A translational colorectal cancer organoid biobank mirrors patients’ tumor histology, molecular profiles, and treatment responses

Abstract Background Colorectal cancer (CRC) exhibits pronounced inter- and intratumoral heterogeneity, emphasizing the need for preclinical models that accurately capture its molecular and histological diversity. Patient-derived organoids (PDOs) represent valuable ex vivo systems to model CRC, yet whether they can preserve subtype-specific features and maintain fidelity upon in vivo transplantation remains unclear. Methods We established a biobank of PDOs from both treatment-naïve and neoadjuvant-treated CRC patients, encompassing the major histological subtypes – micropapillary, medullary, serrated, mucinous, and adenocarcinoma not otherwise specified (NOS). PDOs were comprehensively characterized by histomorphological, genomic, and transcriptomic analyses. To assess in vivo fidelity, PDOs were orthotopically transplanted into immunodeficient mice to generate patient-derived organoid xenografts (PDOXs). PDOX-derived tumors and organoids were analyzed to evaluate the preservation of histological and molecular traits, as well as therapy responses. Results PDOs displayed distinct, subtype-specific morphologies and growth patterns that closely paralleled their respective patient tumor histologies. Orthotopic PDOXs recapitulated the histological architecture, gene expression profiles, and signaling pathway activation of the original tumors. PDOX-derived organoids retained these subtype-specific morphologies, molecular features, and exhibited similar responses to FOLFOX treatment as their corresponding PDOs, confirming both molecular and functional stability of the organoid-xenograft cycle. Conclusion This study establishes orthotopic transplantation of CRC PDOs as a robust and predictive preclinical model that captures the full spectrum of CRC heterogeneity. The model preserves histological and molecular subtype fidelity across in vitro and in vivo contexts and enables functional assessment of therapy response. By bridging patient-derived tumor biology with translational modeling, this platform provides a valuable resource for dissecting CRC pathogenesis and advancing patient-tailored precision oncology.