Aim To evaluate the influence of cone-beam computed tomography (CBCT) filters on diagnosis of simulated endodontic complications.
Methodology Sixteen human teeth, in three mandibles, were submitted to the following simulated endodontic complications: (G1) fractured file, (G2) perforations in the...

Aim To evaluate the influence of cone-beam computed tomography (CBCT) filters on diagnosis of simulated endodontic complications.
Methodology Sixteen human teeth, in three mandibles, were submitted to the following simulated endodontic complications: (G1) fractured file, (G2) perforations in the canal walls, (G3) deviated cast post, and (G4) external root resorption. The mandibles were submitted to CBCT examination (I-Cat((R)) Next Generation). Five oral radiologists evaluated the images independently with and without XoranCat((R)) software filters. Accuracy, sensitivity and specificity were determined. ROC curves were calculated for each group with the filters, and the areas under the curves were compared using ANOVA (one-way) test. McNemar test was applied for pair-wise agreement between all images versus the gold standard and original images versus images with filters (P < 0.05).
Results G1 was the most difficult endodontic complication to diagnosis, followed by G2, G4 and G3. There were no differences between areas under the ROC curves for the filters in all groups; however, Sharpen Super Mild filter had the best results for G1 (0.47), Angio Sharpen Low 3 x 3 for G2 (0.93), Angio Sharpen Low 3 x 3, S9, Shadow and Sharpen for G3 (1.00) and Sharpen 3 x 3 for G4 (1.00). The McNemar test revealed significant differences between all filters with the gold standard (P = 0.00 for all filters) and the originals images (P = 0.00 for all filters) only in G1 group. There were no differences in the other groups.
Conclusions The filters did not improve the diagnosis of the simulated endodontic complications evaluated. Their diagnosis remains a major challenge in clinical practice

Fechado