Presentation Authors: Felix Preisser*, Frankfurt, Germany, Roderick van den Bergh, Amsterdam, Netherlands, Gorgio Gandaglia, Milan, Italy, Piet Ost, Ghent, Belgium, Christian Sucrel, Bucharest, Romania, Prasanna Sooriakumaran, London, United Kingdom, Francesco Montorsi, Milan, Italy, Markus Graefen, Hamburg, Germany, Henk van der Poel, Amsterdam, Netherlands, Alexandre de la Taille, Creteil, France, Alberto Briganti, Milan, Italy, Laurent Salomon, Guillaume Ploussard, Creteil, France, Derya Tilki, Hamburg, Germany
Introduction: To assess the effect of extended pelvic lymph node dissection (PLND) on oncologic outcomes in prostate cancer (PCa) patients with D'Amico intermediate- or high-risk characteristics treated with radical prostatectomy (RP).
Methods: 9,742 patients who underwent RP between 2000 and 2017 with or without PLND were identified in a multi-institutional database of four centers. Only patients with more than 5% risk of lymph node invasion according to the Briganti nomogram were included. Univariable and multivariable Cox regression models tested the effect of extended PLND on biochemical recurrence (BCR), clinical recurrence (CR) and cancer-specific mortality (CSM). All analyses were repeated after propensity score matching. Matching was performed for pathologic primary Gleason, pathologic tumor stage, prostate specific antigen (PSA) and surgical margin.
Results: Overall, 707 (7.3%) patients did not undergo PLND. Of those, 520 vs. 187 harbored D'Amico high- or intermediate-risk characteristics, respectively. Individuals without PLND had significantly lower median PSA and more frequently harbored biopsy Gleason score â‰¤6 as well as pathologic T2 stage. The median number of removed lymph nodes was 14 in the PLND cohort; 17.6% (n=1,714) of the patients with PLND harbored lymph node metastases. BCR-free rates were 64.2% vs. 50.5%, CR-free rates were 90.1 vs. 73.8% and CSM-free rates were 96.5 vs. 93.1%, at 120 months after RP for no PLND vs. PLND, respectively. After adjustment for pathologic co-variables, PLND did not achieve independent predictor status for BCR, CR and CSM (all p>0.05). After 2:1 propensity score matching, BCR-free rates were 62.1 vs. 63.9%, CR-free rates were 88.8 vs. 90.0% and CSM-free rates were 96.9 vs. 96.4% at 120 months after RP for PLND and no PLND, respectively. Multivariable Cox regression models after matching revealed that PLND is no independent predictor for BCR, CR and CSM (all pâ‰¥0.5).
Conclusions: Patients with D'Amico high- or intermediate-risk prostate cancer without PLND at radical prostatectomy did not show worse oncologic outcome compared to patients who underwent PLND. While PLND is the most accurate staging procedure, its therapeutic value remains unclear.