Polyenergetic known-component CT reconstruction with unknown material compositions and unknown x-ray spectra

Area of Science:

• Medical Imaging
• Computational Imaging
• Image Reconstruction

Background:

• Metal artifacts significantly degrade computed tomography (CT) image quality, especially in interventional procedures with surgical tools or implants.
• Existing methods like known-component reconstruction (KCR) require detailed physical and material properties of metal components and CT scanners, posing a practical challenge.

Purpose of the Study:

• To develop a novel CT image reconstruction method that reduces metal artifacts without prior knowledge of component material or scanner energy-dependent characteristics.
• To improve image quality in regions of interest near metal implants.

Main Methods:

• Proposed a modified KCR approach termed Poly-KCR, utilizing a mixed forward model.
• Employed a polyenergetic model for known metal components and a monoenergetic model for background anatomy.
• Jointly estimated a spectral transfer function for components and background attenuation values.

Main Results:

• Poly-KCR effectively reduces metal artifacts, outperforming traditional and model-based iterative reconstruction methods.
• Demonstrated improved performance in both simulated data and physical data, including an implanted cadaver sample.
• Eliminated the need for prior knowledge of component material composition and scanner energy-dependent characterization.

Conclusions:

• Poly-KCR offers a robust solution for metal artifact reduction in CT imaging.
• This approach simplifies the reconstruction process by removing the requirement for extensive system and component characterization.
• The method shows significant potential for enhancing diagnostic accuracy in interventional CT imaging.