Influence of basis images and skull position on evaluation of cortical bone thickness in cone beam computed tomography

Objectives. The aim of this study was to assess the influence of the number of basis images and the orientation of the skull on the evaluation of cortical alveolar bone in cone beam computed tomography (CBCT). Study Design. Eleven skulls with a total of 59 anterior teeth were selected. CBCT images were acquired by using 4 protocols, by varying the rotation of the tube-detector arm and the orientation of the skull (protocol 1: 360(omicron)/0(omicron); protocol 2: 180(omicron)/0(omicron); protocol 3: 180(omicron)/90(omicron); protocol 4: 180(omicron)/180(omicron)). Observers evaluated cortical bone as absent, thin, or thick. Direct observation of the skulls was used as the gold standard. Intra- and interobserver agreement, as well as agreement of scoring between the 3 bone thickness classifications, were calculated by using the k statistic. The Wilcoxon signed-rank test was used to compare the 4 protocols. Results. For lingual cortical bone, protocol 1 showed no statistical difference from the gold standard. Higher reliability was found in protocol 3 for absent (k = 0.80) and thin (k = 0.47) cortices, whereas for thick cortical bone, protocol 2 was more consistent (k - 0.60). In buccal cortical bone, protocol 1 obtained the highest agreement for absent cortices (k - 0.61), whereas protocol 4 was better for thin cortical plates (k = 0.38) and protocol 2 for thick cortical plates (k = 0.40). Conclusions. No consistent effect of the number of basis images or head orientation for visual detection of alveolar bone was detected, except for lingual cortical bone, for which full rotation scanning showed improved visualization.