Single-slice rebinning reconstruction in spiral cone-beam computed tomography

Area of Science:

• Medical Imaging
• Computed Tomography
• Image Reconstruction

Background:

• Multislice computed tomography (CT) utilizes approximate cone-beam algorithms for spiral data reconstruction.
• Existing methods are often limited to small cone angles, necessitating advancements for area detector CT.

Purpose of the Study:

• To identify and evaluate practical, efficient approximate cone-beam CT algorithms for medical applications.
• To extend the utility of these algorithms for medium cone angles typical in area detector CT.
• To assess algorithm performance concerning image artifacts, spatial resolution, contrast resolution, and noise.

Main Methods:

• Investigated two approximate single-slice rebinning algorithms: multirow Fourier reconstruction (MFR) and an extended advanced single-slice rebinning (ASSR) with z-filtering.
• Formulated both MFR and ASSR within a z-filtering framework using optimized spiral interpolation.
• Identified X-ray samples approximating a virtual reconstruction plane for each view.

Main Results:

• The ASSR method, employing tilted reconstruction planes, proved practical and efficient.
• ASSR achieved image quality comparable to single-row systems, even with a 46-row detector and 64 mm table feed.
• Both algorithms accommodate various table feeds up to the detector's maximum height, utilizing all sampled detector data via z-filtering.

Conclusions:

• The ASSR algorithm is a suitable and effective method for cone-beam CT reconstruction with area detectors.
• This approach enhances image quality and data utilization in spiral CT, offering a practical solution for medical imaging.
• The z-filtering framework allows comprehensive use of detector data, improving reconstruction efficiency.