Search

RF transmit field inhomogeneities at 7T usually degrade the image quality in regions such as the cerebellum and brainstem, which has so far delayed its use to study cerebellar ataxias with higher fields. We present an optimized pipeline that maximizes the QSM contrast of nuclei in the cerebellum based on multi-echo MPM data. We show that averaging the individual susceptibility maps increases the signal-to-noise ratio (SNR) and reduces the group standard deviation. We introduced a weighted averaging based on the individual acquisition SNRs. Final maps presented high contrast of the dentate nucleus and a high delineation of its denticulated silhouette.

While iron is essential for normal brain functioning, elevated concentrations are commonly found in neurodegenerative diseases and are associated with impaired cognition and neurological deficits. Currently, only little is known about genetic and environmental factors that influence brain iron concentrations.Heritability and bivariate heritability of regional brain iron concentrations, assessed by R2* relaxometry at 3 Tesla MRI, were estimated with variance components models in 130 middle-aged to elderly participants of the Austrian Stroke Prevention Family Study.Heritability of R2* iron ranged from 0.46 to 0.82 in basal ganglia and from 0.65 to 0.76 in cortical lobes. Age and BMI explained up to 12% and 9% of the variance of R2* iron, while APOE ε4 carrier status, hypertension, diabetes, hypercholesterolemia, sex and smoking explained 5% or less. The genetic correlation of R2* iron among basal ganglionic nuclei and among cortical lobes ranged from 0.78 to 0.87 and from 0.65 to 0.97, respectively. R2* rates in basal ganglia and cortex were not genetically correlated.Regional brain iron concentrations are mainly driven by genetic factors while environmental factors contribute to a certain extent. Brain iron levels in the basal ganglia and cortex are controlled by distinct sets of genes.

Susceptibility maps are usually derived from local magnetic field estimations by minimizing a functional composed of a data consistency term and a regularization term. The data-consistency term measures the difference between the desired solution and the measured data using typically the L2-norm. It has been proposed to replace this L2-norm with the L1-norm, due to its robustness to outliers and reduction of streaking artifacts arising from highly noisy or strongly perturbed regions. However, in regions with high SNR, the L1-norm yields a suboptimal denoising performance. In this work, we present a hybrid data fidelity approach that uses the L1-norm and subsequently the L2-norm to exploit the strengths of both norms.We developed a hybrid data fidelity term approach for QSM (HD-QSM) based on linear susceptibility inversion methods, with total variation regularization. Each functional is solved with ADMM. The HD-QSM approach is a two-stage method that first finds a fast solution of the L1-norm functional and then uses this solution to initialize the L2-norm functional. In both norms we included spatially variable weights that improve the quality of the reconstructions.The HD-QSM approach produced good quantitative reconstructions in terms of structural definition, noise reduction, and avoiding streaking artifacts comparable with nonlinear methods, but with higher computational efficiency. Reconstructions performed with this method achieved first place at the lowest RMS error category in stage 1 of the 2019 QSM Reconstruction Challenge.The proposed method allows robust and accurate QSM reconstructions, obtaining superior performance to state-of-the-art methods.

Deep neural networks are increasingly used for neurological disease classification by MRI, but the networks9 decisions are not easily interpretable by humans. Heat mapping by deep Taylor decomposition revealed that (potentially misleading) image features even outside of the brain tissue are crucial for the classifier9s decision. We propose a regularization technique to train convolutional neural network (CNN) classifiers utilizing relevance-guided heat maps calculated online during training. The method was applied using T1-weighted MR images from 128 subjects with Alzheimer9s disease (mean age=71.9+-8.5 years) and 290 control subjects (mean age=71.3+-6.4 years). The developed relevance-guided framework achieves higher classification accuracies than conventional CNNs but more importantly, it relies on less but more relevant and physiological plausible voxels within brain tissue. Additionally, preprocessing effects from skull stripping and registration are mitigated, rendering this practically useful in deep learning neuroimaging studies. Understanding the decision mechanisms underlying CNNs, these results challenge the notion that unprocessed T1-weighted brain MR images in standard CNNs yield higher classification accuracy in Alzheimer9s disease than solely atrophy.

Purpose: The presence of dipole-inconsistent data due to substantial noise or artifacts causes streaking artifacts in quantitative susceptibility mapping (QSM) reconstructions. Often used Bayesian approaches rely on regularizers, which in turn yield reduced sharpness. To overcome this problem, we present a novel L1-norm data fidelity approach that is robust with respect to outliers, and therefore prevents streaking artifacts. Methods: QSM functionals are solved with linear and nonlinear L1-norm data fidelity terms using functional augmentation, and are compared with equivalent L2-norm methods. Algorithms were tested on synthetic data, with phase inconsistencies added to mimic lesions, QSM Challenge 2.0 data, and in vivo brain images with hemorrhages. Results: The nonlinear L1-norm-based approach achieved the best overall error metric scores and better streaking artifact suppression. Notably, L1-norm methods could reconstruct QSM images without using a brain mask, with similar regularization weights for different data fidelity weighting or masking setups. Conclusion: The proposed L1-approach provides a robust method to prevent streaking artifacts generated by dipole-inconsistent data, renders brain mask calculation unessential, and opens novel challenging clinical applications such asassessing brain hemorrhages and cortical layers.

Background/objective Higher white matter hyperintensity (WMH) load has been reported in Alzheimer's disease (AD) patients in different brain regions when compared to controls. We aimed to assess possible differences of WMH spatial distribution between AD patients and age-matched controls by means of lesion probability maps. Methods The present study included MRI scans of 130 probable AD patients with a mean age of 73.4±8.2 years from the Prospective Dementia Registry Austria Study and 130 age-matched healthy controls (HC) from the Austrian Stroke Prevention Family Study. Risk factors such as hypertension, diabetes mellitus, hypercholesterolemia, coronary artery disease, and smoking were assessed. Manually segmented FLAIR WMH masks were non-linearly registered to a template and voxel-based probability mapping was performed. Results There were no significant between-group differences in cardiovascular risk factors and WMH volume. AD patients showed a significantly higher likelihood of having WMH in a bilateral periventricular distribution than controls before and after correcting for age, sex, cardiovascular risk factors, and ventricular volume (p≤0.05; threshold-free cluster enhancement corrected). There was no significant association between the periventricular WMH volume and cognitive decline of AD patients. Conclusion In AD, WMH were preferentially found in a periventricular location but the volume of lesions was unrelated to cognitive decline in our study irrespective of lesion location.

Purpose To create a realistic in silico head phantom for the second QSM reconstruction challenge and for future evaluations of processing algorithms for QSM. Methods We created a digital whole-head tissue property phantom by segmenting and postprocessing high-resolution (0.64 mm isotropic), multiparametric MRI data acquired at 7 T from a healthy volunteer. We simulated the steady-state magnetization at 7 T using a Bloch simulator and mimicked a Cartesian sampling scheme through Fourier-based processing. Computer code for generating the phantom and performing the MR simulation was designed to facilitate flexible modifications of the phantom in the future, such as the inclusion of pathologies as well as the simulation of a wide range of acquisition protocols. Specifically, the following parameters and effects were implemented: TR and TE, voxel size, background fields, and RF phase biases. Diffusion-weighted imaging phantom data are provided, allowing future investigations of tissue-microstructure effects in phase and QSM algorithms. Results The brain part of the phantom featured realistic morphology with spatial variations in relaxation and susceptibility values similar to the in vivo setting. We demonstrated some of the phantom's properties, including the possibility of generating phase data with nonlinear evolution over TE due to partial-volume effects or complex distributions of frequency shifts within the voxel. Conclusion The presented phantom and computer programs are publicly available and may serve as a ground truth in future assessments of the faithfulness of quantitative susceptibility reconstruction algorithms.

Purpose The aim of the second quantitative susceptibility mapping (QSM) reconstruction challenge (Oct 2019, Seoul, Korea) was to test the accuracy of QSM dipole inversion algorithms in simulated brain data. Methods A two-stage design was chosen for this challenge. The participants were provided with datasets of multi-echo gradient echo images synthesized from two realistic in silico head phantoms using an MR simulator. At the first stage, participants optimized QSM reconstructions without ground truth data available to mimic the clinical setting. At the second stage, ground truth data were provided for parameter optimization. Submissions were evaluated using eight numerical metrics and visual ratings. Results A total of 98 reconstructions were submitted for stage 1 and 47 submissions for stage 2. Iterative methods had the best quantitative metric scores, followed by deep learning and direct inversion methods. Priors derived from magnitude data improved the metric scores. Algorithms based on iterative approaches and total variation (and its derivatives) produced the best overall results. The reported results and analysis pipelines have been made public to allow researchers to compare new methods to the current state of the art. Conclusion The synthetic data provide a consistent framework to test the accuracy and robustness of QSM algorithms in the presence of noise, calcifications and minor voxel dephasing effects. Total Variation-based algorithms produced the best results among all metrics. Future QSM challenges should assess whether this good performance with synthetic datasets translates to more realistic scenarios, where background fields and dipole-incompatible phase contributions are included.

PurposeThe aim of the second quantitative susceptibility mapping (QSM) reconstruction challenge (Oct 2019, Seoul, Korea) was to test the accuracy of QSM dipole inversion algorithms in simulated brain data.MethodsA two-stage design was chosen for this challenge. The participants were provided with datasets of multi-echo gradient echo images synthesized from two realisticin silicohead phantoms using an MR simulator. At the first stage, participants optimized QSM reconstructions without ground-truths available to mimic the clinical setting. At the second stage, ground-truths were provided for parameter optimization.Submissions were evaluated using eight numerical metrics and visual ratings.ResultsA total of 98 reconstructions were submitted for stage 1 and 47 submissions for stage 2. Iterative methods had the best quantitative metric scores, followed by deep-learning and direct inversion methods. Priors derived from magnitude data improved the metric scores. Algorithms based on iterative approaches and Total Variation (and its derivatives) produced the best overall results. The reported results and analysis pipelines have been made public to allow researchers to compare new methods to the current state of the art.ConclusionThe synthetic data provides a consistent framework to test the accuracy and robustness of QSM algorithms in the presence of noise, calcifications and minor voxel dephasing effects. Total Variation-based algorithmsproduced the best results along all metrics. Future QSM challenges should asses if this good performance with synthetic datasets translates to more realistic scenarios, where background fields and dipole-incompatible phase contributions are included.

PurposeTo create a realistic in-silico head phantom for the second Quantitative Susceptibility Mapping (QSM) Reconstruction Challenge and for future evaluations of processing algorithms for QSM.MethodsWe created a digital whole-head tissue property phantom by segmenting and post-processing high-resolution (0.64mm isotropic), multi-parametric MRI data acquired at 7T from a healthy volunteer. We simulated the steady-state magnetization at 7T using a Bloch simulator and mimicked a Cartesian sampling scheme through Fourier-based processing. Computer code for generating the phantom and performing the MR simulation was designed to facilitate flexible modifications of the phantom in the future, such as the inclusion of pathologies, as well as the simulation of a wide range of acquisition protocols. Specifically, the following parameters and effects were implemented: repetition time and echo time, voxel size, background fields, and RF phase biases. Diffusion weighted imaging phantom data is provided allowing future investigations of tissue microstructure effects in phase and QSM algorithms.ResultsThe brain-part of the phantom featured realistic morphology with spatial variations in relaxation and susceptibility values similar to the in vivo setting. We demonstrated some of the phantom’s properties, including the possibility of generating phase data with non-linear evolution over echo time due to partial volume effects or complex distributions of frequency shifts within the voxel.ConclusionThe presented phantom and computer programs are publicly available and may serve as a ground truth in future assessments of the faithfulness of quantitative susceptibility reconstruction algorithms.

Background Deep gray matter structures in patients with Alzheimer disease (AD) contain higher brain iron concentrations. However, few studies have included neocortical areas, which are challenging to assess with MRI. Purpose To investigate baseline and change in brain iron levels using MRI at 3 T with R2* relaxation rate mapping in individuals with AD compared with healthy control (HC) participants. Materials and Methods In this prospective study, participants with AD recruited between 2010 and 2016 and age-matched HC participants selected from 2010 to 2014 were evaluated. Of 100 participants with AD, 56 underwent subsequent neuropsychological testing and brain MRI at a mean follow-up of 17 months. All participants underwent 3-T MRI, including R2* mapping corrected for macroscopic B0 field inhomogeneities. Anatomic structures were segmented, and median R2* values were calculated in the neocortex and cortical lobes, basal ganglia (BG), hippocampi, and thalami. Multivariable linear regression analysis was applied to study the difference in R2* levels between groups and the association between longitudinal changes in R2* values and cognition in the AD group. Results A total of 100 participants with AD (mean age, 73 years ± 9 [standard deviation]; 58 women) and 100 age-matched HC participants (mean age, 73 years ± 9; 60 women) were evaluated. Median R2* levels were higher in the AD group than in the HC group in the BG (HC, 29.0 sec

Objective We investigated R2* relaxation rates as a marker of iron content in the substantia nigra in patients with common tremor disorders and explored their diagnostic properties. Methods Mean nigral R2* rates were measured in 40 patients with tremor‐dominant Parkinson's disease (PD), 15 with tremor in dystonia, 25 with essential tremor, and 25 healthy controls. Results Tremor‐dominant PD patients had significantly higher nigral R2* values (34.1 ± 5.7) than those with tremor in dystonia (30.0 ± 3.9), essential tremor (30.6 ± 4.8), and controls (30.0 ± 2.8). An R2* threshold of 31.15 separated tremor‐dominant PD from controls with a sensitivity and specificity of 67.5% and 72%. The sensitivity and specificity for discrimination between PD and non‐PD tremor patients was 67.5% and 60%. Conclusion Iron content in the substantia nigra is significantly higher in tremor‐dominant PD than in tremor in dystonia, essential tremor, and controls. Because of the considerable overlap, nigral R2* cannot be suggested as a useful diagnostic tool. © 2018 The Authors. Movement Disorders published by Wiley Periodicals, Inc. on behalf of International Parkinson and Movement Disorder Society.

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, 2017. © 2017 International Society for Magnetic Resonance in Medicine.

Purpose Current MRI techniques cannot reliably assess iron content in white matter due to the confounding diamagnetic effect of myelin. The purpose of this study was to validate with histology a novel iron mapping technique that uses the temperature dependency of the paramagnetic susceptibility in multiple sclerosis (MS) brains, where white matter has been reported to show significant variations in iron content. Methods We investigated post mortem brain tissue from three MS patients and one control subject. Temperature-dependent R2* relaxometry was performed between 4°C and 37°C. The resulting temperature coefficient ( TcR2*) maps were compared with immunohistochemical stains for ferritin light chain. Results Good agreement between TcR2* maps and ferritin staining was found by way of visual comparison and quantitative analysis. The highest iron concentrations were detected at the edge of MS lesions and in the basal ganglia. For all regions, except the subcortical U-fibers, there was a significant negative correlation between the TcR2* values and the ferritin count. Conclusion This study provides further evidence that TcR2* may be a reliable measure of white matter iron content due to the elimination of myelin-induced susceptibility changes and is well suited for further research into neurological diseases with distortions of the iron homeostasis. Magn Reson Med 79:1609-1615, 2018. © 2017 International Society for Magnetic Resonance in Medicine.

Purpose Formalin fixation prevents tissue autolysis by crosslinking proteins and changes tissue microstructure and MRI signal characteristics. Previous studies showed high variations in MR relaxation time constants of formalin fixed brain tissue, which has been attributed to the use of different formalin concentrations. Our investigations confirmed the influence of formalin concentration on relaxation times and unexpectedly revealed an influence of vendor specific formalin composition, which has not been investigated so far. Methods We systematically analyzed relaxation times of human brain tissue fixed with 4% and 10% formalin compared with unfixed condition at 3 Tesla MRI. Furthermore, we assessed relaxation times of nine formalin solutions from different vendors and performed comparisons of their magnetic susceptibility by SQUID (superconducting quantum interference device) magnetometry. Results Tissue relaxation times decreased approximately twice as fast using 10% than in 4% formalin fixation. The vendor specific composition of the formalin solutions and concentration dependent paramagnetic effects showed a substantial contribution to differences in relaxation times of formalin. Conclusion Our study demonstrates that differences of the formalin composition have substantial effects on MRI signal characteristics after fixation, which can explain the divergence of reported relaxation times beyond the effect of differences in formalin concentration. Magn Reson Med 79:1111-1115, 2018. © 2017 International Society for Magnetic Resonance in Medicine.

Purpose The 4th International Workshop on MRI Phase Contrast and QSM (2016, Graz, Austria) hosted the first QSM Challenge. A single-orientation gradient recalled echo acquisition was provided, along with COSMOS and the χ33 STI component as ground truths. The submitted solutions differed more than expected depending on the error metric used for optimization and were generally over-regularized. This raised (unanswered) questions about the ground truths and the metrics utilized. Methods We investigated the influence of background field remnants by applying additional filters. We also estimated the anisotropic contributions from the STI tensor to the apparent susceptibility to amend the χ33 ground truth and to investigate the impact on the reconstructions. Lastly, we used forward simulations from the COSMOS reconstruction to investigate the impact noise had on the metric scores. Results Reconstructions compared against the amended STI ground truth returned lower errors. We show that the background field remnants had a minor impact in the errors. In the absence of inconsistencies, all metrics converged to the same regularization weights, whereas structural similarity index metric was more insensitive to such inconsistencies. Conclusion There was a mismatch between the provided data and the ground truths due to the presence of unaccounted anisotropic susceptibility contributions and noise. Given the lack of reliable ground truths when using in vivo acquisitions, simulations are suggested for future QSM Challenges.

Quantitative susceptibility mapping (QSM) reveals pathological changes in widespread diseases such as Parkinson’s disease, Multiple Sclerosis, or hepatic iron overload. QSM requires multiple processing steps after the acquisition of magnetic resonance imaging (MRI) phase measurements such as unwrapping, background field removal and the solution of an ill-posed field-to-source-inversion. Current techniques utilize iterative optimization procedures to solve the inversion and background field correction, which are computationally expensive and lead to suboptimal or over-regularized solutions requiring a careful choice of parameters that make a clinical application of QSM challenging. We have previously demonstrated that a deep convolutional neural network can invert the magnetic dipole kernel with a very efficient feed forward multiplication not requiring iterative optimization or the choice of regularization parameters. In this work, we extended this approach to remove background fields in QSM. The prototype method, called SHARQnet, was trained on simulated background fields and tested on 3T and 7T brain datasets. We show that SHARQnet outperforms current background field removal procedures and generalizes to a wide range of input data without requiring any parameter adjustments. In summary, we demonstrate that the solution of ill-posed problems in QSM can be achieved by learning the underlying physics causing the artifacts and removing them in an efficient and reliable manner and thereby will help to bring QSM towards clinical applications.

This study explored whether autoregulatory shifts in cerebral blood volume induce susceptibility changes large enough to be depicted by quantitative susceptibility mapping. Eight healthy subjects underwent fast quantitative susceptibility mapping at 3T while lying down to slowly decrease mean arterial pressure. A linear relationship between mean arterial pressure and susceptibility was observed in cortical and subcortical structures, likely representing vessels involved in autoregulation. The slope of this relationship is assumed to indicate the extent of cerebral vascular compliance.

Background: Cerebrospinal fluid (CSF) immunoglobulin free light chains (FLC) have been suggested as quantitative alternative to oligoclonal bands (OCB) in the diagnosis of multiple sclerosis (MS). However, little is known on their role in predicting clinical and paraclinical disease progression, particularly in early stages. Objective: To assess the prognostic value of FLC in OCB-positive patients with clinically isolated syndrome (CIS) suggestive of MS and early MS. Methods: We determined FLC kappa (KFLC) and lambda (LFLC) in CSF and serum by nephelometry in 61 patients (CIS ( n = 48), relapsing-remitting multiple sclerosis ( n = 13)) and 60 non-inflammatory neurological controls. Median clinical follow-up time in CIS was 4.8 years (interquartile range (IQR), 1.5–6.5 years). Patients underwent 3T magnetic resonance imaging (MRI) at baseline and follow-up (median time interval, 2.2 years; IQR, 1.0–3.7 years) to determine T2 lesion load (T2LL) and percent brain volume change (PBVC). Results: CSF FLC were significantly increased in CIS/MS compared to controls (all p Conclusion: Our study confirms increased intrathecal synthesis of FLC in CIS/MS which supports their diagnostic contribution. The KFLC/LFLC CSF ratio appears to have a prognostic value in CIS beyond OCB.

Post-mortem MRI of the brain is increasingly applied in neuroscience for a better understanding of the contrast mechanisms of disease induced tissue changes. However, the influence of chemical processes caused by formalin fixation and differences in temperature may hamper the comparability with results from in vivo MRI. In this study we investigated how formalin fixation and temperature affect T1, T2 and T2* relaxation times of brain tissue. Fixation effects were examined with respect to changes in water content and crosslinking. Relaxometry was performed in brain slices from five deceased subjects at different temperatures. All measurements were repeated after 190 days of formaldehyde immersion. The water content of unfixed and fixed tissue was determined using the wet-to-dry ratio following drying. Protein weight was determined with sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). Fixation caused a strong decrease of all relaxation times, the strongest effect being seen on T1, with a reduction of up to 76%. The temperature coefficient of T1 was lower in the fixed than unfixed tissue, which was in contrast to T2, where an increase of the temperature coefficient was observed following fixation. The reduction of the water content after fixation was in the range of 1-6% and thus not sufficient to explain the changes in relaxation time. Results from SDS-PAGE indicated a strong increase of the protein size above 260 kDa in all brain structures examined. Our results suggest that crosslinking induced changes of the macromolecular matrix are responsible for T1 shortening and a decreased temperature dependency. The relaxation times provided in this work should allow optimization of post-mortem MRI protocols for the brain.

Three-dimensional Gradient Echo (GRE) is the main workhorse sequence used for Susceptibility Weighted Imaging (SWI), Quantitative Susceptibility Mapping (QSM) and Susceptibility Tensor Imaging (STI). Achieving optimal phase signal-to-noise ratio requires late echo times, thus necessitating a long repetition time (TR). Combined with the large encoding burden of whole-brain coverage with high resolution, this leads to increased scan time. Further, the dipole kernel relating the tissue phase to the underlying susceptibility distribution undersamples the frequency content of the susceptibility map. Scans at multiple head orientations along with Calculation Of Susceptibility through Multi-Orientation Sampling (COSMOS) are one way to effectively mitigate this issue. Additionally, STI requires a minimum of 6 head orientations to solve for the independent tensor elements in each voxel. The requirements of high-resolution imaging with long TR at multiple orientations substantially lengthen the acquisition of COSMOS and STI. The goal of this work is to dramatically speed up susceptibility mapping at multiple head orientations. We demonstrate highly efficient acquisition using 3D-GRE with Wave-CAIPI, and dramatically reduce the acquisition time of these protocols. Using R=15 fold acceleration with Wave-CAIPI permits acquisition per head orientation in 90 s at 1.1 mm isotropic resolution, and 5:35 minutes at 0.5 mm isotropic resolution. Since Wave-CAIPI fully harnesses the 3D spatial encoding capability of receive arrays, the maximum g-factor noise amplification remains below 1.30 at 3T and 1.12 at 7T. This allows a 30-minute exam for STI with 12 orientations, thus paving the way to its clinical application.

PURPOSE: Background‐field removal is a crucial preprocessing step for quantitative susceptibility mapping (QSM). Remnants from this step often contaminate the estimated local field, which in turn leads to erroneous tissue‐susceptibility reconstructions. The present work aimed to mitigate this undesirable behavior with the development of a new approach that simultaneously decouples background contributions and local susceptibility sources on QSM inversion. METHODS: Input phase data for QSM can be seen as a composite scalar field of local effects and residual background components. We developed a new weak‐harmonic regularizer to constrain the latter and to separate the 2 components. The resulting optimization problem was solved with the alternating directions of multipliers method framework to achieve fast convergence. In addition, for convenience, a new alternating directions of multipliers method–based preconditioned nonlinear projection onto dipole fields solver was developed to enable initializations with wrapped‐phase distributions. Weak‐harmonic QSM, with and without nonlinear projection onto dipole fields preconditioning, was compared with the original (alternating directions of multipliers method–based) total variation QSM algorithm in phantom and in vivo experiments. RESULTS: Weak‐harmonic QSM returned improved reconstructions regardless of the method used for background‐field removal, although the proposed nonlinear projection onto dipole fields method often obtained better results. Streaking and shadowing artifacts were substantially suppressed, and residual background components were effectively removed. CONCLUSION: Weak‐harmonic QSM with field preconditioning is a robust dipole inversion technique and has the potential to be extended as a single‐step formulation for initialization with uncombined multi‐echo data.

Quantitative susceptibility mapping (QSM) is based on magnetic resonance imaging (MRI) phase measurements and has gained broad interest because it yields relevant information on biological tissue properties, predominantly myelin, iron and calcium in vivo. Thereby, QSM can also reveal pathological changes of these key components in widespread diseases such as Parkinson's disease, Multiple Sclerosis, or hepatic iron overload. While the ill-posed field-to-source-inversion problem underlying QSM is conventionally assessed by the means of regularization techniques, we trained a fully convolutional deep neural network - DeepQSM - to directly invert the magnetic dipole kernel convolution. DeepQSM learned the physical forward problem using purely synthetic data and is capable of solving the ill-posed field-to-source inversion on in vivo MRI phase data. The magnetic susceptibility maps reconstructed by DeepQSM enable identification of deep brain substructures and provide information on their respective magnetic tissue properties. In summary, DeepQSM can invert the magnetic dipole kernel convolution and delivers robust solutions to this ill-posed problem.

Quantitative susceptibility mapping (QSM) aims to extract the magnetic susceptibility of tissue from magnetic resonance imaging (MRI) phase measurements. The mapping of magnetic susceptibility in vivo has gained broad interest in several fields of science and medicine because it yields relevant information on biological tissue properties, predominantly myelin, iron and calcium. Thereby, QSM can also reveal pathological changes of these key components in devastating diseases such as Parkinson’s disease, Multiple Sclerosis, or hepatic iron overload. As QSM requires the solution of an ill-posed field-to-source-inversion, current techniques utilize manual optimization of regularization parameters to balance between smoothing, artifacts and quantification accuracy. We trained a fully convolutional deep neural network - DeepQSM - to invert the magnetic dipole kernel convolution. This network is capable of solving the ill-posed field-to-source inversion on real-world in vivo MRI phase data without the need for manual parameter tuning, which proves that this network has generalized the underlying mathematical principle of the dipole inversion. We demonstrate that DeepQSM’s susceptibility maps enable identification of deep brain substructures that are not visible in MRI phase data and provide information on their respective magnetic tissue properties. We illustrate DeepQSM’s clinical relevance in a patient with multiple sclerosis showing its sensitivity to white matter lesions. In summary, DeepQSM can be used to determine the composition of myelin sheets of nerve fibers in the brain, and to assess quantitative information on iron homeostasis and its dysregulation, and will subsequently contribute to a better understanding of these biological processes in health and disease.

Objective It has been suggested recently that cortical pathology in multiple sclerosis (MS) may, at least partly, be caused by factors in cerebrospinal fluid (CSF). We thus hypothesized that MS-related tissue changes in compartments close to the CSF, such as periventricular lesions, might correlate with cortical pathology. Methods We investigated a cohort of 160 patients, comprising 91 with a clinically isolated syndrome (CIS) and 69 with relapsing-remitting MS (RRMS; mean age: CIS: 31.4 ± 9.0; RRMS: 33.0 ± 8.7 years; mean disease duration: CIS: 7.2 ± 15 months; RRMS: 8.0 ± 6.5 years, Expanded Disability Status Scale (median, min-max): CIS: 1, 0–3.5; RRMS: 1.25, 0–4) with 3.0T magnetic resonance imaging. MS lesions were segmented semiautomatically on fluid-attenuated inversion recovery images. To quantify periventricular lesion load (PV-LL), we generated ventricle masks and dilated them by a voxel factor of 3. Lesions within the dilated ventricle margin were classified as periventricular. Cortical thinning was assessed by cortical mean thickness (CMT) and compared to data from 58 healthy controls (HCs; mean age: 29.1 ± 7.4 years). Results Compared to HC, CIS and (even more so) RRMS patients demonstrated significantly reduced CMT. Even after controlling for ventricular volume and total lesion load, increased periventricular lesion occupancy (percentage of PV-LL) significantly correlated with decreased CMT in RRMS (r = –0.295; p = 0.015), but not in CIS (r = 0.032; p = 0.768) patients. Interpretation The correlation between increased periventricular lesion burden and decreased CMT indicative of subpial cortical pathology supports the concept that common CSF-mediated factors might play a role in the accumulation of damage to gray and white matter in MS. Ann Neurol 2015;78:530–539

Background: The extent and clinical significance of brain volume changes in different phases of multiple sclerosis (MS) is still under discussion. Objective: To determine the rate of global and compartmental brain volume changes in patients with a clinically-isolated syndrome (CIS) compared to patients with definite MS, by long-term follow-up and as a predictor of conversion to MS in a routine clinical setting. Methods: We investigated 120 patients (63 CIS and 57 MS) at baseline and after a mean follow-up period of 43 months, including detailed clinical examination and 3-Tesla magnetic resonance imaging (MRI). Our imaging analyses comprised the normalized brain volume (NBV), cortical grey matter (cGMV) and white matter (WMV) volumes using SIENA/X, the percentage of brain volume change (PBVC) using SIENA and the change in the volume of the thalami (TV) and basal ganglia (BGV). We also determined the amount and change of T2-lesion load (T2-LL). Results: At baseline, all the brain volume metrics, except cGMV, were significantly lower; and the T2-LL was significantly higher, in patients with MS rather than CIS. During the follow-up, only the PBVC was higher in MS ( p = 0.008) and this difference was driven by converters from CIS to MS. Quartiles of PBVC did not allow us to predict conversion to MS, but were associated with the degree of disability. Conclusions: PBVC is the most sensitive marker of progressing atrophy and a higher PBVC was generally associated with more active disease; however, it did not serve to predict the course of MS on an individual basis, in this study.

We investigated longitudinal changes in iron concentration in the subcortical gray matter (caudate nucleus, globus pallidus, putamen, thalamus) of patients with clinically isolated syndrome (CIS) and definite multiple sclerosis (MS) and their relation to clinical and other morphologic variables.We followed 144 patients (76 CIS; median Expanded Disability Status Scale [EDSS] 1.0 [interquartile range (IQR) 0.0-2.0]; 68 MS; median EDSS 2.0 [IQR 1.0-3.3]) clinically and with 3T MRI over a median period of 2.9 (IQR 1.3-4.0) years. Iron concentration was determined by R2* relaxometry at baseline and last follow-up.At baseline, subcortical gray matter iron deposition was higher in MS compared to CIS. In CIS, R2* rates increased in the globus pallidus (p0.001), putamen (p0.001), and caudate nucleus (p0.001), whereas R2* rates in the thalamus decreased (p0.05). In MS, R2* rates increased in the putamen (p0.05), remained stable in the globus pallidus and caudate nucleus, and decreased in the thalamus (p0.01). Changes in R2* relaxation rates were unrelated to changes in the volume of respective structures, of T2 lesion load, and of disability.Iron accumulation in the basal ganglia is more pronounced in the early than later phases of the disease and occurs independent from other morphologic brain changes. Short-term changes in iron concentration are not associated with disease activity or changes in disability.

Besides specific iron accumulation in some neurodegenerative disorders, increased iron deposition in cerebral deep gray matter (DGM) is found in multiple sclerosis. As this is considered largely a white matter (WM) disease, we speculated that patients with more severe ischemic WM hyperintensities (WMH) might also have an increased iron concentration in DGM structures and tested this assumption by using magnetic resonance imaging-based quantitative R2* relaxometry. WMH severity was measured in 61 patients with acute transient neurological symptoms (mean age: 71.5 ± 8.3 years) undergoing 3-Tesla magnetic resonance imaging. Despite a 6-year higher age of patients with more severe (i.e., early confluent or confluent) WMH, their DGM R2* rates did not differ from patients with punctate or no WMH. In the globus pallidum, R2* rates were even lower in patients with severe WMH. WMH volume was not correlated with R2* levels in any of the analyzed DGM structures. These findings argue against WM damage per se causing increased DGM iron deposition in multiple sclerosis and suggest no role of iron accumulation in ischemic small vessel disease.

Background: Vascular risk factors (VRF) in multiple sclerosis (MS) patients have been associated with lower brain volumes. It is currently unknown if this association already exists in early MS and how it develops over time. Methods: We identified 82 patients with clinically isolated syndrome (CIS) ( n = 61) or with early relapsing-remitting MS ( n = 21) and assessed their VRF including arterial hypertension, hyperlipidaemia, diabetes mellitus and smoking. We analysed T2-lesion load, normalized brain volume (NBV), cortical grey (cGMV) and white matter volumes (WMV), thalamic and basal ganglia volumes at baseline and follow-up magnetic resonance imaging (MRI) and assessed the percentage of brain volume change (PBVC) using SIENA. Results: Patient mean age was 32.4 (±8.7) years and 54 (65%) were women. Median follow-up period was 42 (29–54) months. In total, 26 patients (31.7%) had one or more VRF (VRF+). At baseline, VRF+ patients had a lower NBV (1530.9 cm3 vs 1591.2 cm3, p = 0.001), a lower cGMV (628.5 cm3 vs 668.6 cm3, p = 0.002) and WMV (752.2 cm3 vs 783.9 cm3, p = 0.009) than VRF-negative patients. Similar results were obtained at follow-up. PBVC was comparable between patients with and without VRF. Conclusion: VRF are associated with lower brain volume already in early MS but do not lead to increased brain volume loss during 3.5 years of follow-up.

Current MRI techniques cannot reliably assess iron content in white matter due to the confounding diamagnetic effect of myelin. The purpose of this study was to validate with histology a novel iron mapping technique that uses the temperature dependency of the paramagnetic susceptibility in multiple sclerosis (MS) brains, where white matter has been reported to show significant variations in iron content.We investigated post mortem brain tissue from three MS patients and one control subject. Temperature-dependent R2* relaxometry was performed between 4°C and 37°C. The resulting temperature coefficient ( TcR2*) maps were compared with immunohistochemical stains for ferritin light chain.Good agreement between TcR2* maps and ferritin staining was found by way of visual comparison and quantitative analysis. The highest iron concentrations were detected at the edge of MS lesions and in the basal ganglia. For all regions, except the subcortical U-fibers, there was a significant negative correlation between the TcR2* values and the ferritin count.This study provides further evidence that TcR2* may be a reliable measure of white matter iron content due to the elimination of myelin-induced susceptibility changes and is well suited for further research into neurological diseases with distortions of the iron homeostasis. Magn Reson Med 79:1609-1615, 2018. © 2017 International Society for Magnetic Resonance in Medicine.