Robot-assisted augmented reality navigation for osteotomy and personalized guide-plate in mandibular reconstruction: a preclinical study

Background: Oral and maxillofacial tumors, particularly those requiring mandibular reconstruction, present significant clinical challenges due to the complexities involved in achieving precise surgical outcomes and ensuring post-operative stability. Traditional methods for mandibular reconstruction, such as freehand bending of titanium plates, often result in errors in osteotomy and reconstruction, compromising both the precision and stability of the procedure. Methods: This study initially developed the Robot-assisted Augmented Reality Osteotomy Navigation System (RARONS) and constructed a virtual surgical plan using imaging data. In this robot-assisted cadaveric osteotomy study, 20 Non-Flap Fibulas (NFFs) and 9 Free Flap Fibulas (FFFs) were included to evaluate the impact of pedicle factors on osteotomy precision. Augmented Reality technology enabled surgeons to intuitively perform procedures according to the virtual plan during the robot-assisted osteotomy process. Additionally, in the mandibular reconstruction experiments using cadavers, 11 NFFs and 9 FFFs were included to assess the influence of pedicle factors on reconstruction accuracy. 11 fibulas were reconstructed using surgical guide-plate methods, while 9 fibulas were reconstructed using freehand methods to evaluate the impact of surgical guide plates on reconstruction precision. Finally, an evaluation framework was established to quantify the precision of osteotomy and reconstruction, with osteotomy errors measured in terms of length, angle, and volume, and mandibular reconstruction errors assessed based on width, height, anteroposterior diameter, and symmetry. Results: In the osteotomy experiments, the Non-Flap Fibula group demonstrated higher accuracy compared to the Free Flap Fibula group, with mean errors in osteotomy length, angle, and volume recorded as 1.08 ± 1.57 mm, 11.78° ± 5.72°, and 10.76 ± 5.94%, respectively. Similarly, in the mandibular osteotomy experiment, the errors in osteotomy length were within clinically acceptable ranges, with mean errors in length, angle, and volume reported as 1.75 ± 1.32 mm, 6.32° ± 2.71°, and 10.61 ± 5.29%, respectively. In the reconstruction experiments, the presence of a pedicle fibula did not significantly affect the error rate in mandibular reconstruction. Additionally, the personalized surgical guide-plate reconstruction group showed superior accuracy in width, anteroposterior diameter, and symmetry compared to the traditional freehand reconstruction group. The posterior mandibular angle and anterior mandibular angle for the personalized group were 1.57° ± 1.07° and 1.16° ± 1.31°, respectively. Conclusions: This study demonstrated that RARONS significantly improved osteotomy accuracy and enhanced mandibular reconstruction precision using personalized pseudo-titanium plates and guide-plates. These advancements lay the foundation for integrating digital osteotomy navigation and personalized surgical guide-plate technology into clinical maxillofacial surgery.