Predictors of persistent PSA after robot-assisted radical prostatectomy: a high-volume single-center experience

This study aimed to evaluate clinical and pathological factors associated with persistent prostate-specific antigen (pPSA) after robot-assisted radical prostatectomy (RARP). This retrospective study included 1,066 patients who underwent RARP between June 2019 and February 2022. Patients were grouped according to postoperative PSA as persistent (≥ 0.1 ng/mL at 4–8 weeks) or undetectable. Clinical and pathological variables were compared, and univariable and multivariable logistic regression analyses were performed to identify independent predictors of pPSA. Persistent PSA was observed in 193 of 1,066 patients (18.1%). Patients with pPSA had significantly higher preoperative PSA levels (12.9 vs. 7.1 ng/mL; p < 0.001), higher biopsy positivity rates and higher International Society of Urological Pathology (ISUP) grades, more advanced clinical and pathological stages, and were more frequently classified as high risk according to the D’Amico system (all p < 0.001). Patients with pPSA more frequently underwent pelvic lymph node dissection and less frequently received bilateral nerve-sparing surgery. Persistent PSA was significantly associated with multiple adverse pathological features, including intraductal carcinoma, lymphovascular invasion, extraprostatic extension, perineural invasion, seminal vesicle invasion, positive surgical margins, advanced pathological stage, nodal involvement, and higher tumor burden (all p < 0.05). In multivariable analysis, preoperative PSA level, seminal vesicle invasion, positive surgical margins, presence of intraductal carcinoma, and postoperative ISUP grade ≥ 4 were identified as independent predictors of pPSA. ROC analysis identified an optimal preoperative PSA cut-off value of 10.0 ng/mL for predicting pPSA, with an AUC of 0.736, sensitivity of 62.7%, and specificity of 73.3%. Persistent PSA may reflect more aggressive tumor biology and can be independently predicted using routinely available clinical and pathological parameters. Early identification of high-risk patients may facilitate individualized follow-up and timely planning of adjuvant or salvage treatment strategies.