Exploring the optimal treatment modality for non-small cell lung cancer after stage N2 surgery based on the SEER database and constructing a predictive model for the beneficiary population.

This study aimed to identify the optimal postoperative adjuvant treatment for operable N2 stage NSCLC and develop a predictive model for predicting survival benefit. A total of 951 N2 stage NSCLC patients from the SEER database were divided into POCT (postoperative adjuvant chemotherapy) and POCRT (postoperative adjuvant chemoradiotherapy) groups. PSM (Propensity Score Matching) was used to minimize bias. The primary endpoint was OS (overall survival). Survival rates between the two groups before and after PSM were compared using Kaplan-Meier survival curves. Robust Cox regression analysis after PSM identified independent prognostic factors, and a logistic regression prediction model was established based on the training set. Model validity was validated using AUC (Area Under Curve) and calibration curves. Clinical decision analysis was employed to evaluate the predictive model's practical value in clinical decision-making. The model incorporated Age, Sex, T stage, and LODDS as predictors. The training set AUC was 0.723, while the internal validation set and external validation set AUC values were 0.710 and 0.729, respectively, indicating good predictive capability. The patients in the POCT group were exploratively divided into high-benefit and low-benefit groups. Survival analysis revealed superior outcomes in the high-benefit cohort receiving adjuvant chemotherapy alone. In summary, this SEER-based observational analysis demonstrates that patients with operable N2-stage NSCLC achieve better outcomes with POCT alone compared to POCRT. Furthermore, the predictive model effectively quantifies patient benefit.