External validation of preoperative nomograms predicting lymph node involvement in patients who underwent extended pelvic lymph node dissection during robot-assisted radical prostatectomy

Objectives: To review the quality of our extended pelvic lymph node dissection (ePLND) and to evaluate the accuracy of preoperative nomograms predicting lymph node involvement (LNI) in patients who underwent ePLND during robot-assisted radical prostatectomy (RARP). Patients and methods: The study included 83 patients who underwent ePLND during RARP between November 2012 and May 2015. We performed ePLND in patients with T3 disease, DAmico high-risk disease, and in those where the risk of LNI was higher than 5% as estimated by the Japan PC Table. Preoperative and postoperative patient data were collected retrospectively. We carried out ROC analyses for external validation of preoperative LNI-predicting nomograms. Results: The median number of lymph nodes removed was 19. LNI was found in 15 of the 83 patients (18%). In the 15 patients with nodal metastasis, a total of 28 positive lymph nodes were detected, 30% of which were in the obturator area, 17% in the external iliac area and 48% in the internal iliac area. In the external validation of LNI prediction by preoperative nomograms using ROC analyses, AUC of the Briganti nomogram was 0.80 whereas that of the Japan PC Table was 0.68. Conclusions: The number of lymph nodes removed and lymph node positivity by ePLND during RARP in our institution were comparable to previous reports based on an open radical prostatectomy series. In this study, the Japan PC Table was inferior to the Briganti nomogram for prediction of LNI, and therefore the development of an updated nomogram for Japanese patients who undergo ePLND is required.