Risk factors and prognostic nomogram for patients with second primary cancers after lung cancer using classical statistics and machine learning

Abstract

Background: Previous studies have revealed an increased risk of secondary primary cancers (SPC) after lung cancer. The prognostic prediction models for SPC patients after lung cancer are particularly needed to guide screening. Therefore, we study retrospectively analyzed the Surveillance, Epidemiology, and End Results (SEER) database using classical statistics and machine learning to explore the risk factors and construct a novel OS prediction nomogram for patients with SPC after lung cancer.Methods: Data of patients with SPC after lung cancer, covering 2000 to 2016, were gathered from the Surveillance, Epidemiology, and End Results (SEER) database. The incidence of SPC after lung cancer was calculated by Standardized incidence ratios (SIRs). Cox proportional hazards regression, Kaplan-Meier methods and log-rank tests were conducted to identify the independent prognostic factors for predicting OS. These significant prognostic factors were used for the development of an OS prediction nomogram. Results: Totally, 10487 SPC samples gathered from the SEER database were divided into training cohort and validation cohort (score construction and internal validation) at random. Multivariate Cox regression analysis showed that age, sex, race, grade, marital status, CT (chemotherapy), TNM stage, radiation, lung cancer surgery and SPC surgery were independent risk factors for OS in SPC patients. The prognostic nomogram we constructed was also for OS. The concordance index (C-index) in the training cohort and validation cohort were 0.715(95%CI:0.710-0.718) and 0.716(95%CI:0.712-0.720). Additionally, calibration curves and decision curve analysis (DCA) curves revealed that the nomogram has excellent clinical utility. In the random forest (RF) model, SPC site and T stage were the two most important variables, which was consistent with the nomogram. And, the random forest (RF) model had slightly better prediction performance than the Cox risk regression model (AUC: 0.748 vs. AUC: 0.719).Conclusions: The prognosis characteristics of SPC following the lung cancer was systematically reviewed. The nomogram we constructed can provide survival predictions to assist clinicians in making individualized decisions for appropriate treatment.