An analytical solution to the dispersion-by-inversion problem in magnetic resonance elastography.

Area of Science:

• Biomedical Engineering
• Medical Imaging
• Rheology

Background:

• Magnetic Resonance Elastography (MRE) quantifies soft tissue stiffness using shear vibrations.
• Current MRE methods often employ inversion-based reconstructions with finite difference operators (FDOs).
• FDOs can introduce biased frequency dispersion in stiffness estimates.

Purpose of the Study:

• To analytically, numerically, and experimentally investigate the impact of FDOs on MRE stiffness estimates.
• To identify the sources of bias in FDO-based MRE reconstructions.
• To provide a method for minimizing inversion artifacts in MRE postprocessing.

Main Methods:

• Analysis of spatial harmonic response to shear vibrations in MRE.
• Application of first- and second-order finite difference operators (FDOs).
• Analytical derivation and experimental validation of FDO dispersion functions.

Main Results:

• FDOs lead to underestimation at low frequencies (high spatial support) and overestimation at high frequencies (low spatial support) due to noise and discretization.
• Second-order FDOs are more sensitive to noise than first-order FDOs.
• Optimized parameter A (wavelength/pixel size/stencil pixels) minimizes FDO artifacts, with specific SNR-dependent ranges provided for first- and second-FDOs.

Conclusions:

• This study offers an analytical solution to the long-standing problem of FDO-induced artifacts in MRE.
• The findings contribute to reducing inversion artifacts, improving the accuracy of MRE stiffness measurements.
• Optimizing FDO parameters based on SNR and wave properties is crucial for reliable MRE postprocessing.