A colorectal cancer risk model based on cell adhesion-related genes for predicting prognosis and immunological features.

Colorectal cancer (CRC) remains among the leading causes of cancer mortality worldwide, with tumor heterogeneity and immunosuppressive microenvironments posing major challenges to effective risk stratification and immunotherapy. Although cell adhesion molecules (CAMs) are well-established drivers of CRC progression through metastasis and immune evasion, their potential as integrated prognostic biomarkers and therapeutic targets remains underexplored. This study systematically characterizes CAM-related gene signatures to develop a robust prognostic framework that refines CRC risk assessment, elucidates immune regulatory mechanisms, and identifies precision therapy opportunities. Using The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) transcriptomic datasets, we identified adhesion-related genes (ARGs) and constructed an 8-gene prognostic model via Least Absolute Shrinkage and Selection Operator (LASSO) and multivariate Cox regression. We evaluated immune infiltration, mutation burden, drug sensitivity, mendelian randomization (MR) causality, and defined novel molecular subtypes through consensus clustering. The 8-gene risk model effectively stratified patients into high- and low-risk groups with markedly different survival outcomes. High-risk patients exhibited immunosuppressive tumor microenvironments, reduced immunotherapy response, and distinct therapeutic vulnerabilities. MR confirmed a causal role for SLAMF1 in CRC risk. This ARG-driven prognostic framework not only enhances precision risk stratification in CRC but also reveals subtype-specific immune evasion mechanisms and therapeutic targets. These findings offer actionable insights for personalized CRC management and underscore CAMs as promising candidates for future functional and translational studies across gastrointestinal malignancies.