The influence of tracker position and anatomical factors on the accuracy of dynamic navigation for periodontal procedures in severe bone loss models

This study evaluates the registration accuracy of dynamic navigation systems in non-surgical periodontal treatment using an in vitro model, focusing on the impact of tracker guide positions, providing a basis for clinical application. An in vitro experiment was conducted using a model with severe periodontal bone defects. Point cloud registration was performed, collecting three-dimensional error data (total error D and directional errors Dx, Dy, Dz) from 1,200 key points. Multiple linear regression analysis assessed the impact of tracker guide positions (anterior, left posterior, right posterior), probing depth, jaw type, tooth position, and tooth surface on registration accuracy. The mean three-dimensional registration error was 0.541 ± 0.176 mm. No significant differences were observed in D across the three tracker guide positions ( p > 0.05). A slight increase in D was observed with increasing probing depth (0.008 mm, p = 0.009), with no clinical significance. Errors were lower in posterior regions compared to anterior (D: 0.046 mm, p = 0.000), and higher in the mandible versus the maxilla (D: 0.050 mm, p = 0.000). Mesial surfaces showed higher errors ( p = 0.010), with directional errors indicating systematic shifts. In vitro, tracker placement did not significantly affect registration accuracy. Dynamic navigation demonstrates the potential to address certain limitations of traditional procedures under in vitro conditions, which warrants further clinical validation.