Abstract P33: Tumor transcriptome deconvolution identifies genes associated with metastatic progression in colorectal cancer

Abstract Most late-stage colorectal cancer (CRC) patients experience lethal metastasis, often to the liver. However, the molecular mechanisms that drive metastatic progression and adaptation to new tissue environments remain challenging to study in patients. To investigate molecular changes associated with CRC liver metastasis, we analyzed genomic and transcriptomic profiles in a discovery cohort of 49 patients with paired primary and metastatic colorectal cancer samples, along with an independent single-cell RNA sequencing (scRNA-seq) dataset from 16 patients with matched primary and liver metastasis samples. We found that noise arising from tissue-specific gene expression signatures obscured conventional bulk-tumor analytical approaches to profile transcriptomic alterations associated with metastatic progression. To address this challenge, we employed a tumor transcriptome deconvolution approach to specifically profile transcriptomic differences between primary and metastatic cancer cells, localized in the colon and liver tissue environments, respectively. We integrated these results with the scRNA-seq cohort to identify high-confidence transcriptomic alterations associated with cancer-cell metastatic progression in CRC. Intriguingly, metastatic cancer cells showed significant upregulation of genes linked to embryonic development. Surprisingly, our analysis also revealed marked downregulation of cell cycle checkpoint pathways, suggesting impaired cell cycle regulation and proliferation in metastatic lesions, which may limit the efficacy of cytotoxic chemotherapy. Overall, our study demonstrates the importance of tumor transcriptome deconvolution when analyzing and interpreting tumor gene expression data from metastatic lesions. Our results highlight specific molecular alterations in metastatic colorectal cancer cells, which could serve as novel biomarkers or therapeutic targets to detect and prevent lethal disease spread. Citation Format: Sinem Kadioglu, Yu Amanda Guo, Simone Rizzetto, Iain Bee Huat Tan, Ker Kan Tan, Anders Jacobsen Skanderup. Tumor transcriptome deconvolution identifies genes associated with metastatic progression in colorectal cancer [abstract]. In: Proceedings of Frontiers in Cancer Science 2025; 2025 Nov 5-7; Singapore. Philadelphia (PA): AACR; Cancer Res 2026;86(13_Suppl):Abstract nr P33.