COMP and CXCL10: a novel immunogenic cell death-derived gene signature for prognostic prediction in colorectal cancer

Background Colorectal cancer (CRC) continues to be a major contributor to cancer-related mortality, with advancements in immunotherapy progress impeded by suboptimal patient responses. Immunogenic cell death (ICD), a distinct process that promotes antitumor immunity, has garnered significant attention in oncological research. However, prognostic models based on ICD for CRC remain insufficiently investigated. This study aimed to analyze ICD-related genes (ICDRGs) in CRC and to develop a prognostic framework that enhances the understanding of disease mechanisms to identify potential prognostic markers. Methods Data from TCGA-CRC, GSE17536 , and ICDRGs were integrated to identify differentially expressed genes (DEGs) between tumor and normal tissues. A weighted gene co-expression network analysis (WGCNA) identified modules associated with immunogenic cell death (ICD) using ssGSEA-derived ICD scores. The intersection of DEGs and module genes formed the candidate set. Prognostic determinants and risk factors were identified to develop the risk and prognostic models. Functional enrichment analyses were conducted on risk groups to elucidate CRC-related pathways. Immune infiltration patterns and MSI statuses were assessed in TCGA-CRC. Additionally, drug sensitivity was compared between high- and low-risk groups and between MSI-H and MSI-L/MSS cohorts. Finally, the expression of prognostic genes was validated in CRC and adjacent non-tumor tissues using qRT-PCR. Results The overlap between 979 differentially expressed genes (DEGs) and 1,217 modular genes in TCGA-CRC identified 47 candidate genes. From this set, five genes retained in the multivariate Cox regression, GPRASP1, FAT4, MEIS2, COMP, and CXCL10, were incorporated into a risk model that showed moderate predictive accuracy in the GSE17536 validation cohort. A nomogram integrating the risk score with age and T, N, and M stages was subsequently established. An immune infiltration analysis revealed a significant association between risk score and regulatory T cells (Tregs), while CXCL10 exhibited a pronounced positive correlation with multiple immune checkpoints. Drug sensitivity assessments revealed significant disparities in responses to standard chemotherapeutic agents between high- and low-risk groups, as well as between MSI-H and MSI-L/MSS cohorts. Furthermore, a qRT-PCR analysis confirmed significant overexpression of COMP and CXCL10 in CRC tissues. Conclusions This study identified COMP and CXCL10 as significantly expressed prognostic genes in colorectal cancer and introduced a novel immune cell death-associated prognostic model. The model exhibited consistent predictive reliability across both training and validation datasets, providing valuable insights for prognosis assessment and therapeutic strategies in CRC.