Tuned‐Aperture Computed Tomography for Detection of Induced Mid‐Buccal/Lingual Alveolar Bone Defects

Abstract

The detection of bone loss in mid-buccal and lingual crests is impossible using conventional radiographs because of the superimposition of overlying anatomy and lack of three-dimensional information. The purpose of this study was to compare the diagnostic efficacy of tuned-aperture computed tomography (TACT) and conventional two-dimensional direct digital radiography (DDR) in an in vitro environment. A total of 45 mandibular molars had 0.8-mm lesions on mid-buccal/lingual crestal areas. Half of the sites received defects, whereas the other half served as controls. Nine DDR images were used to generate TACT slices that were further subjected to iterative restoration (TACT-IR). Eight observers used a confidence rating scale to record diagnoses. Receiver operating characteristic (ROC) analysis was done, and areas under the curves were computed (A(z)) as indicators of diagnostic accuracy. Analysis of variance (ANOVA) was used to test for effects of observer, imaging modality, and location on the detection of lesions. TACT-IR performed significantly better than DDR. There was a significant difference in the accuracy of diagnosis based on observers (P <0.001). TACT-IR appears to be the imaging modality of choice for the detection of small osseous changes on crestal bone in mid-buccal/lingual sites.