Machine learning to predict pediatric choledocholithiasis: A Western Pediatric Surgery Research Consortium retrospective study
Timestops (8)
Tools Used
Topic Overview
Gretchen Floan Sachs MD, Shadassa Ourshalimian MPH, Aaron R. Jensen MD, MEd, MS, FACS, Lorraine I. Kelley-Quon MD, MSHS, FACS, Benjamin E. Padilla MD, FACS, Stephen B. Shew MD, FACS, Katrine M. Lofberg MD, FACS, Caitlin A. Smith MD, FACS, Jonathan P. Roach MD, FACS, Samir R. Pandya MD, FACS, Katie W. Russell MD, FACS, Romeo C. Ignacio Jr. MD, MS, MPath, FACS, Western Pediatric Surgery Research Consortium Choledocholithiasis Investigative Group
Background: The purpose of this study was to accurately predict pediatric choledocholithiasis with clinical data using a computational machine learning algorithm.
Methods: A multicenter retrospective cohort study was performed on children <18 years of age who underwent cholecystectomy between 2016 to 2019 at 10 pediatric institutions. Demographic data, clinical findings, laboratory, and ultrasound results were evaluated by bivariate analyses. An Extra-Trees machine learning algorithm using k-fold cross-validation was used to determine predictive factors for choledocholithiasis. Model performance was assessed using the area under the receiver operating characteristic curve on a validation dataset.
Results: A cohort of 1,597 patients was included, with an average age of 13.9 ± 3.2 years. Choledocholithiasis was confirmed in 301 patients (18.8%). Obesity was the most common comorbidity in all patients. Choledocholithiasis was associated with the finding of a common bile duct stone on ultrasound, increased common bile duct diameter, and higher serum concentrations of aspartate aminotransferase, alanine transaminase, lipase, and direct and peak total bilirubin. Nine features (age, body mass index, common bile duct stone on ultrasound, common bile duct diameter, aspartate aminotransferase, alanine transaminase, lipase, direct bilirubin, and peak total bilirubin) were clinically important and included in the machine learning algorithm. Our 9-feature model deployed on new patients was found to be highly predictive for choledocholithiasis, with an area under the receiver operating characteristic score of 0.935.
Conclusion: This multicenter study uses machine learning for pediatric choledocholithiasis. Nine clinical factors were highly predictive of choledocholithiasis, and a machine learning model trained using medical and laboratory data was able to identify children at the highest risk for choledocholithiasis.
Keywords
Hashtags
Transcript
Pediatric cholelithiasis is becoming increasingly common due to the rising obesity rates among children and adolescents. Unobstructed common bile duct stone can be found in up to 30% of pediatric patients with cholelithiasis during surgery. Making a timely diagnosis of common bile duct stone can be critical for improving patient outcome. The article titled Machine Learning to Predict Pediatric Choleocholithiasis, a Western Pediatric Surgery Research Consortium. retrospective study was published in surgery in 2023. The authors evaluated if a machine learning model could accurately predict pediatric cleoliassis using clinical and laboratory data. This retrospective cohort study included 1,597 pediatric patients who underwent cholecystectomy across 10 institutions between 2016 and 2019. An extra 3's machine learning algorithm was used to evaluate nine clinical features including age, BMI, and specific lab values to predict the presence of cholelithiasis. The machine learning model demonstrated high accuracy with an area under the receiver operating characteristic curve out of 0.95 and a negative protective value of 98%, significantly outperforming previous prediction models. A limitation of the study is the reliance on the retrospective data, which could introduce bias and affect generalizability of the model. Also, machine learning is a novel technology for the medical field. Implementing this machine learning model in clinical practice could improve the accuracy of diagnosing cholelithiasis preoperatively in children and improving patient selection and resource organization and use. It could also reduce the need for invasive procedures and unnecessary imaging.