Accuracy of vertical root fracture detection in endodontically treated teeth using a cone-beam computed tomography with a variable voxel size in an in vitro model

Aim: This study aims to study the effect of different voxel sizes on the accuracy of detecting vertical root fractures (VRFs) in the presence and absence of gutta percha using cone-beam computed tomography (CBCT). Methods: Sixty single-rooted premolars were divided into two groups of thirty each, Group one-obturated and Group two-unobturated. Incomplete VRF was induced in 15 samples of each group employing universal testing machine. Teeth were placed randomly in dry mandibular dental sockets and scanned using Dentium CBCT, at voxel size 0.08 mm and 0.125 mm, the images acquired were examined by two independent observers. Inter-observer variability analysis was performed using Cohen’s Kappa Statistics. Chi-square test was employed for comparing VRF diagnosis and level of interference. Receiver operating characteristic analysis was performed to estimate sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) as well as accuracy. The level of statistical significance was set at P < 0.05. Results: Inter-observer kappa coefficient indicates perfect agreement. In obturated group, statistically significant difference was found with respect to sensitivity, specificity, PPV as well as NPV between scans at voxel sizes 0.08 mm and 0.125 mm. The mean area under curve and accuracy respectively was 0.847, 83.3% for 0.08 mm and 0.698, 66.7% for 0.125 mm voxel ( P < 0.001). No interference due to artifact in VRF detection was observed in 33.3% samples ( n = 10) at voxel 0.08 mm, whereas only 13.3% samples ( n = 4) showed no effect of artifact in VRF detection at 0.125 mm ( P < 0.008). Conclusion: CBCT using smaller voxel size (0.08 mm) favor VRF detection more accurately and the interference due to gutta percha is also least observed in CBCT images obtained using smaller voxel.