BMC Oral Health. 2026 Jul 20. doi: 10.1186/s12903-026-09362-8. Online ahead of print.
ABSTRACT
BACKGROUND: Augmented reality (AR) visualization projects preoperative virtual three-dimensional (3D) models onto real surgical sites and can provide intuitive intraoperative guidance. This in vitro proof-of-concept study evaluated the feasibility of markerless AR-guided endodontic microsurgery by comparing objective outcomes between AR guidance (ARG) and traditional freehand (FH) procedures on customized 3D-printed mandibular models.
METHODS: Customized 3D dental models with artificial periapical lesions (APLs) were generated from cone beam computed tomography (CBCT) images obtained from one patient. A total of 20 dental models (80 mandibular molars and 160 APLs) were included, with 10 models assigned to the ARG group and 10 to the FH group. Surgical planning was performed on virtual models reconstructed from CBCT images. All procedures were performed by one endodontist with more than 10 years of clinical experience after standardized training in both workflows. Surgical time was recorded during each procedure, and postoperative CBCT images were acquired to evaluate bone window marking errors, osteotomy errors, and root resection angle error.
RESULTS: The ARG group required less time than the FH group across all four molar positions. Median (IQR) surgical times ranged from 3.72 (2.99-5.75) to 5.25 (4.58-6.56) min in the ARG group and from 7.02 (6.60-7.66) to 9.13 (8.28-9.75) min in the FH group (all P < 0.05). The ARG group also yielded lower median (IQR) values for bone window marking, osteotomy, and root-resection angle errors. Several tooth-specific comparisons reached statistical significance (P < 0.05).
CONCLUSIONS: Within the limitations of an in vitro proof-of-concept study, markerless AR guidance was feasible and was associated with improved accuracy and time efficiency during simulated mandibular molar endodontic microsurgery. Further clinical studies are required before routine clinical application.
PMID:42477686 | DOI:10.1186/s12903-026-09362-8
