The 15th Asian Congress of Oral and Maxillo-Facial Radiology (ACOMFR 2026) , Seoul, Güney Kore, 30 Nisan - 02 Mayıs 2026, ss.38, (Özet Bildiri)
Introduction: Tooth ankylosis is the fusion of the ce mentum or dentin of the tooth root to the surrounding alveolar bone and partial or complete obliteration of the periodontal ligament (PDL) space. In the present study, aim is to compare the diagnostic accuracy of CBCT images acquired with different voxel sizes and imaging protocols, providing further insight into opti mal imaging protocols for ankylosis detection. Materials and Methods: 30 extracted teeth includ ed for tooth ankylosis simulation. To simulate the PDL space, the roots were coated with Teflon tape. In the simulated ankylosis group, the Teflon tape was carefully removed in assigned locations. The teeth were then embedded in plaster blocks. The samples imaging were performed in four different imaging pro tocols using two CBCT devices (NewTom VGi Evo and KaVo OP 3D Pro). The findings were recorded as “present” or “absent”. Micro-CT imaging (SkyScan 1173) was performed on 20% of the specimens to ver ify the accuracy of the simulation. Fleiss’ Kappa anal ysis was used for the overall agreement. Sensitivity, specificity, positive predictive value (PPV), and nega tive predictive value (NPV) were calculated. Results: The smallest voxel size (0.25 mm) demon strated the most favorable overall results. Although its Cohen’s Kappa coefficient indicated only slight agree ment (0.113), the protocol achieved relatively high sensitivity (71.4%) and acceptable predictive values (PPV: 52.6%, NPV: 60%). The largest voxel size (0.35 mm) showed the weakest performance, with a negative Kappa value (–0.104), the lowest sensitivity (53.8%), and reduced predictive values (PPV: 43.8%, NPV: 45.5%). Conclusions: CBCT imaging demonstrated high sen sitivity in detecting ankylosis, but specificity was con siderably low. Although a moderate level of agreement was observed among imaging protocols, the diagnostic capacity of 0.30 mm voxel size was markedly more limited compared with smaller voxel sizes.