OBJECTIVE: The purpose of this study was to evaluate the effect of the number of iterative restorations (IR) on the diagnostic accuracy of Tuned Aperture Computed Tomography images (TACT) for detection of induced vertical/oblique root fractures in mandibular teeth. STUDY DESIGN: Fractures were induced in 28 single-rooted, endodontically treated cadaver mandibular teeth while another 26 such teeth served as controls. TACT slices reconstructed by using 9 arbitrary projections (angular disparity of 9 degrees -20 degrees) were iteratively restored 1, 2, and 3 times. Eight observers scored their diagnoses by using a confidence rating scale. Ground truth was available from direct examination with transillumination, following extraction. RESULTS: Mean areas under the representative ROC curves for the 3 operations were 0.96 (3 IRs), 0.87 (2 IRs), and 0.86 (1 IRs). Analysis of variance demonstrated significant differences between 3 and 1 to 2 IRs (P <.01), as well as between observers (P <.01). CONCLUSION: Diagnostic accuracy of TACT for vertical/oblique root fracture detection improved after 3 IRs.
OBJECTIVE: The purpose of this study was to evaluate the effect of the number of iterative restorations (IR) on the diagnostic accuracy of Tuned Aperture Computed Tomography images (TACT) for detection of induced vertical/oblique root fractures in mandibular teeth. STUDY DESIGN:Fractures were induced in 28 single-rooted, endodontically treated cadaver mandibular teeth while another 26 such teeth served as controls. TACT slices reconstructed by using 9 arbitrary projections (angular disparity of 9 degrees -20 degrees) were iteratively restored 1, 2, and 3 times. Eight observers scored their diagnoses by using a confidence rating scale. Ground truth was available from direct examination with transillumination, following extraction. RESULTS: Mean areas under the representative ROC curves for the 3 operations were 0.96 (3 IRs), 0.87 (2 IRs), and 0.86 (1 IRs). Analysis of variance demonstrated significant differences between 3 and 1 to 2 IRs (P <.01), as well as between observers (P <.01). CONCLUSION: Diagnostic accuracy of TACT for vertical/oblique root fracture detection improved after 3 IRs.