372 THE NUMBER OF POSITIVE NODES IS THE STRONGEST PREDICTOR OF CANCER SPECIFIC SURVIVAL IN PATIENTS TREATED WITH RADICAL PROSTATECTOMY FOR PATHOLOGICAL T3 PROSTATE CANCER

Abstract

INTRODUCTION AND OBJECTIVES: Previous studies demonstrated that patients (pts) with one or two positive nodes have a significantly better cancer specific survival (CSS) compared to pts with three or more nodes involved at radical prostatectomy (RP). However the impact of the number of positive nodes on CSS have never been tested in pts with locally-advanced PCa. The aim of the study was to examine CSS according to the number of positive nodes in pts harboring pT3 PCa at RP. METHODS: We identified 1562 pT3 pts treated with RP and extended pelvic lymph node dissection between 1988 and 2012 at our institution. Pts were stratified according to the absence of seminal vesicle invasion (SVI) vs presence of SVI and according to the number of positive nodes: 0 (N0), 1-2 (N1a) and 3 (N1b). Kaplan-Meier analyses assessed CSS. Multivariable (MVA) Cox regression analyses was used to test the impact of number of positive nodes on CSS. Covariates included patient age, surgical margin (SM) status, pathological Gleason score (GS), and adjuvant therapy. RESULTS: Overall, SVI was reported in 806 (51.6%) pts and 990 (63.4%), 366 (23.4%), and 206 (13.2%) pts were N0, N1a, and N1b, respectively. Mean follow-up time was 71 months (median 60 months). The mean number of nodes removed was 17 (median: 16, range: 1-66). The 5 and 10 year CSS after RP was 94.5% and 86.5%. After stratification according to the number of positive nodes, 5 and 10 year CSS was 96% and 86% for N0, 93% and 78% for N1a, 78% and 61% for N1b (all p 0.02). At MVA, no difference between N0 and N1a pts was recorded (p 0.3). Conversely N1b pts were 3 times more likely to die for PCa than N0 pts (HR:3.2, p 0.001). After stratification according to stage, the 5 and 10 year CSS was 99% and 89% for N0, 96 and 77% for N1a, 60% and 47% for N1b in the group without SVI (all p 0.04). Conversely, in pts with SVI, no significant difference was observed in CSS at 5 and 10 years between N0 and N1a pts (92vs82% for N0 and 93% and 77% for N1a respectively, p 0.7), while CSS was significantly lower in N1b pts (77% and 60% respectively, p 0.001). At MVA, no significant difference was observed between pts with N0 and N1a both in pts with or without SVI (all p 0.2). Conversely N1b pts had a significantly higher risk of dying of PCa compared to N0 and N1a pts (HR 14.2 and 2.1 for pts with and without SVI respectively, all p 0.02). CONCLUSIONS: In conclusion in pT3 pts, the presence of a high nodal burden is associated with poor prognosis. However, a limited (2 or less) number of nodes involved does not compromise cancer control in pts with locally advanced disease.