Medicine (Baltimore). 2026 May 08. 105(19):
e48526
The proportion of elderly patients with non-small cell lung cancer (NSCLC) undergoing surgical treatment is low. However, the prognostic factors influencing the outcomes of nonsurgical patients have not been systematically studied. Therefore, we aim to construct a prognostic prediction model for this patient population to provide a more accurate survival prediction. We conducted a retrospective analysis of patients with pathologically diagnosed NSCLC. We constructed nomograms for overall survival (OS) and cancer-specific survival (CSS) at 1, 3, and 5 years using least absolute shrinkage and selection operator regression and Cox regression analysis. The performance of the predictive models was evaluated using consistency indices, calibration curves, receiver operating characteristic curves, and decision curve analysis. Both internal and external validations were performed. A total of 18,939 NSCLC patients were included, divided into training and validation sets in a 7:3 ratio. The chi-square test indicated no statistically significant difference between the 2 datasets regarding baseline information (P > .05). Through least absolute shrinkage and selection operator regression and Cox regression analyses, we identified age, sex, marital status, American Joint Committee on Cancer stage, radiotherapy, chemotherapy, and distant metastasis as influencing factors for OS. We used these factors to construct a nomogram for OS. Similarly, we identified independent prognostic factors affecting CSS, which included sex, American Joint Committee on Cancer stage, radiotherapy, chemotherapy, and distant metastasis, and constructed a nomogram for CSS. After construction, we validated the prognostic models for OS and CSS using receiver operating characteristic curves, consistency indices, calibration curves, and decision curve analysis, which demonstrated the accuracy and reliability of our models. Finally, we confirmed the feasibility of using these models in different populations through external validation sets. This predictive model can provide a more accurate prognostic assessment for non-operated NSCLC patients over 70 years old.
Keywords: NSCLC; SEER database; elderly; nomogram; predictive modeling