Surgeon-led prostate cancer lymph node staging: pathological outcomes stratified by robot-assisted dissection templates and patient selection

Abstract

Objectives: To evaluate the perioperative, pathological, and oncological outcomes from surgeon-led pathological staging of pelvic lymph node (LN) metastases at the time of robot-assisted radical prostatectomy (RARP). Patients and Methods: Over the 6-year period of 2006–2012, three distinct pelvic LN dissection (PLND) strategies were used in chronological order at a single cancer referral hospital. Strategies were characterised by both an omission of PLND (pNx) vs inclusion decision threshold, and standard vs extended templates for patients selected for PLND. The three cohorts included: (i) omission vs standard template (04/2006–10/2007), for dominant Gleason score 4–5 or a prostate-specific antigen (PSA) level of >10 ng/mL; (ii) omission/standard vs extended template (11/2007–12/2010), for dominant Gleason score 4–5, PSA level of >10 ng/mL, any single core >7 mm, or >3 ipsilateral positive cores; and (iii) extended template with minimal exceptions (01/2011–08/2012). Standard outcomes data compared included: Clavien–Dindo complication rates, LN metrics (yield, percentage positive), and biochemical recurrence (BCR). A novel metric comprised ‘pNx regret’: the rate of pNx patients upgraded/upstaged. Exploratory analyses included selection criteria for reduced PLND templates, i.e. low-yield subsets. Results: Standard PLND yielded 8–10 LNs and a positive-LN yield of 2.2–6.2%. The addition of an extended PLND (E-PLND) significantly increased the yield to 14–20 LNs and the positive-LN yield to 17.4–18.4% (both P < 0.001). E-PLND had the highest impact on the percentage of positive LNs (%pN1) for high-risk disease (9.3 vs 32.8%, P = 0.002), modest for intermediate risk (4.2 vs 10.9%, P = 0.003), and minimal impact on low risk disease (4.1 vs 0%, P = 0.401). The combined strategies of setting a very low threshold for E-PLND and sending separate LN packets increased the LN yields (18 vs 24, P < 0.001), but did not significantly change the observed %pN1 rates by clinical risk group (P = 0.975). Efforts to reduce the need for E-PLND included omission by clinical criteria, but resulting in ‘pNx regret’ in 16–19%. A third of patients with unilateral disease and positive LNs were found to have contralateral disease. A subset of men with minimal biopsy volume Gleason score 4 + 3 had pN1 rates after E-PLND of three of 14 (21%) compared to minimal biopsy volume Gleason score 3 + 4 pN1 rates after E-PLND of 0 of 31. E-PLND takes about twice as long to perform but with no statistically significant difference in complications (5.0 vs 6.0%, P = 0.511). The 5-year BCR rates were higher for E-PLND, given the selection criteria, but not different for overall survival. Conclusions: The net benefit of E-PLND remains uncertain, and therapeutic impact will probably require a randomised trial, given the strong selection criteria. E-PLND contributes to oncological staging in a significant number of high- and intermediate-risk patients, and should be bilateral. Immediate concerns include longer operative times, but no higher complication rates.