Determining the optimal number of examined lymph nodes for accurate staging of pancreatic cancer: An analysis using the nodal staging score model

Introduction: The aim of this study was to determine the optimal number of examined lymph nodes (ELNs) for accurate staging of pancreatic cancer using the nodal staging score model. Materials and methods: Clinicopathological data for patients with resected pancreatic cancer were collected from SEER database (development cohort [DC]) and Fudan University Shanghai Cancer Center database (validation cohort [VC]). Multivariable models were constructed to assess how the number of ELNs was associated with stage migration and overall survival (OS). Using the beta-binomial distribution, we developed a nodal staging score model from the DC and tested it with the VC. Results: Both cohorts exhibited significant proportional increases from node-negative to node-positive disease (DC: odds ratio [OR], 1.047; P <0.001; VC: OR, 1.035; P< 0.001) and improved OS (DC: hazard ratio [HR], 0.982; P<0.001; VC: HR, 0.979; P< 0.001) as ELNs increased. Nodal staging scores escalated separately as ELNs increased for different tumor (T) stages, with plateaus at 16, 21, and 23 LNs (cut-offs) for T1, T2, and T3 tumors, respectively. Multivariable analysis indicated that examining more LNs than the corresponding cut-off value was a significant survival predictor (DC: HR, 0.813; P< 0.001; VC: HR, 0.696; P= 0.028). Conclusion: The optimal number of ELNs for adequate staging of pancreatic cancer was related to T stage. We recommend examining at least 16, 21, and 23 LNs for T1, T2, and T3 tumors, respectively, as a nodal staging quality measure for both surgery and pathological analysis. (C) 2019 The Authors. Published by Elsevier Ltd.