PATH-26. A NOVEL PLASMA microRNA SIGNATURE FOR GLIOBLASTOMA DETECTION

Abstract

BACKGROUND: Glioblastoma (GBM) is the most common and aggressive primary brain tumor having a median overall survival below 15 months. To improve the prognosis, biomarkers are needed for early detection of GBM progression. In the present study we identified differentially expressed microRNAs (miRNAs) in plasma samples from GBM patients and investigated their potential to function as noninvasive biomarkers. METHOD(S): The expression profile of 754 miRNAs was measured in plasma samples from GBM patients (n=20), patients with brain metastases (n=20), and age- and sex-matched healthy controls (n=20). The expression of selected miRNAs was verified by semiquantitative real-time PCR and significantly deregulated miRNAs were further selected for validation in an independent cohort. RESULT(S): We found that the expression level of five miRNAs (miR-642, miR-31, and miR-10b, miR-146b and miR-206) was significantly decreased in plasma from GBM patients compared to healthy controls. A model based on these miRNAs was evaluated by receiver operating characteristic (ROC) analysis. In addition a comprehensive model was constructed based on plasma miRNA expression profiles using machine learning. This model enabled discrimination of GBM patients from patients with brain metastases, and healthy controls with an accuracy of 87%, specificity of 100%, and a sensitivity of 80%, but the model must be validated in an independent cohort. CONCLUSION(S): We find that altered expression of certain miRNAs correlates with the presence of GBM, suggesting that plasma miRNAs have the potential to function as noninvasive biomarkers for GBM.