Identification of polyamine metabolism-related gene signatures for prognostic and immunological stratification in glioma.

Polyamine metabolism plays an essential role in glioma progression and the tumor microenvironment (TME). However, its prognostic and immunotherapeutic significance remains incompletely understood. We integrated transcriptomic and clinical data from The Cancer Genome Atlas (TCGA) to construct a prognostic risk model via Least Absolute Shrinkage and Selection Operator (LASSO) regression. This model was externally validated using the GSE108474 and GSE43378 cohorts. Multivariable Cox regression was employed to identify key prognostic predictors, while consensus clustering was applied to stratify patients into distinct molecular subtypes. A robust 19-gene signature was established to calculate risk scores, effectively categorizing patients into high- and low-risk groups with significantly divergent survival outcomes. From this signature, five genes were identified as independent prognostic factors via multivariable Cox analysis. Ultimately, three prioritized core candidate biomarkers (AGMAT, PSMC5, and SMS) were extracted, among which SMS exhibited exceptionally high individual diagnostic efficacy. Furthermore, consensus clustering delineated two distinct subgroups (C1 and C2). The C1 subtype exhibited a less immune-infiltrated phenotype but higher PD-L1 expression and elevated TIDE scores, indicating stronger immune evasion and poorer survival. In contrast, the C2 subtype was characterized by robust immune cell infiltration, a more favorable prognosis, and a superior computationally predicted response to immunotherapy. In conclusion, our polyamine metabolism-related gene signature may serve as a promising candidate biomarker for evaluating clinical outcomes and computationally estimated immunotherapy efficacy in glioma. While this risk model and molecular subtyping framework offer novel insights for therapeutic stratification, prospective clinical validation is required.