Influence of Voxel Size in the Diagnostic Ability of Cone Beam Tomography to Evaluate Simulated External Root Resorption Gabriela Salatino Liedke, BDS,* Heloísa Emília Dias da Silveira, BDS, MSc, PhD,* Heraldo Luis Dias da Silveira, BDS, MSc, PhD,* Vinícius Dutra, BDS, PhD, † and José Antônio Poli de Figueiredo, BDS, MSc, PhD ‡ Abstract This study assessed in vitro the diagnostic ability of cone beam computed tomography (CBCT) scans with different voxel resolutions in the detection of simulated external root resorption (ERR). For that purpose, 59 teeth were viewed through iCAT tomography (Imaging Sciences International, Inc, Hatfield, PA) following three protocols in which the variation was the voxel resolu- tion (0.4, 0.3, and 0.2mm). A calibrated examiner blinded to the protocol assessed the images through the i-CAT View software (Imaging Sciences Interna- tional, Inc). The chi-square statistical analysis did not show associations between voxel resolution, section plane, size of cavity, and radicular third. Sensitivity and specificity values were similar. However, likelihood ra- tio values of 6.4 for a 0.4-mm voxel, 16 for a 0.3-mm voxel, and 12 for a 0.2-mm voxel were found. It was concluded that CBCT is a reliable method for the in- vestigation of simulated ERR, and a 0.3-mm voxel appeared to be the best protocol, associating good diagnostic performance with lower X-ray exposure. (J Endod 2009;35:233–235) Key Words Cone beam computed tomography, diagnosis, external root resorption E xternal tooth resorption is a multifactorial process of irreversible structural loss that may lead to the loss of the tooth (1, 2). The diagnosis is based on clinical and radiographic examination (3). Nevertheless, conventional radiographs produce false- negative results in 51.9% and false-positive results in 15.3% of cases investigated (4). It has been shown that lesions under 0.6 mm in diameter and 0.3 mm in depth are not detected by conventional periapical radiography (5, 6). The main difficulty in the radiographic diagnosis is encountered when the lesions are on the buccal or lingual tooth surfaces (7). Aiming to evaluate how the elimination of structure superimposition improves visualization, da Silveira et al (8) conducted an in vitro study and observed that multislice computed tomography (MSCT) scans offer high sensitivity and specificity to the detection of simulated cavities on the buccal surface of axial tooth sections, except for small cavities present in the apical third. Diagnostic efficiency and correct characterization of the location and the size of tooth resorption on the root surface are essential aspects to the definition of the ade- quate therapeutic approach and eventually to success in treatment (6). In this sense, three-dimensional images have proved to be important tools in dental practice. Some studies have reported the use of MSCT for this purpose, although more recently cone beam computed tomography (CBCT) scans have shown other advantages, such as (1) the presence of a collimator, which confines irradiation to the area of study; (2) the isotropic nature of voxels, which affords reconstructions of the same quality of the original image; (3) the speed of conduction (between 10 and 70 seconds); and (4) the low radiation dose the patient is submitted to (around 1/60 of the MSCT dose) (9 –13). Another advantage is that voxel size bears a positive correla- tion with image quality and exposure dose. The aim of imaging in dentistry is to promote the combined advantages of early diagnosis and reduced exposure of patients to x-rays. With this in mind, the present study was designed to evaluate the diagnostic capacity of CBCT scans to detect external root resorption on the buccal root surface. The experimental variables were cavity size, location on the radicular third, visualization plan (frontal, sagittal, and axial), and voxel resolution adopted. Materials and Methods Sixty human mandibular incisors from the Endodontics Department tooth bank of the School of Dentistry were randomly numbered from 1 to 60. The teeth were not sectioned, but their root portions were divided into thirds: cervical, middle, and apical. A total of 180 thirds was obtained. For each root third (cervical, middle, and apical), there were 4 simulation pos- sibilities: small, medium, or large cavity or no cavity. Therefore, there were 12 possible combinations, and, for each combination (third  size), 15 teeth were randomly selected. To simulate external root resorption, the teeth were placed in plaster bases, and cavities of 0.6, 1.2 or 1.8 mm in diameter and 0.3, 0.6 or 0.9 mm in depth (small, medium and large cavities) were drilled according to a protocol reported in the liter- ature (5, 6, 14, 15). The cavities were drilled on the buccal surfaces with high-speed and round burs 0.6, 1.2, and 1.8 mm in diameter (KG, Sorensen, Brazil) and the From the *Department of Surgery and Orthopedics, School of Dentistry, Federal University of Rio Grande do Sul (UFRGS), Brazil; † CDI Imaging Diagnosis Centre; and ‡ Post-Graduate Program in Dentistry, Pontifical Catholic University of Rio Grande do Sul, Rio Grande do Sul, Brazil. Address requests for reprints to Dr José Antônio Poli de Figueiredo, Post-Graduate Program in Dentistry, Pontifical Catholic University of Rio Grande do Sul–PUCRS, Av. Ipiranga 6681 Prédio 6 sala 206, CEP 90619-900 Porto Alegre, RS, Brazil. E-mail address: jose.figueiredo@pucrs.br. 0099-2399/$0 - see front matter Copyright © 2008 American Association of Endodontists. doi:10.1016/j.joen.2008.11.005 Basic Research—Technology JOE — Volume 35, Number 2, February 2009 Voxel Size and CBT to Evaluate External Root Resorption 233