Reduction of Metal Artifacts Produced by Dental Implants in CBCT Images

Abstract

Cone beam CT, which offers clear 3D images with small voxel size, has been widely used in head and neck diagnoses& especially in the field of implantology. Metal artefacts produced by dental implants might be play a principal role in the degradation of the image quality. The purpose of thisstudy was to evaluate the reduction of metal artefacts produced by dental implants in CBCT images using two fields of view, different KVps and with and without the application of artefact reduction (MAR algorithm) with the ultimate goal of finding if the grey level value surrounding an implant as well as image quality are influenced by these variables.

Three implant holes were drilled in the proposed implant sites; canine, premolar and molar regions of an edentulous mandible. As a control group, the mandible was scanned without implant placement using a small FOV (40 x 50mm), and a medium FOV (50 x 80mm). Imaging was performed for each FOV at 70 KVp, 80 KVp, and 90 KVp. Implants were then placed at the drilled implant sites and the mandible was scanned again using the previous parameters. Images were evaluated using Mimics® 10.01software calculating the average grey level surrounding the implants for quantitative assessment, while qualitative assessment involved subjective appraisal of the streak artefacts of the area surrounding the implants. The differences among groups were analyzed using the ANOVA; whereas, post hoc tests was used to find the significance between groups. The differences between pairs were analyzed using paired t test. Pearson's chi-squared test was applied to sets of categorical data. Seventy kVp group showed significantly larger grey level value than both 80 kVp and 90 kVp groups. Eighty kVp and 90 kVp groups showed insignificant difference in grey level values. Molar region showed significantly smaller grey level value than both premolar and canine regions. Premolar and canine regions showed insignificant differences in grey level values. Post-implant group showed significantly larger grey level value than prior implant group. Five*eight FOV group showed significantly larger grey level value than 4*5 FOV group. With and without MAR groups showed insignificant differences in grey level values. Regarding qualitative analysis within coronal views; the relations between streak artefacts and both different region groups and MAR application groups were significant; P ≤ 0.05. No significant differences of the remaining comparisons were found between groups (P > 0.05). Regarding qualitativeanalysis within sagittal views, the relations between streak artefacts and both different region groups and MAR application groups were significant; P ≤ 0.05. No significant differences of the remaining comparisons were found between groups (P > 0.05).There was a statistically significant very good intra-observer and inter-observer reliabilities regarding all analysis.