A novel glioblastoma prognostic assay using droplet digital polymerase chain reaction.

e14026 Background: Glioblastoma (GBM) is an incurable brain tumor for which outcomes have seen little improvement in the past 30 years. Despite limited GBM therapies, accurate prognostic tools can be invaluable in informing best course of treatment. Recently, we have identified the GBM progression gene signature (GBM-PGS), a collection of 31 genes that shows superior accuracy at predicting GBM progression compared to existing biomarkers. From GBM-PGS expression, a GBM progression risk score can be calculated to provide clinicians with patient-tailored prognostic information. However, the question remains as to the best way to measure GBM-PGS to create a clinically applicable test. At our disposal are three approaches to quantify gene expression: quantitative polymerase chain reaction (qPCR), droplet digital PCR (ddPCR), and RNA sequencing (RNA seq). We hypothesize that the combination of high accuracy and sensitivity, while also providing absolute quantification, will deliver ddPCR as the best option for creating a rapid clinical test that is scalable for measuring GBM-PGS. Methods: A ddPCR protocol was developed, and cDNA and primer concentrations were optimized using samples from two GBM cell lines (SF295, LN18). To confirm assay accuracy, GBM-PGS was quantified using ddPCR and compared to GBM-PGS expression by qPCR, a protocol that has previously been thoroughly validated. Finally, GBM-PGS expression was measured in patient tumors using ddPCR, and the GBM-PGS risk score algorithm was trained using an RNA seq data set from the Tissue Cancer Genome Atlas (TCGA). Following training, risk scores were calculated for the patient samples using ddPCR-measured GBM-PGS expression. Risk scores were compared with clinical patient survival to determine if GBM-PGS expression measured by ddPCR could predict patient outcomes. Results: Optimal primer concentration (200 nM) and cDNA concentration (1.2 ng/μl) were identified. GBM-PGS expression measured by ddPCR absolute quantification and qPCR Ct value demonstrated a strong correlation in the SF295 (R2 = 0.91) and LN18 (R2 = 0.88) cell lines. GBM-PGS was measured in two patient samples, and the resulting risk scores were calculated as 26.6 and 29.8 (score > 0 = high progression risk; score < 0 = low progression risk) compared to disease free survival times of 12.9 and 10.5 months, respectively, demonstrating an inverse relationship between patient risk score and survival time. Conclusions: Our results demonstrate that we have developed an accurate ddPCR-based assay capable of measuring GBM-PGS in multiple GBM cell lines, and preliminary results suggest that GBM-PGS quantified using ddPCR accurately predicts patient survival.