Development and Validation of a Comprehensive Risk Prediction Model for Polypoid Lesions of the Gallbladder

Abstract

Polypoid lesions of the gallbladder (PLG) represent localised protrusions of the gallbladder wall, which can be either benign or malignant. Although malignant PLG is relatively rare, its prognosis is adverse. Cholecystectomy is the most efficacious treatment for malignant PLG; however, its suitability varies among PLG patients, and its indications remain controversial. To offer guidance for clinical diagnosis and treatment of PLG, 461 patients were included and classified into three subgroups based on postoperative pathological results. Logistic regression analysis was employed to identify the risk factors for PLG (the number of lesions, gallbladder wall thickness, gallbladder stones, and clinical symptoms), malignant PLG (age, polyp size, colour Doppler blood flow signal, gallbladder volume, and cholecystitis), and adenomatous PLG (CA199, the number of lesions, and gallbladder wall thickness). Multivariate logistic regression analysis was employed to construct clinical prediction models for PLG (model A, containing the number of lesions, gallbladder wall thickness, gallbladder stones, and clinical symptoms), malignant PLG (model B, containing age, polyp size, CA199, and gallbladder volume), and adenomatous PLG (model C, containing CA199, the number of lesions, and gallbladder wall thickness). Subsequently, corresponding nomograms were developed. The AUC values of all models exceeded 0.7, indicating excellent predictive efficacy. Calibration curves and DCA curves affirmed the models' reliability and validity. In conclusion, the models derived from this study demonstrate significant predictive efficacy for PLG, malignant PLG, and adenomatous PLG, respectively. They are anticipated to offer guidance for the diagnosis of PLG and provide a reliable foundation for subsequent treatment strategies.