Background: To determine whether a variable definition of biochemical recurrence (BCR) based on clincopathologic features facilitates early identification of patients likely to suffer from disease progression. The definition of BCR after radical prostatectomy (RP) bears important implications for patient counseling and management; however, there remains a significant debate regarding the appropriate definition.
Methods: The study cohort consisted of 3619 men who underwent RP for localized prostate cancer from 1989 to 2007, with data abstracted from the Cancer of the Prostate Strategic Urologic Research Endeavor (CaPSURE) registry. Patients were stratified into three risk groups according to Cancer of the Prostate Risk Assessment post-Surgical (CAPRA-S) score. Three single threshold PSA cut-points for BCR were evaluated (PSA > or =0.05, > or =0.2 and > or =0.4 ng ml(-1)) as well as a variable cut-point defined by risk group. After reaching the cut-points, patients were followed for further PSA progression.
Results: The proportion of patients with BCR differed by cut-point and risk group, ranging from 7 to 37% (low risk), 22 to 58% (intermediate risk) and 60 to 86% (high risk). The positive-predictive value (PPV) for predicting further PSA progression was 49% for the PSA > or =0.05 ng ml(-1), 62% for the PSA > or =0.2 ng ml(-1), 65% for the PSA > or =0.4 ng ml(-1) and 68% for the risk-adjusted definition. Five-year progression-free survival was 39% for the risk-adjusted definition compared with 45-52% for the other definitions of BCR.
Conclusions: These data suggest that a variable definition of BCR determined by clinicopathologic risk may improve the identification of early recurrence after RP without increasing the overdiagnosis of BCR. By using a risk-adjusted BCR definition, clinicians can better predict future PSA progression and more appropriately counsel patients regarding salvage therapies.