eIRIS: Eigen-analysis approach for improved spine multi-shot diffusion MRI.

Area of Science:

• Magnetic Resonance Imaging
• Medical Physics
• Biomedical Engineering

Background:

• Multi-shot diffusion MRI techniques like IRIS (image-space sampling function) correct motion-induced phase variations using navigator-echo phase maps.
• Geometric distortions between navigator-echo and image-echo, due to differing bandwidths, necessitate B0 map correction in standard IRIS.
• This B0 map requirement adds scan time and complexity to diffusion-weighted imaging.

Purpose of the Study:

• To introduce an enhanced IRIS (eIRIS) method that removes the need for B0 maps in correcting inter-shot phase variations.
• To evaluate the performance of eIRIS against standard IRIS and GRAPPA with a compact kernel (GRAPPA-CK) for diffusion MRI.
• To assess the potential of eIRIS for improving diffusion-weighted image quality in spinal cord applications.

Main Methods:

• eIRIS treats individual shots as virtual coils, employing an eigen-analysis approach to estimate coil sensitivity maps.
• This eigen-analysis is inherently insensitive to geometric mismatches between navigator and image echoes.
• Image reconstruction is performed using the SENSE framework, integrating the estimated sensitivity maps.

Main Results:

• Images reconstructed with eIRIS exhibit no noticeable artifacts, unlike IRIS without B0 correction.
• eIRIS performance is comparable to IRIS with B0 correction and GRAPPA-CK.
• eIRIS demonstrates slightly superior performance over GRAPPA-CK in terms of normalized root-mean-square error and signal-to-noise ratio.

Conclusions:

• The eIRIS method successfully eliminates the requirement for B0 maps in multi-shot diffusion MRI reconstruction.
• eIRIS provides high-quality diffusion-weighted images with reduced artifacts and improved quantitative metrics.
• This technique holds significant potential for advancing spinal cord diffusion MRI research and clinical diagnosis.