Scientific Session 01 — SS01: Neuroradiology - Head, Neck, and SpineMonday, May 6, 2019
2268. Combining Artifact Reduction Algorithms and Monoenergetic Images From SDCT Allows For Reduction of Severest Artifacts From Dental Implants
Laukamp K1,2*, Zopfs D2, Lennartz S2, Borggrefe J2, Große Hokamp N1,2 1. UH Hospitals, Cleveland, Ohio; 2. University Hospital Cologne, Cologne, Germany
Address correspondence to K. Laukamp (email@example.com)
Objective: This study compares reduction of severe metal artifacts from large dental implants and bridges by use of virtual monoenergetic images (VMI), metal artifact reduction algorithms or reconstructions (MAR), and a combination of both methods (MAR-VMI) to conventional CT images (CI), all acquired with a clinical spectral-detector CT (SDCT).
Materials and Methods: Forty-one SDCT datasets of patients for which additional MAR reconstructions due to strongest artifacts from large dental implants and bridges were available and were included in this retrospective study. CI, VMI, MAR, and MAR-VMI in a range of 70–200 keV (10 keV increment) were reconstructed from the same scans. ROI-based objective image analyses were performed by measurement of attenuation (HU) and standard deviation in the most pronounced hypo- and hyperdense artifacts in soft tissue with artifacts (e.g., mouth floor, soft palate). Extent of artifact reduction, diagnostic assessment of soft tissue, and appearance of new artifacts were rated visually on 5-point Likert scales by two radiologists.
Results: The hypo- and hyperattenuating artifacts showed a significant increase and decrease of corrected attenuation/HU values in MAR and MAR-VMI (CI, MAR, MAR-VMI200keV: -369.8±239.6, -37.3±109.6, -46.2±71.0 HU, p<0.001; and 274.8±170.2, 51.3±150.8, 36.6±56.0, p<0.001, respectively). Higher VMI keV values in hyperdense artifacts allowed for an additional artifact reduction; however, this trend was not significant. Regarding attenuation and image noise, artifacts in mouth floor or soft palate were also significantly reduced by MAR and MAR-VMI. Visually, high-keV VMI, MAR, and MAR-VMI reduced artifacts and improved diagnostic assessment of the mouth floor and soft palate; however, overcorrection and new artifacts were detected that mostly did not hamper diagnostic assessment. Overall interrater agreement was excellent (ICC=0.85).
Conclusion: In the presence of severe artifacts due to large oral implants, MAR is a powerful method for artifact reduction. For hyperdense artifacts, MAR supplemented by VMI of 160–200 keV from SDCT yield optimal artifact reduction. This combination of techniques is recommended to improve diagnostic assessment in imaging of the head and neck.