Cardiac diffusion tensor imaging based on compressed sensing using joint sparsity and low-rank approximation.

Area of Science:

• Medical Imaging
• Biophysics
• Magnetic Resonance Imaging

Background:

• Diffusion Tensor Magnetic Resonance (DTMR) imaging and Diffusion Tensor Imaging (DTI) are crucial for noninvasively assessing biological tissue structure.
• Current DTI methods face limitations due to prolonged acquisition times, hindering clinical utility.

Purpose of the Study:

• To develop an advanced Compressed Sensing (CS) reconstruction technique for cardiac DTMR imaging.
• To address the challenge of long acquisition times in DTI by reconstructing images from undersampled k-space data.

Main Methods:

• A novel CS reconstruction method utilizing joint sparsity and rank deficiency was proposed.
• Diffusion-weighted images were arranged into a matrix, exploiting its row sparsity and low-rank properties.
• A first-order fast method was employed to solve the constrained optimization problem.

Main Results:

• The proposed CS method demonstrated lower reconstruction errors for DTI indices (Fractional Anisotropicity - FA, Mean Diffusivities - MD).
• Performance was validated on both simulated and real human cardiac DTMR images.
• The approach outperformed existing CS-DTMR reconstruction techniques in accuracy.

Conclusions:

• The developed CS method effectively reconstructs cardiac DTMR images from undersampled data.
• This technique offers a promising solution for reducing DTI acquisition times while maintaining or improving image quality and metric accuracy.
• The findings support the clinical applicability of faster DTI acquisition protocols.