Novel nomogram to predict biochemical recurrence-free survival after radical prostatectomy

Abstract

Purpose: Conditional survival represents the probability of subsequent survival given that patients have already survived a certain length of time. Several models predict biochemical recurrence (BCR) after radical prostatectomy. However, none of them include postoperative prostate-specific antigen (PSA). We aimed to analyze BCR-free survival evolution over time and develop a nomogram incorporating the postoperative PSA value to predict BCR-free survival.

Material and methods: We included patients treated with robot-assisted radical prostatectomy (RARP) for prostate cancer between 2009 and 2021 and calculated conditional survival. Cox proportional hazard regression analysis was used to assess the predictive variables of BCR. We developed a nomogram predicting BCR-free survival three and five years after RARP. We used c-index and decision curve analyses to compare the nomogram with the Cancer of the Prostate Risk Assessment post-Surgical (CAPRA-S) score.

Results: We included 718 patients. The overall 3- and 5-year BCR-free survival rates were 85.1% and 75.7%, respectively. The 5-year BCR-free survival rates increased to 78.9%, 82.9%, 85.2%, and 84.7% for patients surviving 1, 2, 3, and 4 years without BCR, respectively. We developed a nomogram including the pathological Gleason score and T stage, positive surgical margin, PSA ≥ 0.05 ng/mL at one year, and lymph node involvement to predict BCR at 3 and 5 years postoperatively. Our nomogram presented a higher c-index (0.89) than the CAPRA-S score (0.78; p = 0.001) and a positive net benefit at 3 and 5 years postoperatively in the decision curve analyses.

Conclusion: The 5-year conditional BCR-free survival increased with survival without BCR. The developed nomogram significantly improved the accuracy in predicting BCR-free survival after RARP.

Keywords: Nomograms; Prostatectomy; Prostatic neoplasms; Recurrence; Survival.