Dynamic nomograms combining N classification with ratio-based nodal classifications to predict long-term survival for patients with lung adenocarcinoma after surgery: a SEER population-based study

Abstract

Background: The prognostic roles of three lymph node classifications, number of positive lymph nodes (NPLN), log odds of positive lymph nodes (LODDS), and lymph node ratio (LNR) in lung adenocarcinoma are unclear. We aim to find the classification with the strongest predictive power and combine it with the American Joint Committee on Cancer (AJCC) 8th TNM stage to establish an optimal prognostic nomogram.

Methods: 25,005 patients with T1-4N0-2M0 lung adenocarcinoma after surgery between 2004 to 2016 from the Surveillance, Epidemiology, and End Results database were included. The study cohort was divided into training cohort (13,551 patients) and external validation cohort (11,454 patients) according to different geographic region. Univariate and multivariate Cox regression analyses were performed on the training cohort to evaluate the predictive performance of NPLN (Model 1), LODDS (Model 2), LNR (Model 3) or LODDS+LNR (Model 4) respectively for cancer-specific survival and overall survival. Likelihood-ratio χ² test, Akaike Information Criterion, Harrell concordance index, integrated discrimination improvement (IDI) and net reclassification improvement (NRI) were used to evaluate the predictive performance of the models. Nomograms were established according to the optimal models. They're put into internal validation using bootstrapping technique and external validation using calibration curves. Nomograms were compared with AJCC 8th TNM stage using decision curve analysis.

Results: NPLN, LODDS and LNR were independent prognostic factors for cancer-specific survival and overall survival. LODDS+LNR (Model 4) demonstrated the highest Likelihood-ratio χ² test, highest Harrell concordance index, and lowest Akaike Information Criterion, and IDI and NRI values suggested Model 4 had better prediction accuracy than other models. Internal and external validations showed that the nomograms combining TNM stage with LODDS+LNR were convincingly precise. Decision curve analysis suggested the nomograms performed better than AJCC 8th TNM stage in clinical practicability.

Conclusions: We constructed online nomograms for cancer-specific survival and overall survival of lung adenocarcinoma patients after surgery, which may facilitate doctors to provide highly individualized therapy.

Keywords: Log odds of positive lymph nodes (LODDS); Lung adenocarcinoma; Lymph node ratio (LNR); Nomogram; SEER program.