INTRODUCTION AND OBJECTIVES: Clinicians are always striving to improve their ability to predict prostate cancer (PCa) adverse pathological outcomes pre-operatively. Such information would allow a more informed clinical decision-making, and help in managing patients' expectation correctly. Unfortunately, routinely available tools, such as clinical characteristics and imaging techniques have revealed to be suboptimal in this task. Our objective was to test the performance of the Genomic signature of PCa (Decipher) in predicting adverse pathological outcomes at radical prostatectomy (RP). METHODS: We focused on 1,987 consecutive RP patients who received Decipher Test between February and August 2015. All tumors were re-reviewed by expert pathologists using the ISUP 2005 Gleason grading criteria. Decipher was calculated based on a locked random forest algorithm. Logistic regression analyses were used to evaluate the performance of Decipher in predicting 3 endpoints: >/= pT3b disease, primary Gleason grade (PGG) >/= 4, and lymph node invasion (LNI) and reported per 0.1 unit increase. Covariates consisted of pathologic Gleason score, extraprostatic extension, seminal vesicle invasion, and surgical margins. Discrimination performance of Decipher was characterized using c-index. RESULTS: Overall, 376 (19.0%), 994 (50.0%), and 51 (2.6%) had >/= pT3b disease, PGG >/= 4 and LNI, respectively. Median Decipher risk was 6.8% (interquartile range [IQR] 2.4-17.0%). In multivariable analysis (MVA), Decipher was an independent predictor of >/= pT3b disease (odds ratio [OR] 1.37; 95% confidence interval [CI] 1.05-1.09; p<0.001), PPG >/= 4 (OR 1.32; 95% CI 1.26-1.39; p<0.001), and LNI (OR 1.49; 95% CI 1.26 - 1.79; p<0.001). C-index of the MVA models incorporating Decipher in predicting the aforementioned endpoints was 0.73 (95% CI 0.70-0.76), 0.72 (95% CI 0.70-0.74) and 0.84 (95% CI 0.78-0.89), respectively. Removing Decipher from the model resulted in a c-index drop of 0.1, 0.07, and 0.05, respectively. CONCLUSIONS: The genomic signature of PCa (Decipher) can significantly increase our ability to predict adverse pathological findings at RP. This can greatly improve our treatment strategies, such as the decision to perform a nerve-sparing RP, lymph node dissection, and/or multimodal treatment. This might in turn improve cancer control outcomes in patients with really advanced disease, while sparing those with less aggressive tumors from the morbidity of such procedures.
Abdollah, F., Alter, J., Haddad, Z., Santiago, M., Yousefi, K., Lam, L., … Shah, N. L. (2016). MP07-07 PREDICTING ADVERSE PATHOLOGICAL OUTCOMES IN PROSTATE CANCER (PCA) PATIENTS UNDERGOING RADICAL PROSTATECTOMY (RP): THE ROLE OF GENOMIC SIGNATURE. Journal of Urology, 195(4S). https://doi.org/10.1016/j.juro.2016.02.2210