Pulse Sequences

Hall D                                   Tuesday 13:30-15:30                                                                                                                                             

13:30         3058.     Whole Orbit Soft Tissue Deformation Acquired by Accelerated 3D CSPAMM Tagging During Eye Motion

Computer 15

Marco Piccirelli1,2, Andrea Kaspar Rutz1, Oliver Bergamin2, Peter Boesiger1, Roger Luechinger1

1University and ETH Zurich, Zurich, Switzerland; 2University Hospital Zurich, Switzerland

In complex orbital mechanical disorders, a better comprehension of ocular motion dynamic is needed. The deformation pattern within extraocular muscles (EOM) and orbital connective tissues has not been understood yet. In this work, an accelerated truly three-dimensional tagging acquisition method is proposed enabling the assessment of motion information with whole orbit coverage in a scantime allowing good motion reproducibility. A reduced field-of-view method was incorporated and 3D data sets were acquired sequentially with line tag preparation in each of the three spatial dimensions. Data were post-processed with 3D peak-combination HARP. Tissues within the orbit could be reliably tracked and characterized.


14:00         3059.     Orthogonal TrueFISP Acquisitions Using Paired Reverse Centric Phase Encoding

Computer 15

Jamal Jon Derakhshan1,2, Mark A. Griswold1,2, Jeffrey L. Sunshine2, Jeffrey L. Duerk,12

1Case Western Reserve University, Cleveland, Ohio, USA; 2University Hospitals Case Medical Center, Cleveland, Ohio, USA

Paired reverse centric phase encoding is presented as a way to eliminate saturation banding in interleaved orthogonal TrueFISP imaging. Simulation results demonstrate significant (> 2x) reduction of orthogonal plane saturation artifacts across various base resolutions, flip angles and tissue types. Phantom imaging demonstrates the ability to eliminate both saturation and eddy current artifacts by pairing the reverse centric lines. Human in vivo abdominal scout imaging demonstrates the utility of the new acquisition strategy. Application to interventional MRI is demonstrated by presenting three orthogonal images acquired during real-time guidance of an RF electrode to the porcine adrenal gland in vivo.


14:30         3060.     Artifact-Free Stimulated-Echo Acquisition Mode (STEAM) Cardiac Images with Improved Signal-To-Noise Ratio (SNR)
[Not Available]

Computer 15

Tamer A. Basha1, ElSayed H. Ibrahim1, Nael F. Osman1

1Johns Hopkins University, Baltimore, Maryland, USA

The stimulated echo acquisition mode (STEAM) is currently used in a wide range of applications for imaging tissue parameters. However, when applying STEAM in cardiac imaging, signal loss of the myocardium has been reported due to the intravoxel dephasing of the magnetization during the contraction (or stretching) of the cardiac muscle. Despite of its intrinsically low SNR, STEAM was quite appealing for the assessment of various cardiac functions. In this work, we deal with the SNR and deformation artifactproblems in STEAM technique. First, we introduce a SSFP acquisition technique to increase the SNR then we propose a method for removing the deformation artifacts from the STEAM images.


15:00         3061.     Investigation of Relationship Between Applied Current Amplitude and Measured Current Density Magnitude in a Live Pig

Computer 15

Tim P. DeMonte1, Jia-Hong Gao2, Dinghui Wang3, Weijing Ma3, Michael L.G. Joy3

1Field Metrica Inc., Toronto, Canada; 2University of Chicago, USA; 3University of Toronto, Canada

Current density imaging is an MRI technique used to measure current density vectors in tissue.  Human electro-muscular incapacitation (HEMI) devices are becoming commonly used by law enforcement and military.  The ultimate goal of this work is to achieve better understanding of the effects of HEMI on physiology for enhancement of efficacy and safety.  Specifically, the relationship between applied current amplitude and measured current density magnitude is investigated.  This relationship is expected to be linear over small ranges, but not well understood for larger ranges.  A small range is investigated in an in-vivo pig to establish a method for future work.


New Contrasts & Quantitation

Hall D                                   Tuesday 13:30-15:30                                                                                                                                             

13:30         3062.     Dark Blood BSSFP Imaging Using Magnetization Prepared Random Velocity Encoding

Computer 16

Jamal Jon Derakhshan1,2, Mark A. Griswold1,2, Jeffrey L. Sunshine2, Jeffrey L. Duerk,12

1Case Western Reserve University, Cleveland, Ohio, USA; 2University Hospitals Case Medical Center, Cleveland, Ohio, USA

A new method for generating steady state, short TR, dark blood bSSFP images based on magnetization prepared TrueFISP is presented. Periodically, the magnetization is returned to the z axis with an α/2 pulse. Thereafter, magnetization preparation includes randomly scaled velocity encoding, similar to RF-spoiling. Simulations demonstrate that flowing blood signal can be reduced by > 95% while stationary tissue undergoes much lower loses (~ 24%) based on T2 decay. Phantom imaging results demonstrate stationary and flowing signals consistent with predictions. Human in vivo imaging demonstrates the ability to null blood flow in a short TR magnetization prepared Cartesian bSSFP sequence.


14:00         3063.     Enhanced Contrast in CEST MRI Via Intermolecular Double Quantum Coherences

Computer 16

Shengchun Zhang1, Huijun Sun1, Zhong Chen1, Congbo Cai1, Jianhui Zhong2

1Xiamen University, Xiamen, People's Republic of China; 2University of Rochester, Rochester, New York, USA

A CEST imaging technique based on intermolecular double quantum coherence (iDQC) is proposed. Quantitative analysis and experiments in glucose agarose-gel phantoms demonstrate that, in CEST MRI, iDQC signal is more sensitive to RF saturation than the conventional SQC signal, and thus needs RF saturation pulses of lower power to achieve similar CEST image contrast. Consequently, the method can reduce the potential RF burning in clinic applications, and is expected to facilitate the study of the CEST effect in the system with exchangeable protons of low concentrations.


14:30         3064.     An Investigation of Optimizing and Translating Pulsed-Chemical Exchange Saturation Transfer (CEST) Imaging to a 3 T Clinical Scanner [Not Available]

Computer 16

Phillip Zhe SUN1, Thomas Benner1, Gregory Sorensen1

1A. A. Martinos Center for Biomedical Imaging, Charlestown, Massachusetts, USA

Chemical exchange saturation transfer (CEST) MRI provides a sensitive detection mechanism that allows characterization of dilute labile protons usually undetectable by MRI. Particularly, amide proton transfer (APT) imaging, a variant of CEST MRI, has been shown capable of detecting ischemic tissue acidosis, and may serve as a surrogate metabolic imaging marker. For pre-clinical CEST imaging, long continuous-wave (CW) RF irradiation is often applied so that the steady state CEST contrast can be reached. On clinical scanners, however, specific absorption rate (SAR) limit and hardware design preclude the use of CW irradiation, and instead require an irradiation scheme of repetitive RF pulses (pulsed-CEST imaging). In this work, CW- and pulsed-CEST MRI were systematically compared using a tissue-like pH phantom on an imager capable of both CW and pulsed RF irradiation schemes. The results showed that the maximally obtainable pulsed-CEST contrast is about 95% of CW-CEST contrast, and their optimal RF irradiation powers are equal. Moreover, the pulsed-CEST imaging sequence was translated to a 3 Tesla scanner and detected minor pH difference of 0.6 pH unit using exchangeable amine groups (1.9 ppm). Furthermore, pilot endogenous pulsed-APT imaging of control human volunteers was demonstrated, warranting future APT MRI of stroke patients to fully elucidate its diagnostic value.


15:00         3065.     Relaxometry Changes in a Gel Dosimetry Phantom Due to Continued RF Exposure

Computer 16

Gary Paul Liney1, Mark Godber2, Andrew D. Wilson2, John W. Goodby3

1University of Hull, Hull, UK; 2University of York, York, UK; 3University of York, York, UK

To quantify changes in transverse relaxometry, in phantoms used for MRI gel-dosimetry, due to continued RF heating in the scanner, and to map the distribution of these effects.


13:30         3066.     B1 Correction for Improved Bound Pool Fraction Maps

Computer 17

Nikola Stikov1, Robert F. Dougherty1, John Mark Pauly1

1Stanford University, Stanford, California , USA

The bound pool fraction (f) is an indicator of myelin content in the brain, and cross-relaxation imaging is an efficient method of mapping the f parameter in vivo.  The first step in cross-relaxation imaging is obtaining an accurate T1 map of the brain, but B1 inhomogeneity makes this task difficult.  We incorporated B1 correction in our cross-relaxation procedure, and scanned three subjects with and without this correction.  Our procedure removed variations in the T1 values of white matter across subjects, while reducing the total cross-relaxation scan time.


14:00         3067.     FISPCEST: A Rapid, Acquisition for Dynamic Detection of CEST/PARACEST Activity

Computer 17

Tejas Shah1,2, Meser Ali1, Guanshu Liu1, Mark D. Pagel1, Chris A. Flask1

1Case Western Reserve University, Cleveland, Ohio, USA

We have developed a new FISP Chemical Exchange Saturation Transfer (FISPCEST) pulse sequence to sensitively detect effects. The FISPCEST technique provides <3sec acquisition times which is an order of magnitude less than current CEST techniques. The FISPCEST acquisition combines a single, ~2sec, nonselective CEST preparation and a ~500ms FISP acquisition. The improved temporal resolution is obtained with only a 15% loss in CEST sensitivity in comparison to a spin echo CEST acquisition. The FISPCEST acquisition is adaptable to both endogenous and exogenous (PARA)CEST applications and enables the acquisition of CEST spectra maps and/or multislice CEST images in under 1 minute.


14:30         3068.     Tissue-Dependent Asymmetries in the SSFP Off-Resonance Profile

Computer 17

Karla L. Miller1, Daniel P. Bulte1, Gwenaelle Douaud1, Peter Jezzard1

1Oxford University, Oxford, UK

The SSFP signal is strongly sensitive to off-resonance, with a signal profile for an isochromat population that is theoretically symmetric about the tissue frequency ("on-resonance"). However, the existence of frequency-shifted compartments may lead to an asymmetric profile. We demonstrate this asymmetric response for tissues in the brain, finding a strongly asymmetric response in white matter, a moderately asymmetric response in gray matter and an approximately symmetric response in CSF. This response profile may be useful as a novel marker for tissue content.


15:00         3069.     Dynamic Nuclear Polarization Using a Low Field Multi-Channel MR System

Computer 17

Eugeny Krjukov1, Martyn Paley1

1University of Sheffield, Sheffield, UK

Dynamic nuclear polarisation has been investigated with the free radical carbomyl-PROXYL using a low frequency (360kHz) multi-channel MR system. Enhancement factors of up to 40 were found with 50W or ESR irradiation at 220MHz.


13:30         3070.     Contrast Enhancement by Feedback-Enhanced MRI

Computer 18

Sophia Y. Yang1, Dennis W. Hwang1, Susie Y. Huang2, Lian-Pin Hwang3, Yung-Ya Lin1

1UCLA, Los Angeles, California , USA; 2Harvard Medical School, Boston, Massachusetts, USA; 3National Taiwan University, Taipei, Taiwan

Feedback-enhanced MRI yields robust image contrast that is sensitive to small differences in the underlying microscopic frequency distributions. Important applications of this method include improving the visualization of SPIO nanoparticles through generation of positive contrast and distinguishing small changes in microscopic susceptibility corresponding to tumor and normal tissue. Using an external electronic device can significantly enhance the feedback field and open opportunities for the design of novel imaging pulse sequences in which the feedback interaction is controllable. Examples of in vitro and in vivo tumor detection in human brain tissue and mouse models of lung adenocarcinoma with active feedback will be demonstrated.


14:00         3071.     MRI Manometry Using Gas Filled Microbubbles Exhibiting High Membrane to Gas Synergy: Towards Clinical Relevance

Computer 18

Robert Henry Morris1, Martin Bencsik1, Marie-Pierre Krafft2, Gilles Waton2, Nikolaus Nestlé3, Petrik Galvosas4, Anil Vangala, Yvonne Perrie5

1Nottingham Trent University, Nottingham, UK; 2Institut Charles Sadron, Strasbourg, France; 3BASF Aktiengesellschaft, Ludwigshafen, Germany; 4University of  Leipzig, Leipzig, Germany; 5Aston Universit

MRI manometry is performed in vitro using two alternative contrast agents comprised of compressible microcapsules suspended in a liquid medium presenting high viscosity with little reduction of the diffusion coefficient compared to that of bulk water. The currently available contrast agent utilising standard lipid coated gas microcapsules is shown to be highly unstable in typical clinically relevant pressure conditions, whilst perfluorinated gas microcapsules coated with a perfluoroalkylated lipid will allow in vivo measurements in the future.


14:30         3072.     Visualization of Viscoelastic Properties by Combining US Pulses and MRI

Computer 18

Ole Benjamin Oehms1, Marcus Radicke1, Sarah Wrede1, Meinert Lewerenz1, Andre Engelbertz1, Karl Maier1

1Friedrich Wilhelms Universität, Bonn, Germany

The irradiation of Ultrasound Pulses (30 ms, at 10 MHz) into a sample during a diffusion sensitive MRI sequence leads to signal changes in liquids and tissue. They are caused by the decrease of the acoustic radiation pressure due to damping of the sound wave which leads to a movement in a liquid along the path of sound propagation. This movement leads to a dephasation if it occurs while the diffusion gradient is active which results in a signal diminishment in that region. This diminishment depends on the viscoelastic properties of the sample. First measurements on Water and Glycerine and on a piece of tissue will be presented in the talk.


15:00         3073.     MR Imaging of Transient Shear Waves Induced by Ultrasonic Radiation Force

Computer 18

Remi Souchon1, Rares Salomir1, Olivier Beuf2, Denis Lyonnet3, Jean-Yves Chapelon1, Olivier Rouviere3

1INSERM U556, Lyon, France; 2CNRS UMR 5220, Lyon, France; 3Hospices Civils de Lyon, France

This study reports preliminary wave images and temperature measurements for transient MR elastography (MRE) using ultrasound radiation force. Our initial data suggest that an EPI MRE sequence is likely to provide elasticity images while ensuring patient safety.


13:30         3074.     Improved MREIT Reconstruction Using Sodium MRI

Computer 19

Mark Jason Hamamura1, L Tugan Muftuler1, Orhan Nalcioglu1

1University of California, Irvine, California , USA

In magnetic resonance electrical impedance tomography (MREIT), electrical currents are injected into an object and the resulting magnetic flux density distribution measured using MRI.  These MRI measurements are then used to reconstruct the conductivity distribution within the object.  In this study, we investigated the incorporation of sodium MRI data into the MREIT reconstruction algorithm.  The results demonstrate that this incorporation can improve the accuracy of the reconstructed conductivity maps.


14:00         3075.     SPIO Acid Dissolution Kinetics with MR Susceptometry

Computer 19

Ludovic de Rochefort1, Yi Wang1

1Weill Medical College of Cornell University, New York, New York, USA

Superparamagnetic iron oxides benefit from a very strong magnetic moment at low field due to their superparamagnetic property. Here, we show the feasibility of monitoring chemical reaction of SPIO dissolution by acids with MRI. The magnetic moment destruction is measured continuously as a function of time with MR susceptometry.


14:30         3076.     A Quantitative Approach of Extracting Magnetic Moments in Small Cylindrical Object

Computer 19

Ching-Yi Hsieh1, Yu-Chung Norman Cheng1, Jaladhar Neelavalli1, E. Mark Haacke1

1Wayne State University, Detroit, USA

Our goal is to quantify magnetic moments of a small in-vivo object such as veins in the brain from MR images, without any a priori information. By summing up MR signals within three concentric circles, the magnetic moment of the object obtained from different complex data in the same image can be accurate within 10% of its true value. To achieve this accuracy, a long echo time may be needed. The simulations and experimental results are presented for the gel phantom. The agreement between these two results indicates a promising potential of this method.


15:00         3077.     In Vivo T1ρ-Weighted MR Imaging of Rat Brain Using a Surface Coil at 11.7 Tesla

Computer 19

Su Xu1, Jehoon Yang1, Jun Shen1

1National Institute of Mental Health, Bethesda, Maryland, USA

A sech-based adiabatic spin-lock pulse sequence to obtain T -weighted MR images using a surface transceiver coil was optimized for enhancing tissue contrast. The utility of this technique was demonstrated using in vivo rat brains after focal bicuculline administration and an 11.7 Tesla 89 mm bore vertical microimager. Signal intensity of the lesion in the T 11ρ-weighted images was significantly elevated 50 minutes after administration of bicuculline.


13:30         3078.     Magnetization-Prepared Shells for Efficient T1-Weighted Brain Imaging

Computer 20

Yunhong Shu1, Matthew A. Bernstein1

1Mayo Clinic College of Medicine, Rochester, Minnesota, USA

To maximize the contrast in MP-RAGE brain imaging, it is desirable to select a k-space acquisition order that can sample the center of k-space compactly during the peak contrast difference between WM and GM during the inversion recovery curve. The shells trajectory is a non-Cartesian 3D trajectory with high acquisition efficiency and an inherent centric nature. It provides the flexibility required to synchronize the acquisition of the center of k-space to the contrast maximum contrast. Here we theoretically and experimentally demonstrate the feasibility of combining magnetization preparation with the shells trajectory to achieve T1-weighted brain imaging efficiently.


14:00         3079.     Flow-Independent T2-Prepared Inversion Recovery Black Blood MR Imaging

Computer 20

Chia-Ying Liu1,2, Oliver Wieben1, Jean H. Brittain2, Scott Brian Reeder1

1University of Wisconsin-Madison, Madison, Wisconsin, USA; 2GE Healthcare, Madison, Wisconsin, USA

Black blood prepared MRI is used extensively for cardiac and atherosclerotic plaque imaging. Most black blood sequences employ double inversion recovery, which relies on the inflow of nulled blood. As a result, double IR methods are less effective in the presence of slow flow and in-plane flow. We present a new black-black preparation scheme which employs a T2-prepared sequence in combination with an inversion recovery pulse (T2Prep-IR). Excellent blood suppression independent of flow was demonstrated in the heart and carotid arteries of volunteers.


14:30         3080.     In Vivo Blood T1 Mapping Using Inversion Recovery TrueFISP

Computer 20

Wen-Chau Wu1, Jiongjiong Wang1

1University of Pennsylvania, Philadelphia, USA

In the present study, we demonstrated the feasibility of in vivo blood T1 mapping with an inversion recovery (IR) TrueFISP sequence. The IR TrueFISP signal has been shown to vary with the flip angle, T1 and T2 of static tissue of interest. With continous inflow of flesh blood with undisturbed longitudinal magnitization, the IR TrueFISP curve of blood pool signal approximated standard T1 relaxation. The estimated blood T1 values at 3.0T match well with literature results with minimal sensitivity to variations in flip angle.


15:00         3081.     Practical Optimum Experimental Designs for Fast T1 Relaxometry with SPGR Sequences

Computer 20

Alexey Samsonov1, Andrew L. Alexander1, Youngkyoo Jung1, Aaron S. Field1

1University of Wisconsin, Madison, Wisconsin, USA

Knowledge of the longitudinal relaxation time T1 is required in many quantitative MRI applications.  T1 mapping using variable flip angle SPGR acquisitions is an attractive choice due to its speed.  In this work, we describe a method for automatic selection of T1 mapping flip angles, which explicitly optimizes the performance of T1 mapping for a wide range of T1 values.  The method yielded 3 flip angle designs with performance similar to the previously described 10 flip angle design.  This development may allow more efficient T1 mapping optimized for wide range of tissue types.


13:30         3082.     Standardized Structural Magnetic Resonance Imaging in Multi-Centre Studies Using Quantitative T1 and T2 Imaging

Computer 21

Sean CL Deoni1, Steven CR Williams2, Peter Jezzard1, John Suckling3, Declan GM Murphy2, Derek K. Jones4

1University of Oxford, Oxford, UK; 2Institute of Psychiatry, London, UK; 3University of Cambridge, Cambridge, UK; 4Cardiff University Brain Research Imaging Centre, Cardiff, UK

Multicentre studies are becoming increasingly common as they facilitate the recruitment of greater numbers of subjects while decreasing the economic cost and duration of study.  However, precise matching of structural image quality, necessary to draw meaningful inferences from the data particularly in regards to morphology, becomes difficult as the number and diversity of imaging systems increases.  Here we report on the use of quantitative T1 and T2 imaging for standardizing the structural imaging component of such studies, demonstrating high reproducibility of the measures across different systems.


14:00         3083.     Combining Morphometry and T1 Relaxometry in a Single Imaging Protocol: Measuring T1 with MPRAGE

Computer 21

Olivier Mougin1, Penny Gowland1

1School of Physics and Astronomy, Nottingham, UK

We are using relaxation times to study normal and pathological brain development. Acquisition times for T1 are generally long, so this study aims to make use of the anatomical image that is acquired for morphological information into the relaxometry protocol. Therefore we have optimized the Magnetization Preparation followed by a RApid Gradient Echo (MPRAGE) sequence (which is routinely used for morphology at our site) for the measurement of T1. Study on five subjects at three different fields shows agreement with the literature and gold standard sequences.


14:30         3084.     Fast T1 Mapping in Human Brain Using Inversion Recovery EPI with GRAPPA at 3T and 7T

Computer 21

John Grinstead1, William Rooney2

1Siemens Medical Solutions USA, Inc., Portland, USA; 2Oregon Health & Science University, Portland, USA

Quantitative T1 techniques find a wide range of applications in biological NMR, but the major drawback of these techniques is that they are slow. This is because the sampling requirements are high, not only must the T1 recovery be well sampled, but also spatial encoding is usually desired. This work investigates the combination of inversion recovery echo-planar and parallel imaging techniques for high-speed acquisition of quantitative T1 data sets in human brain at 3T and 7T.


15:00         3085.     Measurement of R1 Dynamics Using a 3D FLASH Variable Flip Angle Sliding Window Technique

Computer 21

Jessica Schulz1,2, Eva Christina Wönne1, Arne Hengerer2, Wolfhard Semmler1, Michael Bock1

1Deutsches Krebsforschungszentrum (dkfz), Heidelberg, Germany; 2Siemens Medical Solutions, Erlangen, Germany

A 3D FLASH variable flip angle method was combined with a sliding window calculation to obtain the relaxation rates R1 dynamically with a temporal resolution of 10 s. In a contrast agent study on tumor-bearing mice the contrast agent-related changes in R1 could be mapped in tumors, and the R1 values were in excellent agreement with reference measurements.


Shared Resources & Quality Control

Hall D                                   Tuesday 13:30-15:30                                                                                                                                             

13:30         3086.     The Neuroimaging Informatics Tools and Resources Clearinghouse (NITRC)

Computer 22

David N. Kennedy1, Robert Buccigrossi2, Jeff Grethe, Christian Haselgrove, Nina Preuss, Keith Wagner, Mark Ellisman

1MGH, Charlestown, Massachusetts, USA; 2Turner Consulting Group, USA

NITRC, a new neuroimaging knowledge environment, is now online (www.nitrc.org). We encourage the fMRI community to try it out and provide feedback on its design, tools, resources, and content. NITRC is a knowledge environment for the fMRI community where tools and resources are presented in a coherent and synergistic environment for the advancement of MRI-based neuroscience research.


14:00         3087.     PAQAP: A Quality Assessment Protocol for MRI

Computer 22

Pieter Vandemaele1, Rik Achten1, Yves De Deene1

1Ghent University Hospital, Ghent, Belgium

Quality assessment in MR imaging provides clinician and researchers with objective measures of the performance of their MR scanner and scan protocols. PAQAP (Pieter’s Automated Quality Assurance Protocol) is a combination of a standard QA phantom and an elaborate software program for full automatic data analysis and report generation. PAQAP provides an easy way to acquire and process QA data on a regular basis with minimal interaction and within a limited time frame by an MR technologist. The system will be implemented and systematically used in a QA program at the experimental MR site of the Ghent University Hospital.


14:30         3088.     MR Image Quality Evaluation Using Weighted Perceptual Difference Model (Case-PDM)

Computer 22

Jun Miao1, Wilbur C. K. Wong1, David L. Wilson1,2

1Case Western Reserve University, Cleveland, USA; 2University Hospital of Cleveland, Cleveland, USA

The perceptual difference model (Case-PDM) is being used to quantify image quality of fast, parallel MR acquisitions and reconstruction algorithms by comparing to slower, full k-space, high quality reference images. In this paper, we create an alternative metric weighted to image features to improve the linear correlation coefficient between human ratings and weighted Case-PDM, across a large set of MR reconstruction test images of varying quality. Our method is robust across subjects and anatomy; that is, scores maintain a high correlation with human ratings even if the test dataset is different from the training dataset.


15:00         3089.     A Novel SNR Estimation Technique Applicable to Clinical Parallel MR Images: Triple Band-Width Single Acquisition Method (TriSAM)
[Not Available]

Computer 22

Yoshio Machida1, Hiroshi Kusahara1, Yoshimori Kassai1

1Toshiba Medical Systems Corporation, Otawara, Japan

We have developed a new technique gTriple band-width Single Acquisition Methodh (TriSAM) in which noise maps can be obtained with originally intended target images with no extra scan time. Application of this technique with parallel imaging to the head images on a volunteer provides the misregistration free noise images. The TriSAM is considered to be one of the most practical SNR estimation approaches for clinical images.



Hall D                                   Tuesday 13:30-15:30                                                                                                                                             

13:30         3090.     Optimal Spoiling of the Transverse Magnetization in the Actual Flip-Angle Imaging (AFI) Sequence for Fast B1 Field Mapping

Computer 23

Vasily L. Yarnykh1

1University of Washington, Seattle, Washington, USA

A recently developed Actual Flip-angle Imaging (AFI) method allows fast B1 mapping based on the spoiled steady-state principle. This study presents theoretical and experimental examination of conditions required for optimal spoiling in the AFI sequence. It was found that the spoiling behavior of the AFI sequence is different from a traditional spoiled gradient echo sequence. To achieve optimal spoiling, appropriate combinations of an RF phase increment and spoiler gradient areas need to be used. The sequence design providing highly accurate B1 measurements and possible sources of errors are described.


14:00         3091.     In-Vivo Assessment of a STEAM Sequence for B1-Mapping

Computer 23

Rudolf Stollberger1, Thomas Birngruber2

1Graz University of Technology, Graz, Austria; 2Medical University of Graz, Austria

RF field inhomogeneities are a main source for image inhomogeneities, spatial dependent SNR and CNR and systematic errors in quantification of MRI data. A STEAM sequences was evaluated in-vivo at 3T for B1-determination in quantitative studies. It could be shown that the sequence is robust and sufficiently accurate for the application in most regions. Some problems occur in the chest from motion artefacts. The acquisition time for a scanning matrix of 128*64 was 47s for TR=800ms and 78s for TR=1300.


14:30         3092.     Impact of the Correction of B1 Inhomogeneities for Dynamic Contrast-Enhanced Imaging at 3 Tesla

Computer 23

Robert Merwa1, Franz Ebner2, Rudolf Stollberger1

1Graz University of Technology, Graz, Austria; 2Medical University of Graz, Graz, Austria

This study was performed in order to evaluate the influence of the B1-inhomogenities for dynamic contrast-enhanced MRI at 3 T. The active RF-field was measured with a stimulated echo sequence whereas the actual flip angle distribution is determined. Using a reference scan and a perfusion scan particular parameters as temporal T1 relaxation time, concentrations and arterial input function can be calculated. The results obtained with the correction of the flip angles show a significant improvement compared to the results obtained without correction. All the measurements were performed on a 3 T System (Siemens Magnetom Trio a Tim System)


15:00         3093.     Rapid RF Flip Angle Imaging

Computer 23

Daniel Kim1, Sohae Chung1, Daniel K. Sodickson1, Leon Axel1

1New York University, New York, New York, USA

The transmit radiofrequency (RF) filed (B1) uniformity plays an important role in determining the image quality in MRI, particularly at high field strengths (&[ge] 3T). Accurate B1 or flip angle maps are needed to compensate for B1 variations through different compensation strategies. Among the existing methods for in vivo B1 mapping, the double angle method (DAM) is most straightforward. However, its image acquisition efficiency is very low due to a need to set TR &[ge] 5 T1s. The purpose of this study is to develop a rapid in vivo B1 mapping method based upon three single-shot image acquisitions.


Image Registration & Alignment

Hall D                                   Wednesday 13:30-15:30                                                                                                                                       

13:30         3094.     Validation of 3D Non-Rigid Whole Body MR Image Registration

Computer 15

Xia Li1, Thomas Yankeelov, Todd Peterson, John Gore, Benoit Dawant

1Vanderbilt University, Nashville, Tennessee, USA

The automatic registration of whole body MR images, which requires non-rigid registration techniques for the

articulated structures, remains a challenge. Although we proposed a promising registration method that permits the automatic registration of MR images for both intra- and inter-subject, one weakness is found in this algorithm: bones can be deformed incorrectly because of the surrounding structures. A modified method was proposed to constrain the deformation of bony structures. However, complete validation is required. In this study, quantitative validation results show the accuracy of our algorithm.


14:00         3095.     Breast MR Registration for Evaluation of Neoadjuvant Chemotherapy Response

Computer 15

Ruparani Chittineni1,2, Min-Ying Su1, Orhan Nalcioglu1

1University of California, Irvine, Irvine, USA

The deformable nature of breast tissue results in significant shape differences between serial studies, making it challenging to chalk out a clear trajectory of the corresponding tumor locations. These studies correspond to MR-based monitoring of chemotherapy for therapy response evaluation. In this abstract we demonstrate the use of constraint based free-form deformations for registration of serial breast MR studies. The algorithm is especially useful in patients with multi-centric or multi-focal lesions. Also, differentiating between therapy-induced inflammation and residual disease becomes amenable. Since, tumor volumes are preserved during the transformation, mis-interpretation of results can be avoided.


14:30         3096.     Image Registration of Mouse Brains Containing Varying Amounts of Extra Cortical CSF

Computer 15

Matthijs C. van Eede1, Jason P. Lerch1, John G. Sled1

1Toronto Centre for Phenogenomics, Toronto, Canada

In recent studies we encountered brains containing varying amounts of extra cortical CSF. Using our groupwise registration method for analysis of the brains, we found this resulted in incorrect alignment. That would lead to incorrect findings. We have developed a modulation strategy to deal with this extra cortical CSF, alleviating the registration problem.


15:00         3097.     Non-Rigid Registration of Diffusion Weighted MRI Using Progressive Principal Component Registration (PPCR)

Computer 15

Andrew Melbourne1, David Hawkes1, David Atkinson1

1University College London, London, UK

Artefacts as a result of patient motion & eddy current distortions often corrupt Diffusion Weighted MR images, reducing the success of subsequent analysis. Registration of images that contain different contrast from each gradient direction may produce inaccurate results. The PPCR scheme allows diffusion direction images to be registered into a common coordinate frame by combining overlapping diffusion contrast using principal components analysis. PPCR registration is compared to data registered using an affine registration of each diffusion direction to the corresponding B0 volume. The use of the PPCR method allows enhanced feature demarcation by removing geometric distortion artefacts.


13:30         3098.     Restoration of Compressed or Constricted Images: A Feasibility Study for Intra- And Inter- Imaging Modality Registration

Computer 16

Ruparani Chittineni1, SeungHoon Ha1, Werner Roeck1, Min-Ying Su1, Orhan Nalcioglu1

1University of California, Irvine, Irvine, USA

Restoring form of compressed or deformed images is of utmost significance. Automatic non-rigid registration techniques have been applied extensively to address non-linear deformations. However, it is interesting to note that such algorithms may fail or be biased towards dominant intensity regions in the images and hence have minimal local registration. We demonstrate and compare the applicability of automatic and landmark based methods for the particular case of addressing deformation in the presence of constriction or applied compression. The near future application of the developed method is for co-registration of breast images acquired using MRI (uncompressed) and scintimammography (under light compression).


14:00         3099.     Registration of 3D MR Images of the Mouse Embryos

Computer 16

Mojdeh Zamyadi1,2, R Mark Henkelman1,2, Shoumo Bhattacharya3, Jurgen E. Schneider3, John G. Sled1,2

1University of Toronto, Toronto, Canada; 2Mouse Imaging Centre, Hospital for Sick Children, Toronto, Canada; 3University of Oxford, Wellcome Trust Centre for Human Genetic, Oxford, UK

We are developing an image registration technique to detect subtle anatomical shape differences between 3D MR images of mouse embryos. In order to assess feasibility, we have used non-linear registration to align a group of genetically identical embryos. We tested the assumption that embryo anatomy is highly conserved among specimens by registering six 3D embryos together. The result of the registration process is shown in form of a final average image consisting of data from the 6 individuals, and the root mean squared (RMS) displacement image which is a representation of the

anatomical variation among the genetically identical embryos. These initial findings suggest that embryo anatomy is highly conserved among specimens and that image registration of 3D MRI data is a feasible approach for subsequently detecting abnormal phenotypes.


14:30         3100.     Quality Control in a Longitudinal Multi Center Alzheimer's Disease Study

Computer 16

Eric Westman1, Andy Simmons2, Sebastian Muehlboeck3, Tony Segerdahl4, Johan Bengtsson4, Lars-Olof Wahlund1, Simon Lovestone2, Christian Spenger4

1Department of Neurobiology, Health Care Sciences and Society, Karolinska Institutet, Stockholm, Sweden; 2MRC Centre forNeurodegeneration Research, Institute of Psychiatry King's College, London, UK; 3McConnell Brain Imagi

Within the InnoMed/AddNeuroMed research project funded by the European Union, sixth frame work program, data has been successfully collected for a multi site MRI study. Quality control and quality assurance are performed on routine basis at data collection centers and at the data coordination centre. The feature-set of the database system covers the entire process from image acquisition, storage, quality control to data querying for analysis. Quality control statistics show that the performance of the participating sites is very high; 97 % of all T1 images passed QC.  


15:00         3101.     Validation of User Independent Planning Tool for Consistent Data Acquisition in Multi-Center Trials

Computer 16

Esben Thade Petersen1,2, Ivan Zimine1,3, Xavier Golay,14, The QUASAR Reproducibility study

1National Neuroscience Institute, Singapore, Singapore; 2Aarhus University Hospital, Aarhus, Denmark; 3Philips Medical Systems, Tokyo, Japan; 4Singapore Bioimaging Consortium, Singapore, Singapore

In this work, we evaluated the accuracy of automatic slice positioning which recently has become available on standard MRI systems. The success of MRI studies often depends on the consistency of the image acquisition and is especially important in longitudinal and multi-center trials. Differences in slice angulations and positioning can easily affect the “subjective” reading by radiologists but also the quantification in DTI, perfusion or volumetric-imaging. Three automatically planned images were acquired in 170 subjects and minor rotation and translation between scans were observed after co-registration of the images, resulting in high consistency for future trials using these tools.


13:30         3102.     Exploring the Relationship Between Natural Fluctuations in Electrical Measures of Brain Activity and the BOLD Response, During Visual Stimulation

Computer 17

Karen J. Mullinger1, Gerda B. Geirsdottir1, Matthew J. Brookes1, Peter F. Liddle1, Richard W. Bowtell1

1University of Nottingham, Nottingham, UK

The correlation of preceding alpha power and driven power with the BOLD response to a visual stimulus has been investigated using simultaneous EEG/fMRI experiments at 3 T. Despite good characterisation of the BOLD and electrical responses no correlation was found between the fluctuations in the alpha power preceding the stimulus or in the driven power and the BOLD response in data from individual subjects. A positive trend was however found when comparing the fractional difference in BOLD response and preceding alpha power in trials falling in the top and bottom quartiles binned according to the preceding alpha power across subjects.


14:00         3103.     Hemispherical Constrained Surface Controller for 3D Navigation

Computer 17

Martin John Graves1, David John Lomas1

1University of Cambridge and Addenbrooke's Hospital, Cambridge, UK

Although there has been significant development of volumetric image acquisition methodologies there has been little development of methods for subsequent reformatting of data beyond standard linear tools. This work describes the development of a 3D constrained surface controller for interrogating volumetric data. The controller allows for intuitive navigation by following an ultrasound-style motion paradigm in which data reformatting is performed over a virtual hemispherical surface around the organ of interest. Constraining the motion to an anatomically consistent surface reduces the possibility of the operator becoming spatially disorientated. The controller was evaluated in comparison to conventional reformatting software.


14:30         3104.     Uncertainty in Simultaneous Estimation of Blood Oxygenation Level and Volume Fraction on the Basis of Spin Dephasing in a Vascular Network

Computer 17

Jan Sedlacik1,2, Jürgen R. Reichenbach1

1University Clinics of the Friedrich Schiller University, Jena, Germany

Blood oxygenation level and volume fraction are essential input parameters of theoretic models of spin dephasing in a vascular network. It is possible to estimate these parameters by fitting the simulated signal to measured signal-time curves. However, if the blood oxygenation level and volume fraction are unknown, they can not be reliably estimated by simply fitting theoretical signal curves to the measured signal decay. The purpose of this work was to unravel this difficulty of a simultaneous estimation of blood oxygenation level and volume fraction.


15:00         3105.     Local Feature-Preserving Selection of Kernel Size for Unwrapping of High-Resolution Phase Images

Computer 17

Julien Milles1, Matthijs J.P. van Osch1, Louise van der Weerd1, Rob J.A. Nabuurs1, Wouter M. Teeuwisse1, Jeroen van der Grond1, Mark A. van Buchem1, Johan H.C. Reiber1

1Leiden University Medical Center, Leiden, Netherlands

The aim of this work is to investigate the use of objective criteria to determine the optimal filtering kernel size in order to perform phase unwrapping while preserving local features on high-resolution phase images by means of a k-space filtering-based algorithm. We propose two local information-based criteria that depict the trade-off involved in phase unwrapping. We study the effect of kernel size on those criteria and show that their analysis allows determining an optimal filtering kernel size that realizes the trade-off between wraps removal and local features preservation.


13:30         3106.     Dual Echo Susceptibility Weighted Imaging (SWI) : Reducing the Error in Making Phase Mask

Computer 18

Yoshiyuki Ishimori1, Masahiko Monma, Yutaka Kouno, Makito Iizuka, Seiichi Sasaki

1Ibaraki Prefectural University of Health Sciences, Ami-machi, Inashiki-gun, Japan

The high pass filter process utilizing in susceptibility weighted imaging has the possibility of insufficient removal of phase wrapping and another artifact on the phase mask. We calculated the phase difference caused in TE interval by dual echo technique and calculated  the time-invariant component of the phase. We used this time-invariant phase for making a phase mask. With this technique, severe phase wrappings were removed and edge enhancement effect around the positive phase portion was reduced.


14:00         3107.     Intelligent Image Repository for MRI Brain Clinical Trials

Computer 18

Gianlorenzo Fagiolo1, Nick Fox2, Derek L. Hill2,3, Adam D. Waldman4, Jo V. Hajnal1

1Imperial College,, London, UK; 2UCL, London, UK; 3London Bioscience Center, London, UK; 4Charing Cross Hospital, London, UK

Longitudinal MRI in which the same subject is scanned on multiple occasions is increasingly being used for clinical trials, and for diagnosis/monitoring of patients, particularly in brain studies. Minor errors in patient information entered into hospital information systems are common and occasionally gross errors occur. The result can be that images to be compared are wrongly identified leading to information loss in clinical trials. In this work, a MRI brain clinical trials image repository with intelligent text recognition and Image-based Subject Identification was developed. The identification proved robust both to scan quality deterioration and to progressing brain atrophy.


14:30         3108.     Rapid Prototyping of a 3D Grid Phantom for MR Image Guided Therapy Quality Assurance

Computer 18

Barbara Holshouser1, David Kittle2, James M. Slater2, Robert D. Pearlstein3

1Loma Linda University Medical Center, Loma Linda, California , USA; 2Loma Linda University, Loma Linda, California , USA; 3Duke University and Medical Center, Durham, USA

We have used rapid prototyping technology to directly fabricate a 3D grid phantom from CAD drawings. The phantom was then used to measure spatial inaccuracies in 3D images acquired in a 12 channel receive only head coil in a 3T MR scanner. The ultimate purpose is to characterize MR imaging related spatial inaccuracies for image guided surgery and radiotherapy.


15:00         3109.     Validation of an Automatic Method for Change Detection in Serial Scalar Images Characterizing Diffusion Properties

Computer 18

Hervé Boisgontier1,2, Vincent Noblet1, Fabrice Heitz1, Lucien Rumbach3, Jean-Paul Armspach2

1Laboratoire des Sciences de l'Image, de l'Informatique et de la Télédétection, UMR CNRS-ULP 7005, Illkirch, France; 2Laboratoire d'Imagerie et de Neurosciences Cognitives, UMR CNRS-ULP 7191, Strasbourg, France; 3CHU Minjo

Studies have already highlighted diffusion property alterations induced by multiple sclerosis. This paper proposes an automatic method for detecting changes between two or more scalar images characterizing diffusion properties. Validation of the proposed approach and comparison of the impact of the scalar index used (mean diffusivity, fractional anisotropy and lattice index) have been done using synthetic simulations. Results show that the performance of the method depends on the index considered. This is mainly explained by the nature of noise, which varies from one index to the other


Motion Artifact Correction

Hall D                                   Wednesday 13:30-15:30                                                                     Chairs: Anja C. Brau and Donglai Huo

13:30         3110.     Motion Artifact Correction with MOJITO: Practical Implications

Computer 19

Candice Anne Bookwalter1, Mark A. Griswold1, Jeffrey L. Duerk1

1Case Western Reserve University, Cleveland, Ohio, USA

It is a well known property of Fourier Transform MRI that rigid body translational motion in image space results in linear phase accumulation in k-space.  This work describes practical simulations and MR experiments using Multiple Overlapping k-space Junctions for Investigating Translating Objects (MOJITO), a correction scheme based on phase differences at trajectory intersections caused by 2D object shifts.  The algorithm allows both detection and correction of motion artifacts caused by 2D rigid body translational motion.  Here, MOJITO is demonstrated using the continuous sampling BOWTIE trajectory.


14:00         3111.     Effect of Motion-Induced Altered Coil Sensitivity on Parallel Imaging Performance

Computer 19

Murat Aksoy1, Roland Bammer1

1Stanford University, Stanford, California , USA

Correction of involuntary patient motion related artifacts remains to be one of the most important topics in MRI. In the specific case of rigid body motion, rotation and translation changes the position of the anatomy under examination with respect to the receiver coil, which, in turn, changes the coil sensitivity “seen” by the anatomy. In this study, we investigated the effects of this change in coil sensitivity with rigid body motion. Our results show that especially at higher reduction factors and in the case of severe patient motion, coil sensitivity profiles have to be modified accordingly to reflect the correct coil sensitivity exposure in order to accomplish accurate parallel image reconstruction.


14:30         3112.     SENSE Motion Correction

Computer 19

Julian R. Maclaren1, Bing Wu1, Philip Bones1, Rick P. Millane1, Richard Watts1

1University of Canterbury, Christchurch, New Zealand

A motion-correction technique using a multiple-shot fast spin echo (FSE) sequence is presented. SENSE is used to reconstruct a complete, albeit noisy, image from data collected in each individual shot. These images are then registered and combined to produce a motion-corrected reconstruction. This technique has the advantage that a regular reconstruction can also be obtained from the same k-space data set. Hence, little is lost by applying this technique and an advantage is gained over a standard FSE acquisition in the case of patient motion.


15:00         3113.     Motion Mitigation for Eye Imaging Using Cloverleaf Navigators

Computer 19

Andre J. W. van der Kouwe1, Thomas Benner1, Ashok Kumar1, Timo van Kerkoerle1, Graham Wiggins1, Giorgio Bonmassar1

1Massachusetts General Hospital, Charlestown, Massachusetts, USA

High-resolution imaging of the human eye in-vivo is exacerbated by involuntary motion. The eye performs small saccadic movements at irregular intervals, and moves during blinking. These movements are manifest in the MR image as blurring, and smearing in the phase encoding direction.  We present a suggestion for dealing with this type of motion during imaging by oversampling in k-space while simultaneously collecting embedded motion-detecting navigators at every repetition time of the running scan. We reconstruct the image off-line by averaging only those lines of k-space that are not corrupted by motion.


13:30         3114.     Requirements on the Accuracy of Navigators for Prospective Motion Correction in High Resolution MR Imaging

Computer 20

Maxim Zaitsev1, Oliver Speck, Jürgen Hennig1

1University Hospital Freiburg, Freiburg, Germany

Prospectively navigated MR imaging is becoming increasingly popular to overcome present measurement time and/or resolution limitations in various imaging applications. Navigated imaging appears to be the only way to overcome the limitation for isotropic resolution in length high-resolution scans, in which involuntarily movements of the order of millimetres are unavoidable. Despite of the popularity of navigators, to our knowledge no data on the required navigator accuracy are available in literature. Here, formalism is developed to analyse statistically the image artefacts introduced by the prospective motion correction based on the imaged sample properties and desired resolution.


14:00         3115.     The Investigation of T2 PROPLLER Motion Estimation Efficiency

Computer 20

Shaorong Chang1, Xiaoli Zhao1, Ajeetkumar Gaddipati1

1GE Healthcare, Waukesha, Wisconsin, USA

T2 PROPELLER reconstructs clinically valuable images with little motion artifacts.  However, studies show that the motion correction efficiency depends on protocol parameter settings.  In this work, the efficiency of PROPELLER motion correction with respect to field of view (FOV), views per blade and k-space coverage is analyzed using invivo data.  Rotation and translation estimation accuracy were studied individually, and the overall effects of motion correction were inspected.  Based on this study, tight FOV, at least 24 views per blade and 1 cm-1 k-space coverage are recommended in a T2 PROPELLER scan to obtain brain images with little motion artifacts.


14:30         3116.     Respiratory Motion-Compensated Radial DCE-MRI of Chest and Abdominal Lesions

Computer 20

Wei Lin1, Junyu Guo1, Mark A. Rosen1, Hee Kwon Song1

1University of Pennsylvania, Philadelphia, Pennsylvania, USA

Accurate assessment of lesion perfusion with dynamic contrast-enhanced MRI (DCE-MRI) in the chest and abdominal regions is difficult due to respiratory motion. In this work, we utilize the self-gating properties of radial imaging with reconstruction flexibility afforded by the golden-angle view order scheme to effectively compensate for respiratory motion. In addition, residual motion within each image is compensated with the autofocusing technique, and motion between images corrected with image correlation. Finally, the resulting data is processed to reduce streaks and enhance SNR, using either k-space weighted image contrast (KWIC) filtering or principal component analysis (PCA). In vivo results demonstrate the effectiveness of our strategy in achieving high quality images of lesions located in the lung and the liver.


15:00         3117.     Image Deformation Recovery Using Overlapping Partial Samples (IDROPS): Model-Based Respiratory Artefact Correction in Free-Breathing Liver MRI

Computer 20

Mark J. White1, David Atkinson1, Liz Charles-Edwards2, Catherine Coolens2, Maria Hawkins2, Keiko Miyazaki3, David Collins3, Martin O. Leach3, David J. Hawkes1

1UCL, London, UK; 2Royal Marsden NHS Foundation Trust, UK; 3Institute of Cancer Research, UK

Respiratory motion needs managing in most liver MRI protocols. Given deformation fields associated with each part of acquired k-space, artefacts in free-breathing liver MRI may be corrected using a general matrix reconstruction method.  iDROPS builds a parameterized non-rigid breathing model from a low-resolution training series, then finds a time-series of parameter values for an imaging acquisition by comparing the parts of k-space which overlap with the training data.  The resulting estimated deformation fields are accurate to 1.5mm near the diaphragm (average of 10 studies), and are demonstrated correcting real continuously-acquired free-breathing liver data.


13:30         3118.     New Approach for Affine Transform Parameters Estimation for Prospective Motion Correction Using External Sensors

Computer 21

Brice Fernandez1,2, Maelene Lohezic1,2, Pierre-Andre Vuissoz2, Jacques Felblinger2

1General Electric Healthcare, Buc, France; 2IADI, Inserm ERI 13, Nancy University, Nancy, France

In most clinical applications, standard external sensors (ECG and respiratory belts) are usually used for triggering. We propose a method in two steps that use external sensors to estimate the position of an organ of interest. The first step consists in the calibration of external sensors output, based on fast imaging and a parametric motion estimation algorithm. The second step is a multivariate linear regression, computed between motion parameters and external sensors. This linear relation is then used to demonstrate the accuracy of the proposed method for prospective correction.


14:00         3119.     Robust Abdominal Imaging with Motion Sensitive Sequences Using Cardiac and Respiratory Double Gating

Computer 21

Ananth J. Madhuranthakam1, Ken-Pin Hwang2, Reed F. Busse3, Jean H. Brittain3, Neil M. Rofsky4,5, David C. Alsop4,5

1GE Healthcare, Boston, Massachusetts, USA; 2GE Healthcare, Houston, Texas, USA; 3GE Healthcare, Madison, Wisconsin, USA; 4Beth Israel Deaconess Medical Center, Boston, Massachusetts, USA; 5Harvard Me

Abdominal imaging with motion sensitive sequences is challenging due to signal loss and artifacts associated with cardio-respiratory motion.  Breath hold acquisitions using cardiac gating tend to decrease the signal loss, however poor patient cooperation and multiple breath holds required for large volume acquisitions can make such studies inefficient. In this work, we present an algorithm to perform prospective cardiac and respiratory double gating and demonstrate artifact free images acquired during free breathing.


14:30         3120.     A Real Time Optical Motion Correction System Using a Single Camera and 2D Marker

Computer 21

Murat Aksoy1, Rexford Newbould1, Matus Straka1, Samantha Holdsworth1, Stefan Skare1, Juan Santos1, Roland Bammer1

1Stanford University, Stanford, California , USA

Correction of motion related artifacts remains one of the most important topics in MRI. In this study, we present a real-time prospective motion correction system that uses a single camera and a 2D marker to detect and correct for rigid body motion. With the camera placed inside the magnet bore and the marker mounted rigidly on patient’s head, it is possible to perform motion correction for both in-plane and through-plane motion using this system with high accuracy and stability.


15:00         3121.     Characterization and Correction of Artifacts from Dynamic Interaction Between Motion and Position Dependent Off-Resonance Patterns

Computer 21

Kiran Kumar Pandey1, Douglas Noll1

1University of Michigan, Ann Arbor, Michigan, USA

Movement during fMRI causes false activations and image registration introduces interpolation errors. Motion reorients the air-tissue interface in the brain and causes dynamically varying off-resonance patterns. These artifacts reduce accuracy of image registration and increase variability in time-series. This study characterized position dependent off-resonance artifacts in a susceptibility phantom and investigated effectiveness of CP gridding and Iterative image reconstruction methods with several fieldmap based off-resonance correction.  Dynamically updated fieldmaps most accurately characterized and compensated for position dependent off-resonance artifacts. The commonly used static fieldmap off-resonance correction method was least accurate.  This trend was also observed in quality of motion correction in phantom data.       


13:30         3122.     Examination of Cardiac-Related Motion in the Lower Thoracic and Lumbar Spinal Cord

Computer 22

Daphne Yau1,2, Chase R. Figley1, Celina Nahanni1, Patrick W. Stroman1

1Queen's University, Kingston, Canada

Spinal cord motion is the dominant source of error in functional magnetic resonance imaging and diffusion tensor imaging. Correcting for spinal cord motion induced by the cardiac cycle improves the sensitivity and reliability of these modalities. However, this motion has not been fully characterized in the lower thoracic and lumbar spinal cord. In the present study, we show minimal displacement of these spinal cord regions in association with the cardiac cycle. In conjunction with previous research, these findings demonstrate a model of spinal cord motion, in which maximal cervical motion progressively diminishes to a minimal displacement in the lumbar region.


14:00         3123.     High Resolution Image Co-Registration via Phase Modulation in the Reciprocal (K) Spatial Domain

Computer 22

James A. Goodman1, Charles S. Springer, Jr2

1Oregon Health and Science University, Portland, Oregon, USA; 2Oregon Health and Science University, Portland, USA

For many applications, it is essential that images be co-registered to a higher resolution than that intrinsic to the image. This is especially important for traditionally low spatial resolution data such as functional or DCE images. Pixel interpolation in the spatial domain can introduce artifacts and are often computationally burdensome. In this study we demonstrate the simplicity and precision of inter-image translation correction via post hoc reciprocal spatial domain (k-space) phase modulation. The co-registration method described here requires no addition image acquisition, field mapping, or pulse sequence modification.


14:30         3124.     Flow Compensation in Non-Balanced SSFP

Computer 22

Francesco Santini1, Oliver Bieri1, Klaus Scheffler1

1University of Basel/University Hospital, Basel, Switzerland

Flow compensation is a critical issue in non-balanced SSFP imaging because high first-order moments are achieved in the dephasing direction. Phase accumulation leads to steady-state disruption and consequent blurring artifacts and signal loss, especially in presence of long-T2 tissues. In this work, a theoretical explanation of the artifacts is given, showing how phase fluctuations influence the amplitude and the stability of the signal, and optimized flow compensated SSFP-FID and SSFP-echo sequences are presented. Finally, In vivo application of the sequence is demonstrated.


15:00         3125.     High-Resolution MR Spectroscopy in Inhomogeneous and Unstable Fields Via Intermolecular Zero-Quantum Coherences

Computer 22

Xi Chen1, Meijin Lin1, Tao Lin1, Zhong Chen1, Jianhui Zhong2

1Xiamen University, Xiamen, People's Republic of China; 2University of Rochester, Xiamen, New York, USA

A new iZQC pulse sequence with stroboscopic acquisition is designed to achieve high-resolution magnetic resonance spectroscopy in inhomogeneous and unstable fields. Primary results suggest potential applications for suppressions of motion-caused t1 noises and inhomogeneous broadenings in in vivo studies.


13:30         3126.     Improvement of Magnetic Field Homogeneity for Cardiac MRI at 3 Tesla

Computer 23

Miriam Rabea Kubach1, Axel Bornstedt1, Michael Schär2, Gerd Ulrich Nienhaus1, Volker Rasche1

1University Ulm, Ulm, Germany; 2The Johns Hopkins University School of Medicine, Baltimore, USA

A cardiac phase resolved B0-map was used to find the optimal shim settings depending on the current motion state. An optimal cardiac phase point for the assessment of the shim coefficients providing sufficient main field homogeneity over the entire cycle was assessed. It was found, that the shim values corresponding to mid diastole achieve the best results all over the cardiac cycle.


14:00         3127.     Application of the Pseudo-Polar Fourier Transform for In-Plane Rotation Correction of MR Images

Computer 23

Irtiza Ali Gilani1, Nadim Jon Shah1,2

1Institute of Neurosciences and Biophysics (Medicine), Research Centre Juelich, Juelich, Germany; 2Institue of Physics, University of Dortmund, Dortmund, Germany

A technique is proposed for in-plane rotational motion correction using the registration of the 2D low-resolution images, termed implicit navigators, inherent to the acquired MR images. Most MR image registration strategies use the iterative correlation approach. The correlation is usually performed in the polar Fourier domain, whereby the interpolation errors in the polar Fourier domain are problematic. In this work the registration is performed by applying the pseudo-polar Fourier transform to the low-resolution images which neither requires interpolation nor iteration. Rotational motion in EPI images obtained at 3T field strength was corrected.


14:30         3128.     Image Based Compensation of Breathing Motion Artifacts in MRI with Continuously Moving Table Acquisitions

Computer 23

Matthias Honal1, Ute Ludwig1, Simon Bauer1, Jochen Leupold1

1University Hospital Freiburg, Freiburg, Germany

MRI with continuously moving table is an efficient method to image arbitrary large body regions within one measurement. However, breathing motion causes problems since standard techniques to avoid breathing motion artifacts such as breath holding, prospective or retrospective gating are often not applicable or very inefficient. This work introduces a method which allows the reconstruction of artifact free images from data acquired during free breathing. Snapshots of arbitrary breathing states are acquired and consistently combined using image registration techniques. Typical breathing motion induced artifacts such as ghosting, blurring and signal cancellations are thus eliminated.


15:00         3129.     Physiological Motion Correction of ASL FMRI Measurement of Rhesus Monkey

Computer 23

Xiaodong Zhang1, Tsukasa Nagaoka1, Robbie Champion1, Timothy Q. Duong1

1Emory University, Atlanta, Georgia, USA

The Arterial Spin Labeling (ASL)-based CBF quantification is significantly vulnerable to the physiological motion because of the pair-wise subtraction of the control and labeling images. In this paper, a robust approach based on the optimal reference phase map was introduced for the physiological motion correction of the ASL functional MRI measurement. The resultant CBF temporal response and CBF activation map were improved dramatically. The method was demonstrated in the non-human primate (rhesus monkey) CBF measurement under the hypercapnia (5% CO2) condition.


13:30         3130.     3D Non-Linear Model-Driven Registration for Motion Corrupted DCE-MRI Data

Computer 24

Angela Caunce1, Giovanni A. Buonaccorsi1, Caleb Roberts1, Geoff JM Parker1

1The University of Manchester, Manchester, UK

Many dynamic contrast enhanced MRI sequences have some form of motion corruption which may jeopardise the estimation of microvascular parameters.  We present a model-driven registration which uses non-linear matrix transforms to bring the time series closer into alignment.  The success of the registration is illustrated using a new alignment measure and by an independent motion assessment.


14:00         3131.     A Fast Flow Compensation Technique for Self-Gated Sequences

Computer 24

Jessica Schulz1,2, Arne Hengerer2, Wolfhard Semmler1, Michael Bock1

1Deutsches Krebsforschungszentrum (dkfz), Heidelberg, Germany; 2Siemens Medical Solutions, Erlangen, Germany

Self-gating utilizes an additional short data acquisition in the pulse sequence for cardiac motion detection. Self-gating increases the TR of a pulse sequence, especially, if additional flow compensation gradients are used. Here, we demonstrate a new flow compensation scheme for self-gating with nearly no time penalty over conventional self-gating sequences without flow compensation.


14:30         3132.     Efficient Motion Artifact Reduction Technique Based on Post-Processing [Not Available]

Computer 24

Sang-Young Cho1, Eung-Yeop Kim1,12,2, Dong-Hyun Kim,1,23

1; 2; 3Yonsei University, Seoul, Republic of Korea

Post-processing motion correction techniques have previously been introduced to reduce motion artifacts. These techniques, which do not need any a priori information, have been shown to reduce translational motion artifacts effectively. One of the main limitations of these techniques however is the long processing time required to find the optimal image. Here, we propose a simple method to reduce this total processing time for metric based motion correction techniques.


15:00         3133.     Relevance of Cardiac-Gating in Longitudenal Diffusion Weighted MRI Studies

Computer 24

Josef Habib1, Jaroslav Hlinka1, Stamatios Sotiropoulos1, Christopher Tench1, Dorothee Auer1, Paul Simon Morgan1

1University of Nottingham, Nottingham, UK

While the presence of artifacts in ungated DW-images and the ability of gating to remedy these have been widely reported, the evaluation of benefits for the clinically important Diffusion-Tensor parameters has received less attention. This study aims to assess the relevance of cardiac-gating in clinical DWI-applications. To this end, the effects of cardiac-gating on precision and accuracy of Fractional Anisotropy and Mean Diffusivity measurements were analyzed, and subsequently put into context with alterations in FA of MS-patients. The observed results suggest that gating produces little benefit in group-studies but should be considered when precise individual acquisitions are needed.


RF Pulse Design

Hall D                                   Thursday 13:30-15:30                                                                                                                                           

13:30         3134.     Time-Optimal 3D Gradient Design for RF Shimming

Computer 15

Maryam Etezadi-Amoli1, Adam B. Kerr1, John M. Pauly1

1Stanford University, Stanford, California , USA

Spoke trajectories that play slice-selective RF subpulses along kz are known to be effective at RF shimming.  In this work, we use gradient time-optimization to achieve acceleration factors of greater than 10% for such spoke trajectories.  We present simulated RF shimming performance using the time-optimal gradients and assuming an eight coil parallel transmit system.


14:00         3135.     Magnitude-Constrained Spokes Design for B1+ Inhomogeneity Correction

Computer 15

Rolf F. Schulte1, Mika W. Vogel1, Adam B. Kerr2, Hans-Peter Fautz1, Patrick Gross1, Yudong Zhu3, Florian Wiesinger1

1GE Global Research, Munich, Germany; 2Stanford University, Stanford, California , USA; 3GE Global Research, Niskayuna, New York, USA

Transmit field (B1+) inhomogeneity is impairing image quality at higher field strengths (B0 3T). A common approach to improve homogeneity is to excite with a multi-channel excitation system in combination with the spokes trajectory, which is a 3D excitation sequence. A slice is excited by regular sinc-Gaussian type of pulses. The in-plane homogeneity is improved by placing a few spokes in-plane. Traditionally, the excitation was constrained in both phase and magnitude,

hence requiring a considerable amount of spokes and consequently a long overall pulse duration. In this work, we introduce a design constraining only the magnitude, hence considerably improving

B1+ homogeneity. The design is solved with a non-linear least squares optimisation, including B0 inhomogeneity and constraining the B1 amplitudes. Validation is performed on 3T with a torso-phantom.


14:30         3136.     Optimal Phase-Relaxed Design of Small-Tip-Angle Parallel Transmission RF Pulses

Computer 15

Dan Xu1, Kevin F. King1, Graeme C. McKinnon1

1General Electric Healthcare, Waukesha, Wisconsin, USA

The existing phase-relaxed parallel transmission RF pulse design relaxes the flat phase constraint in conventional small-tip-angle (STA) design by allowing a predetermined, spatially varying target phase profile to improve the quality of magnitude profile of the resulting transverse magnetization. In this work, the pulse design is reformulated as an optimization problem with a non-quadratic cost function where the phase constraint is completed removed. It is further shown that the gradient vector of the cost function can be expressed as a closed form and a nonlinear conjugate gradient algorithm can be applied to efficiently solve the optimization problem. Bloch simulation results demonstrate that the optimal phase-relaxed design can achieve significantly better magnitude profile and/or lower RF power than the conventional STA design.


15:00         3137.     Ultra-Short 2D RF Pulse for Reduced Field-Of-View SSFP Imaging

Computer 15

Jing Yuan1, Chang-Sheng Mei1,2, Lawrence P. Panych1

1Department of Radiology, Brigham and Women's Hospital, Harvard medical School, Boston, Massachusetts, USA; 2Department of Physics, Boston College, Boston, Massachusetts, USA

Balanced SSFP is a fast imaging sequence providing high signal-to-noise ratio. 2D RF pulse could be used for further shortening the scan time for SSFP by reducing phase encoding number in reduced field-of-view (rFOV). However, the critical requirement of short TE and TR make the use of 2D RF pulse in SSFP challenging. Ultra-short 2D RF pulse is specifically designed for rFOV imaging for SSFP. Sub-pulse duration is as short as 384us for 10mm slice thickness with max slew rate of 150T/m/s. 2.5x acceleration factor has been successfully achieved on GE Signa 1.5T and 3T scanners


13:30         3138.     Segmented 2D-Selective RF Excitations Based on a Weighted Blipped-Planar Trajectory

Computer 16

Jürgen Finsterbusch1,2

1University Medical Center Hamburg-Eppendorf, Hamburg, Germany; 2Neuroimage Nord, Hamburg-Kiel-Lübeck, Germany

2D-selective RF excitations based on a segmented blipped-planar trajectory suffer from a reduced signal amplitude due to the lower flip angles of segments covering outer k-space lines. In applications that involve averaging, like single-voxel spectroscopy, a weighted blipped-planar trajectory can be used that favours larger flip angles over averaging. While segments with high k-space intensity are averaged as usual, a larger flip angle is assigned to low intensity segments so that effectively the same signal amplitude of the segment can be achieved without averaging. This approach considerable improves the signal-to-noise ratio efficiency without degrading the excitation profile.


14:00         3139.     2D-Selective RF Excitations Based on a Half-Fourier Blipped-Planar Trajectory [Not Available]

Computer 16

Jürgen Finsterbusch1,2

1University Medical Center Hamburg-Eppendorf, Hamburg, Germany; 2Neuroimage Nord, Hamburg-Kiel-Lübeck, Germany

2D-selective RF excitations based on a blipped-planar trajectory are well-suited for applications like reduced field-of-view imaging but suffer from the fact that half of the trajectory&#146;s duration needs to be considered in the echo time. Thus, the usage of this trajectory for applications that require a short echo time, like proton-density weighting or single-voxel spectroscopy, is hampered. By applying the half-Fourier method known from slice-selective excitations, this problem can be solved using trajectories that start in outer k-space and end in the centre because only half of a single line of the trajectory contributes to the echo time.


14:30         3140.     Constant Time VERSE for RF Amplitude Reduction in Spectral-Spatial Pulses with Improved Timing Robustness

Computer 16

Peder E. Z. Larson1, Adam B. Kerr2, John M. Pauly2, Daniel B. Vigneron1

1University of California - San Francisco, San Francisco, California , USA; 2Stanford University, Stanford, California , USA

Spectral-spatial RF pulses often must contend with limits on RF power and amplitude, particularly for highly spatially selective pulses, high-field applications, and hyperpolarized 13C.  We present a new method for reducing the peak amplitude in spectral-spatial RF pulses without introducing significant gradient delay sensitivity.  Our method is an adaptation of the VERSE algorithm applicable to spatial subpulses because it maintains a constant pulse duration.  It also includes an additional slew rate limit that reduces the allowable rate for larger RF amplitudes, improving the pulse performance in the presence of gradient delays.


15:00         3141.     High Bandwidth Low Power Spatial Saturation Pulses for 7T

Computer 16

Douglas Arthur Charles Kelley1

1GE Healthcare, San Francisco, California , USA

An extension to the Shinnar-Le Roux algorithm is presented allowing the design of scalable quadratic phase RF pulses, achieving significantly higher bandwidths with similar peak amplitudes and pulse widths. The algorithm is applied to the design of spatial saturation pulses for 7T. The performance of the pulses is verified in a high dielectric permittivity phantom.


13:30         3142.     Selective Adiabatic Refocusing Pulse Pair for 3D RARE

Computer 17

Ziqi Sun1, Jay L. Zweier1

1The Ohio State University, Columbus, USA

A compact selective adiabatic full passage (AFP) pulse pair, consisted of two 90 degree hyperbolic secant (HS1_R20) pulses of alternate frequency sweep (AFS) directions, was successfully applied for spin refocusing in a 3D RARE pulse sequence. In comparison to an amplitude-modulated refocusing pulse and to a single 180 degree HS1_R20 AFP refocusing pulse, the AFS-AFP pulse pair substantially improved signal sensitivity and uniformity across the 3D volume, which is attributed to the effective compensation of the nonlinear phase dispersion and off-resonance effect associated with selective AFP pulses.


14:00         3143.     Selective Adiabatic Pulses for T1-Weighted Contrast Enhancement at 0.38 T

Computer 17

Ziqi Sun1, Sergey Petryakov1, Wael Alzawahra1, Jay L. Zweier1

1The Ohio State University, Columbus, USA

T1-weighted contrast enhancement was achieved in a MnCl2 doped phantom and an isolated rat heart perfused with free radical (TAM) doped St. Thomas solution using selective adiabatic full passage (AFP) inversion and excitation pulses (HS1_R15) in a 3D inversion-recovery (IR) GRE pulse sequence. In comparison to an amplitude modulated selective inversion pulse, the selective adiabatic inversion pulse substantially increased signal sensitivity and image contrast, which is ascribed to the mutual cacellation of the nonlinear phase dispersion between the inversion AFP pulse and the excitation AFP pulse in the IR-GRE3D sequence.


14:30         3144.     The Steady State Properties of Actual Flip Angle Imaging (AFI) [Not Available]

Computer 17

Kay Nehrke1

1Philips Research Europe, Hamburg, Germany

The RF and gradient spoil scheme of the Actual Flip Angle Imaging (AFI) sequence has been studied. It is shown that conventional RF spoiling using a linearly increasing spoil phase shift between successive RF pulses may interfere with the formation of a proper steady state for this specific sequence. Instead, an adapted RF spoil regime is derived, which essentially preserves the quadratic increase of the spoil phase over time. In addition, the spoiler gradient arrangement is adjusted to improve the off-resonance sensitivity of the sequence. Phantom experiments show a significantly improved image quality compared to the original spoiling regime.


15:00         3145.     Regularized B1+ Map Estimation with Slice Selection Effects

Computer 17

Amanda Kay Funai1, Jeffrey A. Fessler1, William Grissom1, Douglas C. Noll1

1University of Michigan, Ann Arbor, Michigan, USA

Spatially varying contrast and signal at high fields make a B1+ map a requirement in parallel transmit excitation.  The standard method for estimating the magnitude, using the double angle formula, ignores noise and neglects effects from slice selection.  This new regularized method incorporates multiple coils and multiple tip angles to estimate the magnitude and phase of the field map and accounts for slice selection effects based on the achieved slice profile.  Phantom and experimental results show a smoother estimate and increased accuracy with the proposed method.  The phase estimate can be used to design pulse sequences in parallel excitation.


Post-Nyquist Imaging

Hall D                                   Thursday 13:30-15:30                                                                                                                                           

13:30         3146.     Accelerated HYPR Reconstruction

Computer 18

Andres Carrillo1, Reed F. Busse2, Jean H. Brittain2, Oliver Wieben3, Yan Wu3, Charles A. Mistretta3, Frank R. Korosec3

1GE Healthcare, Evanston, Illinois, USA; 2GE Healthcare, Madison, Wisconsin, USA; 3University of Wisconsin, Madison, Wisconsin, USA

The use of HYPR in time resolved contrast-enhanced angiography can produce temporal resolutions up to 100 times greater than traditional unaccelerated methods. However, the increase in computational requirements and data throughput can result in long reconstruction times, hampering the use of the technique. We propose an optimized, distributed reconstruction algorithm capable of reducing reconstruction times from several hours to several minutes, enabling the clinical use of the HYPR CE-MRA.


14:00         3147.     L-Map: Exploiting Spatial-Temporal Correlation of Phase in MRI

Computer 18

Kang Wang1, Steven Kecskemeti1, Rafael O'Halloran1, Sean Fain1, Kevin Johnson1, Oliver Wieben1, Jiang Du2, Charles Mistretta1

1University of Wisconsin-Madison, Madison, Wisconsin, USA; 2University of California, San Diego, San Diego, California , USA

Several image reconstruction techniques that are aimed to exploit the spatial-temporal correlation of MR magnitude images, such as Highly Constrained backPRojection (HYPR) and HYPR with Local Reconstruction (HYPR-LR) have been developed recently. These techniques have been developed for magnitude processing and cannot be directly applied to phase sensitive datasets due to phase wrapping and potential signal cancellations in the composite image. Here we propose a novel approach, the “L-map” concept, to overcome these limitations and evaluate the algorithm in simulations.


14:30         3148.     MR Spectroscopic Imaging of Short T2 Tissue Using Complex Division (CD) HYPR-LR Reconstruction

Computer 18

Kang Wang1, Rafael O'Halloran1, Sean Fain1, Steven Kecskemeti1, Oliver Wieben1, Kevin Johnson1, Charles Mistretta1, Jiang Du2

1University of Wisconsin-Madison, Madison, Wisconsin, USA; 2University of California, San Diego, San Diego, California , USA

Highly Constrained Back-Projection (HYPR) and HYPR with Local Reconstruction (HYPR-LR) techniques have been developed to exploit the spatial-temporal correlation for time-resolved MR using magnitude data. Here we propose a complex-valued HYPR-LR method which can be applied to other MR applications where complex images are needed, such as MR spectroscopy. We demonstrate that the Complex Division HYPR-LR, combined with ultrashort echo time (UTE) technique, can provide quantitative spectroscopic information of short T2 tissues, such as cortical bone.


15:00         3149.     Image Reconstruction of Variable Density Undersampled EPI Images

Computer 18

Annie M. Tang1, Man Cheuk Ng1, Edmund Y. Lam1

1The University of Hong Kong, Pokfulam, Hong Kong

We applied a mathematical theory, Compressive Sensing (CS), for image reconstruction to EPI images. With the use of CS, it is possible to undersample the k-space data while preserving image quality. Thus, it allows a more extend coverage of the imaged object per unit time. EPI image qualities resulted from the CS reconstruction and its feasibility of application in functional MRI are discussed. Results indicated that CS outperforms traditional time reduction techniques. The statistical mapping in a typical fMRI experiment is comparable to the fully sampled EPI data.


13:30         3150.     Prediction of HYPR and HYPR LR Performance Based on Image Sparsity and Temporal Correlation

Computer 19

Julia V. Velikina1, Yan Wu1, Charles A. Mistretta1

1University of Wisconsin - Madison, Madison, Wisconsin, USA

We derive analytical formulae for the estimation of the reconstruction error using HYPR and HYPR LR algorithms. We conclude that the performance of these algorithms for each potential clinical application can be accurately predicted based on the a priori knowledge of image sparsity and expected temporal behavior of the image time series.


14:00         3151.     Compressed Sensing MRI with Random B1 Field

Computer 19

Florian M. Sebert1, Yi Ming Zou1, Bo Liu1, Leslie Ying1

1University of Wisconsin - Milwaukee, Milwaukee, Wisconsin, USA

Considerable attention has been focused on the design of practical sampling schemes to apply compressed sensing (CS) to MR. In this abstract, we revisit the sufficient conditions for CS matrices, and show that if the coil sensitivities or the RF excitation profiles can be designed to be spatially random, then the image can be recovered from uniformly undersampled k-space data. Our finding provides an alternative to the random sampling design for reducing the acquisition time using CS. In addition, it has the advantage that the number of samples needed is less than what is needed in the random sampling scheme.


14:30         3152.     Randomly Perturbed Radial  Trajectories for Compressed Sensing MRI

Computer 19

Ali Bilgin1,2, Ted P. Trouard1, Arthur F. Gmitro1, Maria I. Altbach1

1University of Arizona, Tucson, Arizona , USA

The recently introduced Compressed Sensing (CS) theory has the potential to significantly accelerate data acquisition in MRI. Recent results in CS theory indicate that introducing randomization into regularly structured trajectories can be beneficial. We introduce randomly perturbed radial trajectories for CS MRI and illustrate their potential benefits with examples.


15:00         3153.     Data Reordering for Improved Constrained Reconstruction from Undersampled K-Space Data

Computer 19

Ganesh Adluru1, Edward VR DiBella1

1University of Utah, Salt Lake City, USA

Here we propose a new data reordering method to improve the constrained reconstruction techniques from undersampled k-space data in MRI.  In the data reordering technique the undersampled data are reordered in the signal space according to an a priori determined ordering and then appropriate constraints are applied within an iterative reconstruction.  The method is widely applicable to reconstructions in which the data do not perfectly match the constraints being used and it can be used in the contexts which are based on regularization techniques.  The method can be used to accelerate 2D and multi-image data acquisition cases in MRI.


13:30         3154.     SparseSENSE: Randomly-Sampled Parallel Imaging Using Compressed Sensing

Computer 20

Bo Liu1, Florian M. Sebert1, Yi Ming Zou1, Leslie Ying1

1University of Wisconsin - Milwaukee, Milwaukee, Wisconsin, USA

Recent work has applied compressed sensing (CS) to reduce scanning time in conventional Fourier imaging and demonstrated impressive results. In this abstract, we investigate the structure of the sensitivity encoding matrix in parallel imaging, and apply CS to parallel imaging to achieve an even higher reduction in scanning time than what can be achieved by each individual method alone. Our experiments show that the proposed method, named SparseSENSE, can achieve a reduction factor higher than the number of channels.


14:00         3155.     Homotopic L0-Minimization for Highly-Undersampled MRI Reconstruction

Computer 20

Joshua D. Trzasko1, Armando Manduca1

1Mayo Clinic, Rochester, Minnesota, USA

Compressive Sensing and related L1-minimization techniques have recently been demonstrated to accurately and efficiently reconstruct sparse or compressible MR images even at sampling rates far below the Nyquist limit.  In this work, we propose an alternative minimization framework based on homotopic approximation of the L0-recovery problem and show that  accurate MRI reconstructions are possible at sampling rates even lower than are achievable using L1-based methods.


14:30         3156.     Matrix Formulation and Tikhonov Regularization of HYPR Reconstruction

Computer 20

Maximilian Haeberlin1,2, Stefan Skare1, Rexford D. Newbould1, Klaas Paul Pruessmann2, Roland Bammer1

1Stanford University, Stanford, California , USA; 2ETH Zurich, Zurich, Switzerland

A matrix formulation for HYPR reconstruction is presented, allowing to incorporate arbitrary regularization methods in iterative variants of HYPR. It is shown that Tikhonov regularization stabilizes recently proposed CG-HYPR without compromising speed or image quality.


15:00         3157.     Redundant Spatial Harmonic Information in Zeugmatography with Linear Encoding (R-SHIZLE) Theoretically Encodes Intra-Acquisition Decay

Computer 20

Andrew S. Nencka1, Andrew D. Hahn1, Daniel B. Rowe1

1Medical College of Wisconsin, Milwaukee, Wisconsin, USA

Redundant spatial harmonic information is acquired in MRI because of the real-valued nature of the object being imaged. The redundancy is manifested as conjugate symmetry about the k-space origin. This symmetry is broken by both magnetic field inhomogeneity, intra-acquisition T2 decay and noise. In this abstract we simulate the generated k-space signal, including a non-negligible magnetic field inhomogeneity and a spatially varying T2 profile. We then correct the magnetic field inhomogeneity and recover the original T2 profile, with minor error, based upon the broken symmetry of the simulated k-space data. This offers the potential for quantitative T2 or T2* time series instead of usual T2 or T2* weighted image time series.


Image Segmentation

Hall D                                   Thursday 13:30-15:30                                                                                                                                           

13:30         3158.     Spherical Harmonic Representation Based Haptic Rendering for Medical Image Perception

Computer 21

Zhengyi Yang1, Quang Tieng1, Viktor Vegh1, Deming Wang1

1University of Queensland, Brisbane, Australia

Haptics is referred as the technologies of generating artificial sense of touch in the interaction with the objects in virtual space. The use of visual interpretation in conjunction with a haptic device can provide a higher level of perception of medical images towards diagnosis and treatment planning. Spherical harmonics were found to be a feasible representation for haptic rendering with very good feedback characteristics. Spherical harmonics are a compact representation and can be efficiently used for both graphic and haptic representations.Haptic feedback provides another channel for the information flow in medical image perception. Multimodality images can be displayed simultaneously with less confusion by using visual and haptic feedback separately, which is envisioned as a better way for image fusion interaction. Haptically guided medical image exploration can aid the user in various tasks, such as finding and measuring anatomical and pathological features or investigating the spatial relationship between them. Force feedback can improve user performance in image manipulation tasks, such as interactive landmark selection, organ delineation and image registration.


14:00         3159.     Structure-Specific White Matter Analysis of Amyotrophic Lateral Sclerosis

Computer 21

Hui Zhang1, Paul A. Yushkevich1, John H. Woo1, Sumei Wang1, Elias R. Melhem1, James C. Gee1

1University of Pennsylvania, Philadelphia, Pennsylvania, USA

In this study, we evaluated the use of a structure-specific WM analysis (SSWMA) framework to detect the location and magnitude of FA changes in patients with ALS.  The SSWMA framework allows us to focus our analysis on the corticospinal tracts.  The significant FA reductions were identified, which demonstrates the efficacy of the SSWMA framework.


14:30         3160.     Automatic Segmentation of Intra-Abdominal and Subcutaneous Adipose Tissue in 3D Whole Mouse MRI

Computer 21

Petter Ranefall1, Abdel W. Bidar1, David J. Svensson1, Paul David Hockings1

1AstraZeneca, Mölndal, Sweden

We developed fully automatic segmentation of intra-abdominal (IAT) and subcutaneous (SAT) adipose tissue in mice. The novel algorithm outlines the body, extracts fat, and segments IAT and SAT of in vivo high resolution whole mouse 3D images. It uses the Narrowest Passage Transform which codes each pixel with the radius of the narrowest passage on the widest possible 3D path to the body outline, and competitive region growing such that competing classes meet at narrow passages. The new method was tested on 32 fat fed mice independent to those used for algorithm development and shows good correlation to manual segmentation.


15:00         3161.     Accurate Assessment of Muscle/fat Distribution on MR Images of the Thigh

Computer 21

Vincenzo Positano1, Tore Christiansen2, Maria Filomena Santarelli1, Amalia Gastaldelli1, Steffen Ringgaard2, Luigi Landini3

1Institute of Clinical Physiology, Pisa, Italy; 2Aarhus University Hospital, Aarhus, Denmark; 3University of Pisa, Pisa, Italy


13:30         3162.     Structure-Specific Statistical Mapping of White Matter Tracts

Computer 22

Paul A. Yushkevich1, Hui Zhang1, Tony Simon2, James C. Gee1

1University of Pennsylvania, Philadelphia, Pennsylvania, USA; 2M.I.N.D. Institute, University of California, Davis, California , USA

We present a novel technique for analyzing diffusion imaging data in white matter studies. Data associated with six major white matter tracts is analyzed with the help of a geometrical model that exploits the sheet-like structure of these tracts.


14:00         3163.     Development of a Reliable Analysis Method for Measurements of Breast Volume and Fibroglandular Tissue Volume in MRI

Computer 22

Ke Nie1, Siwa Chan2, Ivy Chau2, Tiffany Tseng1, Jeon-Hor Chen1, Orhan Nalcioglu1, Min-Ying Su1

1University of California, Irvine, Irvine, California , USA; 2China Medical University Hospital, Taichung 404, Taiwan

A reliable method using computer-assisted segmentation for the breast and the fibroglandular tissue on breast MRI was developed. Breast segmentation was achieved based on individual woman¡¯s body landmarks for initial estimation, then with a combination of fuzzy C-means, b-spline fitting, and dynamic searching algorithms. 12 cases were selected for reproducibility test. The inter- and intra-operator consistency was evaluated, and the correlation was found to be higher than 0.97. This method may be applied to monitor the density change over time to evaluate the benefits or risks for each individual woman when considering chemoprevention or hormonal replacement therapy.


14:30         3164.     Robust Segmentation and Classification of Heterogeneous Myocardial Infarct Zones

Computer 22

Jay S. Detsky1, Alexander J. Dick1, Graham A. Wright1

1Sunnybrook Health Sciences Centre, Toronto, Canada

Myocardial infarcts can be comprised of heterogeneous "gray zones" that may lead to ventricular arrhythmias.  Image analysis and clustering tools have been developed and applied to inversion-recovery SSFP images to segment and classify the infarct core, gray zone, healthy myocardium, and blood.  Pixel-by-pixel parameter maps are derived from the images and a modification of the fuzzy C-means algorithm is then applied.  This method has been shown to be more robust than the conventional methods for the delineation of infarct gray zones.


15:00         3165.     Segmentation of MR Brain Images with Intensity Correction and Partial Volume Averaging

Computer 22

Sushmita Datta1, Balasrinivasa R. Sajja1, Renjie He1, Joseph M. Dieber1, Ponnada A. Narayana1

1Medical School, University of Texas Health Science Center at Houston, Houston, USA

The classification of brain tissues is important factor to follow changes in tissue volumes. However, the intensity non-uniformity and partial volume averaging (PVA) within voxels compromise accurate volumetric measurements. An iterative segmentation procedure that explicitly considers PVA effect along with bias field correction is proposed. Significant improvement in segmentation was observed with this method.


13:30         3166.     Atlas-Based Segmentation for Quantitative Analysis of Brain Structures in the Rhesus Monkey

Computer 23

Louis Collins1, Alexandre Coimbra2, Marie Holahan2, Richard Hargreaves2, Jacquelynn Cook2, Donald S. Williams2, Stephen Frey1

1McGill University, Montreal, Canada; 2Merck Research Laboratories, West Point, Pennsylvania, USA

Animal models are often used in pre-clinical studies of potential pharmaceuticals to study physiological mechanisms or evaluate efficacy or toxicity. We have designed and implemented an automated atlas-based segmentation (ABS) procedure for the analysis of rhesus macaque brain MRI data. This study briefly describes the image processing pipeline and presents initial results from the analysis of a group of rhesus macaques. This study shows feasibility of automatic regional segmentation of brain MRI data from rhesus monkeys.


14:00         3167.     Knowledge-Based Left Ventricle Segmentation and Partial Volume Calculation in Cardiac Cine MRI

Computer 23

Hae-Yeoun Lee1, Yi Wang1

1Cornell University, New York, USA

In this study, we propose a knowledge-based left ventricle segmentation and partial volume calculation algorithm to segment short-axis cine cardiac MRI. Coil sensitivity of magnitude image is corrected. The intensity statistics of left ventricle and myocardium are estimated by edge classification. Graph searching and expansion are applied to detect the myocardium. We segment the left ventricle using a region-growing scheme and then calculated partial volume effects by a weighting function from statistics of the left ventricle and myocardium. Using 38 subjects, we measure blood volume, ejection fraction, and myocardium mass and compare with manual contour tracing.


14:30         3168.     Towards a New Automated Segmentation Method for Numerous Tissues in Reference to a Whole-Body 3D Anatomical Template

Computer 23

Chu-Fang Lin1, Christopher M. Collins2, George Kesidis1, David J. Miller1

1The Pennsylvania State University, University Park, Pennsylvania, USA; 2Penn State College of Medicine, Hershey, Pennsylvania, USA

We developed a method to identify the location of specific tissues and organs using patient-specific 3D images and with reference to a labeled 3D template of a single “reference” body. Our segmentation approach employs an active contours algorithm and a Markov random field segmentation algorithm (MRF), with optimization based on mean-field annealing (MFA). We have derived satisfactory results when the body shape in patient-specific scans is similar to the labeled template. To obtain robust tissue/organ segmentation for diverse body shapes, we are working on deriving optimal local 3D matching that accounts for changes in tissue location, orientation, and volume.


15:00         3169.     SNR Performance of Automated Geodesic Active Contour Based Liver Segmentation

Computer 23

Dattesh D. Shanbhag1, Ajay Narayanan1, Kajoli Krishnan1, Patrice Hervo2, Rakesh Mullick1

1GE Global Research, Bangalore, India; 2GE Healthcare, Buc, France

SNR performance of geodesic active contour based automated segmentation algorithm is evaluated for 3D MRI of liver. Segmentation was performed on contrast-enhanced, T1-weighted liver volumes obtained from five patients, consequently degraded to lower SNR. The automated segmented liver volumes were compared with manually segmented liver volumes and kappa statistic (&[kappa]) calculated. Degrading the SNR from original ~294 to 12 changed &[kappa] by less than 5%. The results suggest that the algorithm will perform robustly (&[kappa] > 0.8) for SNR ~= 9, translating into faster scans and shorter breath hold times for obtaining 3D MRI liver data for liver volumetry.


13:30         3170.     MR Travel to Scan Image Processing for Real-Time Liver Identification [Not Available]

Computer 24

Sebastian Peter Michael Dries1, Daniel Bystrov1, Vladimir Pekar2, Harald S. Heese1, Peter Koken1, Jochen Keupp1, Peter Börnert1

1Philips Research Europe, Hamburg, Germany; 2Philips Research North America, Markham, Canada

A method to detect the liver region on-the-fly in continuously moving bed magnetic resonance imaging was developed and tested on 3-D image data from ten volunteer survey examinations. The proposed method is based on first detecting the superior and inferior margin of the liver and next adapting a deformable liver model to the data. While the superior margin of the liver was detected with an accuracy of less than one voxel, the detection of the inferior margin was less precise with an accuracy of 4±3 voxels. For the envisioned application of diagnostic scan geometry definition, this detection accuracy was deemed acceptable, taking into account that the shape was captured well and the volume was not underestimated.


14:00         3171.     Robust Anatomy Recognition Approach for Automated Scan Planning of Spine MRI Examinations [Not Available]

Computer 24

Harald Sepp Heese1, Daniel Bystrov1, Vladimir Pekar2, Sebastian P. M Dries1, Rüdiger Grewer1, Chiel J. den Harder3, Rene Bergmans3, Arjan W. Simonetti3, Arianne M. van Muiswinkel3

1Philips Research Europe, Hamburg, Germany; 2Philips Research North America, Markham, Canada; 3Philips Medical Systems, Best, Netherlands

Robust, unattended and fast anatomy recognition is the key pre-requisite for automated MRI scan planning. With its segmental, repetitive anatomy, spine adds the difficulty of level identification to previously existing approaches. The proposed method uses a combination of filtering techniques, prior knowledge on geometric properties of adjacent intervertebral discs as well as a statistical model of the lumbosacral transition to identify and label the spine column in presence of anatomic variability due to posture or disease. Recognition rate and processing speed from an evaluation on 90 cases prove clinical feasibility of the method.


14:30         3172.     Real-Time Myocardial Segmentation in MRI

Computer 24

Qi Duan1, Andrew Francis Laine1, Vinay M. Pai2

1Columbia University, New York, New York, USA; 2SUNY Upstate Medical University, Syracuse, New York, USA

In order to quantitatively evaluate cardiac functional images, segmentation of endocardial and epicardial boundaries of the myocardium is essential. Recent advances of high speed imaging, such as Phase Train Imaging (PTI), can provide large image data sets with very high temporal resolution that can be very valuable for diagnosing and detecting systolic dyssynchrony. An automated real-time segmentation was developed. The proposed method was tested on 414 frames of PTI data. The performance of myocardial segmentation was visually and quantitatively validated. Implemented in Matlab©, current method took less than 1.2 ms per cardiac phase, allowing realization of true real-time online segmentation.


15:00         3173.     Automated Liver Volume Assessment from Contrast Scans: Impact of Pathology

Computer 24

Ajay Narayanan1, Kajoli Banerjee Krishnan1, Rakesh Mullick1, Uday Patil1, Patrice Hervo2

1GE Global Research, Bangalore, India; 2GE Healthcare, Buc, France

Liver volume assessment for transplant, pathologic and metabolic changes using MR has primarily been via semi-automated methods. We have developed an automated method based on statistically driven data adaptive geodesic active contours to extract the liver from fast 3D high resolution T1-weighted, LAVA scans. This study is aimed at evaluating the sensitivity of our method to a range of patient pathologies against expert manual segmentation. The algorithm provides consistent assessment of liver volume compared to the expert manual segmentation (k=0.88) for embedded and focused liver pathology. The relative performance tends to degrade in the presence of peripheral liver pathology.


Spectroscopy Localization, Quantitation, & Methodology

Hall D                                   Monday 14:00-16:00                                                                                                                                             

14:00         3174.     Comparison of Two Approaches to Model the Macromolecule Spectrum for the Quantification of Short TE 1H MRS Spectra at 14.1T

Computer 25

Cristina Cudalbu1, Vladimir Mlynárik, Lijing Xin, Rolf Gruetter,2

1 Ecole Polytechnique Federale de Lausanne, Lausanne, Switzerland; 2Departments of Radiology, Universities of Lausanne and Geneva, Switzerland

The aim of the present study was to assess two approaches based on using LCModel which take the macromolecule contributions into account in the quantification step of in vivo rat brain spectra at 14.1 T: 1) the spectrum of macromolecules measured in vivo; and 2) the built-in LCModel spline baseline. Even if the fitted splines reproduce the in vivo macromolecules very well, the measured macromolecular “baseline” represents an additional prior knowledge and lead to a more accurate and reliable quantification at higher field strengths, which was attributed to an effectively increased spectral resolution of the macromolecule components.


14:30         3175.     Comparison of Spectral Fitting Methods for Overlapping J-Coupled Metabolite Resonances

Computer 25

Ati Gonenc1, Varanasi Govindaraju1, Andrew A. Maudsley1

1University of Miami, Miami, Florida, USA

The results of an automated spectral analysis package capable of performing two dimensional prior-knowledge fitting of multi-TE spectra, one dimensional fitting of TE-averaged spectra, and one dimensional fitting of conventional (single-TE) PRESS spectra, are presented for single-voxel data acquired at 3T. Improved performance and accuracy has been observed for the spectral fitting using the full 2D model in comparison to the TE-averaged and conventional PRESS fit for both in vivo and in vitro data.


15:00         3176.     Angular Dispersion and Average Orientation of Bulk Triglycerides: Influence on the Proton Lineshape from Extramyocellular Fat

Computer 25

Anthony Khuu1, Jimin Ren1, James Murdoch2, Ivan Dimitrov2, Donald Woessner1, A. Dean Sherry1, Craig Malloy1,3

1University of Texas Southwestern Medical Center, Dallas, Texas, USA; 2Philips Medical Systems, Cleveland, USA; 3VA North Texas Healthcare System, Dallas, USA

Accurate determination of intramyocellular lipid content (IMCLs) by single-voxel proton spectroscopy or chemical shift imaging is important in clinical research projects.  However, there is often significant overlap with extramyocellular lipids (EMCLs) due in part to angular dispersion of the strands of extracellular fat.  Simulations of proton spectra generated from known fiber orientations and dispersion, analyzed using conventional symmetric lineshapes, resulted in overestimation of IMCL content. A new fitting algorithm assuming a Gaussian distribution of EMCLs centered from 0 to 90 degrees was developed and used for analysis of spectra acquired from calf muscle of healthy humans at 7T.


15:30         3177.     A Note on the Accurate Model-Based Spectral Fitting of Proton MRS in the Frequency Domain

Computer 25

Zhengchao Dong1,2, Zhishun Wang1,2, Bradley Peterson1,2

1Columbia University, New York, New York, USA; 2New York State Psychiatric Institute, New York, New York, USA

Model based spectral fitting is widely used in quantitative in vivo MRS. However, there is a discrepancy between the discrete Fourier transform (DFT) reconstructed MR spectrum and continuous Fourier transform (CFT) deduced spectral models, which will introduce systematic errors in the spectral quantification, especially when the quantification is based on ratios of spectral peaks. Although this discrepancy can be easily eliminated, they are often overlooked in the long TE MRS spectral fitting. We show here by computer simulation and in vivo MRS data that eliminating the difference between DFT and CFT spectra allows accurate spectral fitting in the frequency domain.


14:00         3178.     Interleaved Dual-Angle Measurements for the Correction of Partial Saturation in 31P MR Spectroscopy

Computer 26

Orlando Lopez1, Damian J. Tyler2, Mark A. Cole2, Carolyn A. Carr2, Daniel J. Stuckey2, Edward Lakatta1, Kieran Clarke2, Richard G. Spencer1

1NIH / National Institute on Aging, Baltimore, Maryland, USA; 2University of Oxford, Oxford, UK

Use of short repetition times (TR), relative to metabolite T1's, is common in NMR spectroscopy of biological samples.  However, accurate quantification of metabolite concentrations from spectral resonances acquired with short TR values requires proper correction for saturation effects.  The goal of this work was to determine whether accurate metabolite measurements could be performed by continuous dual-angle data collection using partially saturated spectra in the setting of chemical exchange.  Simulations appropriate to a hypoxic intervention experiment in the heart were performed.  Results demonstrated the feasibility of using continuous dual-angle acquisition to correct for partial saturation of metabolite resonances in bioenergetic experiments.


14:30         3179.     Reproducibility of 31P Cardiac Magnetic Resonance Spectroscopy at 3T

Computer 26

Lowri E. Cochlin1, Damian J. Tyler1, Yaso Emmanuel1,2, Lucy Hudsmith3, Cameron J. Holloway1,2, Stefan Neubauer3, Kieran Clarke1, Matthew Robson3

1University of Oxford, Department of Physiology, Anatomy and Genetics, UK; 2Department of Cardiovascular Medicine, UK; 3University of Oxford, Department of Cardiovascular Medicine, UK

The purpose of this work was to investigate the application, reproducibility and reliability of cardiac 31P MRS at 3 Tesla, with a view to taking advantage of the increased signal available at 3T by improving spatial resolution and increasing myocardial specificity. We present a technique that routinely provides high quality spectra along with a robust analysis method which is free from potential bias through spectral information or quality. The presented technique is suitable for widespread application both across research groups and across disease models.


15:00         3180.     31P 3D K-Space Weighted MRSI with Adiabatic Excitation: 3D Absolute Quantification of Phosphorus Metabolites in Human Liver

Computer 26

Marek Chmelik1,2, Albrecht Ingo Schmid1,2, Stephan Gruber2, Julia Szendroedi3, Martin Krssak2, Siegfried Trattnig2, Ewald Moser2, Michael Roden,13

1Karl-Landsteiner Institute for Endocrinology and Metabolism, Vienna, Austria; 2Medical University of Vienna, Vienna, Austria; 3Hanusch Hospital, Vienna, Austria

A protocol for measuring absolute concentrations of hepatic phosphorus metabolites by using a 3D k-space-weighted SI sequence with B1 homogeneity insensitive adiabatic pulses was designed and tested. Compared to excitation pulses used in previous studies adiabatic pulses used here do not need exact pulse calibration. In addition k-space weighted acquisition and 3T B0 field lead to relative high spatial resolved spectra (3.65ml) in reasonable time (34 min). The processing tool allows user friendly data processing and quantifying of hundreds of spectra per patient and producing absolute metabolic maps of 31P metabolite concentration distribution in the liver.


15:30         3181.     Automatic 31P MRS Quantification in the Human Brain Based on OASIS-HSVD Algorithm

Computer 26

Xin Wang1, Jing-Huei Lee1

1University of Cincinnati, Cincinnati, USA

A new algorithm termed Optimal Adaptive Separation of Interference Signal Hankel Singular Value Decomposition (OASIS-HSVD) is proposed as an automatic time domain method to adaptively separate distorted baselines and to quantify the target peaks from interested chemical compounds of human 31P MRS data at 4 T.  This algorithm was tested on both simulation and in vivo data. The simulation used the Monte Carlo method for testing the robustness and accuracy of this algorithm. In vivo data was obtained from a 3D 31P MRS from human brain. Both studies demonstrate that the OASIS-HSVD is a promising algorithm for automatic 31P MRS quantification.


14:00         3182.     Lactate Imaging with Hadamard Encoded Slice Selective SelMQC-CSI

Computer 27

Stephen Pickup1, Seung Cheol Lee1, Jerry D. Glickson1

1University of Pennsylvania, Philadelphia, Pennsylvania, USA

We¡¯ve presented here lactate imaging of tumor. Previously published lactate editing technique, the SelMQC sequence, is a doubly frequency selective, double quantum filtering method that is highly selective for lactate and filters out lipid signals that fall in the same chemical shift range as lactate. To enable slice selection, we added Hadamard-encoding inversion pulses in front of the SelMQC sequence with slice gradients. It achieved slice selection without conflict with lactate editing capability. We applied the sequence to a subcutaneously implanted tumor bearing mouse. Multi-slice lactate maps with 1x1x2 mm3 spatial resolution were obtained from the tumor within an hour at 9.4 T.


14:30         3183.     Multi-Volume GABA-Edited Spectroscopy of the Human Brain

Computer 27

Laura Sacolick1, Douglas L. Rothman1, Robin A. de Graaf1

1Yale University, New Haven, Connecticut, USA

The acquisition of multiple, independent volumes within a single repetition time is often desirable, but is typically limited by the inability to provide adequate magnetic field homogeneity across the volumes. Here we present multi-volume GABA-edited MRS in combination with dynamic shim updating to obtain high-quality MR spectra at 4.0 T from three volumes located in the occipital cortex, corpus callosum, and baso-frontal cortex. The three volumes were segmented into gray matter, white matter and CSF, and GABA concentrations were calculated for gray and white matter to be 1.26±0.32 and 0.54±0.15. These and other metabolite concentrations were found to be within the ranges reported in the literature.


15:00         3184.     Elevated Phosphocholine as a Marker of Carcinogenesis in the Woodchuck Model of Hepatocellular Carcinoma

Computer 27

Eilean J. McKenzie1,2, Marco L.H. Gruwel1

1National Research Council - Institute for Biodiagnostics, Winnipeg, Canada; 2University of Manitoba, Winnipeg, Canada

Woodchucks are the only animal model of chronic viral hepatitis and underwent repeated 31P-MRS to follow tumor growth in vivo. It was discovered that phosphomonoester resonances are elevated in tumors compared to controls. Ex vivo analysis by 31P-NMR confirmed elevations to phosphocholine contribute to PME elevation in vivo.


15:30         3185.     Strategies for Reliable Quantification of Intracerebral GABA by 1H-MRS

Computer 27

Wolfgang Bogner1, Stephan Gruber1, Andreas Stadlbauer2, Marc Doelken2, Siegfried Trattnig1, Arnd Doerfler2, Hermann Stefan2, Thilo Hammen2

1MR Center of Excellence, MUW, Vienna, Austria; 2Uniklinik Erlangen-Nuremberg, Erlangen, Germany

Gamma-aminobutyric acid (GABA) is an important inhibitatory neurotransmitter in human brain with anticonvulsive character. Because of an increased interest in GABA metabolism we tested the reliability of a special single-voxel spectroscopy (SVS) editing sequence which allows non-invasive intracerebral measurements of this neurotransmitter in 9 healthy adults on a 3 Tesla Scanner (Siemens TimTrio). Intra- and inter-subject reproducibility was assessed. Reliability of GABA/Cr and GABA/H2O ratios was compared in occipital lobe. Different quantification algorithms were compared (time-domain-line fitting and integration of signals).


14:00         3186.     High Resolution NMR Spectra in Inhomogeneous Fields Via Intermolecular Multiple Quantum Coherences Without Coherence Selection Gradients

Computer 28

Zhong Chen1, Congbo Cai1, Yanqin Lin1, Shuhui Cai1, Jianhui Zhong2

1Xiamen University, Xiamen, People's Republic of China; 2University of Rochester, Rochester, New York, USA

Coherence selection gradient has been thought to be essential for high resolution NMR spectra in inhomogeneous field based on intermolecular multiple quantum coherences (iMQC). However, our experimental results show that it can be omitted if correct phase cycling is applied. This means that the measured line-width of high resolution spectral peaks is not determined by the dipolar correlation distance caused by coherence selection gradient, but only affected by the effects of diffusion and T2 relaxation. This result prompts us to reconsider the iMQC high resolution mechanism.


14:30         3187.     Detection of Low Levels of Brain Galactitol in Galactosemia Using  1H 2D Double-Quantum Spectroscopy

Computer 28

Zhiyue Jerry Wang1,2, Fernando Scaglia1,2, Vernon Reid Sutton1,2, Kerri M. Lamance2, Jill V. Hunter1,2

1Baylor College of Medicine, Houston, USA; 2Texas Children's Hospital, Houston, USA

Galactosemia is an autosomal recessive hereditary disease of galactose metabolism. Patients have elevated urine galactitol even under the dietary restriction of galactose, due to endogenous production of galactose. With conventional proton MRS a low level galactitol (below 0.5 mM) cannot be resolved from other brain metabolites. We demonstrate a two-dimensional double-quantum 1H spectroscopy technique at 3T for detecting low levels of galactitol in the brain of galactosemics under dietary restrictions. This work opens the possibility for further investigations of the role of residual brain galactitol in the well-being of galactosemia patients.


15:00         3188.     An Effective Fast Acquisition Scheme to Achieve High-Resolution MRS with J-Coupling Scaling Via Intermolecular Multiple-Quantum Coherences

Computer 28

Xi Chen1, Meijin Lin1, Jincan Chen1, Tao Lin1, Zhong Chen1

1Xiamen University, Xiamen, People's Republic of China

A series of intermolecular double-quantum filtered (iDQF) sequence with efficient solvent suppression and different scaling factors of J-coupling constants, named iDQF-HOMOGENIZED II (abbreviated as iDH2), are designed to achieve fast acquisition of high-resolution spectra in inhomogeneous fields. Experiments on swine brain tissues were performed to test the feasibility of the new method. The results suggest potential applications for in vivo spectroscopy.


15:30         3189.     A Flexible IMQC Method for Accurate Determination of J-Coupling Constants in Inhomogeneous Fields

Computer 28

Yanqin Lin1, Shuhui Cai1, Yuqing Huang1, Zhong Chen1, Jianhui Zhong2

1Xiamen University, Xiamen, People's Republic of China; 2University of Rochester, Rochester, New York, USA

An improved pulse sequence was developed to scale apparent J coupling constants by a scaling factor ranging theoretically from zero (completely decoupled) to infinity under inhomogeneous fields via intermolecular multiple-quantum coherences. Scaling up the apparent J coupling constants allows more accurate measurement of small J coupling constants, and a completely decoupled homonuclear spectrum can be of considerable help for improving signal separation and thus peak assignment in MRS. The resulting spectrum retains conventional high-resolution NMR spectral information.


Hyperpolarized 13C & Other Nuclei

Hall D                                   Monday 14:00-16:00                                                                                                                                             

14:00         3190.     Reduction of the Myocardial Intracellular Matrix as Measured by Hyperpolarized 13C NMR

Computer 29

Matthew E. Merritt1,2, Crystal Harrison2, Rajesh Pidikiti1, Charles Storey1, Shawn C. Burgess1, Craig R. Malloy1, A. Dean Sherry1,2

1UTSW Medical Center, Dallas, Texas, USA; 2UT Dallas, Richardson, Texas, USA

Infusion of a combination of octanoate and pyruvate is shown to inhibit pyruvate dehydrogenase (PDH) flux as measured using hyperpolarized [1-13C]-pyruvate and 13C NMR in a perfused rat heart. Octanoate,a medium chain fatty acid that freely diffuses into the cell, provides extra reducing equivalents to the mitochondria, ultimately shifting the REDOX state of the heart and increasing the production of lactate as well as blocking the appearance of 13CO8 and H13CO3-.


14:30         3191.     Strategies to Prolong the T1 Times in Hyperpolarized 13C and 15N Biomolecules

Computer 29

Jan-Bernd Hövener1

1HMRI, Pasadena, California , USA

The advantage of hyperpolarization, with signal enhancement approaching 100.000 fold, is limited by the intrinsic lifetime T1. Biomedical applications which can be achieved within a timeframe of 5 T1 (~100s) are limited. A major hurdle is therefore to retain the high polarization over time.

Here, we describe a simple measure which significantly prolongs T1. After deuteration of 4 PASADENA molecules (13C, 15N), or by dissolution in D2O, the T1 was increased 6.5 fold max., ~two-fold in average. This widens the range for applications of long-lived, hyperpolarized biomolecules in-vivo, including oncology, vulnerable plaque, cancer biomarkers and oncogene tracking.


15:00         3192.     In Vivo Hyperpolarized 13C MRS/MRSI Using 13C-Lactate as the Pre-Polarized Substrate

Computer 29

Albert P. Chen1, John Kurhanewicz1, Robert Bok1, Duan Xu1, David Joun1, Vickie Zhang1, Sarah J. Nelson1, Ralph E. Hurd2, Daniel B. Vigneron1

1UCSF, San Francisco, California , USA; 2GE Healthcare, Menlo Park, California , USA

Prior studies have shown the ability to detect the metabolic conversion of the hyperpolarized 13C-pyruvate into 13C-lactate, 13C-alanine and 13C-bicarbonate in animal models. In this study, a hyperpolarized MR probe based on 13C1-lactate as the substrate for in vivo 13C MRS/MRSI studies was developed.   After injection of hyperpolarized 13C1-lactate, metabolic products 13C1-pyruvate and 13C1-alanine were observed in normal rats and in transgenic mouse prostate cancer models.  This study demonstrated the feasibility of using pre-polarized 13C1-lactate to study lactate metabolism as well as tumor lactate uptake in vivo.


15:30         3193.     Establishment of Hyperpolarized 13C MR in the Isolated Perfused Heart

Computer 29

Marie Allen Schroeder1, Lisa C. Heather1, Mark A. Cole1, Kieran Clarke1, George K. Radda1, Damian J. Tyler1

1University of Oxford, Oxford, UK

This metabolic information available from hyperpolarized 13C MR could provide the basis for kinetic modelling of cellular uptake and enzymatic activity; however, understanding of the physiological interactions between the hyperpolarized substrate and tissue is necessary. The isolated perfused heart provides the ideal model system to gain this understanding, as many parameters can be easily controlled. This study demonstrated the methodology we have developed to inject hyperpolarized 1-13C-pyruvate into the perfused heart without affecting heart function, and to monitor pyruvate uptake and metabolism. Further, we have directly compared our perfused heart spectra with analogous in vivo results.


14:00         3194.     The Effect of Hyperpolarized Tracer Concentration on Myocardial Uptake and Metabolism

Computer 30

Damian J. Tyler1, Marie A. Schroeder1, Lowri E. Cochlin1, Kieran Clarke1, George K. Radda1

1University of Oxford, Oxford, UK

Hyperpolarization of 13C-labelled metabolic substrates provides the MR signal necessary to visualize in vivo substrate uptake and metabolism in real time. This dynamic information could provide the basis for the kinetic modeling of cellular uptake and enzymatic activity. However, to achieve this end, an understanding of the physiological interactions between the hyperpolarized substrate and tissue is necessary. In this study, the relationship between initial pyruvate tracer concentration and the MR signal of pyruvate and its metabolic products was examined in the heart of living rats.


14:30         3195.     In Vivo Detection of Hyperpolarized 15N Choline in the Rat

Computer 30

Cristina Cudalbu1, Arnaud Comment2, Kai Uffmann, Ruud B. van Heeswijk, Chiara Perazzolo, Fiodar Kurdzesau2,3, Sami Jannin2, Vladimir Denisov4, Deniz Kirik4, Jacques J. vna der Klink2, Rolf Gruetter,5

1 Ecole Polytechnique Federale de Lausanne, Lausanne, Switzerland; 2Laboratory for Physics of Nanostructured Materials, Ecole Polytechnique Fédérale de Lausanne, Lausanne, Switzerland; 3Paul Scherrer Institute, Villigen, ,

The aim of the present study was to demonstrate the feasibility of detecting hyperpolarized 15N labeled Choline in vivo in the rat.

All the 15N MRS data were acquired on a 9.4T system. 15N choline was polarized at 3.35T and 1.2K. The long T1 combined with the potential to observe hyperpolarized 15N Cho in vivo makes this compound useful for early detection of tumors and also for a potential utilization in the assessment of blood flow. To our knowledge the in vivo detection of hyperpolarized 15N has not been demonstrated to date. We conclude that it is feasible to detect hyperpolarized 15N in live animals.


15:00         3196.     Localized Detection of Hyperpolarized [1-13C]Pyruvate and Its Metabolic Products in Rat Brain

Computer 30

Dinesh K. Deelchand1, Isabelle Iltis1, Malgorzata Marjanska1, Chris Nelson1, Kamil Ugurbil1, Pierre-Gilles Henry1

1University of Minnesota, Minneapolis, Minnesota, USA

This work examines the feasibility of measuring 13C signals from hyperpolarized 13C metabolic products in the rat brain in vivo following i.v. injection of hyperpolarized [1-13C]pyruvate. Enhanced 13C resonances from pyruvate, pyruvate hydrate, lactate and alanine were observed. Comparison of unlocalized 13C spectra with 13C-localized LASER spectra suggests that a significant fraction of [1-13C]lactate signal arises from brain tissue.


15:30         3197.     Detecting the Invisible: DNP-Enhanced Detection of 13C in Carboxyl Resonances of Rat Brain Extracts

Computer 30

Chiara Perazzolo1, Arnaud Comment1, Kai Uffmann1, Ruud B. van Heeswijk1, Fiodar Kurdzesau1,2, Sami Jannin1, Jacques J. van der Klink1, Rolf Gruetter1,3

1Ecole Polytechnique Fédérale de Lausanne, Lausanne, Switzerland; 2Paul Scherrer Institute, Villigen, Switzerland; 3Universities of Lausanne and Geneva,, Switzerland

13C spectra are largely used in MRS to investigate brain metabolism in vivo; hyperpolarization has been used to enhance the signal of a specific 13C tracer with the aim to observe metabolism. Brain extracts may yield important additional insight into the measurement of metabolic rates. The aim of the current study was to establish hyperpolarization of 13C in brain extract for the detection of low concentration 13C with intrinsically low sensitivity. To the best of our knowledge, for the first time a brain extract solution has been successfully hyperpolarized.


14:00         3198.     Hyperpolarized 13C Acetate Detection in the Rat Brain in Vivo

Computer 31

Kai Uffmann1, Arnaud Comment1, Ruud B. van Heeswijk1, Chiara Perazolla1, Sami Jannin1, Fiodar Kurdzesau1,2, J. A. Konter2, Patrick Hautle2, Ben van der Brandt2, Jacques J. van der Klink1, Rolf Gruetter1,3

1Ecole Polytechnique Fédéral de Lausanne (EPFL), Lausanne, Switzerland; 2Paul Scherrer Institute, Villigen, Switzerland; 3Universities of Lausanne and Geneva, Switzerland

Hyperpolarized 1-13C-acetate was infused into three rats, followed by collection of 20 spectra (TR=3s). The decay of the enhanced signal of the tracer could be sampled. In an additional experiment performing spectroscopic imaging the localization of the acetate signal was proven to be in the brain. A T1 relaxation time of 20.7s of the labeled compound was derived by fitting the decay using a biexponential curve. Furthermore a resonance peak at 174.1ppm was observed, which followed the similar decay as the acetate signal. The origin of this peak remains to be investigated.


14:30         3199.     Can Hyperpolarized 89Y Be Used as a Molecular Imaging Agent?

Computer 31

Matthew E. Merritt1,2, Crystal Harrison1,2, Zoltan Kovacs1, Craig R. Malloy1, A Dean Sherry1,2

1UTSW Medical Center, Dallas, Texas, USA; 2UT Dallas, Richardson, Texas, USA

Yttrium-89 is an NMR active nucleus with exceptionally long T1's but low sensitivity compared to protons. It is shown that dynamic nuclear polarization can be used to enhance 89Y sensitivity 250 to 1500-fold. Yttrium has similar coordination chemistry to gadolinium, and could potentially be used as a molecular imaging agent if bound to the appropriate ligand.


15:00         3200.     Direct Optical Hyperpolarization of Liquids

Computer 31

Daniel Robert Elgort1, Remus Albu1

1Philips Research North America, Briarcliff Manor, New York, USA

This novel hyperpolarization method uses light with orbital angular momentum (OAM) to enable the direct hyperpolarization of liquids. The feasibility of this technique was verified experimentally by performing NMR spectroscopy on liquid sample (index matching oil, refractive index of 1.516) at 0.18T.   A proton NMR signal was generated with an SNR of 5.3 (sample size = 30µL, polarization time = 70ms, FID measurement time = 8ms, 200ksamples, 20 averages).  The resulting NMR spectra corresponded well with a standard MR spectroscopy measurement performed on a commercial 1.0T Philips Panorama scanner (PRESS sequence, FOV=1cm3, FA=90&[deg], TE/TR=50/2500ms, NSA=24, BW=8000Hz, NP=16384).


15:30         3201.     Methyl Group Tunnelling and Dynamic Nuclear Polarization

Computer 31

Martyn Paley1

1University of Sheffield, Sheffield, UK

A tunnel magnetic resonance switched field experiment on zinc acetate shows generation of dynamic nuclear polarization without the use of microwave irradiation based on tunnelling methyl groups interacting with a free electron.