Abstract
Chimeric antigen receptor T (CAR-T) cell therapy has achieved transformative success in hematological malignancies; however, its translation to solid tumors remains severely limited by tumor heterogeneity, immunosuppressive microenvironments, and safety concerns such as on-target/off-tumor toxicity. A major contributor to these challenges is the lack of preclinical models capable of faithfully recapitulating human tumor architecture and tumor-immune interactions. Conventional two-dimensional cell cultures and animal models frequently fail to predict CAR-T efficacy, resistance, and toxicity observed in patients. Organoid technology, particularly patient-derived organoids (PDOs) and immune-integrated organoid systems, has emerged as a next-generation platform that bridges this translational gap. By preserving patient-specific genetic, phenotypic, and spatial heterogeneity, organoids provide a physiologically relevant and scalable system for interrogating CAR-T cell behavior in human-like tumor contexts. Recent advances in tumor-immune co-culture, vascularized organoids, and microfluidic organoid-on-a-chip platforms have further expanded their utility for dynamic assessment of CAR-T infiltration, cytotoxicity, cytokine release, and adaptive resistance mechanisms. In this review, we comprehensively examine how organoid-based models are reshaping the CAR-T development pipeline, spanning target discovery and validation, functional efficacy assessment, safety profiling, and optimization of combination therapies. We further discuss emerging applications of organoids as patient-specific "avatars" for personalized CAR-T selection and response prediction. Finally, we highlight current technical limitations and future bioengineering directions required to enable clinical translation. Collectively, organoid platforms represent a transformative tool for accelerating precision development of next-generation CAR-T cell therapies and advancing human-relevant immuno-oncology research.