Predicting pediatric cardiac surgery-associated acute kidney injury using machine learning

Area of Science:

• Pediatric Nephrology
• Cardiovascular Surgery
• Artificial Intelligence in Medicine

Background:

• Cardiac surgery-associated acute kidney injury (CS-AKI) poses significant risks to pediatric patients.
• Early prediction of CS-AKI is vital for timely intervention and improved patient outcomes.
• Developing accurate predictive models is a key goal in pediatric cardiac care.

Purpose of the Study:

• To develop and validate a supervised machine learning (ML) model for predicting moderate to severe CS-AKI.
• To identify CS-AKI risk at postoperative day 2 (POD2) in pediatric patients undergoing cardiac surgery.
• To assess the performance of the ML model using various statistical metrics.

Main Methods:

• A retrospective cohort of 402 pediatric patients undergoing cardiac surgery was analyzed.
• An 80%-20% train-test split was used to develop and validate the ML model.
• The model incorporated demographic, preoperative, intraoperative, and POD0 clinical and laboratory data.

Main Results:

• The ML model demonstrated strong predictive performance with an accuracy of 0.91 and AUROC of 0.88.
• Key predictors included preoperative serum creatinine, surgery duration, and POD0 lactate levels.
• The model achieved a precision of 0.92 and recall of 0.63 for CS-AKI prediction.

Conclusions:

• A supervised ML model effectively predicts moderate to severe CS-AKI in pediatric cardiac surgery patients.
• The model leverages readily available clinical and laboratory data for early risk identification.
• This ML approach shows promise for improving clinical decision-making and patient management.