Metal artifact algorithm reduces radiation exposure, improves accuracy

A study published in the March issue of the American Journal of Roentgenology found an advanced metal artifact reduction (MAR) algorithm improved lesion detection accuracy near hardware while also reducing radiation exposure.

Traditionally, filtered back projection (FBP) is used to reduce artifact distortion and improve visualization of structures near metallic hardware, but the method typically exposes patients to relatively high doses of radiation.

MAR algorithms have gained acceptance and become further available for clinical use in producing quality images with lower doses, but there is still little research on the method.

The published study was based on presentations at the Radiological Society of North America 2016 annual meeting and the Society of Skeletal Radiology 2016 annual meeting.

The team created a CT phantom simulated arthroplasty, which they scanned using clinical dose, low-dose and high-dose methods. The images were analyzed using MAR, advanced modeled iterative reconstruction (ADMIRE) and FBP.

Results were as follows:

  • Accuracy using MAR was significantly higher compared to FBP at all exposures.
  • Sensitivity was higher for MAR than FBP at all exposures.
  • Specificity was high for both techniques at all exposures with no significant difference.
  • Accuracy of low dose MAR was higher than standard dose and high dose FBP.
  • MAR was significantly more sensitive than FBP in detecting smaller lesions and those near the high streak artifact.

“[W]e conclude that the results of this study suggest that using advanced MAR techniques can lead to significant reductions in radiation exposure without compromising the ability to detect lesions when imaging patients with arthroplasties,” wrote corresponding author Naveen Subhas, with the Department of Musculoskeletal Radiology at the Cleveland Clinic, and colleagues.