P08.44 Preoperative serum microRNA profiles as a diagnostic tool in glioma

Introduction: Current diagnosis of glioma is based on clinical suspicion, magnetic resonance imaging (MRI) and surgical biopsy based on WHO criteria for neuropathology. Blood sampling for circulating assessment of biomarkers is a potentially valuable method that could add to diagnostic and prognostic accuracy. We previously showed that miR-10b and miR-21 were significantly upregulated in high-grade glioma patients compared to healthy controls. We hypothesized that large-scale miRNA profiling could improve diagnostic discrimination across all grades of glioma. Materials and Methods: Pre-operative serum samples from 91 glioma patients (45 GBM, 46 Low Grade Glioma, LGG) from the Royal Melbourne and Melbourne Private Hospitals between 2009 and 2013 were included in the cohort and compared to 20 healthy controls. miRNA profiling was performed by Nanostring v.3 analysis of isolated total serum RNA. Statistical output was generated using nSolver and R packages. Results: miR-320e, miR-16-5p and miR-223-3p differed significantly between GBM, LGG and healthy controls. Highly differentially expressed serum miRs in GBM vs LGG included miR-520f-3p, miR-223-3p and miR-150-5p and these were also associated with overall survival. A 9-miRNA gene classifier incorporating miRs with p0.5 in either direction was able to discriminate between GBM and LGG. Conclusions: Profiling of miRNA expression levels in serum shows promise for discovery of diagnostic and prognostic circulating biomarkers in glioma. These preliminary results require validation in prospective data sets including sequential post-operative serum samples.