Metal artifact reduction method based on single spectral CT (MARSS).

Area of Science:

• Medical Imaging
• Computed Tomography (CT)
• Image Reconstruction

Background:

• Metal artifacts in CT imaging significantly degrade image quality and can lead to misdiagnosis in patients with metal implants.
• Existing spectral CT-based metal artifact reduction (MAR) methods show promise but often require multi-spectral data, which is not feasible with conventional CT scanners.

Purpose of the Study:

• To introduce a novel MAR method utilizing single spectral CT (MARSS) data.
• To establish a constrained image reconstruction model incorporating prior information and discrete material solution spaces for improved MAR.

Main Methods:

• The MARSS method employs discrete tomography principles within a constrained reconstruction framework.
• Prior knowledge is integrated to discretize material solution spaces, reducing ill-posedness and simplifying the problem.
• An iterative algorithm alternates between reconstruction and discretization to ensure convergence and accurate representation of the scanned object.

Main Results:

• The MARSS method demonstrated significant reduction of metal artifacts, effectively restoring anatomical structures adjacent to implants.
• Crucially, the method avoided the introduction of new artifacts and structural distortions observed in comparative MAR techniques.

Conclusions:

• MARSS successfully achieves metal artifact reduction using single spectral CT data.
• Both subjective and quantitative evaluations confirm that MARSS substantially enhances image quality compared to existing methods.