Elsevier

NeuroImage

Volume 54, Issue 4, 14 February 2011, Pages 2789-2807
NeuroImage

Quantitative imaging of intrinsic magnetic tissue properties using MRI signal phase: An approach to in vivo brain iron metabolism?

https://doi.org/10.1016/j.neuroimage.2010.10.070Get rights and content

Abstract

Quantitative susceptibility mapping (QSM) based on gradient echo (GRE) magnetic resonance phase data is a novel technique for non-invasive assessment of magnetic tissue susceptibility differences. The method is expected to be an important means to determine iron distributions in vivo and may, thus, be instrumental for elucidating the physiological role of iron and disease-related iron concentration changes associated with various neurological and psychiatric disorders. This study introduces a framework for QSM and demonstrates calculation of reproducible and orientation-independent susceptibility maps from GRE data acquired at 3T. The potential of these susceptibility maps to perform anatomical imaging is investigated, as well as the ability to measure the venous blood oxygen saturation level in large vessels, and to assess the local tissue iron concentration. In order to take into account diamagnetic susceptibility contributions induced by myelin, a correction scheme for susceptibility based iron estimation is demonstrated. The findings suggest that susceptibility contrast, and therewith also phase contrast, are not only linked to the storage iron concentration but are also significantly influenced by other sources such as myelin. After myelin correction the linear dependence between magnetic susceptibilities and previously published iron concentrations from post mortem studies was significantly improved. Finally, a comparison between susceptibility maps and processed phase images indicated that caution should be exercised when drawing conclusions about iron concentrations when directly assessing processed phase information.

Research highlights

►A framework for quantitative susceptibility mapping was introduced and evaluated. ►Susceptibility maps demonstrate rotation-invariant, quantitative anatomical contrast. ►Tissue susceptibility is not only influenced by tissue storage iron. ►Diamagnetic susceptibility contributions from myelin can be corrected. ►Direct phase measurements are unreliable for estimation of tissue iron concentration.

Introduction

It has been known for a long time that brain tissue in healthy adults possesses a relatively high concentration of iron (Spatz, 1922, Stein, 1923). This iron is either bound to hemoglobin and enzymes in the blood pool (heme iron), or is present in the parenchyma as storage iron (non-heme iron) (Haacke et al., 2005). The electronic configuration of hemoglobin is different in the absence and presence of oxygen. This blood oxygen level dependent (BOLD) effect (Ogawa et al., 1990) forms the basis for assessing the cerebral blood oxygenation saturation level, a much more specific parameter to characterize the physiological and pathological states of the brain than heme iron concentration itself (He and Yablonskiy, 2007). Together with cerebral blood flow, the blood oxygenation saturation level enables in principle assessment of the metabolic rate of oxygen (CMRO2). Non-heme iron is predominantly found in a mineralized form (ferrihydrite) associated with the iron storage protein ferritin and its breakdown product, hemosiderin (Harrison and Arosio, 1996, Schenck and Zimmerman, 2004). Histological stains have shown that various neurological and psychiatric disorders, including Alzheimer's disease (Bouras et al., 1997, Hallgren and Sourander, 1960, LeVine, 1997), Huntington's disease (Chen et al., 1993, Dexter et al., 1991), multiple sclerosis (LeVine, 1997), and Parkinson's disease (Chen et al., 1993, Dexter et al., 1991), may be associated with elevated focal accumulations of iron. Increased iron accumulations have also been detected in the vicinity of chronic hemorrhage, cerebral infarction, anemia, thalassemia, hemochromatosis, and Hallervorden-Spatz syndrome (Haacke et al., 2005). Both the physiological role and reasons for the distribution patterns of non-heme iron are not well understood yet (Gerlach et al., 1994, Ke and Qian, 2003, Schenck and Zimmerman, 2004). An important means to investigate disease-related changes that involve non-heme iron deposition is quantitative, non-invasive assessment of iron distributions in brain tissue of patients and normal subjects. Regional quantitative information of iron content is also of utmost importance for early clinical diagnosis, monitoring of treatment, and disease progression (for reviews, see Schenck and Zimmerman, 2004, Utter and Basso, 2008).

Several MR-based techniques have been proposed for detecting and quantifying iron in the brain non-invasively (reviewed by Haacke et al. (2005)). Typically, these approaches exploit the paramagnetic properties of iron, for example, macroscopic field perturbations due to spatial variations in the volume susceptibility, induced by the local tissue iron load. Gradient echo (GRE) phase images allow determination of the local Larmor resonance frequency shifts due to these magnetic field perturbations within tissues. GRE phase is, therefore, an indirect measure of the volume susceptibility distribution in tissue (Ogg et al., 1999). However, the relation between magnetic susceptibility and field perturbation is non-local and depends on both the spatial susceptibility distribution and its orientation with respect to the main magnetic field (Schäfer et al., 2009). In practice, even anatomical visualization of deep gray matter nuclei with elevated iron load in vivo has been proven challenging with conventional MR imaging due to these effects (Deoni and Catani, 2007).

A novel technique, referred to as quantitative susceptibility mapping (QSM), overcomes the non-local relationship between magnetic field perturbation and susceptibility. Here, GRE phase data are processed to extract the underlying magnetic susceptibility distribution, resulting in a novel quantitative anatomical contrast of an intrinsic physical tissue property. Magnetic susceptibility maps are expected to enable quantitative examination of changes in magnetic tissue properties in vivo due to, e.g., elevated iron storage or local blood oxygen saturation.

Recently, ambitious approaches were presented for solving this challenging inverse problem to generate susceptibility maps (de Rochefort et al., 2010, Schweser et al., 2010a, Wharton et al., 2010, and references therein). However, most of the presented studies demonstrated fairly poor fidelity, were performed ex vivo, required excessive data acquisition and prior knowledge about the tissue composition or were restricted to small sections of the brain. One reason for these limitations is the lack of artifact-free pre-processing of GRE phase information, which is essential to resolve inherent phase aliasing and to suppress strong superimposed background contributions induced by air–tissue or tissue–bone interfaces (Chen et al., 2010, Hammond et al., 2008, Schweser et al., 2010d, and references therein).

Our study aims to overcome this obstacle with two novel techniques to improve the quality of the input data to the ill-posed QSM problem. The first technique is a sophisticated approach to determine voxels with unreliable phase information using the condition of spatiotemporal smoothness of the GRE phase data. The second technique, called Sophisticated Harmonic Artifact Reduction for Phase data (SHARP), eliminates background field contributions from GRE phase data without remaining artifacts. As a direct application, the QSM framework is used to investigate the potential of magnetic susceptibility maps for anatomical imaging and iron quantification. For this purpose, a magnetic susceptibility map is calculated without regularization based on data acquired from a single volunteer with his head in different orientations with respect to the main magnetic field (multi angle acquisition, MAA; a method proposed by Liu et al. (2009) and Marques and Bowtell (2005), also referred to as COSMOS) to serve as gold standard for tissue magnetic susceptibility. This map is then compared to susceptibility maps that were calculated from each single angle acquisition (SAA) of the same volunteer using a recently published algorithm (Schweser et al., 2010a). Furthermore, magnetic susceptibility values of different anatomical brain regions are presented that were obtained from volunteer data (n = 5) using the SAA approach. We also explore the potential of QSM for anatomical imaging and measurement of venous blood oxygen saturation level. Since little attention has been paid to the sources besides iron that may affect magnetic susceptibility, we investigate the potential of magnetic susceptibility maps to assess local tissue iron concentration. Myelin, for example, is a spiral membranous structure that is tightly wrapped around axons (van der Knaap and Valk, 2005). It is supposed to be diamagnetic due to its high lipid content (Duyn et al., 2007) and, thus, has an opposite effect on the voxel susceptibility compared to iron. Hence, the presence of myelin may result in an underestimation of tissue iron concentrations based on magnetic susceptibility effects. We examine the contribution of myelin to the magnetic susceptibility of brain parenchyma by exploiting magnetization transfer (MT) images (Gringel et al., 2009). A correction scheme is demonstrated to take into account myelin-induced magnetic susceptibility contributions. Furthermore, anisotropy effects of the highly ordered myelin fiber bundles on the GRE phase that have recently been discussed by several authors (He and Yablonskiy, 2009, Lee et al., 2010, Liu, 2010) are investigated based on the calculated susceptibility maps and the SHARP corrected phase images. Finally, we analyze direct assessment of pre-processed GRE phase data to directly quantify the amount of iron, since several authors rely on this approach (e.g., (Haacke et al., 2009, Xu et al., 2008, Zhang et al., 2010)), and compare the results to findings based on magnetic susceptibility maps.

Section snippets

Theory

The following four subsections introduce the fundamental mathematical ideas behind the QSM framework that is described in detail in the Materials and methods section. The first subsection briefly recapitulates the relation between GRE phase and the underlying magnetic susceptibility distribution, which is essential for QSM and has already been published elsewhere. The physical and mathematical concepts of the SHARP method and the method for determining phase data reliability based on

Data acquisition

All experiments were approved by the local ethics committee and informed written consent was obtained from each recruited subject. A total of five healthy volunteers (3 female, 2 male; 21–26 years) were involved in the study. Axial high-resolution whole-brain GRE phase data were acquired from all volunteers on a 3T whole body MRI scanner (Tim Trio, Siemens Medical Solutions, Erlangen, Germany) using a 12-channel receive head-matrix coil. For acquisition of the GRE data a special dual-echo

Phase processing using the SHARP algorithm

Representative axial and sagittal IFS images computed with the SHARP algorithm are presented in Fig. 3. These images clearly show the IFS induced by the underlying magnetic susceptibility distribution. Although reflecting a non-local effect, the IFS images provide anatomic information in high detail, enabling delineation of many different brain structures and sub-structures.

Fig. 4 illustrates the results obtained with the different phase processing algorithms. Fig. 4b and e shows IFS images of

Discussion

The current study presented a framework for quantitative imaging of magnetic tissue properties, and investigated its applicability for non-invasive estimation of tissue iron concentrations. The obtained magnetic susceptibility maps reveal a novel quantitative contrast, which directly reflects rotation-invariant intrinsic magnetic tissue properties. Susceptibility maps reproduce the underlying anatomy more directly than IFS maps, which suffer from non-locality and orientation dependency. Two

Conclusions

The proposed QSM framework provides reproducible and orientation-independent magnetic susceptibility maps. This novel quantitative anatomical contrast may be instrumental to unravel structural and physiological composition of tissue in vivo. The findings suggest that susceptibility contrast, and therewith also phase contrast, are not only linked to storage iron concentration but are also significantly influenced by other sources, e.g., myelin. Consequently, susceptibility contrast may serve as

Acknowledgments

This research was funded by a Carl Zeiss Foundation Dissertation Fellowship for Ferdinand Schweser and a grant by the German Research Foundation (RE 1123/9-1). We are grateful to Dr. Andreas Schäfer and to Dr. John Schenck for valuable and helpful discussions, and Marie Atterbury for proofreading the manuscript. Furthermore, we would like to thank Moritz Hütten for implementing the MAA extension to the QSM algorithm.

References (115)

  • A. Schäfer et al.

    Using magnetic field simulation to study susceptibility-related phase contrast in gradient echo MRI

    Neuroimage

    (2009)
  • K. Shmueli et al.

    Magnetic susceptibility mapping of brain tissue in vivo using MRI phase data

    Magn. Reson. Med.

    (2009)
  • H.S. Abdul-Rahman et al.

    Fast and robust three-dimensional best path phase unwrapping algorithm

    Appl. Opt.

    (2007)
  • A.M. Abduljalil et al.

    Enhanced gray and white matter contrast of phase susceptibility-weighted images in ultra-high-field magnetic resonance imaging

    J. Magn. Reson. Imaging

    (2003)
  • J. Annese

    A myelo-architectonic method for the structural classification of cortical areas

    Neuroimage

    (2004)
  • J.S. Becker et al.

    Imaging of copper, zinc, and other elements in thin section of human brain samples (hippocampus) by laser ablation inductively coupled plasma mass spectrometry

    Anal. Chem.

    (2005)
  • C. Bouras et al.

    A laser microprobe mass analysis of brain aluminum and iron in dementia pugilistica: comparison with Alzheimer's disease

    Eur. Neurol.

    (1997)
  • J.C. Chen et al.

    MR of human postmortem brain tissue: correlative study between T2 and assays of iron and ferritin in Parkinson and Huntington disease

    Am. J. Neuroradiol.

    (1993)
  • Z. Chen et al.

    An optimised framework for reconstructing and processing MR phase images

    Neuroimage

    (2010)
  • Y.-C.N. Cheng et al.

    Limitations of calculating field distributions and magnetic susceptibilities in MRI using a Fourier based method

    Phys. Med. Biol.

    (2009)
  • S.C.K. Chu et al.

    Bulk magnetic susceptibility shifts in NMR studies of compartmentalized samples: use of paramagnetic reagents

    Magn. Reson. Med.

    (1990)
  • J.F. Collingwood et al.

    In situ characterization and mapping of iron compounds in Alzheimer's disease tissue

    J. Alzheimers Dis.

    (2005)
  • A. Deistung et al.

    ToF-SWI: simultaneous time of flight and fully flow compensated susceptibility weighted imaging

    J. Magn. Reson. Imaging

    (2009)
  • S.C.L. Deoni et al.

    Visualization of the deep cerebellar nuclei using quantitative T1 and ρ magnetic resonance imaging at 3 Tesla

    Neuroimage

    (2007)
  • L. de Rochefort et al.

    Quantitative susceptibility map reconstruction from MR phase data using Bayesian regularization: validation and application to brain imaging

    Magn. Reson. Med.

    (2010)
  • G. Deville et al.

    NMR multiple echoes observed in solid 3He

    Phys. Rev. B

    (1979)
  • D.T. Dexter et al.

    Alterations in the levels of iron, ferritin and other trace metals in parkinson's disease and other neurodegenerative diseases affecting the basal ganglia

    Brain

    (1991)
  • W. Dickinson

    The time average magnetic field at the nucleus in nuclear magnetic resonance experiments

    Phys. Rev.

    (1951)
  • M. Fukunaga et al.

    Layer-specific variation of iron content in cerebral cortex as a source of MRI contrast

    Proc. Natl Acad. Sci. U. S. A.

    (2010)
  • M. Gerlach et al.

    Altered brain metabolism of iron as a cause of neurodegenerative diseases?

    J. Neurochem.

    (1994)
  • X. Golay et al.

    Measurement of tissue oxygen extraction ratios from venous blood T(2): increased precision and validation of principle

    Magn. Reson. Med.

    (2001)
  • T. Gringel et al.

    Optimized high-resolution mapping of magnetization transfer (MT) at 3 Tesla for direct visualization of substructures of the human thalamus in clinically feasible measurement time

    J. Magn. Reson. Imaging

    (2009)
  • E.M. Haacke et al.

    Establishing a baseline phase behavior in magnetic resonance imaging to determine normal vs. abnormal iron content in the brain

    J. Magn. Reson. Imaging

    (2007)
  • E.M. Haacke et al.

    Magnetic properties of tissues: theory and measurement

  • E.M. Haacke et al.

    Imaging iron stores in the brain using magnetic resonance imaging

    Magn. Reson. Imaging

    (2005)
  • E.M. Haacke et al.

    Iron stores and cerebral veins in MS studied by susceptibility weighted imaging

    Int. Angiol.

    (2010)
  • E.M. Haacke et al.

    In vivo validation of the BOLD mechanism: a review of signal changes in gradient echo functional MRI

    Int. J. Imag. Syst. Technol.

    (1995)
  • E.M. Haacke et al.

    Characterizing iron deposition in multiple sclerosis lesions using susceptibility weighted imaging

    J. Magn. Reson. Imaging

    (2009)
  • E.M. Haacke et al.

    Correlation of putative iron content as represented by changes in R2* and phase with age in deep gray matter of healthy adults

    J. Magn. Reson. Imaging

    (2010)
  • E.M. Haacke et al.

    Susceptibility weighted imaging (SWI)

    Magn. Reson. Med.

    (2004)
  • B. Hallgren et al.

    The effect of age on the non-haemin iron in the human brain

    J. Neurochem.

    (1958)
  • B. Hallgren et al.

    The non-haemin iron in the cerebral cortex in Alzheimer's disease

    J. Neurochem.

    (1960)
  • K.E. Hammond et al.

    Development of a robust method for generating 7.0 T multichannel phase images of the brain with application to normal volunteers and patients with neurological diseases

    Neuroimage

    (2008)
  • P.M. Harrison et al.

    The ferritins: molecular properties, iron storage function and cellular regulation

    Biochim. Biophys. Acta (BBA) - Bioenerg.

    (1996)
  • X. He et al.

    Quantitative BOLD: mapping of human cerebral deoxygenated blood volume and oxygen extraction fraction: default state

    Magn. Reson. Med.

    (2007)
  • X. He et al.

    Biophysical mechanisms of phase contrast in gradient echo MRI

    Proc. Natl Acad. Sci. U. S. A.

    (2009)
  • E. Hernández Torres et al.

    The relationship between susceptibility weighted phase and white matter fiber orientation

  • F.T. Hong et al.

    Magnetic anisotropy and the orientation of retinal rods in a homogeneous magnetic field

    Proc. Nat. Acad. Sci. U. S. A.

    (1971)
  • K. Hopp et al.

    Brain iron detected by SWI high pass filtered phase calibrated with synchrotron X-ray fluorescence

    J. Magn. Reson. Imaging

    (2010)
  • D.I. Hoult

    The principle of reciprocity in signal strength calculations?A mathematical guide

    Concepts Magn. Reson.

    (2000)
  • Cited by (596)

    View all citing articles on Scopus

    Author contributions: F.S., A.D., and J.R.R. designed the research; F.S. performed the research; F.S., A.D., and B.W.L. contributed new reagents/analytic tools; F.S. analyzed the data; and F.S., A.D., and J.R.R. wrote the paper.

    View full text