Influence of a Metal Artifact Reduction Filter on the Diagnostic Accuracy of Complete and Incomplete Vertical Root Fractures

This study aimed to assess the influence of the e-vol DX BAR filter on the diagnostic accuracy of complete and incomplete vertical root fractures (VRFs). Twenty single-rooted teeth were selected. The tooth crowns were removed, and the root canals were prepared up to a 40/.06 instrument (ProTaper Next). Each specimen was scanned in a Prexion 3D cone-beam computed tomography (CBCT) device, in a dry human skull, in 4 different situations: no root canal filling, gutta-percha, fiberglass post, and metal post. The specimens were fractured in a universal testing machine by using a customized wedge. Each specimen was reinspected to confirm the presence of incomplete VRFs. Another set of CBCT scans was performed. Then, the fractures were completed in the universal testing machine, the teeth were reinspected, and the CBCT images were acquired again. Images were assessed by using the e-Vol DX software system, with the BAR filter and with the original images, for the diagnosis of VRFs. Statistical analysis was performed by using the DeLong and McNemar tests for the comparison of the area under the receiver operating characteristic curves (AUC), accuracy, sensitivity, and specificity values. The use of the BAR filter did not improve the diagnostic values of AUC, accuracy, sensitivity, and specificity in both fracture patterns (P > .05). Incomplete VRFs presented significantly lower AUC, accuracy, and sensitivity compared with complete VRFs (P