What is the role of AI-driven automation in static surgical guide design? A scoping review
Maria Fernanda Silva Andrade‐Bortoletto, Xijin Du, Eslam Abdelwahab Dawood, Oana Elena Burlacu Vatamanu, Mihai Tarce, Rocharles Cavalcante Fontenele, Deborah Queiroz Freitas, Reinhilde Jacobs
Abstract
OBJECTIVE: This scoping review aims to evaluate the extent of artificial intelligence (AI)- driven automation currently available for designing static surgical guides (SG) for implant placement and its correlation with implant placement accuracy. METHODS: A comprehensive literature search was conducted across PubMed, Scopus, Web of Science, Embase, Cochrane Library, and grey literature up to February 2025. Two reviewers independently screened 518 English-language articles, selecting 140 for eligibility evaluation. After applying exclusion criteria, seven studies were included for full-text review. The SG planning process was classified as manual, semi-automated, or fully automated driven by AI. Implant placement accuracy was assessed based on linear (cervical and apical) and angular deviations, with a detailed review of measurement methods used. Additionally, a market survey was conducted to identify available SG design software, its key features, and the level of automation implemented for each design step. RESULTS: All included studies (n=7) employed a semi-automated software for SG design. The mean deviations in implant placement using SGs were 0.65 mm (0.22 to 1.19 mm) (linear-cervical), 0.95 mm (0.18 to 2.11 mm) (linear-apical), and 2.92° (0.77 to 6.35°) (angular). The software programs used were: coDiagnostiX™ software (Version 9.0, Dental Wings GmbH, Germany), Smop-software (version 2.7.0, Swissmeda AG, Switzerland), 3Shape Implant studio (Version 2021.1.2, 3Shape, Denmark), R2WARE™ (MegaGen implant, Korea), 3-Matic modelling software (Materialise, Belgium) and Blue Sky Plan 4.8 (Blue Sky Bio, USA). CONCLUSIONS: This scoping review found that most surgical guide planning software employs a semi-automated approach requiring human intervention, which has shown clinically acceptable implant placement accuracy. Fully automated (AI-based) designs were not yet validated scientifically.