Abstract Introduction: Over the last decade, chimeric antigen receptor T cell (CAR-T)-based immunotherapy has achieved unprecedented clinical benefits for the treatment of hematological malignancies. However, application of CAR-T therapy to solid tumors remains challenging, in part due to tumor antigen heterogeneity and microenvironmental complexity present in these tumors. Therefore, selecting physiologically-relevant models is critical for validation of the on-target efficacy and off-target effect of CAR-T therapy. Emerging 3D patient-derived tumoroid models better recapitulate the morphological organization, mutational status, and gene expression levels of patient tumors. Hereby, we propose a screening platform to evaluate mesothelin (MSLN)-targeted CAR-T therapy for solid tumors using patient-derived tumoroids. Methods: T cells were isolated from peripheral blood mononuclear cells of healthy human donors. Meso3 (region III of MSLN) or CD19 CARs (served as control) were knocked-in (KI) to activated T cells using electroporation-based CRISPR/Cas9 system followed by CAR Cassette containing adeno-associated virus (AAV) transduction. T-cell receptor (TCR) knockout (KO) and CAR-KI efficiency were validated by flow cytometry. Patient-derived tumoroids with high and low MSLN expression were co-cultured with various effector-to-target ratios of Meso3 CAR, CD19 CAR and Control (TCR KO without CAR KI) T cells. Cytotoxicity was evaluated by caspase-based live cell imaging. Key cytokine secretion was measured using a multiplexing platform 2-3 days post co-culture. Results: T-cell electroporation of CRISPR/Cas9 yielded a high KO efficiency (>90%) of TCR as well as high AAV-mediated CAR cassette-KI efficiency (>70%). The killing efficiency of Meso3 CAR-T cells with different effector to target ratios was successfully captured in a ratio-dependent manner via live cell imaging and analysis platform in the MSLN-high tumoroids, while minimum killing was observed in MSLN-low tumoroids. As expected, control and CD19 CAR-T cells showed no killing towards MSLN-high or -low tumoroids, which do not express CD19. Moreover, the observed killing ability correlated with secretion of cytokines associated with T cell-mediated cytotoxicity. Conclusion: These findings demonstrated the ability of 3D patient-derived models as a screening platform for validations of CAR-T cell therapy in vitro. This platform has potential advantages over commonly used impedance-based and flow cytometry-based killing assay since it better captures the 3D architecture and microenvironment of the tumor origin and allows real-time quantification. With the ability of accurately detecting on and off-target killing, quick turn-around time, and cost-effectiveness compared to in vivo systems, this platform allows for high-throughput screening workflows for CAR-T cell therapies. Citation Format: Xiaoyu Yang, Maryam Mafreshi, Garrett Wong, Vivek Chandra, Colin Paul, Chris Yankaskas, Brittany Balhouse, Jason Sharp, Matt Dallas, Erik Willems, David Kuninger. Evaluation of CAR-T cytotoxicity in solid tumors: A screening workflow using 3D patient-derived tumoroids [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 6787.