Abstract 7741: AI-designed personalized neoantigen vaccine, EVX-01, induces durable <i>de novo</i> T-cell responses in advanced melanoma patients

Abstract Cancer vaccines have recently emerged as promising anti-cancer therapies that complement immune checkpoint blockade by stimulating targeted anti-tumor immunity. Their efficacy critically depends on including tumor antigens that can elicit a potent, cancer-specific T-cell response. To achieve this, Evaxion developed the proprietary AI-Immunology™ platform to identify highly immunogenic tumor antigens and design personalized cancer vaccines based on patient-specific sequencing data, optimized for maximal immune activation.The AI-platform identifies somatic mutations and indels in the tumor mutanome, considering tumor expression, clonality and HLA loss to select neoantigens with highest immunogenic potential. EVX-01, the first vaccine candidate generated from this AI-platform, was evaluated in a phase 2 clinical study (NCT05309421) in advanced melanoma. EVX-01 was administered in combination with pembrolizumab, following a 12-week run-in period with pembrolizumab as monotherapy. Patients received in total 10 EVX-01 doses, split into 6 priming and 4 booster immunizations. Primary endpoint was RECIST 1.1-based response improvement of the objective response rate (ORR). For immunogenicity and biomarker analysis peripheral blood was collected before, during and after treatment and processed to peripheral blood mononuclear cells (PBMC). Vaccine-specific T-cell responses were assessed by Interferon-gamma ELISpot and intracellular cytokine staining and flow cytometry, both ex vivo and after in vitro stimulation. Non-vaccine antigens were tested to investigate potential epitope spreading. At the two-year read out, the ORR was 75% (12/16) in the overall cohort. Of the responders, 92% (11/12) continued to respond at 24 month follow up. The treatment was well-tolerated, and the manufacturing success rate was 100%. Additionally, of 13 patients with stable disease (SD) or partial response (PR) after a the 12-week pembrolizumab induction, 54% (7/13) deepened the response upon initiation of EVX-01. The final immunogenicity analysis showed that EVX-01 priming induced strong vaccine-specific T-cell responses in the 15 assessable patients, with booster doses sustaining response magnitudes until last assessment at week 102. Both CD4+ and CD8+ T cells contributed to the response, with CD4+ T cells predominating. 84% of tested neoantigens elicited a specific immune response and AI-Immunology™ platform-assigned immunogenicity scores correlated with the magnitude of T-cell reactivity. In conclusion, the AI-Immunology™ platform can effectively identify highly immunogenic neoantigens, design manufacturable and safe personalized cancer vaccines with pharmacodynamic activity. Combined with PD-1 blockade, EVX-01 achieves a 75% ORR and induces durable, vaccine-specific T-cell immunity in all treated patients, validating the platforms predictive precision. Citation Format: Mads Lausen, Michail Angelos Pavlidis, Ohrt Andersen Rasmus, Thuesen Nikolas, Nadia Viborg, Georgina V. Long, Adnan Khattak, Paolo A Ascierto, Carolina Cimminiello, Thomas Trolle, Christian Garde, Michael J. Chisamore, Benjamin Wolthers, Stine Friis Thorsen, Birgitte Rønø, Daniela Kleine-Kohlbrecher. AI-designed personalized neoantigen vaccine, EVX-01, induces durable de novo T-cell responses in advanced melanoma patients [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 7741.