New machine learning models improve diagnosis of obstructive jaundice

A comprehensive study analyzing 5,726 patients with obstructive jaundice has identified key etiologies, revealing cholangiocarcinoma, pancreatic adenocarcinoma, and common bile duct stones as predominant causes. Despite traditional markers lacking accuracy, two machine learning models were successfully developed, achieving an area under the receiver operating characteristic curve (AUROC) of 0.862 for differentiating benign and malignant causes. This innovative approach offers clinicians a rapid, reliable diagnostic tool while enhancing decision-making transparency.