Open Access

Mike Wenzel, Christoph Würnschimmel, Luigi Nocera, Claudia Collà Ruvolo, Zhe Tian, Shahrokh F. Shariat, Fred Saad, Alberto Briganti, Markus Graefen, Luis Kluth, Philipp Mandel, Felix Chun, Pierre I. Karakiewicz

• Background: We hypothesized that lymph node dissection (LND) at salvage radical prostatectomy may be associated with lower cancer-specific mortality (CSM) and we tested this hypothesis. Methods: We relied on surveillance, epidemiology, and end results (2004–2016) to identify all salvage radical prostatectomy patients. Categorical, as well as univariate and multivariate Cox regression models tested the effect of LND (LND performed vs. not), as well as at its extent (log-transformed lymph node count) on CSM. Results: Of 427 salvage radical prostatectomy patients, 120 (28.1%) underwent LND with a median lymph node count of 6 (interquartile range [IQR], 3–11). According to LND status, no significant or clinically meaningful differences were recorded in PSA at diagnosis, stage and biopsy Gleason score at diagnosis, except for age at prostate cancer diagnosis (LND performed 63 vs. 68 years LND not performed, p < .001). LND status (performed) was an independent predictor of lower CSM (hazard ratio [HR] 0.47; p = .03). Similarly, lymph node count (log transformed) also independently predicted lower CSM (HR: 0.60; p = .01). After the 7th removed lymph node, the effect of CSM became marginal. The effect of N-stage on CSM could not be tested due to insufficient number of observations. Conclusions: Salvage radical prostatectomy is rarely performed and LND at salvage radical prostatectomy is performed in a minority of patients. However, LND at salvage radical prostatectomy is associated with lower CSM. Moreover, LND extent also exerts a protective effect on CSM. These observations should be considered in salvage radical prostatectomy candidates.