Effect of voxel size on accuracy of cone beam computed tomography–aided assessment of periodontal furcation involvement

ObjectiveTo assess the accuracy of cone beam computed tomography (CBCT) images obtained at different voxel sizes in detecting ligature-induced furcation involvement.Study DesignFurcation lesions were prepared in 36 molar teeth in rats, and 24 intact teeth served as controls. CBCT Kodak 9000 images at three different voxel sizes were acquired: 0.076 mm, 0.100 mm, and 0.200 mm. Four observers assessed 3 sets of images using a 5-point evaluation scale. κ-Coefficients were calculated for intra- and interobserver agreement. Receiver operating characteristic analysis measured the true positive rate (TPR), true negative rate (TNR), and area under the curve (AUC). Accuracy values were compared by using Mood's Median Chi-Square (α < 0.05).ResultsIntra- and interobserver agreement ranged from 0.600 to 0.999 and from 0.366 to 0.573, respectively. Highest median AUC and true positive rate (TPR) values were obtained for voxel size of 0.076 mm. The highest median TNR values were obtained for the voxel size 0.1 mm. There were no significant differences (P ≥ .05) among the median AUC, TPR, or TNR values for the 3 different CBCT voxel sizes.ConclusionsGiven the limitations of this study, all voxel sizes performed similarly in the detection of furcation involvement.