A Six-Immune-Related Genes Prognostic Signature for Glioblastoma

Abstract Background: Glioblastoma (GBM) multiforme is a common malignant brain tumor with high mortality. It is urgently necessary to develop a new treatment because traditional approaches have reached a bottleneck. Purpose : Here we created an immune-related gene (IRGs)-based prognostic signature to comprehensively define the prognosis of glioblastoma (GBM). Methods: Glioblastoma samples were abstracted from the Chinese Glioma Genome Atlas (CGGA) and the Gene Expression Omnibus (GEO). We retrieved IRGs from the ImmProt data resource. Univariate Cox analysis was adopted to determine the prognostically remarkable IRGs for individual with GBM. The prognostically optimal IRGs were determined via LASSO regression, and predictive model created. Besides, the association of specific factors with the overall survival (OS) of individuals with GBM was explored via multivariate Cox-regression. Lastly, we constructed a predictive nomogram integrating the independent predictive factors to determine the one-, two-, and three-year OS likelihoods of individuals with GBM. Additionally,gene set enrichment analysis(GSEA) and single sample GSEA(ssGSEA) were performed to understand the correlation between the risk score and immune activity. Results: Overall, 273 IRGs which exhibited differential expression were identified in GBM tumor in contrast with the non-malignant samples. Of these 273 IRGs, only six were remarkably linked to OS of individuals with GBM, which were employed in constructing the predictive signature. The GBM were categorized into either the high-risk GBM group or the low-risk GBM group. There were remarkable differences between the high-risk GBM and the low-risk GBM groups regarding OS. The AUC for predicting one-,two-, and three-year OS in training set was 0.610,0.698 and 0.694.In line with the AUC of validation set was 0.608,0.692 and 0.678.Besides,the results of ssGSEA showed the score of prognostic signature is closely related to immune activity. Conclusion: Herein, a robust predictive model based on IRGs was created to estimate the diversity of OS likelihoods in GBM patients, as well as aid future clinical research.