Factors Predicting Prostate Specific Antigen Failure Following Radical Prostatectomy: Experience with 961 Patients

Purpose: To determine disease-related predictors for the occurrence of prostate specific antigen (PSA) failure in Iranian prostate cancer patients who underwent radical prostatectomy. Methods: In this cohort study, we enrolled eligible patients with prostate cancer who underwent radical prostatectomy at our center between 2001 and 2018. The primary endpoint was the incidence of postoperative biochemical failure, defined as two consecutive PSA levels ≥ 0.2 ng/dl. Patients with TNM stage ≥ III, Gleason score ≥ 8, or baseline PSA above our calculated cut-off level were considered as high risk. Kaplan–Meier survival method and Cox proportional hazards regression analysis were used for determining the biochemical relapse-free survival and its predictors. Results: Data of 959 patients (age = 61.2 ± 6.4 years) were analyzed with a median follow up of 36 months (range 6 months to 18 years). A total of 97 patients (10.1%) developed biochemical failure at the time of analysis who had a significantly older age and longer follow-up duration (P = 0.024 and P < 0.001, respectively). Preoperative PSA level of 8.85 mg/dl could predict the occurrence of biochemical failure with a sensitivity of 83.2% and specificity of 39.2% (Area under the curve = 0.601, 95% CI: 0.541-0.662; P = 0.001). In the multivariate analysis, higher preoperative PSA, Gleason score ≥ 8, and high-risk TNM stage were independent predictors for biochemical relapse (P = 0.029, P = 0.001, and P = 0.008, respectively). Conclusion: Preoperative PSA, Gleason score, and TNM stage were independent predictors for biochemical failure following radical prostatectomy in prostate cancer patients. We also determined a lower cut-point for PSA that could predict biochemical failure.