Glioblastoma (GBM), the most common and aggressive brain tumor, has a very poor outcome and high tumor recurrence rate. The immune system has positive interactions with the central nervous system. Despite many studies investigating immune prognostic factors, there is no effective model to identify predictive biomarkers for GBM. Genomic data and clinical characteristic information of patients with GBM were evaluated by Kaplan–Meier analysis and proportional hazard modeling. Deseq2 software was used for differential expression analysis. Immune-related genes from ImmPort Shared Data and the Cistrome Project were evaluated. The model performance was determined based on the area under the receiver operating characteristic (ROC) curve. CIBERSORT was used to assess the infiltration of immune cells. The results of differential expression analyses showed a significant difference in the expression levels of 2942 genes, comprising 1338 upregulated genes and 1604 downregulated genes (p < 0.05). A population of 24 immune-related genes that predicted GBM patient survival was identified. A risk score model established on the basis of the expressions of the 24 immune-related genes was used to evaluate a favorable outcome of GBM. Further validation using the ROC curve confirmed the model was an independent predictor of GBM (AUC = 0.869). In the GBM microenvironment, eosinophils, macrophages, activated NK cells, and follicular helper T cells were associated with prognostic risk. Our study confirmed the importance of immune-related genes and immune infiltrates in predicting GBM patient prognosis.
Identification of differential expression of genes in hepatocellular carcinoma by suppression subtractive hybridization combined cDNA microarray