Examined lymph node count and survival in resected non-small cell lung cancer: attenuation of benefit beyond 16 nodes is confined to node-negative disease.

Background: Adequate lymph node (LN) examination during resection of non-small cell lung cancer (NSCLC) underpins accurate staging, but the number of examined lymph nodes (ELNs) beyond which additional retrieval confers no further benefit remains uncertain, and it is not known whether any single threshold applies across nodal stages. This study aims to systematically examine the association between ELN count, staging migration, and long-term survival to identify an evidence-based optimal ELN threshold. Methods: We analysed 61,994 patients undergoing resection for NSCLC in the program (SEER) 17 registry (2010-2019). The detection of nodal metastasis was assessed as the proportion of node-positive cases and the mean number of positive nodes across ELN categories. Non-linearity was examined with restricted cubic splines. A threshold of 16 nodes, proposed in earlier work, was pre-specified and tested using a two-piecewise Cox model rather than derived from these data; the analysis was repeated across candidate breakpoints from 10 to 30 and separately by pathological N stage and extent of resection. Internal validation used split-sample and bootstrap resampling. Results: The median ELN count was 9 (interquartile range, 5-15). Regression analyses included the 59,961 patients with 1 to 30 nodes examined, among whom 18,849 deaths occurred. After full adjustment, each increment of 5 ELNs was associated with higher odds of node-positive disease [odds ratio (OR) 1.19; 95% confidence interval (CI): 1.17-1.22] and lower mortality [hazard ratio (HR) 0.94; 95% CI: 0.92-0.95]. Below 16 nodes, each additional node was associated with a 2.1% reduction in the hazard of death (HR 0.979; 95% CI: 0.976-0.983; P<0.001), whereas above 16 no association was detectable (HR 1.001; 95% CI: 0.994-1.007; P=0.88); the two slopes differed significantly (P<0.001). No association above the breakpoint was found at any candidate value from 12 nodes upwards. On stratified analysis the two-piecewise model improved fit only in node-negative disease (P<0.001) and only after lobectomy (P<0.001); among node-positive patients the association persisted across the full range without a plateau (P=0.32). Conclusions: In node-negative NSCLC treated by anatomical resection, the survival benefit of additional nodal retrieval attenuates beyond approximately 16 examined nodes, and this count may serve as a practical benchmark for auditing the adequacy of nodal examination. No comparable ceiling was identified in node-positive disease, in which more extensive retrieval remained associated with lower mortality. As an observational registry analysis, these findings describe associations rather than causal effects.