A Robust Immunohistochemistry-Based Classification for BRAF V600E-Mutant Colorectal Cancer With Clinical Implications

BRAF V600E-mutant colorectal cancer (CRC) represents a distinct molecular subtype with considerable heterogeneity in tumor biology and therapeutic response. Although gene expression-based classifications (BM1/BM2 subtypes) provide valuable insights into underlying molecular and immune features, their clinical application is limited by the need for high-throughput sequencing. This study aims to establish an immunohistochemistry (IHC)-based classification system to enable practical subtype stratification and aid prognostic and therapeutic evaluation. Using two independent cohorts (public dataset, n = 218; institutional cohort, n = 122), we performed differential expression analysis, machine learning modeling, and clinical feasibility evaluation. Fourteen candidate markers were identified, and a decision tree algorithm selected CD8 and ARHGEF17 as optimal classifiers. Based on these markers, two IHC-based subtypes were established: iBM1 (CD8⁺/ARHGEF17^-) and iBM2 (CD8^- with any ARHGEF17 expression or CD8⁺/ARHGEF17⁺). The IHC-based subtypes showed concordance with transcriptomic BM subtypes (training: 82.69%, κ = 0.55; validation: 72.22%, κ = 0.44; prospective: 83.33%, κ = 0.57). Transcriptomic profiling revealed enrichment of immune activation and epithelial-mesenchymal transition in iBM1, and cell cycle-related pathways in iBM2. In clinical validation, iBM1 was associated with poorer survival but greater sensitivity to immune checkpoint inhibitors. This IHC-based classification provides a practical and accessible approach for BM subtype stratification, reflecting underlying molecular and immune characteristics, and may support prognostic assessment and therapeutic decision-making in BRAF V600E-mutant CRC.