50948255-1646 - AI-Empowered Tooth Segmentation from CBCT for the Fabrication of Dental Splints

Abstract

<p><strong>Background</strong>: AI-empowered teeth segmentation from CBCT is gaining attention, yet its clinical utility requires further investigation. This study investigates the clinical feasibility of using AI-empowered teeth segmentation from CBCT in designing dental splints.</p><p><strong>Methods:</strong> CBCT and maxillary plaster models of 15 patients were utilized. AI-empowered segmentation was performed using BlueSkyPlan, and plaster models were scanned with the intraoral scanner. Splints were designed to cover the crown third of the maxillary teeth, excluding the second and third molars. Different offset values were assigned (0.2 mm for plaster model scans; 0.25 mm, 0.5 mm, 0.75 mm for CBCT segmentation). Digital splints were printed using a Sonic Mighty Revo 3D printer. Clinical assessments were based on a scaled questionnaire for obstruction and stability, with digitization and quantitative space analysis performed using spiral CT scans. Data were analyzed using SPSS and GraphPad Prism.</p><p><strong>Results:</strong> In the obstruction test, the scoring for splints from 3D scan-0.2 mm (median: 3; interquartile range (IQR): 2-3) was higher compared to CBCT-0.25 mm (median: 2; IQR: 1-3; p=0.048), but comparable to CBCT-0.5 mm (median: 3; IQR: 3-3; p=0.20) and CBCT-0.75 mm (median: 3; IQR: 3-3; p=0.32). In the stability test, there was no significant difference found. Space analysis revealed that splints designed from 3D-scanned models had smaller spaces in both 2D (p<0.001) and 3D (p<0.001) compared to those from AI-powered tooth segmentation.</p><p><strong>Conclusion:</strong> It is feasible to fabricate dental splints using AI-powered tooth segmentation with offset values of 0.5 mm and 0.75 mm with acceptable performance. However, the clinical impact of these findings warrants further investigation.</p>