ESMO 2026 Melanoma Slides

[Slide 1] Adverse Event Summary | All-Cause AEs: Intismeran + Pembrolizumab (N = 755) | All-Cause AEs: Placebo + Pembrolizumab (N = 380) | Treatment-Related AEs: Intismeran + Pembrolizumab (N = 755) | Treatment-Related AEs: Placebo + Pembrolizumab (N = 380) Any grade | 742 (98.3) | 366 (96.3) | 723 (95.8) | 342 (90.0) Grade 3-5 | 178 (23.6) | 77 (20.3) | 107 (14.2) | 51 (13.4) Serious | 133 (17.6) | 60 (15.8) | 74 (9.8) | 42 (11.1) Led to discontinuation of any drug | 125 (16.6) | 36 (9.5) | 106 (14.0) | 33 (8.7) Discontinuation of pembrolizumab | 115 (15.2) | 36 (9.5) | 97 (12.8) | 33 (8.7) Discontinuation of intismeran or placebo | 66 (8.7) | 21 (5.5) | 54 (7.2) | 19 (5.0) Discontinuation of both drugs simultaneously | 49 (6.5) | 20 (5.3) | 39 (5.2) | 18 (4.7) Led to death (none related to intismeran) | 8 (1.1) | 0 | 2 (0.3) | 0 aOf the 5 events, 2 were attributed to pembrolizumab (muscular weakness, myocarditis, and myositis in 1 patient and myasthenia gravis in 1 patient); the remaining deaths were due to cardiac failure (n=1), road traffic accident (n=1), and sudden death (n=1). bOf the 3 events, none were attributed to study treatment (acute myocardial infarction (n=1), aspiration pneumonia (n=1), and sudden cardiac death (n=1). As-treated population: all randomized patients who received at ≥1 dose of study treatment. Data cutoff date: June 30, 2026 (median study follow-up, 24.6 mo). Presented by: Georgina V. Long, Melanoma Institute Australia

[Slide 1] Recurrence-Free Survival, ITT Population HR 0.58 (95% CI 0.45–0.75) P < 0.001 (two sided)a 24-month landmark (dashed line): 82.2% (95% CI 79.3–84.8) [intismeran arm, teal]; 70.0% (95% CI 65.0–74.5) [placebo arm, blue] N | Events, n (%) Intismeran + pembrolizumab | 757 | 136 (18.0%) Placebo + pembrolizumab | 380 | 111 (29.2%) No. at risk (months 0/6/12/18/24/30/35): Intismeran + pembrolizumab: 757, 683, 646, 618, 136, 10, 0 Placebo + pembrolizumab: 380, 341, 302, 282, 72, 3, 0 aMultiplicity-adjusted significance level: two-sided 0.024692. ITT: all randomized patients according to the arm to which they were assigned. RFS: time from randomization to the first occurrence of any recurrence or death from any cause. Data cutoff date: June 30, 2026 (median study follow-up, 24.6 mo). Presented by: Georgina V. Long, Melanoma Institute Australia [ESMO Congress Madrid 2026 logo] [Survival data captured as a table — no point estimates read from the curve] Endpoint: Recurrence-free survival (RFS), ITT population Intismeran + pembrolizumab | 757 Placebo + pembrolizumab | 380 HR: 0.58 HR CI: 0.45–0.75 P: <0.001 (two sided); multiplicity-adjusted significance level two-sided 0.024692 {'timepoint': '24 months', 'arm': 'Intismeran + pembrolizumab', 'value': '82.2%', 'ci': '79.3–84.8'} {'timepoint': '24 months', 'arm': 'Placebo + pembrolizumab', 'value': '70.0%', 'ci': '65.0–74.5'} Note: Data cutoff June 30, 2026; median study follow-up 24.6 mo. Presented by Georgina V. Long. Medians not printed.

[Slide 1] LBA1 - INTerpath-001: Phase 3 study of the mRNA-based individualized neoantigen therapy (INT) intismeran autogene (intismeran) plus pembrolizumab (pembro) vs pembro alone for adjuvant treatment of resected stage IIB-IV melanoma (MEL) Speakers Georgina V. Long (Sydney, Australia) Lecture Time 16:30 - 16:42 Abstract Background Many patients (pts) with resected high-risk MEL experience recurrence despite adjuvant therapy. Enhancing immune activation may improve efficacy. Intismeran is an mRNA-based INT designed to expand pre-existing and generate de novo neoantigen-specific antitumor T-cell responses by targeting pt-specific tumor-derived antigens. INTerpath-001 (NCT05933577) is a double-blind phase 3 study of intismeran + pembro as adjuvant therapy for resected stage IIB–IV MEL. Methods Eligible adults with surgically resected stage IIB/IIC (pathologic or clinical) or III/IV cutaneous MEL and a tumor sample suitable for next-generation sequencing were randomized 2:1 to pembro 400 mg (9 cycles, IV Q6W) + intismeran 1 mg or placebo (9 doses, IM Q3W). Pembro started at randomization, intismeran or placebo upon availability. Primary endpoint: RFS. Secondary endpoints: DMFS, OS, and safety. Results 1137 pts were randomized: 757 to intismeran + pembro, 380 to placebo + pembro; 726 (95.9%) received ≥1 dose of intismeran and 365 (96.1%) received ≥1 dose of placebo as planned. Median follow-up as of 30Jun2026 was 24.6 mo. Intismeran + pembro significantly improved RFS and DMFS compared with placebo + pembro (Table). Benefit was similar in stage IIB–IIIB and IIIC–IV MEL and other subgroups. For intismeran + pembro vs placebo + pembro, AEs were grade ≥3 in 23.6% vs 20.3%; 6.5% vs 5.3% discontinued both drugs simultaneously due to an AE. Most common AEs were pyrexia (52.2% vs 8.2%), injection site pain (51.0% vs 9.5%), and fatigue (38.5% vs 29.7%). Table: LBA1 | Intismeran + pembro (N = 757) | Placebo + pembro (N = 380) Recurrence-free survival (RFS) 24-mo rate, % (95% CI) | 82.2 (79.3–84.8) | 70.0 (65.0–74.5) HR (95% CI) | 0.58 (0.45–0.75) | — Two-sided P | <0.001 | — Distant metastasis-free survival (DMFS) 24-mo rate, % (95% CI) | 88.5 (86.0–90.6) | 80.2 (75.6–84.0) HR (95% CI) | 0.62 (0.45–0.85) | — Two-sided P | 0.002 | — Conclusions Intismeran + pembro significantly improved RFS and DMFS compared with placebo + pembro and was generally well tolerated when given as adjuvant therapy for stage IIB–IV MEL. These findings set a new benchmark for adjuvant MEL and are a pivotal advance in personalized medicine and cancer treatment.

[Slide 1] Intismeran mechanism of action • Intismeran is designed to strengthen existing and create new T cell responses1-3 • Pembrolizumab blocks PD-1 to enhance T-cell activity, complementing intismeran's MOA 1 Administer INT 2 mRNA uptake into cells 3 mRNA translated into protein and processed 4 Neoantigen presentation 5 T cell activation and clonal expansion 6 Migration to tumor microenvironment 7 Tumor cell killing 8 Epitopes releaseda Diagram labels: APC (eg, dendritic cell); [illegible small diagram labels]; T cells; Dying tumor cell; Cytokine release [footnotes and references illegible] Presented by: Georgina V. Long, Melanoma Institute Australia [blurred] [Slide 2] Intismeran Design, Manufacture,a and Exposure • 91% of the 1137 randomized patients had the maximum 34 neoantigens identified for intismeran design • >99% of the 757 patients in the intismeran + pembrolizumab arm had their intismeran successfully manufactured • Median time to first intismeran dose was 64 days • Median number of intismeran doses was 9b No. neoantigens in each intismeran Median: 34 (range, 7–34) Pie chart legend: 34 | 30-33 | 25-29 | 20-24 | 15-19 | 10-14 | <10 [footnotes a/b largely illegible: aIntismeran was designed for all randomized patients. The manufacturing workflow was initiated when it was known that a patient was randomized to the intismeran + pembrolizumab arm. bThe median number of placebo administrations and pembrolizumab cycles was also [?]] Presented by: Georgina V. Long, Melanoma Institute Australia

[Slide 1] Intismeran mechanism of action • Intismeran is designed to strengthen existing and create new T cell responses1-3 1 Administer INT 2 mRNA uptake into cells APC (eg, dendritic cell) 3 mRNA translated into protein and processed Endosome | Endoplasmic reticulum | INT mRNA | Polypeptide | Proteasome | Tumor-specific neoantigens | MHC I | MHC II | Endo-lysosomes CD8 T cell (cytotoxic activity) CD4 T cell (memory) 4 Neoantigen presentation 5 T cell activation and clonal expansion T cells 6 Migration to tumor microenvironment 7 Tumor cell killing CD4+ T cell | CD8+ T cell | Cytokine release | Dying tumor cell 8 Epitopes releaseda aHypothetical mechanism of action. 1Weber JS et al. Lancet 2024;403:632–44. 2Ojalvo LS et al. JCO Oncol Adv 2026;3:e2500206. 3Gainor JF, et al. Cancer Discov. 2024;14(11):2209-2223 Presented by: Georgina V. Long, Melanoma Institute Australia [Slide 3] All-Cause AEs, Incidence ≥10% Intismeran + Pembrolizumab (N = 755) | Placebo + Pembrolizumab (N = 380) AE | Intismeran + pembro (%) | Placebo + pembro (%) Pyrexia | 52.2 | 8.2 Injection site pain | 51.0 | 9.5 Fatigue | 38.5 | 29.7 Influenza-like illness | 25.4 | 9.2 Diarrhea | 23.7 | 23.7 Pruritus | 23.4 | 23.4 Headache | 23.3 | 13.2 Arthralgia | 22.6 | 23.2 Injection site reaction | 21.6 | 1.6 Nausea | 21.2 | 15.3 Myalgia | 20.3 | 10.5 Asthenia | 19.9 | 18.2 Injection site erythema | 19.9 | 0.8 Chills | 16.2 | 3.4 Hypothyroidism | 15.8 | 15.5 Rash | 15.1 | 17.6 Cough | 10.9 | 10.3 Dry mouth | 9.7 | 11.3 ~10% greater incidence vs pembrolizumab alone due to intismeran injection Low grade and transient Managed per standard practice Legend: Grade 1–2 | Grade 3–5 Incidence, % As-treated population: all randomized patients who received at ≥1 dose of study treatment. Data cutoff date: June 30, 2026 (median study follow-up, 24.6 mo). Presented by: Georgina V. Long, Melanoma Institute Australia

[Slide 1] INTerpath-001: Phase 3 Study of the mRNA-Based Individualized Neoantigen Therapy (INT) Intismeran Autogene plus Pembrolizumab vs Pembrolizumab Alone for Adjuvant Treatment of Resected Stage IIB-IV Melanoma Georgina V. Long; Muhammad A. Khattak; Helen J. Gogas; Dirk Schadendorf; Ana Arance; Matteo S. Carlino; Piotr Rutkowski; Gareth Rivalland; Antoni Ribas; Janice M. Mehnert; Patrick A. Ott; Reinhard Dummer; Paola Queirolo; Christoffer Gebhardt; Laureen S. Ojalvo; Jinchun Zhang; Dmitri O. Grebennik; Mizuho Fukunaga-Kalabis; Paolo A. Ascierto; Michele Del Vecchio on behalf of the INTerpath-001 Investigators MADRID 2026 ESMO congress

[Slide 1] Results 1137 pts were randomized: 757 to intismeran + pembro, 380 to placebo + pembro; 726 (95.9%) received ≥1 dose of intismeran and 365 (96.1%) received ≥1 dose of placebo as planned. Median follow-up as of 30Jun2026 was 24.6 mo. Intismeran + pembro significantly improved RFS and DMFS compared with placebo + pembro (Table). Benefit was similar in stage IIB–IIIB and IIIC–IV MEL and other subgroups. For intismeran + pembro vs placebo + pembro, AEs were grade ≥3 in 23.6% vs 20.3%; 6.5% vs 5.3% discontinued both drugs simultaneously due to an AE. Most common AEs were pyrexia (52.2% vs 8.2%), injection site pain (51.0% vs 9.5%), and fatigue (38.5% vs 29.7%). Table: LBA1 | Intismeran + pembro (N = 757) | Placebo + pembro (N = 380) Recurrence-free survival (RFS) 24-mo rate, % (95% CI) | 82.2 (79.3–84.8) | 70.0 (65.0–74.5) HR (95% CI) | 0.58 (0.45–0.75) | — Two-sided P | <0.001 | — Distant metastasis-free survival (DMFS) 24-mo rate, % (95% CI) | 88.5 (86.0–90.6) | 80.2 (75.6–84.0) HR (95% CI) | 0.62 (0.45–0.85) | — Two-sided P | 0.002 | —

[Slide 2] DMFS Primary Analysis: ITT 88.5% (95% CI 86.0–90.6) 80.2% (95% CI 75.6–84.0) HR 0.62 (95% CI 0.45–0.85) P = 0.002 (two sided)a | N | Events, n (%) Intismeran + pembrolizumab | 757 | 89 (11.8%) Placebo + pembrolizumab | 380 | 71 (18.7%) Months No. at risk 757 | 708 | 682 | 658 | 149 | 11 | 0 380 | 357 | 342 | 323 | 83 | 3 | 0 aMultiplicity-adjusted significance level: two-sided 0.01608. ITT: all randomized patients according to the arm to which they were assigned. DMFS: time from randomization to first occurrence of distant metastasis or death from any cause. Data cutoff date: June 30, 2026 (median study follow-up, 24.6 mo). Presented by: Georgina V. Long, Melanoma Institute Australia [Survival data captured as a table — no point estimates read from the curve] Endpoint: DMFS (ITT, primary analysis) Intismeran + pembrolizumab | 757 Placebo + pembrolizumab | 380 HR: 0.62 HR CI: 0.45–0.85 P: 0.002 (two sided) {'timepoint': '24 months', 'arm': 'Intismeran + pembrolizumab', 'value': '88.5%', 'ci': '86.0–90.6'} {'timepoint': '24 months', 'arm': 'Placebo + pembrolizumab', 'value': '80.2%', 'ci': '75.6–84.0'} Note: Landmark at 24-month dashed line. Data cutoff June 30, 2026. [Slide 3] Efficacy in Prespecified Subgroups RFS and DMFS benefits observed in most subgroups Recurrence-Free Survival Subgroup | Intismeran + Pembro Events/Patients | Placebo + Pembro Events/Patients | Hazard Ratio (95% CI) All patients | 136/757 | 111/380 | 0.58 (0.45–0.75) Age <65 yr | 79/458 | 61/230 | 0.65 (0.47–0.91) Age ≥65 yr | 57/299 | 50/150 | 0.53 (0.36–0.77) Sex Female | 57/301 | 40/153 | 0.71 (0.47–1.06) Sex Male | 79/456 | 71/227 | 0.54 (0.39–0.74) Australia | 18/89 | 12/39 | 0.54 (0.26–1.15) Europe | 70/415 | 54/209 | 0.64 (0.45–0.92) Latin America | 15/45 | 9/22 | 0.85 (0.37–1.95) United States | 12/84 | 10/35 | 0.50 (0.22–1.16) Rest of world | 23/124 | 26/75 | 0.50 (0.29–0.89) ECOG PS 0 | 120/697 | 97/350 | 0.61 (0.47–0.80) ECOG PS 1 | 16/60 | 14/30 | 0.48 (0.24–0.99) Risk-based disease stage IIB/IIC/IIIA/IIIB | 59/440 | 50/230 | 0.60 (0.41–0.88) Risk-based disease stage IIIC/IIID/IV | 77/316 | 61/150 | 0.57 (0.41–0.79) Distant Metastasis-Free Survival All patients | 89/757 | 71/380 | 0.62 (0.45–0.85) Age <65 yr | 47/458 | 37/230 | 0.65 (0.42–0.99) Age ≥65 yr | 42/299 | 34/150 | 0.61 (0.39–0.95) Sex Female | 31/301 | 27/153 | 0.58 (0.34–0.97) Sex Male | 58/456 | 44/227 | 0.66 (0.44–0.97) Australia | 12/89 | 11/39 | 0.42 (0.19–0.96) Europe | 42/415 | 34/209 | 0.63 (0.40–0.99) Latin America | 11/45 | 9/22 | 0.62 (0.26–1.49) United States | 8/84 | 2/35 | 1.77 (0.38–8.35) Rest of world | 16/124 | 15/75 | 0.63 (0.31–1.28) ECOG PS 0 | 77/697 | 60/350 | 0.65 (0.46–0.91) ECOG PS 1 | 12/60 | 11/30 | 0.47 (0.21–1.06) Risk-based disease stage IIB/IIC/IIIA/IIIB | 41/440 | 33/230 | 0.65 (0.41–1.02) Risk-based disease stage IIIC/IIID/IV | 48/316 | 38/150 | 0.59 (0.39–0.91) Intismeran + Pembro Better | Placebo + Pembro Better ITT: all randomized patients according to the arm to which they were assigned. Data cutoff date: June 30, 2026 (median study follow-up, 24.6 mo). Presented by: Georgina V. Long, Melanoma Institute Australia

[Slide 3] INTerpath-001: Randomized, Double-Blind, Phase 3 Trial Key Eligibility Criteria • Age ≥18 years • Resected cutaneous melanoma • AJCC 8th ed. Stage IIB/IIC (clinical IIB/IIC permitted), III, or IV • FFPE tumor sample suitable for NGS and intismeran generation • No prior systemic therapy • ECOG PS 0 or 1 R 2:1 Intismeran 1 mg IM Q3W for 9 doses + Pembrolizumab 400 mg IV Q6W for 9 cycles Placeboa IM Q3W for 9 doses + Pembrolizumab 400 mg IV Q6W for 9 cycles Stratification factors • Risk-based stage (IIB/IIC/IIIA/IIIB vs IIIC/IIID/IV) • Age (<65 years vs ≥65 years) aNormal saline or dextrose. Placebo was started at random to maintain the blind. ClinicalTrials.gov identifier: NCT05933577. Presented by: Georgina V. Long, Melanoma Institute Australia

[Slide 1] Science enables individualized therapeutic cancer vaccines Neoantigen discovery • Tumor-specific mutations • Potential for enhanced immunogenicity Cancer Cell: Patient A | Patient B | Patient C Neoantigen selection • NGS identifies tumor mutations • Computational prediction of HLA presentation and ranking The mRNA platform • Multiple antigens per mRNA • Robust CD8+ and CD4+ T cell responses • Rapid, individualized manufacturing Lipid nanoparticle (LNP) | mRNA HLA, human leukocyte antigen; mRNA, messenger RNA; NGS, next-generation sequencing. Ascierto PA, Melero I. J Immunother Cancer 2025;13:e013766. Zaidi N et al. Nat Rev Cancer 2025;25:517–33. Haanen JB et al. Nat Med 2026;32:828-40. Weber JS et al. Lancet 2024;403:632–44. Ojalvo LS et al. JCO Oncol Adv 2026;3:e2500206. Presented by: Georgina V. Long, Melanoma Institute Australia [Slide 2] Standardized intismeran workflow 1 Tissue and blood samples 2 Neoantigen identification 3 mRNA design 4 Manufacture 5 Administer ~6-8 weeks Presented by: Georgina V. Long, Melanoma Institute Australia

[Slide 3] Baseline Characteristics | Intismeran + pembrolizumab (N = 757) | Placebo + pembrolizumab (N = 380) Age, median (range) | 61.0 y (18–89) | 61.5 y (18–88) Male sex, n (%) | 456 (60.2) | 227 (59.7%) Race, n (%) Asian | 15 (2.0) | 13 (3.4) Multiple | 5 (0.7) | 4 (1.1) White | 696 (91.9) | 337 (88.7) Other | 10 (1.3)a | 4 (1.1)b Missing | 31 (4.1) | 22 (5.8) Geographic region, n (%) Australia | 89 (11.8) | 39 (10.3) Europe | 415 (54.8) | 209 (55.0) Latin America | 45 (5.9) | 22 (5.8) United States | 84 (11.1) | 35 (9.2) Rest of world | 124 (16.4) | 75 (19.7) ECOG PS 0, n (%) | 697 (92.1) | 350 (92.1) Risk-based disease stage, n (%) IIB/IIC/IIIA/IIIB | 441 (58.3) | 230 (60.5) IIIC/IIID/IV | 316 (41.7) | 150 (39.5) Overall disease stage, n (%) IIA | 1 (0.1)c | 0 IIB | 155 (20.5) | 73 (19.2) IIC | 92 (12.2) | 42 (11.1) IIIA | 50 (6.6) | 32 (8.4) IIIB | 143 (18.9) | 83 (21.8) IIIC | 245 (32.4) | 118 (31.1) IIID | 10 (1.3) | 4 (1.1) IV | 61 (8.1) | 28 (7.4) aIncludes American Indian or Alaska Native (n = 5), Black or African American (n = 3), and Native Hawaiian or other Pacific Islander (n = 2). bIncludes American Indian or Alaska Native (n = 3) and Black or African American (n = 1). cProtocol deviation. ITT population: all randomized patients according to the arm to which they were assigned. Data cutoff date: June 30, 2026 (median study follow-up, 24.6 mo). Presented by: Georgina V. Long, Melanoma Institute Australia

[Slide 3] Summary and Conclusions • INTerpath-001 met both the primary RFS and key secondary DMFS endpoints – Intismeran + pembrolizumab reduced the risk of recurrence or death by 42% and the risk of distant metastasis or death by 38% compared with placebo + pembrolizumab – In a sensitivity analysis, intismeran + pembrolizumab reduced the risk of recurrence or death by 58% after month 6 – Benefit was observed in most prespecified subgroups, including risk-based disease stages – OS was immature and did not meet prespecified significance boundarya; follow-up is ongoing aThe hazard ratio for OS was 0.77 (95% CI, 0.45–1.31); two-sided P = 0.33. Presented by: Georgina V. Long, Melanoma Institute Australia

[Slide 1] INTerpath-001: Randomized, Double-Blind, Phase 3 Trial Key Eligibility Criteria • Age ≥18 years • Resected cutaneous melanoma • AJCC 8th ed. Stage IIB/IIC (clinical IIB/IIC permitted), III, or IV • FFPE tumor sample suitable for NGS and intismeran generation • No prior systemic therapy • ECOG PS 0 or 1 R 2:1 Intismeran 1 mg IM Q3W for 9 doses + Pembrolizumab 400 mg IV Q6W for 9 cycles Placebo(a) IM Q3W for 9 doses + Pembrolizumab 400 mg IV Q6W for 9 cycles Pembrolizumab • Started at randomization, ≤13 weeks after resection Intismeran • Started upon availability to site • Earliest: Cycle 1, day 22 • Latest: Cycle 4, day 1 Stratification factors • Risk-based stage (IIB/IIC/IIIA/IIIB vs IIIC/IIID/IV) • Age (<65 years vs ≥65 years) Primary end point: RFS per investigator Secondary end points: DMFS per investigator, OS, patient-reported outcomes, safety (a)Normal saline or dextrose. Placebo was started at random to maintain the blind. ClinicalTrials.gov identifier: NCT05933577. Presented by: Georgina V. Long, Melanoma Institute Australia

[Slide 1] RFS Sensitivity Analysis: Piecewise Hazard Ratios, ITT 0-6 months: HR 0.99 (95% CI 0.66–1.49) >6 months: HR 0.42 (95% CI 0.30–0.58) | 0-6 months Events (ratea) | >6 months Events (ratea) Intismeran + pembrolizumab | 68 (0.016) | 68 (0.006) Placebo + pembrolizumab | 35 (0.016) | 76 (0.015) Months: 0 | 6 | 12 | 18 | 24 | 30 | 35 No. at risk 757 | 683 | 646 | 618 | 136 | 10 | 0 380 | 341 | 302 | 282 | 72 | 3 | 0 aRate calculated as no. of events divided by person-time in each interval. ITT: all randomized patients according to the arm to which they were assigned. RFS: time from randomization to the first occurrence of any recurrence or death from any cause. Data cutoff date: June 30, 2026 (median study follow-up, 24.6 mo). Presented by: Georgina V. Long, Melanoma Institute Australia [Survival data captured as a table — no point estimates read from the curve] Endpoint: RFS sensitivity analysis — piecewise hazard ratios (ITT) Intismeran + pembrolizumab | 757 Placebo + pembrolizumab | 380 HR: 0-6 months: 0.99; >6 months: 0.42 HR CI: 0-6 months: 0.66–1.49; >6 months: 0.30–0.58 Note: Piecewise HRs split at 6 months (dashed line). No p-values, medians or landmark rates printed. Arm N from numbers-at-risk row at time 0. Data cutoff June 30, 2026; median follow-up 24.6 mo.

[Slide 3] Summary and Conclusions • INTerpath-001 met both the primary RFS and key secondary DMFS endpoints – Intismeran + pembrolizumab reduced the risk of recurrence or death by 42% and the risk of distant metastasis or death by 38% compared with placebo + pembrolizumab – In a sensitivity analysis, intismeran + pembrolizumab reduced the risk of recurrence or death by 58% after month 6 – Benefit was observed in most prespecified subgroups, including risk-based disease stages – OS was immature and did not meet prespecified significance boundarya; follow-up is ongoing • Safety was generally consistent with previous experience, with no potentiation of pembrolizumab-associated AEs with the combination • Intismeran was reliably delivered at a global scale • Ongoing phase 2 and 3 trials are evaluating intismeran for bladder cancer, melanoma, NSCLC, and RCC in the neoadjuvant, adjuvant, or metastatic settings • INTerpath-001 sets a new benchmark for adjuvant therapy in melanoma and establishes intismeran as a landmark advance in personalized cancer medicine aThe hazard ratio for OS was 0.77 (95% CI, 0.45–1.31); two-sided P = 0.33 Presented by: Georgina V. Long, Melanoma Institute Australia

[Slide 2] Neoadjuvant therapy for macroscopic resectable disease EFS, RFS according to pathological response, and Real-world OS NADINA - EFS neo-adjuvant ipi/nivo versus adj nivo macroscopic stage III [KM curve] Neoadjuvant 83.7 | Adjuvant 57.2 (labels at 12 months) Events/N: Neoadjuvant: 28/212; Adjuvant: 72/211 HR= 0.32 ( 99.9% CI: 0.15 - 0.66 ), p <0.0001 Months since Randomization NADINA - RFS neo-adjuvant ipi/nivo versus adj nivo macroscopic stage III [KM curve] Major pathological response 95.1 | Partial pathological response 76.1 | Pathological nonresponse 57.0 (labels at 12 months) Months since Randomization ESMO Guidelines: Neoadjuvant ipi/nivo is recommended for macroscopic resectable st III and IV melanoma Neoajduvant ipi/nivo is associated with 30% severe toxicity as compared to 15% for adjuvant anti-PD1 NADINA have not yet presented mature OS data Inge Marie Svane Blank et al NEJM 2024, Long et al Nature Med 2026 [Survival data captured as a table — no point estimates read from the curve] Endpoint: EFS (NADINA, neoadjuvant ipi/nivo vs adjuvant nivo, macroscopic stage III) Neoadjuvant ipi/nivo | 212 Adjuvant nivo | 211 HR: 0.32 HR CI: 99.9% CI 0.15-0.66 P: <0.0001 {'timepoint': '12 months', 'arm': 'Neoadjuvant ipi/nivo', 'value': '83.7', 'ci': None} {'timepoint': '12 months', 'arm': 'Adjuvant nivo', 'value': '57.2', 'ci': None} {'timepoint': '12 months', 'arm': 'RFS - major pathological response', 'value': '95.1', 'ci': None} {'timepoint': '12 months', 'arm': 'RFS - partial pathological response', 'value': '76.1', 'ci': None} {'timepoint': '12 months', 'arm': 'RFS - pathological nonresponse', 'value': '57.0', 'ci': None} Note: Landmark labels printed at the 12-month dashed line. Partial response label 76.1 is faint (yellow). Source cited on slide: Blank et al NEJM 2024, Long et al Nature Med 2026. [Slide 3] Neoadjuvant therapy for macroscopic resectable disease EFS, RFS according to pathological response, and Real-world OS NADINA - EFS neo-adjuvant ipi/nivo versus adj nivo macroscopic stage III [KM curve] Neoadjuvant 83.7 | Adjuvant 57.2 (labels at 12 months) Events/N: Neoadjuvant: 28/212; Adjuvant: 72/211 HR= 0.32 ( 99.9% CI: 0.15 - 0.66 ), p <0.0001 Months since Randomization NADINA - RFS neo-adjuvant ipi/nivo versus adj nivo macroscopic stage III [KM curve] Major pathological response 95.1 | Partial pathological response 76.1 | Pathological nonresponse 57.0 (labels at 12 months) Months since Randomization Mixed trial and RW data - OS neo-adjuvant CPI resectable stage III/IV [KM curve, OS (%) vs Time (years)] 89 / 83 (at 3 years); 88 / 83 (at 5 years) - PD-1 alone - PD-1 + IO P = 0.04 ESMO Guidelines: Neoadjuvant ipi/nivo is recommended for macroscopic resectable st III and IV melanoma Neoajduvant ipi/nivo is associated with 30% severe toxicity as compared to 15% for adjuvant anti-PD1 NADINA have not yet presented mature OS data Mixed trial and real-world data indicates excellent survival after neoadjuvant CPI ESMO Guidelines: discussions with the patient should consider the RFS benefits but lack of mature OS data Inge Marie Svane Blank et al NEJM 2024, Long et al Nature Med 2026 [Survival data captured as a table — no point estimates read from the curve] Endpoint: EFS (NADINA, neoadjuvant ipi/nivo vs adjuvant nivo, macroscopic stage III) Neoadjuvant ipi/nivo | 212 Adjuvant nivo | 211 HR: 0.32 HR CI: 99.9% CI 0.15-0.66 P: <0.0001 {'timepoint': '12 months', 'arm': 'Neoadjuvant ipi/nivo', 'value': '83.7', 'ci': None} {'timepoint': '12 months', 'arm': 'Adjuvant nivo', 'value': '57.2', 'ci': None} {'timepoint': '12 months', 'arm': 'RFS - major pathological response', 'value': '95.1', 'ci': None} {'timepoint': '12 months', 'arm': 'RFS - partial pathological response', 'value': '76.1', 'ci': None} {'timepoint': '12 months', 'arm': 'RFS - pathological nonresponse', 'value': '57.0', 'ci': None} {'timepoint': '3 years', 'arm': 'Mixed trial/RW OS - upper curve', 'value': '89', 'ci': None} {'timepoint': '3 years', 'arm': 'Mixed trial/RW OS - lower curve', 'value': '83', 'ci': None} {'timepoint': '5 years', 'arm': 'Mixed trial/RW OS - upper curve', 'value': '88', 'ci': None} {'timepoint': '5 years', 'arm': 'Mixed trial/RW OS - lower curve', 'value': '83', 'ci': None} Note: Landmark labels printed at the 12-month dashed line. Partial response label 76.1 is faint (yellow). Source cited on slide: Blank et al NEJM 2024, Long et al Nature Med 2026. Third panel (Mixed trial and RW data - OS, neo-adjuvant CPI, resectable stage III/IV): legend PD-1 alone vs PD-1 + IO, P = 0.04; legend colours not distinguishable in photo, so curve-to-arm mapping not assigned. [Slide 4] Take-home messages The INTerpath-001 results are compelling and mark a crucial moment for personalized cancer vaccines OPEN QUESTIONS • Will improved RFS translate into improved OS with longer FU and when will data be available ? • To which extend is the clinical efficacy of Intismeran driven by neoantigen-dependent immune effects ? • How should adjuvant pembro/intismeran be positioned compared to neoadjuvant ipi/nivo for macroscopic St III/IV ? – For patients without major pathological response ? • Further insight into predictive biomarkers and MOA are urgently needed HURDLES • Intismeran production process is complex and time-consuming which translate into an expensive drug • Intismeran is not available for a significant number of patients with tissue samples unsuitable for sequencing • Health technology assessments have yet to evaluate Intismeran real-world cost-effectiveness and feasibility Inge Marie Svane

[Slide 1] Science enables individualized therapeutic cancer vaccines Neoantigen discovery - Tumor-specific mutations - Potential for enhanced immunogenicity Cancer Cell | Patient A | Patient B | Patient C Neoantigen selection - NGS identifies tumor mutations - Computational prediction of HLA presentation and ranking The mRNA platform - Multiple antigens per mRNA - Robust CD8+ and CD4+ T cell responses - Rapid, individualized manufacturing Lipid nanoparticle (LNP) | mRNA Presented by: Georgina V. Long, Melanoma Institute Australia [footnote abbreviations and references illegible]

[Slide 2] RFS Sensitivity Analysis: mITT HR 0.51 (95% CI 0.39–0.67) 84.2% (95% CI 81.3–86.7) 70.4% (95% CI 65.3–74.9) | N | Events,a n (%) Intismeran + pembrolizumab | 726 | 116a (16.0%) Placebo + pembrolizumab | 365 | 106b (29.0%) Months: 0 | 6 | 12 | 18 | 24 | 30 | 35 No. at risk 726 | 674 | 638 | 611 | 133 | 9 | 0 365 | 333 | 294 | 275 | 71 | 2 | 0 [a3]1 patients in the intismeran + pembrolizumab arm were excluded from the mITT population due to AEs (n=13), recurrence (n=10), or other reason (n=8). RFS events occurred in 20 (64.5%) of the 31 patients who did not [re]ceive intismeran as planned. b15 patients in the placebo + pembrolizumab arm were excluded from the mITT population due to AEs (n=5), recurrence (n=4), or other reason (n=6). RFS events occurred in and 5 (33.3%) of [the 15] patients who did not receive placebo as planned. mITT: all randomized patients who received intismeran or placebo as assigned. RFS: time from randomization to the first occurrence of any recurrence or death from any [ca]use. Data cutoff date: June 30, 2026 (median study follow-up, 24.6 mo). [Pre]sented by: Georgina V. Long, Melanoma Institute Australia [Survival data captured as a table — no point estimates read from the curve] Endpoint: RFS sensitivity analysis, mITT Intismeran + pembrolizumab | 726 Placebo + pembrolizumab | 365 HR: 0.51 HR CI: 95% CI 0.39–0.67 {'timepoint': '24 months', 'arm': 'Intismeran + pembrolizumab', 'value': '84.2%', 'ci': '95% CI 81.3–86.7'} {'timepoint': '24 months', 'arm': 'Placebo + pembrolizumab', 'value': '70.4%', 'ci': '95% CI 65.3–74.9'} Note: No p-value printed (sensitivity analysis). Medians not printed. Footnote left edge cropped; bracketed text reconstructed only where obvious (e.g. 31 excluded = 13+10+8).

[Slide 3] Treatment Disposition and Exposure 1749 patients screened 302 (17.3%) had inadequate tumor sample 125 (7.1%) not disease free 91 (5.2%) did not provide informed consent 94 (5.4%) excluded for other reasons 1137 randomized (2:1 ratio) 757 to intismeran + pembrolizumab | 380 to placebo + pembrolizumab 756 received ≥1 dose of treatment | 379 received ≥1 dose of treatment 756 received ≥1 dose of pembrolizumab | 379 received ≥1 dose of pembrolizumab 726 (96.0%) received ≥1 dose of intismeran | 365 (96.3%) received ≥1 dose of placebo 1 (0.1%) received ≥1 dose of placebo(a) | 228 (30.2%) discontinued treatment | 117 (30.9%) discontinued treatment 114 (15.1%) adverse event | 37 (9.8%) adverse event 82 (10.8%) recurrence/relapse | 68 (17.9%) recurrence/relapse 32 (4.2%) other | 12 (3.2%) other 528 (69.8%) completed treatment | 262 (69.1%) completed treatment (a)The patient for whom intismeran could not be manufactured received placebo to maintain the blind. Presented by: Georgina V. Long, Melanoma Institute Australia

[Slide 2] Efficacy in Prespecified Subgroups RFS and DMFS benefits observed in most subgroups Recurrence-Free Survival Subgroup | Intismeran + Pembro Events/Patients | Placebo + Pembro Events/Patients | Hazard Ratio (95% CI) All patients | 136/757 | 111/380 | 0.58 (0.45-0.75) Age <65 yr | 79/458 | 61/230 | 0.65 (0.47-0.91) >=65 yr | 57/299 | 50/150 | 0.53 (0.36-0.77) Sex Female | 57/301 | 40/153 | 0.71 (0.47-1.06) Male | 79/456 | 71/227 | 0.54 (0.39-0.74) Geographic region Australia | 16/89 | 12/39 | 0.54 (0.26-1.15) Europe | 70/415 | 54/209 | 0.64 (0.45-0.92) Latin America | 15/45 | 9/22 | 0.85 (0.37-1.95) United States | 12/84 | 10/35 | 0.50 (0.22-1.16) Rest of world | 23/124 | 26/75 | 0.50 (0.29-0.89) ECOG PS 0 | 120/697 | 97/350 | 0.61 (0.47-0.80) 1 | 16/60 | 14/30 | 0.48 (0.24-0.99) Risk-based disease stage IIB/IIC/IIIA/IIIB | 59/440 | 50/230 | 0.60 (0.41-0.88) IIIC/IIID/IV | 77/316 | 61/150 | 0.57 (0.41-0.79) Axis: 0.2 | 0.5 | 1.0 | 0.5 ; Intismeran + Pembro Better / Placebo + Pembro Better Distant Metastasis-Free Survival Subgroup | Intismeran + Pembro Events/Patients | Placebo + Pembro Events/Patients | Hazard Ratio (95% CI) All patients | 89/757 | 71/380 | 0.62 (0.45-0.85) Age <65 yr | 47/458 | 37/230 | 0.65 (0.42-0.99) >=65 yr | 42/299 | 34/150 | 0.61 (0.39-0.95) Sex Female | 31/301 | 27/153 | 0.58 (0.34-0.97) Male | 58/456 | 44/227 | 0.66 (0.44-0.97) Geographic region Australia | 12/89 | 11/39 | 0.42 (0.19-0.96) Europe | 42/415 | 34/209 | 0.63 (0.40-0.99) Latin America | 11/45 | 9/22 | 0.62 (0.26-1.49) United States | 8/84 | 2/35 | 1.77 (0.38-8.35) Rest of world | 16/124 | 15/75 | 0.63 (0.31-1.28) ECOG PS 0 | 77/697 | 60/350 | 0.65 (0.46-0.91) 1 | 12/60 | 11/30 | 0.47 (0.21-1.06) Risk-based disease stage IIB/IIC/IIIA/IIIB | 41/440 | 33/230 | 0.65 (0.41-1.02) IIIC/IIID/IV | 48/316 | 38/150 | 0.59 (0.39-0.91) Axis: 0.2 | 0.5 | 1.0 | 2.0 | 3.0 | 9.0 ; Intismeran + Pembro Better / Placebo + Pembro Better ITT: all randomized patients according to the arm to which they were assigned. Data cutoff date: June 30, 2026 (median study follow-up, 24.6 mo). Presented by: Georgina V. Long, Melanoma Institute Australia

[Slide 1] [Left slide] Summary and Conclusions • INTerpath-001 met both the primary RFS and key secondary DMFS endpoints – Intismeran + pembrolizumab reduced the risk of recurrence or death by 42% and the risk of distant metastasis or death by 38% compared with placebo + pembrolizumab – In a sensitivity analysis, intismeran + pembrolizumab reduced the risk of recurrence or death by 58% after month 6 – Benefit was observed in most prespecified subgroups, including risk-based disease stages – OS was immature and did not meet prespecified significance boundarya; follow-up is ongoing • Safety was generally consistent with previous experience, with no potentiation of pembrolizumab-associated AEs with the combination • Intismeran was reliably delivered at a global scale • Ongoing phase 2 and 3 trials are evaluating intismeran for bladder cancer, melanoma, NSCLC, and RCC in the neoadjuvant, adjuvant, or metastatic settings • INTerpath-001 sets a new benchmark for adjuvant therapy in melanoma and establishes intismeran as a landmark advance in personalized cancer medicine aThe hazard ratio for OS was 0.77 (95% CI, 0.45–1.31); two-sided P = 0.33. Presented by: Georgina V. Long, Melanoma Institute Australia [Right slide] Take-home messages The INTerpath-001 results are compelling and mark a crucial moment for personalized cancer vaccines OPEN QUESTIONS • Will improved RFS translate into improved OS with longer follow-up and when will data be available ? • To which extent is the clinical efficacy of intismeran driven by neoantigen-dependent immune effects ? • How should adjuvant pembrolizumab/intismeran be positioned compared to neoadjuvant ipi/nivo for macroscopic stage III/IV ? – For patients without major pathological response ? • Further insight into predictive biomarkers and mechanisms of action are urgently needed. HURDLES • Intismeran production process is complex and time-consuming which translate into an expensive drug • Intismeran is not available for a significant number of patients with tissue samples unsuitable for sequencing • Health technology assessments have yet to evaluate Intismeran real-world cost-effectiveness and feasibility PERSPECTIVES • Potential for broader use across other cancer types and disease stages • Potential for new treatment combinations Inge Marie Svane

[Slide 1] [Madrid 2026 ESMO congress logo] Trial design and patient selection INTerpath-001: - Melanoma – high mutation load, highly immunogenic - Pembrolizumab – potent backbone, with vaccine directing T cell response - High risk population – resected stage IIB – IV melanoma CRC: BNT-122: - MSS/MMRp CRC – modest mutation load, low immunogenicity & often immune-excluded - Vaccine monotherapy – in a tumour type where immunotherapy has so far not succeeded - Very high risk population: - ctDNA positive - Roche Avenio ctDNA test limit of detection ~0.1% vs. Signatera 0.01%1 - Potential under-detection of low-level MRD 1) Chen et al. Mol Diagn Ther 2021 [Slide 2] [Madrid 2026 ESMO congress logo] Taking the brakes off immune cells and tackling immunosuppression in CRC - MMRp CRC: modest immunogenicity, often immune excluded - In the metastatic setting: negligible response rates and PFS benefit from PD1/PDL1 +/- CTLA4 ICIs1 - Still: omission of trial arm with vaccine + ICI appears like a key mistake now - Several next-generation ICIs look more promising in CRC: - Botensilimab/Balstilimab (PD1 + Fc-enhanced CTLA4): 17% RR in 101 MSS CRC (0% in patients with liver metastases)2 - INCA33890 (PD1/TGFBR bispecific AB): 15.2% RR in 105 MSS CRC (12% in patients with liver metastases)3 - We need combination trials with ICIs that overcome immunosuppression in CRC. 1) Chen et al. Jama Onc 2020 2) Bullock et al. Nat Med 2024 3) Moyers GI ASCO 2026

[Slide 1] NEWS [Menu] First personalised cancer vaccine cuts risk of disease returning in major trial [Photo: nurse administering an injection to a seated man in a clinic] The trial also found that the new drug combination reduced the risk of melanoma, a type of skin cancer, spreading by 38% over two years compared to immunotherapy alone. The drug manufacturers MSD and Moderna say the same approach is already being tested against other tumours, including kidney, bladder and lung cancer.

[Slide 1] 8 Results: GC101 TIL Survival vs. ICC • During extended follow-up, PFS benefit maintained with longer follow-up. • Median PFS*: 4.37 vs. 1.61 months, HR = 0.44, p = 0.0004 (ASCO 2026: 4.30 vs. 1.58 months, HR=0.43, p=0.0002). • Even with 59%(ASCO 2026: 57%) chemo cross-over, a survival benefit trend showed in the TIL group. PFS — Primary analysis GC101-TIL VS ICC: Event/N: 32/50 43/49; mPFS(95%CI): 4.37(3.98-5.55) 1.61(1.28-2.79); HR(95%CI): 0.44(0.28-0.70); P value: 0.0004 PFS Rate(%) at 6 month: 34.48 VS 10.24 No.at Risk: GC101-TIL 50 35 26 9 5 3 3 0; ICC 49 21 12 4 3 1 1 0 PFS — Sensitive analysis GC101-TIL VS ICC: Event/N: 22/37 42/47; mPFS(95%CI): 4.21(2.79-5.72) 1.54(1.28-2.66); HR(95%CI): 0.44(0.26-0.74); P value: 0.0013 PFS Rate(%) at 6 month: 29.73 VS 10.05 No.at Risk: GC101-TIL 37 28 15 4 4 3 2 0; ICC 47 19 12 4 3 1 1 0 OS§ — Primary analysis GC101-TIL VS ICC: Events/N: 9/50 15/49; mOS(95%CI): NE(9.49-NE) 12.55(8.74-NE); HR(95%CI): 0.63(0.28-1.41); P value: 0.2571 No.at Risk: GC101-TIL 50 47 38 28 20 11 7 1 0; ICC 49 49 41 30 21 16 12 7 0 * IRC-assessed § Not yet mature CI, confidence interval; HR, hazard ratio; ICC, investigator's choice of chemotherapy; OS, overall survival; PFS, progression-free survival; TIL, tumor-infiltrating lymphocytes. Updated Data cutoff date: July 2, 2026 (ASCO 2026 Data cutoff date : May 12, 2026) [Survival data captured as a table — no point estimates read from the curve] Endpoint: PFS (IRC) and OS (immature) GC101 TIL (PFS, primary analysis) | 50 | 4.37 mo | 3.98–5.55 ICC (PFS, primary analysis) | 49 | 1.61 mo | 1.28–2.79 GC101 TIL (OS) | 50 | NE | 9.49–NE ICC (OS) | 49 | 12.55 mo | 8.74–NE HR: PFS 0.44; OS 0.63 HR CI: PFS 0.28–0.70; OS 0.28–1.41 P: PFS 0.0004; OS 0.2571 {'timepoint': '6 months', 'arm': 'GC101 TIL (PFS primary)', 'value': '34.48%', 'ci': None} {'timepoint': '6 months', 'arm': 'ICC (PFS primary)', 'value': '10.24%', 'ci': None} {'timepoint': '6 months', 'arm': 'GC101 TIL (PFS sensitivity)', 'value': '29.73%', 'ci': None} {'timepoint': '6 months', 'arm': 'ICC (PFS sensitivity)', 'value': '10.05%', 'ci': None} Note: Sensitivity analysis PFS: 4.21 vs 1.54 mo, HR 0.44 (0.26–0.74), P=0.0013, n 37 vs 47. Updated DCO July 2, 2026 (ASCO 2026 DCO May 12, 2026: 4.30 vs 1.58, HR 0.43). OS immature; 59% crossover. [Slide 2] 9 Subgroup Analysis: PFS by Subtype • Compared with Chemo, the TIL group showed PFS benefit across all subtypes. • Median PFS* with GC101 TIL was 4.81, 4.37, and 4.14 months in the mucosal, acral, and cutaneous subgroups. Mucosal subtype GC101-TIL VS ICC: Event/N: 5/6 9/10; mPFS(95%CI): 4.81(2.33-NE) 1.45(0.30-4.04); HR(95%CI): 0.23(0.06-0.88); P value: 0.0215 PFS Rate(%) at 6 month: 33.33 VS NA No.at Risk: GC101-TIL 6 6 4 2 2 1 1 0; ICC 10 4 2 0 Acral subtype GC101-TIL VS ICC: Event/N: 20/31 22/25; mPFS(95%CI): 4.37(2.99-5.55) 1.38(1.22-2.66); HR(95%CI): 0.41(0.22-0.77); P value: 0.0034 PFS Rate(%) at 6 month: 30.96 VS 10.34 No.at Risk: GC101-TIL 31 23 17 5 2 1 1 0; ICC 25 8 5 2 2 0 * IRC-assessed; Subgroup analyses should be interpreted with caution due to the limited sample size. CI, confidence interval; HR, hazard ratio; ICC, investigator's choice of chemotherapy; OS, overall survival; PFS, progression-free survival; TIL, tumor-infiltrating lymphocytes. Updated Data cutoff date: July 2, 2026 (ASCO 2026 Data cutoff date : May 12, 2026) [Survival data captured as a table — no point estimates read from the curve] Endpoint: PFS (IRC) by subtype Mucosal: GC101 TIL | 6 | 4.81 mo | 2.33–NE Mucosal: ICC | 10 | 1.45 mo | 0.30–4.04 Acral: GC101 TIL | 31 | 4.37 mo | 2.99–5.55 Acral: ICC | 25 | 1.38 mo | 1.22–2.66 HR: Mucosal 0.23; Acral 0.41 HR CI: Mucosal 0.06–0.88; Acral 0.22–0.77 P: Mucosal 0.0215; Acral 0.0034 {'timepoint': '6 months', 'arm': 'Mucosal GC101 TIL', 'value': '33.33%', 'ci': None} {'timepoint': '6 months', 'arm': 'Mucosal ICC', 'value': 'NA', 'ci': None} {'timepoint': '6 months', 'arm': 'Acral GC101 TIL', 'value': '30.96%', 'ci': None} {'timepoint': '6 months', 'arm': 'Acral ICC', 'value': '10.34%', 'ci': None} Note: Cutaneous median 4.14 mo (text only). Small subgroups; interpret with caution. [Slide 3] 10 Objective Response: GC101 TIL vs. ICC • Primary analysis: 44.0% vs. 8.2% (ASCO 2026: 42.0% vs. 6.1%). • Sensitivity analysis: 40.5% vs. 4.3% (ASCO 2026: 34.3% vs. 4.3%). • Extended follow-up, target lesion CRs increased from 1 to 4 in GC101 TIL arm. Response, n (%) | Primary analysis: GC101 TIL (N=50) | ICC (N=49) | Sensitivity analysis*: GC101 TIL (N=37) | ICC (N=47) BOR CR | 1(2.0) | 0(0.0) | 1 (2.7) | 0 (0.0) PR | 21(42.0) | 4(8.2) | 14 (37.8) | 2 (4.3) SD | 10(20.0) | 15(30.6) | 11 (29.7) | 15 (31.9) PD | 13(26.0) | 29(59.2) | 10 (27.0) | 30 (63.8) NE | 5(10.0) | 1(2.0) | 1(2.7) | 0 ORR | 22(44.0) | 4(8.2) | 15(40.5) | 2(4.3) DCR | 32(64.0) | 19(38.8) | 26(70.3) | 17(36.2) [Waterfall plots: GC101 TIL§ ; ICC§ — % best percentage change] *Investigator-assessed § Primary analysis BOR, best overall response; Chemo, chemotherapy; CR, complete response; DCR, disease control rate; DOR, duration of response; FAS, full analysis set; ICC, investigator's choice of chemotherapy; IRC, independent review committee; NE, not evaluable; ORR, objective response rate; PD, progressive disease; PPS, per protocol set; PR, partial response; SD, stable disease; TIL, tumor-infiltrating lymphocytes. Updated Data cutoff date: July 2, 2026 (ASCO 2026 Data cutoff date : May 12, 2026) [Slide 4] 13 Conclusions ■ MIZAR-003: First randomized trial of TIL therapy in late-line, anti-PD-1–refractory advanced melanoma — primary endpoint met (median follow-up: 8.3 months) ➢ PFS benefit observed in all the subtypes (by IRC in the primary analysis): • Overall: 4.37m vs. 1.61m (HR 0.44, P = 0.0004); • Mucosal: 4.81m vs. 1.45m (HR 0.23, P = 0.0215); • Acral: 4.37m vs. 1.38m (HR 0.41, P = 0.0034); ➢ Tumor Response(by IRC in the primary analysis): • ORR: 44.0% vs. 8.2%; DCR: 64.0% vs. 38.8%; ➢ OS remains immature, trend favoring TIL despite 59% crossover from the chemotherapy arm; ➢ Safety: No new safety signals, ICU admissions, or treatment-related deaths reported to date. ■ The encouraging Phase II efficacy and safety results, particularly in acral and mucosal melanoma, support further evaluation in a Phase III trial.

[Slide 1] HBI-8000-303: Study Design Advanced, unresectable melanoma naïve to anti-PD1/-L1 therapy -> R 1:1 N=203: HBI-8000 30mg BIW + Nivolumab 480mg IV Q4W N=201: Placebo + Nivolumab 480mg IV Q4W -> Follow-up Primary endpoint: PFS by BIRC Secondary endpoints: -ORR by BIRC -Overall Survival -Safety and Toxicity Treatment until progression, intolerable toxicity, withdrawal of consent, or completion of 24 months of therapy Stratification Factors: Serum LDH; Tumor PD-L1 Expression* • 123 sites in 15 countries • Accrual from July 2021 to April 2024 • Separate open label cohort (N=45) of asymptomatic brain metastases treated with HBI-8000 + Nivolumab (reported at ESMO 2025; 1630P) * PD-L1 Positive: ≥ 1% tumor membrane staining in a minimum of 100 evaluable tumor cells Nikhil I. Khushalani, MD [ESMO Congress Madrid 2026 logo] [Slide 2] Efficacy: Progression-Free Survival, Normal LDH HBI + NIVO (N=135) | Placebo + NIVO (N=133) Events: 66 | 89 Median (months): 20.04 | 7.39 [95% CI]: [8.31, 35.61] | [4.63, 13.04] HR [96% CI]: 0.65 [0.47, 0.90] P-value: 0.0082 Landmark labels at 24 months: 47.2% (95% CI 37.7-56.1) [HBI-8000 + Nivolumab]; 32.6% (95% CI 24.3-41.2) [Placebo + Nivolumab] Y-axis: Probability of Progression-Free Survival; X-axis: Time from Randomization (Months) Legend: HBI-8000 + Nivolumab; Placebo + Nivolumab; Censored N at Risk (0,6,12,18,24,30,36,42,47 mo): HBI-8000: 135 77 61 50 37 26 14 2 0 Placebo: 133 67 53 43 28 18 7 3 0 Nikhil I. Khushalani, MD [ESMO Congress Madrid 2026 logo] [Survival data captured as a table — no point estimates read from the curve] Endpoint: PFS (Normal LDH subgroup) HBI-8000 + Nivolumab | 135 | 20.04 mo | 8.31–35.61 Placebo + Nivolumab | 133 | 7.39 mo | 4.63–13.04 HR: 0.65 HR CI: 0.47–0.90 (96% CI) P: 0.0082 {'timepoint': '24 mo', 'arm': 'HBI-8000 + Nivolumab', 'value': '47.2%', 'ci': '37.7–56.1'} {'timepoint': '24 mo', 'arm': 'Placebo + Nivolumab', 'value': '32.6%', 'ci': '24.3–41.2'} Note: Subgroup (normal LDH), not ITT. Landmark labels small; read at limit of legibility. [Slide 3] Efficacy: Objective Response Rate (BIRC) Response | HBI-8000 + Nivolumab (N=203) | Placebo + Nivolumab (N=201) CR | 29 (14.3%) | 14 (7.0%) PR | 55 (27.1%) | 51 (25.4%) SD | 41 (20.2%) | 49 (24.4%) PD | 54 (26.6%) | 74 (36.8%) ORR | 84 (41.4%) | 65 (32.3%) 95% CI – ORR [a] | [34.5, 48.5] | [25.9, 39.3] OR (96% CI) [b] | 1.50 (0.97, 2.33) P Value [b] | 0.0569 [a]: Clopper-Pearson exact 95% CIs for ORR [b]: OR and its 96% CI, and p value are derived using Cochran-Mantel-Haenszel (CMH) stratified by LDH and PD-L1 status. Nikhil I. Khushalani, MD [ESMO Congress Madrid 2026 logo] [Slide 4] Adverse Events (Clinical; ≥10%) Columns: HBI-8000 + Nivolumab (N=199) Any Grade | Grade 3-4 || Placebo + Nivolumab (N=201) Any Grade | Grade 3-4 Any AE: 196 (98.5) | 142 (71.4) || 196 (97.5) | 91 (45.3) TRAE*: 184 (92.5) | 109 (54.8) || 160 (79.6) | 36 (17.9) Diarrhea: 88 (44.2) | 11 (5.5) || 29 (14.4) | 0 (0) Nausea: 69 (34.7) | 2 (1.0) || 41 (20.4) | 2 (1.0) Fatigue: 42 (21.1) | 4 (2.0) || 42 (21.4) | 1 (0.5) Asthenia: 41 (20.6) | 2 (1.0) || 19 (9.5) | 0 (0) ↓ Appetite: 37 (18.6) | 4 (2.0) || 18 (9.0) | 1 (0.5) Vomiting: 36 (18.1) | 2 (1.0) || 13 (6.5) | 2 (1.0) Arthralgia: 26 (13.1) | 2 (1.0) || 16 (8.0) | 0 (0) ↓ Weight: 26 (13.1) | 3 (1.5) || 6 (3.0) | 0 (0) Hypothyroidism: 24 (12.1) | 1 (0.5) || 18 (9.0) | 0 (0) Vitiligo: 21 (10.6) | 0 (0) || 19 (9.5) | 0 (0) Pruritus: 20 (10.1) | 0 (0) || 34 (16.9) | 0 (0) Rash: 19 (9.5) | 0 (0) || 25 (12.4) | 0 (0) TRAE leading to discontinuation**: 28 (14.1) | 16 (8.0) || 10 (5.0) | 6 (3.0) * Causality of adverse events as assessed by investigator **Study treatment discontinuation of later drug Nikhil I. Khushalani, MD [ESMO Congress Madrid 2026 logo]

[Slide 1] [Screen with speaker video of Toni K. Choueiri at podium; MADRID 2026 ESMO congress logo] Why mRNA? Key Advantages for Vaccine Development [Radial diagram centered on: mRNA Cancer Vaccine] Short research and development cycle Rapid production Highly targeted Lower production costs Fewer side effects Lower risk of insertional mutagenesis Support personalized therapy Lower risk of viral infection Induce stronger immune responses Wang, et al. Front. Immunol. 2023. Toni K. Choueiri, MD (c) 2023 Moderna, Inc. All rights reserved.

[Slide 1] [ESMO Congress Madrid 2026 logo] What determines the immunogenicity of a mutation? Variant in DNA • Is the mutation "real"? • Clonal vs. subclonal? • Expression? (RNAseq) Peptide selection • Proteosomal processing • "Foreignness" • Similarity to Pathogens Peptide-HLA interaction • Binding Strength • Stability of pHLA complex Peptide T cell interaction • TCR repertoire – a vast space • Much harder to model than HLA • Need experimental TCR-pHLA Datasets! [Diagram: Tumor cell; DNA mutations; Antigenic precursors; Proteolytic intermediates; MHC; T cell receptor; Anchor residues of peptide; PEPTIDE; T cell contact residues of peptide] Purroy & Wu CSH Perspect Med 2017 [Slide 2] [ESMO Congress logo] Functional, diverse, and durable immune responses Generation of memory response [UMAP plots: Cluster 0 1 2 3; Patient 3 4 5; UMAP_1 / UMAP_2] [Line plots: Naive, Cytotoxic, AICD, Memory; y-axis % of tetramer+ CD4+ T cells; x-axis Pre 3 8 12 16 20 24 wks; Priming / Boosting] T cell diversification Pt. 4 Priming: wk 3, wk 8, wk 12 Boosting: wk 16, wk 20, wk 24 [pie charts] Persistence of vax-specific T cell responses Pt. 3 Week 16 | 47 months CD4+ T cells: DMSO, mut-VPS16, mut-CASP1, mut-CRY1b, mut-CRY1c CD8+ T cells: DMSO, mut-VPS16, mut-CASP1 [ELISpot well images] Hu Z & Leet DE Nat. Med., 2021 [Slide 3] [ESMO Congress logo] PD-1 inhibition and NeoVax induce distinct sets of circulating TCR clonotypes in all vaccinated patients [Left plot] Number of significantly expanding clones; After Nivo vs After Vaccine; p = 0.00219 Patients: P101, P103, P104, P105, P106, P108, P109, P110, P111 [Middle plot] Patient 103; Nivolumab; NeoVaxM; Vaccine-expanded? Frequecny [sic]; Timepoint: Pre-treatment, Post-Nivolumab, Post-vaccine, W48, W72; LOD Type: Other, Existing, Post-Nivolumab, Post-Vaccine [Pie chart] Median percent of vaccine expanded post-vaccine clones Post-vax clones: Not expanded 75.18%; Expanded 24.82% Blass et al, Cell, 2025 [Slide 4] [ESMO Congress Madrid 2026 logo] How can we improve vaccines further? •Immune Adjuvants •Vaccine Platform Technology - Peptides Link peptides and adjuvant to self-assembling nanoparticles [Diagram: Peptide antigen; CpG adjuvant; DMPC lipid; 22A apoA1 peptide; Lipid nanodisc; Subcutaneous; Lymph node] Link peptides and adjuvant to an albumin-binding phospholipid tail [Diagram: Albumin-binding phospholipid; PEG linker; Peptide antigen; CpG DNA; Membrane insertion; Albumin] Kim et al, Nat Nanotechnol, 2025 [podium: ESMO Congress Madrid Spain 23-27 October 2026]

[Slide 1] ORR PER BICR Encorafenib and binimetinib plus pembrolizumab demonstrated clinically meaningful and statistically significant improvement in ORR vs the control arm Confirmed ORR per BICR, (95% CI) Odds ratio (90% CI): 4.903 (2.870, 8.125) P<0.0001^a [Bar chart] Triplet (n=110): 70.0% (60.9, 77.8); Control (n=110): 32.7% (24.7, 41.9); legend CR / PR All randomized patients | Triplet (n=110) | Control (n=110) Confirmed ORR, n (%) [95% CI] | 77 (70.0) [60.9, 77.8] | 36 (32.7) [24.7, 41.9] Odds ratio (90% CI) | 4.903 (2.870, 8.125) 1-sided P-value | <0.0001 Best overall response, n (%) CR | 12 (10.9) | 5 (4.5) PR | 65 (59.1) | 31 (28.2) SD | 16 (14.5) | 16 (14.5) Non-CR/non-PD | 4 (3.6) | 1 (0.9) PD | 5 (4.5) | 47 (42.7) Not evaluable | 8 (7.3) | 10 (9.1) Responders | n=77 | n=36 TTR, median (range), weeks | 9.1 (7.7-58.3) | 9.7 (7.1-46.1) DOR, median (95% CI), months | NE (20.7, NE) | NE (18.8, NE) Patients with a response >=15 months, n (%) | 43 (55.8) | 25 (69.4) Data cutoff: January 12, 2026. ^aOne-sided P-value. BICR, blinded independent central review; CR, complete response; NE, not estimable; PD, progressive disease; PR, partial response; SD, stable disease; TTR, time to response. Dirk Schadendorf, MD [ESMO Congress Madrid 2026 logo; presenter at podium] [Slide 2] PFS PER BICR Encorafenib and binimetinib plus pembrolizumab demonstrated statistically significant improvement in PFS over the control arm | Triplet (n=110) | Control (n=110) Follow-up, median (95% CI), months | 24.9 (24.6, 27.8) | 30.2 (25.8, 33.1) Events, n (%) | 44 (40) | 77 (70) PFS, median (95% CI), months | 33.0 (20.0, NE) | 3.4 (2.2, 8.3) Hazard ratio (stratified) (90% CI) | 0.33 (0.24, 0.46) 1-sided P-value | <0.0001 12 months: Triplet 70.7%; Control 34.3% 24 months: Triplet 54.8%; Control 26.4% Median PFS: 3.4 mo (95% CI: 2.2, 8.3); Median PFS: 33.0 mo (95% CI: 20.0, NE) No. at risk (0,3,...,42 months): Triplet: 110 95 89 69 62 58 52 46 40 27 19 12 4 0 0 Control: 110 57 44 35 31 31 29 24 21 18 13 10 4 2 0 Data cutoff: January 12, 2026. BICR, blinded independent central review; NE, not estimable Dirk Schadendorf, MD [ESMO Congress Madrid 2026 logo; presenter at podium] [Survival data captured as a table — no point estimates read from the curve] Endpoint: PFS per BICR Encorafenib + binimetinib + pembrolizumab (triplet) | 110 | 33.0 mo | 20.0-NE Placebo + pembrolizumab (control) | 110 | 3.4 mo | 2.2-8.3 HR: 0.33 (stratified) HR CI: 0.24-0.46 (90% CI) P: <0.0001 (1-sided) {'timepoint': '12 mo', 'arm': 'Triplet', 'value': '70.7%', 'ci': None} {'timepoint': '12 mo', 'arm': 'Control', 'value': '34.3%', 'ci': None} {'timepoint': '24 mo', 'arm': 'Triplet', 'value': '54.8%', 'ci': None} {'timepoint': '24 mo', 'arm': 'Control', 'value': '26.4%', 'ci': None} Note: Same slide as 2113638289521471984_s2. DCO January 12, 2026. [Slide 3] OS OS showed a positive trend favoring encorafenib and binimetinib plus pembrolizumab vs control but was immature at final analysis | Triplet (n=110) | Control (n=110) Follow-up, median (95% CI), months | 30.8 (29.0, 32.6) | 30.9 (28.8, 32.7) Events, n (%) | 32 (29.1) | 41 (37.3) OS, median (95% CI), months | NE (NE, NE) | NE (31.8, NE) Hazard ratio (stratified) (90% CI) | 0.71 (0.48, 1.05) 12 months: Triplet 89.5%; Control 78.7% 24 months: Triplet 76.6%; Control 64.2% No. at risk (0,6,...,42 months): Triplet: 110 102 92 84 74 43 11 0 Control: 110 92 77 71 57 37 12 2 Data cutoff: January 12, 2026. NE, not estimable Dirk Schadendorf, MD [ESMO Congress Madrid 2026 logo; presenter video panel] [Survival data captured as a table — no point estimates read from the curve] Endpoint: OS Encorafenib + binimetinib + pembrolizumab (triplet) | 110 | NE | NE-NE Placebo + pembrolizumab (control) | 110 | NE | 31.8-NE HR: 0.71 (stratified) HR CI: 0.48-1.05 (90% CI) {'timepoint': '12 mo', 'arm': 'Triplet', 'value': '89.5%', 'ci': None} {'timepoint': '12 mo', 'arm': 'Control', 'value': '78.7%', 'ci': None} {'timepoint': '24 mo', 'arm': 'Triplet', 'value': '76.6%', 'ci': None} {'timepoint': '24 mo', 'arm': 'Control', 'value': '64.2%', 'ci': None} Note: Same slide as 2113638289521471984_s3. OS immature. [Slide 4] MOST FREQUENT (>=15%)^a ALL-CAUSE TEAES AEs were consistent with the known safety profiles of the individual components of encorafenib and binimetinib plus pembrolizumab Triplet (n=110): Grade 1/2, Grade 3/4 | Control (n=110): Grade 1/2, Grade 3/4 Triplet all-grade frequencies by AE (as listed top to bottom): Nausea 45%; Diarrhea 40%; Alanine aminotransferase increased 33%; Fatigue 33%; Pyrexia 32%; Blood creatine phosphokinase increased 31%; Arthralgia 30%; Lipase increased 29%; Pruritus 26%; Rash 26%; Vomiting 26%; Aspartate aminotransferase increased 25%; Headache 21%; Anemia 20%; Asthenia 18%; Constipation 16%; Blood creatinine increased 15% [Grade 3/4 and control-arm bar values visible (e.g. 2%, 5%, 14%, 1%, 8%, 3%, 12%, 4%, 11%, 7%; control 17%, 19%, 10%, 17%, 14%, 15%, 21%, 14%, 12%, 15%, 15%, 9%, 15%, 15%, 15%, 10%, 7%) but row alignment cannot be reliably assigned from this angled photo] Treatment-related (including immune-related)^b AEs of special interest were reported in 46.4% of patients in the triplet arm and 17.3% of patients in the control arm X-axis: Patients, % (50 - 25 - 0 - 25 - 50) Data cutoff: January 12, 2026 ^aFrequency is based on the triplet arm. ^bIncluding cardiac failure, cytokine release syndrome, diarrhea, hepatic events/hepatitis, myocarditis, pneumonitis/intestinal lung disease, rash, and uveitis (group terms). AE, adverse event; TEAE, treatment-emergent adverse event Dirk Schadendorf, MD [ESMO Congress Madrid 2026 logo]

[Slide 2] Which Cancers Should We Target? ...benefit may extend to immunologically "Cold" tumors [emoji] Autogene Cevumeran in Resected Pancreatic Cancer: Phase 1 trial and 3-year follow-up [Schema] Resection -> Process and transport tissues -> Sequence tumour and normal DNA; Tumour RNA -> Predict and select neoantigens -> Custom manufacture autogene cevumeran (individualized neoantigen-encoding mRNA-lipoplex) Custom manufacture autogene cevumeran Key inclusion criteria - All surgically resectable PDAC - No borderline resectable - No locally advanced or metastatic disease - No neoadjuvant therapy - >=5 neoantigens [Timeline] Screen for eligibility; Week 0 Surgery; 6 Atezolizumab 1 dose; 9-17 Autogene cevumeran Priming doses 1-8; 21-43 mFOLFIRINOX 12 cycles (biweekly); 46 Booster dose 9; Follow-up Rojas, et al. Nature 2023. [Bar chart: Number of neoantigens in autogene cevumeran; Responders (n = 8), Non-responders (n = 8); legend Immunogenic / Non-immunogenic / No data; pie 50% / 50%] [Small KM inset: RFS, Responders (n = 8) vs Non-responders (n = 8); Median RFS: not reached; Median RFS 13.4 months; P = 0.007; HR [illegible]; Median follow-up [illegible] years] Bethune, et al. Nature 2025. Toni K. Choueiri, MD [ESMO Congress Madrid 2026 logo] [Slide 3] Which Cancers Should We Target? ...and beyond, to cancer syndromes Nous-209 Neoantigen Vaccine For Cancer Prevention In Lynch Syndrome Carriers: A Phase Ib/II Trial [Schema] Population: LS carriers n = 45 GAd (week 0); MVA (week 8); Weeks 0, 4, 8, 12, 24, 28, 48, 52/68 Blood collection and colonoscopy PBMCs; P1-P16 Peptide pools; IFNγ; T cell [Charts: SFCs per 10^6 PBMCs Baseline vs Peak; immune response over time n = 27 100%, n = 33, n = 33 [percentages partly illegible]; Negative immune resp / Positive immune resp; No. of adenomas per participant (Baseline vs EoS, LS carriers); Advanced adenomas present] Toni K. Choueiri, MD D'Alise, et al. Nature Medicine 2026. [ESMO Congress Madrid 2026 logo] [Slide 4] What about cancers with low TMB? Subgroup^a | Intismeran plus pembrolizumab (n = 107) Events, % (n/N) | Pembrolizumab (n = 50) Events, % (n/N) | Hazard ratio (95% CI) Overall (N = 157) | 26.2 (28/107) | 46.0 (23/50) | 0.510 (0.294 to 0.887) Disease stage: Stage IIIC (n = 131) | 28.1 (25/89) | 45.2 (19/42) | 0.563 (0.310 to 1.024) Stage IV (n = 22) | 18.8 (3/16) | 33.3 (2/6) | 0.480 (0.080 to 2.898) PD-L1 status: Positive (n = 98) | 25.7 (18/70) | 50.0 (14/28) | 0.439 (0.218 to 0.885) Negative (n = 19) | 28.6 (4/14) | 100 (5/5) | 0.143 (0.033 to 0.619) BRAF status: Mutant^b (n = 61) | 22.0 (9/41) | 55.0 (11/20) | 0.350 (0.145 to 0.847) Wildtype (n = 96) | 28.8 (19/66) | 40.0 (12/30) | 0.641 (0.311 to 1.324) TMB status [boxed]: High^c (n = 109) | 22.8 (18/79) | 36.7 (11/30) | 0.579 (0.273 to 1.227) [boxed] Non-high^d (n = 45) | 38.5 (10/26) | 63.2 (12/19) | 0.515 (0.222 to 1.197) [boxed] ctDNA status: Positive^e (n = 15) | 76.9 (10/13) | 100 (2/2) | NE Negative (n = 109) | 15.8 (12/76) | 51.5 (17/33) | 0.241 (0.115 to 0.507) [Forest plot, Hazard Ratio axis 0.01, 0.1, 1, 5, 10] ^aRFS in the overall population was estimated by a stratified Cox proportional hazards model; an unstratified analysis was used for subgroup analyses. ^bV600K or V600E mutation. ^c>=175 mutations/exome; ^d<175 mutations/exome. ^eHazard ratio for the ctDNA-positive subgroup was NE as all patients in the pembrolizumab had a recurrence event. Subgroups were defined as previously described.^1 1. Weber JS, et al. Lancet. 2024;403(10427):632-644 Carlino et al, ASCO 2026

[Slide 2] Selection of antigens for adoptive cellular therapy TIL therapy (intra- and extracellular antigens) TCR therapy (intracellular antigens) CAR-T therapy (surface antigens) Private (patient-specific) neoantigens Tumour (patient)-specific mutations Low risk of off-target toxicity HLA-restricted Clonal/subclonal - subject to immune editing Public (shared) neoantigens Shared across patients (mutant KRAS, TP53) Evolutionary constraint HLA-restricted Potential for pre-manufactured therapies TAAs: over-expressed, cancer-testis, differentiation antigens Expressed across multiple tumour subtypes (NY-ESO-1, Claudin(s)) Restricted normal tissue expression Intra- and extracellular Potential for pre-manufactured therapies [ESMO Congress Madrid 2026 logo] [Slide 3] PRAME-directed TCR therapy in ICI-resistant cutaneous and uveal melanoma (anzu-cel) [Embedded ASCO 2026 slide:] Anzu-cel Induced Rapid and Deep Responses in Metastatic PD-1-Relapsed Melanoma and Metastatic Uveal Melanoma | Cutaneous Melanoma n=14 | Uveal Melanoma n=16 | Melanoma (other) n=3 | All Melanoma n=33 cORR | 50% (7/14) | 67% (10/15) | 1/3 | 56% (18/32) ORR | 57% (8/14) | 69% (11/16) | 2/3 | 64% (21/33) DCR | 93% (13/14) | 88% (14/16) | 3/3 | 91% (30/33) [Waterfall plot: Best % Change in Sum of Longest Diameter of Target Lesions from Baseline and BOR (RECIST 1.1); BOR legend PD/SD/PR/CR/Ongoing; per-patient values small] • Median time to BOR was 1.4 months (1.2-2.8) • Most responders exhibit shrinkage of >=1 lesion by first scan • Responses observed in target and non-target lesions [Footnotes: small, largely illegible] Data cutoff Sep 24, 2025 2026 ASCO ANNUAL MEETING #ASCO26; PRESENTED BY: Diwakar Davar M.D., Abstract #95[illegible] Andrew J.S. Furness [ESMO Congress Madrid 2026 logo] [Slide 4] Advancement of cell-based approaches for solid tumours? • Proof-of-principle in terms of objective response and clear room for optimism • Antigen selection - successful targeting should evoke effective induction of epitope spreading • Potential benefit to enrich for neoantigen-reactivity in TIL therapy? • Earlier introduction of therapy to mitigate potential for neoantigen editing and development of resistance to all cell-based approaches reliant upon HLA-TCR engagement (e.g. defects in antigen presentation machinery) • Combination therapies to promote endogenous immunity e.g. immune checkpoint blockade and simultaneously address immune regulation of immune effector cell therapies • Combination therapies to maximise potential for epitope spreading e.g. TCR T (or CAR-T and stromal targeting) + neoantigen vaccination + immune checkpoint blockade Andrew J.S. Furness [ESMO Congress Madrid 2026 logo]

[Slide 1] [ESMO Congress Madrid 2026 logo] Efficacy Lifileucel with Pembrolizumab results in 62.5% ORR and CR rate of 29.2% Best Percentage Change from Baseline in Target Lesion SOD Best Overall Response: PD / SD / PR / CR [Waterfall plot, Change from Baseline (%), patients 1A-25 ... 1A-22; n=2 above zero, n=21 below] Investigator-Assessed Response (RECIST v1.1) (n=24)(a) Objective Response Rate, n (%) | 15 (62.5) (95% CI) | (40.6, 81.2) Disease Control Rate, n (%) | 20 (83.3) (95% CI) | (62.6, 95.3) Best Overall Response, n (%) CR | 7 (29.2) PR | 8 (33.3) SD | 5 (20.8) PD | 3 (12.5) NE | 1 (4.2)(b) 95% CI was calculated using the Clopper-Pearson Exact test. aIOV-COM-202 Cohort 1A Full Analysis Set included all patients who received TIL ≥5.58 billion that met the manufacturing product specifications. bPatient 1A-10 is included as NE and did not have a post-dose tumor response assessment. CI, confidence interval; CR, complete response; NE, not evaluable; ORR, objective response rate; PD, progressive disease; PR, partial response; RECIST, Response Evaluation Criteria in Solid Tumors; SD, stable disease; SOD, sum of diameters; TIL, tumor-infiltrating lymphocyte. [Slide 2] [ESMO Congress Madrid 2026 logo] Efficacy Lifileucel with Pembrolizumab demonstrates durable and deepening responses [Swimmer plot by patient 1A-05, 1A-07, 1A-06, 1A-03, 1A-17, 1A-18, 1A-22, 1A-12, 1A-14, 1A-01, 1A-15, 1A-20, 1A-23, 1A-11, 1A-04; X-axis Time (Months) Since TIL Infusion 0–60] Legend: CR Start; PR Start; Ongoing on Study; Completed without PD or Death; Progression; Pembrolizumab Infusion • mDOR not reached(a) (min, max: 4.3 mo, 57.7+ mo) • Median f/u for DOR(b): 26.9 mo • 3 of 15 responders completed 5-year response assessment w/o PD or death (DOR range, 57+ to 57.7+ mo) • In addition, 3 of 15 responders still ongoing at data cut (DOR range, 25.3+ to 26.9+ mo) [Footnotes, partially cropped/legible:] ...V-COM-202 Cohort 1A Full Analysis Set includes all patients who received ≥5.58 billion TILs meeting manufacturing product specifications. ...ch bar is presented for each patient starting from date of TIL infusion up to date of new anti-cancer therapy, end of assessment, death, or data cutoff date, whichever occurs earlier. Patient 1A-14 was in response during assessment phase while still in response. Patient 1A-12 discontinued from study while still in response to continue pembrolizumab off-study. ...tients 1A-03 and 1A-11 withdrew consent during assessment phase ... before moving to long-term follow-up phase due to relocation from the investigative site. ...ased on Kaplan-Meier product-limit estimates. bBased on reverse Kaplan-Meier method. ...complete response; DOR, duration of response; mDOR, median duration of response; Pembro, pembrolizumab; PD, progressive disease; PR, partial response; TIL, tumor-infiltrating lymphocyte. 8 [Speaker video panel; MADRID 2026 ESMO logo] [Slide 3] [ESMO Congress Madrid 2026 logo] Efficacy Lifileucel with Pembrolizumab results in median PFS of 19.8 months and median PFS2 of 34.3 months Y-axis: Progression-free Survival (%); X-axis: Time (Months) Since Pembrolizumab Infusion (0–78) + Censored Median PFS (95% CI): 19.8 (3.6-NR) Median PFS2 (95%CI): 34.3 (6.8-NR) PFS2 suggests feasibility of systemic therapies in later lines with frontline use of lifileucel Patients at risk (0,3,6,...,78 months): PFS 24 20 16 15 15 14 14 11 9 7 6 6 5 4 4 4 3 3 3 3 1 0 PFS2 24 23 18 17 16 16 15 14 12 10 9 8 7 5 5 5 5 5 4 4 2 1 1 1 1 1 1 0 ...COM-202 Cohort 1A Full Analysis Set includes all patients who received ≥5.58 billion TILs meeting manufacturing product specifications. ...nfidence interval; NR, not reached; PFS, progression-free survival; PFS2, progression-free survival 2; TIL, tumor-infiltrating lymphocyte. 10 [Survival data captured as a table — no point estimates read from the curve] Endpoint: PFS and PFS2 (single arm) Lifileucel + pembrolizumab - PFS | 24 | 19.8 mo | 3.6-NR Lifileucel + pembrolizumab - PFS2 | 24 | 34.3 mo | 6.8-NR Note: Single-arm cohort; no comparator. Time measured from pembrolizumab infusion. Number-at-risk rows partially small/blurry; transcribed as read. [Slide 4] [ESMO Congress Madrid 2026 logo] Safety: Treatment-Emergent Adverse Events in ≥30% of Patients(a) Safety of lifileucel with pembrolizumab consistent with disease and safety profiles of study regimen N=27 | Any grade | Grade 3/4 Preferred Terms, n (%) Thrombocytopenia | 24 (88.9) | 18 (66.7) Pyrexia | 22 (81.5) | 4 (14.8) Chills | 21 (77.8) | 3 (11.1) Nausea | 21 (77.8) | 0 Anemia | 19 (70.4) | 13 (48.1) Vomiting | 17 (63.0) | 0 Neutropenia | 15 (55.6) | 15 (55.6) Fatigue | 14 (51.9) | 1 (3.7) Headache | 13 (48.1) | 0 Cough | 12 (44.4) | 0 Decreased appetite | 12 (44.4) | 0 Diarrhea | 12 (44.4) | 2 (7.4) Febrile neutropenia | 11 (40.7) | 10 (37.0) Rash (maculopapular) | 11 (40.7) | 5 (18.5) Alopecia | 10 (37.0) | 0 Dyspnea | 10 (37.0) | 2 (7.4) Hypokalemia | 9 (33.3) | 0 Safety was consistent with the underlying disease and known safety profiles of pembrolizumab, NMA-LD, lifileucel, and IL-2 • AEs consistent with the lifileucel regimen occurred and resolved early • AEs occurring later than 30 days after lifileucel infusion were generally consistent with the safety profile of pembrolizumab monotherapy N=27 (safety analysis set includes all patients who received lifileucel). One patient had a grade 5 TEAE of sepsis. aTEAEs refer to adverse events that occurred from the first dose of pembrolizumab or lifileucel infusion (whichever occurred first) up to 30 days after the last dose of pembrolizumab or lifileucel infusion (whichever occurred later) or up to the start of a new anti-cancer therapy. AE, adverse event; IL-2, interleukin-2; NMA-LD, non-myeloablative lymphodepletion; TEAE, treatment-emergent adverse event. 12

[Slide 1] [Screen with speaker video of Toni K. Choueiri; MADRID ESMO congress logo] How does an mRNA personalized vaccine wo[rk] [cropped] 5 Vaccination 6 Vaccine Uptake by Antigen-Presenting Cells 7 mRNA Translation and Neoantigen Presentation 8 Priming and Expansion of Neoantigen-Sp[ecific ...] [cropped] Toni K. Choueiri, MD Created with BioRender.com [Podium sign: MADRID SPAIN 23-27 OCTOBER 2026] [Slide 3] [Screen with speaker video of Toni K. Choueiri; MADRID 2026 ESMO congress logo] [Left edge, cropped: Moderna COVID-19 Vaccine vial; mRNA-1273; 7 JAN 2021; photo of vaccination] Accelerated research and development timeline for COVID-19 SARS-CoV-2 (mRNA-1273) Isolation/Sequencing: Viral sequence isolated from infected patient 2 DAYS Antigen design: Spike protein designed to elicit robust immune response mRNA design 42 DAYS Vaccine manufacturing + Quality testing 9 MONTHS Clinical trials FDA Emergency use authorization Total Time: 11 MONTHS [Podium sign: MADRID SPAIN 23-27 OCTOBER 2026] [Slide 4] [Screen with speaker video of Toni K. Choueiri; MADRID 2026 ESMO congress logo] Introduction: Evolutio[n ...] [cropped] From Early Concepts to Personali[zed ...] [cropped] 1893: The concept of immunotherapy inspired by observations of spontaneous cancer regressions after infections by William Coley 1968: Lloyd Old discovered that Bacillus Calmette-Guerin prolongs the survival of mice with cancer, ushering in the era of immunotherapy 1984: Hoover's pioneering use of a vaccine made from tumor cell lysates increased responses to autologous tumor cells in patients with colorectal cancer 1991: The first tumor-associated antigen (TAA) MAGE-1 were identified and clinical trials of TAA-based cancer vaccine were conducted since then 1999: Melacine was the first approved-to-market tum[or] lysate vaccine against melanoma [Next item cropped: Pepti[de] vacci[ne] strat[egies] ...] Toni K. Choueiri, MD

[Slide 1] STARBOARD PHASE 3 STUDY DESIGN A randomized, double-blind, placebo-controlled, phase 3 study evaluating encorafenib and binimetinib plus pembrolizumab vs placebo plus pembrolizumab in patients with previously untreated metastatic or unresectable locally advanced BRAF V600E/K melanoma Patient population - Documented evidence of BRAF V600E/K mutation - Unresectable or metastatic cutaneous melanoma - No prior systemic treatment for metastatic or locally advanced disease - Measurable disease (RECIST 1.1) - ECOG PS <=1 R^a 1:1 N~216 Triplet^b (n~108): Encorafenib 450 mg QD + binimetinib 45 mg BID + pembrolizumab 200 mg Q3W Control^c (n~108): Placebo + pembrolizumab 200 mg Q3W Primary endpoint: ORR by BICR^d Key secondary endpoint: PFS by BICR^e Other secondary endpoints: OS, PFS and ORR by investigator, DOR, TTR, QoL, Safety, and PK Stratification: - Prior systemic adjuvant therapy - CPI or BRAF/MEKi: yes vs no - Disease stage - IIIB, IIIC, IIID, IV M1a[0], IV M1b[0] vs IV M1a[1], IV M1b[1], IV M1c[0], IV M1c [1], IV M1d [0] and IV M1d [1] We present here the final analysis of both ORR and PFS by BICR, as well as OS, and safety from the phase 3 portion of STARBOARD NCT04657991 ^aRepresents planned sample size. Patients were randomized between June 21, 2022, and January 16, 2024. The study design was amended due to recruitment issues, resulting in a reduction in the planned sample size and a revision of primary and key secondary endpoints. ^bStudy treatment was administered in 21-day cycles until patient had disease progression or met any of the other protocol-defined criteria for discontinuation. ^cPatients received a maximum of 35 administrations of pembrolizumab 200 mg every 3 weeks with placebos for both encorafenib and binimetinib. ^dORR was calculated with corresponding 2-sided Wilson score 95% CI; the treatment effect was tested at a 1-sided significance level of 0.05. ^eThe difference in PFS was estimated using a stratified Cox proportional hazards model to calculate the hazard ratio, along with the corresponding 2-sided 90% CI; the treatment effect was tested at a 1-sided significance level of 0.05. The Kaplan-Meier method was used to estimate median PFS for each arm, with the corresponding 95% CIs calculated using the Brookmeyer-Crowley method. Pembrolizumab was limited to a maximum of 35 administrations and treatment with triplet therapy continued until disease progression [footnote order approximate; small angled print] BICR, blinded independent central review; BID, twice daily; CPI, immune checkpoint inhibitor; DOR, duration of response; ECOG PS, Eastern Cooperative Oncology Group performance status; MEKi, mitogen-activated protein kinase kinase inhibitor; PK, pharmacokinetics; QD, once daily; Q3W, every 3 weeks; QoL, quality of life; RECIST, Response Evaluation Criteria in Solid Tumors; TTR, time to response Dirk Schadendorf, MD [ESMO Congress Madrid 2026 logo] [Slide 4] SAFETY SUMMARY No new safety signals and no grade 5 treatment-related TEAEs were observed Patients, n (%) | Triplet (n=110) | Control (n=110) TEAE | 110 (100.0) | 104 (94.5) Treatment related | 106 (96.4) | 81 (73.6) Maximum grade 3 or 4 TEAE | 77 (70.0) | 43 (39.1) Treatment related | 63 (57.3) | 24 (21.8) Grade 5 TEAE | 5 (4.5) | 10 (9.1) Treatment related | 0 | 0 Serious TEAE | 58 (52.7) | 35 (31.8) Treatment related | 35 (31.8) | 11 (10.0) TEAE leading to pembrolizumab interruption | 76 (69.1) | 42 (38.2) TEAE leading to pembrolizumab discontinuation | 35 (31.8) | 15 (13.6) TEAE leading to encorafenib/placebo dose reduction or interruption | 91 (82.7) | 51 (46.4) TEAE leading to encorafenib/placebo discontinuation | 47 (42.7) | 16 (14.5) TEAE event leading to binimetinib/placebo dose reduction or interruption | 92 (83.6) | 49 (44.5) TEAE leading to binimetinib/placebo discontinuation | 47 (42.7) | 16 (14.5) [A] larger proportion of grade 3 or 4 TEAEs and discontinuations due to TEAEs were observed in the triplet arm than in the control arm [Data cut]off: January 12, 2026. [TEAE, treat]ment-emergent adverse event [Dirk Scha]dendorf, MD

[Slide 1] [ESMO Congress Madrid 2026 logo] IOV-COM-202: Phase 2, Multicohort, Multicenter Study of Lifileucel + Pembrolizumab in Patients with Solid Tumors Cohort 1A of IOV-COM-202 (NCT03645928) assesses the efficacy and safety of lifileucel + pembrolizumab in patients with ICI-naïve unresectable or metastatic melanoma • Patients may have received BRAF/MEK inhibitor treatment if they are BRAF mutation positive • Patients may have received no prior therapies Eligible patients must have: • ≥1 resectable lesion (≥1.5 cm diameter) and • ≥1 measurable lesion for response assessment per RECIST v1.1 Trial was designed as a proof-of-concept study to support a registrational study in the frontline treatment setting STUDY SCHEMA GMP Manufacturing 1 Enrollment/ Surgical Resection 2 Pembrolizumab(a) 1st dose 3 NMA-LD Day -7 to -6: CY Day -5 to -1: FLU 4 Lifileucel Infusion 1 dose Day 0 5 IL-2 ≤6 doses Day 0 or 1 to Day 3 or 4 6 Pembrolizumab Q3W or Q6W(a) 7 Efficacy Follow up (a)First administration of single-dose pembrolizumab IV 200 mg or 400 mg, followed by pembrolizumab IV 200 mg Q3W or 400 mg Q6W for 24 months or until disease progression or unacceptable toxicity. BRAF, B-Raf proto-oncogene; CY, cyclophosphamide; FLU, fludarabine; GMP, Good Manufacturing Practice; ICI, immune checkpoint inhibitor; IL-2, interleukin-2; IV, intravenous; MEK, mitogen-activated extracellular signal-regulated kinase; NMA-LD, non-myeloablative lymphodepletion; Q3W, every 3 weeks; Q6W, every 6 weeks; RECIST, Response Evaluation Criteria in Solid Tumors. 4

[Slide 1] Phase 1 Multicenter Trial of IMA203CD8 in PRAME+ Solid Tumors Patient Journey SCREENING/MANUFACTURING | TREATMENT/OBSERVATION | FOLLOW UP HLA-A*02:01 Testing (blood) PRAME Testing – Biopsy or archived tissue – No longer required for indications with high PRAME prevalence, including ovarian and uterine cancer, melanoma, and synovial sarcoma Leukapheresis IMA203CD8 Manufacturing (~2 weeks) Lymphodepletion: Fludarabine 30 mg/m² Cyclophosphamide 500 mg/m² days -6 to -3 IMA203CD8 one-time infusion ± Low Dose IL-2ª: 1M IU SUBQ QD, days 3-5, then 1M IU SUBQ BID, days 6-10 Key Objectives Primary: • Tolerability • Determination of RP2D Secondary: • Efficacy • Pharmacokinetics Key Eligibility Criteria • Advanced or metastatic solid tumors • Age ≥18 years • ECOG PS 0-1 • HLA-A*02:01 positive • PRAME positive • No available SOC treatment options • Measurable disease (RECIST 1.1) • Adequate organ function Dose Escalation Scheme ± Low Dose IL-2ª DL3 (n=7) 0.44-1.00×10⁹ total TCR T cells DL4a (n=38) 0.93-1.86×10⁹ total TCR T cells DL4b (n=4) 1.49-2.05×10⁹ total TCR T cells DL4c (n=9) 1.15-3.06×10⁹ total TCR T cells DL5 (n=7) 1.95-4.30×10⁹ total TCR T cells DL6 (n=15) 3.55-8.22×10⁹ total TCR T cells DL7 (n=17) 4.26-12.50×10⁹ total TCR T cells Dose escalation completed Enrichment at relevant DLs (≥DL4c): ~1.2–12.5 billion total TCR T cells Data cutoff: August 7, 2026 [Footnotes small/partly illegible: IMA203-101: NCT03686124; ... ªDose level ≥DL4c is evaluated ±IL-2 ...; anzu-cel, anzutresgene autoleucel ... [illegible]] Prof. Dr. med. Antonia Busse, abstract #7231 [ESMO Congress Madrid 2026 logo] 4 [Slide 2] Tolerability in Advanced Solid Tumors TEAEs in ≥20% of Patients (N=98ª) Preferred term, n (%) | Any grade | Grade ≥3 Patients with any adverse event | 98 (100) | 97 (99) Neutropenia | 88 (90) | 87 (89) Anemia | 78 (80) | 54 (55) Thrombocytopenia | 71 (72) | 36 (37) Nausea | 64 (65) | 2 (2) Lymphopenia | 45 (46) | 45 (46) Rash | 39 (40) | 5 (5) Fatigue | 37 (38) | 6 (6) Transaminase elevation | 37 (38) | 12 (12) Hypokalemia | 34 (35) | 2 (2) Constipation | 33 (34) | 0 Vomiting | 28 (29) | 0 Abdominal pain | 26 (27) | 3 (3) Leukopenia | 25 (26) | 23 (23) Hyponatremia | 23 (23) | 2 (2) Pyrexia | 23 (23) | 0 Hypophosphatemia | 21 (21) | 3 (3) Diarrhea | 20 (20) | 3 (3) Headache | 20 (20) | 1 (1) All TEAEs occurring in at least 20 patients (≥20%) are presented. Grades were determined according to National Cancer Institute Common Terminology Criteria for Adverse Events, version 5.0. Grades for CRS and ICANS were determined according to CARTOX criteria (Neelapu et al, 2018 before protocol v11.0; Neelapu et al, from protocol v11.0 onward) Adverse Events of Special Interest (N=98ª) CRS, any grade, n (%) | 94 (96) Grade 1 | 38 (39) Grade 2 | 47 (48) Grade 3 | 8 (8) Grade 4 | 1 (1) HLH, any grade, n (%) | 7 (7) Grade 1 | 0 Grade 2 | 4 (4) Grade 3 | 2 (2) Grade 4 | 1 (1) ICANS, any grade, n (%) | 8 (8) Grade 1 | 5 (5) Grade 2 | 1 (1) Grade 3 | 2 (2) Late N- ICAHT Beyond Day 28* (N=87) ICAHT, any grade, n (%) | 20 (23) Grade 1-2 | 15 (17) Grade 3-4 | 5 (6) Late neutropenia-based Immune Effector Cell-Associated Hematotoxicity (N-ICAHT) grade was based on the second-lowest absolute neutrophil count between days 28-100 postinfusion and required ≥2 neutropenic measurements during this period, including ≥1 between days 35-60. Measurements before the day 28 visit were not used for late ICAHT grading but were considered when determining neutropenia onset and duration. Day 28 visits could have occurred ±3 days before or after day 28. Assessments were available for n=87 patients. • Most frequent TEAEs were cytopenias associated with lymphodepletion • CRS was mostly grade 1-2 • DLTs in 4 of 98 patients: 1. DL4b: grade 4 neurotoxicity/CRS with grade 3 HLH 2. DL4b: grade 3 CRS 3. DL5: grade 3 ICANSᵇ 4. DL7: grade 4 skin infectionᵇ • Late grade 3-4 N-ICAHT in 5 patients was manageable with supportive treatment; all improved to grade ≤2 at a median of 69 days postinfusion (range, 42–121) • No IMA203CD8-related grade 5 eventᶜ • MTD not reached with doses up to 12.5 billion IMA203CD8 cells ª All patients who started lymphodepletion, including 1 patient who did not receive IMA203CD8 cells. ᵇ 2 events met DLT criteria but did not limit dose escalation: 1 in an enrichment cohort at DL5 and 1 at highest dose, DL7 (1/6 patients), both within protocol-defined dose-finding rules). ᶜ Possibly related grade 5 event as previously reported was determined by the PI to be unlikely related to IMA203CD8 after complete assessment. Patient died from sepsis that was aggravated by immunosuppression from fludarabine/cyclophosphamide (possibly related), a grade 4 HLH event, the toxicity management, and rapidly progressing disease. CARTOX, CAR T-cell therapy–associated toxicity; CRS, cytokine release syndrome; DL, dose level; DLT, dose-limiting toxicity; ICAHT, immune effector cell–associated hematotoxicity; HLH, hemophagocytic lymphohistiocytosis; ICANS, immune effector cell–associated neurotoxicity syndrome; PI, principal investigator; MTD, maximum tolerated dose; TEAE, treatment-emergent adverse event. Data cutoff: August 7, 2026 Prof. Dr. med. Antonia Busse, abstract #7231 [ESMO Congress Madrid 2026 logo] [Slide 3] Tumor Reduction Across Dose Levels and Indications During Dose Escalation Objective responses across all dose levels and biologically distinct tumor types, including ovarian and uterine cancer, melanoma, and synovial sarcoma [Waterfall plot: Best % Change in Sum of Longest Diameter of Target Lesions from Baseline (RECIST 1.1); per-patient tumor-type labels on x-axis (ovarian, uterine, melanoma, uveal melanoma, synovial sarcoma, NSCLC, breast, etc.) - too small to transcribe individually] N | 90 ORR | 51% (46/90) cORR | 41% (36/87) DCR (at week 6) | 82% (74/90) Tumor shrinkage | 83% (71/86) BOR (RECIST 1.1): PD, SD, PR, cPR, CR, cCR ▶ Ongoing response/disease control * Confirmed response based on prior PR that deepened to CR Data cutoff: August 7, 2026 [Footnotes small, partly illegible] Prof. Dr. med. Antonia Busse, abstract #7231 [ESMO Congress Madrid 2026 logo] [Slide 4] Tumor Reduction Across Dose Levels During Dose Escalation by Indication | | Gynecologic Indications (N=27) | Melanoma (N=44) | Synovial Sarcoma (N=12) | Other (N=7) | | Median dose [×10⁹ total TCR T cells] | 3.55 | 1.67 | 1.59 | 1.95 | | ORR | 48% (13/27) | 52% (23/44) | 67% (8/12) | 29% (2/7) | | cORR | 41% (11/27) | 41% (17/42) | 67% (8/12) | 0% (0/6) | | DCR | 70% (19/27) | 82% (36/44) | 100% (12/12) | 100% (7/7) | | Tumor shrinkage | 68% (17/25) | 88% (37/42) | 92% (11/12) | 86% (6/7) | [Waterfall plots by indication: Best % Change in Sum of Longest Diameter of Target Lesions from Baseline (RECIST 1.1)] BOR (RECIST 1.1): PD, SD, PR, cPR, CR, cCR ▶ Ongoing response/disease control * Confirmed response based on prior PR that deepened to CR Data cutoff: August 7, 2026 [Footnotes small, partly illegible] Prof. Dr. med. Antonia Busse, abstract #7231 [ESMO Congress Madrid 2026 logo] 11

[Slide 1] Study Design: MIZAR-003 Treatment Schema 5 Key Eligibility Criteria: • Aged 18 to 75 years; • Histologically confirmed unresectable or metastatic melanoma; • ≥2 prior systemic therapy, anti-PD-1 failure; • Feasible tumor tissue for resection (≥150mm3); • At least one remaining measurable target lesion. N=98, 1:1 GC101 TIL arm: D0 Tumor Resection → D2-D10 Bridging Therapy (As Needed) → Radiological assessment → Preconditioning: D37-39: CTX 20mg/kg/d, D37: HCQ 600mg/d → D42 GC101 TIL infusion → Sintilimab, 100mg, q6w (up to 5 doses) → D84 Radiological assessment q6w → PD PFS (Primary Analysis) from D0; PFS (Sensitive Analysis) from D37-39 ICC arm: Investigator's Choice of Chemotherapy: every 3 weeks - Dacarbazine, temozolomide, paclitaxel, cisplatin, carboplatin → D42 Radiological assessment → D84 → PD PFS (Primary Analysis / Sensitive Analysis) Eligible patients cross over to GC101 TIL therapy after IRC-verified initial PD Study Endpoints Primary endpoint: • PFS (Primary analysis) by IRC (RECIST v1.1) Secondary endpoints: • PFS by Investigator (primary and sensitivity analyses) • ORR, DCR, DOR assessed by both IRC and Investigator (primary and sensitivity analyses) • OS • Safety CTX, cyclophosphamide; DCR, disease control rate; DOR, duration of response; HCQ, hydroxychloroquine; ICC, investigator's choice of chemotherapy; IRC, independent review committee; ORR, objective response rate; OS, overall survival; PD, progressive disease; PFS, progression-free survival; RECIST, Response Evaluation Criteria in Solid Tumors; TIL, Tumor-Infiltrating Lymphocytes. [ESMO Congress Madrid 2026 logo] [Slide 2] Results: Baseline Characteristics 7 • A total of 99 pts were enrolled, over 70% were acral or mucosal melanomas. • As data cutoff, 43 pts in Primary analysis of the GC101 TIL group received TIL infusion (ASCO 2026: 35 pts). • The median infused cell dose was 42.90×10^9 (range, 4.62-53.92×10^9). Characteristic | GC101 TIL (N=50) | ICC (N=49) | Total (N=99) Age (years, median, range) | 58 (29~73) | 59 (28~73) | 58 (28~73) Sex, n (Female, %) | 28(56.0) | 25(51.0) | 53(53.5) Melanoma subtype, n (%) Acral | 31(62.0) | 25(51.0) | 56(56.6) Cutaneous | 12(24.0) | 12(24.5) | 24(24.2) Mucosal | 6(12.0) | 10(20.4) | 16(16.2) Unknown primary | 1(2.0) | 2(4.1) | 3(3.0) Visceral metastasis, n (%) Lung | 27(54.0) | 31(63.3) | 58(58.6) Liver | 9(18.0) | 8(16.3) | 17(17.2) Lymph node | 37(74.0) | 37(75.5) | 74(74.7) Bone | 8(16.0) | 6(12.2) | 14(14.1) Characteristic | GC101 TIL (N=50) | ICC (N=49) | Total (N=99) ECOG performance status, n (%) 1 | 27(54.0) | 37(75.5) | 64(64.6) 0 | 23(46.0) | 12(24.5) | 35(35.4) Gene mutation, n (%) BRAF mutated | 5(10.0) | 2(4.1) | 7(7.1) NRAS mutated | 6(12.0) | 5(10.2) | 11(11.1) Target Lesions Sum of Diameter (mm, Mean ± SD) | 54.92±26.326 | 63.99±29.852 | 59.60±28.418 Number of prior systemic therapy lines (n, median, range) | 2(2,5) | 2(2,5) | 2(2,5) Lactate dehydrogenase (LDH), n (%) ≤ ULN | 37(74.0) | 36(73.5) | 73(73.7) > ULN | 13(26.0) | 13(26.5) | 26(26.3) ECOG, eastern cooperative oncology group; ICC, investigator's choice of chemotherapy; LDH, lactate dehydrogenase; SD, standard deviation; TIL, tumor-infiltrating lymphocytes; ULN, upper limit of normal. Updated Data cutoff date: July 2, 2026 (ASCO 2026 Data cutoff date: May 12, 2026) [ESMO Congress Madrid 2026 logo] [Slide 3] Results: GC101 TIL Survival vs. ICC 8 • During extended follow-up, PFS benefit maintained with longer follow-up. • Median PFS*: 4.37 vs. 1.61 months, HR = 0.44, p = 0.0004 (ASCO 2026: 4.30 vs. 1.58 months, HR=0.43, p=0.0002). • Even with 59% (ASCO 2026: 57%) chemo cross-over, a survival benefit trend showed in the TIL group. PFS - Primary analysis (GC101-TIL VS ICC): Event/N: 32/50 43/49; mPFS(95%CI): 4.37(2.99-5.55) 1.61(1.28-2.76); HR(95%CI): 0.44(0.28-0.70); P value: 0.0004. PFS Rate(%) at 6 month 34.48 VS 10.24 No. at Risk (months 0,2,4,6,8,10,12,14): GC101-TIL 50 35 26 9 5 3 3 0; ICC 49 21 12 4 3 1 1 0 PFS - Sensitive analysis (GC101-TIL VS ICC): Event/N: 22/37 42/47; mPFS(95%CI): 4.31(2.79-5.72) 1.54(1.28-2.66); HR(95%CI): 0.44(0.26-0.74); P value: 0.0013. PFS Rate(%) at 6 month 29.73 VS 10.05 No. at Risk (0,2,4,6,8,10,12,14): GC101-TIL 37 28 15 4 4 3 2 0; ICC 47 19 12 4 3 1 1 0 OS§ - Primary analysis (GC101-TIL VS ICC): Event/N: 9/50 15/49; mOS(95%CI): NE(9.49-NE) 12.65(8.74-NE); HR(95%CI): 0.6?(0.26-1.41) [second digit illegible]; P value: 0.2571 No. at Risk (0,2,4,6,8,10,12,14,16): GC101-TIL 50 47 38 28 20 11 7 1 0; ICC 49 49 41 30 21 16 12 7 0 *IRC-assessed; §Not yet mature CI, confidence interval; HR, hazard ratio; ICC, investigator's choice of chemotherapy; OS, overall survival; PFS, progression-free survival; TIL, tumor-infiltrating lymphocytes. Updated Data cutoff date: July 2, 2026 (ASCO 2026 Data cutoff date: May 12, 2026) [ESMO Congress Madrid 2026 logo] [Survival data captured as a table — no point estimates read from the curve] Endpoint: PFS (IRC, primary analysis); also sensitivity PFS and OS GC101 TIL | 50 | 4.37 mo | 2.99-5.55 ICC | 49 | 1.61 mo | 1.28-2.76 HR: 0.44 HR CI: 0.28-0.70 P: 0.0004 {'timepoint': '6 months PFS', 'arm': 'GC101 TIL', 'value': '34.48%', 'ci': None} {'timepoint': '6 months PFS', 'arm': 'ICC', 'value': '10.24%', 'ci': None} {'timepoint': '6 months PFS (sensitivity analysis)', 'arm': 'GC101 TIL', 'value': '29.73%', 'ci': None} {'timepoint': '6 months PFS (sensitivity analysis)', 'arm': 'ICC', 'value': '10.05%', 'ci': None} Note: Sensitivity PFS: 4.31 (2.79-5.72) vs 1.54 (1.28-2.66), HR 0.44 (0.26-0.74), p=0.0013, events 22/37 vs 42/47. OS not mature: mOS NE (9.49-NE) vs 12.65 (8.74-NE), events 9/50 vs 15/49, HR 0.6? (0.26-1.41) second digit illegible, p=0.2571. Updated DCO July 2, 2026; prior ASCO 2026 DCO May 12, 2026 gave 4.30 vs 1.58, HR 0.43, p=0.0002. 59% chemo crossover.

[Slide 2] Th17 therapy induces PC, GC and memory B cell proliferation [Schema] C57BL/6 — B16F10 (day -7) — 4 Gy (day -1) — Th17 ACT (day 0) — Tissue analysis (day 12) C [Histograms, % Max vs Ki67] Th17 ACT; No Treatment Total B cells (CD19+B220+); Plasma Cell; GC B cell; Memory B Cell D [Bar plots, % Ki67+, NT vs Th17] Total B cells (CD19+B220+) **** Plasma Cell **** GC B cell **** Memory B Cell **** Cole A.C, Knochelmann H.M., et. al., Cancer Cell 2025 [Slide 4] ↑ B cell proliferation/differentiation ↑ GC Formation ↑ MHC-II CD40 B Th17 (CD26+) ↑ IL-21 ↑ Inflammatory cytokines Tumor-specific antibody secretion Ab-mediated tumor destruction Sustained Tumor Killing TRM CD8 — Tissue-resident memory CD8 (TRM CD8) — Direct tumor killing TRM CD4 (Th17) — Tissue-resident memory CD4 (Th17) (TRM CD4) — Local cytokine production & tumor control

[Slide 1] ICI – Clinical data, mothers • Eleven case reports1,2 & thirteen database cases3 o Fifteen conceived while on ICI treatment o Nine initiated treatment during pregnancy (between week 9 and 26) o Two twin pregnancies • Initiated during pregnancy o Some responses during pregnancy, however majority progresses during pregnancy or shortly after delivery • Conceived on therapy o Not recommended during and for at least 5 months after last dose of ICI4 o Generally better prognosis; majority already in complete remission or NED after completion of therapy o Possibility to discontinue ICI treatment in certain cases with good outcomes • All current information from anti-PD-1 and anti-CLTA-4; no information on anti-LAG-3 [Figure: Borgers & Heimovaara et al. Lancet Oncol. 2021 — bar chart of treatment by weeks of gestation (0-40)] Legend: No treatment | Ipilimumab plus IL-2 | Nivolumab | Ipilimumab plus nivolumab | Pembrolizumab | No congenital abnormalities child | Minor congenital abnormalities child Rows: Menzer et al (2018) | Mehta et al (2018) | Burotto et al (2018) | Xu et al (2019) | Bucheit et al (2020) | Haiduk and Ziemer (2021) | Polnaszek et al (2021) X axis: Weeks of gestation Jessica S.W. Borgers, MD References (partially cropped): ...m J Hematol (2022); 2 Baarslag, Heimovaara, Borgers et al. NEJM (2023); 3 Borgers et al. ECM (2025); 4 Prescribing information [Slide 2] ICI – Clinical data, children • >50% of cases C-sections because of intrauterine growth restriction and/or placental insufficiency o Majority born prematurely • (Potential) ir-toxicities o First trimester exposure: two children with congenital hypothyroidism, spontaneously recovered within 6 months o Second and third trimester exposure: pembrolizumab-induced immune-related gastro-enterocolitis (presented at 15w), 1mg/kg prednisolone and infliximab for first [partially obscured: 2 years of life]1 • At most recent follow-up (6 weeks to 2.75 years) o Healthy o Normal developmental milestones o No evidence of disease [Figure: Borgers & Heimovaara et al. Lancet Oncol. 2021 — same gestation-week bar chart as previous slide: Menzer et al (2018), Mehta et al (2018), Burotto et al (2018), Xu et al (2019), Bucheit et al (2020), Haiduk and Ziemer (2021), Polnaszek et al (2021); Weeks of gestation 0-40] Jessica S.W. Borgers, MD 1 Baarslag, Heimovaara, Borgers et al. NEJM (2023) [Slide 3] Immune-related adverse events (irAEs) Mother • irAEs events are common, especially when receiving combination therapy1 • irAEs often treated with steroids o Up to 20 mg/day considered safe during pregnancy o 1-2 mg/kg associated with an increased risk of gestational diabetes, fetal oral–facial clefts (first trimester exposure), and hypothalamic-pituitary-adrenal axis dysfunction2 o Most 2nd or 3rd line immunosuppressive agents other than infliximab (category B) are classified as pregnancy category C or D3 Fetus • Weight of ICI exceeds the limit for simple diffusion -> active transport using neonatal Fc receptors o Most effective for IgG1, closely followed by IgG4 isotypes4 o Not present during the first 14 weeks: protected during organogenesis o Highest rate of transplacental transmission occurs in late second and third trimesters4 [Figure: Martins et al. Nat Rev Clin Onc (2019) — body diagram of irAEs by organ] Jessica S.W. Borgers, MD 1 Martins et al. Nat Rev Clin Onc (2019); 2 Ponticelli et al. J Clin Med (2018); 3 Ponticelli. J Clin Med (2018); 4 Pentsuk et al. Birth Defects Res B Dev Reprod Toxicol (2009) [Slide 4] Targeted therapy • Four published case reports1-4, seven cases in our database • Mothers: o Two women conceived while receiving dabrafenib/trametinib and vemurafenib (abortions) o Nine women initiated treatment during second trimester (week 17, 18, 20, 22, 26) o Six women received vemurafenib, three received dabrafenib o Most women SD/PR during pregnancy, then progressed after delivery o One severe side effect: toxic epidermal necrolysis (ICU admission with intubation)2 • Children: o All born prematurely (week 26, 30, 32, 34, 36) o Fetal growth restriction o Cardiac toxicity? ❖ Neonatal paroxysmal supraventricular tachycardia (treated with beta-blocker) ❖ Atrial septal defect, patent ductus arteriosus, and Wolff-Parkinson-White syndrome o Normal growth and development Jessica S.W. Borgers, MD 1 Maleka et al. (2019); 2 De Haan et al. (2018); 3 Marce et al. (2019); 4 Pagan et al. (2019)

[Slide 1] [MADRID 2026 ESMO congress logo] Targeting PRAME+ Tumors with IMA203CD8, a PRAME-directed TCR T-cell Therapy: Results from a Phase 1 Study Antonia Busse1, Dejka M. Araujo2, Winfried Alsdorf3, Oladapo Yeku4, Tobias A.W. Holderried5,6, Leonel F. Hernandez-Aya7, Sarah Taylor8, Dirk Jaeger9, Manik Chatterjee10, Anthony J. Olszanski11, Eva M. Wagner-Drouet12, Peter Herhaus13, Katrin Wetzko14, Apostolia M. Tsimberidou2, Friederike Schmitz8, Diwakar Davar15, Karine Pozo16, Liane Preussner17, Cedrik M. Britten17, Martin Wermke18 1Charité Medical University Hospital, Berlin, Germany; 2The University of Texas MD Anderson Cancer Center, Houston, TX, USA; 3Department of Hematology and Oncology, University Hospital Hamburg Eppendorf, Hamburg, Germany; 4Massachusetts General Hospital, Boston, MA, USA; 5University of Bonn, University Hospital Bonn, Medical Clinic and Polyclinic III Internal Medicine with a focus on Oncology, Hematology, and Rheumatology, Bonn, Germany; 6Clinical Division of Hematology, Department of Internal Medicine, Medical University of Graz, Graz, Austria; 7University of Miami, Miami, FL, USA; 8University of Pittsburgh Cancer Institute, Pittsburgh, PA, USA; 9University Hospital Heidelberg, Heidelberg, Germany and Translational Oncology/Early Clinical Trial Unit (ECTU); 10Comprehensive Cancer Center Mainfranken, University Hospital Würzburg, Würzburg, Germany; 11Department of Hematology/Oncology, Fox Chase Cancer Center, Philadelphia, PA, USA; 12University Medical Center of the Johannes Gutenberg University Mainz, Mainz, Germany; 13TUM School of Medicine and Health, Department of Internal Medicine III, University Medical Center, Technical University of Munich, Germany; 14University Hospital Dresden, Dresden, Germany; 15University of Pittsburgh Medical Center (UPMC), Hillman Cancer Center, Pittsburgh, PA, USA; 16Immatics US, Houston, TX, USA; 17Immatics Biotechnologies GmbH, Tuebingen, Germany; 18TU Dresden, Carl-Gustav-Carus Faculty of Medicine, NCT/UCC Early Clinical Trial Unit, Dresden, Germany. Prof. Dr. med. Antonia Busse Proffered paper presentation, abstract #7231 [QR code] [Madrid skyline graphic] [ESMO logo] [Stage photo: presenter at ESMO podium (MADRID SPAIN 23-27 OCTOBER 2026), two chairs seated] [Slide 2] Tumor Reduction Across Dose Levels and Indications During Dose Escalation Objective responses across all dose levels and biologically distinct tumor types, including ovarian and uterine cancer, melanoma, and synovial sarcoma N 90 ORR 51% (46/90) cORR 41% (36/87) DCR (at week 6) 82% (74/90) Tumor shrinkage 83% (71/86) [Waterfall plot; y-axis: Best % Change in Sum of Longest Diameter of Target Lesions from Baseline (RECIST 1.1), 25 to -100; per-patient bars labelled with % change from +24 to -100; x-axis labels = tumor types per patient (uterine cancer, acral melanoma, ovarian cancer, uveal melanoma, synovial sarcoma, lung cancer (NSCLC), cutaneous melanoma, breast cancer, mucosal melanoma, melanoma (UNK), etc.)] BOR (RECIST 1.1): PD; SD; PR; cPR; CR; cCR ▶ Ongoing response/disease control * Confirmed response based on prior PR that deepened to CR Data cutoff: August 7, 2026 n=2 patients with ovarian cancer and n=1 patient with melanoma died before first postbaseline scan and n=1 additional patient with melanoma had a non-evaluable target lesion; these patients are not depicted in the plot but were included in the ORR, cORR and DCR calculation. ORR: according to RECIST 1.1 at any postbaseline scan, PD or death at any prior timepoint; Confirmed ORR for patients with ≥2 postbaseline scans per RECIST 1.1, PD or death at any prior timepoint, those with ongoing unconfirmed PR/CR were excluded. For patients achieving a (c)CR with <100% change from baseline, target lesions were lymph nodes that resolved to <10 mm per RECIST 1.1. BOR, best overall response; (c)CR, (confirmed) complete response; DCR, disease control rate; NSCLC, non-small cell lung cancer; (c)ORR, (confirmed) objective response rate; PD, progressive disease; (c)PR, (confirmed) partial response; SD, stable disease; UNK, unknown. Prof. Dr. med. Antonia Busse, abstract #7231 [MADRID 2026 ESMO congress logo] 10 [Slide 3] Tumor Reduction Across Dose Levels During Dose Escalation by Indication | Gynecologic Indications (N=27) | Melanoma (N=44) | Synovial Sarcoma (N=12) | Other (N=7) Median dose [×10^9 total TCR T cells]: 3.55 | 1.67 | 1.59 | 1.95 ORR: 48% (13/27) | 52% (23/44) | 67% (8/12) | 29% (2/7) cORR: 41% (11/27) | 41% (17/42) | 67% (8/12) | 0% (0/6) DCR: 70% (19/27) | 82% (36/44) | 100% (12/12) | 100% (7/7) Tumor shrinkage: 68% (17/25) | 88% (37/42) | 92% (11/12) | 86% (6/7) [Waterfall plots per indication; y-axis: Best % Change in Sum of Longest Diameter of Target Lesions from Baseline (RECIST 1.1), 25 to -100; Baseline line] BOR (RECIST 1.1): PD; SD; PR; cPR; CR; cCR ▶ Ongoing response/disease control * Confirmed response based on prior PR that deepened to CR Data cutoff: August 7, 2026 n=2 patients with ovarian cancer and n=1 patient with melanoma died before first postbaseline scan and n=1 additional patient with melanoma had a non-evaluable target lesion; these patients are not depicted in the plot but were included in the ORR, cORR and DCR calculation. ORR: according to RECIST 1.1 at any postbaseline scan, PD or death at any prior timepoint; Confirmed ORR for patients with ≥2 postbaseline scans per RECIST 1.1, PD or death at any prior timepoint, those with ongoing unconfirmed PR/CR were excluded. For patients achieving a (c)CR with <100% change from baseline, target lesions were lymph nodes that resolved to <10 mm per RECIST 1.1. BOR, best overall response; (c)CR, (confirmed) complete response; DCR, disease control rate; (c)ORR, (confirmed) objective response rate; PD, progressive disease; (c)PR, (confirmed) partial response; SD, stable disease. Prof. Dr. med. Antonia Busse, abstract #7231 [MADRID 2026 ESMO congress logo] 11 [Slide 4] Changes in Tumor Size Over Time All dose levels: Deep and durable responses observed across all indications with ongoing responses beyond 2.5 years (mFU time of mPFS: 12.8 months) [Spider plot; y-axis: Change in Sum of Longest Diameter of Target Lesions from Baseline [%], 100 to -100; x-axis: Months post T-cell infusion 0-45 (axis break between 15 and 24); reference lines PD, Baseline, PR] BOR (RECIST 1.1): PD; SD; PR; cPR; CR; cCR ▶ Ongoing response/disease control ▶ Ongoing Data cutoff: August 7, 2026 BOR, best overall response; (c)CR, (confirmed) complete response; mFU, median follow-up; PD, progressive disease; (c)PR, (confirmed) partial response; SD, stable disease Prof. Dr. med. Antonia Busse, abstract #7231 [MADRID 2026 ESMO congress logo] 12

[Slide 1] CHAIRS: CELESTE LEBBE, VERNON SONDAK cSCC PROGNOSIS | Scenario 1 of 4 · Transplant and cSCC Immunosuppression predicts cSCC death and the next cSCC, beyond transplantation Solid organ transplant recipients • Metastasis at 2 years: 6.2% in the prospective SCOPE-ITSCC cohort (no control arm)1 • For comparison, about 1 to 4% in immunocompetent cohorts2,3 All IS patients, 3,994 tumours • DSD HR 1.7 vs IC; metastasis HR 1.2 (NS)4 • Transplantation predicts recurrence, metastasis and death; CLL predicts recurrence and death4 • Meta-analysis, all hosts: immunosuppression and tumour size predict death5 Haematology • CLL: cSCC 5-year metastasis 18%, death 11%6 • PWH: no outcome cohort; one series of 10 patients with recurrent and metastatic cSCC8 5-year risk of another cSCC after a first one (%), upper bound of published ranges7 Organ transplant | 92 Haematological malignancy | 64 General population | 67 Full ranges: general population 28 to 67%, haematological malignancy 41 to 64%, organ transplant 74 to 92% Abbreviations: IS, immunosuppressed; IC, immunocompetent; HR, hazard ratio (multivariable); DSD, disease-specific death; CLL, chronic lymphocytic leukaemia; PWH, people with HIV References: 1. Eggermont C, JAAD 2024;90:530-6 (SCOPE-ITSCC) 2. Brantsch KD, Lancet Oncol 2008;9:713-20 3. Venables ZC, JAMA Dermatol 2019;155:298-306 4. Klein JC, JAAD 2026;94:151-60 (18,760 tumours) 5. Dessinioti C, Dermatology 2022;238:1026-35 6. Brewer JD, Clin Adv Hematol Oncol 2012 7. Hollestein LM, JEADV 2026 8. Nguyen P, Arch Dermatol 2002;138:758-63 8 [Speaker panel] Celeste Lebbe — Immunocompromised patients, including transplant and HIV [Slide 2] CHAIRS: CELESTE LEBBE, VERNON SONDAK PREVENTION OF cSCC IN SOTR | Scenario 1 of 4 · Transplant and cSCC Photoprotection, mTOR inhibitors and retinoids; nicotinamide fails in SOTR Sun protection first • Daily broad-spectrum sunscreen in OTR: fewer AK and invasive cSCC at 24 months1 • Education at listing and at every visit; clothing, hats, UV avoidance (expert opinion) mTOR inhibitors • TUMORAPA (randomised): sirolimus after a first cSCC, fewer new cSCC (p < 0.001); no benefit after multiple cSCC2,3 • 1,055 SOTR, 2011-2021: mTORi + reduced CNI, cSCC HR 0.63; everolimus HR 0.384 Retinoids (acitretin) • 3 small RCTs in RTR, 1995 to 2003: new cSCC 11% vs 47% at 6 months5; fewer cSCC in crossover6; two doses alike7 • 32 OTR: about 2 fewer cSCC per year; rebound on stopping8; teratogenic, poorly tolerated • EADO 2026: consider after one or more cSCC9; NCCN: effective in some high-risk patients10 Nicotinamide • IC (ONTRAC, n = 386): 23% fewer keratinocyte cancers at 12 months11 • SOTR (ONTRANS, n = 158): no effect, RR 1.0 (0.8 to 1.3)12 • Not a substitute for IS modification At the first cSCC: switch to an mTOR inhibitor and consider acitretin; nicotinamide is not a substitute Abbreviations: OTR/SOTR, (solid) organ transplant recipients; RTR, renal transplant recipients; AK, actinic keratosis; mTORi, mTOR inhibitor; CNI, calcineurin inhibitor; HR, hazard ratio; RCT, randomised controlled trial; RR, rate ratio; IC, immunocompetent; IS, immunosuppression References: 1. Ulrich C, BJD 2009;161(Suppl 3):78-84 2. Euvrard S, NEJM 2012;367:329-39 3. Dantal J, JCO 2018;36:2612-20 4. Gomez-Tomas A, BJD 2026;194:747-57 5. Bavinck JN, JCO 1995;13:1933-8 6. George R, Australas J Dermatol 2002;43:269-73 7. de Sevaux RG, JAAD 2003;49:407-12 8. Harwood CA, Arch Dermatol 2005;141:456-64 9. Stratigos AJ, EADO cSCC guideline part 1, 2026 update 10. NCCN SCC skin cancer v1.2024 11. Chen AC, NEJM 2015;373:1618-26 12. Allen NC, NEJM 2023;388:804-12 10 [Speaker panel] Celeste Lebbe — Immunocompromised patients, including transplant and HIV [Slide 3] CHAIRS: CELESTE LEBBE, VERNON SONDAK JAK INHIBITORS AND cSCC: THE OTHER COHORTS | Scenario 2 of 4 · JAK inhibitor A consistent two- to three-fold excess of cSCC under ruxolitinib, from trials to registries Study | Design and population | cSCC or NMSC finding COMFORT-II, 5 years1 | RCT, myelofibrosis: ruxolitinib 146 vs best available therapy 73 | NMSC 17.1% vs 2.7%; 6.1 vs 3.0 per 100 patient-years RESPONSE, 80 weeks2 | RCT, polycythaemia vera: ruxolitinib 110 vs best available therapy 111 | NMSC 4.4 vs 2.7 per 100 patient-years, mostly with a prior NMSC Lin, JAAD 20223 | Retrospective cohort, 564 PV or MF, 188 exposed, 10 years, Stanford | cSCC adjusted HR 3.24 (1.45 to 7.22); NMSC HR 2.69 (1.03 to 7.02) Polverelli, BJH 20214 | 700 ruxolitinib-treated MF, 20 European centres | NMSC = 51% of second cancers; ruxolitinib > 5 years, HR 2.93 (1.39 to 6.17) MPN-K, Leukemia 20195 | Nested case-control, 647 second cancers vs 1,234 MPN controls | NMSC with ruxolitinib: OR 3.87 (1.18 to 12.75) ORAL Surveillance, NEJM 20226 | RCT, rheumatoid arthritis, 4,362: tofacitinib vs TNF inhibitor | NMSC HR 1.90 (5 mg) and 2.16 (10 mg) vs TNF inhibitor Pasquer, ASH 2024 (our data)7 | Pharmacovigilance (VigiBase) and single-centre MPN cohort, n = 1,004 | cSCC disproportionality PRR 4.88; JAKi independently associated with cSCC, HR 6.52 Abbreviations: NMSC, non-melanoma skin cancer; cSCC, cutaneous squamous cell carcinoma; MF, myelofibrosis; PV, polycythaemia vera; MPN, myeloproliferative neoplasm; HR, hazard ratio; OR, odds ratio; PRR, proportional reporting ratio; JAKi, JAK inhibitor References: 1. Harrison CN, Leukemia 2016;30:1701-7 2. Verstovsek S, Haematologica 2016;101:821-9 3. Lin JQ, JAAD 2022;86:339-44 4. Polverelli N, Br J Haematol 2021;193:356-68 5. Barbui T, Leukemia 2019;33:1996-2005 6. Ytterberg SR, NEJM 2022;386:316-26 7. Pasquer H, ASH 2024, abstract 4513 15 [Speaker panel] Celeste Lebbe — Immunocompromised patients, including transplant and HIV [Slide 4] CHAIRS: CELESTE LEBBE, VERNON SONDAK MERKEL CELL CARCINOMA | Scenario 3 of 4 · Merkel after transplant In transplant recipients, MCC is mostly virus-negative and more lethal EUROPEAN multicentre cohort: 30 SOTR vs 44 IC1 MCPyV-positive tumours (%) Immunocompetent (n = 44) | 91 SOTR (n = 30) | 33 • Virus-negative, UV-driven MCC predominates in SOTR (p = 0.001); SOTR younger at diagnosis (69 vs 78 years) • Worse outcomes than IC: progression SHR 3.35, MCC-specific death SHR 2.55, OS HR 3.26; mTORi switch after diagnosis did not reduce progression FRANCE, kidney transplant recipients2 • Same virus-negative predominance; MCC often arising with cSCC or its precursors in a field of cancerisation USA, Seattle: 37 SOTR vs immunocompetent, 8 on anti-PD-(L)1 3 • Stage III or more at diagnosis: 76% vs 51%; death within 1 year: 24% vs 4% • Anti-PD-(L)1 in 8 kidney transplant recipients: 5 of 8 responded, 2 irreversible graft failures FRANCE, 13 SOTR on anti-PD-(L)1 (9 kidney, 2 liver, 2 heart)4 • 1 complete response, 12 progressions; median PFS 1.6 months; no graft rejection Abbreviations: MCPyV, Merkel cell polyomavirus; SOTR, solid organ transplant recipients; IC, immunocompetent; SHR, subdistribution hazard ratio; HR, hazard ratio; OS, overall survival; mTORi, mTOR inhibitor References: 1. Ferrandiz-Pulido C, JEADV 2022 (European multicentre cohort, 30 SOTR vs 44 IC) 2. Pacaud A, JEADV 2023 (French kidney transplant recipients) 3. Akaike T, Br J Dermatol 2025;193:1221 (Seattle, 37 SOTR) 4. Idoudi, JDP 2026 (French series, 13 SOTR) 24 [Speaker panel] Celeste Lebbe — Immunocompromised patients, including transplant and HIV

[Slide 1] Multicenter Selective Lymphadenectomy Trial – 3 (MSLT-3) – S2601, EORTC-2519 Resectable Stage III[B-D?] Melanoma N = ~1,574 [?] RANDO 1:1 • Continent • AJCC N Stage • NAT Regimen TLND Arm n = 787 [?] ILN Arm n = 787 [?] Neoadjuvant Immunotx 6 weeks TLND Path Response (ILN + TLND*) -> MPR n = 248 [?] -> 2 year RFS -> Non-MPR n = 539 [?] -> SoC Adjuvant Rx ILN Resection Path Response (ILN only*) -> MPR n = 248 [?] -> Salvage TLND (for recurrence) -> 2 year RFS -> Non-MPR n = 539 [?] -> TLND <3 wks + Path Response -> SoC Adjuvant Rx 10 year OS Index Lymph Node Marker Insertion Key: ILN - index lymph node; MPR - major pathological response (complete + near complete); NAT - neoadjuvant therapy; OS - overall survival; RFS - relapse-free survival; SoC Rx - standard of care treatment; TLND - therapeutic lymph node dissection; * and any additional nodes [key text small/angled, partly illegible] NCT07049276 Alexander C.J. van Akkooi [Slide 2] NCT06754904 OMIT Study The Netherlands Informed Consent: IC Weeks: -3 | 0 | 3 | 7 | 10 Patients with macroscopic resectable stage III melanoma 1. with at least 1 LN metastasis A. RECIST 1.1 measurable OR B. PA proven 2. eligible for neoadjuvant treatment 3. additional in-transit (max 3) allowed -> ILN marking placing -> Neoadjuvant immunotherapy nivolumab + ipilimumab -> ILN removal Yes -> MPR = pCR/npCR -> OMIT TLND -> Follow-up No -> No MPR pPR/pNR -> SoC* -> Follow-up [Slide 3] Practice Variation Currently, some centers started offering ILN - Not experienced in ILN delivery - Do not follow PRADO protocol - Local interpretations = non-evidence based - Postcode variation = non-desired

[Slide 1] [ESMO Congress Madrid 2026 logo] Efficacy and Safety of Lifileucel and Pembrolizumab in Patients with Immune Checkpoint Inhibitor-Naïve Advanced Melanoma: Final Results from IOV-COM-202 Study Andrew J.S. Furness, Helen Gogas, Bernard Doger de Speville Uribe, Donald Lawrence, Gino K. In, Sajeve Thomas, Young Ki Hong, Juan Martin-Liberal, Kai He, Marta Sotelo Garcia, Simon Häfliger, Sylvia Lee, Theresa Medina, Friedrich Graf Finckenstein, Morganna Vance, Rana Fiaz, Melissa Catlett, Siyu Zhu, Rongsu Qi, Brian Thompson, Joe Dean, Brian Gastman, Frederick G. Vogt, Patrick Terheyden Andrew Furness The Royal Marsden Hospital London, UK October 23, 2026 [Live video of presenter at podium on left of screen] [Slide 3] [ESMO Congress Madrid 2026 logo] LBA55: Autologous Tumor-Infiltrating Lymphocytes (GC101 TIL) in Advanced Melanoma: Extended Follow-up of a Randomized, Open-Label, Pivotal Phase II Trial Lu Si, Yu Chen, Weizhen Zhang, Jing Chen, Zhengyun Zou, Dongyuan Zhu, Xingxiang Pu, Xianan Li, Xiaoshi Zhang, Jiyan Liu, Quanli Gao, Zhenchao Yuan, Hong Cen, Meiyu Fang, Lingjun Zhu, Ke Li, Huijing Feng, Tongyu Lin, Huajun Jin, Jun Guo Presented By Lu Si, MD Peking University Cancer Hospital [Live video of presenter at podium on right of screen]

[Slide 1] [ESMO Congress logo] Abstract 2080P [Moffitt Cancer Center logo] [ECOG-ACRIN cancer research group logo] [NCI National Clinical Trials Network logo] Host inherited genetic variation and the risk of hepatitis following immune checkpoint immunotherapy of melanoma Ahmad A. Tarhini, MD, PhD; Zhihua Chen, MS; Mohammad Ali Khaksar, MD; Sandra J. Lee, ScD; F. Stephen Hodi, MD; Tingyi Li, MS; Howard Streicher, MD; Vernon K. Sondak, MD; John M. Kirkwood, MD; Xuefeng Wang, PhD; Peter A. Kanetsky, PhD U.S. Intergroup E1609: Study Design and Accrual Patients with resected IIIB, IIIC, M1a, M1b Melanoma (AJCC7) -> RANDOMIZED -> Ipi3 INDUCTION Ipilimumab 3 mg/kg Q 3 week x 4 -> Ipi3 MAINTENANCE Ipilimumab 3 mg/kg Q 3 month x 4 HD-IFN INDUCTION IFN-α2b 20 MU/m2/d IV x1 month -> HD-IFN MAINTENANCE 10 MU/m2 SC TIW x11 months Ipi10 INDUCTION Ipilimumab 10 mg/kg Q 3 week x 4 -> Ipi10 MAINTENANCE Ipilimumab 10 mg/kg Q 3 month x 4 -> FOLLOW UP Stratified by: IIIB, IIIC, M1a, M1b N = 1673 E1609 (NCT01274338) tested ipilimumab at 3 mg/kg (ipi3) and 10 mg/kg (ipi10) vs. high-dose interferon-α (HDI). It demonstrated significant improvement in overall survival (OS) with ipi3 versus HDI. Ipi3 was significantly less toxic than ipi10 and at least as effective. INTRODUCTION - ICIs are part of the standard of care for patients with advanced melanoma. - Immune-mediated hepatitis (IMH) is a common toxicity following ICI and can be life-threatening. - From E1609, 25.4% of pts treated with ipilimumab experienced IMH, including 5.5% Gr3/4. - Predicting the risk of IMH may inform the treatment plan and clinical follow-up. - We hypothesize that inherited genetic variations contribute to susceptibility to IMH. METHODS - We conducted genome-wide genotyping on biospecimens from 744 consenting patients enrolled in E1609. - Illumina Infinium Global Screening Array v.3.0 + Multi-Disease BeadChip with 730059 designed loci was used. - Genotypes were called using Illumina Genome Studio 2.0 and IAAP CLI v1.1.0, quality controlled, and harmonized with the 1000 Genomes Phase 3 (1KGP3) references. - Michigan Imputation Server 2, pipeline v2.0.6 was used to impute the genotypes against the hrc-r1.1 reference panel. - Logistic regression implemented in plink (v2.00a4LM glm with firth fallback) was used to estimate associations between inherited genetic markers and the occurrence of IMH, adjusting for genetic ancestry. - We calculated relevant PRS listed in PGS-Catalog v20240318 containing variant information and weights for existing polygenic scores (PGSs) and examined 17 distinct PRS previously reported to be associated with inflammatory liver disease and generated from non-melanoma patient populations that were not exposed to immunotherapy. - We treated IMH grades as categorical (Kruskal-Wallis test) and as dichotomous (Wilcoxon rank-sum test) variables. - IMH was graded using CTCAE v.4, the adopted version at the time of conduct of the E1609. RESULTS - Following QC screening, 738 patients were included in the final analysis. - Mean age 52.2 years, 66% (n = 489) male, 98% White, 94% non-Hispanic. - IMH of any grade occurred in 28% of patients in this cohort. See Table 1 [Table 1. Baseline characteristics - small, largely illegible] [Figure 1. Examination of PGS-Catalog PRSs & custom PRS & association with grade - box plots] - There was no significant difference in the distribution of the PRS across the grades of IMH (Kruskal-Wallis) with any of the 17 previously reported inflammatory liver disease PRSs. See Figure 1 - From our GWAS analysis, we identified 10 independent SNPs significantly associated with IMH incidence following ipilimumab. - When combined using log(OR) as weights to construct our custom PRS, these SNPs were significantly associated with the severity of IMH grades, both across irAE grades (Kruskal-Wallis p=3.03e-35) and when dichotomized as any IMH vs. none (Wilcoxon p=2.93e-38). See Figures 1, 2 - ROC analyses were performed to evaluate discriminatory performance for our custom PRS (0 vs. all IMH grades; AUC=0.81). See Figure 3 - Figure 4 shows the Manhattan plot of SNPs associated with IMH in this cohort. - In multivariable survival analysis adjusting for primary tumor status (Known, Unknown) and sex, any grade IMH is associated with improved clinical outcomes for both RFS (HR 0.84, 95% CI 0.67-1.07; p=0.16) and OS (HR 0.71, 95% CI 0.52-0.99; p=0.04). See Figure 5 - In multivariable survival analysis adjusting for primary tumor (Known, Unk) and sex, the combination of 10 SNPs trended towards an association with RFS (HR 0.55, 95% CIO 0.09-3.23; p=0.51) & OS (HR 0.36, 0.04-3.65; p=0.39) likely limited by sample size & number of events. See Figure 5 Figure 2. Novel generated PRS & association with IMH by grade [box plots] Figure 3. ROC analyses showing AUC for custom PRS (0 vs. all grades) [ROC curve, AUC 0.81] Figure 4. Manhattan plot of SNPs associated with IMH (X-axis is SNP chromosomal position; Y-axis is the SNP P-value on a -log10 scale) Figure 5. KM plots & multivariable survival analysis [KM Plot of RFS by Grade 0 vs all IMH; KM Plot of OS by Grade 0 vs all IMH; KM Plot of RFS by Tertiles of PGS Score; KM Plot of OS by Tertiles of PGS Score - details illegible] CONCLUSIONS - Our results support an association between inherited genetic variation and the risk of IMH following ICI - We developed a germline-based PRS that predicts susceptibility to ipilimumab-induced hepatitis with meaningful discrimination (AUC = 0.81) - The generated 10-SNP novel PRS may have implications in better informing patient risk and follow-up in clinic [REFERENCES, funding and acknowledgements - illegible] Presented by Ahmad A. Tarhini, MD, PhD at the European Society for Medical Oncology (ESMO) Annual Meeting, Madrid, Spain, October 23-27, 2026

[Slide 1] Family planning • Recommendations should be made on an individual bases o Tumor characteristics & maternal age1 • Early-stage melanoma with low recurrence risk -> pregnancy likely not affected by melanoma • High risk melanoma after adjuvant Tx -> ideally wait three years after completion of adjuvant Tx1 o Most recurrences occur within the first three years2 • High risk melanoma after neoadjuvant ICI Tx -> potentially shorter interval for major pathological respo[nse — cut off at image edge] o Major pathological response is excellent surrogate for relapse-free survival3 Jessica S.W. Borgers, MD [...]2017); 2 Pérez-Morales. Ann Surg Oncol (2023); 3 Lucas. Ann Oncol (2024)

[Slide 1] [ESMO Congress Madrid 2026 logo] Summary - Personalized cancer vaccines are feasible at a large scale - Increasing evidence of clinical activity, positive randomized phase 2 and 3 studies in adjuvant melanoma - The excellent safety profile gives room for combinations - Robust vaccine-induced immunity, but breadth is limited - A reliable marker tying vaccine-induced immunity to clinical response is elusive (TCR-clonotype expansion, epitope spreading?) - Major [illegible - obscured by camera]tunities for improved/novel delivery approaches and tumor antigen [illegible - obscured]covery [Photo: speaker at ESMO podium, audience in foreground]

[Slide 1] HIGHLIGHTS OF PRESCRIBING INFORMATION These highlights do not include all the information needed to use KEYTRUDA safely and effectively. See full prescribing information for KEYTRUDA. KEYTRUDA® (pembrolizumab) injection, for intravenous use Initial U.S. Approval: 2014 ----------RECENT MAJOR CHANGES---------- Indications and Usage (1) 02/2026 Dosage and Administration (2) 02/2026 ----------INDICATIONS AND USAGE---------- KEYTRUDA is a programmed death receptor-1 (PD-1)-blocking antibody indicated: Melanoma • for the treatment of patients with unresectable or metastatic melanoma. (1.1) • for the adjuvant treatment of adult and pediatric (12 years and older) patients with Stage IIB, IIC, or III melanoma following complete resection. (1.1) [illegible - cropped line] [Slide 2] Probability of Event-free Survival Neoadjuvant–adjuvant group Adjuvant-only group P=0.004 by log-rank test Months since Randomization No. at Risk Neoadjuvant–adjuvant group 154 96 69 46 25 17 1 Adjuvant-only group 159 98 67 40 22 10 2 (time points 0, 6, 12, 18, 24, 30, 36) Figure 1. Kaplan–Meier Estimates of Event-free Survival, as Assessed by the Local Investigators. The log-rank test was stratified according to disease stage and lactate dehydrogenase level at randomization. In the intention-to-treat analysis of event-free survival, there were 105 events (38 in the neoadjuvant–adjuvant group and 67 in the adjuvant-only group). All events that occurred before the start of adjuvant therapy were assigned an event date of 84 days. Tick marks indicate censored data. [Survival data captured as a table — no point estimates read from the curve] Endpoint: Event-free survival (local investigator assessment, ITT) Neoadjuvant–adjuvant group | 154 Adjuvant-only group | 159 P: 0.004 (stratified log-rank) Note: N taken from number-at-risk at time 0. No HR, medians or landmark rates printed in this figure crop. [Slide 3] CLINICAL/PATHOLOGIC STAGE | WORKUP | PRIMARY TREATMENT | ADJUVANT TREATMENT Stage III (clinically positive node[s])ff Workup: • Core biopsy preferred or fine-needle aspiration (FNA). If needle biopsy is not possible, excisional biopsy is acceptable. • Imagingp for baseline staging and to evaluate specific signs or symptoms • BRAF mutation testinggg Resectable nodal disease → Wide excision of primary tumort (category 1) + therapeutic lymph node dissection (TLND) or Neoadjuvant therapy, preferably in the context of a clinical trialtt,uu → Systemic therapy options:ll • Preferred regimens ▸ Nivolumabnn (category 1) ▸ Pembrolizumabnn (category 1) ▸ Dabrafenib/trametinibss for patients with BRAF V600-activating mutation (category 1) and/or Locoregional therapy option: • Consider RT to nodal basin in selected high-risk patients based on location, size, and number of involved nodes, gross and/or histologic extracapsular extensionww,xx (category 2B) or Observationll → See Follow-up (ME-10) Unresectable/borderline resectable → See Unresectable pathway on ME-15uu,vv [Slide 4] CLINICAL STAGE | WORKUPgg | INITIAL TREATMENT | RESPONSE ASSESSMENT | SUBSEQUENT TREATMENT | ADJUVANT TREATMENT Stage III (clinical satellite/in-transit)zz Workup: • Core biopsy preferred or FNA. If needle biopsy is not possible, excisional biopsy is acceptable. • Imagingp for baseline staging and to evaluate specific signs or symptoms • BRAF mutation testinggg Limited resectable disease → Neoadjuvant systemic therapyss,aaa,bbb,ccc (for regimens, see ME-6) → Complete excision to clear margins → or Complete excision to clear marginsddd → No evidence of disease (NED) → / Less than complete resection → See Initial Treatment for Unresectable/borderline resectable disease (ME-8) Optionsll • Systemic therapy optionsggg ▸ Preferred regimens ◊ Nivolumab (category 1)hhh ◊ Pembrolizumabhhh ◊ Dabrafenib/trametinibdd if BRAF V600 mutation positive • Observation (ME-11) or Talimogene laherparepvec (T-VEC)/intralesional therapyeee or Systemic therapybbb (First-Line Therapy, MELSYS-1)ccc → Clinical assessment ± imagingfff to determine treatment response or progression → Residual/progressive unresectable disease Residual resectable disease → See Initial Treatment for Limited Resectable Disease (ME-7) NED after local or regional therapy → Consider adjuvant systemic therapy options listed above (category 2B) or Observation (ME-11) NED after systemic therapy → Observation (ME-11), (ME-J 2 of 4) Unresectable/borderline resectable disease → ME-8 Note: All recommendations are category 2A unless otherwise indicated. Clinical Trials: NCCN believes that the best management of any patient with cancer is in a clinical trial. Participation in clinical trials is especially encouraged. Footnotes on ME-7A ME-7 Melanoma

[Slide 1] A Percentage of the best change from baseline(%) [Waterfall plot, y-axis -100 to 100, dashed lines at +20 and -30] Patients (left to right) with best response labels: H01006ME | PD H01002ME | SD H01012ME | SD H01005ME | PD H01015ME | SD H01003ME | SD H01010ME | SD H01009ME | PR H01004ME | PR H01011ME | SD H01008ME | SD H01014ME | SD H01013ME | SD H01001ME | PR Legend: Cutaneous | Acral | Mucosal | Unknown primary

[Slide 1] Multi-omic profiling identifies determinants of clinical benefit in autologous TIL therapy (LM103) for advanced melanoma Chuanliang Cui(1), Li Zhou(1), Xiaojun Cai(1), Haizhen Du(1), Xuan Wang(1), Yue Yang(1), Jiayong Liu(1), Zhichao Tan(1), Yongming Xue(2), Jun Guo(1), Yan Kong(1) (1)Peking University Cancer Hospital & Institute, Beijing, China; (2) Suzhou BlueHorse Therapeutics Co., Ltd FPN2079P 01 BACKGROUND • Advanced melanoma in Asian populations has distinct biological features and limited benefit from standard therapies. • LM103 is an autologous TIL therapy evaluated in patients with refractory disease. Clinical cohort n = 14 57.1% bORR | DCR 85.7% 4.6 mo MEDIAN PFS 19.12 mo MEDIAN OS Data updated on 2026.09.30 QUESTION: Which intrinsic features of the infused TIL product are associated with longer PFS? 02 METHODS • LM103 products from patients stratified by PFS (> median vs ≤ median) • RNA-seq and ATAC-seq; differential testing with DESeq2 • Pathway interrogation by KEGG/GO GSEA and GSVA 03 MULTI-OMIC SEPARATION OF TIL PRODUCTS BY PFS LONGER PFS IS ASSOCIATED WITH A DISTINCT TIL PRODUCT STATE. [Panels A-D: PCA plots and volcano plots, PFS>median vs PFS<=median] PFS>median open peaks | PFS<=median open peaks [ATAC-seq signal profiles] ATAC-seq average profiles and heatmap centered on differential peaks, showing group specific chromatin accessibility patterns Key multi-omic findings 9 genes upregulated / 7 downregulated in PFS > median TILs Several hundred differentially accessible chromatin regions distinguish PFS groups Interpretation: clinical benefit is linked to a product state consistent with enhanced tumour re-infiltration potential, rather than to baseline tumour features. 05 WORKING MODEL AND CLINICAL IMPLICATION PFS > median TIL product -> CXCR6 / CCR6 + open chromatin -> Putative re-infiltration / persistence -> Clinical benefit Trial registration: CTR20233999 | Funding: Study sponsored by Suzhou BlueHorse Therapeutics Co., Ltd. Disclosure: Y. Xue are employees of Blue Horse Medical Technology Correspondence: k-yan08@163.com 04 PATHWAY-LEVEL CONVERGENCE CXCR6 / CCR6 expression and accessibility are higher in PFS> median group [Panels F, G: GSEA-KEGG enrichment based on RNA-seq; GSEA-GO enrichment based on ATAC-seq — small print] PFS > median Cytokine–cytokine receptor interaction NES = 2.11, FDR = 0.000399 Cytokine Wilcoxon p = 0.0152 Cytokine / chemokine pathways are enriched in PFS>median group PFS ≤ median Glycolysis / Gluconeogenesis NES = -2.07, FDR = 0.0564 Glycolysis Wilcoxon p = 0.002 Glycolytic RNA program is enriched in PFS ≤ median TILs. TAKE-HOME MESSAGE • Autologous TIL therapy with LM103 showed manageable safety and encouraging antitumor activity. • Clinical benefit was associated with sustained in vivo clonal persistence and upregulated genes favoring tumor re-infiltration of infused TILs. [Slide 2] Multi-omic profiling identifies determinants of clinical benefit in autologous TIL therapy (LM103) for advanced melanoma Chuanliang Cui(1), Li Zhou(1), Xiaojun Cai(1), Haizhen Du(1), Xuan Wang(1), Yue Yang(1), Jiayong Liu(1), Zhichao Tan(1), Yongming Xue(2), Jun Guo(1), Yan Kong(1) (1)Peking University Cancer Hospital & Institute, Beijing, China; (2) Suzhou BlueHorse Therapeutics Co., Ltd FPN2079P 01 BACKGROUND • Advanced melanoma in Asian populations has distinct biological features and limited benefit from standard therapies. • LM103 is an autologous TIL therapy evaluated in patients with refractory disease. Clinical cohort n = 14 57.1% bORR | DCR 85.7% 4.6 mo MEDIAN PFS 19.12 mo MEDIAN OS Data updated on 2026.09.30 QUESTION: Which intrinsic features of the infused TIL product are associated with longer PFS? 02 METHODS • LM103 products from patients stratified by PFS (> median vs ≤ median) • RNA-seq and ATAC-seq; differential testing with DESeq2 • Pathway interrogation by KEGG/GO GSEA and GSVA 03 MULTI-OMIC SEPARATION OF TIL PRODUCTS BY PFS LONGER PFS IS ASSOCIATED WITH A DISTINCT TIL PRODUCT STATE. [Panels A-D: PCA plots and volcano plots, PFS>median vs PFS<=median] PFS>median open peaks | PFS<=median open peaks [ATAC-seq signal profiles] ATAC-seq average profiles and heatmap centered on differential peaks, showing group specific chromatin accessibility patterns Key multi-omic findings 9 genes upregulated / 7 downregulated in PFS > median TILs Several hundred differentially accessible chromatin regions distinguish PFS groups Interpretation: clinical benefit is linked to a product state consistent with enhanced tumour re-infiltration potential, rather than to baseline tumour features. 05 WORKING MODEL AND CLINICAL IMPLICATION PFS > median TIL product -> CXCR6 / CCR6 + open chromatin -> Putative re-infiltration / persistence -> Clinical benefit Trial registration: CTR20233999 | Funding: Study sponsored by Suzhou BlueHorse Therapeutics Co., Ltd. Disclosure: Y. Xue are employees of Blue Horse Medical Technology Correspondence: k-yan08@163.com [Section 04 PATHWAY-LEVEL CONVERGENCE and Take-home message largely obscured by presenter standing in front of poster]