Click here to

  1. Home
  2. Meeting Program
  3. All Sessions
  4. Abstracts
Session: [Return to Session]

Optimized Practice for Metal Artifact Reduction

J Dumas*, A Bruget, R Dal, L De Marzi, V Calugaru, R Ferrand, Institut Curie, Paris FR

Presentations

PO-GePV-I-54 (Sunday, 7/10/2022)   [Eastern Time (GMT-4)]

ePoster Forums

Purpose: For Radiotherapy planning, to determine via qualitative and quantitative tools which options provide an optimal image quality given the localization and the type of artifact.

Methods: The study was carried out on a Siemens Somatom Open RT CT scanner with four cylindrical phantoms including a homogeneous 18cm diameter, the small and the large CIRS phantoms and the Cheese phantom. Each configuration was built with and without metal artifacts, whose diameter increased with the diameter of the phantom, ranging from small Titanium balls to large steel and titanium bars 2cm in diameter. We compared each configuration with and without artifact correction carried out with iMAR software, Dual energy CT (DECT) and both of them. The evaluation was based on the Image Quality Index, the Histogram representation of image subtraction and an experimented physician performed visual analysis. We also evaluated the Hounsfield units’ variation, distortion, Signal to noise and Contrast to noise ratio. In order to quantify the dose improvement from the DECT we used the data provided by the Dose Archiving System in our institution.

Results: IQI distortion reduced by 12% with iMAR and by 13% when DECT was used except on bone ROI where DECT did not improved distortion. Image subtraction showed a strong improvement with iMAR, enhanced with DECT. SNR and CNR were improved in all ROI cases, as 2% to 30%, depending on ROI mass density. Hounsfield units’ improvements were more or less significant depending on the size of the material causing artifact, ROI mass density and phantom size. Visual expertise also showed better results with iMAR, improved by DECT. Doses were lower or equal to standard procedure.

Conclusion: iMAR algorithm and DECT were efficient depending on the strength of the artifact, the size of the material and must be adapted to the size of the phantom.

Keywords

CT, Image Artifacts

Taxonomy

Not Applicable / None Entered.

Contact Email

Share:

Additional Links