EFFECTIVENESS OF ARTIFICIAL INTELLIGENCE IN PREDICTING SOFT TISSUE OUTCOMES IN PATIENTS UNDERGOING ORTHOGNATHIC SURGERY: A SYSTEMATIC REVIEW

Abstract

Predicting soft tissue changes after orthognathic surgery is essential but challenging, as traditional methods often lack accuracy in capturing individual patient outcomes. Recent advancements in artificial intelligence (AI) offer promising solutions for achieving more precise predictions. This systematic review examines the effectiveness of AI models in forecasting postoperative soft tissue changes in patients undergoing orthognathic surgery. A comprehensive literature search was performed in PubMed, Scopus, and ScienceDirect, identifying 22 relevant studies published from 2014 to April 2025 that compared AI predictions with actual postoperative results. These studies primarily utilized AI methods such as deep learning, finite element modeling (FEM), and landmark-based techniques, achieving prediction accuracy within errors ranging from 0.55 mm to 2.91 mm. Higher accuracy was generally found in stable midface regions compared to more mobile or lateral facial areas. Deep learning techniques consistently showed superior performance compared to traditional and biomechanical methods. Cone-beam computed tomography (CBCT) was the most frequently used imaging technique, and voxel-based image registration provided the highest accuracy in aligning images. AI models have shown strong potential to significantly improve individualized predictions of soft tissue outcomes in orthognathic surgery. However, due to substantial variability in study methodologies, there is an urgent need for standardized protocols and further multicenter studies to ensure reliability and widespread clinical applicability.