The use of artificial intelligence (AI) and machine learning (ML) is transforming the prediction of mandibular bone growth

Abstract

Introduction: Accurately predicting mandibular bone growth is essential for orthodontics and maxillofacial surgery, as it influences both treatment planning and patient outcomes. Traditional approaches, which depend on linear models and clinician expertise, often lead to errors and inconsistencies. AI and machine learning (ML) present more advanced solutions by analyzing complex datasets, offering improved predictive accuracy. This systematic review evaluates the performance of AI and ML models in predicting mandibular growth, comparing them to conventional methods.   Methodology: The review followed PRISMA guidelines, considering studies up to July 2024, from databases like PubMed, Embase, Scopus, and Web of Science. Out of 31 studies, 6 met the inclusion criteria, focusing on AI algorithms and prediction accuracy. The risk of bias was assessed using the QUADAS-2 tool.   Results: AI and ML models, such as the LASSO model, demonstrated high accuracy, with an average error of 1.41 mm for skeletal landmarks. However, some AI models, particularly deep learning models, performed worse than traditional methods.   Discussion: Challenges arose due to dataset variability and the complexity of AI models, which could hinder clinical application. Nonetheless, AI and ML show considerable potential for improving predictive accuracy in mandibular growth, though further research is needed to refine methods and enhance clinical usability.