Yoshitaka Bito¹, Yuko Kawai², Koji Hirata¹, Toshihiko Ebisu³, Yosuke Otake¹, Satoshi Hirata¹, Toru Shirai¹, Yoshihisa Soutome¹, Hisaaki Ochi¹, Masahiro Umeda², Toshihiro Higuchi⁴, and Chuzo Tanaka^4
1Central Research Laboratory, Hitachi, Ltd., Kokubunji-shi, Tokyo, Japan, ²Medical Informatics, Meiji University of Integrative Medicine, Nantan-shi, Kyoto, Japan, ³Neurosurgery, Nantan General Hospital, Nantan-shi, Kyoto, Japan, ⁴Neurosurgery, Meiji University of Integrative Medicine, Nantan-shi, Kyoto, Japan

Diffusion tensor spectroscopic imaging (DTSI), using diffusion-weighted echo-planar spectroscopic imaging with a pair of bipolar diffusion gradients (DW-EPSI with BPGs), was applied to measure diffusion tensor images (DTIs) of N-acetylaspartate (NAA), creatine (Cr), and choline (Cho) in normal rat brains. The DTIs of NAA and Cr were successfully measured, however, the DTI of Cho was deteriorated by gradually decreasing the signal of Cho. The measured DTIs of NAA and Cr are very similar to the DTI of water in most brain regions but show differences in the detail, i.e., cortex and corpus callosum. These results suggest that this DTSI technique is effective in investigating microstructures of tissue.

Stefan Posse^1,2, Elena Ackley¹, Jingjing Michele Zhang³, and Tongsheng Zhang^1
1Neurology, University of New Mexico, Albuquerque, NM, United States, ²Electrical and Computer Engineering, University of New Mexico, Albuquerque, NM, United States, ³Department of Biomedical Physics, David Geffen School of Medicine at UCLA, University of California, Los Angeles, CA, United States

Diffusion tensor spectroscopic imaging (DTSI) in human brain was developed using high-speed proton-echo-planar-spectroscopic-imaging (PEPSI) with ECG gating, correction of movement-related phase errors using spatially localized echo-planar navigator signal acquisition and compensation of the variability in T1-saturation, which results from cardiac gating. DTSI was implemented on a 3 Tesla clinical scanner equipped with a 32-channel array coil. Data in a phantom and in 3 healthy volunteers were acquired using a 32x32 spatial matrix, 1 cc voxel size, bmax = 1734 s/mm2, and 6 gradient directions. Apparent diffusion coefficients and fractional anisotropy values of Cho, Cr, NAA and tissue water measured in a supraventricular slice were in the ranges reported in previous studies using single voxel methods and using diffusion tensor MRI. 3D DTSI is currently under development.

Angèle Lecocq¹, Yann Le Fur¹, Andrew A. Maudsley², Sulaiman Sheriff², Mohammad Sabati², Virginie Callot³, Monique Bernard⁴, Maxime Guye¹, and Jean-Philippe Ranjeva^1
1CRMBM, CNRS, Aix-Marseille Univ, MARSEILLE, France, ²Department of radiology, Miller School of Medicine, University of Miami, MIAMI, Florida, United States, ³Aix-Marseille University, MARSEILLE, France, ⁴Aix-Marseille Université, MARSEILLE, France

One major MRSI limitation is the artifact related to magnetic susceptibility effects leading to metabolite map distortions and signal drop out. The use of a short echo time helps to reduce signal loss and minimize phase differences. This susceptibility artifact could also be minimized by the choice of the MRSI data orientation. We aimed here to evaluate the impact of MRSI orientation on quality of whole brain of major metabolites maps between acquisitions performed along the AC-PC line and along the AC-PC+15°line using a recent whole brain MRSI sequence called Echo Planar Spectroscopy Imaging (EPSI) with short echo time (20ms).

Patrick J. Bolan¹, Gregory J. Metzger¹, Steen Moeller¹, and Michael Garwood^1
1Radiology, CMRR, University of Minnesota, Minneapolis, MN, United States

This work explores the feasibly of using reconstructions based on Spectral Localization by Imaging (SLIM) to separate the spectral signals from the aqueous and lipid compartments in the breast. Fat fraction and B₀ maps derived from high-resolution water-fat imaging are used to create a geometric model of the breast. A conventional CSI spectral data set is then reconstructed using SLIM and the geometric model to create discrete spectra for the water and lipid compartments. The feasibility of this approach is demonstrated through simulation and in vivo measurements.

Rajakumar Nagarajan¹, Daniel J.A. Margolis¹, Stevan S. Raman¹, Nidhi Jain¹, Robert E. Reiter², and M. Albert Thomas^1
1Radiological Sciences, University of California Los Angeles, Los Angeles, CA, United States, ²Urology, University of California Los Angeles, LOS ANGELES, CA, United States

Due to its prevalence in the male population as well as its unpredictable clinical course, early detection and diagnosis of prostate cancer have become a priority for many health care professionals. The one-dimensional (1D) MRS combined with two-dimensional (2D) or three-dimensional (3D) spatial encoding is popularly known as MR Spectroscopic Imaging (MRSI) and a drawback of 1D MRS is the overlap of metabolite signals due to the limited spectral dispersion. The inherent acceleration in the echo-planar spectroscopic imaging (EPSI) can be further accelerated using non-uniform undersampling (NUS) imposed on incremented spatial and spectral dimensions of the 4D EP-JRESI sequence. The CS-reconstructed data using the split-Bregman non-linear reconstruction algorithm showed an excellent reproduction of multiple 2D J-resolved spectra with the 2D diagonal and cross peaks at the expected locations. The 25% undersampled CS-reconstructed EP-JRESI data using the endorectal “receive” coil showed excellent 2D J-resolved spectral quality demonstrating the clinically acceptable time (6-12 minutes) with 1ml voxel resolution.

Itthi Chatnuntawech¹, Berkin Bilgic¹, and Elfar Adalsteinsson^2,3
1EECS, Massachusetts Institute of Technology, Cambridge, MA, United States, ²EECS, MIT, Cambridge, MA, United States, ³Harvard-MIT Division of Health Sciences and Technology, Cambridge, MA, United States

In this work, a two-step model-based method that leads to an accurate reconstruction from undersampled spectroscopic imaging data is proposed. This method takes advantage of a fast water reference scan to estimate a subset of (non-linear) unknowns, leaving only a few, linear unknowns to be determined in the next step. Then, a regularized optimization problem with prior knowledge on the structure of the data is formulated to reconstruct the spectroscopic imaging data. This method reduces acquisition time by undersampling while preserving high reconstruction quality. The proposed method yields significantly lower root mean square error than that of the conventional method which finds the minimum-norm solution without regularization.

Chris Hanstock¹ and Myrlene Gee^1
1Biomedical Engineering, University of Alberta, Edmonton, Alberta, Canada

A challenge for serial MRS studies is volume placement reproducibility. Typically, this has relied on visual registration which is prone to significant operator bias. Therefore the consistency of placement can be relatively poor both intra- and inter-centre. As a result the comparison of data from study to study has been prone to huge variability with studies for the same volunteer cohort. Our new Talairach Atlas-based automated protocol allows the exact volume placement prior to the scan and then exact co-registration for subsequent scan sessions. Moreover this procedure can be utilized across scanner platforms for multi-centre studies giving identical volume placement.

Dinesh K. Deelchand¹, Xiaoping Wu¹, Pierre-Gilles Henry¹, and Pierre-Francois Van de Moortele^1
1Center for Magnetic Resonance Research, University of Minnesota, Minneapolis, MN, United States

In this study we demonstrate with phantom experiments the feasibility of selectively exciting arbitrary shaped voxel using 2D-RF pulse design (segmented radial k-space trajectory) with parallel RF excitation at 16.4 T while achieving sufficient excitation bandwidth for 1H MR spectroscopy. A novel simulation guided segment-wise RF pulse design is introduced to tackle the issue of gradient imperfection and is shown to preserve the excitation quality in the presence of gradient waveform deviations.

Meng Gu¹, Priti Balchandani^1,2, Sulaiman Sheriff³, Mohammad Sabati³, Daniel Spielman¹, and Andrew Maudsley^3
1Radiology, Stanford University, Stanford, CA, United States, ²Radiology, Mount Sinai School of Medicine, New York, NY, United States, ³Radiology, University of Miami, Miami, FL, United States

MR spectroscopic imaging with localization schemes using conventional RF pulses suffer from B1 inhomogeneity and chemical shift localization error due to bandwidth limitations. Adiabatic refocusing RF pulses have been introduced to overcome the B1 inhomogeneity. However, conventional hyperbolic secant adiabatic refocusing pulses are limited by long pulse-width, prohibiting acquisition with short TE. Here, we present a localization scheme using adiabatic SLR refocusing pulses with broad bandwidth and short pulse-width. This localization scheme was incorporated into a volumetric MRSI sequence with echo planar readout. In vivo study showed high quality spectra and metabolite maps were obtained throughout the brain.

Tom W.J. Scheenen^1,2, Pascal Sati³, Steve Li⁴, Jun Shen⁴, and Daniel S. Reich^3
1Radiology, Radboud University Nijmegen Medical Center, Nijmegen, Netherlands, ²Lab of Functional and Molecular Imaging, National Institute of Neurologiocal Disorders and Stroke, National Institutes of Health, Bethesda, MD, United States, ³Translational Neuroradiology Unit, Neuroimmunology Branch, National Institute of Neurological Disorders and Stroke, National Institutes of Health, Bethesda, MD, United States, ⁴MRS Core Facility, National Institute of Mental Health, National Institutes of Health, Bethesda, MD, United States

³¹P-MRSI of the human brain with a dual ¹H-³¹P TxRx volume coil at 7T was explored in five healthy volunteers. The combined coil setup allowed ¹H irradiation of water during TR to enhance the signal of ³¹P metabolites. This Nuclear Overhauser Enhancement was explored in phantoms and quantified in several deep brain areas of healthy volunteers, allowing the acquisition of good quality spectra of the whole brain with a spatial resolution of 10.6 cc in 21 minutes.

Sandro Romanzetti¹, Daniel P. Fiege¹, and Nadim Jon Shah^1,2
1Institute of Neuroscience and Medicine - 4, Forschungszentrum Jülich, Jülich, Germany, ²JARA - Faculty of Medicine, RWTH Aachen University, Aachen, Germany

In this work we present a novel centric 3D sequence called 3DTWIRL. This sequence is based on Twisted Projection Imaging (TPI) technique and as such it combines the ultra-short echo time capability of projection imaging (PI) with a constant sampling density of k-space. The new method, however, requires the calculation of only the innermost TPI cone. The k-space is then sampled by rotating this cone by the appropriate polar and azimuthal angles. Initial results show that the 3DTWIRL sequence has a reduce sensitivity to off-resonances and ease of implementation.

Keith Wachowicz^1,2, Amr Heikal³, and B. Gino Fallone^2,4
1Oncology, University of Alberta, Edmonton, Alberta, Canada, ²Medical Physics, Cross Cancer Institute, Edmonton, Alberta, Canada, ³Physics, University of Alberta, Edmonton, Alberta, Canada, ⁴Physics and Oncology, University of Alberta, Edmonton, Alberta, Canada

The implications of compressed sensing (CS) reconstruction in spectroscopic imaging on spatial resolution have not been quantitatively investigated in the literature. Modulation transfer analyis in this work is used to demonstrate that dramatic loss of modulation transfer across the range of spatial frequencies can occur with CS. A modification to the CS algorithm (CMaCS) involving further constraints on the randomization of k-space and mapping of unpaired conjugate data is introduced and shown to recover much of this lost modulation.

Peter Adany¹, Phil Lee², and In-Young Choi^3
1Hoglund Brain Imaging Center, University of Kansas Medical Center, Kansas City, KS, United States, ²Department of Molecular & Integrative Physiology, Hoglund Brain Imaging Center, University of Kansas Medical Center, Kansas City, KS, United States, ³Department of Neurology, Hoglund Brain Imaging Center, University of Kansas Medical Center, Kansas City, KS, United States

An extended non-Fourier based spectral localization technique is proposed for fast and accurate measurements of neurochemicals from non-rectangular, arbitrary shaped brain regions of interest (e.g., gray and white matter) in the human brain in vivo. The proposed spectral localization by imaging (SLIM)-based technique takes into account both inhomogeneous coil sensitivity with the use of multiple-channel receiver coils and B0 inhomogeneity. Thus, full recovery of the MRS signal quality, which is otherwise compromised, is achievable to assess gray and white matter differences of metabolite concentrations with a reduced number of phase-encoding steps and cross-compartmental contamination.

Itthi Chatnuntawech¹, Berkin Bilgic¹, Borjan Gagoski², Trina Kok¹, Audrey Peiwen Fan¹, and Elfar Adalsteinsson^1,3
1EECS, Massachusetts Institute of Technology, Cambridge, MA, United States, ²Fetal-Neonatal Neuroimaging & Developmental Science Center, Boston Children's Hospital, Harvard Medical School, Boston, MA, United States, ³Harvard-MIT Division of Health Sciences and Technology, Cambridge, MA, United States

Specific physiological abnormalities could be detected by irregular change of metabolite concentration in specific brain regions. The combination between fully sampled spectroscopic imaging data and segmented structural image has been used to estimate metabolite value at each voxel. This abstract presents an N-compartment-model method with polynomial masks to obtain metabolite maps from undersampled spectroscopic imaging data. With the assumption that metabolite value within the same tissue type is slowly varying, the information of tissue boundaries is obtained from segmented structural image. Then, a regularized reconstruction with priors is formulated to reconstruct the metabolite maps. By acquiring only a subset of k-space samples, the acquisition process is sped up, while high reconstruction quality is retained via prior knowledge of tissue boundary and structure of data.

Hamed Mojahed^1,2, Fernando Arias-Mendoza², and Truman R. Brown^3
1Department of Biomedical Engineering, Columbia University, New York, NY, United States, ²Department of Radiology, Columbia University, New York, NY, United States, ³Department of Radiology and Radiological Science, Center for Biomedical Imaging, Medical University of South Carolina, Charleston, SC, United States

A new three dimensional zero J-modulation echo planar chemical shift imaging (3D ZJ-EPSI) pulse sequence is implemented on a conventional 3T MR scanner. Water suppressed ¹H spectra of the human brain using a CHESS pulse (RF 80°-80°-145°) is acquired using a Sense head coil in quadrature mode. Lipid suppression is achieved using ten saturation bands around the skull. Spatial resolution of 7.5×7.5×6 mm³ is achieved in 10:28 min with TR=1500 ms and TE<1 ms. Having an echoless data and zero J-modulation increases the SNR, improves the T₂*, and makes it easier to quantify spectra in the fast 3D ZJ-EPSI sequence.

Ivica Just Kukurová^1,2, Ladislav Valkovic^1,3, Martin Gajdosik¹, Martin Krssák⁴, Stephan Gruber¹, Tibor Liptaj², Siegfried Trattnig¹, and Marek Chmelík^1
1MR Centre of Excellence, Department of Radiology, Medical University of Vienna, Vienna, Austria, ²Department of NMR and MS, Faculty of Chemical and Food Technology, Slovak University of Technology, Bratislava, Slovakia, ³Department of Imaging Methods, Institute of Measurement Science, Slovak Academy of Sciences, Bratislava, Slovakia, ⁴Department of Internal Medicine III, Medical University of Vienna, Vienna, Austria

The study aim was to simultaneously assess the composition of fatty acids from the extramyocellular lipids, bone marrow and subcutaneous adipose tissue in the calf at 7T and analyze their saturation profiles. Right calf of four volunteers was scanned using a 2D-CSI sequence with FID and long echo-time acquisition. Ratios of unsaturated fatty-acids (UFA) to CH3 group, polyunsaturated fatty-acids (PUFA) to CH3 group, and PUFA to UFA+PUFA were calculated in: subcutaneous fat, bone marrow and calf muscles. The new 2D-CSI sequence was found suitable for detecting lipids in various tissues for their composition analysis with no extra time needed.

Rajakumar Nagarajan¹, Christian Roberts², Cathy C. Lee^3,4, Theodore Hahn^3,4, and M. Albert Thomas^1
1Radiological Sciences, University of California Los Angeles, Los Angeles, CA, United States, ²Exercise and Metabolic Disease Research Laboratory, School of Nursing, University of California Los Angeles, Los Angeles, California, United States, ³Department of Medicine, University of California Los Angeles, Los Angeles, California, United States, ⁴Geriatrics, Research, Education and Clinical Center (GRECC ), VA Greater Los Angeles Healthcare System, Los Angeles, California, United States

Type 2 diabetes (T2D) has progressed into a major cause of preventable death in recent decades, expected to reach 21 million cases in the U.S. in 2010 with an estimated additional 7 million undiagnosed. In obese individuals with insulin resistance (IR) and in patients with T2D, skeletal muscle insulin-stimulated glucose uptake is markedly blunted. Intramyocellular lipids (IMCL) in skeletal muscle can be measured non-invasively by single voxel (SV) based 1H magnetic resonance spectroscopy (MRS) or multi-voxel based magnetic resonance spectroscopic imaging (MRSI). However, SV-based MRS and conventional MRSI recording 1D MRS suffer from severe overlap of the IMCL and extramyocellular lipid (EMCL) signals hampering accurate quantitation of saturated and unsaturated portions of IMCL. Also, the total time required for SV-MRS acquisition in multiple voxels would be impractical. We have evaluated a novel four-dimensional multi-echo echo-planar spectroscopic imaging (MEEP-COSI) to record multi-voxel based 2D MRS in T2D, obese and healthy subjects. Significantly increased intramyocellular lipids (IMCL) and extramyocellular lipids (EMCL), and decreased IMCL unsaturation indices and choline were observed in this pilot study.

Richard Anthony Edward Edden^1,2, Jamie Near³, C. John Evans⁴, Nicolaas A. J. Puts^2,5, and Peter B. Barker^1,2
1Russell H. Morgan Department of Radiology and Radiological Science, Johns Hopkins University, Baltimore, MD, United States, ²F. M. Kirby Center for Functional Brain Imaging, Kennedy Krieger Institute, Baltimore, MD, United States, ³Douglas Mental Health University Institute and Department of Psychiatry, McGill University, Montreal, Quebec, Canada, ⁴CUBRIC, School of Psychology, Cardiff University, Cardiff, Wales, United Kingdom, ⁵Russell H. Morgan Department of Radiology and Radiological Science, The Johns Hopkins University School of Medicine, Baltimore, MD, United States

In this abstract, we investigate factors that influence the multiplet pattern observed in J-difference editing of GABA. Density matrix simulations were applied to investigate the shape of the 3 ppm multiplet as a function of the editing sequence’s slice-selective refocusing pulse properties, in particular bandwidth, transition width, and flip angle. Conclusion: The 3 ppm GABA multiplet pattern observed in the MEGA-PRESS experiment depends quite strongly on the properties of the slice selective refocusing pulses used. Under some circumstance the central peak can be quite large; this does not necessarily indicate inefficient editing, or a subtraction artifact, but should be recognized as a property of the pulse sequence itself.

Assaf Tal¹ and Oded Gonen^2
1Radiology, NYU Langone School of Medicine, New York, NY - New York, United States, ²Radiology, New York University, New York, NY - New York, United States

B0 field instabilities lead to sizable localization errors in prolonged phase-encoded spectroscopic imaging, in contrast to line-broadening observed in single-voxel studies. We demonstrate these errors in vivo and in a phantom, and propose an efficient and quick way to correct for them.

Jullie W. Pan¹ and Hoby P. Hetherington^2
1Neurosurgery, Yale University School of Medicine, New Haven, CT, United States, ²Yale University, New Haven, CT, United States

In the brain, the amygdala is believed to function in the processing of emotion and memory, and has been shown to be dysfunctional in many psychological and psychiatric disorders such as post-traumatic stress and bipolar disorders. The MR evaluation of the amygdala can be challenging however, because of its inferior temporal-frontal location lateral to the cavernous sinus, and is particularly problematic for ultra-high field studies. In this study we describe implementation of high degree shimming to perform J-refocused spectroscopic imaging using in the region of the midbrain and amygdala at 7T.

Ouri Cohen^1,2, Assaf Tal¹, and Oded Gonen^3
1Center for Biomedical Imaging, NYU, New York, NY, United States, ²Biomedical Engineering, Columbia University, New York, NY, United States, ³Center for Biomedical Imaging, New York University, New York, NY, United States

Transverse Hadamard spectroscopic imaging (T-HSI) has been shown to overcome the intrinsic signal-to-noise loss and increased bleed of chemical shift imaging by virtue of its approaching-ideal point-spread-function. However, because it uses a superposition of pulses, it is less suited for higher fields where the available B1 is smaller. In this work we demonstrate a method that overcomes this limitation and allows maintaining the benefits of T-HSI despite the higher (3T) field strength.

Meijin Lin¹, Anand Kumar¹, and Shaolin Yang^1,2
1Department of Psychiatry, University of Illinois at Chicago, Chicago, IL, United States, ²Department of Radiology, University of Illinois at Chicago, Chicago, IL, United States

Two-dimensional (2D) localized chemical shift correlated spectroscopy (L-COSY) is one of the simplest and the most useful methods applied for 2D magnetic resonance spectroscopy (MRS). L-COSY spectra provide more separated peaks with cross-peaks for spectral quantification. Keeping the cross-peak intensities from attenuation is critical for reliable and accurate quantification of metabolites. In this abstract, however, we demonstrated that using limited-bandwidth (BW) of radiofrequency (RF) pulses for slice selection might attenuate the intended cross-peaks in L-COSY spectra. We further demonstrated using semi-localization by adiabatic selective refocusing (sLASER) pulses for localization in COSY can maintain the intended cross-peaks from attenuation.

Adrian Tsang¹, Rob Stobbe¹, and Christian Beaulieu^1
1Biomedical Engineering, University of Alberta, Edmonton, Alberta, Canada

Selective detection of sodium signal from nuclei in ordered environments that exhibit non-zero residual quadrupole interactions is possible with a double-quantum magic-angle (DQ-MA) sequence. We show that DQ-MA signal from sodium is only detected in xanthan gum (ordered macromolecular environment) and not in either saline or 4% agar gel (disordered). We then demonstrate the first DQ-MA images of human brain acquired in 18 minutes at 4.7T in 3 healthy volunteers. This selective sodium measurement may provide novel insight into alterations of the microscopic ordered tissue microstructure of the brain.

Yi Sun^1,2, Djaudat Idiyatullin³, Donald R. Nixdorf^4,5, X.Frank Walboomers⁶, Egbert Oosterwijk², Michael Garwood³, and Arend Heerschap^1
1Radiology, Radboud University Medical Center Nijmegen, Nijmegen, Netherlands, ²Urology, Radboud University Medical Center Nijmegen, Nijmegen, Netherlands, ³Center for Magnetic Resonance Research, University of Minnesota, Minnespolis, MN, United States, ⁴Department of of Diagnostic & Biological Sciences, School of Dentistry, University of Minnesota, Minneapolis, MN, United States, ⁵Department of Neurology, Medical School, University of Minnesota, Minneapolis, MN, United States, ⁶Dentistry, Radboud University Medical Center Nijmegen, Nijmegen, Netherlands

Dental X-ray has several disadvantages, including an increased risk for meningioma, necessitating to consider other imaging options. It has been demonstrated that 1H MRI by the SWIFT technique can offer high-resolution diagnostic images of human teeth. However, dental MRI focuses on 1H, while the main content of human tooth is crystalline calcium phosphate. Therefore we explored the feasibility of 31P SWIFT and ZTE MRI of teeth to obtain direct information of phosphate constituents. We demonstrate that this is possible and in combination with 1H MRI can identify the anatomy of tooth and has the potential to determine bone phosphate density.