Construction And External Validation Of A Survival Prediction Model For Early Stage Non-small Cell Lung Cancer Patients After Surgery Based On The SEER Database

Objective: Nomogram have been widely used to predict cancer prognosis.Therefore,we attempted to develop a new survival prediction model for patients with early-stage non-small cell lung cancer.The purpose of this study was to construct a prediction model that can well predict the overall survival of patients with early-stage non-small cell lung cancer(NSCLC).Methods: A total of 26941 patients with stage Ⅰ and Ⅱ non-small cell lung cancer were sampled from the Surveillance,Epidemiology and End Results(SEER)database and randomly divided into training and validation groups in a ratio of 7:3.Cox proportional risk models were used to identify covariates associated with OS.Nomogram were constructed using significant covariates,and the constructed Nomogram were validated internally by 1000 Bootstrap replicate samples and externally by using 505 patients recruited from Jiaxing First Hospital.The predictive accuracy and discriminative power of the Nomogram were then examined by C-index and calibration curves.Finally,the accuracy of the Nomogram was compared with the tumour-lymphnode-metastasis(TNM)clinical staging to derive the corresponding area under the curve and the composite improvement index.Results: Age,gender,histological type,tumour size,anatomical site,degree of differentiation,AJCC stage,number of lymph node dissections,positive lymph nodes,radiotherapy,and type of surgery were identified as independent factors affecting OS by univariate and multifactorial factor analysis.Norman plots were then created based on these factors,with a C-index of 0.726(95% CI,0.718-0.735)and 0.721(95% CI,0.709-0.734)for the training and test sets,respectively,and a C-index of 0.758(95% CI,0.691-0.825)in external validation,indicating that the model has good discrimination.The calibration plots showed the best agreement between the survival probabilities predicted by the Norman plots and those actually observed.In addition,the DCA analysis showed a significant increase in the net benefit of the new model compared to the TNM staging system.Compared to the TNM staging system,the new model has an IDI improvement index of 2.3% 、 4.3% and 4.9% at 1,3 and 5 years respectively,indicating that the new model has better overall discriminatory power.Conclusion: We have developed and externally validated a survival prediction model for patients with early stage non-small cell lung cancer.This new Norman chart is superior to the traditional TNM staging system and can guide clinicians in making the best clinical decisions.