Pathologically node-positive prostate cancer after Robot-Assisted Radical Prostatectomy: current patterns of nodal involvement and clinical implications

Background. pN1 after RARP represents a heterogeneous entity in which prognosis depends on tumor biology, nodal burden, and early postoperative PSA kinetics. ePLND remains the reference for accurate nodal staging, though detailed real-world data on the anatomical topography of positive nodal packets and its clinical correlates are limited. We aimed to describe the distribution of metastatic lymph nodes after RARP+ePLND and to investigate predictors of PSA persistence. Methods. We performed a single-centre observational cohort study including consecutive pN1 patients treated with RARP plus ePLND at Santa Chiara Hospital (Trento, Italy) between 2015 and 2024. PSA persistence was defined as PSA ≥0.1 ng/mL at two consecutive postoperative assessment 4-8 weeks after surgery. Clinicopathologic, perioperative, and follow-up variables were extracted from a prospectively maintained database. Comparisons between PSA-persistent and non-persistent groups used Mann–Whitney U and χ²/Fisher’s exact tests. Multivariable logistic regression explored independent predictors of PSA persistence. Results. Among 1,523 RARP procedures, 625 patients underwent ePLND; 111 were pN1 and formed the study cohort. Median nodal yield was 21 (IQR 16–25); 72.0% had ≤2 positive nodes and 27.0% had bilateral nodal involvement. Major complications (Clavien–Dindo ≥3) occurred in 6.3%. Station-specific mapping was feasible in 84/111 patients (75.7%); the most frequently involved regions were obturator (40.5%), fossa of Marcille (39.3%), external iliac (36.9%), and internal iliac (28.6%), whereas preprostatic (6.0%) and para-aortic (1.2%) involvement was rare. Median follow-up was 59 months (IQR 30–82). PSA persistence occurred in 65.7%. At follow-up, cancer-specific survival was 96.3% and overall survival 88.9%; most patients received multimodal adjuvant treatment (RT+ADT 60.3%). On multivariable analysis, ISUP ≥4 (OR 4.47), preoperative PSA >20 ng/mL (OR 3.30), and number of positive nodes (OR 1.49 per node) independently predicted PSA persistence, while nodal yield did not. Conclusions. In pN1 patients undergoing RARP with ePLND, metastatic nodes clustered predominantly in the obturator region and Marcille’s fossa, with frequent involvement of external/internal iliac stations. PSA persistence was common and was independently driven by aggressive tumour biology and increasing nodal burden rather than the extent of nodal yield. Integrating nodal burden, anatomical topography, and early PSA response may improve postoperative risk stratification and guide personalized multimodal management.