MicroRNA signatures predict prognosis of patients with glioblastoma multiforme through the Cancer Genome Atlas

Abstract

MicroRNAs (miRNAs) play major roles in various biological processes and have been implicated in the pathogenesis and malignant progression of glioblastoma multiforme (GBM). The aim of this study was to assess the predictive values of miRNAs for overall survival (OS) of patients with GBM. MiRNA expression profiles and clinical information of 563 GBM patients were obtained from the Cancer Genome Atlas. The most significantly altered miRNAs were identified and miRNA expression profiles were performed, through principal component analysis, the least absolute shrinkage and selection operator method. The survival analysis was performed using the Cox regression models. Additionally, receiver operating characteristic (ROC) analysis was used to assess the performance of survival prediction. We used the bioinformatics tools to establish the miRNA signature for biological relevance assessment. A linear prognostic model of three miRNAs was developed and the patients were divided into high risk and low risk groups based this model. The area under the ROC curve (AUC) for the three miRNA signature predicting 5-year survival was 0.894 (95%CI, 0.789-1.000) in the testing set and0.841 (95%CI, 0.689-0.993) in all GBM patients. High risk patients had significantly shorter OS than patients with low risk (P< 0.001). The results from this study support a three miRNA signature for outcome prediction of GBM. These results provided a new prospect for prognostic biomarker of GBM.

Keywords: TCGA; glioblastoma multiforme; microRNA; overall survival; prognosis.

Figure 1. MiRNA selection using the least absolute shrinkage and selection operator (LASSO) binary logistic regression model

(A) Tuning parameter (λ) selection in the LASSO model. The area under the receiver operating characteristic (AUC) curve was plotted versus log (λ). (B) LASSO coefficient profiles.

Figure 2. The ROC curves for the three microRNA signature in TCGA GBM cohort

The ROC curve for predicting 5-year survival in GBM with an AUC of 0.841 (95%CI, 0.689-0.993) in the training set (A), an AUC of 0.894 (95%CI, 0.789-1.000) in the testing set (B), and an AUC of 0.854 (95%CI, 0.744-0.964) in all GBM patients (C), respectively.

Figure 3. Kaplan Meier curves for the three microRNA signature in TCGA GBM cohort

Individual patient was scored according to the three miRNAs signature. The Kaplan-Meier curves for GBM risk groups obtained from the TCGA cohort divided by the cut-off point. The OS of high risk group is significantly lower than that of low risk group in all GBM patients (C). The P values of the log-rank tests are <0.001.

Figure 4. The top 20 enriched functional analysis from Gene ontology (GO) analysis