Abstract 5492: Application of Optim.AI™ platform for predictive evaluation of immunotherapy combinations in breast cancer

Abstract PURPOSE: Successful pre-clinical evaluation of immunotherapy within ex vivo settings require complex systems to accurately represent tumor heterogeneity and the surrounding tumor microenvironment (TME). With expanding immunotherapy combinations being investigated in clinical trials, particularly for breast cancer (BC), there is a growing need for reliable methodologies for functional immune-oncology drug testing in a patient- and subtype-specific manner. We have developed Optim.AI™, a hybrid experimental-analytics platform designed to predict and rank actionable treatment options, with successful application in clinical settings to guide cytotoxic and targeted therapies. However, its potential for evaluating immunotherapy responses remains underexplored. In this study, we present an integrated platform that combines short-term 3D tumor-immune cell co-culture with high content imaging and Optim.AI™ to evaluate immunotherapy-focused drug sets in HER2-positive and triple-negative BC subtypes. METHODS: BC cells were pre-labeled with cell tracker and seeded in a 384-well format for 3-4 days to allow spheroid formation, and co-cultured with PBMCs. Optim.AI™ combinatorial drug testing was then applied to the co-culture system with up to 12 FDA-approved chemotherapies, targeted agents, monoclonal antibodies, and antibody-drug conjugates. Utilizing real-time fluorescent apoptosis signal as the phenotypic readout, high-content analysis of tumor-specific cell death was evaluated at specific timepoints and used for Optim.AI™ analytics and predictive ranking of combinatorial immunotherapy sensitivities. RESULTS: Quantitative, real-time, image-based analysis allowed for phenotypic measurement of complex tumor-immune cell interactions within the 3D co-culture system, including immune cell migration, infiltration into spheroids, and subsequent tumor-specific killing. Co-culturing BC spheroids with PBMCs at an optimized effector-to-target ratio induced significant antibody-dependent cellular cytotoxicity, with measurable responses observed within 72 hours post-treatment. Through Optim.AI™, we observed distinct sensitivities towards the different antibody-based therapies, in HER2-positive and triple-negative BC spheroids. These findings underscore the potential of Optim.AI™ to effectively evaluate and rank immunotherapy-based treatment combinations in a physiologically relevant, 3D TME model. CONCLUSIONS: We present an integrated platform combining high-content screening of a 3D tumor-immune cell co-culture system with Optim.AI™ and demonstrated both visualization and quantification of tumor-specific responses in the context of the TME. Future applications to ex vivo patient samples hold promise for personalized treatment regimens and identifying more effective immunotherapy combinations for BC and other solid malignancies. Citation Format: Sharon Pei Yi Chan, Jhin Jieh Lim, Masturah Mohamed Abdul Rashid, Edward Kai-Hua Chow. Application of Optim.AI™ platform for predictive evaluation of immunotherapy combinations in breast cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 5492.