Machine learning to predict pediatric choledocholithiasis: A Western Pediatric Surgery Research Consortium...

Area of Science:

• Pediatric Gastroenterology
• Computational Medicine
• Medical Informatics

Background:

• Pediatric choledocholithiasis, or common bile duct stones, poses diagnostic challenges.
• Accurate prediction is crucial for timely intervention and management in children.

Purpose of the Study:

• To develop and validate a machine learning algorithm for predicting pediatric choledocholithiasis.
• To identify key clinical factors that accurately predict the presence of common bile duct stones in children.

Main Methods:

• A multicenter retrospective cohort study involving 1,597 pediatric patients undergoing cholecystectomy.
• Utilized an Extra-Trees machine learning algorithm with k-fold cross-validation on clinical, laboratory, and ultrasound data.
• Model performance evaluated using the area under the receiver operating characteristic curve (AUC).

Main Results:

• Choledocholithiasis was confirmed in 18.8% of patients.
• Key predictors included common bile duct stones on ultrasound, increased common bile duct diameter, and elevated liver enzymes and bilirubin levels.
• A nine-feature model achieved a high predictive performance with an AUC of 0.935.

Conclusions:

• Machine learning effectively predicts pediatric choledocholithiasis using readily available clinical data.
• The developed model identifies children at high risk, aiding clinical decision-making.
• This approach enhances the diagnostic accuracy for common bile duct stones in pediatric populations.