A five-miRNA signature with prognostic and predictive value for MGMT promoter-methylated glioblastoma patients

Abstract

Although O(6)-methylguanine DNA methyltransferase (MGMT) promoter methylation status is an important marker for glioblastoma multiforme (GBM), there is considerable variability in the clinical outcome of patients with similar methylation profiles. The present study aimed to refine the prognostic and predictive value of MGMT promoter status in GBM by identifying a micro (mi)RNA risk signature. Data from The Cancer Genome Atlas was used for this study, with MGMT promoter-methylated samples randomly divided into training and internal validation sets. Data from The Chinese Glioma Genome Atlas was used for independent validation. A five miRNA-based risk signature was established for MGMT promoter-methylated GBM to distinguish cases as high- or low-risk with distinct prognoses, which was confirmed using internal and external validation sets. Importantly, the prognostic value of the signature was significant in different cohorts stratified by clinicopathologic factors and alkylating chemotherapy, and a multivariate Cox analysis found it to be an independent prognostic marker along with age and chemotherapy. Based on these three factors, we developed a quantitative model with greater accuracy for predicting the 1-year survival of patients with MGMT promoter-methylated GBM. These results indicate that the five-miRNA signature is an independent risk predictor for GBM with MGMT promoter methylation and can be used to identify patients at high risk of unfavorable outcome and resistant to alkylating chemotherapy, underscoring its potential for personalized GBM management.

Keywords: MGMT promoter methylation; glioblastoma; miRNA; predictive model; prognosis.

Figure 1. A five-miRNA signature is closely associated with MGMT promoter-methylated GBM prognosis

A five-miRNA signature divides patients into two groups with significantly different prognosis in training A., validation B., combined C., and independent D. datasets.

Figure 2. The five-miRNA signature has prognostic significance in different cohorts

Kaplan-Meier survival analysis of 150 GBM patients with MGMT promoter methylation based on the five-miRNA signature and stratified by the following clinicopathologic risk factors: age A., B., gender C., D., Karnofsky Performance Score E., F., and isocitrate dehydrogenase 1 mutation status G., H..

Figure 3. Comparison of prognosis between unmethylated GBM samples and high- and low-risk GBM patients with MGMT promoter methylation

A.-C. Survival among GBM patients in different groups stratified by MGMT promoter methylation status and the five-miRNA signature. D.-F. Kaplan-Meier survival curves based on alkylating chemotherapy and five-miRNA signature. *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 4. Quantitative model based on three independent prognostic factors

A. Predictive risk model for GBM patients with MGMT promoter methylation. B. Comparison of sensitivity and specificity for predicting 1-year survival using the risk model, which includes age, alkylating chemotherapy, and the five-miRNA signature in a combined set of 150 GBM samples with MGMT promoter methylation.