Deep learning in dental implantology: A systematic review

Abstract: Artificial intelligence (AI) has the potential to revolutionize dental implantology by improving the accuracy, efficiency, and safety of the dental implant placement process. The purpose of this systematic review is to evaluate AI applications and their performance in this field. Totally, six electronic databases, including Medline (via PubMed), Embase (via Ovid), Scopus, Google Scholar, Web of Science, and arXiv, were searched for relevant studies without any limitations in language and publication date. Our PICO question was, “What are the applications of AI approaches and their performances in the field of dental implantology?” The quality of the studies that met the eligibility criteria was evaluated using the QUADAS-2. Meta-analyses were not conducted due to heterogeneity in reporting. Of the 33 included studies, 15 focused on detecting and classifying implant systems, while ten studies focused on predicting implant prognosis. Four studies focused on treatment planning, and a further three studies focused on the classification and identification of peri-implantitis. Moreover, one study both classifies implant type and detects peri-implantitis. The artificial neural network is the most commonly used AI model. The studies utilized five data modalities, including panoramic radiographs, periapical radiographs, cone beam computed tomography (CBCT), simulated X-ray images from 3D models, and clinical data. Only one-third of the included studies showed a low risk of bias in all the domains. Overall, this review provides evidence of the potential for AI to improve dental implantology and highlights the need for continued research and development in this area. Future studies should focus on enhancing the comparability of results across studies and developing standardized reporting and evaluation methods to increase the robustness of the evidence base.