Applications of artificial intelligence in diagnosis and treatment planning of orthodontics: a narrative review

Abstract

Artificial intelligence (AI) has recently gained significant interest in orthodontics due to its ability to enhance diagnostic accuracy, guide treatment planning, and improve therapeutic outcomes. This review aimed to explore the relevance and applications of AI across various aspects of orthodontics. A comprehensive literature search was performed from January 2010 to 1 March 2025 in databases including PubMed, EMBASE, Web of Science, Scopus, and Cochrane. Letters to the editor, case reports, systematic reviews, and animal studies were excluded. Artificial intelligence models, especially those using deep learning, have been integrated into multiple orthodontic fields, including landmark identification, malocclusion classification, treatment planning, growth prediction, and risk assessment. They have also achieved notable success in segmenting two-dimensional and three-dimensional anatomical structures, aligner therapy, evaluating facial asymmetry, localizing impacted canines, and identifying clefts. While several investigations highlight the high accuracy of AI models, others emphasize the need for clinician oversight, recommending that these tools serve as a supportive tool rather than a replacement for clinical judgment. AI-based algorithms may enhance treatment quality, decrease procedural time and operator variability, and reduce human error. However, further clinical trials are needed to validate and optimize the accuracy and reliability of these models in orthodontics.