Evaluation of Metal Artifacts in Cone Beam Computed Tomography by Metal Supported Porcelain Crowns Using Different FOV and Localizations: An In Vitro Study

Introduction

Cone beam computed tomography (CBCT) provides detailed and accurate information for diagnosis and treatment planning of dentomaxillofacial region diseases. CBCT provides high-resolution image of bone and radiation dose is lower than Computed Tomography (CT) (1, 2). CBCT is used successfully in dentistry such as endodontics, orthodontics and surgical procedures (3). Despite the benefits, CBCT could potentially have some complications including metal streak artifacts. Artifacts caused by metals in the scan area affect CT and CBCT images quality (4). Metal artifacts are among the most important factors leading to impaired diagnostic quality of CBCT images (5). These artifacts formed in CBCT reduce the contrast of the image, compromise the image quality by obscuring vital the structures that are desired to be evaluated in the image. Furthermore, interpretation of these images is more time consuming for the operator (6).

In recent years, a wide range of CBCT devices with different exposure parameters have been presented.  Radiation dose is affected by exposure time and tube current and x-ray quality is altered by tube voltage, filtration and FOV (field of view). These factors affect quality of image through noise, contrast resolution and artifacts (7). Tomographic images have variable densities according to localization of the object in FOV (8).Artifacts in images will increase when higher number of objects are located outside of FOV (9).Size of FOV affects CBCT quality and artifact formation based on projection data discontinuity effect (10, 11). Metal-supported porcelain crowns (MSPC) and bridges may be present in the dentition of patients undergoing CBCT scans. Metal artifacts caused by these prostheses may adversely affect the image quality. The aim of this study is to evaluate metal artifacts in CBCT images caused by MSPCs in 4 different scanning protocols using two different FOVs (18x16 and 8x8 cm) and two different localizations measuring them spatially and dimensionally.

Materials and Methods

Twenty MSPCs including 9 molars, 6 premolars and 5 incisors were used in this in vitro study. These MSPCs  obtained from spare ones which were originally fabricated for clinical use. Block of wax was used to stabilize the crowns during the scan (figure 1A). A 18x16 cm cardboard box was used (figure 1B) to mount each MSPCs in the gantry for CBCT scanning(figure 1C). 18x16 cm and 8x8 cm FOVs were selected for scanning.  For localization, two scanning points were identified as central and peripheral for each FOV. The peripheral point was assigned as a midpoint in top level of both FOV. Scout images were taken before actual scan in both FOVs to adjust the position of the object in FOV in lateral and coronal directions. For peripheral position, MSPC was placed at the peripheral point in the cardboard box according to the preferred FOV. Later, in the first lateral scout image taken, MSPC was adjusted to be at midpoint of top and anterosuperior of FOV. Within this setting, image was taken by moving the gantry with imaging computer. In the coronal image, which was the second scout imaging, MSPC was adjusted to be at the midpoint at mediolateral of top point.  The central point was determined as midpoint in middle level of both FOVs. These points were checked in lateral and coronal scout images. Despite peripheral position, central position adjustments were performed by choosing the midpoint of top-bottom and antero-superior in lateral scout image and choosing the midpoint of top-bottom and mediolateral in coronal scout images.  During the scan, bucco-lingual positions of the crowns were adjusted to simulate a dental arch shape. Each crown was scanned 4 times and a total of 80 CBCT images were obtained for evaluation. Images were obtained by NewTom 5G (Quantitative Radiology, Verona, Italy) CBCT device using a standard protocol of 110 kVp and a maximum of 20).  Exposure time was the same in all the scans (18hs). Contrast adjustment were similar to clinical x-rays for patient images. Voxel values ​​were the same in all images (0.25mm). CBCT images were evaluated using NNT software (Quantitative Radiology, Verona, Italy).

For each MSPC, cross-sections were taken in the middle of the crown (thickness: 0.05 mm) to see the bucco-lingual direction. The area around 2.5 mm of the MSPC cross-sectional image was first evaluated. Next, area with metal artifacts within this measured area was measured. Later, artifact-area to total-area ratio was converted to form of a percentage (figure 2). Furthermore, lengths of metal streak artifacts were measured. The cross-sectional image in the bucco-lingual direction was recorded for each MSPC by measuring the maximum linear dimension originated from the metal crown edge (figure 3).   In order to determine the end point of metal streaks, these evaluations were conducted in a dark room and hyperdensed border was considered the end point. To obtain a cross-sectional image parallel to long axis of tooth, each MSPC was embedded in a wax block parallel to the surface. The Kolmogorov-Smirnov and Shapiro Wilk tests were used to test the normality of the data. The results indicated that our data do not show normal distribution.  For statistical evaluation, data collected were compared with Kruskal-Wallis analysis and Mann Whitney U test (P˂0.05). To define intraobserver error rate, measurements were repeated on randomly selected 30% of all images intraclass correlation coefficients (ICC). 

Results   

ICC scores were 0.87-0.92 for all measurements.Area measured from artifacts and linear size measurements means are presented in Table I. In groups where metal artifacts occurred, there was no statistically significant difference between groups with different FOV and localizations (P=0.121) (Table II). However, there was a statistically significant difference in linear size measurements of metal artifacts occurring at different FOV and localizations in MSPC CBCT scans (P=0.000) (table II). Artifacts of linear size measurements in FOV of 18×16 cm localization peripheral was statistically significant difference than other groups (Table III). A statistically significant difference was found between the FOV of 18x16 cm-localization central with FOV of 8x8 cm localization-central scan artifact dimensional measurements  (P=0.046) (Table III).

Table I: Mean value of percentage of measured area and linear dimension measurements.

Table II: Kruskal Wallis test for linear dimension measurements and percentage of area measured of metal artifacts

TableIII: Mann Whitney U test for linear dimension measurements