Artificial Intelligence and Digital Technologies in Orthognathic and Reconstructive Maxillofacial Surgery: Data...

Orthognathic and reconstructive maxillofacial surgery addresses severe dentofacial deformities, post-traumatic defects, and defects following oncologic resection. Across the pathway from preoperative imaging and virtual planning to soft-tissue prediction and intraoperative guidance, artificial intelligence and related digital technologies are increasingly being investigated to support decision-making, simulation, and plan transfer. This structured narrative review synthesizes evidence across that surgical pipeline and uses data availability as an organizing framework for interpreting evidence maturity. Predefined searches of PubMed/MEDLINE, Embase, and Scopus identified English-language peer-reviewed articles published from January 2018 to April 2026, supplemented by selected foundational studies. Within the literature reviewed, automated cephalometric and three-dimensional landmark detection has undergone the most extensive quantitative evaluation, supported by several systematic reviews and meta-analyses and by multicenter retrospective evaluations, including one study with independent external test sets. Artificial intelligence-assisted diagnosis, osteotomy planning, and virtual surgical planning show increasing technical capability, including low-millimeter reposition-vector prediction, although external validation and patient-centered outcomes remain limited. Soft-tissue prediction has advanced through deep learning and finite-element modeling, with selected studies reporting comparable geometric accuracy and substantially faster computation. In contrast, intraoperative augmented reality, navigation, and robot-assisted craniomaxillofacial surgery remain supported mainly by small clinical series, cadaveric studies, and preclinical validation. These technologies demonstrate plan-transfer feasibility in selected settings but do not yet establish broad clinical effectiveness or routine superiority over established workflows. Overall, evidence is more mature when datasets are structured, accessible, and standardized, and less mature when paired longitudinal imaging or intraoperative tracking data are sparse. This association should be interpreted as an organizing hypothesis rather than proof of causality.