MP02-03 NEXT-GENERATION SEQUENCING REVEALS TRANSCRIPT CLUSTERS WITH PROGNOSTIC POTENTIAL FOR PROSTATE CANCER

measured using RT-qPCR. Results from cohort A were used to develop a model with (combinations of) the four genes based on the comparative CT method, and the model was validated in the independent cohort B. RESULTS: The HOXC6-DLX1 score model resulted in the highest area-under-the-curve (AUC) of 0.75 (95% confidence interval: 0.70-0.80) for the diagnosis of high-grade PCa upon prostate biopsy, with a specificity of 36% and 91% sensitivity. The HOXC6-DLX1 score outperformed PCA3 in both cohorts (cohort A: AUC 0.75 vs. 0.65; cohort B: 0.73 vs. 0.62). Furthermore, the HOXC6-DLX1 score was significantly correlated with the Gleason score upon biopsy. Subgroup analysis confirmed the added value of this urine test for the prediction of high-grade PCa in patients with low serum PSA values (