Sign in Subscribe
Algorithms

Metal Artifact Reduction in Computed Tomography Imaging

Strategic Insights

— 1 min read

Researchers from GE HealthCare Technology & Innovation Center have developed a new framework for simulating metal artifacts in computed tomography (CT) imaging, which could improve the accuracy of metal artifact reduction (MAR) algorithms. The new framework, called a hybrid training database and evaluation benchmark, is designed to provide a comprehensive evaluation of MAR methods for X-ray CT imaging.

Key Takeaways:

  • The lack of artifact-free ground truth data and paired training datasets with and without artifacts hinders the development of MAR algorithms.
  • The researchers propose a simulation-based approach to generate a large training database for deep-learning based MAR algorithms.
  • The simulation tool, called CatSim, is capable of realistic simulation of metal artifacts on clinical CT data, with a mean CT number deviation of less than 2% compared to real data.
  • A comprehensive evaluation benchmark for MAR is defined, covering metrics such as CT number accuracy, noise, image sharpness, streak amplitude, structural integrity, and the effect on range in proton therapy.
  • The benchmark is applied to a numerical and a deep-learning based MAR algorithm, demonstrating the potential of the new framework.
  • The simulated metal scenarios cover a wide range of clinically relevant use cases, including small metal implants and large metal implants.
  • The simulation tools and benchmark are made publicly available for the development and evaluation of MAR algorithms.
  • The research is the first comprehensive evaluation benchmark covering a large number of clinically realistic metal artifact scenarios.

Statistics:

  • 14,000 metal scenarios were simulated in the head, thorax, and pelvis regions.
  • The simulation tool achieves a mean CT number deviation of less than 2% compared to real data.
  • The benchmark covers a range of clinical scenarios, including small metal implants (fiducial markers) and large metal implants (hip replacements).

Sources:

  • A hybrid training database and evaluation benchmark for assessing metal artifact reduction methods for X-ray CT imaging. Medical Physics, 2025; 52(10).
  • GE HealthCare Technology & Innovation Center.
  • NewsRx. Findings on Information Technology Described by Eri Haneda and Colleagues (A hybrid training database and evaluation benchmark for assessing metal artifact reduction methods for X-ray CT imaging). Information Technology Newsweekly. October 21, 2025; p 270.

Read more