Pathological reporting of cystectomy lymph nodes: a retrospective analysis of experience in Paris

Abstract

Abstract Purpose Pathological evaluation of pelvic lymph node (LN) dissection (PLND) is important for management of cystectomy patients. However, challenges such as unclear interobserver variability of LN counting remain. Here, we assess interobserver variability of LN measures and their clinical utility, with a focus on variant histology. Methods We retrieved radical cystectomy cases with PLND between 2010 and 2016 and reevaluated pathological parameters; number of total and metastatic LN, LN density (LND), length of metastatic LN and metastases, extranodal extension (ENE). Results We report 96 patients: median age of 71a, 34 cases pN+, 36 cases with any extent of variant histology, median follow-up 10 months. Perivesical LN were only rarely identified, but frequently metastatic (4/9). Variant histology (34 cases) frequently exhibited LN metastasis (53% of pN+ cases). Interobserver variance was poor for total LN (kappa = 0.167), excellent for positive LN (0.85) and pN staging (0.96), and mediocre for LND (0.53). ROC analysis suggests that both LND and the sum of LN metastasis length may predict outcome (AUC 0.83 and 0.75, respectively). Conclusion Our study confirms the notion of LND as a prognostic measure, but cautions due to strong interobserver variance of LN counts. The sum length of LN metastases could be a measure that is independent of LN counts. We find that microscopically identified perivesical LN merit particular attention. In summary, our study highlights current challenges in pathological reporting of PLND, confirms previous observations and forms a basis for further studies.