K-Bayes reconstruction for perfusion MRI. I: concepts and application

Area of Science:

• Medical Imaging
• Neuroimaging
• Biophysics

Background:

• Magnetic resonance imaging (MRI) is crucial in science and medicine, but high-resolution applications are limited.
• Perfusion MRI offers vital blood flow data but suffers from low resolution and artifacts, hindering disease detection.
• Artifacts like partial volume effects and aliasing stem from limited k-space sampling and discrete Fourier transform (DFT) reconstruction.

Purpose of the Study:

• To develop and evaluate a novel Bayesian modeling procedure (K-Bayes) for reconstructing perfusion MRI data.
• To improve the resolution and reduce artifacts in perfusion MRI images.
• To enhance the diagnostic power of perfusion MRI for detecting subtle changes in brain perfusion.

Main Methods:

• Developed the K-Bayes approach, integrating a k-space signal process model with a Markov random field prior.
• Incorporated high-resolution segmented structural MRI information into the prior distribution.
• Validated K-Bayes using both simulation studies and in vivo human brain perfusion MRI data, comparing results against standard DFT reconstruction.

Main Results:

• K-Bayes reconstructed images showed significant qualitative and quantitative improvements over DFT.
• Demonstrated reduced bias and increased precision in K-Bayes reconstructed perfusion MRI data.
• Achieved greater effect sizes and higher resolution in K-Bayes images compared to DFT, validated in human brain scans.

Conclusions:

• The K-Bayes method offers a superior approach for perfusion MRI reconstruction.
• This technique effectively mitigates resolution limitations and artifacts inherent in standard MRI reconstruction.
• K-Bayes holds promise for advancing neuroimaging research and clinical diagnosis by providing more accurate perfusion information.