Abstract
BACKGROUND: Colorectal cancer (CRC) is a major cause of cancer-related deaths worldwide. Understanding the genetic and molecular alterations in CRC can improve patient outcomes. Circulating tumor cells (CTCs) are crucial in cancer metastasis and progression. Analyzing the differentially expressed genes (DEGs) between CTCs and CRC may provide us with new therapeutic strategies. Therefore, this study aims to analyze these DEGs to construct a prognostic risk model that predicts the outcomes of CRC patients and guides clinical treatment. METHODS: We analyzed The Cancer Genome Atlas (TCGA) database to identify 1,727 DEGs between CRC and normal samples, and GSE82198 data to find 3,564 DEGs between CTCs and primary CRC samples. Using enrichment analysis, least absolute shrinkage and selection operator (LASSO) regression, and stepwise Cox regression, we derived eight model genes to construct a prognostic risk model. Various algorithms were employed in the immune microenvironment analysis. Integrating clinical factors with risk grouping, we developed a nomogram. We assessed chemotherapy sensitivity and epithelial-mesenchymal transition (EMT) scores in high-/low-risk groups and explored model gene expression at the single-cell level. RESULTS: We constructed a prognostic risk model for CRC based on eight DEGs of CTCs. The model effectively predicted treatment outcomes and correlated closely with actual prognosis. Through immune microenvironment analysis, we revealed differences in immune cell infiltration and checkpoint gene expression among different risk groups. Moreover, patients in the high-risk group showed higher sensitivity to chemotherapy drugs compared to those in the low-risk group. CONCLUSIONS: The prognosis model based on CTCs' DEGs can effectively predict patient outcomes, facilitating precision treatment for patients. This model holds significant guiding implications for immunotherapy and chemotherapy in CRC, offering potential strategies for the clinical treatment of CRC.