AI for dental imaging: impressive in vitro, but what about in practice?
Authors
Affiliations (4)
Affiliations (4)
- AI and Digital Health Team, University of Birmingham, Birmingham, United Kingdom. [email protected].
- Special Care Dental Services, Birmingham Community NHS Trust, Birmingham, United Kingdom. [email protected].
- University of Birmingham, Birmingham, United Kingdom.
- University Hospitals Birmingham NHSFT, Birmingham, United Kingdom.
Abstract
Alam M K, Alftaikhah S A A, Issrani R et al. Applications of artificial intelligence in the utilisation of imaging modalities in dentistry: a systematic review and meta-analysis of in-vitro studies. Heliyon 2024; https://doi.org/10.1016/j.heliyon.2024.e24221 . A systematic review of in vitro studies utilising artificial intelligence (AI) in dental imaging. Searches were carried out across multiple databases: CINAHL, Cochrane Library, Embase, Google Scholar, IEEE Xplore, PubMed/MEDLINE, Scopus, and Web of Science as well as hand-searching references from eligible articles. Studies were eligible if i) classed as in vitro, defined as simulations or laboratory tests outside a living organism, ii) studied the performance of AI techniques and iii) involved analysis of dental imaging. Studies not in English or with insufficient data were excluded. An adapted version of the CONSORT bias tool was used for the assessment of studies. Outcome measures were extracted including: odds ratios, true positive rate, true negative rate, positive predictive value, and negative predictive value. A meta-analysis using a fixed-effects model assessed accuracy with a 95% confidence interval. Heterogeneity and overall effect tests were applied to evaluate the reliability of the meta-analysis. After screening, nine studies were identified, eight of which focused on Cone Beam Computed Tomography (CBCT) imaging. Endpoints included caries detection, segmentation tasks and virtual 3D model creation. Across the nine studies, and when pooled in meta-analysis, AI performance was shown to be superior to reference standards. This systematic review of in-vitro studies highlights the potential of AI to improve the speed or quality of dental imaging tasks. However clinical studies are required to ensure evidence from laboratory studies can translate into clinical practice.