Fast Robust Dejitter and Interslice Discontinuity Removal in MRI Phase Acquisitions: Application to Magnetic Resonance Elastography

Abstract

MRI phase contrast imaging methods that assemble slice-wise acquisitions into volumes can contain interslice phase discontinuities (IPDs) over the course of the scan from sources, including unavoidable physiological activity. In magnetic resonance elastography (MRE), this can alter wavelength and tissue stiffness estimates, invalidating the analysis. We first model this behavior as jitter along the z-axis of the phase of 3D complex-valued wave volumes. A two-step image processing pipeline is then proposed that removes IPDs. First, constant slicewise phase shift is removed with a novel, non-convex dejittering algorithm. Then, regional physiological noise artifacts are removed with novel filtering of 3D wavelet coefficients. Calibration of two pipeline coefficients, the dejitter parameter α and the wavelet band high-pass coefficient ω c , was first performed on a finite-element method brain phantom. A comparative investigation was then performed, on a cohort of 48 brain acquisitions, of four approaches to IPDs: 1) the proposed method; 2) a "control" condition of neglect of IPDs; 3) an anisotropic wavelet-based method; and 4) a method of in-plane (2D) processing. The present method showed medians of |G * | = 1873 Pa for a multifrequency wave inversion centered at 40 Hz which was within 6% of methods 3) and 4), while neglect produced |G * | estimates a mean of 17% lower. The proposed method reduced the value range of the cohort against methods 3) and 4) by 29% and 31%, respectively. Such reduction in variance enhances the ability of brain MRE to predict subtler physiological changes. Our theoretical approach further enables more

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

“…Therefore, we focused, in theory, on unsmoothed raw wave data assuming white Gaussian noise and analyzed the effect of noise suppression by FDO kernel size. Clearly, noise, as defined as unwanted signal in MRE, 15 can have different sources including compression waves 30 or slice jittering 31 making noise anisotropic and non‐Gaussian. Furthermore, SNR given by in this study is normally addressed by other methods such as octahedral shear strain, 24 shear strain invariant noise, 32 or wavelet‐based methods, 29 all of which provide values different from those obtained by Equation ().…”

Elastography Validity Criteria Definition Using Numerical Simulations and MR Acquisitions on a Low-Cost Structured Phantom

“…The displacement fields are then fed into the MRE reconstruction pipeline, which solves the inverse problem and retrieves the mechanical properties for comparison with groundtruth values. Simulations are generally based on analytical formulations [5,14,17,21] or finite element methods (FEM), either with custom-developed tools [22][23][24] or dedicated commercial softwares [20,[24][25][26][27][28][29][30]. Simulations are very useful to assess the error linked to reconstruction algorithms, however they do not suffice to reflect all potential limitations of an MRE experiment (transducer, B 0 and B 1 inhomogeneities, SNR, motion sensitivity, susceptibility issues, and heterogeneity of the material at very small scales).…”

TURBINE‐MRE: A 3D hybrid radial‐Cartesian EPI acquisition for MR elastography

“…This artifact manifests as spatially low‐order phase terms, which, if not corrected before MRE processing, will alter the apparent wavelength along the slice direction, leading to a severe error in stiffness estimation. This artifact may be corrected by high‐pass filtering the 2D slices of complex wave data, as done in this and previous studies, 21,22 or by alternative processing techniques 27,28 . In 3D TURBINE acquisition, the phase errors are present between EPI blades in k‐space instead of 2D slices (image space) in a dispersed and incoherent manner; therefore, they are not as visually apparent and have less impact on the stiffness estimation.…”

“…This artifact may be corrected by high-pass filtering the 2D slices of complex wave data, as done in this and previous studies, 21,22 or by alternative processing techniques. 27,28 In 3D TURBINE acquisition, the phase errors are present between EPI blades in k-space instead of 2D slices (image space) in a dispersed and incoherent manner; therefore, they are not as visually apparent and have less impact on the stiffness estimation. We did not attempt to correct the phase error at this stage because it did not visually appear to corrupt the TURBINE images.…”