How Different Iterative and Filtered Back Projection Kernels Affect Computed Tomography Numbers and Low Contrast...

Area of Science:

• Medical Imaging Physics
• Radiological Technology

Background:

• Computed tomography (CT) image quality is influenced by reconstruction algorithms.
• Filtered back projection (FBP) and iterative reconstruction (IR) are common CT reconstruction techniques.
• Understanding the impact of different kernels and reconstruction strengths on CT numbers and detectability is crucial for accurate diagnosis.

Purpose of the Study:

• To evaluate the effect of various iterative and filtered back projection kernels on computed tomography (CT) numbers.
• To assess the impact of these reconstruction techniques on low contrast detectability.

Main Methods:

• Five distinct scans were acquired using a Catphan 600 phantom at six different tube potentials.
• Scans were reconstructed using all available filtered back projection (FBP) body kernels and iterative reconstruction (IR) techniques.
• Consistent dose levels and imaging parameters were maintained across all acquisitions.

Main Results:

• CT numbers and contrast-to-noise ratios (CNR) were quantified for each reconstruction method.
• The study analyzed how kernel choice and the strength of iterative reconstruction influenced CT numbers and CNR.
• Significant variations in CT numbers were observed between kernels with similar sharpness for specific densities.

Conclusions:

• Iterative reconstruction generally improved contrast-to-noise ratio (CNR) in computed tomography (CT) imaging.
• However, in specific scenarios, iterative reconstruction led to a decrease in CNR.
• CT number variations can be substantial across different kernels, even those with comparable sharpness, impacting image interpretation.