Identification of a Specific Gene Module for Predicting Prognosis in Glioblastoma Patients

Introduction: Glioblastoma (GBM) is the most common and malignant variant of intrinsic glial brain tumors. The poor prognosis of GBM has not significantly improved despite the development of innovative diagnostic methods and new therapies. Therefore, further understanding the molecular mechanism that underlies the aggressive behavior of GBM and the identification of appropriate prognostic markers and therapeutic targets is necessary to allow early diagnosis, to develop appropriate therapies and to improve prognoses. Methods: We used a weighted gene co-expression network analysis (WGCNA) to construct a gene co-expression network with 524 glioblastoma samples from The Cancer Genome Atlas (TCGA). A risk score was then constructed based on four module genes and the patients' overall survival (OS) rate. The prognostic and predictive accuracy of the risk score were verified in the GSE16011 cohort and the REMBRANDT cohort. Results: We identified a gene module (the green module) related to prognosis. Then, multivariate Cox analysis was performed on 4 hub genes to construct a Cox proportional hazards regression model from 524 glioblastoma patients. A risk score for predicting survival time was calculated with the following formula based on the top four genes in the green module: risk score = (0.00889 × EXP(CLEC5A)) + (0.0681 × EXP(FMOD)) + (0.1724 × EXP(FKBP9)) + (0.1557 × EXP(LGALS8)). The 5-year survival rate of the high-risk group (survival rate: 2.7%, 95% CI: 1.2–6.3%) was significantly lower than that of the low-risk group (survival rate: 8.8%, 95% CI: 5.5–14.1%). Conclusions: This study demonstrated the potential application of a WGCNA-based gene prognostic model for predicting the survival outcome of glioblastoma patients.