Outcome predictors of radical prostatectomy in patients with prostate-specific antigen greater than 20 ng/ml: a European multi-institutional study of 712 patients.

BACKGROUND Prostate cancer (PCa) patients with pretreatment prostate-specific antigen (PSA) >20 ng/ml have a high risk of biochemical and clinical failure and even cancer-related death after local therapy. Pretreatment predictors of outcome after radical prostatectomy (RP) in this patient group are necessary. OBJECTIVE Our aim was to assess how the use of additional high-risk factors (biopsy Gleason score [bGS] > or = 8 or clinical stage 3-4) can improve prediction of treatment failure and cancer-related death after RP in patients with PSA >20. DESIGN, SETTING, AND PARTICIPANTS In a retrospective multicentre cohort study from six European centres between 1987 and 2005, 712 patients with PSA >20 ng/ml underwent RP and bilateral pelvic lymphadenectomy. MEASUREMENTS Subgroups were analysed to determine the relationship between the number of high-risk factors and histopathology, biochemical progression-free survival, clinical evidence of progressive disease, prostate cancer-specific mortality (PCSM), and overall mortality. Kaplan-Meier analysis with log-rank test and Cox multivariable analysis were applied. RESULTS AND LIMITATIONS Median follow-up was 77 mo. The number of high-risk factors was significantly associated with unfavourable histopathology. Among patients with only PSA >20 ng/ml, 33% had pT2 PCa, 57.9% had bGS <7, 54% had negative surgical margins, and 85% were lymph node negative (pN0), whereas among patients with all three high-risk factors, 4.5% had pT2 PCa, 2.3% had bGS <7, 20.5% had negative margins, and 49% were pN0 (p<0.001). The strongest predictor of progression and mortality was bGS. PSA >20 ng/ml associated with bGS < or =7 resulted in 10-yr PCSM of 5%; when associated with bGS > or =8, PCSM was 35%. The main limitations of the study were retrospective design and varying treatment modalities. CONCLUSIONS PCa patients with PSA >20 ng/ml have varying risk levels of disease progression and PCSM. Considering additional risk factors further stratifies this group into four subgroups that can guide the clinician in preoperative patient counselling.