Identification of Poor Prognosis-Associated Fibroblast Subpopulation Signature Genes Utilizing the Scissor Algorithm to Classify Colorectal Cancer Subtypes and Evaluate the Immune Landscape.

BACKGROUND/AIMSColorectal cancer (CRC) shows high heterogeneity. Conventional bulk transcriptome-based classification methods fail to capture the complex tumor microenvironment of CRC, limiting precision therapy advances. The Scissor algorithm integrates single-cell/bulk transcriptomic data with clinical information to identify prognosis-linked cell subpopulations, offering new insights into tumor heterogeneity resolution.

METHODSWe integrated bulk RNA-seq data, single-cell RNA-seq data, and clinical information from The Cancer Genome Atlas and Gene Expression Omnibus databases. The Scissor algorithm was employed to screen fibroblast subpopulations strongly associated with prognosis. Multidimensional analyses, including signature gene identification, functional enrichment analysis, Gene Set Variation Analysis (GSVA)-based subtyping, survival analysis, immune landscape assessment, cell-cell communication, mutational profiling, and drug sensitivity prediction, were conducted.

RESULTSWe identified Scissor+ fibroblast subpopulations significantly correlated with prognosis, whose signature genes were associated with pro-metastatic pathways such as extracellular matrix remodeling and transforming growth factor-beta signaling. GSVA scoring stratified samples into Scissor_high and Scissor_low subtypes, with the former being associated with worse survival outcomes, immunosuppressive microenvironment features (including Treg and M2 macrophage enrichment), and stronger immune evasion tendencies. Cell communication analysis revealed that the Scissor_high subtype strongly interacted with many cell types, with remarkable enrichment in the ANNEXIN, SPP1, and NIF signaling pathways. Drug sensitivity predictions suggested that patients with this cell subtype may respond better to specific anticancer agents (e.g., AZD3759, Erlotinib, Gefitinib).

CONCLUSIONSThis study was the first to identify Scissor+ fibroblast subpopulations markedly associated with poor prognosis in CRC, revealing their ability to activate pro-metastatic pathways and immune-activated but functionally exhausted characteristics. Moreover, their predictive value in CRC therapy was revealed. The study provided new perspectives for CRC prognosis evaluation and personalized immune-targeted combination therapies.