Registration accuracy, time, and osteotomy outcomes across four strategies for task-autonomous robotic maxillofacial osteotomy
Related Experiment Videos
Shiwei Song1, Heng Liu2, Chen Liu1
1State Key Laboratory of Oral & Maxillofacial Reconstruction and Regeneration, National Clinical Research Center for Oral Diseases, Shaanxi Key Laboratory of Stomatology, Digital Dentistry Center, School of Stomatology, The Fourth Military Medical University, No.145 Changle West Road, Xincheng District, Xi'an, Shaanxi 710032, China.
Objectives:
To compare four registration strategies for task-autonomous robotic maxillofacial osteotomy.
Methods:
A maxillofacial model was constructed from imaging data. The maxilla and mandible were registered separately using splint-based landmark registration (SL), dental-landmark registration (DL), dental-surface registration (DS), and bone-surface registration (BS). The target-registration error (TRE) was measured in three osteotomy regions, and registration time was recorded. Osteotomies were then performed using a task-autonomous robotic system. The actual and planned osteotomy surfaces were compared using the root mean square (RMS) deviation, mean signed deviation, and angular deviation.
Results:
All four strategies achieved mean TREs of less than 0.8 mm. In the Le Fort I region, TRE was lower for BS than for SL (0.398 ± 0.126 vs. 0.580 ± 0.098 mm; P = 0.0045). Mandibular TREs showed a significant strategy-by-region interaction (P = 0.0031); in the genioplasty region, DS and BS yielded lower TREs than SL and DL. In the mandibular angle region, DS yielded lower TREs than SL and DL, and BS yielded lower TREs than DL. SL had the shortest registration time for both jaws. In the genioplasty region, DL resulted in greater RMS and angular deviations than the other strategies, whereas in the mandibular angle region, DS resulted in lower angular deviation than SL and DL.
Conclusions:
Registration strategy significantly influenced TRE, registration time, and mandibular osteotomy accuracy outcomes. All four strategies achieved submillimetric TREs. Overall, surface-based registration (BS and DS) yielded TREs that were lower than or comparable to those of landmark-based registration (SL and DL) and showed favorable performance for selected mandibular osteotomy accuracy outcomes. However, no consistent difference in accuracy was identified between DS and BS.
Clinical Significance:
Surface-based registration may support accurate robot-assisted osteotomy when adequate dentition or bony surfaces are accessible.