Abstract: Objective · To explore the prognostic factors for elderly patients after lung cancer surgery, and establish a prognostic nomogram to predict their prognosis with multiple factors. Methods · The surveillance, epidemiology and end results (SEER) database was used to obtain clinical data of elderly patients after lung cancer surgery, and Kaplan-Meier method was performed to calculate overall survival (OS) and lung cancer-specific survival (LCSS). Cox univariate and multivariate analysis was displayed to determine the independent predictors of postoperative lung cancer in elderly patients, and the independent predictors were selected to draw nomogram. In addition, the internal validation and effectiveness of the model were performed by Bootstrap method, concordance index (C-index) and ROC curve, the discriminative ability was compared with tumor node metastasis classification (TNM classification) system. Results · A total of 14 891 patients were included in the study. The age, gender, the primary site of tumor, tumor pathological type, TNM classification and the number of regional lymph nodes cleaning were independent predictors of elderly postoperative patients with lung cancer. The prediction model could accurately predict the prognosis of patients, and the accuracy was higher than that of TNM classification system (C-index=0.720 vs C-index=0.686). Conclusion · This study establishes a reliable prognostic model for elderly patients after lung cancer surgery based on SEER. The predictive performance is excellently demonstrated by internal validation.

Key words: lung cancer, elderly patient, prognosis, nomogram, SEER database