Vertical root fractures (VRFs) are difficult to diagnose by conventional periapical

Transcription

1 Detection of Vertical Root Fractures by Using Cone Beam Computed Tomography with Variable Voxel Sizes in an In Vitro Model Senem Yig it Özer, DDS, PhD Abstract Introduction: The aim of this study was to compare the diagnostic accuracy of cone beam computed tomography (CBCT) scans with different voxel resolutions in the detection of simulated vertical root fracture (VRF). Methods: Sixty teeth, 30 with root fractures and 30 without, were examined through i-cat tomography at 4 different voxel resolutions (0.125, 0.2, 0.3, 0.4 mm). Three observers scored the presence of VRF in all modalities. Sensitivity, specificity, likelihood ratio, and accuracy were calculated for each modality. Results: Chi-square statistical analysis showed no significant difference among the resolutions or observers. However, positive likelihood ratio values of 24.5 for a mm voxel, for a 0.2-mm voxel, 13.2 for a 0.3-mm voxel, and 13 for a 0.4-mm voxel were found. Accuracy results confirmed the likelihood ratios as 0.97, 0.96, 0.93, and 0.92, respectively. Conclusions: CBCT scans were reliable in detecting simulated VRF, and a 0.2-mm voxel was the best protocol, considering the lower x-ray exposure and good diagnostic performance. (J Endod 2011;37:75 79) Key Words Cone-beam computed tomography, diagnosis, vertical root fracture, voxel From the Department of Operative Dentistry and Endodontics, Faculty of Dentistry, Dicle University, Diyarbakir, Turkey. Address requests for reprints to Dr Senem Yiğit Özer, Dicle University, Faculty of Dentistry, Department of Operative Dentistry and Endodontics, 21280, Diyarbakir, Turkey. address: senemygt@hotmail.com /$0 - see front matter Copyright ª 2011 American Association of Endodontists. doi: /j.joen Vertical root fractures (VRFs) are difficult to diagnose by conventional periapical radiography (CPR), and they can be overlooked if the x-ray beam does not pass along the fracture line. Thus, 2 or more radiographs taken at different angles are recommended (1, 2). Two-dimensional (2D) intraoral radiography systems fail to provide information regarding teeth and adjacent structures in the third dimension, and the superimposition of other structures further limits sensitivity (3). Thus, a need exists for an imaging system that provides 3-dimensional (3D) information in treatment assessment of root canal systems. Recently a 3D diagnostic imaging system, cone beam computed tomography (CBCT), has been reported to be more accurate than CPR for detecting longitudinal root fractures with high image quality, permitting direct visualization of fracture lines, which are masked in CPR (4, 5). CBCT scans have been found to be more accurate in detecting VRFs in many studies (5 10). CBCT has been found to perform better than CPR and other intraoral techniques in detecting VRFs; however, many studies have concluded that the radiation doses from CBCT were higher than from conventional dental radiography techniques (11). Briefly, digital images are composed of elements called pixels. A pixel is a picture element, a square in a 2D matrix. Each pixel has a specific size, intensity value, and location within the matrix. A voxel is a volume element, and it is the cubed form of a pixel, having a third dimension. Three-dimensional images are composed of voxels instead of the pixels in 2D digital images. The size of each voxel is determined by its height, width, and thickness, and a voxel is the smallest element of the 3D radiograph image volume (12, 13). CBCT images use isotropic voxels, meaning they have the same height, width, and thickness. The isotropic nature of the voxels affords the same quality as the original image in reconstructions. Image quality has been described as the visibility of diagnostically important structures in the CT image (14, 15), and a number of factors affect image quality such as the milliamperage settings, voxel size, field of view (FoV), and detector type of the scanner (15 17). Diagnosis with CBCT offers 3D images without superposition of adjacent structures, and it is also a noninvasive technique with high accuracy (18 23). However, the radiation dose is equivalent to that needed for 4 15 panoramic radiographs (11),andunfortunately,imagequalityisdirectlyrelatedtoradiationexposure (24). Voxel size has been reported to have a positive correlation with image quality (contrast and resolution) and exposure dose (6, 25). ALARA is the acronym for as low as reasonably achievable and is a basic principle for diagnostic radiology. Keeping the dose as low as possible while still obtaining the needed information constitutes the basis of this principle (26). It is known that with smaller voxel sizes, radiation exposure would be higher (12, 24, 27). Without sacrificing image quality and adopting the ALARA principle, the ability to select various voxel settings would be helpful in reducing the radiation dose to the patient. Thus, we compared the diagnostic accuracy of different voxel sizes for the detection of experimentally induced VRFs. Materials and Methods The experimental group consisted of 60 recently extracted human maxillary premolars, inspected with a stereomicroscope (MT4200 binocular; Meiji Techno JOE Volume 37, Number 1, January 2011 Detection of Vertical Root Fractures with CBCT with Variable Voxel Sizes 75

2 TABLE 1. VRFs Identified from a Total of 60 Teeth with Different Voxel Resolutions by 3 Observers Observer 1 Observer 2 Observer 3 Total Voxel size (mm) True False True False True False True False X 2 test P =.7 P =.06 Significance of c 2 test for observers voxels is defined as P value. (UK) Ltd, Axbridge, Somerset, UK) for the absence of VRFs. An access opening was made for each tooth, and the root canals were prepared with the ProTaper rotary system (Dentsply/Maillefer, Tulsa, OK) up to size F3. The teeth were divided into 2 groups of 30, with root fractures or intact. Root fractures were generated in the vertical plane according to a protocol reported in the literature (6). Five tooth roots broke into more than 3 fragments and were excluded from the study. Two fragments for 30 teeth were relocated with glue (Scotch Gel Universal; 3M, Leiden, The Netherlands). One human mandible was coated with 3 layers of dental wax buccally and lingually to simulate soft tissue and to reduce artifacts in the image (28). An observer who was not involved in preparation placed the teeth into sockets of the mandible at the most suitable position mesiodistally. They were held in place in the alveolus socket with boxing wax and were coded during imaging sessions. CBCT Scans and Data Treatment All images were evaluated by 3 observers (2 endodontists and 1 radiologist) at different sessions. CBCT images were scanned by using i-cat CBCT (120 KvP, 5 ma; Imaging Sciences, Hatfield, PA). Scans of the mandible were made with 4-cm FoV selection at 4 resolutions: mm voxel (4 seconds for acquisition), 0.2-mm voxel (4 seconds for acquisition), 0.3-mm voxel (7 seconds for acquisition), and 0.4- mm voxel (7 seconds for acquisition). Four-second and sevensecond acquisition periods with shorter scanning time were selected to reduce radiation dose, even though i-cat has longer scanning acquisition times with the same voxel sizes. After each scan with different voxel sizes, the scanner was calibrated. The mandible was placed on a horizontal plane and fitted to the chin support, with its occlusal plane parallel to the plate. By using the reference light lines in displaying the x and y coordinates, the chin support was moved upward, so that the y- line was placed between the central incisors sockets. Data sets were in DICOM 3 file format. The images were analyzed with the software for the CBCT machine (XoranCAT, version ; Xoran Technologies, Ann Arbor, MI). The images for each tooth in the 4 resolutions were reconstructed separately. All images were displayed on a 19-inch flat-panel screen (EIZO Flexscan S inch LCD monitor; Eizo Nanao Corp, Ishikawa, Japan). Detection of VRFs was decided by direct visualization of a radiolucent line on the root surface. The image enhancement mode of the software was not used for standardization of image quality. Observers had been trained and calibrated for reading CBCT in a previous pilot study. Each observer evaluated the 4 voxel resolutions (0.125, 0.2, 0.3, 0.4 mm) by using axial, sagittal, and coronal directions, and a single score was obtained for each tooth. Observers were not informed that some teeth were intact without fractures. Observation time was not restricted, and at a 1-week interval, the same observations were repeated. Weighted kappa coefficients were calculated to assess intraobserver and interobserver agreement. The 4 voxel resolutions (0.125, 0.2, 0.3, 0.4 mm) were analyzed independently by c 2 test at 5% significance. Sensitivity, specificity, positive likelihood ratio, and accuracy were evaluated for diagnostic performance according to the different voxel resolutions. The fracture/non-fracture assessment was scored binarily: 1 if a fracture was present or 0 if it was not. Fracture/non-fracture scores were evaluated as follows: correct identification of a non-fractured root (true negative [TN]), correct identification of fracture site in a fractured root (true positive [TP]), identification of a fracture in a nonfractured root (false positive [FP]), incorrect identification of a fracture site in a fractured root (false negative [FN]), and failure to identify a fracture in a fractured root (false negative [FN]). Sensitivity and specificity among the 4 modalities (0.125-, 0.2-, 0.3-, 0.4-mm voxel) were calculated as follows: Sensitivity = (TP/TP) + FN, Specificity = (TN/TN) + HP, Positive Predictive Value = (TP/TP) + FP, Negative Predictive Value = (TN/TN) + FN, Accuracy = (TP + TN)/(TP + TN + FP + FN), Positive Likelihood Ratio = Sensitivity/(1 Specificity). Results Intraobserver kappa values were above 0.86, indicating very good intraobserver agreement. Table 1 shows the number of true (TP + TN) and false (FP + FN) readings among the 3 observers and 4 different voxel resolutions. On the basis of the observed c 2 test results, no difference (P =.06) was found among voxel sizes or observers (P =.7). Table 2 shows the results of diagnostic performance tests according to voxel resolution. The results indicate that voxel sensitivity and specificity values were similar; the highest accuracy was attained with mm (97%) and in descending order, 0.2 mm (96%), 0.3 mm TABLE 2. Results of Diagnostic Performance Test with Different Variables at 4 Voxel Resolutions Voxel size (mm) Test Sensitivity Specificity Positive predictive value Negative predictive value Accuracy (%) Positive likehood ratio (%) Yiğit Özer JOE Volume 37, Number 1, January 2011

3 Figure 1. CBCT images of a fractured root with 4 different voxels in 3 planes. (1a) Axial plane, mm voxel; (1b) sagittal plane, mm voxel; (1c) coronal plane, mm voxel; (2a) axial plane, 0.2-mm voxel; (2b) sagittal plane, 0.2-mm voxel; (2c) coronal plane, 0.2-mm voxel; (3a) axial plane, 0.3-mm voxel; (3b) sagittal plane, 0.3-mm voxel; (3c) coronal plane, 0.3-mm voxel; (4a) axial plane, 0.4-mm voxel; (4b) sagittal plane, 0.4-mm voxel; (4c) coronal plane, 0.4-mm voxel. (93%), and 0.4 mm (92%). Furthermore, the positive likelihood ratio results are better with mm and 0.2-mm voxel sizes (24% 24.25%) than with 0.3-mm and 0.4-mm voxel sizes (13% 13.2%). Discussion The diagnosis of VRF is usually problematic because it often necessitates prediction rather than a definite identification (28). Because radiographs and CBCT scans acquired by using different imaging parameters cannot exactly visualize these fractures (29, 30), the location and defect size cannot always be objectively assessed until the tooth has been extracted or a simultaneous flap operation has been performed for direct visualization. Also, determining the position and extent of the fracture might be helpful to decide when to recommend extraction or treatment possibilities. The detection of VRF is usually not performed JOE Volume 37, Number 1, January 2011 Detection of Vertical Root Fractures with CBCT with Variable Voxel Sizes 77

4 during treatment, which leads to bone loss, pain, and malfunction of the involved area discomforting the patient maybe for years (31). Thus, identification of the problem in early stages as soon as the problem gets more complex is quite important. The present study compared the diagnostic accuracy of CBCT scans with different voxel resolutions in the detection of simulated VRFs. The dental CBCT, i-cat, chosen for this study has been shown to be more accurate than digital and conventional 2D intraoral techniques in detecting VRFs (5 7, 10). Hassan et al (6) compared 5 different CBCTs for the detection of VRFs and reported that the Next Generation i-cat was the most accurate system (with a voxel of 0.25 mm), followed by the Scanora 3D (voxel of 0.2 mm), Accuitomo-XYZ (voxel of 0.25 mm), Newtom 3G (voxel of 0.2 mm), and Galileos 3D (voxel of 0.3 mm). Generally, performance was better with 0.25-mm and 0.2-mm voxels than with the 0.3-mm voxel size. When Wenzel et al (7) evaluated the diagnostic accuracy of CBCT and a photostimulable storage phosphor plate system (PSP) to detect VRFs, they used an i-cat scanner with mm and 0.25-mm voxel sizes. Higher accuracy was reported with the smaller voxel sizes; mm voxels showed a specificity of 98 and sensitivity of 87. A 0.25-mm voxel resolution was more accurate than the periapical PSP system, and high-resolution scanning was recommended in cases of suspected VRF that could not be visualized in periapical images. Also, in an in vitro study to detect VRF by using CBCT, Kamburoğlu et al (10) compared 2 different CBCT units with varying resolutions of 0.19-, 0.1-, and 0.3-mm voxel sizes. The 0.19-mm and 0.1-mm voxel resolutions achieved better results than the 0.3-mm resolution, but with higher reconstruction times, meaning more radiation exposure. Liedke et al (25) also used i-cat in assessing diagnostic ability in the detection of simulated external root resorption with different voxel resolutions (0.2, 0.3, 0.4 mm); the best results were achieved with a voxel size of 0.2 or 0.3 mm. According to the ALARA principle, the 0.3-mm voxel size with a shorter scanning time was recommended to reduce the patient s x-ray exposure. Tanimoto et al (24) reported that choosing a small voxel size without changing the radiation dose increases the resolution. In the present study, no significant differences were found among observers or voxel sizes, and the sensitivity/specificity results were high for all groups (sensitivity/specificity for mm, 98/96; 0.2 mm, 97/ 96; 0.3 mm, 93/93; 0.4 mm, 91/93). The values revealed that all groups were similar in terms of sensitivity and specificity. The positive likelihood ratio showed that the mm and 0.2-mm voxel resolutions were better than the 0.3-mm and 0.4-mm voxel resolutions (24.5, 24.25, 13.2, and 13, respectively) (Fig. 1). Accuracy results also showed that the mm and 0.2-mm resolutions were more successful than the 0.3-mm and 0.4-mm resolutions in detecting VRFs, although the latter 2 voxel resolutions are also acceptable under clinical conditions (Fig. 1). However, there were cases identified as False-Negative in readings that failed to identify a fracture in a fractured root in all voxel sizes. Those might have resulted from low signal-to-noise ratio as a result of decreasing acquisition time to reduce the radiation dose. The 4 voxel resolutions showed similar results in detecting simulated VRFs. However, accuracy was higher and the decision was easier with mm and 0.2-mm voxels. The best voxel selection would be 0.2 mm, with a shorter scanning time and reduced radiation exposure to the patient. Conclusions CBCT is a reliable imaging system to detect VRFs, and considering the ALARA principle, 0.2-mm resolution is the best choice for diagnostic use. References 1. Andreasen FM, Andreasen JO. Root fractures. In: Andreasen JO, Andreasen FM, eds. Textbook and color atlas of traumatic injuries to the teeth. 3rd ed. Copenhagen, Denmark: Munsksgaard; 1994: Kositbowornchai S, Nuansakul R, Sikram S, Sinahawattana S, Saengmontri S. Root fracture detection: a comparison of direct digital radiography with conventional radiography. Dentomaxillofac Radiol 2001;30: Mora MA, Mol A, Tyndall DA, Rivera EM, Hill C. In vitro assessment of local computed tomography for the detection of longitudinal tooth fractures. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2007;103: Bernardes RA, de Moraes IG, Duarte MAH, Azevedo BC, de Azevedo JR, Bramante CM. Use of cone-beam volumetric tomography in the diagnosis of root fractures. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2009;108: Hassan B, Metska ME, Ozok AR, van der Stelt P, Wesselink R. Detection of vertical root fractures in endodontically treated teeth by a cone beam computed tomography scan. J Endod 2009;35: Hassan B, Metska ME, Ozok AR, van der Stelt P, Wesselink PR. Comparison of five cone beam computed tomography systems for the detection of vertical root fractures. J Endod 2010;36: Wenzel A, Neto FH, Frydenberg M, Kirkevang LL. Variable-resolution cone-beam computerized tomography with enhancement filtration compared with intraoral photostimulable phosphor radiography in detection of transverse root fractures in an in vitro model. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2009; 108: Hannig C, Dullin C, Hulsmann M, Heidrich G. Three-dimensional, non-destructive visualization of vertical root fractures using flat panel volume detector computer tomography: an ex vivo in vitro case report. Int Endod J 2005;38: Nair MK, Nair UDP, Gröndahl HG, Webber RL, Wallace JA. Detection of artificially induced vertical radicular fractures using tuned aperture computed tomography. Eur J Oral Sci 2001;109: Kamburoğlu K, Murat S, Yüksel SP, Cebeci AR, Horasan S. Detection of vertical root fracture using cone-beam computerized tomography: an in vitro assessment. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2010;109:e Scarfe WC, Farman AG, Sukovic P. Clinical applications of cone-beam computed tomography in dental practice. J Can Dent Assoc 2006;72: Farman A, Scarfe WC. The basics of maxillofacial cone beam computed tomography. Semin Orthod 2009;15: Hatcher DC, Aboudara CL. Diagnosis goes digital. Am J Orthod Dentofacial Orthop 2004;125: Sohaib SA, Peppercorn PD, Horrocks JA, Keene MH, Kenyon GS, Reznek RH. The effect of decreasing mas on image quality and patient dose in sinus CT. Br J Radiol 2001;74: Cohnen M, Fischer H, Hamacher J, Lins E, Kotter R, Modder U. CT of the head by use of reduced current and kilovoltage: relationship between image quality and dose reduction. AJNR Am J Neuroradiol 2000;21: Gündogdu S, Mahmutyazicioglu K, Ozdemir H, Savranlar A, Asil K. Assessment of image quality of a standard and three dose-reducing protocols in adult cranial CT. Eur Radiol 2005;15: Rustemeyer P, Streubühr U, Suttmoeller J. Low-dose dental computed tomography: significant dose reduction without loss of image quality. Acta Radiol 2004;45: Arai Y, Tammisalo E, Iwai K, Hashimoto K, Shinoda K. Development of a compact computed tomographic apparatus for dental use. Dent Maxillofac Radiol 1999; 28: Cotton TP, Geisler TM, Holden DT, Schwartz SA, Schindler WG. Endodontic applications of cone beam volumetric tomography. J Endod 2007;33: Nair MK, Nair UP. Digital and advanced imaging in endodontics: a review. J Endod 2007;33: Trope M, Pettigrew J, Petras J, Barnett F, Tronstad L. Differentiation of radicular cyst and granulomas using computerized tomography. Endod Dent Traumatol 1989;5: Estrela C, Bueno MR, Leles CR, Azevedo BC, Azevedo JR. Accuracy of cone beam computed tomography, panoramic and periapical radiographic for the detection of apical periodontitis. J Endod 2008;34: Estrela C, Bueno MR, Azevedo B, Azevedo JR, Pecora JD. A new periapical index based on cone beam computed tomography. J Endod 2008;34: Tanimoto H, Arai Y. The effect of voxel size on image reconstruction in cone-beam computed tomography. Oral Radiol 2009;25: Liedke GS, da Silveira HED, da Silveira HLD, Dutra V, de Figueiredo JAP. Influence of voxel size in the diagnostic ability of cone beam tomography to evaluate simulated external root resorption. J Endod 2009;35: Farman AG. Alara still applies. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2005;4: Yiğit Özer JOE Volume 37, Number 1, January 2011

5 27. Ngan DC, Kharbanda OP, Geenty JP, Darendeliler MA. Comparison of radiation levels from computed tomography and conventional dental radiographs. Aust Orthod J 2003;19: Berman LH, Hartwell G. Diagnosis. In: Cohen S, Hargreaves KM, eds. Pathways of the pulp. 9th ed. St Louis, MO: Mosby; 2006: Patel S, Dawood A, Ford TP, et al. The potential applications of cone beam computed tomography in the management of endodontic problems. Int Endod J 2007;40: Nesari R, Rossman LE, Kratchman SI. Cone-beam computed tomography in endodontics: are we there yet? Compend Contin Educ Dent 2009;30: Fuss Z, Lustig J, Katz A, Tamse A. An evaluation of endodontically treated vertical root fractured teeth: impact of operative procedures. J Endod 2001; 27:46 8. JOE Volume 37, Number 1, January 2011 Detection of Vertical Root Fractures with CBCT with Variable Voxel Sizes 79