The impact on diagnostic yield of the scan mode of cone beam CT images in artificial external root resorption

Abstract

Introduction: Root resorption is an undesired but frequent side effect of orthodontic treatment. Several studies have already aimed to evaluate the performance of CBCT on this topic. However, none have addressed the peculiarities of the most common daily orthodontic scenario. The aim of the study was to evaluate the influence of CBCT scans on the diagnosis of artificial external root resorption in the apical third of anterior teeth. Materials and methods: One hundred extracted human anterior teeth were randomly assigned to 2 uneven groups (51 as the control and 49 as the experimental group). A limited area of the apical third of the root of the teeth of the experimental group was selected and a buffer solution was used to induce tooth subsurface demineralization. Before CBCT image acquisition, each tooth was coated with an approximately 0.3 mm thick layer of utility wax to simulate the radiographic appearance of the periodontal space and placed into an empty mandibular anterior socket of a partially edentulous dry human mandible. The mandible was placed into a polystyrene box filled with water prior to the CBCT examination to simulate soft-tissue attenuation and scattering. The CBCT images were obtained on an i-CAT unit (i-CAT Next Generation, Imaging Sciences International, Hatfield, PA) according to three protocols: (I) half scan (180°), 0.40 mm voxel size; (II) full scan (360°), 0.40 mm voxel size; and (III) full scan (360°), 0.125 mm voxel size. The 300 resultant CBCT DICOM volumes were imported into InVivo software (InVivo5, Anatomage, San Jose, CA) for evaluation by three blinded, previously calibrated observers using a fivepoint confidence rating scale. Cohen's kappa was used to calculate observers' agreement. vi The diagnostic values of sensitivity (Sn), specificity (Sp) and accuracy (Ac) were performed by pooling observer responses for every image modality. Receiver operating characteristic (ROC) were built and the areas under the curve (AUC) were calculated. The Sn, Sp and Ac values were compared by Cochran's Q test. The AUC values were compared by Mann-Whitney U test. Results: The observers' agreement ranged from 0.63 to 0.71, which was interpreted as a substantial agreement. Protocol III (0.125 mm voxel size) displayed the highest Sn (81.63 %), Ac (80.67%) and AUC (0.807). There were statistically significant differences between protocol III and the other two protocols (p < 0.001). The specificity of protocol I (84.97 %) was greater than that of protocols II (69.93 %) and III (79.74 %); however statistically significant difference was only found between protocols I and II (p = 0.005). Conclusion: Although there was no difference in accuracy between the degrees of rotation (half and full scan) within the same voxel size (0.4 mm), there was a considerable difference between those and the smallest voxel size (0.125 mm). In fact, it may be suggested that a more dedicated, high resolution scan should be acquired when one intends to investigate the early stage of external root resorption during orthodontic ...