OBJECTIVES: To assess the ability of preoperative prostate-specific antigen level, Gleason score and stage to predict prostate cancer outcomes beyond biochemical recurrence, specifically castration-resistant prostate cancer, metastases and prostate cancer-specific mortality in radical prostatectomy patients.

METHODS: We carried out a retrospective study of 2735 men in the Shared Equal Access Regional Cancer Hospital database treated by radical prostatectomy from 1988 to 2011 with data available on pathological stage, grade and preoperative prostate-specific antigen. We used Cox hazards analyses to examine the predictive accuracy (c-index) of the preoperative prostate-specific antigen (log-transformed), path Gleason score ( < /= 7, 3 + 4, 4 + 3 and 8-10) and path stage grouping (pT2 negative margins; pT2 positive margins; pT3a negative margins; pT3a positive margins; pT3b; vs positive nodes) to predict biochemical recurrence, castration-resistant prostate cancer, metastases and prostate cancer-specific mortality.

RESULTS: Median follow up was 8.7 years, during which, 937 (34%) had biochemical recurrence, 108 (4%) castration-resistant prostate cancer, 127 (5%) metastases and 68 (2%) prostate cancer-specific mortality. For the outcomes of biochemical recurrence, castration-resistant prostate cancer, metastases and prostate cancer-specific mortality, the c-indices were, respectively: prostate-specific antigen 0.65, 0.66, 0.64 and 0.69; Gleason score 0.66, 0.83, 0.76 and 0.85; and pathological stage group 0.69, 0.76, 0.72 and 0.80.

CONCLUSIONS: Gleason score can predict with very high accuracy prostate cancer-specific mortality in patients undergoing radical prostatectomy. Thus, Gleason score should be given more weight in nomograms to predict prostate cancer-specific mortality. Furthermore, men with a high Gleason score should be given special consideration for adjuvant treatment or referral to clinical trials because of a higher risk of prostate cancer-specific mortality.