Identification of immune-related gene signature for predicting prognosis of glioblastoma patients.

BACKGROUNDGlioblastoma (GBM) is the most common and aggressive primary brain tumor with a poor prognosis due to its immunosuppressive microenvironment. Reliable prognostic biomarkers are still lacking. This study aims to construct an immune-related gene signature to predict the prognosis of GBM patients and explore the molecular characteristics of different immune subtypes.

METHODSThe Cancer Genome Atlas (TCGA), the Chinese Glioma Genome Atlas (CGGA), and Gene Expression Omnibus Series (GSE) 108474 datasets were integrated to classify cold and hot tumor subtypes via single-sample gene set enrichment analysis (ssGSEA) and partitioning around medoids (PAM) clustering. Prognostic genes were identified via weighted gene co-expression network analysis (WGCNA) and screened with 101 machine learning (ML) algorithms. A risk-scoring model was developed, and gene expression was validated via single-cell transcriptomics and immunohistochemistry (IHC).

RESULTSSeven core genes were identified, with area under the curve (AUC) values of 0.75-0.88 for distinguishing cold/hot subtypes. The model exhibited stable predictive performance in external validation cohorts (AUC =0.55-0.69). The survival rate of the high-risk cold tumor cohort was significantly lower than that of the low-risk hot tumor cohort.

CONCLUSIONSOur study identifies key genes and molecular mechanisms underlying the immune microenvironment of GBM. The prognostic model developed lays a theoretical foundation for personalized immunotherapy and stratified treatments.