2-Immune+ | 46 | 12 | 0.27 (0.08) HER2-Immune-DRD+ | 10 | 2 | 0.27 (0.11) Dato-DXd +Duvalumab Signature | N | N pCR after Block A | Est. pCR Rate (s.d.) HR+HER2- | 42 | 4 | 0.17 (0.06) HR-HER2- | 64 | 21 | 0.43 (0.06) HR+HER2-Immune-DRD- | 25 | 0 | 0.03 (0.03) HR-HER2-Immune-DRD- | 23 | 2 | 0.12 (0.05) HER2-Immune+ | 47 | 20 | 0.65 (0.09) HER2-Immune-DRD+ | 11 | 3 | 0.22 (0.09) T-DXd + Rilvegostomig Subtype | Estimated pCR Rate (95% CI): R+T-DXd | Control | P(>CC) HR-HER2- | 59% (45%-73%) | 61% (44%-76%) | 0.45 HR+HER2-Immune- | 9% (3%-18%) | 15% (6%-28%) | 0.17 HR-HER2-Immune- | 24% (7%-50%) | 33% (17%-69%) | 0.26 HR+HER2-Immune+ | 66% (40%-87%) | 65% (35%-89%) | 0.51 HR-HER2-Immune+ | 72% (56%-85%) | 70% (49%-86%) | 0.55 Oliveira et al, SABCS 2025; Khoury K et al, Nature Medicine 2024; Shatsky R et al, Nature Medicine 2024; O'Sullivan et al, ASCO 2026

[Slide 2] Efficacy- ORR and DCR - The doublet arm showed an ORR of 69.6% and a DCR of 95.7%. - For the triplet cohorts, the overall ORR was 84.2% with a DCR of 100%, wherein the 6.0 mg triplet subgroup achieved an ORR of 95.8% and a DCR of 100%. [Waterfall plot: Best Reduction From Baseline in Target Lesion (%), -110 to 10; legend: TQB2102 7.5mg+TQB2450; TQB2102 7.5mg+TQB2450+Cape; TQB2102 6mg+TQB2450+Cape; x-axis Patients] Treatment response | Doublet arm: TQB2102 7.5mg/kg +TQB2450 (N=25) | Triplet arm: TQB2102 7.5 mg/kg +TQB2450+Cape (N=37) | TQB2102 6.0 mg/kg +TQB2450+Cape (N=24) | Total (N=61) BOR CR, n (%): 0 | 1 (3.0) | 2 (8.3) | 3 (5.3) PR, n (%): 16 (69.6) | 24 (72.7) | 21 (87.5) | 45 (78.9) SD, n (%): 6 (26.1) | 8 (24.2) | 1 (4.2) | 9 (15.8) PD, n (%): 1 (4.3) | 0 | 0 | 0 ORR, n (%) 95% CI: 16 (69.6) (47.1%, 86.8%) | 25 (75.8) (57.7%, 88.9%) | 23 (95.8) (78.9%, 99.9%) | 48 (84.2) (72.1%, 92.5%) cORR, n (%) 95% CI: 14 (60.9) (38.5%, 80.3%) | 23 (69.7) (51.3%, 84.4%) | 22 (91.7) (73.0%, 99.0%) | 45 (78.9) (66.1%, 88.6%) DCR, n (%) 95% CI: 22 (95.7) (78.1%, 99.9%) | 33 (100) (89.4%, 100.0%) | 24 (100) (85.8%, 100.0%) | 57 (100) (93.7%, 100.0%) mTTR, months (min, max): 2.60 (1.18, 2.96) | 1.38 (1.25, 2.92) | 1.45 (1.22, 4.14) | 1.41 (1.22, 4.14) Abbreviations: BOR, best of response; CR, complete response; PR, partial response; SD, stable disease; PD, progressive disease; NE, not evaluable; ORR, objective response rate; cORR, confirmed objective response rate; DCR, disease control rate; CI, confidence interval. Presenter: Miao-Zhen Qiu [Slide 3] Efficacy- PFS - As of 03 Sep 2026, the median PFS was 12.68 months(95% CI, 5.78-NR), for the doublet arm and 11.04 months (95% CI, 9.07-NR) for the triplet arm; median PFS for the triplet arm was immature. - In the triplet arm, the 6-month PFS rate was 79.9%, with 73.1% for the 7.5 mg/kg dose group versus 90.2% for the 6 mg/kg dose group. - The OS data was not mature. Doublet arm (TQB2102 7.5 mg/kg + TQB2450) TQB2102 7.5mg+TQB2450: Events 11; Censor (%) 14 (56.0); Median (95% CI) 12.68 (5.78, NR) mPFS: 12.68 mo mPFS follow-up: 13.6 mo No. at risk (0,3,6,9,12,15 mo): 25, 18, 14, 11, 8, 0 Triplet arm (TQB2102 7.5 mg/kg & 6.0 mg/kg + TQB2450 + Cape) TQB2102+TQB2450+Cape: Events 25; Censor (%) 36 (59.0); Median (95% CI) 11.04 (9.07, NR) mPFS: 11.04 mo mPFS follow-up: 8.51 mo No. at risk (0,3,6,9,12,15 mo): 61, 51, 37, 22, 11, 0 [Y-axis Progression-free Survival (%); X-axis Time since randomization (months)] Presenter: Miao-Zhen Qiu Survival table Endpoint: PFS Arm | N | Median | Median CI Doublet: TQB2102 7.5 mg/kg + TQB2450 | 25 | 12.68 mo | 5.78-NR Triplet: TQB2102 7.5 & 6.0 mg/kg + TQB2450 + Cape | 61 | 11.04 mo | 9.07-NR Landmark rates: 6 mo: Triplet arm (overall) 79.9% | Triplet 7.5 mg/kg 73.1% | Triplet 6 mg/kg 90.2% [Slide 4] Conclusions - TQB2102 plus benmelstobart (TQB2450) and capecitabine demonstrates promising antitumor activity and a manageable safety profile as first-line treatment for patients with HER2-positive advanced GC/GEJC. - Compared with data from the standard-of-care regimen, our study demonstrates prominent efficacy benefits, particularly in the 6mg cohort. - The 6.0 mg/kg dose of TQB2102 provides an optimal benefit-risk balance and has been selected as the recommended phase III dose. - A phase III trial is planned. Abbreviations: GC, gastric cancer; GEJC, gastro-esophageal junction cancer; HER2, human epidermal growth factor receptor 2. Presenter: Miao-Zhen Qiu

[Slide 2] Conclusions The study met its predetermined endpoint for OS with the longest median OS reported to date in HER2+ LM. Complete responses were noted in a subset of patients. IT trastuzumab/pertuzumab should be considered in the management of HER2+ LM. Kamran Ahmed [Slide 3] Response Rate Response ▲ LM Progression ■ CR ● PR □ SD Ongoing follow-up → Best LM Response: CR 5 (17%) PR 5 (17%) and SD in 17 (57%) Individual Patients Time from Treatment Initiation (Months) Kamran Ahmed

Results - Lymphoid organisation and spatial immune phenotype Translational analysis set (n=76) Spatial Immunophenotype: baseline to EOT Paired-evaluable patients: n= 64 Baseline: Inflamed n=29 (45%); Excluded n=21 (33%); Desert n=14 (22%) EOT: Inflamed n=40 (63%); Excluded n=13 (20%); Desert n=11 (17%) Legend: More inflamed | Unchanged | Less inflamed Tertiary Lymphoid strucutres (TLS) detection: baseline to EOT Paired-evaluable patients: n= 64 Baseline: TLS present n=12 (19%); TLS absent n=52 (81%) EOT: TLS present n=27 (42%); TLS absent n=37 (58%) Legend: Gained | Unchanged | Lost Shift towards an inflamed phenotype and more frequent TLS detection Joana M Ribeiro

[Slide 1] Background - Gastric and gastro-esophageal junction cancer (GC/GEJC) represents a substantial global disease burden, with a notable subset of HER2-positive advanced cases. - The KEYNOTE-811 trial established anti-HER2 and immunotherapy as first-line standard-of-care for patients with HER2-positive advanced GC/GEJC1, supporting evaluation of novel bispecific HER2-ADC-based combinations. - TQB2102 is a novel bispecific HER2 ADC targeting domains ECD 2/4, conjugated to a deuterated Topo I inhibitor (DAR=6). - This is the first phase II study of a bispecific HER2 ADC combination as first-line therapy for HER2-positive advanced GC/GEJC. [Molecule diagram: Anti-Her2 ECD2; Anti-Her2 ECD4; VH/VL; Cleavable linker; Topo1 inhibitor; DAR 5.8 ~ 6.0] References: 1. Janjigian YY, et al. Lancet 2023. Abbreviations: GC, gastric cancer; GEJC, gastro-esophageal junction cancer; HER2, human epidermal growth factor receptor 2; ADC, antibody-drug conjugate; DAR, drug-to-antibody ratio. Presenter: Miao-Zhen Qiu [Slide 4] Efficacy- PFS - As of 03 Sep 2026, the median PFS was 12.68 months(95% CI, 5.78-NR), for the doublet arm and 11.04 months (95% CI, 9.07-NR) for the triplet arm; median PFS for the triplet arm was immature. - In the triplet arm, the 6-month PFS rate was 79.9%, with 73.1% for the 7.5 mg/kg dose group versus 90.2% for the 6 mg/kg dose group. - The OS data was not mature. Doublet arm (TQB2102 7.5 mg/kg + TQB2450): mPFS: 12.68 mo; mPFS follow-up: 13.6 mo; No. at risk 25, 18, 14, 11, 8, 0 Triplet arm (TQB2102 7.5 mg/kg & 6.0 mg/kg + TQB2450 + Cape): mPFS: 11.04 mo; mPFS follow-up: 8.51 mo; No. at risk 61, 51, 37, 22, 11, 0 [In-plot event/censor legend too small in this photo] Presenter: Miao-Zhen Qiu Survival table Endpoint: PFS Arm | N | Median | Median CI Doublet: TQB2102 7.5 mg/kg + TQB2450 | 25 | 12.68 mo | 5.78-NR Triplet: TQB2102 7.5 & 6.0 mg/kg + TQB2450 + Cape | 61 | 11.04 mo | 9.07-NR Landmark rates: 6 mo: Triplet arm (overall) 79.9% | Triplet 7.5 mg/kg 73.1% | Triplet 6 mg/kg 90.2%

neoHIP: INTEGRATION OF ER AND IMMUNE BIOMARKERS HR status | Immune status | Arm A: THP | Arm B: THP + pembro | Δ pCR (B − A) HR− | PD-L1+ or sTILs+ | 64% (7/11) | 86% (6/7) | +22% HR− | PD-L1− and sTILs− | 71% (5/7) | 88% (7/8) | +17% HR+ | PD-L1+ or sTILs+ | 38% (5/13) | 80% (4/5) | +42% HR+ | PD-L1− and sTILs− | 36% (5/14) | 25% (3/12) | −11% Page D, ESMO 2026 3-biomarker rule for "pembro eligibility" Pembro eligible: +30% pCR with pembro Pembro not eligible: -11% pCR with pembro • Exploratory analysis • Very small subgroups • Data-driven biomarker integration • Unbalanced biomarker availability between arms: 55% of pembro arm / 78% of control arm • No test for interaction shown • However, nicely hypothesis-generating ER−/HER2+ Typically HER2-enriched, intrinsically immune-enriched, more sensitive to ICI → conventional immune biomarkers (PD-L1, TILs) not further discriminating ER+/HER2+ Heterogeneous, ≥60% luminalA/B, mostly immune-cold → conventional immune biomarkers (PD-L1, TILs) offer a first discriminating signal HER2DX: integrated biomarker Clinical (T, N) + 4 gene modules: immune (IGG) · luminal · proliferation · HER2 amplicon How would HER2DX perform in immunotherapy trials? Prat A, J Natl Cancer Inst 2014; Dieci MV, Ann Oncol 2019; Griguolo G, NPJ Precis Oncol 2021; Prat A, EBioMedicine 2022 Maria Vittoria Dieci

[Slide 1] Study Design This is a multicenter, open-label, phase I/II trial conducted in China (ChiCTR2400094683) Key eligibility criteria • ECOG PS of 0-1 • Aged 18-75 years • Histologically confirmed advanced and/or metastatic solid tumors • At least a measurable lesion as per RECIST 1.1 • Relapsed or progressed after standard treatment Phase I dose escalation BION design 1.2 mg/kg, Q3W -> 2.4 mg/kg, Q3W -> 4.0 mg/kg, Q3W -> 6.0 mg/kg, Q3W -> 8.0 mg/kg, Q3W -> 10.0 mg/kg, Q3W 4.0 mg/kg, Q2W -> 6.0 mg/kg, Q2W Phase II dose expansion 4.0 mg/kg Q3W 6.0 mg/kg Q3W 8.0 mg/kg Q3W 6.0 mg/kg Q2W Lung Cancer | Breast Cancer | GC | NPC | HNSCC | ESCC | Liver cancer | Others Primary endpoint: safety, tolerability, and recommended phase II dose (RP2D) (phase I) and objective response rate (phase II) Hongxia Wang, MD [Slide 2] Conclusion • SYS6043 had a favorable safety profile. ■ The most common TRAEs were hematological toxicities and gastrointestinal toxicities. • SYS6043 showed promising durable antitumor activity in patients with previously treated breast cancers ■ In HR+/HER2- breast cancer (n = 82), ORR 53.7%, DCR 92.7%, mPFS and mDoR not reached ■ In TNBC (n = 32) , ORR 75.0%, DCR 93.8%, mPFS 7.2 months, mDoR 5.5 months ■ In HER2+ breast cancer (n = 23), ORR 69.6%, DCR 95.7%, mPFS not reached, mDoR 7.0 months • Data support further evaluation of SYS6043 at 6 mg/kg Q3W in patients with pretreated breast cancers. Hongxia Wang, MD

[Slide 1] Testing This Approach in Clinical Trial: NCT 05809752, IND 27748: Apheresis Scheme Donor apheresis → Monocytes / T Cells → + GMCSF + IL-4 → immature dendritic cell (iDC) + class II peptides → + IFNγ + LPS (TLR4) → Type 1-polarized MATURE dendritic cell (DC1) → DC1 co-culture with naive CD4 T Cells → + IL-2 → iDC co-culture with primed CD4 T Cells → Restimulation with iDC specific class II peptide vs. irrelevant control peptide → IFNγ (pg/ml) peptide vs control (>1.4) Dr. Peter Forsyth [Slide 2] Testing this Approach: Phase 1 Clinical Trial (NCT 05809752) a LMD Diagnosis → LMD Standard of Care + Ommaya Placement → Screening & Staging: MRI, Blood Draw, CSF Analysis; Apheresis & cDC1 Production → Enroll → Cycle 1 Weekly Intrathecal Dendritic Cell Therapy (1e6, 2e6, 5e6, or 10e6 cDC1) HER2/3 cDC1 Week# 1–6; Serum and CSF Sampling (pre and post cDC1 when possible) → Re-Staging: MRI, Blood Draw, CSF Analysis → Continue to Cycle 2 of cDC1 if permissable e CSF pleocytosis (number of nucleated cells; Patients #01, #02, #03, #05, #06, #08) f CSF Cytokines (IFN-γ, IL-1β, IL-12, IL-8, TNF-α, IL-6) by patient and treatment dose g CSF Chemokines/cytokine (CCL3, CCL4, CCL8, CXCL8, CXCL9) - N = 15 women with LMD (9 HER2+ & 6 TNBC) - 100% had headaches; DLT in two @ 5 x 10^7 IT cDC1s - RP2D = 1 x 10^7 IT cDC1s - mPFS = 2.83 mos (95% CI 1.28-NA) - mOS = 16.24 mos (95% CI 11.21-N/A) Brian Czerniecki, MD PhD, Chair, Dept of Breast Oncology, Moffitt Cancer Center Pawel Kalinski MD, PhD, Vice Chair, Translational Research in Development in Women's Health, University of Pittsburgh Law et al. (manuscript under revision) Dr. Peter Forsyth [Slide 3] Patient Characteristics ( Data lock May 7, 2026) Patient Number | Dose | Age | Breast Cancer Subtype | CSF cyto. | F/U from C1D1 (mos.) | Status (A=Alive, D=Dead) Disease Status | Survival Months #01 | 1 x 10^[illegible] | 48/F | "TNBC", HER2 scRNA | + | 14 | !D, PD (RT induced leuko.) | 14.70 #02 | 1 x 10^[illegible] | 33/F | HER2+, ER+, PR- | - | 30 | A, PD now SD (on H&P) | 26.93 #03 | 1 x 10^[illegible] | 41/F | HER2+, HR- | - | 23 | D, PD | 23.24 #04 | 2 x 10^[illegible] | 46/F | TNBC | + | 5 | D, PD | 5.98 #05 | 2 x 10^[illegible] | 48/F | TNBC | + | 23 | A, SD | 20.84 #06 | 2 x 10^[illegible] | 59/F | HER2+, HR- | - | 16 | D, PD | 16.24 #07 | 1 x 10^7 | 59/F | HER2+, HR- | - | 22 | A, SD | 19.13 #08 | 1 x 10^7 | 48/F | HER2+, ER+/PR<1% | - | 21 | A, SD (thalamic stroke) | 18.41 #09 | 1 x 10^7 | 62/F | TNBC | + | 2 | !D, Systemic PD | 2.30 #10 | 5 x 10^7 | 58/F | HER2+, ER+ | - | 10 | !D, (PD + RT induced leuko.) | 11.21 #11 | 5 x 10^7 | 41/F | HER2+, ER low+ (3%), PR- | - | 3 | D, PD | 3.78 #12 | 5 x 10^7 | 65/F | TNBC | - | 5 | A, on treatment | NA #13 | 1 x 10^7 | 43/F | ER-/PR+ HER2+ | - | 4 | A, PD now SD (on H&P) | 2.83 #14 | 1 x 10^7 | 34/F | TNBC | + | 3 | A, PD | 2.30 #15 | 1 x 10^7 | 35/F | TNBC | + | 1 | D, PD + hydro | 0.99 cyto. = cytology; DEC = Dose Expansion Cohort; hydro= hydrocephalus; ! = died non-LMD cause; Cut-off date:05/07/2026 Dr. Peter Forsyth [Slide 4] Overall Survival ( OS), progression Free Survival (PFS) PFS (Strata: All) | OS (Strata: All) - mPFS = 2.83 mos. (95% CI 1.28-NA) - mOS = 16.24 mos . (95% CI 11.21-N/A) - RP2D = 1 x 10^7 IT cDC1s Cut-off date: 05/07/2026 Dr. Peter Forsyth Survival table Endpoint: PFS and OS (single arm, N=15) Arm | N | Median | Median CI All (PFS) | 15 | 2.83 mo | 1.28–NA All (OS) | 15 | 16.24 mo | 11.21–N/A

Mind-reading is not a validated tool What does it take to accept intensive chemotherapy? It depends who you ask Minimal chance of cure required (median) to accept intensive chemotherapy General public 50% Cancer nurses 50% GPs 25% Cancer doctors 10% Patients with cancer 1% Doctors believed 71% of women with breast cancer prioritised keeping their breast Patients actually: 7% Slevin ML et al. BMJ 1990;300:1458-60; Mulley AG et al. BMJ 2012;345:e6572 Claire Gallois

[Slide 1] 10:15 - 11:45 Rapid oral 1: Breast cancer, advanced stage CHAIRS: VESNA BJELIC-RADISIC, GÉRALDINE GEBHART, WILLIAM GRADISHAR Background • HER2CLIMB-05 investigated the efficacy and safety of adding tucatinib (Tuc) to trastuzumab (T) and pertuzumab (P) as first-line (1L) maintenance therapy in patients with human epidermal growth factor receptor 2–positive (HER2+) metastatic breast cancer (mBC) HER2CLIMB-05 Study Design Key Eligibility Criteria • Centrally confirmed HER2+ MBC • No evidence of progression after THP (4 to 8 cycles) • ECOG PS of 0 or 1 • No or asymptomatic BM confirmed by contrast-enhanced MRI at screening R 1:1 Randomization was stratified by: • Diagnosis: de novo or recurrent • HR status: positive or negative • Presence or history of BM: yes or no 1L Maintenance Therapy TUC 300 mg PO BID + HP* Once every 21 days ± ET (n = 326) PBO PO BID + HP* Once every 21 days ± ET (n = 328) Study treatment continues until unacceptable toxicity, disease progression, consent withdrawal, or study closure. No crossover from PBO to TUC was allowed. Endpoints Primary • Investigator-assessed PFS per RECIST v1.1 Secondary • OS (key secondary) • PFS per BICR • CNS-PFS • Safety • HRQoL (circled) • Pharmacokinetics *H (6 mg/kg IV or 600 mg SC) and P (420 mg IV) or fixed-dose combination (SC) of H (600 mg), P (600 mg), and hyaluronidase (20,000 units). 1L = first-line; BICR = blinded independent central review; BM = brain metastases; CNS-PFS = central nervous system progression-free survival; ECOG PS = Eastern Cooperative Oncology Group performance status; ET = endocrine therapy; H = trastuzumab; HER2+ = human epidermal growth factor receptor 2-positive; HRQoL = health-related quality of life; IV = intravenous; MRI = magnetic resonance imaging; OS = overall survival; P = pertuzumab; PBO = placebo; PFS = progression-free survival; PO BID = orally twice a day; R = randomization; RECIST = Response Evaluation Criteria in Solid Tumors; SC = subcutaneous; T = taxane; TUC = tucatinib. Véronique Diéras [Speaker panel] Veronique Dieras — Health-related quality of life from HER2CLIMB-05: A phase 3 study of tucatinib versus placebo in combination with trastuzumab and pertuzumab as 1L maintenance therapy for HER2+ mBC [Slide 2] 10:15 - 11:45 Rapid oral 1: Breast cancer, advanced stage CHAIRS: VESNA BJELIC-RADISIC, GÉRALDINE GEBHART, WILLIAM GRADISHAR Results: TTF-PRO - GHS/QoL [Prespecified secondary endpoint] Time to first PRO deterioration excluding disease progression and death (TTF-PRO) | TUC (n = 307) | PBO (n = 316) Events (%) | 169 (55.0%) | 142 (44.9%) Median, months (95% CI) | 11.8 (6.8, 17.4) | 18.7 (14.5, -) HR (95% CI) | 1.29 (1.03, 1.62) Number at risk (Time, months 0-42) 1 | 316 237 179 145 121 98 68 49 34 21 8 4 1 1 0 2 | 307 201 161 141 116 97 70 51 29 14 3 0 • TUC showed a 29% increased risk of deterioration compared with control • Limitations: Considering only PRO deterioration, analysis does not account for influence of disease progression and death; PRO assessment schedule was dependent on disease progression [Speaker panel] Veronique Dieras — Health-related quality of life from HER2CLIMB-05: A phase 3 study of tucatinib versus placebo in combination with trastuzumab and pertuzumab as 1L maintenance therapy for HER2+ mBC Survival table Endpoint: TTF-PRO GHS/QoL (time to first PRO deterioration excluding disease progression and death; prespecified secondary) Arm | N | Median | Median CI TUC | 307 | 11.8 mo | 6.8, 17.4 PBO | 316 | 18.7 mo | 14.5, - HR: 1.29 HR CI: 1.03, 1.62 [Slide 3] 10:15 - 11:45 Rapid oral 1: Breast cancer, advanced stage CHAIRS: VESNA BJELIC-RADISIC, GÉRALDINE GEBHART, WILLIAM GRADISHAR Results: TTF-PRO-DP-D (left) and TTC-PRO-DP-D (right) Global Health Status / Quality of Life TTF-PRO-DP-D = Time to first PRO deterioration, disease progression, or death [post hoc exploratory] | TUC (n = 307) | PBO (n = 316) Events (%) | 231 (75.2%) | 249 (78.8%) Median, months (95% CI) | 6.2 (4.9, 7.6) | 6.1 (4.9, 7.0) HR (95% CI) | 1.00 (0.83, 1.19) Number at risk 1 | 316 218 164 125 105 77 61 36 29 19 7 4 1 1 0 2 | 307 195 156 123 107 86 63 41 24 7 1 0 TTC-PRO-DP-D = Time to confirmed PRO deterioration (≥2 consecutive assessments), disease progression, or death [post hoc exploratory] | TUC (n = 307) | PBO (n = 316) Events (%) | 181 (59.0%) | 217 (68.7%) Median, months (95% CI) | 14.6 (12.4, 18.5) | 10.3 (8.3, 12.4) HR (95% CI) | 0.81 (0.67, 0.99) Number at risk 1 | 316 251 207 163 140 106 80 47 37 20 9 4 1 1 0 2 | 307 236 207 172 163 130 102 67 43 14 3 0 Véronique Diéras [Speaker panel] Veronique Dieras — Health-related quality of life from HER2CLIMB-05: A phase 3 study of tucatinib versus placebo in combination with trastuzumab and pertuzumab as 1L maintenance therapy for HER2+ mBC Survival table Endpoint: TTF-PRO-DP-D GHS/QoL (post hoc exploratory); TTC-PRO-DP-D also shown Arm | N | Median | Median CI TUC | 307 | 6.2 mo | 4.9, 7.6 PBO | 316 | 6.1 mo | 4.9, 7.0 HR: 1.00 HR CI: 0.83, 1.19 [Slide 4] 10:15 - 11:45 Rapid oral 1: Breast cancer, advanced stage CHAIRS: VESNA BJELIC-RADISIC, GÉRALDINE GEBHART, WILLIAM GRADISHAR Conclusions • For the prespecified secondary endpoint GHS/QoL TTF-PRO definition, Tuc had a shorter time to deterioration with an increased hazard relative to control • Alternate definitions for events sought to address TTF-PRO limitations by incorporating disease progression/death and investigating "confirmed" PRO deterioration • Both modifications showed stronger treatment effects than PRO-only "first" deterioration events • Results support integration of PROs with objective disease outcomes to provide more holistic view of health-related QoL across Tuc and placebo groups Key Takeaways: • Taken together, these findings suggest that the substantial PFS benefit of Tuc is accompanied by variable treatment effects on PROs, with modest effects for confirmed composite deterioration event definitions, consistent with a favorable benefit-risk profile for tucatinib as 1L maintenance therapy in HER2+ mBC. • Treating disease progression and death as events, and requiring confirmation of deterioration, addresses limitations of PRO-only first deterioration endpoints and provides complementary and important information regarding the patient experience. Véronique Diéras [Speaker panel] Veronique Dieras — Health-related quality of life from HER2CLIMB-05: A phase 3 study of tucatinib versus placebo in combination with trastuzumab and pertuzumab as 1L maintenance therapy for HER2+ mBC

[Slide 3] 10:15 - 11:45 Rapid oral 2: Breast cancer, advanced stage CHAIRS: ANTONIO LLOMBART CUSSAC, SHERENE LOI, PEDRAM RAZAVI Safety TRAEs | N=62 n (%) TRAEs | 62 (100) Grade ≥3 TRAEs | 25(40.3) Serious TRAEs | 8(12.9) TRAEs leading to treatment interruption | 28(45.2) TRAEs leading to dose reduction | 5(8.1) TRAEs leading to discontinuation | 0 TRAEs leading to death | 0 TRAEs, % AST increased Neutrophil count decreased WBC decreased ALT increased Hypertriglyceridemia Hypercholesterolemia Anemia Hyperglycemia Asthenia PLT decreased GGT increased Grade ≥3 · Grade 1-2 TRAE, treatment-related adverse event; WBC, white blood cell; AST, aspartate aminotransferase; ALT, alanine aminotransferase; GGT, gamma-glutamyltransferase; PLT, platelet. Dr Yanchun Meng Yanchun Meng Bulumtatug fuvedotin (BFv, 9MW2821), a nectin-4-directed antibody-drug conjugate (ADC), in patients with locally advanced or metastatic triple negative breast cancer (la/mTNBC) previously treated with topoisomerase I inhibitor (TopI)-based ADCs: Resu... ALICANTE AUDITORIUM - HALL 6 [Slide 4] 10:15 - 11:45 Rapid oral 2: Breast cancer, advanced stage CHAIRS: ANTONIO LLOMBART CUSSAC, SHERENE LOI, PEDRAM RAZAVI Conclusion • BFv demonstrated promising antitumor activity with a manageable safety profile in heavily pretreated patients with la/mTNBC who had all received prior Top1i-based ADCs. • The first pivotal trial for this indication worldwide is underway. Dr Yanchun Meng Yanchun Meng Bulumtatug fuvedotin (BFv, 9MW2821), a nectin-4-directed antibody-drug conjugate (ADC), in patients with locally advanced or metastatic triple negative breast cancer (la/mTNBC) previously treated with topoisomerase I inhibitor (TopI)-based ADCs: Resu... ALICANTE AUDITORIUM - HALL 6

[Slide 1] 15:15 - 15:45 Keynote lecture 1 CHAIR: ANNA SOPHIE BERGHOFF Solid Metastasis - update • CNS-active systemic therapy can justify deferring local treatment in selected asymptomatic patients. • Stereotactic radiotherapy (SRS) is supported in selected patients with 5–20 metastases; lesion count alone is insufficient. • Fractionated cavity SRS and preoperative SRS are reshaping postoperative strategy. • ADCs have become clinically relevant intracranial therapeutics, especially in HER2-positive disease. • Radionecrosis, neurocognition and patient-reported outcomes matter more as survival improves. ALMERIA AUDITORIUM - HALL 9 [Slide 2] 15:15 - 15:45 Keynote lecture 1 CHAIR: ANNA SOPHIE BERGHOFF Solid Metastasis - SRS versus hippocampal-avoidance whole brain radiotherapy (WBRT) 8.3 mo Median OS — SRS 8.5 mo Median OS — HA-WBRT 0.8% Lesion-level local recurrence — SRS 12.3% Lesion-level local recurrence — HA-WBRT • SRS improved patient-reported symptoms and functional interference. • More distant intracranial failures and radiographic radionecrosis after SRS. • Selection process? PMID: 41712219 ALMERIA AUDITORIUM - HALL 9 [Speaker video inset] Wolfgang Wick CNS disease as the next frontier in oncology care: Glioblastoma, metastases, and leptomeningeal disease [Slide 3] 15:15 - 15:45 Keynote lecture 1 CHAIR: ANNA SOPHIE BERGHOFF Solid Metastasis – Her2 positive breast cancer: pooled data from DESTINY trials Trastuzumab deruxtecan: prospective CNS activity 263 Patients with baseline brain metastases 58.9% 12-month CNS-PFS 71.7% CNS ORR in measurable disease • Activity observed in both stable and active brain metastases. • T-DXd and tucatinib–trastuzumab–capecitabine provide distinct sequencing options. • Monitor interstitial lung disease/pneumonitis and treatment history. PMID: 39241960 ALMERIA AUDITORIUM - HALL 9 [Speaker video inset] Wolfgang Wick CNS disease as the next frontier in oncology care: Glioblastoma, metastases, and leptomeningeal disease [Slide 4] 15:15 - 15:45 Keynote lecture 1 CHAIR: ANNA SOPHIE BERGHOFF Solid Metastasis – algorithm • Large symptomatic lesion / mass effect → surgery or urgent focal treatment. • Limited-volume disease → SRS; consider fractionation by lesion size/locat[obscured by overlay] • Numerous [obscured] [illegible] ted patients with reliab[obscured] • Actionable driver, asymptomatic lesions → consider systemic-first strategy. • Extensive disease unsuitable for SRS → consider HA-WBRT + memantine or supportive care. [Overlay box:] Primary scope: intervention should optimally preserve neurological function with the least long-term toxicity ALMERIA AUDITORIUM - HALL 9 [Speaker video inset] Wolfgang Wick CNS disease as the next frontier in oncology care: Glioblastoma, metastases, and leptomeningeal disease

[Slide 1] 15:15 - 15:45 Keynote lecture 1 CHAIR: ANNA SOPHIE BERGHOFF Glioblastoma • TRIDENT phase III: earlier TTFields start did not improve OS • Intracranial B7-H3 CAR-T: encouraging phase I biological and clinical activity • Personalized vaccines: feasibility and immunogenicity; survival efficacy unproven • Advances in spatial biology and neuron - glioma cell interactions outpace clinical translation ➢ No new broadly effective treatment has replaced the current standard of care. ALMERIA AUDITORIUM - HALL 9 [Speaker video inset] Wolfgang Wick CNS disease as the next frontier in oncology care: Glioblastoma, metastases, and leptomeningeal disease [Slide 2] 15:15 - 15:45 Keynote lecture 1 CHAIR: ANNA SOPHIE BERGHOFF [Stacked bar chart, PFS-6 Yes/No by arm; y-axis 0%-100%] APG101 n = 26 p = 0.88 Atezolizumab n = 42 p = 0.66 Palbociclib n = 41 p = 0.48 Temsirolimus n = 46 p = 0.01 SOC n = 54 p = 0.83 PFS-6: Yes / No H0: p = 0.231 [Overlay box:] Activity of mTOR inhibition in newly diagnosed GB w/o MGMT promotor hypermethylation in tumors with phospho mTOR > 150 PMID: 30277538 ALMERIA AUDITORIUM - HALL 9 [Speaker video inset] Wolfgang Wick CNS disease as the next frontier in oncology care: Glioblastoma, metastases, and leptomeningeal disease [Slide 3] 15:15 - 15:45 Keynote lecture 1 CHAIR: ANNA SOPHIE BERGHOFF [Figure panels] a: Informed consent, Leukapheresis, Ommaya reservoir placement; One treatment cycle: TX103 infusion; Optional cycle: TX103 infusion; timeline D-21, D-5, D0, D14, D28, D42, D55, D77, Every 2 months; DLT monitoring; GBM FFPE B7-H3 testing; CAR-T cell manufacturing; PET-MRI; MRI* b: 26 patients assessed for eligibility — Excluded 11 patients: 6 patients ineligible for study (3 B7-H3 expression <30%; 1 pathological reassessment as grade 3 glioma; 1 unable to provide slides for B7-H3 testing; 1 tumor hemorrhagic apoplexy, risk for trial); 5 patients withdrew consent (4 incomplete screening; 1 pandemic travel restriction). Enrolled 15 patients — TX103: 3 patients DL1: 2 × 10^7; 6 patients DL2: 6 × 10^7; 6 patients DL3: 1.5 × 10^8. 15 safety and efficacy evaluation; 13 pharmacokinetics and immune response a (swimmer plot): Patient IDs 01007, 01013, 01002, 01023, 01003, 01026, 01025, 01012, 01027, 01019, 01022, 01024, 01006, 01008, 01018; x-axis Time from first infusion (months) 0-39; legend: CAR-T infusion (ICV), CAR-T infusion (ITC), Death, Surgery, Continued survival, Bevacizumab, BRAF-targeted therapy, Chemotherapy, Complete response, First infusion to response evaluation, Not evaluable for response, Partial response, Stable disease, Survival after progression, Without measurable disease b (OS): Median OS: 19.1 months (8.93, not reached) Number at risk (censored), time 0/10/20/30/40 months: 15 (0) | 10 (0) | 5 (2) | 2 (3) | 0 (5) c (PFS): Median PFS: 1.87 months (1.17, 2.77) Number at risk (censored), time 0/10/20/30/40 months: 15 (0) | 1 (1) | 1 (1) | 1 (1) | 0 (2) d: Best overall response (PD, SD, CR) waterfall — Best change in SPD from baseline (%) e: Patient 01007 serial MRI, CAR-T infusion timepoints, Time from first infusion (weeks) PMID: 42562965 ALMERIA AUDITORIUM - HALL 9 [Speaker video inset] Wolfgang Wick CNS disease as the next frontier in oncology care: Glioblastoma, metastases, and leptomeningeal disease Survival table Endpoint: OS and PFS (single-arm phase I, intracranial B7-H3 CAR-T TX103) Arm | N | Median | Median CI TX103 (all dose levels) - OS | 15 | 19.1 months | 8.93, not reached TX103 (all dose levels) - PFS | 15 | 1.87 months | 1.17, 2.77 [Slide 4] 15:15 - 15:45 Keynote lecture 1 CHAIR: ANNA SOPHIE BERGHOFF Glioblastoma - A PERSONALIZED NEOANTIGEN VACCINE TO REPROGRAM THE IMMUNE LANDSCAPE OF GLIOBLASTOMA • 39 dexamethasone-free patients with nexly diagnosed glioblastoma have been treated with up 20 (neo) peptides in poly-ICLC after surgery and radiotherapy • 4 cohorts (1-3: temozolomide free plus pembrolizumab, MGMT unmeth & 4: with temozolomide and pembrolizumab) J Clin Oncol 2026;44 (suppl 16 abstr 2006). ALMERIA AUDITORIUM - HALL 9

136P Modeling temporal therapeutic evolution in advanced breast cancer through integrated trial-level and real-world analyses F Miglietta et al. [full author list and affiliations: small print, partly illegible]; Istituto Oncologico Veneto IRCCS; University of Padova BACKGROUND Therapeutic benefit in advanced breast cancer (ABC) may evolve over time as a result of TREATMENT ECOSYSTEM MATURATION, through evolution in: control-arm prognosis, treatment sequencing, post-progression treatment access, and biologic reassessment. We investigated TEMPORAL THERAPEUTIC EVOLUTION across BC subtypes (HR+/HER2-, HER2+, TNBC) using integrated trial-level and real-world (RW) analyses. METHODS TRIAL-LEVEL: We analyzed 116 treatment arms from pivotal ABC trials using temporal regression and explainable ML modeling (SHAP-based interpretation) to evaluate experimental-control ΔmPFS and ΔmOS (Δ=delta) adjusted for treatment platform, treatment line, prior/subsequent therapies, clinicopathological features. RW-LEVEL: Findings were contextualized within a large multicentric ambispective RW cohort of 1938 ABC pts (1990-2025) analyzed using subtype-adjusted Cox models for OS. RESULTS EXPERIMENTAL-CONTROL ΔmPFS AND ΔmOS ACROSS ABC TRIALS (N=116 TREATMENT ARMS) Distinct ABC subtypes showed divergent patterns of therapeutic ecosystem evolution: HR+/HER2- progressive expansion of experimental-control separation HER2+ parallel evolution of control and experimental outcomes TNBC modern-era expansion of PFS benefit ΔmPFS: HR+/HER2- +0.28 mos/yr (0.091-0.467), p=0.004; HER2+ +0.90 mos/yr (-0.315-2.107), p=0.147; TNBC +0.30 mos/yr (-0.007-0.608), p=0.056 ΔmOS: HR+/HER2- +0.51 mos/yr (0.002-1.017), p=0.049; HER2+ -0.17 mos/yr (-2.361-2.029), p=0.882; TNBC +0.28 mos/yr (-0.277-0.827), p=0.329 *Change in experimental-control benefit per year (months/year) MAIN CONTRIBUTORS TO THE PREDICTED OS BENEFIT SHAP analysis identified the main contributors to predicted OS benefit at the trial-level. Mid-year of recruitment 2.46; CDK4/6i exposure 1.39; Post-PD crossover 0.94 (Mean |SHAP| value) REAL-WORLD OS VALIDATION (N=1938 PATIENTS) Real-world OS improved over time, but temporal trends were not fully explained by treatment evolution alone; primary-to-metastatic phenotype instability suggests a concurrent contribution of biological evolution. Temporal trend of primary-to-metastatic phenotype switch: OR/year 1.03 (95%CI 0.99-1.06). Overall switch rate 21.5% HR/year (95%CI); p-value: Overall 0.965 (0.956-0.974); p<0.001 | HR+/HER2- 0.955 (0.944-0.967); p<0.001 | HER2+ 0.967 (0.946-0.990); p=0.004 | TNBC 0.985 (0.967-1.004); p=0.129 CONCLUSIONS Therapeutic benefit in ABC is increasingly shaped by evolving control-arm prognosis, post-progression therapies, and biologic reassessment, beyond intrinsic drug activity. Our findings highlight the growing complexity of contemporary control arm selection and equitable post-progression treatment access, supporting ecosystem-aware trial design and interpretation.

CHAIRS: KOMAL JHAVERI, BARBARA PISTILLI [partially cropped] Low pCR rate replacing entire/part of conventional NACT with ADCs in HR+ EBC What's the optimal strategy/combination ? SOLTI-VALENTINE: HER3-DXd [Embedded study schema, small text] VALENTINE SOLTI - Parallel, randomized, non-comparative, open-label, phase II trial (NCT05569811); N=120; arms: HER3-DXd 5.6 mg/kg q21d x 6 cycles; HER3-DXd + Letrozole 5.6 mg/kg q21d x 6 cycles + letrozole 2.5mg QD +/- OFS; Standard Multi-agent Chemotherapy EC or AC q14d or q21d x 4 cycles followed by weekly paclitaxel 80mg/m2 x 12 weeks; Primary Endpoint: Rate of pCR (ypT0/is ypN0) at surgery; Secondary Endpoints: ORR, CelTIL score change, Ki67 drop, PAM50, ROR, Safety, iDFS [small eligibility text largely illegible] [Embedded results table] | HER3-DXd N=50 | HER3-DXd + LET N=48 | Chemotherapy N=24 | Overall N=122 pCR rate N | 2 | 1 | 1 | [illegible] % (95%CI) | 4.0% (0.5-13.7) | 2.1% (0.1-11.1) | 4.2% (0.1-21.1) | 3.3% (0.9-8.2) ORR N | 35 | 39 | 17 | 91 % (95%CI) | 70.0% (55.4-82.1) | 81.3% (67.4-91.1) | 70.8% (48.9-87.4) | 74.6% (65.9-82.0) PD N (%) | 0 | 1 (2.1%) | 1 (4.2%) | 2 (1.6%) [Waterfall plot, illegible labels] Low pCR rate with 6 cycles of HER3-DXd, same of standard NACT (4%), but fewer severe AEs Addition of letrozole did not increase pCR I-SPY 2.2: DATO-DXd +/-Durvalumab; T-DXd+Rilvegostomig [I-SPY 2.2 sequential multiple-assignment schema: Screen, Randomize, Block A, Block B, Block C, Surgery, Control arm; small text illegible] Dato-DXd Signature | N | N pCR after Block A | Est. pCR Rate (s.d.) HR+HER2- | 53 | 3 | 0.08 (0.04) HR-HER2- | 50 | 15 | 0.25 (0.06) HR+HER2-Immune-DRD- | 36 | 1 | 0.03 (0.02) HR-HER2-Immune-DRD- | 11 | 3 | 0.15 (0.06) HER2-Immune+ | 46 | 12 | 0.27 (0.08) HER2-Immune-DRD+ | 10 | 2 | 0.27 (0.11) Dato-DXd +Duvalumab Signature | N | N pCR after Block A | Est. pCR Rate (s.d.) HR+HER2- | 42 | 4 | 0.17 (0.06) HR-HER2- | 64 | 21 | 0.43 (0.06) HR+HER2-Immune-DRD- | 25 | 0 | 0.03 (0.03) HR-HER2-Immune-DRD- | 23 | 2 | 0.12 (0.05) HER2-Immune+ | 47 | 20 | 0.65 (0.09) HER2-Immune-DRD+ | 11 | 3 | 0.22 (0.09) T-DXd + Rilvegostomig Subtype | Estimated pCR Rate (95% CI) R+T-DXd | Control | P(>CC) HR-HER2- | 59% (45%-73%) | 61% (44%-76%) | 0.45 HR+HER2-Immune- | 9% (3%-18%) | 15% (6%-28%) | 0.17 HR-HER2-Immune- | 24% (7%-50%) | 33% (17%-69%) | 0.26 HR+HER2-Immune+ | 66% (40%-87%) | 65% (35%-89%) | 0.51 HR-HER2-Immune+ | 72% (56%-85%) | 70% (49%-86%) | 0.55 Low pCR rate with 4 cycles of Dato-DXd +/- Durvalumab or T-DXd + Rilvegostomig in unselected lu[obscured] Immune-enriched HR+ breast cancers derived greater benefit of ADC–IO comb[obscured] making this combination worthy of further evaluation in biomarker-selected lumina[obscured]

ESMO congress (Madrid 2026) ADC Resistance | Contemporary Challenge in 2026 BRE 354: Ph2 IIT of Patritumab Deruxtecan in MBC Parts B and Z: Prior ADC Treated Part A: HER2- MBC (N=60); Topoisomerase I exatecan-naive** -> Part A Analysis: Activity and HER3 expression Part B: HR+/HER2- MBC (N=20); Post T-DXd and/or SG Part B: mTNBC (N=20); Post Dato-DXd and/or SG -> Part B Efficacy Analysis: ORR and PFS-6 Part Z: HER2+ MBC (N=21); PD on prior T-DXd -> Part Z Efficacy Analysis: ORR and PFS-6 Endpoints: Overall response rate (ORR) Progression-free survival at 6 months (PFS-6) Duration of response (DoR) PFS, clinical benefit rate (CBR) Safety and tolerability *All study participants received HER3-DXd 5.6 mg/kg IV q3w. **Prior exposure to sacituzumab govitecan (SG) was permitted. In Part A, 6 pts with mTNBC received prior SG. Participants Part | A: HR+/HER2- MBC or mTNBC, Topoisomerase I exatecan-naive (N=60) | B: HR+/HER2- MBC or mTNBC, Prior ADC (N=40) Baseline Characteristics Median age (years, range) | 57.5 (32, 80) | 57 (29, 78) Female (n, %) | 59 (98.3) | 40 (100) White (n, %) | 46 (76.7) | 31 (77.5) Prior Regimens Prior ADC (n, %) | 6 (10)* | 40 (100) ADC as latest therapy (versus an earlier line of tx) | 5 (8.3) | 25 (62.5) Median prior metastatic regimens (n, range) | 3 (1, 7) | 4 (1, 8) 4+ prior metastatic regimens (n, %) | 22 (36.7) | 23 (57.5) Baseline HER3** Baseline HER3 ≥50% (n, %) | 40 (66.7) | 23 (57.5) Baseline HER3 <50% (n, %) | 7 (11.7) | 13 (32.5) HER3 data unavailable (n, %) | 13 (21.7) | 4 (10.0) Hamilton E. ESMO Breast 2026 @ErikaHamilton9

[Slide 2] Gene-expression profiling of HR+/HER2- BC arising in gBRCA1/2 mutation carriers identifies distinct molecular features with potential therapeutic implications 1844P Gaia Griguolo1,2, Michele Bottosso1,2, Giusy Landa1,2, Fabio Girardi2, Giovanni Faggioni2, Federica Miglietta1,2, Giorgio Bonomi1,2, Cristina Falci2, Alice Menichetti2, Eleonora Mioranza2, Rocco Cappellesso3, Angelo Paolo Dei Tos3,4, Maria Vittoria Dieci1,2, Valentina Guarneri1,2 1. DiSCOG, University of Padova, Padova, IT; 2. Oncology 2, IOV IRCCS, Padova, IT; 3. Pathological Anatomy Unit, Department of Integrated Diagnostics, Padova University Hospital, Padova, IT; 4. Department of Medicine, University of Padova, Padova, IT BACKGROUND Increasing evidence suggests reduced sensitivity to endocrine therapy and CDK4/6 inhibitors (CDK4/6i) in HR+/HER2- breast cancer (BC) arising in germline BRCA1/2 mutation (gBRCAmut) (pathogenic/likely pathogenic) carriers. Whether the pathogenetic mechanisms underlying these tumors translate into distinct biological features compared with sporadic HR+/HER2- BC remains unclear. METHODS We identified patients (pts) with stage I-III HR+/HER2- BC with known gBRCA1/2 mutations treated at our institution (2002-2024) and matched them 1:1 for age and histologic grade with pts without known gBRCA1/2 mutations. RNA from FFPE tumor samples was analyzed using Breast Cancer 360 panel (nCounter). PAM50 intrinsic subtypes and gene signatures were calculated, and single-gene expression was compared. RESULTS Study cohort: 33 gBRCAmut pts (6 BRCA1, 27 BRCA2) and 33 matched pts without known gBRCA1/2 mutations Table 1. Study cohort characteristics | gBRCAmut (N=33) Median | Range | gBRCA wt/unknown (N=33) Median | Range | p-value Age at BC diagnosis: median (range) | 45 | 28-69 | 47 | 30-77 | 0.691 ER expression: median (range) | 90 | 10-100 | 90 | 25-100 | 0.417 PgR expression: median (range) | 65 | 0-95 | 85 | 1-100 | 0.009 Ki67 expression: median (range) | 31 | 2-75 | 30 | 10-75 | 0.634 | N | % | N | % | p-value Histotype NST | 28 | 84.8% | 32 | 97.0% | 0.197 Histotype Lobular | 5 | 15.2% | 1 | 3.0% | Histological Grade G2 | 11 | 33.3% | 11 | 33.3% | 1.00 Histological Grade G3 | 22 | 66.7% | 22 | 66.7% | Stage at first BC diagnosis Stage I | 15 | 45.5% | 2 | 6.1% | 0.001 Stage II | 14 | 42.4% | 24 | 72.7% | Stage III | 4 | 12.1% | 7 | 21.2% | PAM50 intrinsic subtype distribution according to gBRCAmut status gBRCA1/2 carriers presented a numerically higher prevalence of PAM50 Basal-like subtype (30.3% vs 18.2%, not significant) gBRCAmut (N=33): 30.3% | 36.4% | 33.3% gBRCA wt/unknown (N=33): 18.2% | 42.4% | 39.4% Basal-like | Luminal B | Luminal A p=0.517 Distribution of PAM50 continuous signatures differs significantly according to gBRCAmut status Higher expression of Basal and lower expression of Luminal-B and HER2-enriched signatures in gBRCAmut tumors Basal-like p < 0.001 | HER2-enriched p = 0.017 | Luminal-A p = 0.230 | Luminal-B p = 0.029 | Normal-like p = 0.068 Differentially expressed genes according to gBRCA status (SAM analysis, 5% FDR) 58 genes were significantly underexpressed and 400 significantly overexpressed in gBRCAmut BCs (examples below) Underexpressed in gBRCAmut: PGR, RB1 Overexpressed in gBRCAmut - Genes associated with resistance to CDK4/6i: CCNE1, AURKA, MYC Overexpressed in gBRCAmut - Immune-related genes: CD274 (PD-L1), PDCD1 (PD-1) CONCLUSIONS HR+/HER2- BCs arising in gBRCA1/2 carriers are characterized by increased expression of Basal-like PAM50 signature and genes associated with CDK4/6i resistance and immune activation. If confirmed, HR+/HER2- BCs arising in gBRCA1/2 carriers may require tailored therapeutic strategies. Contact gaia.griguolo@unipd.it for permission to reprint and/or distribute Disclosures: GG reports Advisory Board fees: Pfizer, Eli Lilly; Speaker fees: Eli Lilly, Novartis; Research Funds: Fondazione Roche; Travel Support: Gilead, Eli Lilly, Pfizer, Novartis, Daiichi Sankyo, AstraZeneca

51P – ISOLATED INTRACRANIAL PROGRESSION IN METASTATIC BREAST CANCER (MBC): OUTCOME OF STEREOTACTIC RADIOTHERAPY (SRT) WITHOUT CHANGE IN SYSTEMIC TREATMENT Michele Bottosso(1), Gaia Griguolo(1,2), Giusy Landa(1,2), Marina La Commare(1,2), Laura Trovò(1,2), Tommaso Giarratano(1), Anna Gaspari(1,2), Giorgio Bonomi(1,2), Luisa Bellu(3), Cristina Falci(1), Federica Miglietta(1,2), Vittoria Aldegheri(4), Grazia Vernaci(1), Maria Vittoria Dieci(1,2), Francesco Pasqualetti(2,3), Valentina Guarneri(1,2) (1)Division of Oncology 2, Istituto Oncologico Veneto IRCCS, Padova, IT (2)Department of Surgery, Oncology and Gastroenterology, University of Padova, Padova, IT (3)Radiotherapy Unit, Istituto Oncologico Veneto IRCCS, Padova, IT (4)Radiology Unit, Istituto Oncologico Veneto IRCCS, Padova, IT Background • Optimal management of isolated brain progression in MBC patients (pts) with controlled extracranial disease remains poorly defined. • When feasible, multiple treatments with SRT while continuing ongoing systemic therapy represents a widely adopted approach, especially in HER2+ MBC, although supporting evidence is limited. Methods • We retrospectively identified MBC pts with isolated intracranial progression treated with SRT without systemic treatment modification. • Each SRT course was analyzed independently. • Intracranial progression-free survival (icPFS) was defined as time from SRT to further intracranial progression or death (censored at extracranial progression or last follow-up). Conclusions • In pts with isolated brain progression, treatment with SRT while continuing the ongoing systemic treatment yields clinically meaningful disease control, especially in cases of HER2+ MBC and limited intracranial burden. • Repeated locoregional treatments may remain effective in pts with prior durable benefit. Disclosures: presenting author reports personal fees from MSD, AstraZeneca, Pfizer, Novartis, Eli Lilly, Gilead and Daiichi Sankyo, and travel support from Eli Lilly, Daiichi Sankyo and Gilead. Printing costs were covered by DiSCOG-UNIPD. michele.bottosso@iov.veneto.it Results A total of 73 SRT treatments were identified, of which 20 retreatments after previous RT during the same systemic therapy line (Table 1). Table 1. Clinicopathological characteristics at SRT treatment | N=73 (%) Median age at SRT, years (range) | 53 (31-82) Breast cancer subtype HR+/HER2- | 29 (39.7) HER2+ | 40 (54.8) Triple negative | 4 (5.5) Karnofsky Performance Status 90-100 | 62 (84.9) 60-80 | 11 (15.1) Number of progressive brain lesion 1 | 27 (37.0) >1 | 46 (63.0) Line systemic therapy 1 | 25 (34.2) >1 | 48 (65.8) Previous RT within the same line No | 53 (72.6) Yes | 20 (27.4) At a median follow-up of 26.5 months, median icPFS was 7.5 months (95% CI 5.4-9.5; Figure 1). Figure 1. Intracranial PFS after SRT (icPFS probability vs Months from SRT, 0-30) HER2-positivity, presence of a single progressive brain lesion and first brain RT within a specific line of systemic therapy were independently associated with improved icPFS at multivariable analysis (Table 2). Hormone receptor status, line of therapy, performance status, and systemic treatment duration before intracranial progression were not prognostic. Table 2. Multivariable analysis of factors associated with icPFS following SRT | N (%) | Median icPFS (95%CI) | HR (95%CI) at multivariable analysis | p-value HER2-status Positive | 42 (57.5) | 8.3 (6.7-9.9) | 0.31 (0.16-0.59) | <0.001 Negative | 31(42.5) | 4.3 (3.2-5.6) | Ref. | Number of brain lesions 1 | 27 (37.0) | 9.3 (6.8-11.8) | 0.37 (0.19-0.74) | 0.005 >1 | 46 (63.0) | 5.4 (4.6-6.2) | Ref. | Previous RT within the same line No | 53 (72.6) | 8.0 (7.1-8.9) | 0.51 (0.27-0.96) | 0.036 Yes | 20 (27.4) | 5.0 (3.8-6.2) | Ref. | In cases receiving SRT after previous RT treatment during the same line, an icPFS >6 months following the previous RT was associated with improved outcomes (median icPFS 8.0 vs. 3.5 months; p=0.006; Figure 2). Figure 2. Patients who received multiple SRT courses during the same line of systemic therapy (swimmer plot; legend: First SRT, Subsequent SRT, PD with systemic therapy change, Systemic therapy ongoing; Months from first SRT)

[Slide 3] EFS analysis of the 24 patients underwent RC+PLND [Three KM panels; Event-free survival rate (%); Time from neoadjuvant therapy (months) 0-60] Panel 1: pCR vs no-PCR: labels 100%, 84.6%, 84.6%, 75%; P=0.904 Panel 2: HER2+ vs HER3+ [legend as printed]: labels 93.8%, 85.9%, 83.3%, 55.6%; P=0.103 Panel 3: T2N0M0 vs >T2N0M0: labels 92.9%, 92.9%, 87.5%, 54.7%; P=0.112 > RC + PLND patients: 1-/2-/3-year EFS = 100% / 90.6% / 78.1%. > pCR patients: 1-/2-/3-year EFS = 100% / 84.6% / 84.6%. > HER2 3+ vs HER2 2+: no significant EFS difference (3-year EFS 55.6% vs 85.9%). > T2N0M0 vs >T2N0M0: 3-year EFS 92.9% vs 54.7%, but difference not statistically significant. A Phase II Multicenter Study of Neoadjuvant Disitamab Vedotin Combined with Gemcitabine in MIBC Patients with HER2 2-3+: Interim Prognostic Outcomes Survival table Endpoint: Event-free survival (24 patients who underwent RC+PLND) All RC+PLND patients | 24 pCR HER2 3+ HER2 2+ T2N0M0 >T2N0M0 P: pCR vs no-pCR P=0.904; HER2 P=0.103; stage P=0.112 Landmark rates: 1 year: All RC+PLND 100% | pCR 100% 2 years: All RC+PLND 90.6% | pCR 84.6% 3 years: All RC+PLND 78.1% | pCR 84.6% | HER2 3+ 55.6% | HER2 2+ 85.9% | T2N0M0 92.9% | >T2N0M0 54.7% [Slide 4] Conclusion > The updated interim analysis of the GUARD-01 trial demonstrates that neoadjuvant GD regimen yields a high pCR rate and durable long-term EFS survival benefit in HER-2 overexpressing MIBC patients. > Improved safety profiles broadens the eligible population for neoadjuvant treatment. A Phase II Multicenter Study of Neoadjuvant Disitamab Vedotin Combined with Gemcitabine in MIBC Patients with HER2 2-3+: Interim Prognostic Outcomes

[Slide 1] 18F-FES PET–defined whole-body ER heterogeneity identifies endocrine responsiveness and predicts survival in ER+/HER2– Metastatic Breast Cancer (MBC): results of a prospective international clinical trial Alessandra Gennari, Simone Nardin, Etienne Brain, Arnoldo Piccardo, Andrea Decensi, Matteo Lambertini, Federica Matteucci, Rachel Wuerstlein, Romain-David Seban, Francesco Fiz, Beatrice Ruffilli, Davide Corradengo, Gian Mauro Sacchetti, Valeria Fontana, Oriana Nanni, Michela Palleschi, Valentina Rossi, Francesca D'Avanzo, Luca Boni, Nadia Harbeck, on behalf of the ET-FES Collaborative Group Alessandra Gennari, MD, PhD 24 October 2026 [Slide 3] PROGRESSION-FREE AND OVERALL SURVIVAL Median Progression-free survival Group A: 18.1 months (95%CI 11.3-25.3) Group B/C: 12.0 months (95%CI 6.6-22.9) HR 1.36 (95%CI 0.88-2.10); p=0.162 Progression-free survival (%); Years 0–11 No. at Risk (No. Cumulative Censors) Group A: 92 (0) 53 (0) 39 (0) 29 (0) 25 (0) 21 (1) 11 (6) 9 (7) 7 (9) 4 (12) 2 (14) 0 (16) Group B+C: 31 (0) 16 (0) 9 (0) 7 (0) 6 (0) 3 (1) 2 (2) 1 (2) 0 (3) Median Overall survival Group A: 69.6 months (95% CI, 48.3-85.1) Group B/C: 42.5 months (95% CI, 27.0-70.4) HR = 1.87 (95% CI 1.16–3.01); p = 0.010 Overall survival (%); Years No. at Risk (No. Cumulative Censors): Group A 92 (0) ... ; Group B+C 32 (0) ... [remaining OS at-risk values too small to read reliably] Gennari A. Survival table Endpoint: Progression-free survival Arm | N | Median | Median CI Group A (100% concordant) | 92 | 18.1 months | 11.3-25.3 Group B/C | 31 | 12.0 months | 6.6-22.9 HR: 1.36 HR CI: 0.88-2.10 P: 0.162

ER and HER2 crosstalk: Rationale for dual blockade of pathways [Panels a, b, c: schematic diagrams of estrogen/ER and HER2 signalling, labels too small to transcribe] • 50-70% of HER2+ BC co-express HR • In the absence of HER2 blockade, HER2+ ER+ breast cancers have lower levels of ER, may have constitutively activated ER, and are less sensitive to endocrine therapy • Blockade of HER2 in HER2+ breast cancer can lead to upregulation of ER-regulated gene transcription • Pietras, et al were the first to demonstrate that combined receptor blockade (anti-HER2 antibody and tamoxifen) resulted in greater combined efficacy compared to either agent alone in HR+/HER2+ BC Pegram M et al. NPJ Breast Cancer. 2023;9. Pietras, R. J. et al. Oncogene 1995;10:2435–2446. Guo S et al. Mol Cell Biol. 2004;24:8681-90. Konecny G et al. J Natl Cancer Inst. 2003;95:142-53. Xia W et al. Proc Natl Acad Sci. 2006;103:7795-7800; Wang Y et al. Breast Cancer Res 2011;13:R121. Sara A. Hurvitz

[Slide 2] HER2-PET Redefines the Tumor Landscape and Stratifies PFS Redefining HER2 Landscape: Correspondence Between Pathological HER2, HER2-PET, and HR Status (N=89) HER2 Status: HER2-IHC 0 7 | HER2-low 54 | HER2-positive 28 HER2-PET Status: HER2-PET-N 17 | HER2-PET-P 72 HR Status: HR+ 49 | HR- 40 Progression-Free Survival (PFS) according to HER2–PET Status | HER2-PET N | HER2-PET [P; cropped] Events | 14/17 | 27/72 Median PFS (95% CI), months | 5.1 (3.6–NR) | 13.6 (11.9–N[R; cropped]) Hazard ratio (95% CI): 4.09 (2.07-8.11); P<0.[cropped] Progression-free survival (%) ; Time (months) 0 5 10 15 20 Patients at risk HER2-PET N | 17 | 10 | 2 | 0 | 0 HER2-PET P | 72 | 61 | 38 | 13 | 2 Survival table Endpoint: PFS Arm | N | Median | Median CI HER2-PET N | 17 | 5.1 | 3.6–NR HER2-PET P | 72 | 13.6 | 11.9–NR (right edge cropped; 'NR' inferred from partial 'N') HR: 4.09 HR CI: 2.07-8.11 [Slide 3] PFS according to HER2 Status and HER2–PET/CT Group | mPFS (months) 95% CI HER2-positive+ HER2-PET P | NR (NR – NR) HER2-low/zero+ HER2-PET P | 11.9 (10.3 – NR) HER2-positive+ HER2-PET N | 9.0 (3.6 – NR) HER2-low/zero+ HER2-PET N | 4.8 (2.9 – NR) Progression-free survival (%) ; Time (months) 0 5 10 15 20 Patients at risk | | | | | | HR (95%CI) | P value HER2-positive +HER2-PET P | 21 | 20 | 17 | 7 | 1 | Reference | - HER2-low/zero +HER2-PET P | 51 | 41 | 21 | 6 | 1 | 4.2 (1.0 - 17.8) | 0.05 HER2-positive +HER2-PET N | 7 | 5 | 1 | 0 | 0 | 7.6 (1.4 - 41.7) | 0.02 HER2-low/zero +HER2-PET N | 10 | 5 | 1 | 0 | 0 | 16.6 (3.5 - 77.7) | < 0.001 Survival table Endpoint: PFS Arm | N | Median | Median CI HER2-positive + HER2-PET P | 21 | NR | NR – NR HER2-low/zero + HER2-PET P | 51 | 11.9 | 10.3 – NR HER2-positive + HER2-PET N | 7 | 9.0 | 3.6 – NR HER2-low/zero + HER2-PET N | 10 | 4.8 | 2.9 – NR

15:30 - 16:30 Rapid fire session 2 CHAIRS: STAVROULA CHANTE, MARIA LAPUENTE Development of a patient-initiated survivorship follow-up model 1 Increasing service demand Increasing pressure on traditional medical oncology follow-up Patient-initiated follow-up established internationally 2 A new model of follow-up First patient-initiated survivorship follow-up programme for early breast cancer in Ireland Introduced in May 2024 Replaced routine medical oncology appointments Patients initiate their own review throughout follow-up 3 Patient-centred care Remote, individualised care Rapid access to CNS's (medical oncology and breast care) Promotes patient empowerment and self-management Brid O' Riordan Brid O' Riordan Patient empowerment and self-management outcomes from a clinical nurse specialist-led early breast cancer survivorship programme SEGOVIA AUDITORIUM - PALACIO MUNICIPAL

[Slide 1] Pathology AI Digital 70-Gene Signature MammaPrint Pathology AI Technology Steps Overview • AI Digital Model is trained on 70,000 FFPE slides with routine MammaPrint microarray results available • Preprocessing (Quality Control - QC) ensures that there is sufficient invasive tumor content in each FFPE slide • AI Digital MammaPrint used Self-Supervised Features Recognition (left), and Microarray MammaPrint-supervised learning models (right) to produce the Digital MammaPrint Index [Flow: Input WSI → Preprocessing (dTu%) module → Pathology SSL: Self-Supervised Learning (Feature Sets Extraction) → Core AI Multiple Instance Learning (digital MammaPrint Pattern Recognition); End-to-End 'one shot' (digital MammaPrint Label Prediction) → Ensemble module → Digital MammaPrint Index] Laura J van 't Veer Laura Van't Veer — AI driven digital pathology for the 70 gene recurrence signature: Development and validation in early stage breast cancer [Slide 2] Pathology AI Digital MammaPrint Performance Pathology AI Independent Performance in 1,000 slides • 1,000 additional FFPE slides Whole Slide Images with routine MammaPrint microarray results available • Preprocessing (invasive tumor) • Read-out Digital MammaPrint index by Ensemble Model • Result: 95% correlation • Test is extensively validated for Precision, Reproducibility, Limit of Detection etc • In-Vitro Diagnostic Regulatory review pending (IVDR, CE, etc) [Scatter: AI Digital MammaPrint Predicted Index (DART Score) vs GT Index (MP_Index) MammaPrint Microarray; Correct / Incorrect; correlation 95%] Laura J van 't Veer [Slide 3] Clinical Validation Pathology AI Digital MammaPrint Validation Results Clinical Outcome All Patients – BCSS: Digital MammaPrint Low risk vs High risk; HR 0.12 (95% CI 0.04–0.39) p < 0.0001 Number at risk (years 0, 2, 4, 6, 8, 10, 12): Low 187 187 184 171 162 113 21; High 183 178 171 159 145 105 11 ER+/HER2- Patients – BCSS: HR 0.10 (95% CI 0.03–0.36) p < 0.0001 Number at risk: Low 181 181 178 166 157 109 20; High 91 91 88 79 70 52 4 • 372 RASTER patients H&E-stained Whole Slide Images (WSI) • 189 patient tumors were (51%) D-Low and 183 (49%) D-High • D-High patients 74% had received chemo, D-Low pts 24% had received chemo • 10-year BCSS was 98.1% for D-Low versus 86.7% for D-High (all patients) In multivariable analyses, AI Digital MammaPrint was the strongest predictor of BCSS HR 0.084 (95%CI 0.02-0.31), p<0.001, outperforming clinical factors. Similar results for the DRFI endpoint Laura J van 't Veer Survival table Endpoint: Breast cancer-specific survival (BCSS), all patients Digital MammaPrint Low risk | 189 Digital MammaPrint High risk | 183 HR: 0.12 HR CI: 0.04–0.39 P: <0.0001 Landmark rates: 10 years: D-Low 98.1% | D-High 86.7% [Slide 4] AI Digital MammaPrint Conclusions and Utility • First time and successful Clinical Validation of AI Digital Pathology MammaPrint • AI Digital Pathology MammaPrint demonstrates prognostic value for DRFI and BCSS • Highly concordant with the Microarray MammaPrint Assay • AI-based MammaPrint assay is feasible from routine FFPE H&E-stained Whole Slide Images for Early Breast Cancer • AI-based MammaPrint Assay can be integrated with other AI features and forms the foundation for further image-based biomarker development Laura J van 't Veer

Survival Outcomes in Metaplastic Breast Cancer: A Stage and Receptor-Matched Real-World Analysis Milad Nourianpour, Salama Bin Hendi, Safa Al-Dulaimi, Farah Abou Fakhr, Shahd Abdelghany, Saeed Rafii Mohammed Bin Rashid University of Medicine and Health Sciences (MBRU), Dubai, UAE; Mediclinic City Hospital, Dubai, UAE FPN 48P | ESMO Congress 2026 - Madrid Background • Metaplastic breast cancer (MpBC) is a rare, heterogeneous and clinically aggressive subtype. • Its outcomes relative to conventional breast cancer within the same stage and receptor phenotype are less well described. Aim: compare 5-year overall survival (OS) between MpBC and non-MpBC overall, within each AJCC stage and within triple-negative (TN) disease. Methods Data: TriNetX Research Network, 112 healthcare organisations; analyses run 22 April 2026. Population: females ≥18 years with breast cancer (ICD-10-CM C50). MpBC: ICD-O-3 8560/3, 8562/3, 8570–8572/3, 8575/3, 8980–8982/3. Non-MpBC: C50 without these codes. Stage and receptors: AJCC stage and ER/PR/HER2 status recorded from 1 month before to 6 months after diagnosis. Outcome: all-cause death from day 1 to 5 years after first breast cancer diagnosis. Primary analysis: 1:1 propensity score matching (PSM) on age at diagnosis. Stage and TN analyses: cohorts restricted to each stage or to TN disease; not propensity-matched. Statistics: Kaplan–Meier, log-rank test, Cox hazard ratio (HR) with 95% CI; proportional-hazards assumption met in all analyses (p>0.5). Study cohorts TriNetX Research Network · 112 HCOs; Females ≥18 years with breast cancer MpBC n=596 | Non-MpBC n=1,242,018 1:1 PSM on age → 552 vs 552; Mean age 58.7 ± 12.7 years in both arms Stage recorded 412 of 596 MpBC (69%) | ER/PR/HER2 recorded 273 of 596 MpBC (46%) Primary analysis: 5-year OS, age-matched cohorts 89.3% | 72.6% HR 2.76 (95% CI 1.93–3.94) Log-rank p<0.001 MpBC (n=551; 122 deaths) | Non-MpBC, age-matched (n=550; 40 deaths) 72.6% vs 89.3% (5-year OS, MpBC vs non-MpBC) | HR 2.76 (95% CI 1.93–3.94; p<0.001) | 122 vs 40 (deaths within 5 years (n=551 vs 550)) Stage-stratified 5-year OS Stage I: 95.5% vs 82.1%; HR 4.36 (2.66–7.14) p<0.001; MpBC (n=105), Non-MpBC (n=35,233) Stage II: 91.3% vs 81.1%; HR 2.26 (1.56–3.29) p<0.001; MpBC (n=179), Non-MpBC (n=25,873) Stage III: 77.3% vs 63.2%; HR 1.80 (1.17–2.76) p=0.007; MpBC (n=77), Non-MpBC (n=16,165) Stage IV: 46.4% vs 27.2%; HR 1.75 (1.14–2.68) p=0.010; Median OS 20.4 vs 51.5 mo; MpBC (n=40), Non-MpBC (n=8,360) Triple-negative disease: 79.3% vs 73.7%; HR 1.32 (0.97–1.78) p=0.076; MpBC TN (n=193), Non-MpBC TN (n=5,245) How MpBC presented AJCC stage (%), among patients with a recorded stage — MpBC (n=412) | Non-MpBC (n=93,882): Stage I 26.0 | 40.2; Stage II 45.1 | 31.2; Stage III 18.7 | 18.9; Stage IV 10.2 | 9.6 Receptor phenotype (%), among patients with recorded status — MpBC (n=273) | Non-MpBC (n=43,035): Triple-negative 71.4 | 12.4; HR+/HER2− 21.2 | 64.6; HER2+ 7.3 | 23.0 Hazard ratios across all analyses Analysis | MpBC n (deaths) | Non-MpBC n (deaths) | HR (95% CI) | p Overall (age-matched) | 551 (122) | 550 (40) | 2.76 (1.93–3.94) | <0.001 Stage I | 105 (16) | 35,233 (1,222) | 4.36 (2.66–7.14) | <0.001 Stage II | 179 (28) | 25,873 (1,724) | 2.26 (1.56–3.29) | <0.001 Stage III | 77 (21) | 16,165 (2,554) | 1.80 (1.17–2.76) | 0.007 Stage IV | 40 (21) | 8,360 (3,197) | 1.75 (1.14–2.68) | 0.010 Triple-negative | 193 (44) | 5,245 (905) | 1.32 (0.97–1.78) | 0.076 HR (log scale) · >1 = higher mortality with MpBC Conclusions • MpBC had significantly worse 5-year OS than age-matched non-MpBC (72.6% vs 89.3%; HR 2.76). • The survival gap was present at every AJCC stage and was largest in stage I (HR 4.36). • MpBC most often presented at stage II and was predominantly triple-negative. • Within TN disease the difference was smaller and not significant (HR 1.32; 95% CI 0.97–1.78), suggesting the TN phenotype explains part of the excess risk. • MpBC-specific research and treatment strategies are needed. Limitations • Retrospective EHR data; diagnosis and morphology coding may be misclassified. • Stage recorded for 69% and receptor status for 46% of MpBC patients. • Primary analysis matched on age only; treatment and comorbidity not analysed. Stage and TN analyses unmatched. • All-cause mortality; median follow-up 53 vs 40 months (MpBC vs matched non-MpBC). Disclosure: the first author (M. Nourianpour) and the presenting author (S. Rafii) have no conflicts of interest to declare. Funding: none. Contact: Dr Saeed Rafii, saeed.rafii@mediclinic.ae

Preclinical Models Validate Efflux and Impaired Internalization as Mechanisms of T-DXd Resistance • ABCB1 or ABCC1 overexpression in NCI-N87 cells significantly reduced intracellular payload exposure and T-DXd sensitivity. • siRNA-mediated MAN2A2 silencing in KPL4 cells decreased T-DXd binding and internalization, reducing intracellular payload exposure and T-DXd sensitivity. Efflux pump overexpression impacts payload exposure in vitro Parental ABCC1 OE ABCB1 OE ABCG2 OE ABCC1 ABCB1 ABCG2 b-Actin HER2 b-Actin N87 Payload exposure Intracellular DXd (% of parental ±SD) 0.0755 <0.0001 <0.0001 N=3 biological replicates with N=4 technical replicates ANOVA with Dunnett's post-hoc Sensitivity IC50 (µg/mL) EMax Parental N87 0.06819 -67.989 ABCC1 OE [illegible] 20.929 ABCB1 OE 2.087 -33.342 ABCG2 OE 0.4553 -41.902 % Viability (mean ± SD) T-DXd (µg/ml) Representative curves from N=3 biological replicates, N=4 technical replicates MAN2A2 downregulation reduces T-DXd internalization in vitro Binding Trastuzumab binding (% ± SD) 0.0475 Ctrl siMAN2A2 Internalization Fabfluor-pH Red Normalized Signal Control siMAN2A2 KPL4 Time (h) Payload exposure Intracellular DXd (% ± SD) <0.0001 Ctrl siMAN2A2 Sensitivity % Viability (mean ± SD) Control siMAN2A2 T-DXd concentration (µg/mL) Paolo Tarantino, MD, PhD Paolo Tarantino Paired-biopsy multi-omic analysis reveals target-independent mechanisms of antibody–drug conjugate (ADC) resistance in metastatic breast cancer (MBC)

1780P - Safety, social costs and quality of life (QoL) in HER2-positive (HER2+) breast cancer (BC) patients (pts) treated with Pertuzumab (P) and Trastuzumab (T) Fixed Dose combination (FDC): results from the prospective, observational, multicenter GIM30-RAPID study G. Buono1, M. Piezzo1, R. Caputo1, O. Tinelli2, V. Di Lauro1, F. Carlino3, A. Piccolo3, G. Pavone3, L. Leo4, F. Vitiello4, A. Fabi5, P.C. Sacco6, S. Della Rossa7, L. Gerratana7, F. Puglisi7, M. Orditura3, M. Managò8, A. Marcellusi8, C. De Angelis1,2, M. De Laurentiis1,2 [affiliations: Istituto Nazionale Tumori – IRCCS "Fondazione G. Pascale", Naples; Scuola Superiore Meridionale, Naples; AORN Sant'Anna e San Sebastiano, Caserta; Ospedaliero Nola/Pollena, ASL NA3SUD, Naples; Fondazione Policlinico Universitario A. Gemelli IRCCS, Rome; Azienda Ospedaliera S Giuseppe Moscati, Avellino; Centro di Riferimento Oncologico di Aviano (CRO) IRCCS, Aviano; University of Udine; University of Milan] BACKGROUND • PT+ chemotherapy (CT) represents the standard of care for pts with early (high-risk neo-/adjuvant) and advanced (I line) HER2+ BC 1,2,3. • PT FDC subcutaneous (SC) formulation demonstrated non-inferiority to intravenous (IV) administration in HER2+ BC pts 4. • The PHASTER study had previously shown economic/organizational advantages of SC over IV 5. • We previously reported initial findings from the GIM30 study [social costs and quality of life (QoL)] 6. Here we show the updated analysis, including adverse events (AEs). METHODS • GIM30-RAPID (NCT06835348) is a prospective, observational, multicenter study which enrolled pts with HER2+ BC after CT+PT IV who were switching or had already switched to FDC SC with at least 6 cycles of therapy expected to complete pre-established oncological treatment. • The primary objective was safety (number and percentage of G3/4 adverse events according to NCI CTCAE v.5). • Secondary objectives included: indirect social costs, measured as loss of productivity for pts and caregivers related to absence from work (absenteeism) and reduced productivity while being at work (presenteeism), and QoL assessed using the EORTC QLQ-C30. • Questionnaires were administered at screening (S), after cycle 3 (Treatment) and at the end of study (EoS). RESULTS • Overall, 90 patients were enrolled across 6 Italian cancer centers (March 2023–November 2024). Baseline characteristics are shown below. Characteristics | n = 90 Median age, years (range) | 56 (26-82) Switch to PT FDC SC - n (%): Pre-enrollment 69 (76.7%); At enrollment 21 (23.3%) Disease status at first diagnosis - n (%): Stage I/II 32 (35.6%); Stage III 19 (21.1%); Stage IV 35 (38.9%); Unknown 4 (4.4%) Disease status at current diagnosis - n (%): Early BC 40 (44.4%); Metastatic BC 48 (53.3%); Unknown 2 (2.2%) HER2 status - n (%): IHC 2+, ISH Amplified 15 (16.7%); IHC 3+ 75 (83.3%) ER status - n (%): Positive 55 (61.1%); Negative 35 (38.9%) PgR status - n (%): Positive 40 (44.4%); Negative 50 (55.6%) ECOG PS - n (%): 0-1 77 (85.6%); 2 2 (2.2%); Unknown 11 (12.2%) SAFETY I Adverse events [n=83* in 37 pts (41.1%)] Distribution by CTCAE grade: G1: 76/83 (91.6%); G2: 5/83 (6.0%); G3: 1/83 (1.2%); G4: 1/83 (1.2%) 76/83 adverse events (91.6%) were grade 1. Both grade 3–4 LVEF reductions were treatment-related. *AEs numbers and percentages were updated after DB cleaning Clinically relevant adverse events Adverse event | G1 | G2 | G3 | G4 | Total n (%) Arthralgia | 7 | 0 | 0 | 0 | 7 (8.4) Nausea | 6 | 1 | 0 | 0 | 7 (8.4) Diarrhea | 6 | 0 | 0 | 0 | 6 (7.2) Fatigue | 4 | 1 | 0 | 0 | 5 (6.0) Onychopathy | 5 | 0 | 0 | 0 | 5 (6.0) Asthenia | 4 | 0 | 0 | 0 | 4 (4.8) Paresthesia | 4 | 0 | 0 | 0 | 4 (4.8) Headache | 3 | 0 | 0 | 0 | 3 (3.6) Pain | 3 | 0 | 0 | 0 | 3 (3.6) Injection-site reaction | 1 | 0 | 0 | 0 | 1 (1.2) Dyspnea | 1 | 0 | 0 | 0 | 1 (1.2) LVEF reduction** | 0 | 0 | 1 | 1 | 2 (2.4) Other events | 32 | 3 | 0 | 0 | 35 (42.2) TOTAL | 76 | 5 | 1 | 1 | 83 (100.0) ** Both G3 and G4 LVEF reductions were treatment-related. INDIRECT SOCIAL COSTS I Productivity and annual costs Absenteeism and presenteeism (days): Patient presenteeism / Patient absenteeism / Caregiver absenteeism — Screening 13.3 / 13.5 / 7.83; Treatment (C3) 10.0 / 10.6 / 5.6; End of study 3.7 / 4.4 / 1.8 Patient and caregiver work-loss days declined across assessments, alongside lower estimated indirect costs Annual indirect social cost per patient: [line chart; labels 4.984,08 €; 3.597,61 €; 2.475,83 €; 1.857,10 €; 1.563,55 €; 1.123,51 €; 944,70 €; 617,00 €; 1.158,93 €; 692,09 €; 361,84 €; 105,00 €; series: Pts absenteeism cost, Pts presenteeism cost, Caregiver absenteeism cost, Total annual indirect costs; Screening / Treatment / EoS] QUALITY OF LIFE I EORTC QLQ-C30 (n = 78) A. Functional scales and global health status [line chart: Physical, Role, Emotional, Cognitive, Social functioning, Global health status; Screening / Treatment / End of study] B. Symptom scales [line chart: Pain, Fatigue, Nausea/vomiting, Dyspnea, Insomnia, Appetite loss, Constipation, Diarrhea] An overall preservation of QoL was observed, with no statistically significant changes in Global Health Status (p = 0.828) or C30 Summary Score (p = 0.198). A statistically significant, although clinically modest, increase in pain scores was observed [19.4 (screening) vs. 25.1 (End Of Study); p = 0.027]. CONCLUSIONS Our results confirm a favorable safety profile for PT FDC SC, demonstrating high tolerability and rare serious AEs. These findings, combined with reduced indirect social costs and sustained QoL levels, support the SC formulation as a safe and efficient option for the treatment of pts with HER2+ BC. BIBLIOGRAPHY 1) Gianni L. et al., Lancet Oncol. 2012 Jan;13(1):25-32. doi: 10.1016/S1470-2045(11)70336-9 2) Baselga J. et al., N Eng J Med 2012 Jan 12;366(2):109-19. doi: 10.1056/NEJMoa1113216 3) von Minckwitz G. et al., N Engl J Med. 2017 Jul 13;377(2):122-131. doi: 10.1056/NEJMoa1703643 4) Tan R.A. et al., Lancet Oncol 2021 Jan;22(1):85-97. doi: 10.1016/S1470-2045(20)30536-2 5) Munzone E. et al., Drugs and Therapy Perspectives, 2023, 39(12), pp. 432–446 6) Buono G. et al., ESMO Breast Cancer 2025, 486P Funding: The project was partially funded by Roche (no-profit study) Disclosures: [small-print disclosure list — partially illegible] Contact: Giuseppe Buono, MD at giuseppe.buono@istitutotumori.na.it

TUXEDO-3: Key outcomes Cohort 1 (mBC with active BMs) 23.8% IC-ORR as per RANO-BM criteria Bartsch R et al. Lancet Oncol, 2025 Median OS 7.9 months Breast cancer phenotype: HER2 9 (42.9); Luminal 5 (23.8); Triple-negative 7 (33.3) Cohort 2 (aNSCLC with active BMs) 30.0% IC-ORR as per RANO-BM criteria Fuereder T et al. Lancet Oncol, 2025 Median OS 6.7 months Lung cancer phenotype: Non-squamous 17 (85.0); Squamous 3 (15.0); KRAS G12C 1 (5.0); EGFR 4 (20.0) Cohort 3 (LMD from solid tumors) 65.0% 3-month OS rate Preusser M et al. Nat Med, 2025 Median OS 10.5 months Number at risk: 20 19 17 13 12 10 10 7 6 4 3 1 0 Primary tumor location: Breast 12 (60.0); Lung 6 (30.0) Breast cancer phenotype: HER2 1 (8.3); Luminal 5 (41.7); Triple-negative 6 (50.0) Preusser et al, ESMO 2026

1796TiP BREnnA: Phase II/III randomised study of ZN-A-1041 or tucatinib (TUC), in combination with trastuzumab + capecitabine (HX), in patients (pts) with HER2-positive locally advanced or metastatic breast cancer (HER2+ LA/mBC) with or without central nervous system (CNS) metastases Nancy U. Lin1, Giampaolo Bianchini2, Erika Hamilton3, Weize Huang4, Adam Knott5, Julia Maués6, Filippo Montemurro7, Volkmar Müller8, Mahesh Shivhare5, Gregory Vidal9, Binghe Xu10, Eleonora Restuccia7, Javier Cortés11-14 1Dana-Farber Cancer Institute, Boston, MA, USA; 2IRCCS San Raffaele Hospital, Milan, Italy; 3Sarah Cannon Research Institute, Nashville, TN, USA; 4Genentech, Inc. South San Francisco, CA, USA; 5Roche Products Ltd, Welwyn Garden City, UK; 6Guiding Researchers and Advocates to Scientific Partnerships (GRASP), Baltimore, MD, USA; 7F. Hoffmann-La Roche Ltd, Basel, Switzerland; 8University Medical Center Hamburg-Eppendorf, Hamburg-Eppendorf, Germany; 9West Cancer Center and Research Institute, Germantown, TN, USA; 10Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, People's Republic of China; 11International Breast Cancer Center (IBCC), Pangaea Oncology, Quiron Group, Barcelona, Spain; 12IOB Madrid, Institute of Oncology, Hospital Beata María Ana, Madrid, Spain; 13Universidad Europea de Madrid, Faculty of Biomedical and Health Sciences, Department of Medicine, Madrid, Spain; 14Medica Scientia Innovation Research (MEDSIR), Barcelona, Spain and Ridgewood, NJ, USA Background (partly obscured by presenter) • CNS metastases are a common site of progressive disease in patients with HER2+ advanced BC, with nearly half of patients developing brain metastases and half of deaths being due to CNS progression. • Existing HER2-targeted therapies have shown varying degrees of CNS efficacy, but have limitations such as gastrointestinal or pulmonary toxi[...] preventive effect and observed disease p[...] • ZN-A-1[...] highly HER2-selective tyrosine kinas[...] blood–brain barrier penetration [...] wide therapeutic index. [remainder obscured] Figure 1. Study design Patients with previously treated HER2+ LA/mBC with or without CNS metastases† Stage 1* N=~234: 120 patients for dose selection ZN-A-1041 (dose 1 BID) + HX n=40 | n=~38 ZN-A-1041 (dose 2 BID) + HX n=40 | n=~38 TUC + HX (control arm) n=40 | n=~38 | n=~208 Interim dose selection analysis Stage 2* N=~416: ZN-A-1041 (selected dose) + HX n=~208 Additional accrual during follow-up of initial 120 enrolled patients -> Progressive disease -> Follow-up Randomisation: • Stage 1: three arms in a 1:1:1 ratio • Stage 2: two arms in a 1:1 ratio Stratification factors: • CNS metastasis status (yes vs. no) • Prior T-DXd exposure (yes vs. no) • World region (USA/Canada vs. Europe vs. Asia vs. RoW) Endpoints:‡ • Primary endpoint: PFS in the FAS • Key secondary endpoints: PFS in patients with CNS metastases at baseline and OS in the FAS • Other secondary endpoints: ORR, DoR and CBR in the FAS and in patients with CNS metastases at baseline; PROs; PK; and safety Table 1. Key eligibility criteria Key inclusion criteria: • Pathologically documented inoperable LABC, or mBC • Centrally confirmed HER2+ status • Measurable disease per RECIST v1.1 in Stage 1 (non-measurable disease allowed in Stage 2) • Previously treated or untreated active or stable CNS or leptomeningeal metastases permitted • ≥1 prior line of HER2-targeted treatment for LA/mBC and prior HER2-targeted ADC treatment (if appropriate) in any treatment setting • Prior TKI in the (neo)adjuvant setting permitted, provided there is a >12-month interval between TKI completion and LA/mBC development • Adequate organ and bone marrow function • ECOG performance status of 0 or 1 • Baseline LVEF ≥50% Key exclusion criteria: • Concurrent anti-cancer treatment, or treatment with an investigational therapy within 28 days prior to study treatment initiation • Previously untreated intracranial lesion of >2.0 cm, except if the patient has no neurological symptoms that would affect trial participation and safety* • Clinically significant cardiovascular disease or risk • Clinically significant electrolyte abnormalities, or family history of sudden unexplained death or long QT syndrome • Concomitant use of any agent known to strongly inhibit or induce CYP3A4 or CYP2C8 activity, or oral coumarin-derivative anticoagulants Figure 2. Participating countries* (world map) *Countries participating as of 10 September 2026; the projected number of centres is 330 across 28 countries. Statistical analysis • Approximately 234 patients will be enrolled in Stage 1 and 416 will be enrolled in Stage 2. • All analyses will be based on combined data from Stages 1 and 2. • The null hypothesis for the primary endpoint will be tested at a one-sided significance level of 0.025. • Null and alternative hypotheses with appropriate type I error control will be tested for key secondary efficacy analyses. • An independent data monitoring committee will be in place. Contact information: • Email: global.rochegenentechtrials@roche.com. • Phone: 1-888-662-6728 (USA only). • ClinicalTrials.gov: https://clinicaltrials.gov/ct2/show/NCT07413939. Acknowledgements: The authors thank the patients and their families, participating study investigators and clinical sites. This study is funded by F. Hoffmann-La Roche Ltd, Basel, Switzerland. Research support in the form of third-party medical writing assistance for this poster, furnished by Katie Wilson, PhD, of Nucleus Global (an Inizio Company), was provided by F. Hoffmann-La Roche Ltd. Presented at the ESMO Congress 2026, Madrid, Spain, 23–27 October 2026.

174P: A Multi-LLM Framework for Oncology Clinical Decision Support: A Breast Cancer Case Series Muhammad Shaheer Shaikh, Judith Pérez-Granado, Mark Gramling, Bruno Larvol The Larvol Group LLC, San Francisco/CA/USA Correspondence: judith.perez@larvol.com INTRODUCTION Access to timely specialist input in cancer care remains variable, particularly in community settings without routine tumour board support. Second opinions lead to clinically meaningful treatment changes in a significant proportion of oncology cases1, yet access is limited by geography, capacity and time. We built a multi-large language model (LLM) framework to address this gap: a second-opinion tool that processes real-world cases and returns a structured, consensus-style treatment ranking. CONCLUSIONS • LLM consensus matched the expert's first choice in 80% of cases and matched on both choices in 40%. • The first or second choice matched in 18 of 20 cases (90%). • Agreement among the four LLMs was high: full first-choice consensus in 85% of cases, at least three of four agreeing in 90%. • Only 2 of 20 cases showed no match: in one the expert combined two listed options; in the other the first and second rankings were inverted. • These findings support multi-LLM consensus as adjunctive clinical decision support in breast cancer. • Validation with multiple expert oncologists and additional tumor types is needed. METHODS 23 real-world breast cancer cases; 3 excluded for incomplete data · 20 analysed; HR+/HER2- · HR+/HER2+ · TNBC Run by four independent LLMs: ChatGPT-5.2 | Grok-4 | Gemini-3 Pro | Claude Sonnet 4.6 1 Clinical review: Real-world cases processed with standardised prompts. 2 Option normalisation: Regimens from all four LLMs consolidated into one option set per case. The same option set goes to both Four LLMs - Preference simulation: Each LLM assigns probability-weighted preferences simulating 100 U.S. community oncologists. Oncologist - Expert assessment: Independently ranks a first and a second choice from the same option set. Statistical analysis: LLM vs. expert: first- and second-choice agreement. Between LLMs: Kendall's τ and Jensen-Shannon divergence, pairwise across all six model pairs. RESULTS LLM consensus vs. expert oncologist · 20 evaluable cases: 80% matched the expert's first choice | 90% matched the first or second choice | 40% matched on both choices Case-level distribution of agreement (n = 20): 40% | 40% | 10% | 10% Both choices matched (8/20); First choice matched only (8/20); Second choice matched only (2/20); Neither matched (2/20) Agreement between the four LLMs · 20 evaluable cases: 85% of cases reached full 4/4 consensus | 0.666 mean Kendall's τ, higher = more similar | 0.161 mean JS divergence, lower = more similar Concordance by model Model | Top-1 | Top-2 | Kendall τ | JS div. ChatGPT-5.2 | 16/20 (80%) | 17/20 (85%) | 0.728 | 0.133 Claude Sonnet 4.6 | 14/20 (70%) | 15/20 (75%) | 0.624 | 0.192 Gemini-3 Pro | 15/20 (75%) | 16/20 (80%) | 0.628 | 0.180 Grok-4 | 17/20 (85%) | 17/20 (85%) | 0.683 | 0.138 Mean of four models | 78% | 81% | 0.666 | 0.161 Pooled LLM consensus | 16/20 (80%) | 18/20 (90%) Top-1: the model's top-ranked option matched the expert's first choice. Top-2: it matched the expert's first or second choice. Kendall τ and Jensen-Shannon (JS) divergence are between-LLM measures computed pairwise across all four model pairs. The pooled consensus combines the four models into one ranking per case; the row above it is their arithmetic mean. 1. Lipitz-Snyderman A, et al. Clinical value of second opinions in oncology. Cancer Med. 2023;12:8063-8072. ChatGPT-5.2, Claude Sonnet 4.6, Gemini-3 Pro and Grok-4 were used to run the analysis described in this poster. Claude was used to assist in drafting the text; the authors reviewed and edited the content and took full responsibility for it. The authors declare no conflicts of interest.

[Slide 2] 7P Impact of ctDNA detected BRCA reversion mutations on outcomes of patients with BRCA-mutated metastatic breast cancer treated with antibody-drug conjugates Jesus Yaringaño, Alessandra Spata, Federica Giugliano, Jessica Paparo, Joana M Ribeiro, Thomas Grinda, Cyril Roussel-Simonin, Alessandro Viansone, Jean Zeghondy, Chayma Bousrih, Tarek Ben Ahmed, Theophraste Henry, Antoine Italiano, Fernanda Mosele, Patricia Kannouche, Veronica Goldbarg, Barbara Pistilli, Elie Rassy Gustave Roussy; VHIO Background Patients with breast cancer and BRCA1/2 pathogenic variants have homologous recombination repair deficiency (HRD) which confers sensitivity to DNA-damaging agents. As topoisomerase I inhibitor (TOP1i)-based antibody-drug conjugates (ADC) induce replication-associated DNA damage, there is a strong biological rationale for their enhanced activity in BRCA-mutated tumors (Zander et al. 2010, Coussy et al. 2020). However, clinical outcomes reported with TOP1i-based ADCs in this population have been inconsistent (Dent et al. 2025, Bardia et al. 2021, Tarantino et al. 2025). These discordant observations raise the possibility that dynamic changes in BRCA biology, rather than baseline mutation status alone, may influence the efficacy of TOP1i-based ADC. Given that BRCA reversion mutations (BRCA-RM) restore homologous recombination repair, we hypothesized that BRCA-RMs may influence the efficacy of subsequent TOP1i-based ADCs in patients with metastatic breast cancer (mBC). Methods This retrospective study used individual-participant data from women with mBC treated at Gustave Roussy Cancer Center and having undergone a liquid biopsy for somatic molecular profiling within the cross-sectional STING platform (NCT04932525) between January 2021 and December 2025. Patients were eligible if they had mBC with pathogenic or likely pathogenic germline BRCA1 or BRCA2 pathogenic variants and had been treated with PARPi (olaparib or talazoparib), followed by a TOP1i-based ADC (sacituzumab govitecan or trastuzumab deruxtecan), either immediately or after intervening systemic therapies. These patients were included if they had a liquid biopsy performed at any point after PARPi progression and prior to ADC initiation. Liquid biopsies were analyzed using the FDA-approved FoundationOne Liquid CDx hybrid capture next-generation sequencing panel. BRCA-RMs were defined as secondary somatic alterations predicted to restore the open reading frame (ORF) of a deleterious germline BRCA1/2 mutation. The primary endpoint was PFS, defined as the time from initiation of ADC to radiological or clinical disease progression or death from any cause. Secondary endpoints included: (i) Overall survival (OS), defined as time from ADC initiation to death from any cause, and (ii) primary resistance, defined as PFS ≤ 3 months. Tumor response was assessed according to physician assessment. Cox proportional hazards models were used to estimate hazard ratios (HRs) and 95% CI. Multivariable models were adjusted for clinically relevant covariates, including ADC subtype, line of ADC, mBC subtype, and presence of visceral M1. Results A total of 35 patients with germline BRCA1/2-mutated mBC were included (Table 1). BRCA reversion mutations were detected in 9 patients (25.7%). After a median follow-up of 18 months, 30 patients (85.7%) have experienced disease progression and 22 had died. The median PFS was 5.53 months (95% CI 3.5-7.5), and median OS was 13.4 months (95% CI 10.73-NR). Table 1: Clinicopathological characteristics of study population (N° of patients) Age at diagnosis (years) | Median 40 | Range 27 – 80 Metastatic stage | Recurrence 29 (82.9%) | De novo 6 (17.1%) HER2 status | Low 15 (42.9%) | Negative 20 (57.1%) Breast cancer subtype | Triple negative 17 (48.6%) | HR-positive 18 (51.4%) PARP inhibitor target | gBRCA1 12 (34.3%) | gBRCA2 23 (65.7%) PARP inhibitor | Olaparib 26 (74.3%) | Talazoparib 9 (25.7%) PARP inhibitor line of treatment | 1: 8 (22.9%) | 2: 15 (42.9%) | ≥3: 12 (34.3%) Best response to PARP inhibitor | Complete response 4 (11.4%) | Partial response 17 (48.6%) | Stable disease 6 (17.1%) | Progressive disease 8 (22.9%) BRCA reversion mutation | Yes 9 (25.7%) | No 26 (74.3%) ADC | Sacituzumab govitecan 21 (60%) | Trastuzumab deruxtecan 14 (40%) ADC line of treatment | 2: 3 (8.6%) | 3: 10 (28.6%) | ≥4: 22 (62.9%) Number of lines between PARPi and ADC | Median 2 | Range 1 – 4 Best response to ADC | Complete response 4 (11.4%) | Partial response 14 (40.0%) | Stable disease 1 (2.9%) | Progressive disease 16 (45.7%) Figure 2: Survival Outcomes according to BRCA-RM status PFS: No BRCA-RM mPFS: 6.12 months (95% CI 3.83-11.6); Yes BRCA-RM mPFS: 2.60 months (95% CI 2.37-NA); p=0.015; HRuniv 2.85 (95% CI 1.19-6.82); HRmult 2.68 (95% CI 1.00-7.20) OS: No BRCA-RM mOS: 14.93 months (95% CI 10.73-NA); Yes BRCA-RM mOS: 7.70 months (95% CI 5.70-NA); p=0.46; HRuniv 1.44 (95% CI 0.54-[illegible]); HRmult 1.57 (95% CI 0.4[illegible]) Primary resistance to ADC therapy was significantly more frequent in patients with BRCA-RM [illegible] those without reversion events (77.8% vs 34.6%; Fisher's exact test p = 0.0498). Figure 1: Swimmer plot of patients with BRCA-RM Conclusion Our findings suggest that BRCA-RM detected in ctDNA after progression to PARPi [illegible] with primary resistance and shorter progression-free survival in patients trea[ted with TOP1i-] based ADC. Pending validation in larger cohorts, dynamic assessment of BRCA [illegible] using ctDNA could represent a potential biomarker to explore treatment sequenc[ing]. Contact: jesusyaringano@vhio.net; elie.rassy@gustaveroussy.fr

[Slide 1] 3484P GUSTAVE ROUSSY CANCER CAMPUS GRAND PARIS | PRISM Communicating Complexity in Precision Oncology: Qualitative Insights to Inform Patient-Facing Digital Twins P. Zyumbileva, E. Martin, S. Ball, S. Everhard, E. Mussault, S. Dogan, J. Vibert, M.A. Franzoi, I. Vaz-Luis 1 Cancer Survivorship Group, Université Paris-Saclay, UVSQ, Gustave Roussy, Inserm, CESP, Villejuif, France 2 Data and Partnerships Department, Unicancer, Paris, France 3 National PRecISion Medicine Center in Oncology, Gustave Roussy, Villejuif, France BACKGROUND Precision oncology increasingly relies on complex data, making treatment decisions harder for patients to understand. This creates communication gaps limiting shared decision-making (SDM). Patient-facing digital twins (PDTs) integrating individual health data may support communication, but their development requires understanding how complexity is communicated in practice. Objective: We conducted a qualitative, multi-source study with patients, caregivers and healthcare providers to better understand how complexity is communicated in precision oncology. METHODS [small text, partially legible] Using a design-thinking co-creation approach grounded in SDM frameworks and socio-technical theory, we conducted an exploratory qualitative phase integrating observations, interviews, and co-creation workshops to inform PDT development. Data included 31 observations of oncology consultations across five centres in France ... 10 in-depth interviews with participating healthcare professionals (HCPs), focus groups with 24 participants ... [illegible] OBSERVED CONSULTATIONS Oncologists observed: 10 in 5 centres across France Patients: 27 women (87%), 4 men (13%) Mean age: 63.8 years (range 43–81) Duration: ≈ 800 min (13h20) in total, Nov 2025 – Apr 2026; mean 25.8, median 20 min Consultations per centre, n (%) [bar chart; largest bar 13; others 5, 5, 5, 3] Cancer type, n (%) [bar chart; largest bar 18; others small - illegible] IN-DEPTH INTERVIEWS 10 oncologists, post-observation interviews Experience in oncology: < 1 to ≈ 30 years Semi-structured, Dec 2025 – Apr 2026 Duration: ≈ 350 min in total; median ≈ 34 min (range 15–52) Interview guide, 8 topics: 1 Background & experience; 2 Introducing precision medicine; 3 Level of detail & uncertainty; 4 Checking understanding; 5 Patient values & priorities; 6 Who leads the decision; 7 Handling emotions; 8 Missing tools & improvements FOCUS GROUPS 24 participants: 12 patients with diverse tumour types, 2 caregivers, 10 HCPs Groups: FG1 patients and caregivers; FG2 and FG3 HCPs Topics: communication of precision medicine; shared decision-making Duration: ≈ 73–93 min per group RESULTS: FIVE THEMES 1 Personalizing Precision Medicine through Considering the Patient Needs and Life Context 2 Translating the Complexity of Precision Oncology 3 Negotiating Shared Decision-Making in Clinical Encounters 4 Caregivers as Communication and Decision Mediators 5 Supporting Precision Medicine Communication Through Information Tools [sub-themes and quotes: small text, illegible] PATIENT-FACING DIGITAL TWIN DESIGN IMPLICATIONS Personalised information delivery Uncertainty communication Caregiver inclusion Tools facilitating patient-clinician dialogue CONCLUSION Communication is a critical but underexplored dimension of precision oncology. [remaining text small - partially illegible] [Slide 2] 4661P Prognostic value of tumor-infiltrating lymphocytes in residual disease (TIL-RD) after neoadjuvant chemo-immunotherapy (NAC-IO) for triple-negative breast cancer: a real-world, multicentre study T.H. Hoang, J. R. Dixon Douglas, T. Grinda, S. Michiels, G. Gessain, N. Joyon, Y.H. Park, Y.-L. Choi, E. Agostinetto, G. Arcuri, Y. Ozaki, N. Yamashita, T. Osako, C. Martin Abreu, D. Trapani, A. Concardi, C. Criscitiello, B. Pistilli, F. André, J. M. Ribeiro 1 Institut Gustave Roussy, Villejuif, France, 2 Samsung Medical Center (SMC), Seoul, Republic of Korea, 3 Institute Jules Bordet, Brussels, Belgium, 4 The Cancer Institute Hospital of Japanese Foundation for Cancer Research (JFCR), Koto-ku, Japan, 5 Hospital Universitario de Canarias, San Cristobal De La Laguna, Spain, 6 European Institute of Oncology (IEO), IRCCS, Milan, Italy, 7 IRCCS Humanitas Research Hospital, Rozzano, Italy BACKGROUND Residual disease (RD) after neoadjuvant chemo-immunotherapy (NAC-IO) remains an unmet need in TNBC. Approximately one-third of patients have RD, with a highly heterogeneous risk of relapse. Higher TIL-RD levels have been associated with better outcomes in TNBC after neoadjuvant chemotherapy. Their prognostic value beyond residual cancer burden (RCB) following NAC-IO remains unclear. OBJECTIVE To investigate the prognostic value of TIL-RD and event-free survival (EFS) in TNBC patients with residual disease after NAC-IO. METHODS Patients with stage I-III TNBC who had RD after NAC-IO at 6 international centers (Jan 2022 – June 2024; n=205) TILs were assessed on surgical specimens by one local breast pathologist at each centre, following International TILs Working Group recommendations. Excluded: TIL-RD not available (n = 24) 181 patients included in the final analysis • TIL-RD was analysed continuously and using a prespecified 30% cut-off. • Primary endpoint: event-free survival (EFS) from surgery. • Kaplan–Meier estimates and centre-stratified Cox models were used. Multivariable models were adjusted for residual burden ypT and ypN, as RCB was not available at all centres. RESULTS Table 1: Clinicopathological characteristics Characteristics | Overall n=181 | TIL-RD <30% n=147 | TIL-RD ≥30% n=34 | p-value Age, years; median (IQR) | 54 (44–63) | 55 (45–63) | 51 (42–55) | 0.050 Clinical stage, n (%) | | | | 0.623 Stage I | 4 (2.2%) | 4 (2.7%) | 0 | Stage II | 130 (71.8%) | 105 (71.4%) | 25 (73.5%) | Stage III | 47 (26.0%) | 38 (25.9%) | 9 (26.5%) | Germline mutation, n (%) | | | | 0.720 No mutation | 142 (89.3%) | 115 (89.1%) | 27 (90.0%) | gBRCA1/2 mutation | 14 (8.8%) | 12 (9.3%) | 2 (6.7%) | Other mutation | 3 (1.9%) | 2 (1.6%) | 1 (3.3%) | Unknown | 22 | 18 | 4 | Pathological T category, n (%) | | | | 0.025 ypT1 | 140 (77.3%) | 107 (72.8%) | 33 (97.1%) | ypT2 | 30 (16.6%) | 29 (19.7%) | 1 (2.9%) | ypT3 | 10 (5.5%) | 10 (6.8%) | 0 | ypT4 | 1 (0.6%) | 1 (0.7%) | 0 | Pathological N category, n (%) | | | | 0.452 ypN0 | 125 (69.1%) | 100 (68.0%) | 25 (73.5%) | ypN1 | 37 (20.4%) | 29 (19.7%) | 8 (23.5%) | ypN2 | 17 (9.4%) | 16 (10.9%) | 1 (2.9%) | ypN3 | 2 (1.1%) | 2 (1.4%) | 0 | Residual cancer burden, n (%) | | | | 0.004 RCB I | 48 (33.6%) | 32 (27.6%) | 16 (59.3%) | RCB II | 84 (58.7%) | 73 (62.9%) | 11 (40.7%) | RCB III | 11 (7.7%) | 11 (9.5%) | 0 | Unknown | 38 | 31 | 7 | Systemic adjuvant treatment, n (%) | | | | - None | 6 (3.3%) | 6 (4.1%) | 0 (0.0%) | Pembrolizumab + capecitabine | 54 (30.0%) | 50 (34.2%) | 4 (11.8%) | Pembrolizumab | 48 (26.7%) | 35 (24.0%) | 13 (38.2%) | Capecitabine | 53 (29.4%) | 41 (28.1%) | 12 (35.3%) | Pembrolizumab + olaparib | 4 (2.2%) | 2 (1.4%) | 2 (5.9%) | Olaparib | 4 (2.2%) | 4 (2.7%) | 0 (0.0%) | Other (e.g. Clinical Trial...) | 11 (6.1%) | 8 (5.5%) | 3 (8.8%) | Unknown | 1 | 1 | 0 | • High TIL-RD (≥30%) was associated with younger age (51 vs 55 years, p=0.05) and lower residual tumour burden (ypT1: 97% vs 73%, p=0.025; RCB I: 59% vs 28%, p=0.004). • Among 150 patients with paired samples, me[obscured] at baseline (IQR 5-30%) and 5% in RD (IQR 2–20%; Wilcox[obscured] • TIL-RD showed inter-centre heterogeneit[obscured] 0% to 15% across the six centres (Kruskal–Wallis[obscured] Figure 1. Changes in TILs fr[obscured] baseline to residual d[obscured] [Median follow-up / two-year results and Table 2 partially obscured by presenter] CONCLUSIONS • High TIL-RD was associated with improved EFS in univariable analysis. • Although not statistically significant after adjustment for ypT and ypN, an independent effect of TIL-RD is not excluded given the association between TIL-RD and residual tumour burden. • TIL-RD may stratify risk within RCB/yp stage groups and evaluation in an expanded cohort is ongoing. • Inter-centre heterogeneity in TIL-RD highlights the need to improve reproducibility of TIL assessment in the real world. REFERENCES 1. Pusztai L, Denkert C, O'Shaughnessy J, et al. Event-free survival by residual cancer burden with pembrolizumab in early-stage TNBC: exploratory analysis from KEYNOTE-522. Ann Oncol. 2024;35(5):429-436. 2. Luen SJ, Salgado R, Dieci MV, et al. Prognostic implications of residual disease tumor-infiltrating lymphocytes and residual cancer burden in triple-negative breast cancer patients after neoadjuvant chemotherapy. Annals of Oncology. 2019;30(2):236-242. 3. [partially obscured] ... Durvalumab in Combination With Neoadjuvant Chemotherapy in Early Triple-Negative Breast Cancer: Long-Term Analysis from the GeparNuevo Trial. J Clin Oncol. ... No conflicts of interest declared. Contact: thihoai.hoang@gustaveroussy.fr BEST POSTER

[Slide 1] Clinical Validation Pathology AI Digital MammaPrint Validation Results Clinical Outcome All Patients – BCSS | ER+HER2- Patients – BCSS [KM panel legends, HRs and number-at-risk rows illegible] • 372 RASTER patients H&E-stained Whole Slide Images (WSI) • 189 patient tumors were (51%) D-Low and 183 (49%) D-High • D-High patients 74% had received chemo, D-Low pts 24% had received chemo • 10-year BCSS was 98.1% for D-Low versus 86.7% for D-High (all patients) In multivariable analyses, AI Digital MammaPrint was the strongest predictor of BCSS HR 0.084 (95%CI 0.02-0.31), p<0.001, outperforming clinical factors [faint line illegible] Survival table Endpoint: Breast cancer-specific survival (BCSS), RASTER cohort, by AI Digital MammaPrint risk D-Low | 189 D-High | 183 Landmark rates: 10 years: D-Low 98.1% | D-High 86.7% [Slide 3] Build a Digital Assay for the 70-Gene Recurrence Signature Pathology AI Technology Outline – Step 1 Whole Slide Image by Scanner: High Definition (Pixels) Breast Cancer Core Biopsies FFPE H&E stained Whole Slide Image | Digital Feature Annotation and QC Whole Slide Image Preprocessing Module Preprocessing ensures that each slide contains sufficient invasive tumor cells with relevant tumor and microenvironment • Tissue Detection (Thresholding) • Patch Extraction (256x256 pixels) • Tissue Type Classification (TTC trained on 191 partially annotated slides) • An AI TTC algorithm that classifies each patch into seven tissue types: invasive, infiltrated-invasive, non-invasive, stroma, tumor-stroma, lymphocyte, and rest

[Slide 1] MADRID 2026 ESMO congress Large language model-assisted chemotherapy side effect management in solid tumor patients: the ALLIANCE randomized controlled trial Young-Joon Kang, MD, PhD Division of Breast Surgery, Department of Surgery, Incheon St. Mary's Hospital College of Medicine, The Catholic University of Korea, Republic of Korea LBA19 | Proffered Paper: AI and digital health | ESMO Congress 2026, Madrid, 24 Oct 2026 [Slide 3] Where the chain broke The chain we assumed Identify more toxicity ✓ Yes - [illegible] vs [illegible] events recorded Manage more of it ✗ Not shown - Dose reduction unchanged; advice adopted [illegible] in part Better quality of life ✗ No benefit shown - [illegible] Where it broke ✗ Used in fewer than half of the monthly reports, only when the clinician felt the need ✗ Where a routine existed, the routine was used; the LLM added where there was none On-request LLM support for the clinician was not useful in real-world practice. Young-Joon Kang

[Slide 1] PROMPT: prospective, matched tissue and plasma in 301 patients [Schema] (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 Cohort 301 matched tissue–plasma pairs Fresh biopsy n = 161 · archival n = 140 Matched leukocyte DNA n = 262 ctDNA tumour fraction median 5% (IQR 0–25%) Assays Tissue TSO500 (523 genes), tumour content > 30%; copy number based on relative coverage and z-score Plasma and leukocyte DNA 57-gene custom hybrid capture — 931 probes, 163 kb; 26 genes fully CDS-tiled, 21 hotspot/partial, 6 CNV backbone Median on-target depth ≈ 3,200× Analysis Tissue as comparator — PPA / NPA / PPV / NPV, Wilson 95% CI SNV / indel: VAF ≥ 1%, CHIP-subtracted on leukocyte DNA Biallelic loss: gene-median log2 < –0.70 (pre-specified) in tumor fraction ≥ 20% Niven Mehra [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) [Oncoprint; rows: ctDNA fraction %, biopsy, BRCA2, BRCA1, ATM, CDK12, CHEK2, PALB2, NBN, RAD51B, RAD51C, RAD51D, PIK3CA, PIK3R1, AKT1, PTEN tissue, PTEN ctDNA] Right-margin per-gene counts (concordant | tissue-only | ctDNA-only): 14|5|3; 0|1|0; 16|4|1 CH 4; 5|4|1 CH 1; 6|0|0 CH 4; 4|0|0; 0|0|1; 0|2|0; 1|2|0; 0|1|0; 3|3|1 CH 1; 1|7|2; 0|1|0; 57 bi | 47 sub-thr; 17 bi | 28 sub Legend: 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); tissue loss below the biallelic call; ctDNA clonal biallelic; ctDNA subclonal / single-copy; not evaluable (ctDNA fraction < 20%); excluded, coverage dropout [small footnote text partly illegible: 174 patients with ≥1 alteration in these genes, grouped by biopsy timepoint and sorted within group by ctDNA fraction (high → low)] In the 17 patients where biallelic PTEN loss was called in tissue - and ctDNA fraction was sufficient - 8 biallelic loss was concordant, Niven Mehra

Decision support in multiple myeloma: AI models compared with expert oncologists LARVOL ONCO DATA AI Mina Nikanjam¹*, Judith Pérez-Granado²*, Carolina Schinke³, Meera Mohan⁴, Sabarinath V. Radhakrishnan⁵, Donald A. Barkauskas⁶, Bruno Larvol², Anita D'Souza⁵, Razelle Kurzrock⁵,⁷ ¹UC San Diego; ²The LARVOL Group; ³University of Arkansas; ⁴MD Anderson Cancer Center; ⁵Medical College of Wisconsin; ⁶University of Southern California; ⁷WIN Consortium. *Equal contribution Corresponding: judith.perez@larvol.com INTRODUCTION [partly covered by hand] ...pared treatment recommendations ...our LLMs with those of expert myeloma ...ologists across nine clinical cases. METHODS Four experts and ChatGPT, Gemini, Claude and Grok assessed nine cases using a shared treatment list. Experts assigned 75/25, 50/50 or 100% preference weights. LLM weights totalled 100%. Pairwise Jensen–Shannon (JS) divergence measured agreement. RESULTS Pairwise JS divergence across nine cases Comparison | MD 1 | MD 2 | MD 3 | MD 4 | ChatGPT | Claude | Gemini | Grok MD 1 | — | 0.4727 | 0.3479 | 0.2495 | 0.3662 | 0.3998 | 0.6013 | 0.4054 MD 2 | 0.4727 | — | 0.3288 | 0.3847 | 0.3608 | 0.3260 | 0.3607 | 0.2362 MD 3 | 0.3479 | 0.3288 | — | 0.2324 | 0.3130 | 0.3870 | 0.4871 | 0.2923 MD 4 | 0.2495 | 0.3847 | 0.2324 | — | 0.2362 | 0.2966 | 0.4379 | 0.2792 ChatGPT | 0.3662 | 0.3608 | 0.3130 | 0.2362 | — | 0.0620 | 0.2074 | 0.1369 Claude | 0.3998 | 0.3260 | 0.3870 | 0.2966 | 0.0620 | — | 0.1771 | 0.1318 Gemini | 0.6013 | 0.3607 | 0.4871 | 0.4379 | 0.2074 | 0.1771 | — | 0.1802 Grok | 0.4054 | 0.2362 | 0.2923 | 0.2792 | 0.1369 | 0.1318 | 0.1802 | — Lower JS divergence indicates greater similarity. CONCLUSIONS Agreement with experts varied: highest for Grok, lowest for Gemini. LLM–physician divergence resembled physician–physician divergence, while LLMs agreed more closely with each other. Major comorbidities challenged recommendations. These pilot findings support further study, with validation required before clinical use. [Background poster, out of focus: LARVOL poster titled approximately 'A Multi-LLM Framework for Oncology Clinical Decision Support: A Breast Cancer Case Series' with results tiles '80%' and '90%' - remainder illegible]

[Slide 2] About Us Who is CPE? Cancer Patients Europe is a pan-European and pan-cancer type association. CPE's Mission Represent and empower the voices of cancer patients and cancer survivors towards policymakers and other stakeholders, with the ultimate objective of co-creating policies that drive prevention, high-quality cancer care, treatment, and survivorship in Europe. CPE's Vision To reduce the burden of cancer on patients and survivors, care givers, healthcare systems and the society as a whole. [Slide 3] Why Europe Needs CPE CPE is the only pan-European cancer patient organisation led entirely by cancer survivors and carers. Its membership provides a unique link between European-level and national cancer organisations, is inclusive of all cancer types and supports all aspects of the patient journey, including issues faced by their carers. As such, it can rightfully claim to be the true voice of Cancer Patients in Europe. [Slide 4] Future Projects in 2027 • HTA – Patient survey and focus group on JCA, alternative evidence to RCT, patient involvement in Eastern Europe • Prostate Cancer and Genomic Testing • Breast Cancer and Genomic Testing • AMR and Cancer • European Semester on Health with a Pilot in Cancer • Smoking Cessation and Cancer • PROs in the Middle East • Lancet Commission and other journals' initiatives [two group photos of attendees]

Overall survival Fig.1 Overall survival in overall patients | | Trastuzumab rezetecan | Physician's choice | Events, n (%) | 66 (23.8) | 77 (28.1) 12-month OS rate % (95% CI) | 90.6 (86.2-93.6) | 81.4 (75.9-85.8) Hazard ratio (95% CI) | 0.67 (0.48-0.94) Legend: Trastuzumab rezetecan; Physician's choice; Censor Y axis: Overall survival (%); X axis: Time (months) No. at Risk Trastuzumab rezetecan: 277 268 257 233 193 153 121 85 48 15 3 Physician's choice: 274 258 240 195 163 137 92 61 28 11 2 0 Fig.2 Overall survival in chemo-naive patients | | Trastuzumab rezetecan | Physician's choice | Events, n (%) | 43 (19.1) | 55 (24.6) 12-month OS rate % (95% CI) | 92.0 (87.3-95.1) | 83.0 (76.9-87.6) Hazard ratio (95% CI) | 0.62 (0.41-0.93) No. at Risk Trastuzumab rezetecan: 225 218 209 187 150 113 86 52 27 8 0 Physician's choice: 224 210 196 157 127 104 63 39 14 4 2 0 Overall survival data remained immature. Further analysis of overall survival, including an interim analysis at 274 OS events (maturity: ~52%) and a final analysis at 392 OS events (maturity: ~74%) are planned. Presenter: Ying Fan MD Survival table Endpoint: Overall survival (overall patients); chemo-naive subgroup also shown Trastuzumab rezetecan (overall) | 277 Physician's choice (overall) | 274 Trastuzumab rezetecan (chemo-naive) | 225 Physician's choice (chemo-naive) | 224 HR: 0.67 (overall); 0.62 (chemo-naive) HR CI: 0.48-0.94 (overall); 0.41-0.93 (chemo-naive) Landmark rates: 12 mo: Trastuzumab rezetecan (overall) 90.6% (CI 86.2-93.6) | Physician's choice (overall) 81.4% (CI 75.9-85.8) | Trastuzumab rezetecan (chemo-naive) 92.0% (CI 87.3-95.1) | Physician's choice (chemo-naive) 83.0% (CI 76.9-87.6)

[Slide 1] ET-FES STUDY ER+/HER2- MBC pts, Ist line therapy, N = 147 -> 18F-FES PET/CT + 18F-FDG-PET/CT -> 18F-FES SUV≥2 -> Endoc[rine therapy] (cropped) -> 18F-FES SUV<2 -> R 1:1 -> Endoc[rine therapy] (cropped) | CT or ET PRIMARY OBJECTIVE To compare the efficacy of first-line ET vs CT in patients with ER+/HER2- MBC and 18F-FES SUV<2 at baseline CT/PET scan. SECONDARY OBJECTIVES • to evaluate DCR/PFS/OS with ET in patients with 18F-FES SUV≥2; • to compare the DCR/PFS/OS with ET in patients with 18F-FES SUV≥2 with 18F-FES SUV <2; • to correlate ER expression in the primary tumor and overall 18F-FES-uptake in metastases; • to assess PFS and OS in all patients and by 18F-FES SUV value. ET-FES HETEROGEN[EITY] (cropped) GROUP A | 100% Concorda[nt] (cropped) GROUP B | Partially conco[rdant] (cropped) GROUP C | 100% Discorda[nt] (cropped) Gennari A. [Slide 3] PROGRESSION-FREE AND OVERALL SURVIVAL [Left: Kaplan-Meier PFS, y-axis Progression-free survival (%), x-axis Years 0-11] Median Progression-free survival Group A: 18.1 months (95%CI 11.3-25.3) Group B/C: 12.0 months (95%CI 6.6-22.9) HR 1.36 (95%CI 0.88-2.10); p=0.162 No. at Risk (No. Cumulative Censors) Group A | 92 (0) | 53 (0) | 39 (0) | 29 (0) | 25 (0) | 21 (1) | 11 (6) | 9 (7) | 7 (9) | 4 (12) | 2 (14) | 0 (16) Group B+C | 31 (0) | 16 (0) | 9 (0) | 7 (0) | 6 (0) | 3 (1) | 2 (2) | 1 (2) | 0 (3) [Right: Kaplan-Meier OS, y-axis Overall survival (%), cropped at right edge] — Gr[oup A] 69.6 (95% CI ... [cropped] HR = 1[cropped] No. at Risk (No. Cumulative Censors) Group A | 92 (0) | 90 (0) | 80 (1) | 70 (1) | 57 (2) | 48 (5) | 33 [cropped] Group B+C | 32 (0) | 28 (1) | 23 (1) | 17 (1) | 14 (1) | 11 (2) | 8 [cropped] Gennari A. Survival table Endpoint: PFS (Group A 100% FES/FDG concordant vs Group B/C partially/fully discordant) Arm | N | Median | Median CI Group A | 92 | 18.1 months | 11.3-25.3 Group B/C | 31 | 12.0 months | 6.6-22.9 HR: 1.36 HR CI: 0.88-2.10 P: 0.162 [Slide 4] Antitumor activity of T-DXd across the whole HE[R2] expression spectrum in Advanced Breast Cance[r] (cropped) The DAISY study Cohort 1: HER2-overexpressing (n = 68) Confirmed ORR: 70.6% (95% CI: 58.3–81.0) [Waterfall plot: Best change from baseline in target lesions (%); legend Progressive disease, Stable disease, Partial response, Complete response, HR-positive, HR-negative] Cohort 2: HER2-low (n = 72) Confirmed ORR: 37.5% (95% CI: 26.4–49.7) [Waterfall plot] Cohort 3: H[ER2-zero] (cropped) — Confirmed [ORR ...] (cropped) 25% of patients show little benefit while being exposed to potentially serious toxicities >50% of patients show limited benefit while being exposed to potentially serious toxicities Mosele et al. Nat Med 2023

2458P Clinical validation of a novel non-invasive liquid biopsy assay (STELLA) for the differential diagnosis of benign and malignant breast lesions Y. Yin1, H. Wang2, X. Huang1, J. Wang1, X. Zha1, Y. Mao2, C. Liu2, B. Yang3, L. Wang3, P. Liu3, C. Zheng3, S. Wu3, Z. Zhang3, K. Zhao3, Y. Zhang3, G. Wang3, C. Li3, Y. Shi3 1Department of Oncology, Jiangsu Province Hospital/The First Affiliated Hospital of Nanjing Medical University, Nanjing, China, 2Breast Disease center, The Affiliated Hospital of Qingdao University, Qingdao, China, 3Burning Rock Biotech, Shanghai, China BACKGROUND Accurate differentiation between benign and malignant breast nodules remains critical for personalized clinical management. STELLA is a non-invasive, tumor-agnostic DNA methylation-based assay (methylation detected via the SPIRAL platform) that enables accurate quantification of ctDNA levels for multiple cancers. Here, we validate the clinical performance of STELLA in differentiating malignant from benign breast lesions. METHODS The STELLA panel and algorithm were adopted from previous studies.1 Blood samples were prospectively collected from patients with malignant and benign breast lesions prior to any anti-tumor therapy (PERCEIVE-BREAST, NCT06979921). All samples underwent cfDNA methylation sequencing to generate methylation signatures via STELLA. Sensitivity and specificity were assessed overall and stratified by clinical stage and molecular subtype. We further assessed the performance of STELLA in BI-RADS category 4 breast lesions. Table 1. Baseline characteristics of all patients included in the study. Characteristic | Benign (n=106) | Cancer (n=410) Age Median (IQR) | 41.5 (32, 50) | 51.0 (43, 58) Sex Female | 106 (100%) | 410 (100%) Pathology Type (Cancer only): IDC 372 (90.7%); ILC 11 (2.7%); Others 27 (6.6%) Stage (Cancer only): 0 (DCIS) 15 (3.7%); I 113 (27.6%); II 190 (46.3%); III 75 (18.3%); IV 17 (4.1%) Molecular Subtype (Cancer only): HR+HER2- 225 (54.9%); HR+HER2+ 89 (21.7%); HR-HER2+ 44 (10.7%); HR-HER2- 43 (10.5%); Unknown 9 (2.2%) RESULTS A total of 516 patients were analyzed, including 395 with invasive breast cancer, 15 with in situ carcinomas and 106 with benign breast lesions (Table 1). The sensitivity was 65.9% (270/410) in all breast cancer at a specificity of 97.2% (103/106). Sensitivity increased with disease stage, showing 32.7% (37/113) for Stage I, 77.4% (147/190) for Stage II, 85.3% (64/75) for Stage III, and 100.0% (17/17) for Stage IV. In addition, STELLA detected 5 out of 15 patients with in situ carcinomas (33.3%) (Figure 1). Fig 1. Sensitivity (overall and by AJCC stage) and specificity of STELLA in differentiating malignant from benign breast lesions. Benign 97.2% 103/106 | Cancer 65.9% 270/410 | 0 33.3% 5/15 | I 32.7% 37/113 | II 77.4% 147/190 | III 85.3% 64/75 | IV 100% 17/17 STELLA showed the highest sensitivity in the HR+/HER2+ subtype, at 89.9% (80/89), followed by HR−/HER2− at 81.4% (35/43), HR−/HER2+ at 79.5% (35/44), and HR+/HER2− at 51.6% (116/225) (Figure 2). Among patients with BI-RADS category 4 findings on ultrasound, mammography, or MRI, the sensitivity of STELLA was 55.5%, 59.7%, and 63.3%, respectively, with corresponding specificities of 98.3%, 100%, and 95.0% (Table 2). Fig 2. The sensitivity of STELLA by molecular subtype and AJCC stage. Table 2. Performance of STELLA in the BI-RADS category 4 patients. BI-RADS 4 | Total (n) | Cancer (n) | Benign (n) | Sensitivity (95%CI) | Specificity (95%CI) Ultrasound | 242 | 182 | 60 | 55.5% (48.2-62.6) | 98.3% (91.1-99.7) Mammography | 157 | 129 | 28 | 59.7% (51.1-67.7) | 100.0% (87.7-100.0) MRI | 69 | 49 | 20 | 63.3% (49.3-75.4) | 95.0% (76.3-99.1) CONCLUSION These findings underscore the potential of STELLA, a DNA methylation-based, tumor-agnostic cfDNA method, for the differential diagnosis of benign and malignant breast lesions. Clinically, the high specificity of STELLA may support its use as an adjunct to imaging for identifying a subset of BI-RADS 4 patients with a high probability of malignancy. Future studies will explore the prognostic value of STELLA in patients with breast cancer. REFERENCE 1. Wu S, et al. Integrated genetic and epigenetic profiling in the same DNA molecules. AMP 2025 Annual Meeting. Poster TT057. DISCLOSURE STATEMENT The first and presenting author has no conflicts of interest to declare. CONTACT Corresponding author: Yongmei Yin, ymyin@njmu.edu.cn.