CT metal artifact reduction method correcting for beam hardening and missing projections

Area of Science:

• Medical Imaging
• Radiology
• Computational Science

Background:

• Metal artifacts in computed tomography (CT) scans degrade image quality.
• Artifacts arise from beam hardening and photon starvation caused by high-density implant materials.
• Existing methods struggle with diverse implant materials, necessitating improved solutions.

Purpose of the Study:

• To develop and validate a novel CT metal artifact reduction (MAR) method.
• To address artifact reduction for a wide range of clinical implant materials.
• To compare the proposed MAR method against conventional techniques.

Main Methods:

• Developed a novel physics-based correction algorithm for low-Z implants (e.g., titanium) to reduce beam hardening.
• Employed regularized iterative reconstruction for high-Z implants (e.g., dental fillings, gold) by treating projections as missing data.
• Combined both algorithms for scans with mixed implant materials.
• Implemented a conventional projection interpolation method for comparison.
• Conducted a blinded, randomized evaluation with ten radiation oncologists ranking scan quality.

Main Results:

• The proposed MAR method was ranked superior to original reconstruction and conventional projection interpolation for both low-Z and high-Z implants.
• 90% of reviewers ranked the method best for low-Z implants.
• All reviewers ranked the method first for high-Z implants.
• Statistically significant improvements were observed compared to original reconstruction and projection interpolation (p < 0.0001 for high-Z).

Conclusions:

• A combined approach using material-specific algorithms effectively reduces CT metal artifacts.
• The developed method offers significant improvements in image quality for patients with various implants.
• This technique holds promise for enhancing diagnostic accuracy in CT imaging.