An Immune Subtype Classification System Enables the Development of Strategies to Predict and Enhance Immunotherapy Responses in Colorectal Cancer

Abstract Colorectal cancer is the second leading cause of cancer-related mortality globally. Although immunotherapeutic approaches are effective in a subset of patients with colorectal cancer, the majority of colorectal cancer cases receive limited benefits from immunotherapy. This study developed an immune subtype classification system based on diverse immune cells and pathways. A model constructed through machine learning based on immune subtypes could accurately predict the sensitivity of patients with colorectal cancer to immunotherapy. Validation of this model across public datasets and clinical samples confirmed its high precision and reliability. Furthermore, drug screening based on the immune subtypes identified the insulin-like growth factor 1 receptor inhibitor I-OMe-AG-538 (AG-538) as a potent enhancer of antitumor immunity. Mechanistic investigations revealed that AG-538 induced reactive oxygen species–dependent DNA damage and downregulated the expression of multiple repair genes, triggering cyclic GMP-AMP synthase/stimulator of interferon gene–mediated type I IFN signaling within tumor cells. This signaling cascade increased tumor immunogenicity and refined the tumor immune microenvironment, thereby enhancing the efficacy of immune checkpoint blockade treatment. In summary, these findings present a predictive model for immune response and highlight the potential of AG-538 combined with anti–PD-1 antibodies as a chemoimmunotherapeutic strategy. Significance: The identification of immune subtypes in colorectal cancer facilitated the construction of a model to determine immunotherapy sensitivity in patients and uncovered an effective chemoimmunotherapeutic approach, paving the way for personalized treatment.