The nomograms for predicting overall and cancer-specific survival in elderly patients with early-stage lung cancer: A population-based study using SEER database

PURPOSE: Lung cancer is the leading cause of death from cancer and the number of operable elderly lung cancer patients is increasing, with advanced age being associated with a poorer prognosis. However, there is no easy and comprehensive prognostic assessment method for these patients. METHODS: Clinicopathological data of patients aged 65 years or older with TNM stage I-II lung cancer from 2004 to 2018 were downloaded from the SEER database. Patients from 2004 to 2015 were randomized into a training group (n = 16,457) and a validation group (n = 7,048). Data from 2016 to 2018 (n = 6,231) were used for external validation. Two nomogram prognostic models were created after independent prognostic factors connected to both overall survival (OS) and cancer-specific survival (CSS) in the training set by using univariate and multivariate Cox proportional hazards regression analysis. In turn, overall survival (OS) and cancer-specific survival (CSS) were predicted for patients at 1, 3, and 5 years. Based on the concordance index (C-index), calibration curves, area under the receiver operating characteristics (ROC) curve (AUC), the time-dependent area under the ROC curve, the validity, accuracy, discrimination, predictive ability, and clinical utility of the models were evaluated. Decision curve analysis (DCA) was used to assess the clinical value of the models. RESULTS: A total of 29,736 patients were included. Univariate and multivariate analyses suggested that age, race, gender, marriage, disease grade, AJCC stage, T-stage, surgery, radiotherapy, chemotherapy, and tumor size were independent risk factors for patient prognosis. These 11 variables were included in nomogram to predict OS and CSS of patients. C-indexes of OS for the training, validation and external validation sets were 0.730 (95% CI, 0.709–0.751), 0.734 (95% CI, 0.722–0.746), and 0.750 (95% CI, 0.734–0.766), respectively. The AUC results for the training and validation sets indicated good accuracy for this nomogram. The calibration curves demonstrated a high degree of concordance between actual and anticipated values, and the DCA demonstrated that the nomograms had better clinical application than the traditional TNM staging approach. CONCLUSION: This study identified risk factors for survival in operable elderly lung cancer patients and established a new column line graph for predicting OS and CSS in these patients. The model has good clinical application and can be a good clinical decision-making tool for physicians and patients.