Prediction of Palatoplasty Timing for Infants With Cleft Lip and Palate Using Machine Learning Algorithm

Area of Science:

• Pediatric Surgery
• Computational Biology
• Growth and Development

Background:

• Infants with cleft lip and palate (CLP) require precise surgical timing for optimal outcomes.
• Predicting the ideal timing for palatoplasty is complex, influenced by various patient factors.
• Machine learning (ML) offers a novel approach to analyze complex data for predictive modeling in CLP.

Purpose of the Study:

• To develop and validate a machine learning algorithm for predicting palatoplasty timing in infants with nonsyndromic CLP.
• To assess the impact of birth weight and cheiloplasty timing on infant growth trajectories.
• To evaluate the accuracy of ML models in forecasting personalized palatoplasty schedules.

Main Methods:

• Trained tree-based ML models using cleft type, age, height, and weight data from 111 infants with nonsyndromic CLP.
• Classified subgroups based on birth weight (low vs. normal) and cheiloplasty timing (early vs. late).
• Compared growth trajectories and analyzed prediction accuracy using metrics like root mean square error.

Main Results:

• Low birth weight infants demonstrated significant catch-up growth, normalizing by T2 compared to normal birth weight peers.
• Early cheiloplasty correlated with earlier palatoplasty, faster growth rates, and improved weight z-score increases post-surgery.
• The CatBoost ML algorithm accurately predicted palatoplasty timing with a mean error of 1.6 months.

Conclusions:

• ML-assisted prediction can accurately forecast palatoplasty timing for infants with CLP.
• Early surgical intervention and appropriate weight management positively influence growth outcomes in CLP patients.
• This predictive tool can assist clinicians in establishing personalized palatoplasty schedules, potentially improving long-term results.