Accelerated model-based quantitative diffusion MRI: A feasibility study for musculoskeletal application.

Area of Science:

• Magnetic Resonance Imaging
• Diffusion Tensor Imaging
• Image Reconstruction

Background:

• Accelerated imaging techniques are crucial for reducing scan times in Diffusion Tensor Imaging (DTI).
• Model-based reconstruction offers potential for improving image quality and quantitative accuracy in accelerated MRI.

Purpose of the Study:

• To develop and validate a model-based reconstruction technique for diffusion quantification.
• To enable accelerated acquisition of multi-b-value 2D echo planar imaging (EPI) data.
• To assess the feasibility of acceleration factors greater than three in a clinical setting.

Main Methods:

• A model-based method minimizing a cost function comparing synthetic and measured k-space data.
• Incorporation of a total variation (TV) constraint for regularization.
• A non-random undersampling pattern with a fully sampled central k-space region for phase correction.
• Acceleration factors corresponding to the number of b-values were employed.

Main Results:

• Qualitative analysis showed preservation of high-frequency information and image quality (S0 and ADC maps).
• Quantitative analysis demonstrated significantly superior performance of the model-based technique compared to conventional SENSE acceleration.
• The technique enabled acceleration factors up to R=3.65 without compromising diffusion data fidelity.

Conclusions:

• Model-based reconstruction combined with a suitable undersampling pattern facilitates significant acceleration of 2D EPI.
• This approach allows for diffusion quantification with high fidelity at accelerated scan times.
• The technique holds promise for improving the clinical utility of diffusion MRI.