Quantitative magnetic susceptibility mapping without phase unwrapping using WASSR

Deep learning-based Lorentzian fitting of water saturation shift referencing spectra in MRI

PURPOSE

Water saturation shift referencing (WASSR) Z-spectra are used commonly for field referencing in chemical exchange saturation transfer (CEST) MRI. However, their analysis using least-squares (LS) Lorentzian fitting is time-consuming and prone to errors because of the unavoidable noise in vivo. A deep learning-based single Lorentzian Fitting Network (sLoFNet) is proposed to overcome these shortcomings.

METHODS

A neural network architecture was constructed and its hyperparameters optimized. Training was conducted on a simulated and in vivo-paired data sets of discrete signal values and their corresponding Lorentzian shape parameters. The sLoFNet performance was compared with LS on several WASSR data sets (both simulated and in vivo 3T brain scans). Prediction errors, robustness against noise, effects of sampling density, and time consumption were compared.

RESULTS

LS and sLoFNet performed comparably in terms of RMS error and mean absolute error on all in vivo data with no statistically significant difference. Although the LS method fitted well on samples with low noise, its error increased rapidly when increasing sample noise up to 4.5%, whereas the error of sLoFNet increased only marginally. With the reduction of Z-spectral sampling density, prediction errors increased for both methods, but the increase occurred earlier (at 25 vs. 15 frequency points) and was more pronounced for LS. Furthermore, sLoFNet performed, on average, 70 times faster than the LS-method.

CONCLUSION

Comparisons between LS and sLoFNet on simulated and in vivo WASSR MRI Z-spectra in terms of robustness against noise and decreased sample resolution, as well as time consumption, showed significant advantages for sLoFNet.

Phosphates form spectroscopically dark state assemblies in common aqueous solutions

Phosphates and polyphosphates play ubiquitous roles in biology as integral structural components of cell membranes and bone, or as vehicles of energy storage via adenosine triphosphate and phosphocreatine. The solution phase space of phosphate species appears more complex than previously known. We present nuclear magnetic resonance (NMR) and cryogenic transmission electron microscopy (cryo-TEM) experiments that suggest phosphate species including orthophosphates, pyrophosphates, and adenosine phosphates associate into dynamic assemblies in dilute solutions that are spectroscopically "dark." Cryo-TEM provides visual evidence of the formation of spherical assemblies tens of nanometers in size, while NMR indicates that a majority population of phosphates remain as unassociated ions in exchange with spectroscopically invisible assemblies. The formation of these assemblies is reversibly and entropically driven by the partial dehydration of phosphate groups, as verified by diffusion-ordered spectroscopy (DOSY), indicating a thermodynamic state of assembly held together by multivalent interactions between the phosphates. Molecular dynamics simulations further corroborate that orthophosphates readily cluster in aqueous solutions. This study presents the surprising discovery that phosphate-containing molecules, ubiquitously present in the biological milieu, can readily form dynamic assemblies under a wide range of commonly used solution conditions, highlighting a hitherto unreported property of phosphate's native state in biological solutions.

Accelerating quantitative susceptibility and R2* mapping using incoherent undersampling and deep neural network reconstruction

Quantitative susceptibility mapping (QSM) and R2* mapping are MRI post-processing methods that quantify tissue magnetic susceptibility and transverse relaxation rate distributions. However, QSM and R2* acquisitions are relatively slow, even with parallel imaging. Incoherent undersampling and compressed sensing reconstruction techniques have been used to accelerate traditional magnitude-based MRI acquisitions; however, most do not recover the full phase signal, as required by QSM, due to its non-convex nature. In this study, a learning-based Deep Complex Residual Network (DCRNet) is proposed to recover both the magnitude and phase images from incoherently undersampled data, enabling high acceleration of QSM and R2* acquisition. Magnitude, phase, R2*, and QSM results from DCRNet were compared with two iterative and one deep learning methods on retrospectively undersampled acquisitions from six healthy volunteers, one intracranial hemorrhage and one multiple sclerosis patients, as well as one prospectively undersampled healthy subject using a 7T scanner. Peak signal to noise ratio (PSNR), structural similarity (SSIM), root-mean-squared error (RMSE), and region-of-interest susceptibility and R2* measurements are reported for numerical comparisons. The proposed DCRNet method substantially reduced artifacts and blurring compared to the other methods and resulted in the highest PSNR, SSIM, and RMSE on the magnitude, R2*, local field, and susceptibility maps. Compared to two iterative and one deep learning methods, the DCRNet method demonstrated a 3.2% to 9.1% accuracy improvement in deep grey matter susceptibility when accelerated by a factor of four. The DCRNet also dramatically shortened the reconstruction time of single 2D brain images from 36-140 seconds using conventional approaches to only 15-70 milliseconds.

Coil combination using linear deconvolution in k-space for phase imaging

Background

The combination of multi-channel data is a critical step for the imaging of phase and susceptibility contrast in magnetic resonance imaging (MRI). Magnitude-weighted phase combination methods often produce noise and aliasing artifacts in the magnitude images at accelerated imaging sceneries. To address this issue, an optimal coil combination method through deconvolution in k-space is proposed in this paper.

Methods

The proposed method firstly employs the sum-of-squares and phase aligning method to yield a complex reference coil image which is then used to calculate the coil sensitivity and its Fourier transform. Then, the coil k-space combining weights is computed, taking into account the truncated frequency data of coil sensitivity and the acquired k-space data. Finally, combining the coil k-space data with the acquired weights generates the k-space data of proton distribution, with which both phase and magnitude information can be obtained straightforwardly. Both phantom and in vivo imaging experiments were conducted to evaluate the performance of the proposed method.

Results

Compared with magnitude-weighted method and MCPC-C, the proposed method can alleviate the phase cancellation in coil combination, resulting in a less wrapped phase.

Conclusions

The proposed method provides an effective and efficient approach to combine multiple coil image in parallel MRI reconstruction, and has potential to benefit routine clinical practice in the future.

Overview of quantitative susceptibility mapping

Magnetic susceptibility describes the magnetizability of a material to an applied magnetic field and represents an important parameter in the field of MRI. With the recently introduced method of quantitative susceptibility mapping (QSM) and its conceptual extension to susceptibility tensor imaging (STI), the non-invasive assessment of this important physical quantity has become possible with MRI. Both methods solve the ill-posed inverse problem to determine the magnetic susceptibility from local magnetic fields. Whilst QSM allows the extraction of the spatial distribution of the bulk magnetic susceptibility from a single measurement, STI enables the quantification of magnetic susceptibility anisotropy, but requires multiple measurements with different orientations of the object relative to the main static magnetic field. In this review, we briefly recapitulate the fundamental theoretical foundation of QSM and STI, as well as computational strategies for the characterization of magnetic susceptibility with MRI phase data. In the second part, we provide an overview of current methodological and clinical applications of QSM with a focus on brain imaging. Copyright © 2016 John Wiley & Sons, Ltd.

Iron quantification with susceptibility

Iron is an essential trace element involved in a variety of biological mechanisms in the human body. Disturbances of iron homeostasis have been observed in several inflammatory and degenerative diseases, which have raised strong interest in non-invasive iron mapping techniques. Numerous MRI techniques have been proposed so far, mostly based on the field changes induced by the magnetic properties of iron. Each of these approaches has a specific sensitivity for iron and its microstructural environment. Quantitative susceptibility mapping is the latest development and provides a direct measure of bulk susceptibility. However, field changes induced by iron are not always directly related to the concentration of iron, but rather reflect the structure of iron compounds and its cellular distribution. This review provides an overview of the most relevant iron compounds in the human body, their magnetic properties and their cellular distribution. In addition, MRI methods based on direct or indirect susceptibility changes are presented and discussed with respect to technical aspects and clinical applicability. Copyright © 2016 John Wiley & Sons, Ltd.

Simultaneous mapping of water shift and B1 (WASABI)-Application to field-Inhomogeneity correction of CEST MRI data

PURPOSE

Together with the development of MRI contrasts that are inherently small in their magnitude, increased magnetic field accuracy is also required. Hence, mapping of the static magnetic field (B₀ ) and the excitation field (B₁ ) is not only important to feedback shim algorithms, but also for postprocess contrast-correction procedures.

METHODS

A novel field-inhomogeneity mapping method is presented that allows simultaneous mapping of the water shift and B₁ (WASABI) using an off-resonant rectangular preparation pulse. The induced Rabi oscillations lead to a sinc-like spectrum in the frequency-offset dimension and allow for determination of B₀ by its symmetry axis and of B₁ by its oscillation frequency.

RESULTS

Stability of the WASABI method with regard to the influences of T₁ , T₂ , magnetization transfer, and repetition time was investigated and its convergence interval was verified. B₀ and B₁ maps obtained simultaneously by means of WASABI in the human brain at 3 T and 7 T can compete well with maps obtained by standard methods. Finally, the method was applied successfully for B₀ and B₁ correction of chemical exchange saturation transfer MRI (CEST-MRI) data of the human brain.

CONCLUSION

The proposed WASABI method yields a novel simultaneous B₀ and B₁ mapping within 1 min that is robust and easy to implement. Magn Reson Med 77:571-580, 2017. © 2016 International Society for Magnetic Resonance in Medicine.

Foundations of MRI phase imaging and processing for Quantitative Susceptibility Mapping (QSM)

Quantitative Susceptibility Mapping (QSM) is a novel MRI based technique that relies on estimates of the magnetic field distribution in the tissue under examination. Several sophisticated data processing steps are required to extract the magnetic field distribution from raw MRI phase measurements. The objective of this review article is to provide a general overview and to discuss several underlying assumptions and limitations of the pre-processing steps that need to be applied to MRI phase data before the final field-to-source inversion can be performed. Beginning with the fundamental relation between MRI signal and tissue magnetic susceptibility this review covers the reconstruction of magnetic field maps from multi-channel phase images, background field correction, and provides an overview of state of the art QSM solution strategies.

Appraising the Role of Iron in Brain Aging and Cognition: Promises and Limitations of MRI Methods

Age-related increase in frailty is accompanied by a fundamental shift in cellular iron homeostasis. By promoting oxidative stress, the intracellular accumulation of non-heme iron outside of binding complexes contributes to chronic inflammation and interferes with normal brain metabolism. In the absence of direct non-invasive biomarkers of brain oxidative stress, iron accumulation estimated in vivo may serve as its proxy indicator. Hence, developing reliable in vivo measurements of brain iron content via magnetic resonance imaging (MRI) is of significant interest in human neuroscience. To date, by estimating brain iron content through various MRI methods, significant age differences and age-related increases in iron content of the basal ganglia have been revealed across multiple samples. Less consistent are the findings that pertain to the relationship between elevated brain iron content and systemic indices of vascular and metabolic dysfunction. Only a handful of cross-sectional investigations have linked high iron content in various brain regions and poor performance on assorted cognitive tests. The even fewer longitudinal studies indicate that iron accumulation may precede shrinkage of the basal ganglia and thus predict poor maintenance of cognitive functions. This rapidly developing field will benefit from introduction of higher-field MRI scanners, improvement in iron-sensitive and -specific acquisition sequences and post-processing analytic and computational methods, as well as accumulation of data from long-term longitudinal investigations. This review describes the potential advantages and promises of MRI-based assessment of brain iron, summarizes recent findings and highlights the limitations of the current methodology.

A multicontrast approach for comprehensive imaging of substantia nigra

We characterize the contrast behavior of substantia nigra (SN) in both magnetization transfer (MT) imaging, which is believed to be sensitive to neuromelanin (NM), and susceptibility weighted imaging (SWI). Images were acquired with a MT prepared dual echo gradient echo sequence. The first echo was taken as the MT contrast image and the second was used to generate the SWI image. SN volumes were segmented from these two types of images using a thresholding method. The spatial and signal characteristics of the extracted SWI and MT volumes were compared. Both images showed the presence of SN but the volumes of the SN identified in the two are spatially incongruent. The MT volume was more caudal than the SWI volume and with only a 12% overlap between the two volumes. Considering the SN volumes in each hemisphere separately, the average distances between the centers of mass of the volumes from the two types images are 5.1±1.1mm and 4.1±1.2mm, respectively. The frequency offsets (homodyne filtered phase/echo time) for the volumes derived from MT (NM) images and SWI images are 0.09±0.32radians/s and -1.12±0.57radians/s (p<0.0001), respectively. The MT contrasts for the two volumes are 0.16±0.02 and 0.10±0.03 (p<0.001), respectively. Our results indicate that the two contrasts are sensitive to different portions of the SN, with MT seeing the more caudal portion of the SN than SWI, likely due to variations of NM and iron content in the SN. Despite the small overlap, these regions are complementary. Our results provide a new understanding of the contrast behavior of the SN in the two imaging approaches commonly used to image it and indicate that using both may yield a more comprehensive visualization of the SN.

Reproducibility of quantitative susceptibility mapping in the brain at two field strengths from two vendors

PURPOSE

To assess the reproducibility of brain quantitative susceptibility mapping (QSM) in healthy subjects and in patients with multiple sclerosis (MS) on 1.5 and 3T scanners from two vendors.

MATERIALS AND METHODS

Ten healthy volunteers and 10 patients were scanned twice on a 3T scanner from one vendor. The healthy volunteers were also scanned on a 1.5T scanner from the same vendor and on a 3T scanner from a second vendor. Similar imaging parameters were used for all scans. QSM images were reconstructed using a recently developed nonlinear morphology-enabled dipole inversion (MEDI) algorithm with L1 regularization. Region-of-interest (ROI) measurements were obtained for 20 major brain structures. Reproducibility was evaluated with voxel-wise and ROI-based Bland-Altman plots and linear correlation analysis.

RESULTS

ROI-based QSM measurements showed excellent correlation between all repeated scans (correlation coefficient R ≥ 0.97), with a mean difference of less than 1.24 ppb (healthy subjects) and 4.15 ppb (patients), and 95% limits of agreements of within -25.5 to 25.0 ppb (healthy subjects) and -35.8 to 27.6 ppb (patients). Voxel-based QSM measurements had a good correlation (0.64 ≤ R ≤ 0.88) and limits of agreements of -60 to 60 ppb or less.

CONCLUSION

Brain QSM measurements have good interscanner and same-scanner reproducibility for healthy and MS subjects, respectively, on the systems evaluated in this study.

Quantitative susceptibility mapping (QSM): Decoding MRI data for a tissue magnetic biomarker

In MRI, the main magnetic field polarizes the electron cloud of a molecule, generating a chemical shift for observer protons within the molecule and a magnetic susceptibility inhomogeneity field for observer protons outside the molecule. The number of water protons surrounding a molecule for detecting its magnetic susceptibility is vastly greater than the number of protons within the molecule for detecting its chemical shift. However, the study of tissue magnetic susceptibility has been hindered by poor molecular specificities of hitherto used methods based on MRI signal phase and T2* contrast, which depend convolutedly on surrounding susceptibility sources. Deconvolution of the MRI signal phase can determine tissue susceptibility but is challenged by the lack of MRI signal in the background and by the zeroes in the dipole kernel. Recently, physically meaningful regularizations, including the Bayesian approach, have been developed to enable accurate quantitative susceptibility mapping (QSM) for studying iron distribution, metabolic oxygen consumption, blood degradation, calcification, demyelination, and other pathophysiological susceptibility changes, as well as contrast agent biodistribution in MRI. This paper attempts to summarize the basic physical concepts and essential algorithmic steps in QSM, to describe clinical and technical issues under active development, and to provide references, codes, and testing data for readers interested in QSM.