Method for determining slice sensitivity profile of iterative reconstruction CT images using low-contrast sphere...

Area of Science:

• Medical Imaging
• Radiological Physics

Background:

• Accurate measurement of the slice sensitivity profile (SSP) is crucial for assessing computed tomography (CT) image quality.
• Iterative reconstruction (IR) algorithms in CT present challenges for traditional SSP measurement methods.

Purpose of the Study:

• To introduce and validate a novel method for measuring the SSP in CT images reconstructed with IR algorithms.
• To compare the proposed method with existing techniques and assess its accuracy across different IR algorithms and scanners.

Main Methods:

• A phantom with a low-contrast sphere was scanned, and mean CT values were plotted to create a profile.
• A numerical object function was generated, and its profile was convolved with a modeled SSP (Gaussian and cosine product).
• Root mean square error (RMSE) minimization was used to optimize SSP model parameters, comparing the convolved profile with the measured profile.

Main Results:

• The proposed method showed good agreement with the standard coin method for filtered back projection (FBP) images.
• Accurate SSP measurements were achieved for four IR algorithms across two scanners, with RMSE < 0.7 HU.
• The method demonstrated good agreement between measured and reconstructed profiles for all IR algorithms.

Conclusions:

• The proposed method is effective and accurate for obtaining valid SSPs in CT images reconstructed using IR algorithms.
• This technique offers a reliable approach for quantitative image quality assessment in modern CT systems.
• The findings support the use of this method for routine quality control and algorithm development in CT imaging.