Development and Validation of a Preoperative Prediction Model for Neoplastic Gallbladder Polyps

Area of Science:

• Gastroenterology and Hepatology
• Surgical Oncology
• Medical Imaging

Background:

• Gallbladder polypoid lesions (GPLs) require evaluation to distinguish neoplastic polyps (NP) from benign ones.
• Accurate preoperative identification of NP is crucial for appropriate patient management and surgical planning.
• Existing risk factor studies for NP have not resulted in a practical predictive model.

Purpose of the Study:

• To develop and validate a practical preoperative prediction model for neoplastic polyps (NP) in gallbladder polypoid lesions (GPLs).
• To utilize simple and easily accessible clinical variables for predicting NP.
• To improve clinical decision-making in the management of GPLs.

Main Methods:

• Retrospective analysis of 621 patients with GPLs who underwent cholecystectomy.
• Development of a logistic regression model incorporating age, polyp size, polyp number, gallbladder wall thickness, and polyp echo characteristics.
• Validation of the model using training, internal, and external datasets with assessment of discrimination, calibration, and clinical utility.

Main Results:

• Key predictors for NP included age, polyp size, polyp number, gallbladder wall thickness, and polyp echo characteristics.
• The nomogram model achieved high predictive accuracy with AUCs of 0.886 (training), 0.836 (internal validation), and 0.867 (external validation).
• The model demonstrated good calibration across all datasets and significant clinical benefit at a threshold probability of 0.6.

Conclusions:

• A practical preoperative prediction model incorporating accessible clinical variables effectively identifies neoplastic polyps in gallbladder polypoid lesions.
• The developed nomogram model shows excellent performance and aids in clinical decision-making for GPL management.
• This tool can assist clinicians in stratifying risk and determining the optimal management strategy for patients with GPLs.