External Validation of the Pathologic Nodal Staging Score for Prostate Cancer: A Population-based Study

Background: We sought to externally validate our pathologic nodal staging score (pNSS) model, which allows for quantification of the likelihood that a pathologically node-negative patient will not have lymph node (LN) metastasis after radical prostatectomy for prostate cancer (PCa) in a population-based cohort. Patients and Methods: We analyzed data from 50,598 patients treated with radical prostatectomy and pelvic LN dissection using the Surveillance, Epidemiology, and End Results database. We estimated the sensitivity of pathologic nodal staging using a β-binomial model and developed a novel pNSS model, which represents the probability that a patient's PCa has been correctly staged as node negative as a function of the number of examined LNs. These findings were compared against those from the original cohort of 7135 patients. Results: The mean and median number of LNs removed was 6.5 and 5, respectively (range, 1-89; interquartile range, 2-8), and 96.9% of the patients (n = 49,020) had stage pN0. Similar to the original cohort, the probability of missing a positive LN decreased with the increasing number of LNs examined. In both the validation and the original cohort, the number of LNs needed to correctly stage a patient's disease as node negative increased with more advanced tumor stage, higher Gleason sum, positive surgical margins, and higher preoperative prostate-specific antigen levels. Conclusion: We have confirmed that the number of examined LNs needed for adequate nodal staging in PCa depends on the pathologic tumor stage, Gleason sum, surgical margins status, and preoperative prostate-specific antigen. We externally validated our pNSS in a population-based cohort, which could help to refine decision-making regarding the administration of adjuvant therapy.