Selection of high‐risk individuals for esophageal cancer screening: A prediction model of esophageal squamous cell carcinoma based on a multicenter screening cohort in rural China

Abstract The mortality benefit of esophageal squamous cell carcinoma (ESCC) screening has been reported in several studies; however, the results of ESCC screening programs in China are suboptimal. Our study aimed to develop an ESCC risk prediction model to identify high‐risk individuals for population‐based esophageal cancer screening. In total, 86 745 participants enrolled in a population‐based esophageal cancer screening program in rural China between 2007 and 2012 were included in the present study and followed up until December 31, 2015. Models for identifying individuals at risk of ESCC within 3 years were created using logistic regressions. The area under the receiver operating curve (AUC) was determined to estimate the model's overall performance. A total of 298 individuals were diagnosed with ESCC within 3 years after baseline. The model of ESCC included the predictors of age, sex, family history of upper gastrointestinal cancer, smoking status, alarming symptoms of retrosternal pain, back pain or neck pain, consumption of salted food and fresh fruits and disease history of peptic ulcer or esophagitis (AUC of 0.81; 95% confidence interval: 0.78‐0.83). Compared to the current prescreening strategy in our program, the cut‐off value of 10 in the score‐based model could result in 3.11% fewer individuals subjected to endoscopies and present higher sensitivity, slightly higher specificity and lower number needed to screen. This score‐based risk prediction model of ESCC based on eight epidemiological risk factors could increase the efficiency of the esophageal cancer screening program in rural China.