Abstract C044: A Comparative Evaluation of Preclinical Models—Cell Lines, Organoids, and PDX—for Informed Model Selection

Abstract Background Preclinical models are critical to drug development, and model selection can significantly influence study outcome, particularly with systematic and comprehensive evaluation and comparison of commonly used in vitro and in vivo models. In this study, leveraging Crown’s large multi-omics data collection of PDXs, organoids, and cell lines, we applied computational approaches to evaluate their clinical relevance, optimizing model choice for efficient preclinical study. Methods We assessed the similarity between 1948 PDXs, organoids (126 PDO and 399 PDXO), and 591 cell lines relative to 1537 TCGA data, covering 17 TCGA cancer types, across multiple omics layers. Gene expression and mutation data of TCGA patient samples were used as a clinical reference for benchmarking model fidelity. These data were obtained from the GDC portal and uniformly processed for cross-platform comparability. These included (1) transcriptional and proteomic similarity analyses, (2) mutational concordance assessment of driver mutations via whole-exome sequencing (WES) data, and (3) pathway activity comparison using Gene Set Variation Analysis (GSVA). These analyses were integrated into a weighted similarity scoring system. Automated pipelines implement this framework to guide model selection for drug discovery, incorporating expression patterns, mutation profiles, and pathway dysregulation. Results We found that PDX and organoid models showed stronger transcriptional and proteomic concordance with TCGA patient samples than cell lines. PDX models better represent stromal and immune-related signatures of the tumor microenvironment (e.g., angiogenesis, ECM remodeling), as reflected in both gene expression and pathway activity, while PDO models better preserve patient-specific tumor heterogeneity. Across all models, 80-90% retention of cancer-specific driver mutations was observed. Our integrated scoring system suggested that for rapid mechanistic studies or large-scale screens, cell lines offer practical advantages despite lower clinical concordance, while PDXs are preferred for TME-dependent studies, and PDOs are more optimal for studies requiring patient-specific heterogeneity. Conclusion This study offers a comprehensive comparative analysis of widely used in vitro and in vivo preclinical models and provides robust guidance for model selection. Our findings demonstrate the complementary strengths of PDXs, organoids, and cell lines, supporting more informed and context-driven choices in preclinical oncology research. Citation Format: Zixuan Xie, Wubin Qian, Sheng Guo, Xiaobo Chen. A Comparative Evaluation of Preclinical Models—Cell Lines, Organoids, and PDX—for Informed Model Selection [abstract]. In: Proceedings of the AACR-NCI-EORTC International Conference on Molecular Targets and Cancer Therapeutics; 2025 Oct 22-26; Boston, MA. Philadelphia (PA): AACR; Mol Cancer Ther 2025;24(10 Suppl):Abstract nr C044.