Abstract
1740
Objectives: When performing PET/CT scans of patients with dental implants, the implants produce streak and star artifacts in the CT images. These artifacts negate the utility of CT for the spatial localization of PET findings and can propagate, through the CT-based attenuation correction, into the corrected PET images. We evaluated the clinical efficacy for PET/CT of an algorithm that reduced the metal artifacts. Methods: Fifty patients with suspected head and neck malignancies and with dental implants were scanned with a Siemens Biograph™ 6 PET/CT system. CT images through the patient’s dental implants were reconstructed at the resolution of diagnostic CT with a Metal Artifact Reduction (MAR) algorithm that uses projection-completion based on an intermediate image to interpolate regions of the sinogram acquired through metal. The intermediate image is obtained with a MAP reconstruction that uses multi-modal priors. Attenuation correction factors were calculated from the CT images. Three experienced PET/CT readers evaluated the PET and CT images for their depiction of malignancy, when present, for image low- and high-contrast resolution, and for artifact reduction. Results: The high resolution, iterative algorithm reduced the dental artifact and increased the clinical utility of the CT images for the localization of malignancies compared with the other methods. PET images that were corrected with attenuation factors from CT images reconstructed with the MAR algorithm exhibited artifacts that were more subtle and, in most cases, of no clinical significance. Conclusions: The results of this study demonstrate that the high resolution, iterative MAR algorithm enhances the structural and spatial content of clinical CT images in the presence of metal artifacts.
Research Support (if any): Catherine Lemmens receives funding from the Flanders Institute for the Promotion of Innovation through Science and Technology.
- Society of Nuclear Medicine, Inc.