Related Experiment Video
Updated: Jan 16, 2026
Guided Endodontics: Three-Dimensional Planning and Template-Aided Preparation of Endodontic Access Cavities
Published on: May 24, 2022
Comparative Analysis of AI-Generated and Manually Designed Approaches in Accuracy and Design Time for Surgical Path
Chen Chen1,2, Xining Zhang1,2, Li Qin1,2
1State Key Laboratory of Oral & Maxillofacial Reconstruction and Regeneration, Key Laboratory of Oral Biomedicine, Ministry of Education, Hubei Key Laboratory of Stomatology, School & Hospital of Stomatology, Wuhan University, Wuhan, China.
Aim:
To compare the accuracy and design time of artificial intelligence (AI)-generated and manually designed (MD) surgical pathways for osteotomies and root-end resections in dynamic navigation (DN)-aided endodontic microsurgery (EMS).
Methodology:
Fifty-one surgical pathways were analysed, each planned using both AI and MD methodologies. Accuracy was assessed using the DCARE Navigation System (v3.2, MedNav Ltd) and AutoCAD (2023, Autodesk Inc.), evaluating five parameters: start deviation, end deviation, angular deviation, root-end resection length deviation, and root-end resection angulation deviation. Design time was measured from the point of CBCT dataset import to the finalisation of the surgical pathway design. Mann-Whitney U test was used to compare the accuracy and design time of the AI and MD groups, whereas the rank-based ANCOVA was used to assess deviations according to tooth type, jaw type, and root number. Statistical significance was set at p < 0.05.
Results:
Compared with the MD group, the AI group exhibited significantly smaller root-end resection length deviations (AI: 0.01 [0.01, 0.02] mm; MD: 0.02 [0.01, 0.03] mm; p = 0.029) but significantly larger root-end resection angulation deviations (AI: 3.48 [1.01, 7.48]; MD: 0.35 [0.16, 0.73]; p < 0.001). There were no significant differences in the start deviation, end deviation, angular deviation, root-end resection length deviation, or root-end resection angulation deviation across tooth type, jaw type, or root number. The design time was significantly shorter in the AI group than in the MD group (55 [21, 74] s vs. 379 [215, 553] s; p < 0.001).
Conclusions:
A clinically operational AI-based surgical path design approach is capable of minimising manual interventions and delivering time-efficient, accurate results for clinical use. The integration of AI with DN-aided EMS may contribute to the development of increasingly autonomous surgical procedures.
