Abstract Background This study aimed to develop and validate a novel lymph node staging system integrating anatomical location and quantitative characteristics, evaluate its prognostic prediction efficacy in non-small-cell lung cancer (NSCLC), and establish a multivariate prognostic model. Methods A total of 23,676 patients with NSCLC from the SEER database (2010–2015) were enrolled. Optimal cutoffs for lymph node parameters (NPLN, LNR, LODDS) were determined using X-tile software. Composite variables (N-NPLN, N-LNR, N-LODDS) were constructed by integrating N staging. Independent prognostic factors were screened via Cox regression, and a nomogram was developed. Performance was assessed using the receiver operating characteristic curves, calibration curves, and decision curve analysis. Results N-LODDS staging demonstrated optimal prognostic prediction, significantly outperforming N-LNR and N-NPLN. The nomogram incorporating N-LODDS, tumor size, and nine independent prognostic factors showed superior discrimination and calibration (5 year area under the curve 0.740; 95% confidence interval 0.731–0.749) in both training and validation cohorts, with significant advantages over the TNM staging system (all P <0.001). Conclusion The N-LODDS staging system significantly improves prognostic accuracy by integrating anatomical and quantitative lymph node features, providing a novel tool for personalized NSCLC management. Future multicenter prospective studies are needed to validate its clinical utility.
Chen et al. (2026) studied this question.