Development of a Combined Risk Factors Prediction Model for Esophageal Squamous Cell Carcinoma: A Secondary Analysis of the Linxian Nutrition Intervention Trial Cohort

Background Early screening and detection is essential to reduce morbidity and mortality in esophageal cancer (EC), particularly for individuals at high risk. To reduce the burden and high costs of whole-population screening, we developed a risk score model for individualized risk assessment of esophageal squamous cell carcinoma (ESCC) incidence. Methods The study was conducted using the Linxian Nutrition Intervention Trial cohort. Cox regression and the points system method were used to build a score-based model for ESCC risk prediction. The receiver operating characteristic (ROC) curve and calibration curve were used to examine the distinction and calibration of the models. Results A total of 29,408 participants were included in final analysis. During the 10-year follow-up period, 1386 ESCC new cases were identified. Cox regression showed that increasing age, smoking, family history of esophageal cancer, dysphagia, fresh vegetable consumption (≤ 1 time/day), low body mass index (BMI < 18.5kg/m ² ), not drinking tap water (versus untreated natural water), and tooth loss were independent risk factors of ESCC incidence. The risk score based on 8 risk factors ranged from 0 to 59 points. Compared to subjects with a risk score < 20 points, the ESCC incidence risk increased by 201% for 20 to 39 points and 615% for score of over 39 points. The area under the curve (AUC) value of the risk score estimating ESCC incidence within 3 years was 0.70 (95% CI: 0.67–0.72). Conclusions Our model effectively stratified the risk of ESCC, demonstrating a potential application in high-risk population identification and ESCC prevention.