Predictors of adverse outcomes for patients with high-risk and very high-risk prostate cancer undergoing radical prostatectomy: Results from the Canadian High-risk Prostate Cancer Collaboration
Aurinjoy Gupta1, Syed Mustafa2, Rodney H. Breau3, Jonathan Izawa4, Ross Mason2, Fred Saad5, Bobby Shayegan6, Alan I. So7, Ricardo A. Rendon2.
1Northern Ontario School of Medicine, Thunder Bay, ON, Canada; 2Department of Urology, Dalhousie University, Halifax, NS, Canada; 3Division of Urology, University of Ottawa, Ottawa, ON, Canada; 4Division of Urology, Western University, London, ON, Canada; 5Division of Urology, Université de Montréal, Montreal, QC, Canada; 6Division of Urology, McMaster University, Hamilton, ON, Canada; 7Department of Urologic Sciences, University of British Columbia, Vancouver, BC, Canada
Introduction: Radical prostatectomy (RP) for the management of patients with localized prostate cancer (PCa) has varied outcomes depending on initial clinical risk group. Even within the same risk groups, the short- and long-term outcomes for patients undergoing treatment for high- and very high-risk diseases are very variable. We describe predictors of adverse outcomes in patients undergoing RP for high- and very high-risk PCa.
Methods: This multi-institutional, retrospective study describes outcomes of patients from seven Canadian academic institutions who underwent RP for high-/very high-risk PCa. Primary endpoints were postoperative detectable prostate-specific antigen (PSA) (≥0.1 ng/mL), pathologic stage ≥T3a, lymph node disease, development of metastases, and castrate-resistant prostate cancer (CRPC). Logistic regression and Cox proportional-hazards models were used to identify prognostic indicators.
Results: A total of 702 patients who underwent RP between 2005 and 2016 were evaluated. Preoperative characteristics are shown in Table 1. Short-term outcomes included postoperative detectable PSA, pathological T stage ≥3a, and lymph node involvement. Predictors of postoperative detectable PSA and pathological lymph node involvement are shown in Figures 1 and 2, respectively. Postoperatively, 17.7 % of patients were no longer high-/very high-risk; 38.9% of Gleason scores were downgraded, while 18.7 % were upgraded. Postoperative outcomes are shown in Table 2. Undetectable PSA was achieved in 72.6% (510) of patients, development of metastases occurred in 10.4%, and development of CRPC in 5.7%, with 5.3% all-cause mortality at a median followup of 3.6 years. Development of metastases was predicted by clinical node involvement (hazard ratio [HR] 4.39, p<0.05), pathological node involvement (HR 2.08, p<0.05), and detectable postoperative PSA (HR 5.79, p<0.001). The same features were predictive of CRPC, with clinical node involvement being the strongest predictor (HR 10.7, p<0.01).
Conclusions: When evaluating patients preoperatively, PSA ≥20 ng/mL is the biggest risk factor for detectable postoperative PSA. Extended pelvic lymph node dissection was protective against postoperative detectable PSA. Clinical and pathological node-positive status, as well as detectable postoperative PSA, are strongly predictive for the development of metastases and CRPC. We hypothesize that in the era of novel therapies, patients with these risk factors should be considered for escalated treatment.
