Related Experiment Video
Updated: Jun 8, 2026
Reliability of Artificial Intelligence-Based Cone Beam Computed Tomography Integration with Digital Dental Images
Published on: February 23, 2024
Diagnostic ability of a cone-beam computed tomography scan to assess longitudinal root fractures in prosthetically
Saulo Leonardo Sousa Melo1, Eduardo Antunes Bortoluzzi, Murillo Abreu
1Department of Dentomaxillofacial Radiology, School of Dentistry, University of Campinas, São Paulo, SP, Brazil. slsmelo@uol.com.br
Introduction:
This study evaluated the influence of cast-gold posts on the diagnostic ability of a cone beam computed tomography (CBCT) system in assessing longitudinal root fractures. In addition, the influence of gutta-percha and variations in voxel resolution were assessed.
Methods:
One hundred eighty endodontically prepared teeth were divided into 3 experimental and 3 control groups and placed in a dry human skull. The teeth in the experimental groups were artificially fractured. Certain experimental and control groups were filled with gutta-percha cones. Other experimental and control groups were filled with cast-gold posts. All the teeth were viewed by using a tomography scan with 2 voxel resolution protocols (0.3-mm and 0.2-mm). A calibrated examiner, blinded to the protocol, assessed the images by using the nominated scan software.
Results:
The kappa values obtained for intraobserver reproducibility were 0.84 and 0.93 for 0.3-mm and 0.2-mm voxel resolution, respectively. The presence of gutta-percha or posts reduced the overall sensitivity and specificity in both voxel resolutions, but with no significant association. The CBCT specificity values were similar and did not depend on the voxel resolution adopted. In contrast, the sensitivity values were significantly higher for 0.2-mm voxel resolution in nonfilled (P < .05), gutta-percha (P < .001), and overall (P < .001) groups.
Conclusions:
The CBCT diagnostic ability was not influenced by the presence of posts or gutta-percha, and the 0.3-mm voxel resolution images were demonstrated not to be a reliable protocol for the investigation of longitudinal root fractures.
