An autonomous robotic osteotomy system may more accurately reproduce planned recipient socket geometry for tooth autotransplantation than static guide-assisted preparation, with the greatest gains observed in double-rooted models.
Researchers compared robot-assisted and static guide-assisted osteotomy in 40 3D-printed mandibular models representing single-rooted and double-rooted donor tooth anatomies. Models were randomly allocated to the 2 techniques, with 20 specimens per group and 10 of each root morphology in each group. Recipient sockets were planned by offsetting the donor root surface outward by 0.5 mm and eliminating undercuts along the planned insertion axis. The robotic system autonomously executed preplanned nonlinear, surface-conforming milling trajectories, whereas the static guide-assisted approach used guided pilot drilling followed by freehand connection, enlargement, and refinement.
Apex deviation was the prespecified primary outcome. Other assessments included platform and angular deviation, root-mean-square surface deviation, volumetric overlap measured with the Dice similarity coefficient, over-removal rate, and preparation time. The researchers also evaluated whether the relative performance of the techniques differed between single-rooted and double-rooted anatomies.
Robot-assisted preparation reduced mean apex deviation to 0.94 mm vs 1.96 mm with static guide-assisted preparation. The difference was most pronounced in double-rooted models, where apex deviation was 0.85 mm vs 2.38 mm, respectively. Among single-rooted models, apex deviation was 1.03 mm vs 1.53 mm, respectively, but the between-group difference was not statistically significant.
Angular deviation also favored robotic preparation at 1.86° vs 7.32° overall. The difference was observed in both root-morphology subgroups, with angular deviations of 2.29° vs 5.99° in single-rooted models and 1.43° vs 8.64° in double-rooted models. In contrast, platform deviation was comparable between approaches, at 0.79 mm with robotic preparation and 0.78 mm with static guidance. The researchers suggested that the robotic advantage primarily involved maintaining the planned trajectory throughout the preparation depth rather than improving initial targeting.
Measures of agreement between the prepared and planned socket geometry further favored the robotic approach. Root-mean-square surface deviation was 0.20 mm vs 0.33 mm, and Dice similarity was 88% vs 63%. The over-removal rate was 8% vs 96%, with differences in these measures observed across both root-morphology subgroups.
Overall preparation time was similar between techniques, at approximately 759 seconds with robotic preparation vs 833 seconds with static guidance. However, the pattern differed according to anatomy. Robotic preparation took longer in single-rooted models, at approximately 591 vs 461 seconds, but was shorter in double-rooted models, at approximately 928 vs 1,205 seconds. The researchers noted that preparation time covered only the active preparation phase and excluded registration, calibration, and seating verification, meaning the findings should not be interpreted as showing equivalent total chairside time.
The findings were limited by the standardized in vitro design. The resin models did not reproduce soft tissues, bleeding, limited mouth opening, restricted visibility, or other factors that could affect surgical accessibility and workflow performance during intraoral surgery. The digitization process could also introduce distortion or reconstruction artifacts. The static guide-assisted workflow also included operator-dependent freehand refinement, whereas robotic accuracy depended on registration and calibration. Additionally, all osteotomies were performed by 1 operator, preventing assessment of interoperator variability or whether robotic assistance reduces dependence on surgical experience.
The results suggest that autonomous, surface-conforming robotic osteotomy may improve full-depth geometric fidelity and limit unnecessary material removal during recipient socket preparation, particularly when root anatomy is more complex. However, clinical studies are needed to determine whether these geometric improvements translate into fewer trial insertions, shorter extra-alveolar time, improved primary stability, or favorable periodontal healing.
“In this standardized in vitro model, an autonomous, multi-axis robotic osteotomy executing nonlinear, surface-conforming toolpaths achieved substantially better apical and angular accuracy and more faithfully reproduced the planned socket morphology than a static guide-assisted workflow, while maintaining a similar overall preparation time,” wrote co-first author Chen Liu, of the Digital Dentistry Center, School of Stomatology at The Fourth Military Medical University in Xi’an, China, and colleagues.
The study was supported by the National Natural Science Foundation of China (grant 82501241). The authors declared no competing interests.
