Introduction Precision medicine emphasizes accurate patient stratification. However, existing molecular frameworks often fail to fully capture the intrinsic heterogeneity of breast cancer (BRCA). This study aims to establish a novel immune-based classification system using immunological gene sets to elucidate tumor heterogeneity and identify predictive biomarkers for optimizing clinical management. Methods Non-negative Matrix Factorization (NMF) analysis of Gene Set Variation Analysis enrichment scores, transformed from TCGA, GEO, and CCLE transcriptomic profiles, revealed distinct immune subtypes. These subtypes were comprehensively characterized for their immune, mutational, and molecular features. Therapeutic sensitivity was assessed using computational prediction algorithms and validated via in vitro experiments. Finally, subtype-specific hub genes were identified through co-expression network analysis (WGCNA) and experimentally confirmed. Results Three distinct immune subtypes were identified. Cluster II exhibited a survival advantage, high immune infiltration, and strong correlation with Basal-like tumors. In contrast, Cluster I aligned with Luminal subtypes and indicated poor prognosis. Therapeutic analysis revealed that Cluster II is most likely to benefit from immunotherapy. Additionally, five identified hub genes further confirmed the stability and accuracy of this immune stratification system. Conclusions This study demonstrated that classifying identification of immune patterns for individual tumor patients could reveal the complexity of tumor immune microenvironment and further optimize precision immunotherapy.
Liu et al. (Tue,) studied this question.