Abstract
BACKGROUND: Prognostic heterogeneity in stage II/III colorectal cancer (CRC) challenges clinical management, yet effective prognostic stratification is still lacking. To address this, we developed a novel machine learning-based signature focused on immunosenescence. METHODS: This study developed a machine learning-based immunosenescence signature (MALISS) using transcriptomic data from 1296 patients. The final 30-gene model was derived via a CoxBoost-Lasso algorithm and validated across multiple independent cohorts. RESULTS: The MALISS signature effectively stratified patients into high- and low-risk groups with distinct progression-free survival. Functional analysis identified NR1D2 as a key gene promoting tumor migration through cellular senescence. The high-risk group was characterized by a unique mutational landscape, an altered tumor microenvironment, and differential drug sensitivity. Furthermore, a prognostic nomogram integrating MALISS with clinical biomarkers demonstrated improved predictive performance. CONCLUSION: MALISS serves as a robust tool for risk stratification and provides valuable insights into tumor biology, offering a promising approach to address prognostic heterogeneity in stage II/III CRC.