Objectives
This study aimed to assess the accuracy of different fields of view (FOVs) and voxel sizes of cone-beam computed tomography (CBCT) for detection of simulated periodontal defects in comparison with periapical (PA) radiography.
Materials and Methods
This in vitro study was conducted on 60 periodontal defects including furcation and wall defects, dehiscence, and fenestration that were artificially created in 5 mandibles of freshly slaughtered sheep. After recording the defect sites and types as the gold-standard, they underwent PA radiography. Thirty CBCT images were also obtained in three FOVs and voxel sizes of 13 x 17 cm2/ 300 µm, 10 x 10 cm2/ 180 µm, and 5 x 5 cm2/ 90 µm, and evaluated by a radiologist and a periodontist. The inter-observer agreement was assessed by the kappa coefficient, and the parameters were calculated and compared using the Chi-square test with STATA 14.
Results
The inter-observer agreement was generally 85.4% for all periodontal defects; this value was the highest (88.1%) with the smallest FOV and voxel size (5 x 5 cm2 and 90 µm). By a reduction in FOV and voxel size, the inter-observer agreement significantly increased (P < 0.001). The mean sensitivity, specificity, PPV, and NPV were 87.8%, 94.9%, 94.6%, and 88.4%, respectively, for CBCT. These values for PA radiography were lower than the corresponding values for CBCT.
Conclusion
Sensitivity, specificity, PPV, NPV, and inter-observer agreement increased in use of smaller FOV for detection of simulated periodontal defects, indicating an increase in diagnostic accuracy by a reduction in FOV and voxel size.