An immune-related gene pair signature predicts the prognosis and immunotherapeutic response in glioblastoma

Abstract

Background: Glioblastoma (GBM) has the feature of aggressive growth and high rates of recurrence. Immunotherapy was not included in standard therapy for GBM due to lacking the predictive biomarkers. In the present study, we performed an immune-related gene pair (IRGP) signature to predict the prognosis and immunotherapy response of GBM. Methods: A total of 160 GBM patients from TCGA were included. ssGSEA was conducted to evaluate the immune infiltration level. Univariate Cox, LASSO regression analysis, ROC analysis, and Kaplan-Meier survival analysis were applied to construct and evaluate the risk model. Moreover, the association between immune infiltration and the risk score was assessed. Finally, the expression of immune checkpoints between different risk groups was explored. Results: According to the normal/tumor, high-/low-immunity group, we identified 125 differentially expressed immune-related genes. Subsequently, a prognostic model including 22 IRGPs was established. The area under the ROC curve to predict 1, 3, and 5-year was 0.811, 0.958, and 0.99 respectively. According to the optimal cut-off value of the 3-year ROC curve, patients were classified into high- and low-risk groups. The Kaplan-Meier analysis result indicated that patients in the low-risk group have longer survival time. The risk score was an independent prognostic predictor (P < 0.001). Moreover, PDCD1 was positively correlated with the risk score (P < 0.01). We also found that patients with high PDCD1 expression had worse survival. Conclusions: The IRGP signature was built to predict the prognosis of GBM patients. This signature can serve as a tool to predict the response to immunotherapy in GBM.

Keywords

• Glioblastoma (GBM)
• IRGP
• Prognostic model
• Immune infiltration
• Immune checkpoints