Purpose: We sought to determine if absolute prostate specific antigen (PSA) value after 6 months of androgen deprivation therapy (ADT) is predictive of subsequent survival in patients with prostate adenocarcinoma.
Materials and methods: We performed a retrospective review of men receiving care within the Veterans Health Administration who initiated ADT for prostate adenocarcinoma. We used low- (≤0.2 ng/ml), intermediate- (>0.2 to 4 ng/ml) and high-risk (>4 ng/ml) absolute PSA values after 6-9 months of ADT, previously described in Southwest Oncology Group trial 9346. The primary endpoints were all-cause mortality and prostate cancer-specific mortality (PCSM). Kaplan-Meier survival curves for each PSA category were estimated and log-rank test was conducted. We employed Cox regression analysis adjusted for covariates and inverse propensity score weights associated with PSA categories to estimate the PSA category association with PCSM and all-cause mortality.
Results: We identified 9,170 patients in our cohort. Following ADT induction, 3,508 patients had low, 3,419 had intermediate and 2,243 had high PSA values. Two- and 5-year survival rates for low, intermediate and high PSA groups were 93.9% and 85.2% vs 88.6% and 71.2% vs 63.6% and 38.6%, respectively (p <0.0001). Patients in the high and intermediate PSA categories had a 15-fold and 3-fold higher risk of PCSM compared to those with PSA <0.2 ng/ml (p <0.0001).
Conclusions: Absolute PSA in hormone-sensitive prostate cancer after 6-9 months of ADT is a predictor of overall mortality and PCSM. This measure can rapidly assess the efficacy of new interventions in phase 2 clinical trials.
Keywords: leuprolide; mortality; prostate-specific antigen; prostatic neoplasms; survival analysis.