ARTIFICIAL INTELLIGENCE IN ORTHODONTIC ASSESSMENT: ACCURACY OF SEGMENTATION AND LANDMARK LOCALIZATION: A SYSTEMATIC REVIEW

Abstract

Introduction Artificial intelligence is a fast-developing technical field in which machines simulate the human intellect. it works on the principle of recognising the pattern of the data and then making the decision with small human assistance.  The advanced techniques of AI, which are being used in orthodontics more often. Due to the gap in the literature, the generalisability of the findings is limited. Thus, these factors emphasise the need for a comprehensive review of the use of AI in orthodontic assessment. AIM: The aim of the current review is to evaluate the analytical performance and reliability of AI-based tools in orthodontic patients compared with traditional diagnostic techniques. Material and methods: In the current review, a thorough search was conducted in the following databases:  PubMed (MEDLINE), Cochrane CENTRAL, Embase, ProQuest, and OVID from inception till 30th August 2024. The MeSH keywords such as AI, machine learning, deep learning, neural networks, convolutional neural networks, automated diagnosis, computer-aided diagnosis, orthodontics, cephalometric analysis, malocclusion, automated landmark detection, dental imaging, diagnostic imaging, diagnosis, analysis, CBCT, and radiographic analysis were used in combination. QUADAS II tool was used to assess the accuracy of the methodology of the included studies. Results: The records identified were 91. Only 7 studies were included after thorough screening and removal of duplicates. Conclusion: AI-based tools have been proven to be promising adjuncts for improving efficiency and consistency, as these tools have high accuracy in orthodontic diagnostic tasks such as landmark detection and CBCT analysis.