Development and external validation of a nomogram to predict the risk of Upper gastrointestinal precancerous lesions in a non-high-incidence area.

Background: Upper gastrointestinal precancerous lesions (UGPL) is the major preventable disease in non-high-incidence area. A prognostic nomogram was constructed to predict and identity susceptible population of UGPL before endoscope screening.

Methods: We recruited 300 ,016 eligible participants for upper gastrointestinal cancer (UGC) screening aged 40-74 years from two cities in Hunan province from 2012 to 2019. Individuals at high risk of UGC on basis of questionnaire estimation underwent endoscopic screening. Participants in two cities accepting endoscopy were used as training and external validation cohorts, respectively. A nomogram was developed based on independent prognostic factors of UGPL determined in multivariable logistic regression analysis.

Results: Of 35, 621 with high risk for UGC, 10, 364 subjects undertook endoscopy (participation rate of 29.1%). The detection rate for UGPL was 4.55%. The nomogram showed that age, gender, mental trama, picked food, and atrophic gastritis history in a descending order were significant contributors to UGPL risk. The C-index value of internal and external validation of the model is 0.612 and 0.670, respectively. The calibration data for UGPL showed optimal agreement between the nomogram prediction and actual observation. Furthermore, high-risk and low-risk group divided based on score from the nomogram predicted a significantly distinct detection rate.

Conclusion: The nomogram provides screening workers a simple and accurate tool for identifying individuals at a higher risk of UGPL as primary screening before endoscopy among Chinese population in non-high-risk areas, thus reducing the incidence of UGC by improving the UGPL detection.