CT artifact-reduction software gets put through its paces

An independent review of four CT vendors’ metal artifact-reduction (MAR) solutions has shown that all four appreciably improve image quality. However, the reviewers noted additional artifacts and degradation of image quality, leading them to recommend carefully evaluating each software package’s algorithm case by case in clinical practice.

In a study running in the August edition of the British Journal of Radiology, Karin Andersson, MSc, of Örebro University in Sweden and colleagues describe their work using the various MAR techniques to image a phantom simulating a patient with bilateral hip implants.

The team imaged the phantom on Philips’s Ingenuity Core system, Toshiba’s Aquilion One Vision Edition, GE’s Discovery 750HD and Siemens’s Somatom Definition Flash.

In addition to acquiring CT images with the specific MAR technique available with each respective scanner, they also acquired 120-kVp CT images—aka “uncorrected images”—without using any MAR technique on each scanner.

They further obtained images using MAR algorithms for single-energy CT data or dual-energy CT (DECT) data and by monoenergetic reconstructions of DECT data. These data were visually graded by five radiologists using 10 image-quality criteria.

The reviewers’ key findings:

  • The MAR algorithms, in general, improved the image quality based on the majority of the criteria with a statistical improvement in overall image quality.
  • Degradation of image quality, such as new artifacts, was seen in some cases.
  • A few monoenergetic reconstruction series improved the image quality for one of the DECT scanners, but it was only improved for some of the criteria.
  • Monoenergetic reconstructions resulted in worse image quality for the majority of the criteria for the other DECT scanner.

On the last point, the researchers concluded that, in general, monoenergetic reconstructions are insufficient for reducing metal artefacts.

In their study report, the authors summarize the scan parameters for each CT machine. The journal posted the paper online for its subscribers earlier this year and is now offering it to all for free.