Validating the accuracy of multispectral metal artifact suppressed diffusion-weighted imaging

Area of Science:

• Musculoskeletal MRI
• Medical Imaging Physics

Background:

• Diffusion-weighted imaging (DWI) quantifies water diffusion (ADC) for microstructural analysis in tissues.
• Conventional DWI in total hip arthroplasty (THA) is challenged by metallic artifacts from echo-planar imaging (EPI).
• Multispectral imaging (MSI) techniques like MAVRIC reduce artifacts, and PROPELLER FSE with DWI-MAVRIC aims to further mitigate distortion and dropout.

Purpose of the Study:

• To assess the quantitative accuracy of DWI-MAVRIC compared to standard diffusion sequences.
• To evaluate the repeatability and reliability of DWI-MAVRIC in musculoskeletal imaging applications.

Main Methods:

• Diffusion phantom scans using DWI-EPI and DWI-MAVRIC on two 1.5T MRI scanners.
• Acquisitions performed in axial and coronal planes with varying spatial offsets to assess positioning effects.
• Intraday, interday, and inter-scanner repeatability evaluated; gradient nonlinearity correction (GNC) impact assessed.

Main Results:

• DWI-MAVRIC demonstrated excellent correlation (r=1.00) but slightly reduced linearity (slope=0.91-0.94) versus DWI-EPI (slope=0.99-1.01).
• Minor ADC bias (>5%) observed in DWI-MAVRIC at lowest ADC values; no variation with DWI-MAVRIC parameters.
• DWI-MAVRIC showed inferior repeatability (intraday: 13.3, interday: 44.7, inter-scanner: 110 μm²/s) compared to DWI-EPI (3.18, 19.2, 20.2 μm²/s).

Conclusions:

• DWI-MAVRIC offers good linearity and acceptable intra- and interday repeatability for musculoskeletal applications.
• The technique shows promise for artifact reduction in THA imaging, despite lower repeatability than conventional DWI.