Impact of the extent of negative lymph nodes in gastric adenocarcinoma undergoing primary surgical resection: An institutional report

Abstract

Background: Sub-total/total gastrectomy with lymph node dissection (LND) remains an effective therapeutic strategy for resectable gastric adenocarcinomas (GACs). Despite the prognostic significance of positive lymph nodes (PLNs) defined in N-status, few have appraised the impacts of negative lymph nodes (NLNs) and the percentage of NLN (=number of NLNs/number of total lymph nodes [TLNs], %), as well as the extent of TLNs to be dissected in GACs.

Methods: We retrospectively analyzed 62 GAC patients (mean age of 67.1 years; 41 men) undergoing primary sub-total/total gastrectomy from a single institute. Candidate variables, including the number of NLNs (≤9 and >9) and the percentage of NLN (≤37.5, 37.5-80.6 and >80.6, %), were evaluated to determine their prognostic impacts and hazard ratios (HRs).

Results: Under the multivariate Cox proportional-hazards regression model, tumor length exceeding 4 cm (p = 0.017; HR = 2.828), perineural invasion (p = 0.037; HR = 3.182), and lower percentage of NLN (p = 0.016 and p = 0.060; HRs = 1.000, 0.327, and 0.333 for subgroups ≤37.5, 37.5-80.6, and >80.6, respectively) were three independent predictors with elevated HRs for poor prognosis. GAC patients with the percentage of NLN > 80.6 were highly related to those with NLNs > 9 (p < 0.001), and GAC patients with NLNs > 9 were highly related to those with TLNs > 15 (p < 0.001). For all 62 GAC or 42 N(+) GAC patients, those who underwent LND with TLNs>15 tended to have more PLNs (p = 0.018, p = 0.003) and more NLNs (p < 0.001, p = 0.029) than did those with TLNs ≤ 15. Among the 42 GAC patients with TLNs > 15, a lower percentage of NLN (p = 0.026 and p = 0.015; HRs = 1.000, 0.272, and 0.180 for subgroups ≤37.5, 37.5-80.6, and >80.6, respectively) remained an independent predictor of poor prognosis.

Conclusion: The percentage of NLN could predict the prognosis of GAC patients properly. However, an accurate percentage of NLN needs a minimal requirement of TLNs > 15 to detect an adequate number of PLNs and sufficient number of NLNs.