Abstract
Glioblastoma multiform (GBM) is a malignant central nervous system cancer with dismal prognosis despite conventional therapies. Scientists have great interest in using immunotherapy for treating GBM because it has shown remarkable potential in many solid tumors, including melanoma, non-small cell lung cancer, and renal cell carcinoma. The gene expression patterns, clinical data of GBM individuals from the Cancer Genome Atlas database (TCGA), and immune-related genes (IRGs) from ImmPort were used to identify differentially expressed IRGs through the Wilcoxon rank-sum test. The association between each IRG and overall survival (OS) of patients was investigated by the univariate Cox regression analysis. LASSO Cox regression assessment was conducted to explore the prognostic potential of the IRGs of GBM and construct a risk score formula. A Kaplan-Meier curve was created to estimate the prognostic role of IRGs. The efficiency of the model was examined according to the area under the receiver operating characteristic (ROC) curve. The TCGA internal dataset and two GEO external datasets were used for model verification. We evaluated IRG expression in GBM and generated a risk model to estimate the prognosis of GBM individuals with seven optimal prognostic expressed IRGs. A landscape of 22 types of tumor-infiltrating immune cells (TIICs) in glioblastoma was identified, and we investigated the link between the seven IRGs and the immune checkpoints. Furthermore, there was a correlation between the IRGs and the infiltration level in GBM. Our data suggested that the seven IRGs identified in this study are not only significant prognostic predictors in GBM patients but can also be utilized to investigate the developmental mechanisms of GBM and in the design of personalized treatments for them.