Patient‐Derived Organoid and Single‐Cell Multiomics Reveal Evolution Driven by
CEACAM6
/
KRT19
‐Mediated Heterogeneity
Changwen Jing, Zhuo Wang, Haixia Cao, Yuan Zhang, Yuchun Huang, Yuetong Yu, Bingzhe Li, Rong Ma ABSTRACT
Colorectal cancer (CRC) remains a leading cause of cancer mortality globally, with therapeutic efficacy hindered by tumor heterogeneity and drug resistance. While organoid technology offers unprecedented opportunities to model tumor complexity, systematic platforms for chemo‐targeted combination screening and molecular mechanisms underlying chemoresistance remain underexplored in Chinese populations. Here, we established a biobank of 17 patient‐derived CRC organoids (85% success rate) from treatment‐naïve surgical specimens, preserving histopathological fidelity and molecular diversity of parental tumors. High‐throughput drug screening across eight clinical regimens revealed different responses. To dissect chemoresistance mechanisms, six organoids representing extreme phenotypes (three sensitive and three resistant, based on IC50 difference for irinotecan + raltitrexed) were selected from this cohort for single‐cell transcriptomic analysis. Single‐cell transcriptomics of irinotecan + raltitrexed‐treated organoids identified chemoresistance‐associated clonal expansion of epithelial subpopulations (cluster_0), characterized by overexpression of REG4/TFF1/TFF3 and epithelial keratins (KRT19/KRT8/KRT18). Intercellular network analysis revealed intra‐epithelial communication as a resistance hub, mediated by 32 ligand‐receptor pairs. Integration with TCGA‐COAD cohort ( n = 378) demonstrated intratumoral heterogeneity (ITH) score as an independent prognostic determinant, outperforming stemness/angiogenesis metrics. Machine learning consensus identified CEACAM6 and KRT19 as dual regulators of mortality and ITH through epithelial plasticity networks. Our multi‐omics framework establishes organoid‐guided chemoresistance mechanisms while proposing ITH quantification targeting as precision strategies. This study bridges critical gaps in CRC therapeutic personalization, offering clinically actionable insights for overcoming treatment failure.