Integrated analysis of single-cell RNA sequencing and bulk RNA sequencing reveals T-cell-associated prognostic risk models and the regulation of the tumour immune microenvironment in colorectal cancer

Objective: Colorectal cancer (CRC), a highly prevalent gastrointestinal malignancy, has experienced a paradigm shift in therapeutic strategies driven by recent advances in immunotherapy. Nevertheless, the precise impact of T-cell subpopulation heterogeneity and functional evolution within the tumor microenvironment (TME) on clinical prognosis warrants further elucidation. This study aims to construct a robust T-cell-related prognostic model for CRC by integrating single-cell RNA sequencing (scRNA-seq) and bulk transcriptomic data, thereby evaluating its utility in guiding clinical decision-making. Methods: We utilized scRNA-seq data to comprehensively analyze T-cell subsets and cellular heterogeneity in patients with CRC. Bulk transcriptomic profiles retrieved from public repositories, including The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) databases, were integrated to identify core T-cell signature genes for prognostic model development. Furthermore, the CIBERSORT algorithm was applied to evaluate immune infiltration landscapes across distinct risk strata and to explore their associations with immunotherapeutic responsiveness. Results: A total of 77,167 high-quality single-cell transcriptomes were successfully acquired, yielding 15,450 T lymphocytes that were further classified into central memory T cells, regulatory T cells (Tregs), effector memory T cells, effector T cells, and helper T cells. Integrative bioinformatics analyses identified six core prognostic genes (HSPA1A, TIMP1, HSPA8, ZG16, LGALS4, and CXCL13), based on which a stable prognostic risk model was established. This model effectively stratified CRC patients into high- and low-risk groups, exhibiting superior predictive accuracy across both training and external validation cohorts. Moreover, the high-risk group demonstrated pronounced stromal enrichment and an immunosuppressive TME phenotype characterized by elevated Treg infiltration. Conclusion: This study successfully identifies and validates a novel prognostic signature capable of accurately predicting overall survival in patients with CRC while reflecting the immunological status of the TME, thereby offering a robust scientific framework for personalized therapeutic strategies and immune stratification in clinical practice.