Application of a machine learning-based PANoptosis-immune-related gene risk score model in prognostic stratification and immunotherapy benefit prediction for glioblastoma.

This study developed a risk score model using PANoptosis and immune-related genes to predict glioblastoma (GBM) prognosis. Utilizing TCGA data and 66 PANoptosis regulatory network-related genes, patients were clustered into two subtypes. Machine learning identified 11 core PANoptosis-Immune-Related Genes (PIRGs). Single-cell analysis revealed their dysregulated expression in GBM. External validation confirmed that high-risk patients exhibited poorer survival, a dysfunctional tumor immune microenvironment (TIME), and reduced sensitivity to radiotherapy and temozolomide. This group displayed enriched immune activation pathways, a lower immunophenoscore (IPS), and differential drug sensitivity. High TIDE scores indicated a potential benefit from immune checkpoint inhibitors.