Machine learning effectively differentiates gallbladder polyp types.

A recent study demonstrated that a machine learning-based model significantly improves the differentiation between non-neoplastic and neoplastic gallbladder polyps. Analyzing data from 744 patients, the combined model achieved an AUC of 0.910, compared to lower scores from individual models, thus reducing unnecessary cholecystectomies. Notably, the polyp short diameter emerged as a critical predictor of neoplastic characteristics. This advancement offers a promising non-invasive tool in managing gallbladder polyps.