Related Experiment Video
Updated: May 24, 2026
Reliability of Artificial Intelligence-Based Cone Beam Computed Tomography Integration with Digital Dental Images
Published on: February 23, 2024
Evaluation of imaging protocols for detecting tooth ankylosis using different CBCT devices: an in vitro study
Şelale Özel1, Taha Emre Köse2, Samed Şatır3
1Department of Oral and Maxillofacial Radiology, Faculty of Dentistry, Altınbaş University, Zuhuratbaba mah. İncirli cd. 11/A Bakırköy, İstanbul 34158, Türkiye.
Background:
To compare the diagnostic accuracy of CBCT images acquired with different imaging protocols for detecting tooth ankylosis.
Methods:
Thirty extracted permanent teeth (20 single-rooted, 10 three-rooted) were prepared; ankylosis was simulated in 20 by selectively removing Teflon from roots previously coated to mimic the periodontal ligament. Samples were embedded in stone plaster and imaged using 4 protocols on 2 CBCT units. CBCT images were primarily evaluated by a blinded radiologist (present/absent). Micro-CT was performed on 20% of specimens to verify the model. Agreement and accuracy metrics (Fleiss/Cohen κ, sensitivity, specificity, PPV, NPV) were calculated.
Results:
Protocol 2 (0.25 mm voxel) showed the most favorable performance (κ = 0.113; sensitivity 71.4%; PPV 52.6%; NPV 60%). Protocol 3 (0.35 mm voxel) performed worst (κ = -0.104; sensitivity 53.8%; PPV 43.8%; NPV 45.5%). Except for Protocol 3, sensitivity was high (≥71.4%) while specificity remained low (35.7%-42.9%).
Limitations:
The study has its own limits. Different imaging protocols using in clinical practice to detect ankylosis can be investigated as well to provide different approaches assessing ankylosis.
Conclusion:
CBCT demonstrated high sensitivity but low specificity for ankylosis detection, indicating a propensity for false positives. Smaller voxel sizes (0.25-0.30 mm) improved detection, whereas 0.35 mm was least reliable.
