Multi-echo dipole inversion for magnetic susceptibility mapping.

PURPOSE: Reconstructing tissue magnetic susceptibility (QSM) from MRI phase data involves solving multiple consecutive ill-posed inverse problems such as phase unwrapping, background field removal, and field-to-source inversion. Multi-echo acquisitions present an additional challenge, as the magnetization field is typically computed from the multiple phase data prior to reconstructing the susceptibility map. Processing the multiple phase data introduces errors during the field estimation, violating assumptions of the subsequent inverse problems, manifesting as streaking artifacts in the susceptibility map. To address this challenge, we propose a multi-echo field-to-source forward model that forgoes the field estimation step. Moreover, we propose a fully general underestimation correction step to recover susceptibility sources that were regularized away during the field-to-source inversion. METHODS: The multi-echo forward model and correction step were validated on the QSM Challenge 2.0 datasets and compared to the standard single field-to-source model in in vivo human brains using different types of deconvolution algorithms. RESULTS: On the QSM Challenge 2.0 datasets the multi-echo forward model and correction step attain state-of-the-art results on all metrics by a wide margin. Experiments in in vivo brains show that the multi-echo model is in agreement with the single field-to-source model and that the proposed forward model and correction step can be used with any available dipole inversion method. CONCLUSION: A multi-echo field-to-source forward model forgoes the need to fit multi-echo phase data and achieves state-of-the-art results on the QSM Challenge 2.0 data. Underestimated low-frequency susceptibility distributions can be partially recovered using a correction step.

Brain Imaging Abnormalities in Mixed Alzheimer's and Subcortical Vascular Dementia.

BACKGROUND: A large proportion of Alzheimer's disease (AD) patients have coexisting subcortical vascular dementia (SVaD), a condition referred to as mixed dementia (MixD). Brain imaging features of MixD presumably include those of cerebrovascular disease and AD pathology, but are difficult to characterize due to their heterogeneity. OBJECTIVE: To perform an exploratory analysis of conventional and non-conventional structural magnetic resonance imaging (MRI) abnormalities in MixD and to compare them to those observed in AD and SVaD. METHODS: We conducted a cross-sectional, region-of-interest-based analysis of 1) hyperintense white-matter signal abnormalities (WMSA) on T2-FLAIR and hypointense WMSA on T1-weighted MRI; 2) diffusion tensor imaging; 3) quantitative susceptibility mapping; and 4) effective transverse relaxation rate (R2*) in N = 17 participants (AD:5, SVaD:5, MixD:7). General linear model was used to explore group differences in these brain imaging measures. RESULTS: Model findings suggested imaging characteristics specific to our MixD group, including 1) higher burden of WMSAs on T1-weighted MRI (versus both AD and SVaD); 2) frontal lobar preponderance of WMSAs on both T2-FLAIR and T1-weighted MRI; 3) higher fractional anisotropy values within normal-appear white-matter tissues (versus SVaD, but not AD); and 4) lower R2* values within the T2-FLAIR WMSA areas (versus both AD and SVaD). CONCLUSION: These findings suggest a preliminary picture of the location and type of brain imaging characteristics associated with MixD. Future imaging studies may employ region-specific hypotheses to distinguish MixD more rigorously from AD or SVaD.

The Effects of Wearing a 3-Ply or KN95 Face Mask on Cerebral Blood Flow and Oxygenation.

BACKGROUND: The SARS-CoV-2 virus has impacted life in many ways, one change being the use of face masks. Their effect on MRI-based measurements of cerebral oxygen levels with quantitative susceptibility mapping (QSM) and cerebral blood flow (CBF) is not known. PURPOSE: This study investigated whether wearing a face mask leads to changes in CBF and cerebral venous oxygen saturation measured with MRI. STUDY TYPE: Repeated-measures cohort study. POPULATION: A total of 16 healthy volunteers (eight male, eight female; 22-36 years) were recruited for the 3-ply study. Ten of the 16 participants (five male, five female; 23-36 years) took part in the KN95 study. FIELD STRENGTH/SEQUENCE: A 3 T, single-delay 3D gradient-and spin-echo pseudo-continuous arterial spin labeling (pCASL) scan for CBF quantification, and gradient-echo for QSM and oxygenation quantification. ASSESSMENT: Gray matter CBF and magnetic susceptibility were assessed by masking the pCASL CBF map and the QSM map to the T1 -weighted gray matter tissue segmentation. Venous oxygenation was determined from venous segmentation of QSM maximum intensity projections. STATISTICAL TESTS: Paired Student's t-tests and Cohen's d effect sizes were used to compare the face mask and no face mask scans for gray matter CBF, gray matter magnetic susceptibility, and cerebral venous oxygen saturation. Standard t-tests were used to assess whether the order of scanning with and without a mask had any impact. A statistical cut off of P < 0.05 was used. RESULTS: The 3-ply masks increased gray matter CBF from an average of 43.99 mL/(100 g*min) to 46.81 mL/(100 g*min). There were no significant changes in gray matter magnetic susceptibility (P = 0.07), or cerebral venous oxygen saturation (P = 0.36) for the 3-ply data set. The KN95 masks data set showed no statistically significant changes in gray matter CBF (P = 0.52) and magnetic susceptibility (P = 0.97), or cerebral venous oxygen saturation (P = 0.93). DATA CONCLUSION: The changes in blood flow and oxygenation due to face masks are small. Only CBF increased significantly due to wearing a 3-ply mask. EVIDENCE LEVEL: 2 TECHNICAL EFFICACY: Stage 3.

Recovering SWI-filtered phase data using deep learning.

PURPOSE: To develop a deep neural network to recover filtered phase from clinical MR phase images to enable the computation of QSMs. METHODS: Eighteen deep learning networks were trained to recover combinations of 13 SWI phase-filtering pipelines. SWI-filtered data were computed offline from five multiorientation, multiecho MRI scans yielding 132 3D volumes (118/7/7 training/validation/testing). Two experiments were conducted to show the efficacy of the networks. First, using QSM processing, local fields were computed from the raw phase and subsequently filtered using the SWI-filtering pipelines. The networks were then trained to invert the filtering operation. Second, the trained networks were fine-tuned to recover unfiltered local fields from filtered local fields computed by applying QSM processing to the SWI-filtered phase. Susceptibility maps were computed from the recovered fields and compared with gold standard multiple orientation sampling reconstructions. RESULTS: Susceptibility maps computed from the raw phase using standard QSM processing have a normalized root mean square error (NRMSE) of 0.732 ± 0.095. Susceptibility maps computed from the recovered phase obtained NRMSEs of 0.725 ± 0.095. The network trained using all 13 processing methods generalized well, obtaining NRMSEs of 0.725 ± 0.89 on filters it has not seen, while matching the reconstruction accuracy of networks trained to recover a single filter. CONCLUSION: It is feasible to recover SWI-filtered phase using deep learning. QSM can be computed from the recovered phase from SWI acquisition with comparable accuracy to standard QSM processing.

The influence of iron oxidation state on quantitative MRI parameters in post mortem human brain.

A variety of Magnetic Resonance Imaging (MRI) techniques are known to be sensitive to brain iron content. In principle, iron sensitive MRI techniques are based on local magnetic field variations caused by iron particles in tissue. The purpose of this study was to investigate the sensitivity of MR relaxation and magnetization transfer parameters to changes in iron oxidation state compared to changes in iron concentration. Therefore, quantitative MRI parameters including R1, R2, R2∗, quantitative susceptibility maps (QSM) and magnetization transfer ratio (MTR) of post mortem human brain tissue were acquired prior and after chemical iron reduction to change the iron oxidation state and chemical iron extraction to decrease the total iron concentration. All assessed parameters were shown to be sensitive to changes in iron concentration whereas only R2, R2∗ and QSM were also sensitive to changes in iron oxidation state. Mass spectrometry confirmed that iron accumulated in the extraction solution but not in the reduction solution. R2∗ and QSM are often used as markers for iron content. Changes in these parameters do not necessarily reflect variations in iron content but may also be a result of changes in the iron's oxygenation state from ferric towards more ferrous iron or vice versa.

DECAES - DEcomposition and Component Analysis of Exponential Signals.

Combinations of multiple exponentially decaying signals are found across many disciplines of science. Decomposition of these multi-exponential signals into their individual components provides insight into the various contributors to the signal. Magnetic resonance images, for instance, can be acquired with multiple gradient or spin echoes to provide voxel by voxel multi-exponential T2* or T2 decays, respectively. With their millions of voxels, these images make the task of decomposition into individual exponentials computationally challenging. Current implementations take several hours, which is prohibitively long in many settings, such as on-scanner calculation for clinical applications. Here, we present a fast approach for the decomposition of multi-exponential signals. The method is applied to multi echo spin echo MRI scans and computes myelin water maps of the whole brain in under 2min, and luminal water maps of the prostate in under 1min.

Myelin water imaging and R2* mapping in neonates: Investigating R2* dependence on myelin and fibre orientation in whole brain white matter.

R2* relaxation provides a semiquantitative method of detecting myelin, iron and white matter fibre orientation angles. Compared with standard histogram-based analyses, angle-resolved analysis of R2* has previously been shown to substantially improve the detection of subtle differences in the brain between healthy siblings of subjects with multiple sclerosis and unrelated healthy controls. Neonates, who are born with very little myelin and iron, and an underdeveloped connectome, provide researchers with an opportunity to investigate whether R2* is intimately linked with fibre-angle or myelin content as it is in adults, which may in future studies be explored as a potential white matter developmental biomarker. Five healthy adult volunteers (mean age [±SD] = 31.2 [±8.3] years; three males) were recruited from Vancouver, Canada. Eight term neonates (mean age = 38.6 ± 1.2 weeks; five males) were recruited from the Children's Hospital of Chongqing Medical University neonatal ward. All subjects were scanned on identical 3 T Philips Achieva scanners equipped with an eight-channel SENSE head coil and underwent a multiecho gradient echo scan, a 32-direction DTI scan and a myelin water imaging scan. For both neonates and adults, bin-averaged R2* variation across the brain's white matter was found to be best explained by fibre orientation. For adults, this represented a difference in R2* values of 3.5 Hz from parallel to perpendicular fibres with respect to the main magnetic field. In neonates, the fibre orientation dependency displayed a cosine wave shape, with a small R2* range of 0.4 Hz. This minor relationship in neonates provides further evidence for the key role myelin probably plays in creating this fibre orientation dependence later in life, but suggests limited clinical application in newborn populations. Future studies should investigate fibre-orientation dependency in infants in the first 5 years, when substantial myelin development occurs.

Rapid solution of the Bloch-Torrey equation in anisotropic tissue: Application to dynamic susceptibility contrast MRI of cerebral white matter.

Blood vessel related magnetic resonance imaging (MRI) contrast provides a window into the brain's metabolism and function. Here, we show that the spin echo dynamic susceptibility contrast (DSC) MRI signal of the brain's white matter (WM) strongly depends on the angle between WM tracts and the main magnetic field. The apparent cerebral blood flow and volume are 20% larger in fibres perpendicular to the main magnetic field compared to parallel fibres. We present a rapid numerical framework for the solution of the Bloch-Torrey equation that allows us to explore the isotropic and anisotropic components of the vascular tree. By fitting the simulated spin echo DSC signal to the measured data, we show that half of the WM vascular volume is comprised of vessels running in parallel with WM fibre tracts. The WM blood volume corresponding to the best fit to the experimental data was 2.82%, which is close to the PET gold standard of 2.6%.

Pathological Insights From Quantitative Susceptibility Mapping and Diffusion Tensor Imaging in Ice Hockey Players Pre and Post-concussion.

Myelin sensitive MRI techniques, such as diffusion tensor imaging and myelin water imaging, have previously been used to reveal changes in myelin after sports-related concussions. What is not clear from these studies, however, is how myelin is affected: whether it becomes degraded and possibly removed, or whether the myelin sheath loosens and becomes "decompacted". Previously, our team revealed myelin specific changes in ice hockey players 2 weeks post-concussion using myelin water imaging. In that study, 45 subjects underwent a pre-season baseline scan, 11 of which sustained a concussion during play and received follow-up scans: eight were scanned within 3 days, 10 were scanned at 14 days, and nine were scanned at 60 days. In the current retrospective analysis, we used quantitative susceptibility mapping, along with the diffusion tensor imaging measures axial diffusivity and radial diffusivity, to investigate this myelin disruption. If sports-related concussive hits lead to myelin fragmentation in regions of lowered MWF, this should result in a measurable increase in magnetic susceptibility, due to the anisotropic myelin fragmenting into isotropic myelin debris, and the diamagnetic myelin tissue being removed, while no such changes should be expected if the myelin sheath simply loosens and becomes decompacted. An increase in radial diffusivity would likewise reveal myelin fragmentation, as myelin sheaths block water diffusion out of the axon, with little to no changes expected for myelin sheath loosening. Statistical analysis of the same voxels-of-interest that were found to have reduced myelin water fraction 2 weeks post-concussion, revealed no statistically significant changes in magnetic susceptibility, axial diffusivity, or radial diffusivity at any time-point post-concussion. This suggests that myelin water fraction changes are likely due to a loosening of the myelin sheath structure, as opposed to fragmentation and removal of myelin debris.

Quantitative susceptibility mapping: Report from the 2016 reconstruction challenge.

PURPOSE: The aim of the 2016 quantitative susceptibility mapping (QSM) reconstruction challenge was to test the ability of various QSM algorithms to recover the underlying susceptibility from phase data faithfully. METHODS: Gradient-echo images of a healthy volunteer acquired at 3T in a single orientation with 1.06 mm isotropic resolution. A reference susceptibility map was provided, which was computed using the susceptibility tensor imaging algorithm on data acquired at 12 head orientations. Susceptibility maps calculated from the single orientation data were compared against the reference susceptibility map. Deviations were quantified using the following metrics: root mean squared error (RMSE), structure similarity index (SSIM), high-frequency error norm (HFEN), and the error in selected white and gray matter regions. RESULTS: Twenty-seven submissions were evaluated. Most of the best scoring approaches estimated the spatial frequency content in the ill-conditioned domain of the dipole kernel using compressed sensing strategies. The top 10 maps in each category had similar error metrics but substantially different visual appearance. CONCLUSION: Because QSM algorithms were optimized to minimize error metrics, the resulting susceptibility maps suffered from over-smoothing and conspicuity loss in fine features such as vessels. As such, the challenge highlighted the need for better numerical image quality criteria. Magn Reson Med 79:1661-1673, 2018. © 2017 International Society for Magnetic Resonance in Medicine.

Rapid two-step dipole inversion for susceptibility mapping with sparsity priors.

Quantitative susceptibility mapping (QSM) is a post-processing technique of gradient echo phase data that attempts to map the spatial distribution of local tissue magnetic susceptibilities. To obtain these maps, an ill-posed field-to-source inverse problem must be solved to remove non-local magnetic field perturbations. Current state-of-the-art algorithms which aim to solve the dipole inversion problem are plagued by the trade-off between reconstruction speed and accuracy. A two-step dipole inversion algorithm is proposed to bridge this gap. Our approach first addresses the well-conditioned k-space region, which is reconstructed using a Krylov subspace solver. Then the ill-conditioned k-space region is reconstructed by solving a constrained l1-minimization problem. The proposed pipeline does not incorporate a priori information, but utilizes sparsity constraints in the second step. We compared our method to well-established QSM algorithms with respect to COSMOS in in vivo volunteer datasets. Compared to MEDI and HEIDI the proposed algorithm produces susceptibility maps with a lower root-mean-square error and a higher coefficient of determination, with respect to COSMOS, while being 50 times faster. Our two-step dipole inversion algorithm without a priori information yields improved QSM reconstruction quality at reduced computation times compared to current state-of-the-art methods.

Susceptibility-sensitive MRI of multiple sclerosis lesions and the impact of normal-appearing white matter changes.

Susceptibility-sensitive magnetic resonance imaging (MRI) has gained importance in multiple sclerosis (MS) research because of its versatility, high resolution and excellent sensitivity to changes in tissue structure and composition. In particular, mapping of the resonance frequency of the MR signal and quantitative susceptibility mapping (QSM) have been explored for the description of MS lesions. Many current studies utilizing these techniques attribute increases in the MR frequency or QSM to elevated tissue iron content, in addition to myelin loss. However, this common interpretation is inconsistent with recent histopathological studies. Here, we investigate the nature of MR frequency shifts related to MS lesions by comparing post-mortem MRI data with histology, and contrast them with numerical simulations of the MR signal. We demonstrate that iron accumulation is not the driving source of the MR frequency or QSM image contrast in our sample; rather, most chronic MS lesions are characterized by advanced loss of both myelin and iron. Moreover, our results suggest that the appearance of MS lesions on MR frequency maps and QSM depends on changes in the non-lesional white matter surrounding the lesions. Understanding and accounting for these changes is essential for the quantitative interpretation of MR frequency or QSM data in white matter.