Evaluation of Post Robotic Prostatectomy Genomic Analysis for Predicting Prostate Cancer Specific Mortality

Abstract OBJECTIVES: Decipher genomic classification (DGC) has been developed to predict disease severity in prostate cancer. This study seeks to evaluate how well individual genes within the DGC predict the clinical outcome of post-prostatectomy disease-specific mortality (DSM). MATERIALS AND METHODS: Proportional hazards regression analyses were used to evaluate 9 genomic group signatures and 36 individual genes as potential predictors of DSM in a prospectively maintained database of 197 patients who underwent robotic prostatectomy (RALRP) with DGC testing of the post prostatectomy specimens with median 9-year follow-up. RESULTS AND CONCLUSIONS: Patient T stage was T2 62.9% and T3/T4 37.1%, and Gleason Grade Group (GGG) 1 (8.1%), GGG 2 (52.8%), GGG 3 (18.3%), GGG 4 (5.6%), GGG 5 (15.2%). Mean DGC risk score was 0.51 (SD 0.25), with patients categorized as high (38.1%), average (19.3%), and low risk (42.6%). In a multivariable proportional hazards regression model only expression of Aurora Kinase A (AURKA) and Androgen Receptor (AR) Signaling Activity were significant; DSM was 6.2 times greater for every 25-unit increase in AURKA(95%CI 1.45-26.56, p=0.014), and 0.15 times lower for every 25-unit increase in AR Signaling Activity (95% CI 0.03-0.78, p=0.024). Conclusions were unchanged after covariate adjustment for T-stage in the model, which was not significant (p=0.09). A threshold of ≥ 71 of AURKA was found to be the best threshold for predicting DSM (p=0.010). Expression of AURKA, a critical component of progression to NEPC, in post prostatectomy genomic analysis was found to be the best predictor of disease specific mortality