Vascular Compliance

                   906.       In Vivo Transit Time MR-Measurements of Pulse Wave Velocity in the Murine Aorta at 17.6 Tesla

Marco Parczyk¹, Volker Herold¹, Gert Klug¹, Thomas Schulze-Till¹, Wolfgang Bauer¹, Eberhard Rommel¹, Peter Jakob¹

¹University of Wuerzburg, Wuerzburg, Germany

Aortic stiffness increases in an early state of arteriosclerosis, assessable by pulse wave velocity (PWV) MR-measurements. Up to now only studies in larger animals and humans have been reported in literature. The feasibility of in vivo PWV MR-measurements by two non-invasive MR-imaging techniques is presented. Because pulse wave and flow velocities are similar to velocities in humans, but dimensions are about 20-times smaller, the challenges in this project were especially the high temporal and spatial resolutions needed.

                  907.       In Vivo Measurement of Local Pulse-Wave Velocity in Mice with MRI at 17.6 T

Volker Herold¹, Gert Klug¹, Marco Parczyk¹, Christian Ziener¹, Thomas Weber¹, Susanta Sarkar², Wolfgang Rudof Bauer¹, Eberhard Rommel¹, Peter Michael Jakob¹

¹University of Wuerzburg, Wuerzburg, Germany; ²GlaxoSmithKline (GSK), King of Prussia, Pennsylvania, USA

Pulse wave velocity (PWV) is an important parameter for the evaluation of the arterial stiffness and cardiovascular risk. Several diseases such as hypertension and arteriosclerosis are associated with vascular remodeling and arterial stiffening. Mouse models of human diseases are increasingly used to investigate patho-physiological mechanisms of the cardiovascular system. A non invasive method is presented to assess local PWV in the ascending and descending aorta of mice with MR-Microscopy at 17.6 T. The results demonstrated the feasibility of high field MR microscopy to quantify local pulse wave velocity as a measure of local aortic stiffness.

                  908.       Carotid Wall Shear Rate Measured with Spiral Fourier Velocity Encoding

Joao Luiz Azevedo de Carvalho¹, Jon Fredrik Nielsen¹, Krishna S. Nayak¹

¹University of Southern California, Los Angeles, California , USA

Fourier velocity encoding (FVE) has been proposed as a method for non-invasively measuring fluid shear rate and hence vascular wall shear stress, an important factor implicated in atherogenesis. The scan-time of 2DFT FVE is prohibitively long for clinical use, but the recently introduced spiral FVE method shows promise as it is substantially faster. In this work, we investigate the feasibility of using spiral FVE for estimating shear rates near the carotid artery walls in clinically practical scan times. We present: (1) a phantom validation of spiral FVE against high-resolution 2DFT phase contrast; (2) evaluation of resolution requirements; (3) in vivo demonstration.

                  909.       Intracranial Compliance and Pressure Measurement Based on MR Flow Quantification and Brain
                                Circulation Model Circuit: Sensitivity to Hyperventilation and Hydrocephalus

Kagayaki Kuroda¹, Kosuke Maruhashi¹, Moyuru Ohya¹, Masatoshi Honda², Hideki Atsumi³, Koichi Oshio⁴, Mitsunori Matsumae³

¹Tokai University, Hiratsuka, Japan; ²Tokai University Hospital, Isehara, Japan; ³Tokai University, Isehara, Japan; ⁴Keio University, Shinanomachi, Japan

A noninvasive technique to evaluate brain compliance index (BCI) and intracranical pressure index (ICPI) based on MR flow quantification and an inverse analysis of brain-circulation-equivalent electrical circuits was developed. The technique was applied to healthy volunteers (N = 6) with normal breathing and hyperventilation during scanning. The BCI significantly decrease with hyperventilation in 5 volunteers. The technique was also applied to patients (N = 3) with hydrocephalus. Both BCI and ICPI were significantly lower in those patients than in the normal volunteers. These results indicated that the technique is sensitive to the change or abnormality of the brain physical properties.

                  910.       Compliance and Anatomy of the Neo-Aorta in Children with Hypoplastic Left Heart Syndrome (HLHS)

Michael Helle¹, ², Inga Voges¹, Michael Jerosch-Herold², Christopher Hart¹, Traudel Hansen¹, Hans-Heiner Kramer¹, Carsten Rickers¹

¹Christian-Albrechts-Universität, Kiel, Germany; ²Brigham & Women's Hospital, Boston, USA

The purpose of this study was to assess aortic anatomy and aortic compliance (AC) in children with hypoplastic left heart syndrome (HLHS) using cardiac MRI at 3 Tesla.Cine MRI and contrast enhanced time-resolved MR-angiography were performed in all patients for measuring selected diameters of the neo-aorta and for a selected determination of the AC.Diameters of the aortic root, the ascending aorta and the aortic arch were significantly increased. We found a significant smaller aortic isthmus in children with HLHS, whereas the dimensions of the descending aorta were not significantly changed. AC was decreased in the ascending aorta as well as the aortic arch.

                   911.       Quantitative 2D and 3D Phase Contrast MRI: Optimized Analysis of Blood Flow and Vessel Wall
                                  Parameters

Aurélien F. Stalder¹, Max F. Russe¹, Alex Frydrychowicz¹, Jelena Bock¹, Jürgen Hennig¹, Michael Markl¹

¹University Hospital Freiburg, Freiburg, Germany

The purpose of this work was the development of an optimized quantitative analysis method in order to derive flow and wall parameters from CINE phase contrast (PC) MR data. The data analysis strategy combines “Green’s theorem” and B-spline interpolation with their finite difference property to provide an optimal quantification of several blood flow and vessel wall parameters. Calculation of the local blood flow velocity derivatives onto the vessel contour using B-spline interpolation allowed a direct and reliable estimate of time-resolved segmental WSS vectors independent of any restrictive global assumptions regarding the flow profile.

Vessel Wall Imaging

                 946.       The Effect of Lipid Core Position on Carotid Fibrous Cap Stress Levels

Samuel Alberg Kock¹, Jens Vinge Nygaard, Anders K. Niemann¹, Anette Klærke, William Paaske, Erling Falk, Won Yong Kim¹

¹Aarhus University Hospital Skejby, Aarhus N, Denmark

To determine if not only the size of atherosclerotic lipid cores, but also their position influence carotid fibrous cap stress levels, three computational models of a carotid bifurcation with varying lipid core placement were created using geometry based on MRI scans of a patient with a high-grade stenosis. Computational fluid structure interaction simulations were performed on the models revealing proximally based lipid cores to exhibit far greater stress levels than both distally and centrally located lipid cores. Computational analyses may yield valuable additional information concerning fibrous cap stress levels which may support current methods of diagnostics.

                   947.       Ultra-Short TE Imaging Protocol for Detection of Aortic Calcification

Daniel A. Herzka¹, Reza Nezafat², Juergen Rahmer³, Warren J. Manning², Peter Boernert³

¹Philips Research North America, Briarcliff Manor, New York, USA; ²Beth Israel Deaconess Harvard Medical School, Boston, Massachusetts, USA; ³Philips Research Europe, Hamburg, Germany

Ultra-short TE sequences are advantageous because they can capture signal from tissues with short T2, while preventing signal loss from sources of artifact such as off-resonance and susceptibility boundaries. We optimized a UTE protocol for visualization of calcified atherosclerotic lesions in vivo in human aortas. The protocol was tested on volunteers and a patient with known atherosclerosis. UTE yielded images with calcium deposits clearly co-localized with atherosclerotic lesions observed with black-blood FSE, and made possible detection of calcification on the luminal side of the aorta. These preliminary results indicate that calcium imaging with MR is feasible and worth further investigation.

                   948.       Signal Features of the Atherosclerotic Plaque at 3.0T Versus 1.5T: Impact on Automatic Classification

William Sean Kerwin¹, Fei Liu¹, Hunter Underhill¹, Vasily Yarnykh¹, Chun Yuan¹

¹University of Washington, Seattle, Washington, USA

We evaluated signal differences between 1.5T and 3T MRI of carotid atherosclerotic plaque using an automatic classifier. Despite being trained on only 1.5T data, subjects scanned at both field strengths yielded highly similar classification results for the presence and average areas of calcification, lipid-rich core, hemorrhage, and fibrous tissue. These results suggest that signal properties are sufficiently similar at 3T relative to 1.5T that classifier performance is not significantly affected.

                  949.       Long Segment Dark Blood Carotid Artery Imaging with Pencil-Beam-Excitation and Diffusion
                                Preparation at 3T
                                 Axel Bornstedt¹, Peter Bernhardt¹, Vinzenz Hombach¹, Markus Kunze¹, Jochen Spiess¹, Nico Merkle¹, Volker Rasche^{1
                                1}Uniklinik Ulm, Ulm, Germany

Robust long segment (~150 mm) 3D dark blood vessel visualisation of the carotid arteries at 3T is accomplished with pencil beam excitation and diffusion prepared gradient echo imaging.

                   950.       Magnetic Resonance Imaging Parameters of Atherosclerotic Plaque Burden Successfully Predict
                                 Manifested Cardiovascular Disease

Venkatesh Mani¹, Hamza El Aidi¹, Mark Woodward¹, Paul Muntner¹, Silvia H. Aguiar¹, Karen Beth Weinshelbaum¹, Hiroaki Taniguchi¹, John E. Postley², Valentin Fuster¹, Zahi A. Fayad¹

¹Mount Sinai School of Medicine, New York, New York, USA; ²Columbia University, New York, New York, USA

The purpose of this study was to evaluate if MRI plaque burden measures in conjunction with traditional risk factors improve predictive capacity for cardiovascular disease (mCVD) and may therefore be useful in pre identification of patients at risk for cardiovascular disease. MR measures of plaque burden were obtained from 296 patients and ROC curves were used to determine predictive capacity of MR derived parameters for predicting mCVD. Combining MR parameters with traditional risk factors provided highest retrospective predictive capacity for mCVD.

                   951.       Longitudinal Observation of Carotid Artery Intraplaque Hemorrhagic Volume on Magnetic Resonance
                                 Direct Thrombus Imaging

James Qiupeng Zhan¹, Alan R. Moody², General Leung², Radhakrishnan Ravikumar², Susan Crisp²

¹Sunnybrook Health Sciences Centre, Toronto , Canada; ²Sunnybrook Health Sciences Centre, Toronto, Canada

A study was conducted to measure carotid artery intraplaque haemorrhage (IPH) volume over time using Magnetic Resonance Direct Thrombus Imaging (MRDTI) and correlate it with the occurrence of prior ipsilateral ischemic events.  Patients who received multiple MRDTI scans and were diagnosed with complicated plaque were classified as either symptomatic (having previous events) or asymptomatic. Two trained observers analyzed the MRDTI data and determined IPH volumes. The intra-and inter-reader coefficients of variation were 6.2% and 8.2%, with correlation coefficients of 0.987 and 0.972. Relative mean change in IPH volume was 86.31±266.44% for asymptomatic patients and 10.23±54.87 % for symptomatic patients (P<0.01).

                   952.       Improvements in Spatial Resolution Using a Novel 8-Element Carotid Phased Array Coil at 3T

Niranjan Balu¹, Vasily Yarnykh¹, Cecil Hayes¹, Joshua Scholnick¹, Dongxiang Xu¹, Baocheng Chu¹, Chun Yuan¹

¹University of Washington, Seattle, USA

Carotid plaque imaging demands high SNR and high resolution. While current four-element phased array (PA) coils provide adequate SNR over a limited FOV, PA coils with additional elements can improve SNR and coverage. Using a  novel eight-element PA (8PA) coil , upto 60% improvement in SNR and CNR were observed with a high-resolution (0.63mm in-plane) protocol.  An ultra-high resolution (0.27mm in-plane) T1w black-blood quadruple inversion recovery sequence was implemented to assess the resolution improvement made possible by the 8PA coil.  The 8PA coil enabled ultra-high resolution imaging with SNR improvements greater than 1.35 times with better vessel wall delineation.

                  953.       Time Dependence of Necrotic Core and Fibrous Cap Quantitative Measurements with Gadobenate
                                 Dimeglumine Enhanced Carotid Plaque MRI at 3T

Kevin DeMarco¹, Xiaohai Ma¹, John Brooks¹, David Zhu¹, Vasily Yarnykh²

¹Michigan State University, East Lansing, Michigan, USA; ²University of Washington, Seattle, Washington, USA

Contrast-enhanced T1 weighted images (CE T1WI) demonstrate lipid rich necrotic core (LR-NC) with higher contrast-to-noise ratio (CNR) at 1.5T using gadodiamide.  This study extends in vivo carotid plaque imaging to 3T and evaluates the effect of post-injection time on appearance of LR-NC using gadobenate dimeglumine by comparing CNR of LR-NC at 5 and 10 minutes after contrast injection in 43 outpatients with carotid stenosis.  LR-NC CNR was similar at both time points, but with higher signal-to-noise ratio at 10 minutes.  LR-NC volume measurement at both time points employing automated algorithm in MRI-PlaqueView is under investigation as is comparison with histology.

                   954.       Magnetic Resonance Intraplaque Hemorrhage is Associated with Cerebrovascular Outcomes in
                                 Asymptomatic Male Patients with Non-Severe Stenosis

Navneet Singh¹, Alan R. Moody¹, General Leung¹, Ravikumar Radhakrishnan¹, James Zhan¹, Robert Magissano¹

¹Sunnybrook Health Sciences Centre, University of Toronto, Toronto, Canada

Intraplaque hemorrhage (IPH) is an emerging marker of plaque instability. We investigated the correlation of MR detected IPH to cerebrovascular outcomes in a group of asymptomatic males with non-severe stenosis. A GE 1.5T MR and an 8 channel neurovascular coil array (USA Instruments, USA) using a 3D T1-weighted, fat-suppressed spoiled gradient echo sequence was used. The group with IPH (n=30) had six events (2 strokes, 1 amF, 3 TIAs) compared to no events in the MR IPH negative group (n=47) (RR = 9.60, 95%CI 1.21 to 75.9, p=0.0320).

                   955.       Fibrous Cap Projection Length: A Better Biomarker of Plaque Vulnerability Than Lipid Core Size

Dongxiang Xu¹, Niranjan Balu¹, Hunter R. Underhill¹, Jianming Cai², Chun Yuan¹

¹University of Washington, Seattle, Washington, USA; ²Chinese PLA General Hospital, Beijing, USA

Atherosclerotic disease has become one of the leading causes of death and major disability in the United States. In the past years, with the rapid development of using high resolution magnetic resonance imaging (MRI) technology in assessing atherosclerotic components, more and more evidences have shown that plaque composition is the decisive factor determining plaque vulnerability. Cai et al[1] using gadolinium-based contrast enhanced MRI showed that post contrast T1-weighted images can provide accurate quantitative measurements of the intact fibrous cap (FC) in advanced carotid atherosclerotic plaques in vivo. Based on this observation, an automatic FC detection method was developed and validated in our previous research [2]. In this study, we further explore lesion index Normalized Fibrous cap Projection Length Index (NFPLI). Our preliminary trial result has shown its more predictive power in plaque vulnerability than other plaque measurements.

                   956.       Rapid 3D Vessel Wall Imaging at 3T:  Optimization of Diffusion Preparation and Comparison to Other
                                  Protocols

Mahender K. Makhijani¹, Gerald S. Pohost², Krishna S. Nayak¹

¹University of Southern California, Los Angeles, California , USA; ²University of Southern California, USA

Multi-contrast high-resolution imaging is used to characterize carotid plaque components. Standard multi-slice methods have long scan times and are not suited for plaque quantitation. Diffusion preparation has been recently proposed as a means for blood suppression in 3D vessel wall imaging. We characterize the tradeoffs of this approach when applied in conjunction with DIR and evaluate its performance at 3T in-vivo for bilateral carotid imaging and compare it with standard protocols and 3D SSFP approach. Cardiac gated 3D carotid vessel-wall datasets with 0.5x0.5x2.5 mm3 resolution over a 16x3.2x5cm3 FOV, and vessel wall CNR > 18, were obtained in 100 seconds

                   957.       Detection of Plaque Vascularity Using Delayed Contrast Enhanced MRI: Correlation with Contrast
                                 Enhanced Ultrasound

Radhakrishnan Ravikumar¹, Alan R. Moody¹, General Leung¹, James Q. Zhang¹, Peter N. Burns¹, Susan Crisp¹, Marilyn Robertson Horton¹, Robert Maggisano¹

¹Sunnybrook Health Sciences Centre, Toronto, Canada

Plaque neovascularity, in addition to nurturing plaque growth can also be a causative factor for intraplaque hemorrhage thus predisposing the patient to increased risk of cerebrovascular events. Contrast enhanced ultrasound has been effectively used in identifying plaque neovascularity. The main purpose of this study was to see whether contrast enhanced MRI can help identify neovessels within the plaque and to compare the MR findings with contrast ultrasound findings. Delayed enhancement uptake patterns visualised on MR correlated well with neovascularity positive areas detected on contrast ultrasound.

                958.       High-Resolution T1- And T2-Weighted Black Blood Inner Volume 3D Fast Spin Echo Imaging for
                             Characterizing Vessel Wall Components in Vivo

Dimitris Mitsouras¹, Robert V. Mulkern¹, ², Christopher D. Owens¹, Tianxi Cai³, Amanda G. Whitmore¹, Hale Ersoy¹, Michael S. Conte¹, Mark A. Creager¹, Frank J. Rybicki¹

¹Harvard Medical School & Brigham And Women's Hospital, Boston, Massachusetts, USA; ²Children's Hospital, Boston, Massachusetts, USA; ³Harvard School of Public Health, Boston, Massachusetts, USA

Saphenous vein femoral-popliteal bypass graft imaging using a high spatial resolution (0.312mm in-plane) black-blood inner-volume 3DFSE sequence in 14 subjects revealed significantly larger wall area measured from T1W than T2W images (median ratio 1.52, median difference 5.45mm², P<0.001). This significant difference was due to an increased outer wall boundary. T₂ relaxometry of two specimens revealed shorter T₂ values in an outer vs. an inner wall layer, accounting for the in vivo finding of differing areas measured from different contrast weightings. Correlative histology attributed the shorter T₂ in the outer layer to collagen-rich fibrous tissue vs. myofibroblasts embedded in proteoglycan-rich matrix.

                   959.       Gender Differences in Aortic Wall Thickness Regression by Magnetic Resonance Imaging (MRI) and Its
                                  Association to HDL Profile: The Plaque Follow Up Study by the National Institute of  Aging (NIA)

Gustavo Khattar Godoy¹, Veronica R.S Fernandes¹, Hossein Bahrami¹, Christopher Sibley¹, Ilan Gottlieb¹, David A. Bluemke¹, Joao A.C Lima¹

¹Johns Hopkins University, Baltimore, USA

The aim of the study was to assess the relationship between changes in aortic wall thickness measured by MRI with lipid profile and gender,of 114 participants of a randomized clinical trial using lipid-lowering drugs for 2 years.Were analized 3 segments of the thoracic aorta (ascending,arch and descending),using a double inversion recovery black blood fast spin-eco sequence with ECG-gating and T1-Weighted post-gadolinium.Were found a significant lower rate reduction of the aortic thickness in men when compared to women.Greater HDL levels were related to a decrease in ascending aortic wall thickness ,after adjustments for variables.

                   960.       Quantitative Comparison of Carotid Plaque Composition Between 1.5 and 3.0T Field-Strengths

Hunter R. Underhill¹, Vasily L. Yarnykh¹, Thomas S. Hatsukami¹, Jinnan Wang¹, Niranjan Balu¹, Cecil Hayes¹, Minako Oikawa¹, Wei Yu¹, Dongxiang Xu¹, Baocheng Chu¹, Bradley T. Wyman², Nayak L. Polissar³, Chun Yuan¹

¹University of Washington, Seattle, USA; ²Pfizer, Groton, USA; ³The Mountain-Whisper-Light Statistical Consulting, Seattle, USA

We sought to assess the effects of field-strength on the quantification of carotid atherosclerotic disease.  Participants with 16-79% carotid stenosis underwent high-resolution carotid MRI at both 1.5T and 3.0T.  There was strong agreement between field-strengths in quantitative measurements of plaque morphology and detection of plaque components.  However, the increased magnetic susceptibility of calcification and a stronger effect of paramagnetic ferric iron in hemorrhage at 3.0T may introduce a quantitative bias in measurements of these components.  As such, 3.0T imaging may improve the detection of calcification, but more sensitive imaging techniques may need to be used for hemorrhage evaluation at 3.0T.

                   961.       Improved Motion-Sensitized Driven Equilibrium (IMSDE) Blood-Suppression Sequence for
                                 Atherosclerosis  Plaque Imaging at 3T

Jinnan Wang¹, ², Vasily L. Yarnykh¹, Baocheng Chu¹, Chun Yuan¹

¹University of Washington, Seattle, Washington, USA

In this study, an improved version of motion sensitization driven equilibrium (iMSDE) sequence was proposed for carotid artery vessel wall imaging. By adding a new refocusing pulse, the iMSDE sequence is less sensitive to the B1 inhomogeneity at high field strength. Both phantom test and in vivo imaging has demonstrated that iMSDE sequence can  dramatically increase signal level, when comparing with traditional MSDE sequence, without losing flow suppression efficiency.

                   962.       High-Resolution Ultra-Short TE Imaging of Ex Vivo Human Carotid Plaques Correlates with CT

Daniel A. Herzka¹, Juergen Rahmer², Reza Nezafat³, Ray Chan¹, Wei Liu¹, Peter Boernert²

¹Philips Research North America, Briarcliff Manor, New York, USA; ²Philips Research Europe, Hamburg, Germany; ³Beth Israel Deaconess Harvard Medical School, Boston, Massachusetts, USA

Ultra-short TE (UTE) sequences have the advantage that they are able to capture signal components with very short T2. We investigate the use of single and multi-echo UTE sequences for the detection of calcium deposits in ex vivo human carotid samples with known atherosclerosis. UTE images correlated significantly with high-resolution CTs but with better soft-tissue contrast. Though low in intensity, the signal from calcifications is detectable with UTE MRI and may be complementary to current multicontrast atherosclerotic tissue characterization approaches. Further work is required to assess the feasibility of these acquisitions in vivo.

                   963.       Passive Targeting of Atherosclerosis with Paramagnetic Lipid Nanoparticles in a Mouse Model of
                                 Vulnerable and Stable Plaques

Glenda Sibylle van Bochove¹, Leonie E M Paulis¹, Dolf Segers², Willem J M Mulder³, Rob Krams⁴, Gustav J. Strijkers¹, Klaas Nicolay¹

¹Eindhoven University of Technology, Eindhoven, Netherlands; ²Erasmus Medical Center, Rotterdam, Netherlands; ³Mount Sinai School of Medicine, New York, New York, USA; ⁴Imperial College London, London, UK

Recently, a mouse model has become available where both stable and vulnerable atherosclerotic plaque phenotypes are induced by the placement of a tapered cast around the right carotid artery of apoE-/- mice. The aim was to evaluate the contrast generated in these plaques by injection of Gd3+ liposomes and Gd3+-micelles to obtain insight in plaque permeability. Both stable and vulnerable plaques appeared non-permeable for liposomes, while accumulation of micelles was observed in both lesion types. Therefore, liposomes are a good candidate for targeting endothelial markers and micelles may also be suitable for targeting factors inside the atherosclerotic plaque.

                   964.       Molecular Imaging of Atherosclerosis Using PEG-Micelles Targeted by an ApoE Derived Peptide

Eik Leupold¹, Esad Vucic², Willem Mulder², Margitta Dathe¹, Zahi Adel Fayad²

¹Leibniz Institute of Molecular Pharmacology (FMP), Berlin, Germany; ²Mount Sinai School of Medicine, New York, USA

We use Gadolinium labeled PEG-micelles as contrast agents for MRI with ApoE knockout mice as atherosclerosis models. Targeting is achieved by an ApoE derived peptide (A2). The PEG-micelles consist of maleimide-PEG-DPPE, Gd-DTPA-BSA and rhodamine-DPPE. The micelles are produced by lipid-film hydration and A2 is coupled via maleimide/sulfhydryle conjugation.The micelles are 18nm in diameter and have a high stability. In vivo application (50mmol Gd/kg) results in a high, long lasting enhancement of atherosclerotic plaque in MRI using a 9.4T system. The maximum enhancement of 129±46% is reached at 24h, compared to 30±15% in mice treated with control PEG-micelles. Confocal microscopy was performed on sections taken from the thoracic aorta. A co-localization of A2 PEG micelles with macrophages in atherosclerotic plaques could be shown.A2 modified PEG-micelles containing Gd are a new contrast agent that is excellently suited for the MRI based diagnosis of atherosclerosis.

                  965.       Comparison of Synthetic HDL Contrast Agents for Atherosclerosis Imaging

David Peter Cormode¹, Rohith Chandrasekar², Karen C. Briley-Saebo¹, Alessandra Barazza¹, Willem J. Mulder¹, Edward A. Fisher³, Zahi Adel Fayad¹

¹Mount Sinai School of Medicine, New York, USA; ²The Cooper Union for the Advancement in Science and Art, New York, USA; ³New York University, New York, USA

We have previously reported a macrophage specific MRI contrast agent based on HDL. To make this agent more versatile, we have formed Gd-labeled, synthetic HDL using a 37 or 18 amino acid peptide that mimics apoA-I, the main protein constituent of HDL. The effectiveness of these peptide-agents for detecting macrophages in the abdominal aorta of apoE knockout mice is assessed via MRI studies. Targeting to macrophages is confirmed by confocal microscopy. In addition, the biodistribution and pharmacokinetics are reported and the therapeutic properties of these particles in terms of removing cholesterol from macrophages are investigated.

                   966.       Imaging Vasulcar Injury Using a Novel Gadolinium-Based Contrast Agent

Pauliina Lehtolainen¹, Manfred Junemann-ramirez, Panagiotis Kyrtatos¹, Anthony N. Price¹, Kenjiro Ikuta², Yoshiki Katayama², John F. Martin, Mark F. Lythgoe

¹ Institute of Child Health and Department of Medicine, University College London, London, UK; ²Kyushu University,, Fukuoka, Japan

The vasculature is one of the most promising targets for a site -specific MRI strategy, as the diagnosis of vascular disease in its early stages is essential to a successful treatment intervention. Here we report on an endothelial-lesion specific contrast agent, evans-blue chelated gadolinium (EB-DTPA-Gd) for imaging the vascular damage and regeneration of vascular wall. Vascular injury generated by angioplasty was clearly distinguished by EB-DTPA-Gd accumulation by T1-weighted MR images using 9.4T scanner.

                   967.       Comparison of Gadofluorine M and Gd-DTPA Relaxivities for Quantitation and Characterization of
                                 Atherosclerotic Plaque in Mouse at 11.7T

Haiying Tang¹, Richard Kennan¹, Ching H. Chang¹, Bernd Misselwitz², Donna Suresch¹, Dan Zhou¹, Brett Connally¹, Michael Klimas¹, Donald S. Williams¹, Richard Hargreaves¹, Haiying Liu¹

¹Merck Research Laboratories, Rahway, New Jersey, USA; ²Bayer Schering Pharma AG, Berlin, Germany

Recent advances in MRI technology and novel imaging contrast agents have made MRI an important imaging modality for detecting and characterizing atherosclerosis. Gadofluorine M has been reported to target the extracellular matrix of plaque, and is an important marker of plaque staging. Comparisons of relaxivities and dose response in plaque are made between Gd-DTPA and Gadofluorine M at 11.7 Tesla. The results confirm that Gadofluorine M greatly helps in the identification and quantitation of plaque burden in major arteries of mouse model of atherosclerosis with improved sensitivity and efficiency, and may provide better characterization of plaque components at different stages.

                   968.       Imaging of Macrophage Infiltration in an Abdominal Aortic Aneurysm Mouse Model

Gregory Harrison Turner¹, Alan R. Olzinski¹, Roberta E. Bernard¹, Karpagam Aravindhan¹, Heather W. Karr¹, Robert N. Willette¹, Peter J. Gough¹, Beat M. Jucker¹

¹GlaxoSmithKline, King of Prussia, Pennsylvania, USA

Abdominal aortic aneurysms result from a vascular inflammatory process involving macrophage recruitment.  Abdominal aortic aneurysms (AAA) can be consistently produced in hyperlipidimic ApoE-/- mice by continuous infusion of Angiotensin-II (Ang-II).  Administration of USPIO was used as an imaging biomarker for the distribution of macrophage within an aneurysm. These results demonstrate the feasibility of using a USPIO contrast agent as a surrogate for acute inflammatory processes in the aorta of ApoE-/- mice.

                   969.       GdAAZTA-C17 (Q=2) Labeled High Density Lipoproteins (HDL) for the in Vivo Detection of
                                  Atherosclerotic Plaque

Karen Briley-Saebo¹, Simonetta Geninatti², David Cormode¹, Alessandra Barazza¹, Wei Chen¹, Edward Fisher³, Silvio Aime², Zahi A. Fayad¹

¹Mount Sinai School of medicine, New York, USA; ²University of Torino, Torino, Italy; ³New York University, New York, USA

To increase CNR and reduce the dose administered, high relaxivity gadolinium chelates with two water exchange sites (q=2) were integrated into high-density lipoproteins (HDL). The q=2 HDL adduct was characterized relative to GdDTPA-DMPE (q=1) HDL, and the MR efficacy evaluated in mouse models of atherosclerosis. The results indicate that the q=2 lipid integrates into the lipid core. The q=1 lipid, however, formed a micelle that interacted with the surface. Similar MR arterial enhancement was observed after administration of 0.048 mmol Gd/Kg q=1 HDL and 0.018 mmol/Kg q=2 HDL. The q=2 HDL adduct may allow for low dose detection of atherosclerosis.

Stephanie Elaine GarWai Chiu¹, Alan R. Moody¹, James Qiupeng Zhan¹, Radhakrishnan Ravikumar¹, General Leung¹

¹Sunnybrook Health Sciences Centre, Toronto, Canada

Intraplaque hemorrhage and plaque neovascularization are recognized as contributors to atherosclerotic plaque vulnerability, but current animal models do not consistently or spontaneously produce these types of lesions.  As a first step towards building upon the commonly used hypercholesterolemic rabbit model, a low dose of recombinant human vascular endothelial growth factor (rhVEGF) was administered to a group of rabbits.  These injections have been shown to increase both intramural endothelial cell and macrophage density in hypercholesterolemic rabbits. MRI using both an intravascular and an extravascular contrast agent was performed on the rabbits in an attempt to detect these changes non-invasively.

Cardiac T2 & T2*

Magalie Viallon¹, Jean-Noël Hyacinthe¹, Dominique Didier¹, Pierre Croisille²

¹Hopital Universitaire de Genève, GENEVA, Switzerland; ²Hopital Cardiologique Louis Pradel, LYON, France

T1 or T2 weighted TSE and T1 SE dark blood (DB) MR imaging (eventually combined with fat saturation (FS) or inversion recovery(IR)) are involved to characterize pathologies like DAVD, tumours, sarcoïdosis, myocarditis,  to determine the area-at-risk in acute myocardium infarction (MI) or to detail cardiac morphology in congenital disease. But Dark-blood TSE methods are subject to artifacts (signal loss due to incoherent cardiac motion from RR length changes), resolution is restricted by the breathhold duration and remain low. In free breathing paediatric patients TSE DB techniques is dramatically hindered by the respiratory motion. PROPELLER then BLADE were new encoding strategies implemented to correct for intra and inter scan motion in brain morphological acquisition and offer a solution in uncooperative patients (Parkinson, stroke or Alzheimer diseases). We investigate here the capabilities of BLADE to circumvent motion sensitivity in DB cardiac morphology imaging and improve image quality.

                  1009.     Detecting Right Ventricular Involvement in Reperfused Myocardial Infarction with T2-Weighted
                                Cardiac MR and IR-Prepared SSFP with Delayed Contrast Enhancement

Yuesong Yang¹, Ram Vijayaraghavan¹, Jay Detsky¹, John J. Graham¹, Warren Foltz¹, Alexander Dick¹, Graham A. Wright¹

¹Sunnybrook Health Sciences Centre, University of Toronto, Toronto, Canada

Right ventricular (RV) myocardial infarction and dysfunction are independent indicators of poor prognosis in patients with acute myocardial infarction (AMI). Conventional DE-MRI has a limited role in the detection of the RV involvement in reperfused MI due to the thinned RV wall and pericardial fat tissue. We hypothesize that a T2-weighted cardiac MR technique with fat saturation is a better method to identify RV involvement in the acute stage of MI, and that the DE-MRI technique based on IR-prepared SSFP is a better technique to demonstrate RV involvement in the chronic stage of MI.

                  1010.     Optimization of T2 and T2* Measurement in Myocardium at 3.0 T

Jared Guthrie Cobb¹, ², Huairen Zeng², ³, Cynthia Paschal²

¹Vanderbilt University , Nashville, Tennessee, USA; ²Vanderbilt University, Nashville, Tennessee, USA; ³Vanderbilt University Medical Center, Nashville, Tennessee, USA

High field clinical scanners are increasingly available to researchers and clinicians and possess a significant SNR advantage over 1.5T scanners.  New accelerated imaging techniques designed to optimize spatial and temporal acquisition may allow researchers to detect subtle changes in myocardial tissue, including changes that may be indicators of differences in oxygen utilization such as variations in T2 and T2*. In pursuit of this goal, we developed optimized breath hold scans to measure T2 and T2* in the myocardial septum at 3.0T.

                  1011.     Single-Shot SSFP for Imaging of Edematous Myocardium in Patients

Jordin D. Green¹, ², James Reavley Clarke², Matthias G. Friedrich²

¹Siemens Medical Solutions, Calgary, Canada; ²University of Calgary, Calgary, Canada

Cardiovascular MR can be used to visualize myocardial edema, a characteristic of patients with acute myocardial infarcts, but can be challenging in difficult patients. SSFP is T₂/T₁-weighted in the steady state with high imaging efficiency. The purpose of this study was to demonstrate the feasibility of using single-shot SSFP to image myocardial edema. The study was conducted in six patients. Results were compared to a conventional T₂-weighted STIR sequence. Of the 32 myocardial segments positive for edema according to the STIR sequence, 24 were positive using SSFP. Of the 52 negative for edema according to STIR, 45 were negative using SSFP.

                  1012.     Enhanced Detection of Myocardial Edema with Spectrally Selective Inversion Recovery-Prepared
                                 T2-Weighted Imaging

Myra Sabene Cocker¹, Oliver Strohm¹, Jordin Daniel Green, ¹², Steven M. Shea³, Hassan Abdel-Aty¹, Matthias G. Friedrich¹

¹Libin Cardiovascular Institute at the University of Calgary, Calgary, Canada; ²Siemens Medical Solutions, Calgary, Canada; ³Siemens Corporate Research Inc., Baltimore, Maryland, USA

AASPIR (Asymmetric Adiabatic Spectral Inversion Recovery) was compared with STIR (Short T1 Inversion Recovery) for T2-weighted cardiovascular magnetic resonance (CMR) imaging of global and regional myocardial edema.  Our findings support the implementation of AASPIR instead of STIR to visualize myocardial edema, as AASPIR allows for increased SNR and improved image quality.

                  1013.     T2* Measurements in Myocardial Iron Overload: Comparison of Error Models on Optimized
                                 Analysis Protocol

Marco Borri¹

¹University of Turin, Turin, Italy

Iron-induced heart failure is the main cause of death in transfusion-dependent anemia. In the presence of tissue iron, the shortening of the relaxation time constant T2* is used for detection of cardiac iron. It has been shown that patients with short T2* - corresponding with myocardial iron loading - have greater risk of systolic dysfunction. T2* measurements are performed with standardized protocols. An error estimation would help the clinical comparison of T2* measurements, especially in single patient’s follow-up. In this work an optimized analysis protocol is indicated and five different error models are compared.

                  1014.     Comparing Myocardial T2* and T2 Measurements in Thalassemia Patients

Taigang He¹, Peter D. Gatehouse¹, Gillian C. Smith¹, Raad H. Mohiaddin¹, Dudley J. Pennell¹, David N. Firmin¹

¹Royal Brompton Hospital and Imperial College London, London, UK

The aim of this study is to compare myocardial T2* and T2 measurements in vivo in order to establish the relationship between them. In total 137 thalassemia patients were scanned on a 1.5T MRI scanner. It indicates that T2* of 20ms is a useful indicator to identify patient with cardiac iron. It also demonstrates that mmyocardial T2* measurement correlated linearly with T2 measurement in TM patients with iron overload. These findings suggests that both T2* and T2 measurements can be used for assessment of iron overload in the heart for transfusion dependent diseases such as thalassemia.

Philipp Ehses¹, Nicole Seiberlich¹, Peter Nordbeck¹, Florian Fidler², Peter Michael Jakob¹, ², Wolfgang Rudolf Bauer¹

¹University of Würzburg, Würzburg, Germany; ²Research Center Magnetic-Resonance-Bavaria, Würzburg, Germany

Cardiovascular T2* magnetic resonance imaging is a valuable tool in the diagnosis of heart disease. In this work, a multi-spiral method for cardiac T2* determination is presented. A multi-spiral sequence is one that acquires each spiral arm multiple times after an excitation in order to generate multiple T2* contrasts. The spiral trajectory was chosen because it inherently refocuses motion- and flow-induced phase errors, which can be beneficial for cardiac applications.

                  1016.     Postmortem Insitu MRI as an Adjunct to Autopsy for the Diagnosis of Myocardial Infarction

Christian Jackowski¹, ², Marcel Warntjes¹, ³, Anders Persson¹, Michael Thali², Johan Berge⁴

¹University of Linköping, Linköping, Sweden; ²University of Bern, Bern, Switzerland; ³Division of Clinical Physiology, University of Linköping, Linköping, Sweden; ⁴Department of Forensic Medicine, Linköping, Sweden

Autopsy diagnosis of myocardial infarction (MI) is still challenging because many lethal ischemic events do not lead to visible myocardial alterations. In these cases the lack of survival time prohibits myocardial reactions after ischemia, such as edema and inflammatory responses. Postmortem non-contrast enhanced insitu MRI (pm-MRI) may support the diagnosis of peracute MI. Being extremely sensitive to water distribution alterations within the myocardial tissue pm-MRI can accurately visualize areas of decreased micro-circulation. The extent of the ischemia is shown by a lowered signal in T2-weighted and proton density weighted images that have been optimized for the post-mortem conditions. As pm-MRI also reliably demonstrates acute, subacute and chronic infarction it might even serve as an alternative in cases in which traditional autopsy is refused for different reasons.

                  1017.     Diffusion-Weighted Imaging Improves the Quantification of Myocardial Oxygenation When
                                 Arrhythmias Are Present

Kyle Stephan McCommis¹, Ioannis Koktzoglou², Haosen Zhang¹, Debiao Li², Robert J. Gropler¹, Jie Zheng¹

¹Washington University School of Medicine, St. Louis, Missouri, USA; ²Northwestern University, Chicago, Illinois, USA

Double-inversion-recovery (DIR) and diffusion-weighted (DW) prepared T2 images were obtained to determine the myocardial oxygen extraction fraction (OEF) in stenotic canines. These T2 and hyperemic OEF values were determined in the left-anterior descending (LAD) stenosis-subtended region and the remote normal left-circumflex (LCX) region. No significant T2 or OEF differences were found between the two methods. However, the DW technique shows better image quality and OEF accuracy when irregular EKG-triggering or arrhythmias occurs.

                  1018.     Myocardial Iron Distribution in Thalassemia: An in Vivo Study with Black Blood T2* Imaging

Taigang He¹, Peter D. Gatehouse¹, Dudley J. Pennell¹, David N. Firmin¹

¹Royal Brompton Hospital and Imperial College London, London, UK

A black blood T2* technique was employed to investigate myocardial iron distribution in 55 thalassemia patients. The study was conducted on a 1.5T MRI scanner. Images were obtained in the mid-ventricular with 8 echo times and compared for each patient. This is the first in vivo study to demonstrate that myocardial iron deposition is uneven and dominantly in the epimyocardial region in iron overloaded thalassemia patients. This finding agrees well with previously published autopsy reports. These data suggest that T2* measurements in-vivo should use large transmural regions of interest.

                  1019.     Longitudinal Analysis of Heart and Liver Iron in Thalassemia Major

Leila Noetzli¹, Nilesh Ghugre¹, Thomas Coates¹, John Wood¹

¹Childrens Hospital Los Angeles, Los Angeles, California , USA

The relationship between heart and liver iron in patients with thalassemia major has been complex and ambiguous.  There has been evidence of high liver iron being associated with high heart iron, but there seems to be no correlation between the two at any given time.  To address this issue, we retrospectively analyzed patients who had three or more MRIs to estimate their liver and heart iron.  Through this longitudinal analysis, we found that the majority of patient’s hepatic iron concentration (HIC) versus cardiac R2* trajectories follow a counterclockwise hysteresis loop.  In particular, heart iron lags with respect to liver iron.  This finding is helpful to understand the complex relationship between heart and liver iron movement.

MR Safety:  Fields

Zhangwei Wang¹, Timothy J. Mosher¹, Christopher M. Collins¹

¹Penn State University, Hershey, Pennsylvania, USA

Little information has been published regarding the SAR distributions in extremities using dedicated extremity coils. This has led to gross extrapolations from data calculated for other parts of the body, such as the head, to determine operating limits in extremity imaging. Here we evaluate what RF power levels can be used in imaging of the human knee with a dedicated extremity coil without exceeding IEC or FDA limits on SAR or temperature.

                  1046.     SAR Evaluation of 7.0 Tesla Perfusion Imaging with Arterial Spin Labeling Coil

Shumin Wang¹, Hellmut Merkle¹, Lalith Talagala¹

¹LFMI/NINDS/NIH, Bethesda, Maryland, USA

We examined the safety of neck labeling coils at 7.0 Tesla by using both numerical simulations and experimental data. This study combines the SAR contribution of a CASL neck labeling coil during RF labeling and that of a volume transmit coil during image acquisition. Results indicate that CASL perfusion with a neck labeling coil will be possible at even higher fields without exceeding the SAR guidelines.

                  1047.     9.4 T RF Heating: In Vivo Thermoregulatory Temperature Response in Porcine Models

Devashish Shrivastava¹, Robert Schlentz¹, Jeramy Kulesa¹, Lance DelaBarre¹, Carl Snyder¹, Timothy Hanson¹, J. Thomas Vaughan¹

¹University of Minnesota, Minnesota, USA

In vivo thermoregulatory response to RF heating at 9.4 T was studied by measuring temperatures in the heads of eight human-sized porcine models. Temperatures were measured in the scalp and brain by fluoroptic thermometry. Continuous wave RF power was delivered for 2.5-3.4 hours to four anesthetized animals. A four loop head coil was used, which was tuned to 400 MHz. The SAR was maintained close to 3 W/kg. Sham RF was delivered to the other four anesthetized animals to understand the effect of anesthesia on temperatures. Run-away heating response was seen in three out of four RF heated animals.

                  1048.     Proton Resonance Frequency Shift Based NMR Thermometry for Ultra-High Field RF Safety Appl

Devashish Shrivastava¹, Lance DelaBarre¹, Shalom Michaeli¹, Carl Snyder¹, Timothy Hanson¹, J. Thomas Vaughan¹

¹University of Minnesota, Minnesota, USA

The relative variation in the slopes of proton resonance frequency shift related phase changes was studied over 34.6 and 40.6  C in a non-perfused porcine brain at 7 T. This was done to develop an MR thermometry technique with sub-degree celsius accuracy to measure RF heating in porcine models at ultra-high fields. The phase-change slope varied between (-0.01 ppm/ C) and (-0.015 ppm/ C) in the porcine brain.

                  1049.     Subjective Acceptance of 7T: Initial Experience in the First 210 Subjects

Jens Matthias Theysohn¹, ², Stefan Maderwald¹, ², Oliver Kraff¹, ², Christoph Moenninghoff¹, ², Wolfgang P. Becker¹, ², Patrick Kokulinsky¹, ², Michael Forsting¹, ², Mark E. Ladd¹, ², Susanne C. Ladd¹, ²

¹University Duisburg-Essen, Essen, Germany; ²University Hospital Essen, Essen, Germany

The introduction of ultra high field MRI systems for clinical human imaging leads to a raised consciousness regarding subjective patient acceptance and appearance of side-effects. We present our data collected from the first 210 human subjects undergoing a 7T MRI examination. Following the examination, potential sources of discomfort and side effects were rated on a 10-point scale and documented. General acceptance was high and side-effects tolerable. We believe that the willingness to undergo such an examination will be even higher if medical benefit for the individual is expected. No critical situation occurred.

                  1050.     Lowering the Imager Significantly Reduces the Field Exposure of MRI Occupational Workers

Adnan Trakic¹, Hua Wang¹, Feng Liu¹, Hector Sanchez Lopez¹, Ewald Weber¹, Stuart Crozier¹

¹The University of Queensland, Brisbane, Australia

In MRI, healthcare workers can be exposed to strong static and time-varying magnetic fields outside the imager, which can lead to the stimulation of electric fields in the body. Tissue of the central and peripheral nervous system (CPNS) in the head and torso is particularly susceptible. Reported is a simple solution that can notably reduce the head/trunk exposure of MRI operators to both static and low-frequency magnetic fields. The numerical results indicate that the upper body CPNS exposure can be reduced by factors of up to 50 or more, when the scanner is lowered by 1 m in height relative to the normal operator position.

                  1051.     Simple Analytical Equation of the Induced E-Field

Esra Abaci¹, Emre Kopanoglu¹, Vakur Behcet Erturk¹, Ergin Atalar¹

¹Bilkent University, Ankara, Turkey

During MRI, due to the time varying magnetic field, electric field is induced. To derive simplified expression of induced E-field in a cylindrical homogenous volume with radius  ρ ₀ and conductivity s the gradient fields are assumed to be uniform. The analytical E-field expression is simplified using low frequency based assumptions. The result satisfies the expected conditions, e.g. the divergence of E-field is zero and also it is in line with our intuitive understanding on what E field should be. With this simplified expression, electric field behavior inside the body can easily be defined without an additional computational work.

                  1052.     Induced Magnetic Forces in the Human Head During MRI Procedures: A Group Analysis

Ruiliang Wang¹, Dardo Tomasi¹, Gene-Jack Wang¹, Elisabeth C. Caparelli¹, Rita Z. Goldstein¹, Nora D. Volkow², Joanna S. Fowler¹

¹Brookhaven National Laboratory, Upton, USA; ²National Institutes of Health, Bethesda, USA

Understanding the complex distribution of magnetic field gradients and the induced magnetic forces in human head during magnetic resonance imaging is an important safety issue, particularly for high filed MRI. The magnetic force acting on biological tissues that are exposed to an external magnetic field is proportional to tissue susceptibility and the spatial distribution of the static magnetic field. The aim of this work was to map the magnetic force acting on biological tissues of the human head when the subject’s heads are placed in the homogeneous magnetic field of an MRI scanner for a group of healthy subjects

                  1053.     Tailoring of Gradient Coils for Numerical Exposure Evaluations Based on Experimentally Measured B-Field

Feng Liu¹, Adnan Trakic¹, Hector Sanchez Lopez¹, Ewald Weber¹, Stuart Crozier¹

¹The University of Queensland, Brisbane, Australia

Assessment of MRI worker exposures to pulsed magnetic fields produced by gradient coils has recently attracted a lot of awareness in the field of occupational health and safety. To accurately model the exposures, a full three-dimensional distribution of the magnetic field in the vicinity of the magnet end is required. Unfortunately, for many MRI installations, the coil pattern that generates this magnetic field is often not provided by the manufacturer. A method is presented in which the prediction of a current distribution that generates a nearly identical magnetic field pattern is constrained by a number of experimentally measured magnetic field sample points outside the gradient set of interest. The method takes into consideration other important descriptors such as field uniformity in the working volume, gradient coil geometry, driving current, gradient strength, active shielding etc. To demonstrate the application of the method, current density and matching magnetic field distributions of x- and z-axis gradient coils are derived. This enables robust, accurate evaluations of exposures of tissue-equivalent numerical worker models without pre-knowledge of gradient coil patterns.

                  1054.     Tissue-Mimicking Phantoms for a Combined Magnetic Resonance (MR)/ultra-Wideband (UWB)
                                 Radar Technique  [Not Available]

Florian Thiel¹, Florian Schubert¹, Werner Hoffmann¹, Frank Seifert¹

¹Physikalisch-Technische Bundesanstalt (PTB), Berlin, Germany

Our research is aimed towards the synergetic technological development of ultra-broadband (UWB) sounding combined with magnetic-resonance imaging (MRI), to access innovative fields of application such as the imaging of cardiac blood vessels and heart beat monitoring, accurate modelling of electromagnetic wave propagation through heterogeneous, malignant and benign, biological tissue for high-field MRI, and fast and precise identification and localisation of breast cancer. As our first step towards this ambitious aim we report on moveable multilayer tissue-like phantoms specifically designed for testing and development of a UWB-MRI combination setup.

                  1055.     MRI Safety: Quantitative Comparison of RF-Heating on Different MR Scanners Based on the
                                 High Frequency B1-Field

Florian Fidler¹, Toni Hippmann², Marcus Warmuth³, Philipp Ehses⁴, Peter Nordbeck³, Michael T. Friedrich⁵, Wolfgang Geistert⁵, Walter Kullmann², Peter Michael Jakob, ¹⁴, Wolgang Rudolf Bauer³

¹Research Center Magnetic-Resonance-Bavaria, Würzburg, Germany; ²Fachhochschule Würzburg-Schweinfurt, Schweinfurt, Germany; ³University Hospital Würzburg, Würzburg, Germany; ⁴University of Würzburg, Würzburg, Germany; ⁵Biotronik GmbH & Co. KG, Berlin, Germany

The understanding of heating effects in MRI, especially in patients with medical implants, is one of the most important issues in MRI safety. Local electric fields are known as the cause of implant heating, but these are in general not accessible. The vast majority of the electric field is induced from eddy currents generated by the high frequency magnetic B1-field. The purpose of this work was to give a quantitative comparison of SAR maps and therefore heating results on different MRI systems and a simulation based on the knowledge of the B1-field.

RF Transmit Array Hardware

Xu Chu¹, Juan Sabate², Yudong Zhu²

¹GE Global Research Center, Shanghai, People's Republic of China; ²GE Global Research Center, Niskayuna, New York, USA

Current variation in a transmit coil due to load change can be substantial in a setup using a conventional RF power amplifier. The capability of a newly developed ultra-low output impedance RF power amplifier in suppressing such current variation is explored. The new amplifier, when connected to drive a transmit coil with L-type input-matching network, acts approximately as a current source, suppressing current variation due to load change and/or EM coupling. Meanwhile, the output-matching network of the MOSFET transforms the input impedance of the coil into the optimum load of the MOSFET, hence maximizing the available output power.

                  1090.     An Integrated Multi-Channel RF Transmitter for Continuous Arterial Spin Labeling with Multiple Label Coils

Adam Martin Winchell¹, ², Ralf Berthold Loeffler¹, Yong Zhang¹, Ruitian Song¹, Kathleen J. Helton¹, Lawrence L. Wald³, Claudia Maria Hillenbrand¹

¹St. Jude Children's Research Hospital, Memphis, Tennessee, USA; ²University of Memphis and UTHSC Joint Biomedical Engineering Program, Memphis, Tennessee, USA; ³Massachusetts General Hospital, Charlestown, Massachusetts, USA

A complete territorial brain perfusion system was developed using three independent surface-coils for the left, right carotid and vertebral arteries.  Three independent trigger signals controlled by the sequence program gated a single external RF pulse to apply flow-driven adiabatic spin inversion for imaging.  This low cost approach was demonstrated in both phantom and volunteer measurements.  The volunteer measurements produced well-defined and localized territorial perfusion images which correspond with expected normal physiological conditions.

                  1091.     RF Current Source Development for Parallel Transmit Arrays Using a High Power MOSFET

WonJe Lee¹, Eddy B. Boskamp, Thomas M. Grist, Krishna N. Kurpad

¹Uinversity of Wisconsin-Madison, Madison, USA

RF current source technology is gaining ground in parallel transmit arrays in order for B1 field pattern and local SAR control in high fields. In this work we present a RF current source development using a high power MOSFET by means of driven current amplitude and suppression of current induced by a neighboring element. Experimental results demonstrate improved craven current efficiency and reliability against load impedance changes. Simultaneously, current induced by a neighboring element is suppressed by a factor of 17 dB at the closest loop to loop distance by shunting the output capacitance of the chosen MOSFET.

                  1092.     Design of a Strip Transmit Coil/Array for Low Field Open MR

Bing Wu¹, ², Jun Gao¹, Jiabin Yao¹, Cunli Zhang¹, Xiaoliang Zhang², ³

¹GE Healthcare, Beijing, People's Republic of China; ²UCSF, San Francisco, California , USA; ³UCSF/UC Berkeley Joint Graduate Group in Bioengineering, San Francisco & Berkeley, California , USA

RF coils with microstrip transmission line structure have shown the advantages for high and ultrahigh field MRI due to the excellent high frequency performance. In this work, we explore the feasibility of transmit body coil or coil array design using microstrip design technique for low-field open MR applications. Result shows that the proposed microstrip transmit (Tx) coil or parallel transmit array appears to be more compact and efficient than the conventional saddle-type transmit coils.

                  1093.     An 8 Channel TX-RX Head Array for Improved SNR at 3T

Frank Resmer¹, Markus Klarhöfer², Titus Lanz¹

¹Rapid Biomedical GmbH, Rimpar, Germany; ²University of Basel, Basel, Switzerland

The design of an 8 channel TX-RX Array for 3T is described. It provides an SNR comparable to a volume coil of similar dimensions. Sufficient decoupling between elements is vital with TX-RX arrays. So far TX-RX arrays have been built with strip lines or gap designs to decouple individual elements. These approaches provide good intrinsic decoupling but also reduce the coil sensitivity. The coil presented here uses a conventional design with a capacitive decoupling network and so combines a high SNR with good element decoupling.

                  1094.     Efficiency of a 3T Whole Body 16 Channel TEM Transmit Array

Ed Boskamp¹, Scott Lindsay¹, Patrick Gross², Hans-Peter Fautz², Mika Vogel², John Lorbiecki¹, Yudong Zhu³

¹GE Healthcare, Waukesha, Wisconsin, USA; ²GE Research, Munich, Germany; ³GE Research, Niskayuna, New York, USA

A 3T whole body 16 channel TEM style transmit array was built and compared to an 8 loop whole body transmit array. The new 16 channel version has an efficiency of 0.231 uT /ãW  of B+1 in the center of the coil when driven in an emulated homogeneous birdcage mode. The new array displays lower temperature than the 8 loop design. This array is connected to an 8 channel transmit chain. Each channel is sending power via splitter to 2 rungs, which can be opposite (180 degree phase difference) or neighboring (22.5 degree phase difference)

                  1095.     7 Tesla Localized RF Excitation/Reception Using a Highly Coupled Coil and Without B₁ Measurements

Tamer S. Ibrahim¹, YiK-Kiong Hue¹, Lin Tang²

¹University of Pittsburgh, Pittsburgh, Pennsylvania, USA; ²University of Oklahoma, Norman, Oklahoma, USA

The presented results demonstrate that by properly modeling the load, transmit/receive array, and the excitation/reception scheme, an RF power-controlled B1 shimming can be 1) guided with simulations that require a minimum of computational time required (seconds) and 2) efficiently implemented without any B1 measurements.  Localized excitation/reception is demonstrated using a 7T system.

                 1096.     Evaluation of a Phase Shifting Matrix for Body Coil B1+ Shimming in a Dielectric Phantom

Elisabeth Beermann¹, Patrick Gross²

¹GE Healthcare, Solingen, Germany; ²GE Global Research, Munich, Germany

One challenges in high field MRI is the B1-field inhomogeneity due to particular patient RF permittivity. This results in artifacts in the MR images and thus can compromise their diagnostic value. RF shimming and parallel transmit offer a potential solution to this challenge.  In order to interface an 8-channel system with a 16-channel transverse-electromagnetic (TEM) body-array, the use of a phase shifting or Butler matrix was investigated and B1 shimming in a dielectric phantom was performed. The results for a dielectric phantom thus indicate that the advantages of using a phase shifting matrix are similar to those in a non-dielectric phantom previously published. 

                  1097.     Preamp-Like Decoupling and Amplitude Modulation in CMCD Amplifiers for Transmit Arrays

Jeremiah Aaron Heilman¹, Natalia Gudino, Matthew J. Riffe, Markus Vester², Mark A. Griswold

¹Case Western Reserve University, Cleveland, Ohio, USA; ²Siemens Medical Solutions, Erlangen, Germany

Development of current-mode class-D amplifier, including improvements for amplitude modulation, greater efficiency, and decoupling.

Transceiver & Control Systems

                  1130.     A Simple Vector Modulator Approach to Phase and Amplitude Control for B1 Shimming

Ke Feng¹, Xiaojun Chen¹, William A. Grissom², Douglas C. Noll², Mary Preston McDougall¹, Steven M. Wright¹

¹Texas A&M University, College Station, Texas, USA; ²University of Michigan, Ann Arbor, Michigan, USA

B1 shimming is an alternative approach to Transmit SENSE as a method of compensating for full-wave effects at high fields. In order to perform B1 shimming, independent amplitude and phase control of the overall RF pulse is required for each element or input port of a multi-port coil. Digital attenuators and phase shifters with sufficient resolution are quite expensive. This paper presents an inexpensive alternative which combines a vector modulator with digital potentiometers to realize a simple and scalable system to control an array for B1 shimming. In addition, by replacing the digital potentiometers with more expensive fast digital-to-analog boards, the system is capable of full modulation for transmit SENSE. 

                  1131.     RF Switching Matrix Enables 128 Channel Architecture and Dynamic Element-To-Receiver Routing

Yuan Ma¹, Jennifer A. Black¹, Vanish K. Dabra¹, William Peterson¹

¹GE Healthcare, Waukesha, Wisconsin, USA

A high linearity, low-noise radio frequency switching matrix has been created to enable 128-channel architecture and dynamic element-to-receiver routing. This was accomplished using RF simulation tools, careful component selection, and programmable intelligence control.

                  1132.     A Wire-Free, Radio-Frequency, Shielded Projection Window for MRI Suites

David Ian Hoult¹, Patricia Gervai¹, Uta Sboto-Frankenstein¹

¹National Research Council Institute for Biodiagnostics, Winnipeg, Canada

The investigation by fMRI of neural mechanisms associated with visual stimuli is now relatively common. However, the absence of high-quality, wire-free projection windows in MRI suites, particularly at the back of the magnet, often forces investigators to place video projectors inside the shielded room. This poses potential hazards including the risk of projector malfunction and over-heating. Thus a salt-solution window with good optical properties is proposed as a solution to this problem.

                  1133.     Reconfiguration of a “standard” Biospec Spectrometer for Simultaneous 2-Channel Acquisitions:
                                Application for Mouse Brain MRI and MRS

Adrian Rengle¹, Hélène Ratiney¹, Adriana Bucur¹, Sophie Cavassila¹, Olivier Beuf¹

¹INSA-Lyon, Université Lyon 1, Villeurbanne, France

In the field of small animal imaging, the interest for phased-array coil imaging is growing but high field MR experimental systems with multiple receiver channels are still rare and the upgrade of existing systems is relatively expensive. In this work, a standard 4.7T Bruker Biospec Avance II spectrometer was modified to allow simultaneous two-channel acquisitions. Modifications were validated on imaging and spectroscopy on metabolite solution phantom as well as on mice brain using a home-made two-channel array coil. Compared to a single-channel surface coil, the mean SNR was improved by about 20%. Modifications realized for proton multiple-channel acquisitions could also be applied for any X-nucleus. Compared to quadrature detection coils, two-channel coils offer the ability to use parallel acquisitions techniques.

                  1134.     Dynamic Downconversion Module for MR Applications

Jennifer A. Black¹, Yuan Ma¹, Vanish K. Dabra¹

¹GE Healthcare, Waukesha, Wisconsin, USA

A small down-conversion module was developed for 1.5T and 3T MR imaging subsystems to allow for dynamic amplification and increased system modularity. This module places the frequency of the MR signal within the frequency range of an analog-to-digital receiver board. Noise figure, linearity, and amplification re optimized in this 1-inch by 2.5-inch module, which can be used for “mix-on-coil” surface coil development or within large switching matrices.

                  1135.     Efficient Tune and Match with Multiple Transmit Coils

John P. Strupp¹, Edward J. Auerbach¹, Ark Gozubuyuk¹, Gregor Adriany¹, Kamil Ugurbil¹, Pierre Francois Van de Moortele¹

¹University of Minnesota, Minneapolis, Minnesota, USA

An RF switch box with a multi-channel cable assembly was engineered to facilitate the efficient tune and matching of multi-channel transmit coils as used in ultra high field applications.  A particular challenge was finding components that were either made of non-magnetic materials or that could be easily modified to be so.  Typically the system can decreases coil setup time from tens of minutes to just minutes when compared to the traditional method using just an RF sweeper probe tuner and a handful of 50 ohm terminators.

                  1136.     Direct MRI Detection at 3T and 9.4 T Using 16-Bit High-Speed Digital Receiver

Andrzej Jesmanowicz¹, James S. Hyde¹

¹Medical College of Wisconsin, Milwaukee, Wisconsin, USA

A new 16-bit digital receiver with three LTC2208 A/D converters sampling at 100 MHz at 3T and 120 MHz at 9.4T was evaluated for off-line real-time image acquisition.  Tested were: the dynamic range improvement, the maximum usable MRI frequency range and the ability to abandon a reference frequency shift between slice selection pulse and readout time.  At 35T (1500 MHz) the dynamic range increase together with drop of sensitivity makes the dynamic range of this converter equal to the old, 2 MSPS 16-bit A/D converters used in commercial scanners.

                  1137.     Influence of RF Synthesizer Phase Noise on MR Imaging Stability

Martin Nisznansky¹, Joerg Stapf¹, Markus Vester¹, Philipp Hoecht¹, Jan Bollenbeck¹, Wilfried Schnell¹

¹Siemens AG Medical Solutions, Erlangen, Germany

phase noise, RF synthesizer, EPI, fMRI, image-to-image stability, SNR

                  1138.     Ultra Low Noise Preamplifier for Transmit/receive Coils: Upgrading MR Systems Without Further
                                Modifications of the Hardware

Johannes Sell¹, Florian Fidler², Toni Hippmann³, Michael Ledwig², Peter Michael Jakob¹, ²

¹University of Wuerzburg, Wuerzburg, Germany; ²Research Center Magnetic-Resonance-Bavaria, Wuerzburg, Germany; ³FH Wuerzburg-Schweinfurt, Germany

A concept for a noise matched low noise preamplifier with integrated transmit bypass is proposed. It has a single wire connection and can be integrated into many existing MR systems without the need of additional wiring. Successful tuning/matching of transmit/receive coils has been demonstrated. The design is adjustable for a wide frequency range with minor changes.

                  1139.     A 500 W, Broadband, Non-Magnetic RF MOSFET Amplifier for MRI Use

David Ian Hoult¹, Glen Kolansky¹

¹National Research Council Institute for Biodiagnostics, Winnipeg, Canada

Initial experiences in building a push-pull, 500W, broadband, non-magnetic MOSFET amplifier are described. For use with Cartesian feedback, the amplifier can be placed at the magnet bore entrance. Strategies for dissipating 1 kW of heat are described and cold-plate liquid cooling is advocated. Direct Q-factor measurements confirm that the MOSFET dynamic output resistance alone (without feedback) is not high enough to provide good current blocking in coupled phased-array coils. A 4:1 Guanella toroidal transformer and a pancake autotransformer are advocated to provide broadband power performance while coping with the MOSFETs’ large drain-source capacitance.

                  1140.     Scalable Low-Profile Linear Amplifier Module for Parallel Transmission in MRI

Xiaojun Chen¹, Ke Feng¹, Mary Preston McDougall¹, Steven M. Wright¹

¹Texas A&M University, College Station, Texas, USA

Parallel transmission has been suggested as a method to improve RF excitation, in paticular, multi-dimensional and spatially selective excitation at high fields. Both B1 shimming and Transmit SENSE require multiple channels of RF amplifiers, though at lower power levels than may be required with single channels. Particularly as the number of channels increases, the power level requirement per channel drops to values that lend themselves to modular compact construction. This abstract reports our progress in developing a low-profile, linear, high-gain amplifier with a high degree of noise blanking. The module reported here generates 45 watts per channel and offers high efficiency, high linearity, low cost and plug-in modality which make it attractive for scalable parallel transmission systems. It could also serve as a driver stage for whole body Transmit SENSE. This work is designated to be used with a vector modulator designed in-house.

                  1141.     On the Operation of Bluetooth Devices Inside the MR Faraday Cage

General Leung¹, Garry Liu¹, Kevan J T Anderson¹

¹Sunnybrook Health Sciences Centre, Toronto, Canada

Bluetooth technology provides an attractive means of communicating with devices confined within the Faraday cage of the MR magnet. Unfortunately, the Faraday cage significantly attenuates a Bluetooth signal such that communication becomes unreliable. To circumvent this, a USB Bluetooth transceiver can be placed within the waveguide enabling broadcasting of the Bluetooth signal inside the Faraday shield. The purpose of this study was to quantify the communication reliability of Bluetooth devices compared with image noise contamination by compromising the waveguide.

                  1142.     A Versatile USB-Based Control System for Instrumentation in a Magnetic Field

Derek Foreman¹, Glen Kolansky¹, David Ian Hoult¹

¹National Research Council Institute for Biodiagnostics, Winnipeg, Canada

Massively-parallel digital control, from outside the MR suite, of instrumentation (e.g. multiple Cartesian feedback transceivers) in a strong magnetic field close to an MR magnet creates problems of complexity, cross-talk, interference and speed reduction. To solve these problems, a  low-cost, 40 MByte per second USB-based, bi-directional, non-magnetic serial link requiring only 2 coaxial or fibre-optic cables has been implemented, together with a custom programming language that describes the computer-instrument interface hardware.

                  1143.     Quantitative Mechanical Stimulator for FMRI and MicroPET Studies

Wei-Chieh Wong¹, Yen-Yu Shih¹, ², Yun-Chen Chiang¹, ², Chien-Hsiang Huang¹, ², Chen Chang², Fu-Shan Jaw¹

¹Institute of Biomedical Engineering, National Taiwan University, Taipei, Taiwan; ²Institute of Biomedical Sciences, Academia Sinica, Taipei, Taiwan

The present study developed a fully MR-compatible, computer-controlled, pneumatic stimulator to deliver quantitative tactile stimulation. Mechanical stimulation of the hindpaw of rat significantly increased BOLD signal intensity in the primary somatosensory cortex of the hindlimb region (S1HL). The results validated the feasibility of this stimulator and were coincident with the microPET experiment.

Karl-Heinz Herrmann¹, Enrico Wagner¹, Andreas Deistung¹, Ines Krumbein¹, Jürgen R. Reichenbach¹

¹Friedrich-Schiller-University, Jena, Germany

Animal experiments are increasingly conducted on clinical MR systems. For rodents, however, the default trigger systems of whole-body scanners are usually not sensitive enough to detect the tiny motions. Therefore, an MRI-compatible low-cost solution for triggering the respiratory motion of small rodents was constructed.  An optical motion detector was used in conjunction with a robust analog electronic circuit, which converts the original optical into an electrical signal, compensates slow drifts and offsets and finally converts the motion signal into a TTL trigger for the clinical whole body MRI scanner.  The trigger was successfully applied in mouse experiments.

Byeong-Ho Eom¹, Mark Steven Cohen², Inseob Hahn¹, Konstantin I. Penanen¹

¹California Institute of Technology, Pasadena, California , USA; ²University of California, Los Angeles, Los Angeles, California , USA

We present here a description of a novel approach to MRI operating at less than the earth's field and show that the SNR potential is comparable to that of traditional high field instruments.

Shinya Handa¹, Kazuya Taniguchi¹, Katsumi Kose¹, Tomoyuki Haishi²

¹University of Tsukuba, Tsukuba, Japan; ²MRTechonology Inc., Tsukuba, Japan

A local RF shielding method has been developed using a conducting plate and a LC balun circuit for a whole hand MRI system. Using an artificial external noise, the RF shielding performance was quantitatively evaluated. The result demonstrated that the whole hand imaging could be performed without a shielded room.

                  1147.     Intuitive Interface for MR Scan Plane Prescription

Rahul Sarkar¹, ², Chrishnika de Almeida¹, Noureen Syed¹, Sheliza Jamal¹, Jeff Orchard¹

¹University of Waterloo, Waterloo, Canada; ²Sunnybrook Research Institute, Toronto, Canada

We present a novel 3D user interface model for scan plane prescription in MRI.  The interface consists of a hand-held panel that is manipulated by the user in free space, resulting in analogous real-time prescription of the scan plane in the reference volume.  By providing 3D input directly, the interface aims to reduce the cognitive effort associated with conventional interfaces that use the mouse for scan plane prescription.  We describe a basic implementation of this model and the results of a user study that compared the effectiveness of the panel interface with a traditional mouse-based interface.

Thermal Therapy & HIFU

Andrew Webb¹, Eun-Joo Park, Thomas Neuberger, Nadine Smith

¹Penn State University, University Park, Pennsylvania, USA

An MRI technique for non-invasively mapping temperature increases close to the surface of an ultrasound transducer is presented. Using conventional sequences, image artifacts caused by the magnetic susceptibility of the ultrasound transducer preclude this type of measurement. Using susceptibility-compensating sequences, temperature measurements can be obtained ~1 mm from the transducer surface. Applications to phased array cymbal tranducers used for non-invasive insulin delivery are shown.

                  1219.     MR-Temperature Maps of a HIFU CMUT

Serena H. Wong¹, Ronald D. Watkins¹, Mario Kupnik¹, Kim Butts-Pauly¹, Butrus T. Khuri-Yakub¹

¹Stanford University, Stanford, California , USA

Noninvasive surgeries with high intensity focused ultrasound (HIFU) under MR guidance are becoming more popular.  Though piezoelectric transducers have been used traditionally for these applications, capacitive micromachined ultrasonics transducer(CMUT) have shown advantages in ease of fabrication, efficient performance, and minimal self-heating. We demonstrate heating of a gel phantom using HIFU from an unfocused CMUT and also monitoring of this heating using MR-temperature maps in a 3.0 T scanner.

Selaka B. Bulumulla¹, Arjun Arunachalam¹, Keith J. Park¹, Thomas K. Foo¹, Yudong Zhu¹

¹GE Global Research Center, Niskayuna, USA

Overlapping coil arrays have been used as phased array receivers in parallel receive schemes as well as transmit elements in parallel transmit schemes. In this work, we consider surface coil arrays excited by continuous wave RF sources as RF hyperthermia applicators for oncology. Using numerical analysis and experimental work, we study the heating patterns from loop elements and demonstrate the feasibility of selective heating from surface coils. The resulting array has the potential to function as parallel transmitters and phased array receivers for diagnostic imaging as well as an applicator for hyperthermia oncology therapy.

Anil Shetty¹, Roger Price, Jon Schwartz, Rajesh Uthamanthil¹, James Wang, Andrew Elliott¹, John D. Hazle¹, R Jason Stafford¹

¹The University of Texas M.D.Anderson Cancer Center, Houston, USA

A controlled minimally invasive alternative for brain tumor treatments is nanoshell mediated heating with intratumoral fiber placement. These particles are configured to preferentially absorb near-infrared (NIR) light and to emit heat through the process of surface plasmon resonance. Preferential damage was seen on the tumor side in a canine brain tumor model, with minimal damage in the contralateral normal brain. The tumor heated up to a higher lethal temperature (~18¢ªC), compared to the contralateral control side. These preliminary results prove the efficacy of these nanoshells to deliver lethal damage to the tumor in a minimally invasive manner.

                  1222.     Validation of Percutaneous MRI-Guided Laser Ablation with Real-Time MR Temperature Monitoring
                                 in a Canine Prostate Model  [Not Available]

Anil Shetty¹, Roger McNichols, Ashok Gowda, Sherry Klumpp¹, Andrew Elliott¹, Agatha Borne¹, David Brammer¹, John D. Hazle¹, R Jason Stafford¹

¹The University of Texas M.D.Anderson Cancer Center, Houston, USA

The goal was to validate the efficacy of using a MR-registrable perineal template to percutaneously guide and execute multi-fiber laser treatment in a canine prostate model with real-time MR temperature monitoring. The maximum discrepancy between the actual signal void from the laser catheter and the center of the predicted trajectory was less than 1 mm. Real-time monitoring detected a maximum temperature rise of ~ 50 ¢ªC. This minimally invasive approach to laser ablation of tissue in the prostate  performed well, with a maximal positional error of less than 1 mm at an insertion depth of 6 cm.

                  1223.     Perfusion Calculation Based on MR-Temperature Maps and Focused Ultrasound Heating. Theoretical
                                 and Experimental Study

Iulius Dragonu¹, Philippe Lourenço de Oliveira¹, Christophe Laurent, ¹², Baudouin Denis de Senneville¹, Charles Mougenot, ¹³, Chrit Moonen¹, Bruno Quesson¹

¹Laboratory for Molecular and Functional Imaging, Bordeaux, France; ²Saint André Hospital, Service de Chirurgie Digestive, Bordeaux, France; ³Philips Medical Systems, Suresnes, France

The objectives of the present work were to propose a method for quantitative estimation of perfusion rate, thermal diffusivity and energy absorption coefficients and to evaluate the pertinence of the bio-heat transfer equation (BHTE) to model the temperature distribution in presence of perfusion. Ex vivo kidneys were heated with HIFU under 3D thermometry, varying the input flow. Excellent concordance was observed between the theoretical description and the experimental results. This method allows for automatic determination of tissue thermal properties which influence the efficiency of thermotherapy in highly perfused organs.

                  1224.     Contribution of Temperature Dependent T₁-Change, Slice Thickness and Positioning to an Artifact
                                 in Temperature Images of FUS Heating

Viola Rieke¹, Kim Butts Pauly¹

¹Stanford University, Stanford, California , USA

The PRF-change with temperature is frequenctly used to monitor focused ultrasound (FUS) ablation. Because the heating area with FUS is small, partial volume effects are common. In addition, temperature dependend T1-changes can influence PRF based temperature measurements. In this work, we are investigating how partial volume effects, slice position, and temperature dependent T1-changes can influence PRF-based temperature measurements during FUS ablation.

                  1225.     Referenceless Multi-Coil Reconstruction

Viola Rieke¹, Rexford Newbould¹, Roland Bammer¹, Kim Butts Pauly¹

¹Stanford University, Stanford, California , USA

Referenceless PRF thermometry estimates the phase due to a temperature rise in the heating region from the background phase and, therefore, does not need a baseline image for image subtraction. To estimate the background phase the phase images have to be smooth after phase unwrapping. We describe a phase reconstruction method for multi-coil acquisitions that prevents phase discontinuities to allow the combined images to be reconstructed with the referenceless method.

                   1226.     3D Navigated Real-Time Thermometry for Abdominal Imaging

Max Köhler¹, Gregory Maclair², Baudouin de Senneville², Chrit Moonen², Mario Ries²

¹Philips Medical Systems, Finland; ²IMF, CNRS / Univ. Bordeaux2, Bordeaux, France

This work presents a combination of slice tracking combined with 2D image registration and look-up table based phase corrections for 3D navigated real-time MR-thermometry on abdominal organs.

                  1227.     Real-Time Correction of Respiratory-Induced Field Disturbances for PRFS-Based MR-Thermometry
                                in the Human Breast

Silke Hey¹, Gregory Maclair¹, Baudouin de Senneville¹, Yasmina Berber¹, Bruno Quesson¹, Chrit Moonen¹, Mario Ries¹

¹IMF, CNRS / Univ. Bordeaux2, Bordeaux, France

Real-time MR-thermometry based on the proton resonance frequency (PRF) technique allows monitoring of the local temperature evolution during radio-frequency, laser, cryogenics or focused ultrasound thermal ablation. Especially in the human breast, the presence of magnetic field fluctuations induced by the respiratory cycle can lead to thermometry artefacts if no correction is applied. For this purpose a look-up-table-based multi-baseline correction algorithm based on pencil-beam navigator data is applied to MR-thermometry to correct for the periodic B0-field changes. The feasibility of the correction method in real-time is demonstrated experimentally in a healthy volunteer using pencil-beam navigators for respiratory control.

                  1228.     A Hybrid PRF/T1 Pulse Sequence for Rapid, Simultaneous Temperature Tracking in Soft and Adipose
                                Tissues

Nick Todd¹, Dennis L. Parker

¹University of Utah, Salt Lake City, Utah, USA

Here we present a hybrid PRF/T1 pulse sequence capable of measuring those two temperature dependent parameters rapidly and simultaneously.  The PRF information is acquired in the usual, using a gradient echo sequence. The T1 information is extracted by varying the flip angle between consecutive scans and using the relation between signal intensity, T1 and flip angle.  This method allows the TR of the sequence to remain fixed, and keeps scan time short enough for real time temperature tracking.  Results from an oil/water heating experiment demonstrate the pulse sequences ability to monitor temperature changes in both water-based and fat-based substances.

                  1229.     Hot Spot Tracking for Focused Ultrasound Surgery of Liver Using Filtered Venography

Daisuke Kokuryo¹, Etsuko Kumamoto², Atsuya Okada³, Takamichi Murakami⁴, Susumu Fujii⁵, Toshiya Kaihara¹, Kagayaki Kuroda⁶, ⁷

¹Graduate School of Engineering, Kobe University, Kobe, Japan; ²Information Science and Technology Center, Kobe University, Kobe, Japan; ³The Center of Imaging Assisted Minimally Invasive Therapy, Iseikai Hospital, Osaka, Japan; ⁴School of Medicine, Kinki University, Sayama, Japan; ⁵Faculty of Science and Technology, Sophia University, Japan; ⁶Graduate School of General Science and Technology, Tokai University, Hiratsuka, Japan; ⁷Molecular Imaging Research Group, Institute of Biomedical Research and Innovation, Kobe, Japan

A target tracking technique using relative displacements of gravity points of cross sectional images of blood vessels was proposed to guide the ultrasound focus in the liver which moves with respiration. Experiments with healthy volunteer livers demonstrated that the average error of the target position estimation was within 3mm. The error was less than 4mm even with a position prediction process for considering time delay between the time points of position estimation and transducer setup. These results suggested that the proposed technique was sufficient for guiding the focus and hence the imaging slab position and orientation for the "self-reference" thermometry.

                  1230.     Temperature Monitoring with MURPS in a Reduced Field-Of-View

Mohammed Hassan Aljallad¹, ², Jing Yuan³, Magda Pilato³, Lawrence Patrick Panych³

¹University of Massachusetts Lowell, Lowell, Massachusetts, USA; ²Brigham and Women's Hospital, Boston, Massachusetts, USA; ³Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts, USA

MURPS was implemented for reduced field-of-view (rFOV) imaging using 2D RF excitation so that no additional time would be needed to acquire three MURPS slices compared to the time to acquire a single, full field-of-view slice without MURPS. 

                  1231.     Auto-Tracking Self-Reference Temperature Mapping During HIFU Transmission: Ex-Vivo Experiments
                                and Motion Simulations

Yu-Shun Wang¹, Teng-Yi Huang¹, Chih-Ching Wu², ³, Hsu-Hsia Peng², Wen-Shiang Chen, ³⁴, Wen-Yih Isaac Tseng⁵

¹National Taiwan University of Science and Technology, Taipei, Taiwan; ²National Taiwan University, Taipei, Taiwan; ³National Health Research Institutes, Taipei, Taiwan; ⁴National Taiwan University Hospital, Taipei, Taiwan; ⁵Medical College of National Taiwan University, Taipei, Taiwan

MR temperature mapping by proton resonance frequency (PRF) shift has been shown useful to monitor the treatment process of high-intensity- focused-ultrasound (HIFU) thermotherapy. However, the conventional reference subtraction method for temperature mapping has one major drawback. The accuracy of temperature mapping may be affected by tissue movement (e.g. respiration). To solve this problem, we proposed an alternative method which tracked the heating region automatically and estimated its background phase by applying 2D polynomial fitting. In this study, the proposed method was applied on ex-vivo porcine liver experiments. Moreover, simulated motion was added into the acquired image to test the efficiency of our method. The result shows our method is a robust and reliable method for temperature monitoring.

                  1232.     Analysis and Optimization of the IR-GRE Delayed Enhancement MRI Sequence Used for Imaging Post
                                RF Ablation Scars in the Posterior Wall of the Left Atrium

Sathya Vijayakumar¹, ², Eugene Gennaidy Kholmovski¹, Nassir Marrousche¹, Edward DiBella¹

¹University of Utah, Salt Lake City, USA

Radio Frequency (RF) ablation of posterior wall of the left atrium (LA) has become an accepted form of treatment for atrial fibrillations. Recently Delayed Enhancement MRI (DE-MRI) using an Inversion Recovery 3D Gradient Echo (IR-GRE) sequence has been proposed to visualize and assess the extent of post RF ablation scars in the LA. In this study, we analyze the performance of this sequence with respect to flip angles, inversion times and image contrast. The importance of phase images to distinguish scar and fat is also established.

                  1233.     Improved Precision of MR Temperature Mapping of Mobile Organs Using Magnetic Field Modeling

Gregory Maclair¹, ², Baudouin Denis de Senneville¹, Mario Ries¹, Bruno Quesson¹, Pascal Desbarats², Jenny Benois-Pineau², Chrit Moonen¹

¹Laboratory for Molecular and Functional Imaging, Bordeaux, France; ²LaBRI, Talence, France

Real-time thermometry provides real-time temperature monitoring inside the human body and is an interesting tool to control interventional therapies based on thermal ablation. The PRF technique gives an estimate of temperature change at instant t by evaluating phase shifts between dynamically acquired images and reference data sets. The proposed approach for MR-thermometry consists in modeling the contribution of phase changes induced by respiratory motion in the abdomen during a learning step by assuming that demagnetization field changes can be approximated in first order with a linear term. Subsequently, during the intervention, the necessary phase reference maps are calculated in real-time.

                  1234.     Combining 2D RF Excitation, Parallel Imaging and UNFOLD in Focused Ultrasound Heating Experiment

Chang-Sheng Mei¹, ², J. Yuan², B. Madore², N. McDannold², L. P. Panych²

¹Boston College, Chestnut Hill, Massachusetts, USA; ²Harvard Medical School, Brigham and Women's Hospital, Boston, Massachusetts, USA

In many clinical MRI-Guided Focused Ultrasound procedures, the focus is often much smaller than FOV. In addition, for acoustic wave to transfer to the ROI, there must be water in between the sonication source and the heated subject, resulting in even larger FOV and more scan time imaging unwanted region. To improve temporal resolution, we use technique that combines three different approaches at fast imaging: parallel imaging, UNFOLD and 2D RF excitation [3]. Results are shown in phase images where the FOV was fitted to a heated target, the FUS focus region, allowing temporal resolution to be increased by 8 fold.

                  1235.     A New Spectrum-Based Model for MR Thermometry

Xinyi Pan¹, Cheng Li¹, Kui Ying¹

¹Tsinghua University, Beijing, People's Republic of China

In this work, a novel spectrum based model is proposed for MR PRF thermometry. The new model describes the signal under temperature change accurately and can be solved by using a modern spectrum analysis algorithm ¨C the extended Prony method, which offers a good spectral resolution compared to conventional FFT analysis. With this model, the absolute temperature map can be obtained using a multi-echo GRE sequence without suffering from the disturbances caused by the fat component, inter-view motion and field drift. The results of phantom experiment are quite consistent with the real temperature measurement by the thermocouple probe.

                  1236.     Absolute Temperature Imaging with Non-Linear Fat/Water Signal Fitting

Kevin Michael Johnson¹, Venkata Chebrolu¹, Scott B. Reeder¹

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

Previous work has shown improved absolute temperature mapping using fat as an internal reference in spectroscopic imaging in tissues that contain both water and fat. We propose a rapid method signal model based method for fat-water temperature mapping, capable of producing absolute temperature images with far fewer time points than spectroscopic method.  Evaluations in a fat/water phantom show excellent agreement with measurements made using a temperature probe.

                  1237.     Rapid MR Temperature Imaging Based on Model-Predictive Filtering of Undersampled Data

Nick Todd¹, Ran Niu, Mikhail Skliar, Dennis L. Parker

¹University of Utah, Salt Lake City, Utah, USA

Thermal therapies under MR guidance would benefit from an acceleration in the rate at which MR temperature maps are acquired.  For procedures that induce rapid tissue heating, current MR techniques to monitor temperature cannot simultaneously provide adequate temporal resolution, spatial resolution and FOV coverage.  We present an algorithm that combines thermal model predictions with undersampled k-space data to reconstruct accurate temperature maps.  When applied to proton resonance frequency (PRF) data, accurate temperature maps are created using one-sixth of the full measurement data.  Reducing the number of acquired k-space lines allows for faster scan times, greater volume coverage or higher spatial resolution.

                  1238.     Cramer-Rao Lower Bound for Model-Based PRF Temperature Mapping

Cheng Li¹, Xinyi Pan¹, Kui Ying¹

¹Tsinghua University, Beijing, People's Republic of China

In this work, the expression for Cram¨¦r-Rao Lower Bound (CRLB) for model based PRF Temperature mapping is derived by exploiting the characteristic that the temperature estimation intrinsically is a frequency difference estimation of damped complex exponential signal. With the CRLB, the noise performance of two algorithms to estimate the temperature, Prony algorithm and LM algorithm based on ML criterion, is evaluated and compared. By investigating the CRLB dependence on the imaging SNR, number of echoes, echo times, fat/water ratio, the choice of imaging parameters is discussed. The CRLB analysis theoretically provides insight into how the imaging parameters and estimation algorithm affect the noise performance in temperature mapping.

                  1239.     Preclinical Testing of a Second-Generation MRI-Guided Focused Ultrasound System for Transcranial
                                Brain Tumor Ablation

Nathan McDannold¹, Eyal Zadicario², Magdalini Pilatou¹, Ferenc Jolesz¹

¹Harvard Medical School and Brigham and Women's Hospital, Boston, Massachusetts, USA; ²InSightec, Haifa, Israel

This work tested the performance of a second-generation focused ultrasound device developed for noninvasive thermal ablation of brain tumors. Tests were performed using cadaver skulls and tissue-mimicking phantoms. MRI-based thermometry was used to evalu-ate the focal heating during high-power sonications (for ablation), lower power sonications (to verify the focal point localization), and heating adjacent to the skull bone. The device, which uses cavitation-enhanced heating, expanded the range of targetable regions in the brain and increased the ratio between focal heating and skull-induced heating. This ratio was 15.6±5.6 when cavitation was evident and 1.6±0.8 when it was not.

                  1240.     MR Acoustic Radiation Force Imaging: Comparison of Encoding Gradients

Jing Chen¹, ², Ron Watkins¹, Kim Butts Pauly¹

¹Stanford University, Stanford, California , USA

MR acoustic radiation force imaging (MR-ARFI) is a novel method toguide HIFU interventions. It measures the tissue displacement inducedby the acoustic radiation force, which is most suitable forapplications with little temperature elevation. In this work, threedifferent gradient sets for displacement encoding are implemented andcompared for MR-ARFI. By using the repeated bipolar gradients, thenonlinear background phase was reduced, and the SNR was significantlyenhanced. Displacement on the order of submicrons could be detectedwith the improved encoding gradients.

                  1241.     Monitoring of HIFU Treatment Effectiveness by MR Imaging: Advantage of the Magnetization Transfer
                                 Ratio Over the Thermal Dose as Evidenced by Overheating Experiments

Hsu-Hsia Peng¹, Teng-Yi Huang², Hsiao-Wen Chung¹, Chih-Ching Wu¹, Wen-Shiang Chen³, Wen-Yih Isaac Tseng⁴

¹National Taiwan University, Taipei, Taiwan; ²National Taiwan University of Science and Technology, Taipei, Taiwan; ³National Taiwan University Hospital, Taipei, Taiwan; ⁴Medical College of National Taiwan University, Taipei, Taiwan

We performed a deliberately designed overheating HIFU treatment experiment where tissues were heated over 100 °C, to investigate the use of the magnetization transfer ratio (MTR) for evaluation of treatment extent in the presence of boiling-induced air bubbles and the consequent susceptibility artifacts. The thermal dose as conventionally used to estimate cell damage failed to outline the heated areas reliably, whereas the MTR measured after termination of HIFU treatment correctly indicate the heated extents. Geometrical consistency between MTR maps and optical images of the heated spot suggests the importance of simultaneous measurements of temperature and MTR changes for HIFU heating.

                  1242.     Acute and Chronic Magnetization Transfer Ratio Observations in Canine Cryoablation

Andrew B. Holbrook¹, ², Sonal Josan¹, Donna M. Bouley¹, Bruce Daniel¹, Kim Butts Pauly¹

¹Stanford University, Stanford, California , USA

MR-guided cryoablation is a promising minimally invasive treatment of prostate cancer.  However, it is not clear what images acquired after ablation provide in terms of tissue assessment.  Besides contrast-enhanced (CE) imaging, magnetization transfer (MT) imaging is a potential way of assessing tissue in-vivo.  Cryolesions were created in three canines, each imaged three times over 2-3 weeks with MT and CE protocols before being euthanized for histological analysis.  MT contrast could be seen immediately and persisted throughout the chronic experiment, even as hemorrhage decreased and the lesions changed.  MT contrast could complement CE imaging for assessing the prostate as it heals.

Sonal Josan¹, Maurice van den Bosch¹, Jarrett Rosenberg¹, Bruce Daniel¹, Kim Butts Pauly¹

¹Stanford University, Stanford, California , USA

MRI guided cryoablation is a promising minimally invasive therapy for prostate tumors. Contrast enhanced(CE), DWI and MT images can be used to predict the extent of tissue necrosis. The purpose of this work was to evaluate the appearance of the acute canine prostate cryolesion on conventional T1 and T2 weighted images, and compare those to the CE images. The signal enhancement on T2w images was correlated to the freeze area, and the lesion boundaries on T1 and T2 images were compared to the CE lesion.

B₀ Inhomogeneity Correction (Global/Local)

                 1250.     Multiple Resonant Frequency Offset Acquisitions for Imaging of Metallic Implants

Kevin M. Koch¹, R F. Busse², T A. Lewein¹, H G. Potter³, R S. Hinks¹, K F. King¹

¹GE Healthcare, Waukesha, Wisconsin, USA; ²GE Healthcare, Madison, Wisconsin, USA; ³Hospital for Special Surgery, New York, New York, USA

Most metallic implants used in bone and joint arthroplasty induce severe spatial perturbations to the B₀ magnetic field utilized in MR. These perturbations distort slice-selection and frequency encoding techniques utilized in conventional 2D MR imaging and tremendously hinder MR diagnosis of complications from arthroplasty.   A method is presented whereby multiple 3D fast-spin-echo images are collected using discrete offsets in RF transmission and reception frequency.  It is demonstrated that these images can be combined into a composite image that is devoid of slice-plane distortion and possesses greatly reduced distortions in the readout direction, even in the immediate vicinity of metallic implants.

                  1251.     Dynamic Compensation of B0 Field Inhomogeneities Restores Complex FMRI Time Series Activation
                                 Power

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

¹Medical College of Wisconsin, Milwaukee, USA

Fluctuations in the main magnetic field (B0) often occur during fMRI experiments as a result of subject respiration or motion.  The effect is temporal noise in both the magnitude and phase, which can reduce the significance of statistical tests for functional activity.  If a single image in a time series is chosen as a reference, all other images in the series can be corrected using their inherent phase information such that the B0 field offset at each acquisition point is time invariant.  This reduces or removes effects of temporal variations in B0 homogeneity, providing higher quality fMRI time series. 

                  1252.     Towards Parcellated Dynamic Shimming

Michael Poole¹, Richard Bowtell¹

¹University of Nottingham, Nottigham, UK

The need for a homogeneous magnetic field in MRI is well established, especially at high static magnetic field strengths where susceptibility-induced image distortions and signal losses become excessively large. Dynamic shim updating has been shown to improve magnetic field homogeneity to a greater extent than conventional whole volume shimming. Previous work showed by simulation that the magnetic field inhomogeneity can be further reduced if shimming is performed over a series of compact, cuboidal sub-volumes. In this work we corroborate the simulated results with experimental data obtained at 7T and demonstrate a fast, robust automatic field mapping and shim calculation routine.

                  1253.     Dynamic B0 Shimming in the Heart at 3T

Jeff A. Stainsby¹, Venkat Ramanan², Graham A. Wright²

¹GE Healthcare, Toronto, Canada; ²Sunnybrook Health Sciences Centre, Toronto, Canada

Dynamic shim updating has been recently demonstrated to provide benefit in neuro applications. This study investigates the feasibility of performing dynamic shim updates as a function of both spatial location and cardiac phase in cardiac imaging at 3T. Dynamic shim updating was shown to reduce the mean frequency offset across multiple slices and cardiac phases from 28.5Hz to 2.3Hz and in-plane linear field variations from –3.35 and 3.05Hz/cm to –0.19 and 1.15Hz/cm respectively.

                  1254.     Sense Shimming (Sensh); a Fast Approach for Determing Field Inhomogeneities Using Coil Sensitivity
                                 Information

Daniel Nicolas Splitthoff¹, Maxim Zaitsev¹

¹University Hospital Freiburg, Freiburg, Germany

The pursuit for ever higher field strengths and faster data acquisitions has led to the construction of coil-arrays with high numbers of elements. With the SENSE approach, it has been shown by Pruessmann et al., how the sensitivity of those elements can be used for image encoding. We will here present a proof of principle of a method which can be considered a special case of the SENSE method and completely abstains from using encoding gradients. The FID data thus obtained can be used for determining field inhomogeneities; the method has therefore been termed SENSH for SENsitivity Shimming.

                   1255.     Dependence of the View-Angle-Tilting Technique on the Slice Orientation Angle in Correcting
                                  Susceptibility Artifacts

Kwan-Jin Jung¹, ², Chan-Hong Moon², Hua Peng¹

¹Univ. of Pittsburgh & Carnegie Mellon Univ., Pittsburgh, Pennsylvania, USA; ²University of Pittsburgh, Pittsburgh, Pennsylvania, USA

The view-angle-tilting (VAT) technique has been known to be effective in correcting the susceptibility artifacts in slice selective MR imaging. However, it has been discovered that VAT works only when the slice orientation is orthogonal to the longer axis of the object with susceptibility. This discovery has been theoretically analyzed and confirmed by computer simulation and experimentation at 3T on a phantom and a human head.

                  1256.     Correction of Susceptibility Artifacts in Slice Selective MR Imaging at a Specific Slice Orientation Angle

Kwan-Jin Jung¹, ², Chan-Hong Moon²

¹Univ. of Pittsburgh & Carnegie Mellon Univ., Pittsburgh, Pennsylvania, USA; ²University of Pittsburgh, Pittsburgh, Pennsylvania, USA

The effects of susceptibility in slice selective MR imaging include slice offset in the slice direction, and the readout shift in the readout direction. When the projection of the slice offset onto the readout axis has the same length as the readout shift but with an opposite polarity, the geometric shift is corrected on the reconstructed image. This correction condition is met when the slice orientation is made oblique to the longer axis of the object with susceptibility by an angle of arctan(GS/GR). This discovery has been theoretically analyzed and confirmed by a computer simulation and experiments at 3T on a phantom and a human head.

                  1257.     Inhomogeneity Correction at 7 Tesla Using Masked Mean Filtering of Fast and Low Resolution Gradient
                                Echo Reference Data

Michael Schildt¹, ², Kai Zhong², Klaus Dietz Tönnies², Oliver Speck²

¹Leibniz-Institute for Neurobiology, Magdeburg, Germany; ²"Otto von Guericke" University, Magdeburg, Germany

High resolution and high contrast T1 images are widely used as anatomical reference for functional imaging studies. The anatomical information is obtained from 3D gradient echo data (e.g. MPRAGE). Typically, images acquired at high field strength of 7 Tesla suffer from severe B1 inhomogeneity resulting in erroneous segmentation due to strong intensity variation of voxels representing similar tissue (e.g. brain gray or white matter). In this study fast and low resolution gradient echo reference data with minimal intrinsic contrast are used to estimate the B1 inhomogeneity by masked mean filtering and correspondingly correct the MPRAGE intensity variation.

                   1258.     Reduction of Artifacts in Susceptibility Weighted Imaging

Zhaoyang Jin¹, ², Ling Xia², Yiping Peter Du³

¹Hangzhou Dianzi University, Hangzhou, People's Republic of China; ²Zhejiang University, Hangzhou, People's Republic of China; ³University of Colorado Denver School of Medicine, Denver, Colorado , USA

In susceptibility weighted imaging, sub-optimal selection of the size of low-pass filter results in large residual background phase in the phase mask in regions with severe field inhomogeneity and artifacts in the susceptibility weighted images. The minimum-intensity projection (mIP) can also cause signal loss in peripheral regions of the brain. In this study, we demonstrate the dependency of the severity of artifacts in the phase mask on the size low-pass filter. We also present a volume-segmented mIP approach to eliminate the signal drop-off in peripheral region of the brain.

High Field Imaging Methods

Guillaume Madelin¹, Matilde Inglese¹, Niels Oesingmann²

¹New York University, New York, New York, USA; ²Siemens Medical Solutions USA, New York, New York, USA

Double Inversion Recovery (DIR) MRI combines two inversion pulses in order to simultaneously suppress signals from tissues with different T1 relaxation times. In the brain, DIR allows to selectively image gray matter (GM) by nulling the signal from white matter and cerebrospinal fluid at the time of the excitation pulse. Imaging GM structures is important in the study of many neurological disorders such as Alzheimer's disease, epilepsy and multiple sclerosis. This study demonstrates the feasability of 2D and 3D DIR MRI with chemical shift fat inversion recovery (csFatIR) at 3T and 7T.

Joseph S. Gati¹, L Martyn Klassen¹, Ravi S. Menon¹

¹Robarts Research Institute, London, Canada

Balanced steady-state free precession (bSSFP) pulse sequences have found widespread use at clinical field strengths because of short acquisition times, unique contrast and increased signal to noise characteristics compared to similarly acquired gradient-recalled echo (GRE) methods. To date there has been no published demonstration of the advantages of bSSFP over GRE at an ultrahigh magnetic field strength that quantifies the advantages in vivo. This study performs the simple task of comparing signal to noise ratio and contrast in optimized bSSFP images versus acquisition matched GRE images in mouse brain at 9.4 T.

Curtis Andrew Corum¹, Djaudat Idiyatullin¹, Steen Moeller¹, Michael Garwood¹

¹University of Minnesota, Minneapolis, Minnesota, USA

We demonstrate high resolution T1 and proton density weighted imaging in human brain with SWIFT at 4 T. SWIFT has several unique properties. The gradient updating is very sooth so SWIFT is quieter, even than quiet radial sequences. The data can be motion corrected, similar to PROPELLER or other radial sequences, and further is immune to other common gradient based errors such as incidental phase encoding.SWIFT is 40dB quieter than equivalent fast cartesian 3d GRE imgaing sequences. We report our progress in SWIFT at 4 T for human brain imaging and look forward to developing SWIFT's potential for T1 and PD pediatric neuroimaging.

Peter Börnert¹, Wouter M. Teeuwisse², Holger Eggers¹, Mark A. van Buchem², Matthias JP van Osch²

¹Philips Research Europe, Hamburg, Germany; ²Leiden University Medical Center, Netherlands

High resolution spiral imaging could find an interesting application in ultra-high field imaging. Reverse and long TE forward spiral imaging providing strong T2* weighting are of interest for susceptibility weighted imaging showing high anatomical detail. In this work, the basic applicability of forward and reverse spiral imaging at 7T is investigated with respect to performance, necessary corrections and contrast properties. In-vivo results are shown and discussed underlining the great potential of spiral sampling at 7T.

Roland Bammer¹, Anne Marie Saywer¹, Jung Jiing Hsu¹, Gary H. Glover¹, Rexford David Newbould¹

¹Stanford University, Stanford, California , USA

Over the last few years concerns have been raised about lacking contrast in SE T1-weighted sequences. Specifically, 3T’s ability to provide adequate GM/WM contrast in the brain and spine was criticized. Nevertheless, it is well known that the T1 relaxation times of semi-solid tissue will increase with field strength, whereas T1 of CSF as well as T2 relaxation times remain almost unchanged across B0. Recently, it was even shown that T1 dispersion in brain tissue increases with increasing B0. T1 prolongation and dispersion actually benefit T1w imaging. The objective of this study was therefore to determine the optimal T1 contrast at 1.5T, 3T, and 7T and to find out whether or not there is truly a loss in tissue contrast with increasing B0.

Oliver Speck¹, Matthias Weigel²

¹Otto-von-Guericke-University, Magdeburg, Germany; ²University Hospital, Freiburg, Germany

The high SNR achievable at 7T can be utilized to increase the spatial resolution of morphologic T2-weighted images. However, the very high RF-power requirements can dramatically limit the efficiency of multi-spin-echo methods (TSE). HyperTSE has been proposed to reduce SAR without compromising image quality. We demonstrate that with further optimization of the RF-pulses and the flip angle variations, T2-weighted imaging at 7T with high resolution and high volume coverage is feasible without increased scan times due to SAR limitations. Without prolongation of the pulse lengths a four-fold improvement in temporal efficiency compared to a standard TSE implementation is possible.

EPI - Corrections

Magnus Mårtensson¹, ², Anders Nordell¹, ², Bo Nordell¹, ², Stefan Skare³

¹Karolinska University Hospital, Stockholm, Sweden; ²Karolinska Institute, Stockholm, Sweden; ³Stanford University, Stanford, California , USA

A new sequence has been developed to be used as a calibration scan with an altering compensation blip factor for phase encoding, followed by an entropy calculation to find the best compensation blip factor. This method can be used both with single and multi-shot EPI sequences with equally good ghost reduction. Using this calibration scan for each scan plane angle in the exam is a robust technique that results in a major improvement of ghosting artifact suppression for oblique EPI scans. The presented method has great potential when implemented in oblique EPI such as fMRI, DWI, PWI and cardiac imaging.

Magnus Mårtensson¹, ², Mathias Engström¹, ², Bo Nordell¹, ²

¹Karolinska University Hospital, Stockholm, Sweden; ²Karolinska Institute, Stockholm, Sweden

In oblique EPI acquisitions data is sampled in a Cartesian coordinate system, but phase offsets, due to system time delays and eddy currents, will cause the phase encoding lines to be non-equidistantly separated, resulting in non-Cartesian data and thus ghosting. This study proposes a non-Cartesian approach for Cartesian data sets, using entropy calculations to regrid data in order to reduce ghosting artifacts.

Holger Eggers¹, Peter Koken¹, Peter Boernert¹

¹Philips Research Europe, Hamburg, Germany

Gradient system and receive coil imperfections give rise to differences in phase and amplitude between the signal from odd and even echoes in bipolar multi-gradient-echo imaging. A reliable water-fat separation with generalized multi-point Dixon methods based on data from such acquisitions requires a suitable compensation for these imperfections. In this work, a spatially resolved estimation and subsequent correction of phase and amplitude errors is suggested. It is demonstrated to substantially reduce artifacts in whole-body water-fat imaging with continuous table movement.

Fred J. Frigo¹, Sangwoo Lee¹, R. Scott Hinks¹

¹GE Healthcare, Waukesha, Wisconsin, USA

Inherent in the data collection scheme for Echo Planar Imaging (EPI) are phase errors that lead to image artifacts when left un-corrected.  One method for reducing phase errors is to use the nearest-neighbor EPI phase correction approach.   For SENSE based image reconstruction, the phase characteristics between channels must be preserved after EPI phase correction.  The proposed method of applying nearest-neighbor EPI phase correction for each spatial location and each channel provides a robust and highly effective phase correction technique for SENSE based EPI that can accommodate double oblique scan geometries and has demonstrated excellent image quality.

Yung-Chin Hsu¹, Ching-Han Hsu¹, Wen-Yih Isaac Tseng²

¹National Tsing Hua University, Taiwan; ²National Taiwan University College of Medicine and Hospital, Taiepi, Taiwan

In DWI, diffusion gradients usually incur residual eddy currents which lead to gross geometry distortions. Previous retrospective methods usually assume the deformation field is linear which may not be true for high b-value DWI images. In the present study, we propose a generalized algorithm incorporating with two DWI images which are acquired with opposite diffusion gradients. With this algorithm, higher order of the deformation field can be taken into consideration. The proposed method is applicable to DSI or QBI dataset where high b-value is used and data points are symmetric about the origin in the q-space.

Relaxometry & Quantitation

Weidong Tang¹, Stanley J. Reeves¹, Donald B. Twieg²

¹Auburn University, Auburn, Alabama, USA; ²University of Alabama at Birmingham, Birmingham, Alabama, USA

This paper proposes a new reconstruction method for single-shot parameter assessment by retrieval from signal encoding (SS-PARSE). The cubic convolution interpolation is used to reconstruct higher resolution images with limited data from short read-out time. A polynomial approximation is suggested to address the problematic exponential time function in SS-PARSE model. By this approximation, reconstruction algorithm based on conjugate-gradients can be accomplished by FFTs. Results with this method show sharper edges and smoother reconstructions in areas known to be smooth.

Keiko Miyazaki¹, David J. Collins¹, Dow-Mu Koh¹, David J. Hawkes², Martin O. Leach¹, Matthew R. Orton¹

¹Institute of Cancer Research, Sutton, UK; ²University College London, London, UK

The variable flip angle method allows relatively accurate estimation of T1 in a short time period compared to lengthy inversion recovery methods. Here, we propose a simple and efficient method of obtaining the optimal pair of flip angles which would enable accurate T1 estimations in a sample with a large T1 range by using an optimization scheme. This is of great importance in numerous areas in MRI, particularly in DCE-MRI studies where accurate determination of T1 is crucial in obtaining quantitative parameters which are widely used to monitor disease progression and regression.

Eleanor F. Cox¹, Penny A. Gowland¹

¹University of Nottingham, Nottingham, UK

T₂ and T₂¢ was measured in the brain at 1.5, 3 and 7T using a gradient echo-spin echo (GESE) sequence and T₂ was also measured with single shot SE-EPI. Both sequences were used to measure values of T₂ in grey (both frontal and occipital) and white matter in four healthy volunteers at all three field strengths. Inverse T₂ contrast was seen in the occipital region of the brain. The T₂ values measured agreed with previously published data and the GESE sequence was calibrated against SE-EPI results using phantoms.

Eleanor F. Cox¹, Caroline L. Hoad¹, Penny A. Gowland¹

¹University of Nottingham, Nottingham, UK

It is possible to measure T₂ and T₁ simultaneously using a T₂-prepared balanced turbo field echo sequence (T₂-prep bTFE) which has particular application in the abdomen at 3T. The T₂-prep bTFE sequence was calibrated against EPI using gel phantoms and was then used to measure T₂ and T₁ in the kidneys, liver and spleen. The results were comparable to previously published results using spin echo and inversion recovery FSE. The ability to measure T₂ and T₁ in a reasonable time has important applications in studying changes of gastrointestinal (GI) contents, such as viscosity and dilution, in the GI tract.

                 1413.     B1 Insensitive Saturation Recovery T1 Measurements at 3 Tesla Using Water Suppression Enhanced
                          Through  T1 Effect (WET) Saturation Pulse

Frederic Courivaud¹, ², Henrik B.W. Larsson³, ⁴

¹Philips Medical Systems, Oslo, Norway; ²Rikshospitalet, Oslo, Norway; ³Glostrup Hospital, University of Copenhagen, Copenhagen, Denmark; ⁴Norwegian University of Science and Technology (NTNU), Trondheim, Norway

Quantitative measurements of MR relaxation properties (T1 and T2) are relevant parameters for diagnosis and prognosis of various diseases. T1 measurement is particularly important in the development of T1-based contrast enhanced perfusion measurements. However, most of T1 measurements methods at high field require B1 mapping correction before signal analysis. We demonstrate here the use of optimized four-pulse Water suppression Enhanced through T1 effects (WET) for B1 insensitive saturation recovery T1 measurements at 3T, without the use of B1 mapping. T1 values in accordance with published results and WM/GM segmentation of the brain are shown.

Christian Graff¹, Eric Clarkson¹, Kenneth L. Weiss², Maria I. Altbach¹

¹University of Arizona, Tucson, Arizona , USA; ²University of Cincinnati, Cincinnati, Ohio, USA

T2* measurements are important for the quantification of physiological events or pathologies related to susceptibility changes in tissue. When fat is present, the T2* signal decay is modulated by the chemical shift of the various hydrogen types that make up the fat molecule leading to improper T2* estimation. In this work we present a signal equation model for extracting T2* for fat, minimizing the effects of the intra-molecular chemical shift.

Zhongliang Zu¹, Yanming Yu¹, Qi liu¹, Xuna Zhao², Min Chen³, Shanglian Bao¹

¹Peking University, Beijing, People's Republic of China; ²Philips MRI Center, Beijing, People's Republic of China; ³Beijing Hospital, Beijing, People's Republic of China

DESPOT2 is a rapid T2 mapping technique based on acquisition of a pair of or several SSFP images and prior knowledge of T1. However, recent research has reported that in biological tissues the steady-state signals of SSFP deviate from theoretical predictions based on Bloch equations, which could be attributed to magnetization transfer (MT). Therefore, the accuracy of DESPOT2 is doubted. We analysed the MT effect on T2 measurement. Optimized imaging parameters were proposed to avoid the MT effect on T2 mapping.

Dirk K. Müller¹, Thies H. Jochimsen¹, Torsten Schlumm¹, Harald E. Möller¹

¹Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany

A challenging aspect of quantitative magnetization-transfer (qMT) is the complex model describing a solid and a semi-solid pool exchanging magnetization, causing long scanning times and difficult post processing. We present a fast fitting algorithm for qMT imaging which greatly reduces post-processing times. Based on the model of Ramani, a combination of brute force and Levenberg-Marquardt algorithms leads to a dramatic reduction of processing times which could eventually be used for online reconstruction in future clinical applications.

Ana-Maria Oros-Peusquens¹, Maija Laurila¹, N. Jon Shah¹

¹Institute of Medicine, Research Centre Juelich, Juelich, Germany

We report the results of a 12-volunteer study on the mapping of the T1  relaxation time at 1.5T, 3T and 4T and its field and age dependence. Three nearly identical whole-body scanners have been employed, operating at field strengths of 1.5T (Siemens Avanto), 3T (Siemens Trio) and 4T (Siemens/Bruker MedSpec). Close to whole-brain T1 mapping was performed using TAPIR, a sequence based on the Look-Locker method, which uses an interleave of slice and time point read-outs for fast multi-slice acquisition whilst maintaining good temporal resolution. The whole-brain histogram of T1 values was fitted using a superposition of three Gaussian distributions. The dependence of the centroids of the WM and GM peaks on field strength was investigated and compared to empirical formulae by Bottomley or Fischer. The age dependence of the centroids was also investigated; no clear behaviour was observed for the WM; a clear decrease of the T1 of GM was found at all three fields.

Christian Labadie¹, ², Jing-Huei Lee³, Harald E. Möller⁴

¹Max Planck Institute  for Human Cognitive and Brain Science, Leipzig, Germany; ²University of Leipzig, Leipzig, Germany; ³University of Cincinnati, Cincinnati, USA; ⁴Max Planck Institute for Human Cognitive and Brain Science, Leipzig, Germany

A CONTIN-based procedure to estimate the inversion-pulse efficiency in the context of multi-exponential longitudinal relaxation is proposed and compared to a mono-exponential three-parameter NLLS fit.  Simulations using continuous T₁ distributions show that the three-parameter NLLS fit underestimates the efficiency when a small T₁ peak is present below 400 ms.  Such underestimation of the efficiency is apparent in the white matter of the human brain at 4 Tesla.  The novel procedure enables to adequately estimate the inversion-pulse efficiency necessary to perform a reliable multi-exponential analysis of spin-lattice relaxation. 

Peter Roland Seevinck¹, Clemens Bos², Chris J.G. Bakker¹

¹University Medical Center, Utrecht, Netherlands; ²Philips Medical Systems, Best, Netherlands

To quantify ultrashort T2* components associated with the increasing interest in high field MR as well as in (super-)paramagnetic contrast agents (iron-oxide or holmium), we propose a post-processing methodology, based on the incorporation of S0 (S at t=0ms) in the fitting algorithm, which is generally applicable for T2* relaxometry and does not increase scan time. Sub-millisecond T2* species introduced by holmium-loaded microspheres in gel and ex vivo rabbit liver were assessed qualitatively and quantitatively using S0-fitting and results were compared to quantitative UTE. S0-fitting is accurate and time efficient and therefore suitable for in vivo studies.

Kevin Harkins¹, Jean-Philippe Galons¹, Timothy Secomb¹, Theodore Trouard¹

¹University of Arizona, Tucson, USA

The presence of compartmental T2 differences in tissue has been discount due to the experimental measurement of monoexponential T2 decay in tissue. To assess the role of exchange, T2 relaxation and noise on fitting of signal decay, we have developed a finite element model consisting of square cells separated from extracellular space by a permeable membrane.  In the presence of physiological membrane permeability and reasonably high SNR, tissues with considerably different T2 relaxation times in the intra and extracellular spaces can exhibit monoexponential behavior.

Julien Sénégas¹, Clemens Bos², Hannes Dahnke¹

¹Philips Research Europe, Hamburg, Germany; ²Philips Medical Systems, Best, Netherlands

Usually, reproducibility of relaxation time mapping is assessed, by comparing relaxation time values across regions of interests in multiple patients. However, generally, little is known on the precision of the individual experiment performed in a specific subject. Here, a maximum likelihood method is presented that calculates a map of the relaxation time standard deviation, from a single mapping experiment, accounting for the noise level of the input images and the sequence parameters. The proposed method was evaluated in simulations and in vivo in T₂ mapping of the knee. Knowing the relaxation time standard deviation, the significance of (localized) relaxation time changes over experiments can be more precisely stated.

Matthew T. Latourette¹, James E. Siebert¹

¹Michigan State University, East Lansing, Michigan, USA

T2* relaxometry gains increasing importance with its new clinical role measuring regional iron concentrations for managing iron overload patients.  Using Monte Carlo simulations, this study evaluates the performance of five methods for parameter estimation for T2* quantitation from multi-echo gradient echo data.  The simulations show that nonlinear least squares fitting of data corrected by S_CÖ|S²-2_ς ²| and maximum likelihood estimation provide the most accurate and precise results. Earlier literature R2* values published for clinical decision thresholds can still be used. Caution should be given for results based upon methods with bias and higher variance especially when number of subjects is small.

Mina Kim¹, ², Joseph Gillen, ², Jinyuan Zhou, ², Peter Christiaan van Zijl, ²

¹, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA; ²Kennedy Krieger Institute, Baltimore, Maryland, USA

Chemical exchange saturation transfer (CEST) imaging exploits exchange-based magnetization transfer (MT) between solute and water protons. CEST measurements generally require an asymmetry analysis of the water saturation spectrum (Z-spectrum) with respect to the water frequency, a process exquisitely sensitive to magnetic field inhomogeneities. We show that direct water saturation images acquired with a brief low-power RF saturation pulse can be used to measure the precise water frequency in each voxel, allowing proper centering of Z-spectra independent of spatial B0 variations. After validation in a phantom, the approach is demonstrated in vivo to detect glycogen in human muscle at 3T.

Mara Cercignani¹, ², Daniel C. Alexander², Rebecca S. Samson², Mark R. Symms², Gareth J. Barker³

¹Fondazione Santa Lucia, Roma, Italy; ²UCL, London, UK; ³Institute of Psychiatry, KCL, London, UK

Here, we address the design of an optimal acquisition protocol for quantitative MT at 3.0T, and assess the performances of this technique at 1.5T and at 3.0T. This involves selecting the most appropriate acquisition parameters, such as TR and flip angle, as well as the optimal sampling scheme, under some field strength dependant constraints. We compare 3 acquisition protocols (2 at 3T and 1 at 1.5T) performing a formal analysis of advantages and limitations of higher field strength using both numerical simulations and real data, in order to determine whether the SNR benefits at 3T outweigh the disadvantages.

Nicholas P. Blockley¹, Alexander G. Gardener¹, Susan T. Francis¹, Penny A. Gowland¹

¹University of Nottingham, Nottingham, UK

The relaxation properties of blood are key to many functional MRI techniques. The transverse relaxation rate, R₂^*, is of particular relevance to BOLD imaging and MR angiography. In this work we investigate the dependence of R₂^* on the oxygenation and the concentration of ProHance contrast agent in whole blood. The variation of R₂^* with contrast agent concentration was non linear. These experiments were performed at field strengths of 1.5T, 3.0T and 7.0T allowing the trends in relaxivity as a function of magnetic field strength to be studied.

LEI JIANG¹, Nicholas Blockley¹, Catherine Ludman², Sue Francis¹, Penny Gowland¹

¹University of Nottingham, Nottingham, UK; ²Nottingham University Hospital NHS Trust, Nottingham, UK

The relaxation properties of blood are key to many MRI techniques. In this work we measure the R1 and R2* relaxivities of a blood pool contrast agent Gadofosveset (Vasovist™, Schering) in whole human blood at 37oC at 1.5 T, 3.0 T, and 7.0 T. We found that the R1 relaxivity reduced with field strength. The R2* relaxivity was non linear, but showed greater sensitivity to contrast agent concentration at higher field strength.

Maja C. Sohlin¹, Jan Sedlacik², Jürgen R. Reichenbach², Lothar R. Schad³

¹German Cancer Research Center, Heidelberg, Germany; ²Friedrich Schiller - University, Jena, Germany; ³Faculty of Medicine Mannheim, University Heidelberg, Germany

A signal decay model valid in the static dephasing regime is commonly used to map the oxygen extraction fraction (OEF) in the brain. In this work, the model was examined in a phantom study under non-static conditions, i.e. small vessel radius. Our measurements show that water diffusion must be taken into account to obtain proper OEF values for small vessel diameters. Both methods become unstable when more than two fit parameters are used, this can be a severe problem in vivo. Furthermore, using T2 as a fit parameter in the static model seems to give misleading results.

Andrew M. Prantner¹, G L. Bretthorst, Jeffry J. Neil, Joel R. Garbow, Joseph J H Ackerman

¹Washington University in St Louis, St Louis, Missouri, USA

Longitudinal relaxation rates in vivo are typically modeled as a monoexponential function because few TI values are collected. When the number of TI values is increased, multiple R1 components can be resolved in peripheral nerve but not in brain gray matter. High-resolution IR data on in vivo rat brain using either 64 or 128 exponentially spaced TI values indicates that there is a ubiquitous distribution of voxels in rat brain gray matter that are best modeled as a biexponential function because of magnetization transfer.

Yu-Guang Meng¹, Hao Lei¹

¹Wuhan Institute of Physics and Mathematics, Chinese Academy of Science, Wuhan, People's Republic of China

In this work, a novel fast z-shimming method was proposed to acquire quantitative T₂^* maps in the presence of macroscopic susceptibility artifacts. The method is based on acquisition of two gradient-echo images in the same scan, one at very short echo time without compensation and the other at long echo time but with compensation for in-slice field inhomogeneities. The feasibility of the method was demonstrated in vivo.

Richard D. Dortch¹, ², Thomas E. Yankeelov¹, Mark D. Does¹

¹Vanderbilt University, Nashville, Tennessee, USA

Multiexponential T₂ (MET₂) analysis is capable of decomposing bulk NMR signal into components that represent underlying tissue compartments. The goal of this preliminary study was to determine whether MET₂ analysis could be applied in a rat glioblastoma tumor model in vivo. Results showed tumor signal to be biexponential, with a short-T₂ component (16 ± 12 ms) representing 6 ± 8 % of the signal and a long-T₂ component (75 ± 12 ms) representing the remaining signal fraction. Thus, MET₂ holds promise as a non-invasive tool for characterizing tumor microenvironment in vivo.

Nadim Jon Shah¹, Irtiza Ali Gilani¹, Heiko Neeb¹, A M. Oros-Peusquens¹

¹Institute of Neurosciences and Biophysics (Medicine), Research Centre Jülich, Jülich, Germany

Measurement of the spin-lattice relaxation time constant, T1, can provide important insights into pathological processes in the human brain such as multiple sclerosis or hepatic encephalopathy. Acquisition times, however, tend to be long and therefore reductions through line sharing are desirable. TAPIR, a Look-Locker based multislice, multishot sequence for rapid T1 mapping  is capable in vivo quantification with high spatial resolution by interleaving the slices and time-points. The objective of this work is to accelerate TAPIR using a multi-channel, line-sharing method which acquires sufficient k-space data for high quality T1 mapping. The method is generally applicable to relaxation time mapping.

Kai Tobias Block¹, Martin Uecker¹, Jens Frahm¹

¹Biomedizinische NMR Forschungs GmbH, Goettingen, Germany

The present work describes a novel reconstruction method for radial fast spin-echo MRI which obtains a spin-density and relaxivity map directly from the measured data without calculating intermediate images. This allows for an efficient T2 quantification from a single radial data set. The method is based on an inverse formulation of the problem and involves a modeling of the received MRI signal. Because a solution is found by a numerical optimization procedure, the approach makes optimal use of all data acquired. Experimental data for the human brain in vivo is presented to demonstrate the feasibility of the method.

Carl Ganter¹, Rene Botnar², Marcus Settles¹

¹Technical University Munich, Munich, Germany; ²King’s College London, London, UK

Rapid T1 measurements like Look-Locker or IR-TrueFISP are very susceptible to low SNR values, especially near the zero-crossing of the magnetization. But even for very low SNR, reliable results can still be obtained, if the measured signal is fitted against its expectation value according to the Rician distribution, using noise a additional fit parameter. This is particularly helpful, when noise is not easily accessible in the experiment, e.g. if parallel imaging is used.

Jan Scholz¹, Heidi Johansen-Berg², Sean Deoni²

¹FMRIB Centre, Oxford, UK; ²FMRIB Centre, UK

T1 relaxation times provide quantitative assessment of pathology and have the potential to provide a measure of structural changes due to disease, recovery or learning. We acquired high-resolution whole-brain quantitative T1 maps with the DESPOT1 protocol in multiple subjects on several occasions to access the intra- and inter-session reproducibility of T1 measurements. These values will be useful to determine power and sample size required to detect changes in T1 values over time. In combination with other markers of structural properties these T1-maps might help to elucidate the processes underlying structural changes in the human brain.

Dehe Weng¹, ²

¹1Siemens Mindit Magnetic Resonance Ltd, Shenzhen, People's Republic of China; ²Institute of Biophysics, Chinese Academy of Sciences, Beijing, People's Republic of China

Introduction  New technique for multiple species separation based on T1 was introducedMethod  Description theory of this methodResult  Separation result images with phantomDiscussion  Advantage and disadvantage of new technique

Jessy J. Mouannes¹, Wanyong Shin², Maulin Shah¹, Timothy J. Carroll¹

¹Northwestern University, Chicago, USA; ²National Institute of Health, Bethesda, USA

An accelerated method for brain T1 mapping before and after contrast agent injection  for MR perfusion quantification is presented, which has the potential to substitute for the conventional, time-consuming T1 mapping technique that applies least square fitting to compute T1 value for each voxel. By including this new method in our fully automatic reconstruction chain for cerebral blood flow and cerebral blood volume quantification using the Bookend technique, we can facilitate the widespread dissemination of this technique and its clinical use.

Image Reconstruction

Thomas Sangild Sørensen¹, David Atkinson¹, Redha Boubertakh, ¹², Tobias Schaeffter², Michael Schacht Hansen³

¹University College London, London, UK; ²King's College London, London, UK; ³University College London, UK

This presentation describes an implementation of non-Cartesian SENSE and kt-SENSE accelerated on commodity graphics hardware. This inexpensive hardware platform is now fully programmable and very suited for solving reconstruction problems. We show that for both SENSE and kt-SENSE the reconstruction time per frame is now below the acquisition time providing non-Cartesian reconstruction with only minimal delay between acquisition and subsequent display of images. This is demonstrated by four-fold and eight-fold undersampled real-time radial imaging reconstructed in 25 ms to 55 ms per frame.

Michael Schacht Hansen¹, David Atkinson¹, Thomas Sangild Sorensen¹

¹University College London, London, UK

This project demonstrates that modern commodity graphics cards (GPUs) can be used to perform fast Cartesian SENSE and k-t SENSE reconstruction. Specifically, the SENSE inversion is accelerated by up to two orders of magnitude and is no longer the time-limiting step. The achieved reconstruction times are now well below the acquisition times thus enabling real-time, interactive SENSE imaging, even with a large number of receive coils. The fast GPU reconstruction is also beneficial for datasets that are not acquired in real-time. We demonstrate it can be used for interactive adjustment of regularization parameters for k-t SENSE in the same way that one would adjust window and level settings. This enables a new way of performing imaging reconstruction, where the user chooses the setting of tuneable reconstruction parameters, in real-time, depending on the context in which the images are interpreted.

Eric Allen Borisch¹, Roger C. Grimm¹, Phillip J. Rossman¹, Clifton R. Haider¹, Stephen J. Riederer¹

¹Mayo Clinic, Rochester, Minnesota, USA

Real-time 3D MR reconstruction, enabled through parallel imaging techniques such as SENSE and GRAPPA, places significant communication and computational challenges on a reconstruction system. We have designed and implemented a distributed, scalable cluster-based image reconstruction system that has enabled real-time 3D reconstruction and visualization during data acquisition. The hardware and software techniques used will be described, as well as reconstruction performance and benchmarks.

Justin P. Haldar¹, Sam S. Stone¹, Haoran Yi¹, Stephanie C. Tsao¹, Bradley P. Sutton¹, Wen-mei W. Hwu¹, Zhi-Pei Liang¹

¹University of Illinois at Urbana-Champaign, Urbana, Illinois, USA

Advanced algorithms for image reconstruction are becoming increasingly common, but their utility is limited by computational requirements.  In this work, we show that significant improvements in reconstruction speed can be achieved by performing data-parallel computations on graphics processing units (GPUs).   Specifically, we leverage the resources of a single NVIDIA GeForce 8800 GTX to achieve computational performance of more than 150 GFLOPS, hundreds of times faster than what is reported on a single modern central processing unit (CPU).

Peter A. Wassenaar¹, Joseph Dunbar¹, Donald William Chakeres¹, Darlene Meeks¹, Michael V. Knopp¹, Petra Schmalbrock¹

¹Ohio State University, Columbus, USA

Inversion Recovery dual turbo gradient echo (IR-dTFE) imaging at ultrahigh field strength can generate images with three types of contrast simultaneously. These are T1 weighting and TOF- arteriography on the first echo and BOLD venography on the second echo. For the MRA information to be useful it is necessary to reconstruct MIP images depicting the vasculature in three dimensions. However, while arteries and veins are well depicted on the original images, standard MIP fails because of increased skull and parenchymal signal. We present here a method for processing 7T IR-dTFE images allowing for MIP visualization of arteries and veins.

Patrick Le Roux¹

¹GEHC, Palaiseau, France

The non-CPMG sequence permits to acquire phased object in Spin Echo mode without relying on the Carr-Purcell-Meiboom-Gill constrain. But it still suffers from some magnitude  modulation when the initial phase of the magnetization varies. It is shown that a slight modification of the reconstruction process, projection direction or change of the reference axis,  permits to reduce this anisotropy effect to less than 1%.

Bruno Madore¹

¹Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts, USA

As the field of dynamic imaging is expanding and new methods are introduced, reliable methodologies to assess the performance of these methods would be highly desirable. We propose here a 2D Modulation Transfer Function (MTF) approach, whereby the performance of a given method gets tested at every combination of spatial and temporal frequencies. As an example, the performance of two dynamic imaging methods, kt-SENSE and UNFOLD-SENSE, was tested.

Jason Kraig Mendes¹, Dennis L. Parker¹

¹University of Utah, Salt Lake City, USA

Imaging time constraints often prevent the acquisition of a full k-t space for dynamic objects.  In many instances, only part of the k-t space is acquired and the missing information can be approximated using various techniques.  UNFOLD is a method which assumes that more than one spatial point can share the same temporal bandwidth without overlap.  This work presents a method based on the theory of UNFOLD that allows some degree of overlap in the temporal bandwidth.  This can be accomplished with a Taylor series approximation of the signal intensities from each spatial point in the image.

Nicolae Cîndea¹, ², Freddy Odille¹, ², Gilles Bosser³, ⁴, Jacques Felblinger², ³, Pierre-André Vuissoz¹, ²

¹Nancy University, Nancy, France; ²Inserm, Nancy, France; ³Centre Hospitalier Universitaire, Nancy, France; ⁴Institut Régional de Réadaptation, Nancy, France

Heart MRI reconstructions in free-breathing acquisition is a challenge. Usual techniques assume identical rescaled copies of heart cycle independently of RR length. In this work, a method for cardio-respiratory acquisition and reconstruction in free-breathing taking RR into account is presented. This method uses a Reproducing Kernel Hilbert Space (RKHS) interpolation scheme for retrospective gated reconstruction in 3 dimensions (cardio-respiratory phase and instantaneous heart rate). With this scheme, it is possible to study  the cardio-respiratory interactions and particularly respiratory modulation of heart rate and diastolic filling. Clinical cardiac images were reconstructed at a fixed cardio-respiratory phase for two different RR values.

Martin Uecker¹, Kai Tobias Block¹, Jens Frahm¹

¹Biomedizinische NMR Forschungs GmbH am MPI für biophysikalische Chemie, Goettingen, Germany

Image reconstructions from undersampled data such as in parallel imaging require the inversion of an ill-posed linear (or non-linear) system. In these cases the quality of the reconstruction critically depends on the choice of the regularization which requires a trade-off between substantial noise and still visible undersampling artefacts. Here, we propose an automatic stopping criterion which terminates the iteration as soon as all undersampling artefacts are fully removed. This is accomplished when the energy density of the inferred k-space positions equals the energy density of the directly measured positions. In an image without undersampling artefacts both quantities should be equal.

Volker Rasche¹, Axel Bornstedt¹, Vinzenz Hombach¹

¹University of Ulm, Ulm, Germany

The variable gridding kernel extent technique is applied for the reconstruction of data acquired on a Cartesian sampling pattern with gradually decreasing dasmpling density in phase encoding (PE) dierection. Examples of carotid artery images show a possible reduction of imaging time of about 70% under the compromise of grdually decreasing spatial resolution along PE direction.

Dimitris Mitsouras¹, Onur Afacan², Dana H. Brooks², Frank J. Rybicki¹

¹Harvard Medical School & Brigham And Women's Hospital, Boston, Massachusetts, USA; ²Northeastern University, Boston, Massachusetts, USA

Variable density k-space trajectories, such as spirals, have been recently used in many applications to enhance acquisition speed. In such schemes k-space sampling density is reduced below the Nyquist rate, typically as a function of distance from the k-space center. While their use and diversity is rapidly increasing, a systematic comparison and characterization of their tradeoffs has not been performed to-date. Moreover, various reconstruction methods have been proposed for such trajectories. This work provides such an analysis of different variable density trajectories and different reconstruction methods using well-founded metrics such as SNR, root-mean-square error, and point-spread function analysis.

Jaeheung Yoo¹, Jong Chul Ye¹

¹KAIST, Daejeon, Republic of Korea

For dynamic MR imaging , radial trajectory is often employed since it is more robust to motion artifacts due to the over-sampled k-space center region. Furthermore, scan time can be further reduced using view undersampling. However, this view undersampling often causes visually annoying streaking artifacts. In radial k-t FOCUSS [1], high spatio-temporal cine imaging without streaking artifact were successfully obtained from undersampled radial trajectory using FOCUSS(FOCal Underdetermined System Solver) [2] that exploits the sparsity of fact x-f support of cardiac cine. We further showed that radial k-t BLAST/SENSE is an approximation of our radial k-t FOCUSS that is asymptotically optimal from compressed sensing perspective. However, the radial k-t FOCUSS was computationally expensive due to the iterative application of the projection and backprojection steps. Application of gridding algorithm to reduce the computational complexity was, however, not often successful due to the propagation of gridding artifacts during FOCUSS iteration. The main contribution of this paper is to eliminate the necessity of gridding or backprojection/reprojection steps by showing that the non-regular sampling structure can be easily incorporated as a filtering process during k-t FOCUSS iteration in regular grid. This implicit gridding operation incurs only minimial computational overhead. Experimental results demonstrate that the modified radial k-t FOCUSS significantly improves the radial k-t FOCUSS algorithm both in reconstruction speed as well as image quality.

Uygar Sümbül¹, Juan Manuel Santos¹, John Mark Pauly¹

¹Stanford University, Stanford, California , USA

A very fast reconstruction algorithm is presented for non-Cartesian parallel dynamic MRI. The algorithm constrains the images both spatially(via the coil sensitivities) and temporally(via estimating how rapidly individual pixel values change) to achieve alias-free reconstructions. The only major computations are 2 gridding and 2 Fourier operations per each frame. Dynamic whole heart imaging experiments show high temporal resolution of rapidly moving cardiac structures.

Christine S. Law¹, Chunlei Liu¹, Gary H. Glover¹

¹Stanford University, Stanford, California , USA

We propose a sliding window self-calibrated parallel imaging method that reduces thermal noise and provides SNR gain over conventional SENSE reconstruction. This method calculates sensitivity profiles dynamically by using a sliding window approach: combining fully sampled data of adjacent frames in an interleaved acquisition, i.e., sensitivity profiles are updated synchronously with image acquisition.  No spatial smoothing is performed so as to retain thermal noise in sensitivity profiles.  This sliding window technique is especially applicable to acquisition of high spatial-resolution images where thermal noise predominates.

High Field Spectroscopic Quantitation & Methodology

Ronald Ouwerkerk¹

¹Johns Hopkins University School of Medicine, Baltimore, Maryland, USA

The fully adiabatic 1H-MRS LASER sequence was modified for use with lower B1s. Refocusing pulses were optimized for minimum threshold and maximum bandwidths. The LASER sequence with Optimized Pulses (LASEROP) was tested in simulation using Bloch equations, which showed a B1 threshold of about 20 μ]T for 4 kHz BW. On phantoms with body coil excitation and with body coil loads the sequence produced excellent spectra of a citrate-containing phantom at TE 39 ms requiring only 21 μ T peak B1 field. This is particularly useful for 1H-MRS of the prostate where the LASER sequence can provide spectra that are much easier to quantify.

Tom WJ Scheenen¹, Dennis WJ Klomp¹, Pierre-Francois van de Moortele², Gregor Adriany², Arend Heerschap¹

¹Radboud University Nijmegen Medical Centre, Nijmegen, Netherlands; ²University of Minnesota, Minneapolis, Minnesota, USA

The semi-LASER pulse sequence is a combination of conventional non-adiabatic slice-selective excitation and double slice-selective refocusing by two pairs of adiabatic full passage pulses. Combined with an efficient 16 channel head coil the sequence can produce a 3D spin echo localized volume of interest at an echo time of 30 ms. With a small chemical shift displacement error and the relative insensitivity for B1 inhomogeneities, the semi-LASER technique is able to provide useful MR spectra over a large part of the brain at 7T without additional lipid or outer volume suppression.

Yan Li¹, ², Duan Xu¹, Albert P. Chen¹, Daniel B. Vigneron¹, Sarah J. Nelson¹

¹University of California, San Francisco, California , USA

The purpose of this study was to estimate the relaxation times of Cho, Cr and NAA at both 3T and 7T, to evaluate the effects of relaxation times in the long echo time acquisition, to examine the differences in SNR between 3T and 7T and to investigate how the acquisition parameters influence the quality of the spectra. Statistical significance was found for both T1 and T2 values of metabolites between 3T and 7T for all metabolites except for the Cho T1 value. After corrections for relaxation parameters, the SNR ratios between 7T and 3T were on average 1.53. 

Tom WJ Scheenen¹, Pierre-Francois van de Moortele², Gregor Adriany², Andre van der Kouwe³, ⁴, Ricardo Otazo⁵, Stefan Posse⁵

¹Radboud University Nijmegen Medical Centre, Nijmegen, Netherlands; ²University of Minnesota, Minneapolis, Minnesota, USA; ³Athinoula A. Martinos Center for Biomedical Imaging, Charlestown, Massachusetts, USA; ⁴Massachusetts General Hospital, Boston, Massachusetts, USA; ⁵University of New Mexico, Albuquerque, New Mexico, USA

Non-uniform spatial sensitivity, increased chemical shift displacement errors, lipid contamination due to B1-inhomogeneity, and increased RF power deposition are major challenges for metabolite imaging at ultra-high field (7T). In this study at 7T we combine Proton-Echo-Planar-Spectroscopic-Imaging with paired adiabatic refocusing pulses and B1-compensated outer volume suppression (OVS) to enable short TE (20 ms) metabolite mapping. We demonstrate the feasibility of high spatial resolution (0.25 cc voxel size) metabolite mapping in central and peripheral regions of the human brain in a clinically feasible measurement time using a 16-channel line array headcoil. 

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 measure in vivo T1 relaxation times of the neurochemical profiles at 14.1T in rat brain. T1 measurements were accomplished using a progressive saturation technique, which was validated with an adiabatic inversion recovery. The T1 measured at 14.1 T is similar (~10%) to those measured at 9.4 and 11.7T suggesting that for metabolites, T1 increases are of minimal consequence beyond 9.4 Tesla.

Changho Choi¹, Chenguang Zhao¹, Ivan Dimitrov¹, ², Ana Stan¹, Carol Tamminga¹

¹University of Texas Southwestern Medical Center, Dallas, USA; ²Philips Medical Systems, Cleveland, Ohio, USA

Spectrally-selective refocusing has been applied to separate the signals between glutamate (Glu) and glutamine (Gln) in human brain selectively at 7.0 T. The C4-proton resonances of Glu and Gln, which differ by ~0.1 ppm » 30 Hz), was selectively refocused using a 50-ms long Gaussian 180° pulse implemented within PRESS. Ignoring T₁ and T₂ effects, an optimal echo time of 137 ms gave Glu and Gln multiplets at 2.35 and 2.45 ppm with peak amplitude 67% and 53% with respect to 90°-acquisition respectively. Preliminary in vivo results from the human brain are presented. 

Trina Kok¹, Eva-Maria Ratai², ³, Florian Eichler, ³⁴, Elfar Adalsteinsson¹

¹Harvard-MIT Division of Health Sciences and Technology, Cambridge, Massachusetts, USA; ²A.A. Martinos Center for Biomedical Imaging, Cambridge, Massachusetts, USA; ³Harvard Medical School, Boston, Massachusetts, USA; ⁴A.A. Martinos Center for Biomedical Imaging,, Cambridge, Massachusetts, USA

The combination of segmented structural data with low-SNR spectroscopy has been used to provide brain metabolite estimates in separate tissue compartments [1-3]. To correct for comparison across subjects under different scanning conditions, expressing metabolite signals as ratios to the sum of the metabolite signals from Cr and PCR (Cr') is a common technique. We extend compartmental analysis of absolute metabolite measures to ratio measures, and demonstrate its use in a study of NAA+NAAG (NAA') changes in X-linked adrenoleukodystrophy at 7T

Jimin Ren¹, ², Ivan Dimitrov¹, ³, A. Dean Sherry¹, Craig Malloy¹, ⁴

¹University of Texas Southwestern Medical Center, Dallas, USA; ²Advanced Imaging Research Center, Dallas, USA; ³Philips Medical Systems, Dallas, USA; ⁴VA North Texas Health Care System, Dallas, USA

1H MR spectra were obtained from subcutaneous adipose tissue and bone marrow in the tibia in 10 healthy subjects at 7T.  The ultra-high field spectroscopy offers a simple method to obtain detailed information about fatty acid composition, due to well-resolved proton resonances for all major structural fragments including –CH3, -(CH2)n-, -CH2-COO, -CH2-CH2-COO, -CH=CH-, -CH=CH-CH2-CH=CH, -CH2-CH2-CH=CH-,  and in some instances from the methine and methylene protons of glycerol backbone. The composition of marrow and adipose fat were similar, ~24-26% polyunsaturated, 46-49% monounsaturated and 28-29% saturated. The estimated fraction of 16-carbon and 18-carbon fatty acids was about 32% and 68%, respectively.

Dennis Klomp¹, Andreas Bitz², Arend Heerschap¹, Tom Scheenen¹

¹Radboud University Nijmegen, Medical Center, Nijmegen, Netherlands; ²Erwin L. Hahn Institute for Magnetic Resonance Imaging, Essen, Germany

To meet the demand for a better sensitivity in ¹H MRSI of the prostate the magnetic field strength of 7T can be employed. We introduce the endorectal RF coil as a tranceiver, validate its safety for in vivo use and present a new pulse sequence optimzed for 3D MRSI of the human prostate at 7T. Safe use of the coil was validated with SAR calculations, MR thermometry and in vivo temperature monitoring. The pulse sequence was designed and validated with quantum mechanical simulations, phantom and in vivo measurements.

Silvia Mangia¹, Michael Garwood¹, Dennis Sorce¹, Kamil Ugurbil¹, ², Shalom Michaeli¹

¹University of Minnesota, Minneapolis, Minnesota, USA; ²3.  Max Planck Institute for Biological Cybernetics, Tubingen, Germany

The phase modulation introduced by J-evolution complicates the analysis of spectra, and impairs the possibility of estimating relaxation rates of J-coupled spins. The phantom and in vivo results presented in this work demonstrate that rotating frame relaxation measurements (T_1ρ and T_2ρ) performed with specific configurations of adiabatic full passage pulses are not affected by J-evolutions. This notably simplifies the estimation of relaxation rates in the rotating frame also for resonances other than singlets. Since the analysis of the relaxation rates measured in the rotating frame can provide quantitative information on the spin dynamics, our results imply that dynamics of several compounds relevant for brain function and metabolism (as for instance glutamate) can be probed by adiabatic T_1ρ and T_2ρ relaxation measurements.

Audrey Jennifer Chang¹, Sergey Cheshkov¹, Subhendra Sarkar¹, Richard Briggs¹

¹UT Southwestern Medical Center, Dallas, Texas, USA

MR spectroscopy has become a valuable tool in the investigation of the pathologies underlying many neurodegenerative disorders.  The goal of this study is to investigate the reproducibility of metabolite T2 relaxation time measurements in normal subjects at 3T, and to develop a standard protocol for acquisition of T2 relaxation measurements in neurodegenerative disease patients within a clinically feasible scan time (1 hr).  In addition to influencing the ability to detect significant changes in cerebral metabolite concentrations, information about variation in time of T2 relaxation measurements might provide insight into molecular mechanisms of disease progression. 

Mari A. Smith¹, Jason A. Koutcher¹, Kristen L. Zakian¹

¹Memorial Sloan-Kettering Cancer Center, New York, USA

The goal of this study was to implement in vivo detection of lactate by J-difference editing 1H MRS on a 3 Tesla human scanner for potential applications in patients. The current study extends the previous work cited in the literature regarding BASING editing for lactate detection by testing the sequence in the presence of high lipid concentrations in vivo at 3 T. The results demonstrate the ability of the BASING J-difference technique to detect lactate in the presence of strong lipid signals and indicate successful lactate editing in the presence of lipid at 3 T.

Gaurav Verma¹, Saadallah Ramadan², ³, Scott Lipnick¹, Nagarajan Rajakumar¹, M. A. Thomas¹

¹UCLA, Los Angeles, California , USA; ²Brigham & Women's Hospital, Boston, Massachusetts, USA; ³Harvard Medical School, Boston, Massachusetts, USA

The purpose of this study is to develop a turbo spin echo (TSE) based sequence for recording multi-voxel based 2D COSY MR Spectra. Through incorporating an echo train in each TR rather than a single acquisition, which has been routinely used in conventional MR Spectroscopic Imaging (MRSI), the data acquisition was accelerated by a factor of 4 or more, thus reducing the total duration required for recording 4 dimensional (4D) MRSI (2D Spatial, 2D Spectral) to times reasonable for adding this sequence to any clinical MRSI protocol.

Enrique Frias-Martinez¹, Nagarajan Rajakumar¹, S. Ramadan¹, S. Banakar¹, X. Liu¹, A. Singhal¹, Michael Albert Thomas¹

¹UCLA, Los Angeles, California , USA

Using a 3T MRI scanner, spatially resolved two-dimensional (2D) MR spectra including L-COSY and JPRESS, and one-dimensional (1D) PRESS spectra were recorded in the occipital white/gray matter regions of healthy human volunteers. ProFit algorithm was used to quantify the 2D MRS raw data and LC-Model for the 1D signal. Eight healthy volunteers and a white matter phantom have been investigated so far. Preliminary results indicate that COSY-ProFit provides better specificity that JPRESS-ProFit, and both of them provide better specificity and a higher number of detected metabolites than 1D PRESS processed using LC-Model. 

Andrew A. Maudsley¹, Colin Studholme², Varanavasi Govindaraju¹

¹University of Miami, Miami, Florida, USA; ²University of California San Francisco, San Francisco, California , USA

A method for analysis of single-subject metabolite images is presented that uses knowledge of the local tissue distributions and of the tissue-specific metabolite values generated from a group of comparison subjects. By incorporating spatial normalization procedures that convert all metabolite image results to a standardized space, this procedure enables voxel-based analyses without the requirement for accurate spatial alignment on the sulcal level. 

Kevin Wayne Waddell¹, Parham Zanjanipour², Subechhya Pradham¹, James M. Joers¹, Edward Brian Welch³, Peter R. Martin¹, Malcolm J. Avison¹, John C. Gore¹

¹Vanderbilt University, Nashville, USA; ²University of Kentucky, Lexington, USA; ³Vanderbilt University, Nashville, Tennessee, USA

The relationships of N-acetyl aspartate (NAA) and choline-containing compounds in the anterior cingulate and cerebellar vermis of chronic alcoholics have previously been studied using proton spectroscopy. Advances in clinically available high-field scanners have made it possible to measure levels of the major inhibitory neurotransmitter, gamma-aminobutyric acid (GABA), in these areas. We used these methods to establish normative fronto-cerebellar concentrations of metabolites using J-difference spectroscopy. We found that ratios of GABA* with respect to creatine are 140% higher on average than their cerebellar-vermian counterparts. Corresponding ratios of glutamate and NAA were also elevated in the anterior cingulate.

Mary C. Stephenson¹, Kay E. Head¹, Andrew M. Peters¹, Dorothee P. Auer, Peter G. Morris¹

¹University of Nottingham, Nottingham, UK

Activation in the brain increases energy use but, due to the haemodynamic response there is a delay of a few seconds in the supply of additional glucose and oxygen.  This study uses ³¹P spectroscopy to explore the possible role of PCr in meeting energy demand during this first few seconds of activation.  Three seconds after the onset of intense visual stimulation we measured small increases in PCr, decreases in P_i and pH.  These are consistent with the energy shuttle hypothesis but not a buffering role for PCr.

Peng Qu¹, Joseph Gillen¹, Peter B. Barker¹, ²

¹Johns Hopkins University, Baltimore, Maryland, USA; ²Kennedy Krieger Institute, Baltimore, Maryland, USA

Blipped phase-encoding (BPE) has previously been used to accelerate echo-planar spectroscopic imaging sequences. However, it apparently has not been applied to conventional magnetic resonance spectroscopic imaging (MRSI). This abstract describes the implementation of BPE-MRSI on 1.5 and 3.0T systems, and compares the results to those from sensitivity-encoded MRSI.

Borjan Aleksandar Gagoski¹, Michael Hamm², Jonathan Polimeni³, ⁴, Gunnar Krueger⁵, Eva-Maria Ratai³, ⁴, Graham Wiggins³, ⁴, Uwe Boettcher⁶, Joonsung Lee¹, Florian Eichler³, ⁴, Stefan Roell⁷, Elfar Adalsteinsson¹, ⁸

¹Massachusetts Institute of Technology, Cambridge, Massachusetts, USA; ²Siemens Medical Solutions, Charlestown, Massachusetts, USA; ³A. A. Martinos Center for Biomedic Imaging, Charlestown, Massachusetts, USA; ⁴Harvard Medical School, Boston, Massachusetts, USA; ⁵Siemens Medical Solutions, Lausanne, Switzerland; ⁶Siemens Medical Solutions, Erlangen, Germany; ⁷Siemens Medical Solutions, Malvern, Pennsylvania, USA; ⁸Harvard-MIT Division of Health Sciences and Technology, Cambridge, Massachusetts, USA

Coil arrays with large number of receive elements are well suited for the time-efficient CSI acquisitions, as they provide important SNR gain for metabolite detection. In this work we make use of the encoding efficiency of the spiral CSI to quickly encode 3D volumetric brain acquisitions with small voxel sizes. We use a custom-built 32-channel coil array to maximize SNR and fast online gridding reconstruction. We present in-vivo results demonstrating that combining the spiral CSI with large receive coil arrays yields large-volume, high-resolution spectroscopic imaging data.

Shang-yueh Tsai¹, Yi-Ru Lin², Teng-Yi Huang³, Stefan Posse⁴, ⁵, Fa-Hsuan Lin⁶, ⁷

¹Graduate Institute of Biomedical Electronics and Bioinformatics, National Taiwan University, Taipei, Taiwan; ²Dept. of Electronic Engineering, National Taiwan University of Science and Technology, Taiwan; ³Dept. of Electrical Engineering, National Taiwan University of Science and Technology, Taiwan; ⁴Dept. of Psychiatry, University of New Mexico, Albuquerque, New Mexico, USA; ⁵Dept. of Physics and Astronomy, University of New Mexico, Albuquerque, New Mexico, USA; ⁶Department of Radiology, Massachusetts General Hospital, Boston, Massachusetts, USA; ⁷MGH-HMS-MIT Athinoula A. Martinos Center for Biomedical Imaging, Charlestown, Massachusetts, USA

Proton Echo Planar Spectroscopic Imaging (PEPSI) has been developed to reduce the scan time of 2-dimensional MRSI to the order of several minutes. However the spectral quality may still be influenced by motion. Radial acquisition is known to be less sensitive to motion artifacts compared with Cartesian sampling scheme. Here a radial PEPSI (rPEPSI) sequence is implemented and compared with regular PEPSI. Spectra and metabolite maps quantified with LCModel from rPEPSI and PEPSI are comparable. The concentrations and fitting errors are at similar level. In this preliminary report, we demonstrate the feasibility to combine radial acquisition and PEPSI technique.

Hidehiro Watanabe¹, Nobuhiro Takaya¹, Fumiyuki Mitsumori¹

¹National Institute for Environmental Studies, Tsukuba, Japan

A ISIS CT-PRESS sequence was implemented on a 4.7-T whole-body spectrometer for peak detection of glutamate, GABA and glutamine in the human brain with higher sensitivity. Three diagonal peaks of GABA C2H (2.28 ppm), Glu C4H (2.35 ppm) and Gln C4H (2.44 ppm) were resolved on CT-PRESS spectra of a phantom containing a brain mixture. Signal to noise ratio of Glu C4H on a CT-PRESS spectrum was 2.24 times higher than that obtained by localized CT-COSY. In volunteer studies, these three peaks were resolved on CT-PRESS spectra of parieto-occipital regions with sensitivity improvement by a factor of 1.7.

Wolfgang Dreher¹, ², Dmitry Ebel¹, ², Dieter Leibfritz¹, ²

¹University of Bremen, Bremen, Germany; ²Center for Advanced Imaging (CAI), Bremen, Germany

Different ways to optimize CT-PRESS measurements and maximize the SNR on a 3 T human head scanner are examined. Short tc values corresponding to the mean echo time are realized by using a short TE PRESS module and an additional rectangular 180 . Strong spoiler gradients make phase cycling obsolete and enable weighted averaging in t1. A local RF transmitter adjustment accounts for spatial B1 inhomogeneities and avoids severe signal losses by deviation from the nominal flip angles. Truncation artifacts can be reduced below the noise level for in vivo measurements by using the S-TRAF function for apodization in t1.

Yi-Yu Shih¹, ², Dominik von Elverfeldt², Martin Buechert², Hsiao-Wen Chung¹, Juergen Hennig²

¹National Taiwan University, Taipei, Taiwan; ²University Hospital Freiburg, Freiburg, Germany

The Asc concentration was successfully detected in 71 of 76 in vivo human brain spectra using PRESS sequence on 3T MR-scanner in our study. To further evaluate the reliability in quantification, three of the spectra with different line-widths were chosen to add artificial Asc spectrum (virtual titration) with different concentrations and SNR levels, followed by analysis via LCModel. The results show the good linearity between added and estimated concentrations of Asc. Other neighboring metabolites (mI and Glx) are only slightly affected (within 10% concentration change) if the line-width was less than 0.043 ppm. Therefore, it is feasible to detect Asc using standard clinical MRS acquisition in combination with LCModel.

Andrew A. Maudsley¹, Claudia Domenig¹

¹University of Miami, Miami, Florida, USA

A signal normalization method for MRSI in the brain is described that uses a tissue water MRSI, obtained using an interleaved measurement, as an internal reference. When combined with spatial registration functions, this procedure enables voxel-based comparisons between repeated studies in the same subject as well as comparisons between subjects. Results for analysis of individual voxel results for intra-subject and inter-subject comparisons are presented, as well as for tissue regression analyses over different brain volumes.

Dmitry Ebel¹, ², Christian Schuster¹, ², Wolfgang Dreher¹, ², Dieter Leibfritz¹, ²

¹University of Bremen, Bremen, Germany; ²Center for Advanced Imaging (CAI), Bremen, Germany

Spectroscopic Missing-Pulse SSFP (spMP-SSFP) is a powerful method for fast 3D spectroscopic imaging. However, further improvements are hampered by specific absorption rate (SAR) and timing limitations. In this study, optimized RF and gradient waveforms were applied within spMP-SSFP to improve saturation of extracranial lipid signals. The number of applicable OVS slices was increased when applied on the human brain at 3 Tesla using VERSE-like 90  pulses, calculated by genetic algorithms, with reduced SAR and improved off-resonance behavior. This approach enables better defined regions of interest and will allow to implement spMP-SSFP at higher magnetic fields.

Jullie Pan¹, Xenophon Papadametris¹, Susan Spencer, Dennis Spencer¹, Hoby Hetherington¹

¹Yale University, New Haven, Connecticut, USA

We have developed and implemented methods to identify and reconstruct spectroscopic imaging pixels using either a common anatomical atlas (MNI standard brain) and other imaging modalities (e.g. PET,CT and SPECT) to define the target locations. Co-registration driven selection and registration provides for precise positioning of the target voxels and the elimination of user bias in voxel selection. We have applied these methods to investigate: 1) the extent of asymmetric subcortical metabolic abnormalities in patients with temporal lobe epilepsy and 2) the relationship between regions with abnormal intracranial EEG (electrodes identified by CT) and decreased NAA in patients with neocortical epilepsy.

Changho Choi¹, Nicholas J. Coupland², Peter Seres², Chenguang Zhao¹, Sanjay Kalra², Philip G. Tibbo²

¹University of Texas Southwestern Medical Center, Dallas, USA; ²University of Alberta, Edmonton, Canada

PRESS echo time dependence of brain glutamate (Glu) selectivity was investigated, with density-matrix simulation, for TE₁ and TE₂ between 30 and 80 ms, incorporating the neighboring resonances of glutamine, N-acetylaspartate and glutathione. The numerical calculation indicated that Glu signal intensity and resolution are both maximized at {TE₁, TE₂} = {35, 75} ms, its peak amplitude being 73% with respect to 90°-acquisition. An in vivo brain spectrum from these subecho times is presented, in comparison with a spectrum from the previously-proposed echo time {40, 40} ms.

Esin Ozturk-Isik¹, Wei Bian¹, Ilwoo Park¹, ², Albert P. Chen¹, Jason C. Crane¹, Daniel B. Vigneron¹, ², Susan M. Chang¹, Sarah J. Nelson¹, ²

¹University of California at San Francisco, San Francisco, California , USA; ²University of California at San Francisco and Berkeley, San Francisco, California , USA

Lactate is an important metabolite that has been detected in high-grade brain tumors, ischemia and stroke cases. Lactate detection requires a special editing scheme like the J-difference technique using dual BASING pulses due to coresonant lipid peaks. The J-difference based lactate detection requires two cycles of spectral data acquisition which results in doubled scan time. In this study, we propose to use a fast data acquisition method, the ellipsoidal SENSE technique, with BASING pulses to acquire 3D lactate edited MRSI of brain tumor patients with the clinically feasible scan time of 9 min at 3T.

Molecular Imaging Agents:  Applications & Detection

M. Meser Ali¹, M. D. Pagel¹

¹Case Western Reserve University, Cleveland, Ohio, USA

PARACEST MRI contrast agents suffer from poor temporal resolution and limited detection sensitivity.  Agents with longer in vivo retention times may compensate for poor temporal resolution, and agents that accumulate at high concentrations within in vivo tissues may compensate for limited sensitivity.  The PARACEST agent EuDOTA-OBnS2-Gly2-COOH has both of these advantages relative to EuDOTAMGly, due to the addition of two o-benzyl moieties.  This was demonstrated by the detection of the PARACEST effect of EuDOTA-OBnS2-Gly2-COOH in the in the inferior vena cava, kidney, and liver of a normal mouse.

                  1646.     Heteronuclear Relaxivity of Commercial Gadolinium Contrast Agents

Cristina Gabellieri¹, Geoffrey S. Payne¹, Martin O. Leach¹, Thomas R. Eykyn¹

¹The Institute of Cancer Research, Sutton, UK

Development of hyperpolarisation techniques have led to increasing interest in non-proton MR for metabolic and molecular imaging. Potential biomarkers with nuclei with long relaxation times are needed to effectively use the high non-equilibrium polarization attained. Here we report relaxivities of two Gd-based contrast agents on ¹³C in glycine and ¹⁵N in choline in aqueous solutions. We also confirm that extracellular contrast agents exert little influence on the relaxation properties of intracellular metabolites by looking at cell suspension by ³¹P NMR of . This may be important in the future use of hyperpolarized tracers to relax the extracellular parent signal and better observe the intracellular daughter compounds. 

                  1647.     Improved Molecular Imaging of Sparse Neovascular Biomarkers with a Novel Lipophilic Gd-DOTA Chelate
                                on Targeted Nanoparticles

Patrick M. Winter¹, Shelton D. Caruthers¹, Anne H. Schmieder¹, Todd A. Williams¹, Dennis Riley, William McGhee, Samuel A. Wickline¹, Gregory M. Lanza¹

¹Washington University, St. Louis, Missouri, USA

The purpose of this study was to compare the relaxivity and in vivo signal enhancement of targeted nanoparticles formulated with two different paramagnetic chelates, Gd-DTPA-BOA and Gd-DOTA-Amide-PE. Nanoparticles were targeted to the α  _ν β ₃-integrin for molecular imaging of angiogenesis induced by tumor growth. The Gd-DOTA-Amide-PE nanoparticles had 50% higher relaxivity and produced 74% higher enhancement in a rabbit tumor model. These results suggest that paramagnetic chelates with improved relaxivity may facilitate increased sensitivity with molecular imaging agents.

                  1648.     In Vivo Imaging of  α _vβ ₃ - Targeted Micellar Nanoparticles in an NSCLC Tumor Xenograft Model by
                                 Off-Resonance Saturation Method  [Not Available]

Chase W. Kessinger¹, Chalermchai Khemtong¹, Jimin Ren¹, Erik A. Bey¹, David A. Boothman¹, A. Dean Sherry¹, Jinming Gao¹

¹University of Texas Southwestern Medical Center at Dallas, Dallas, Texas, USA

Here we describe the use of off-resonance saturation (ORS) method to image α _v β ₃ - targeted superparamagnetic polymeric micelles (SPPM) for the molecular imaging of cancer.  The ORS contrast is achieved by water signal attenuation with an RF pulse set at an off-resonance position to bulk water.  The ORS method was demonstrated in a human NSCLC tumor xenograft model in nude mice in vivo.  The ORS contrast can be turned “on” and “off”, to produce an accurate contrast image of superparamagnetic nanoparticles targeting to tumor neovasculature.

                  1649.     Intracellular Uptake of Targeted Paramagnetic Contrast Agent Leads to Quenching of T₁ Contrast

Maarten B. Kok¹, Sjoerd Hak¹, Willem J.M Mulder², Daisy van der Schaft¹, Gustav J. Strijkers¹, Klaas Nicolay¹

¹Eindhoven University of Technology, Eindhoven, Netherlands; ²Mount Sinai School of Medicine, New York, USA

This study shows that the effective relaxation enhancement introduced by an internalized paramagnetic contrast agent depends crucially on its intracellular localization. The effective T1 relaxivity of internalized áíâ3-targeted liposomes was much lower than that of non-targeted liposomes. This finding has important implications for the interpretation of target specific MRI in vivo, as significant local uptake of contrast agent by target tissue might go undetected.

                  1650.     Novel Method to Characterize Paramagnetic Contrast Agents Association to 2-Dimensional Surfaces

H.M.H.F. Sanders¹, H.P. Huinink¹, S.J.F. Erich¹, O.C.G. Adan¹, M. Merkx¹, G. J. Strijkers¹, K. Nicolay¹

¹Eindhoven University of Technology, Eindhoven, Netherlands

Receptor mapping of cells lining blood vessels is a major application of molecular MR imaging using targeted contrast agents, for example for atherosclerosis or angiogenesis characterization. Here we show that very high-resolution MRI depth profiling by using a saturation recovery pulse sequence, combined with mathematical modelling, is capable of quantifying the surface water longitudinal relaxation rate induced by a surface-bound contrast agent. This technique is helpful in studying the behavior of targeted contrast agents and will yield valuable insights for the optimization of the MRI-based detection of target-associated contrast agents.

                  1651.     Imaging Targeted Delivery of Liposomes to Tumor Vasculature

Maria Mikhaylova¹, Amin Hajitou², Yoshinori Kato¹, Saraswati Sukumar, Dmitri Artemov¹, Marie-France Penet¹, Wadih Arap², Renata Pasqualini², Zaver M. Bhujwalla¹

¹JHU ICMIC Program, The Johns Hopkins University, School of Medicine, Baltimore, Maryland, USA; ²The University of Texas, MD Anderson Cancer Center, Houston, Texas, USA

We recently identified a novel peptide that binds specifically to breast cancer vasculature.  Here we have examined the potential of this peptide for targeting multifunctional liposomes to tumor vasculature in an MDA-MB-231 human breast cancer model.  In vivo magnetic resonance  imaging revealed specific accumulation of targeted liposomes in tumor vasculature starting at 3h following intravenous injection compared to liposomes with a control peptide.  This finding was supported by optical imaging and immunofluorescent localization.  These data suggest that the newly identified peptide may provide a means to deliver therapeutic and analytical cargo to breast cancers under image-guidance.

                  1652.     Targeted Herceptin-Dextran Nanoparticles for Noninvasive Imaging of Her2/neu Receptor by MRI

Yun-Ming Wang¹, Ting-Jung Chen¹, Chiao-Yun Chen², Gin-Chung Liu²

¹Kaohsiung Medical University, Kaohsiung, Taiwan; ²Kaohsiung Medical University Hospital, Kaohsiung, Taiwan

The fabrication and characterization of dextran-coated magnetic nanoparticles CLIO-Herceptin as nanoprobes for MR imaging to target specific receptors were reported. From the results, we found that the CLIO-Herceptin had well-dispersed in different pH value solutions, low protein absorption from medium, no hysteresis, high saturation magnetization and low cytotoxicity of different cells. Moreover, CLIO-Herceptin had ability to target the SKBR-3, BT-474, MDA-MB-231 and MCF-7 cells proved by in vitro and in vivo MR imaging studies.

                  1653.     Molecular MRI of the Estrogen Receptor in Human Breast Cancer Cells

Adi Pais¹, Chidambaram Gunanathan¹, David Milstein¹, Hadassa Degani¹

¹Weizmann Institute of Science, Rehovot, Israel

The estrogen receptor (ER) serves as an important prognostic factor and a marker for predicting response to hormonal therapy. We present two novel contrast agents targeted to ER, composed of a gadolinuimpyridiniumtetraacetate conjugated to estradiol (EPTA-Gd), or to tamoxifen (TPTA-Gd) both with a micromolar binding affinity to ER. These agents exhibited agonistic effect on cell proliferation, disparity in inducing ER degradation and high relaxivities in solution. Concentration dependent studies of T1 and T2 relaxation in viable ER+ cells and in the same cells with null ER indicated a specific binding to ER which reflected the level of this receptor.

Kejia Cai¹, Todd A. Williams¹, Shelton D. Caruthers¹, ², Gregory M. Lanza¹, Samuel A. Wickline¹, Patrick M. Winter¹

¹Washington University, Saint Louis, Missouri, USA; ²Philips Medical Systems, Andover, Massachusetts, USA

The metabolic syndrome is highly associated with cardiovascular disease. JCR:LA-cp diabetic rats were treated with benflurorex for 15 weeks to study the effects on cardiovascular disease. ávâ3-integrin targeted gadolinium nanoparticles were used for molecular imaging of angiogenesis, a hallmark of atherosclerosis. Benfluorex treatment decreased MR signal enhancement by about 60%, indicating suppression of angiogenesis, and possibly inhibition of atherosclerosis. These results suggest that molecular imaging with targeted nanoparticles can detect cardiovascular disease associated with metabolic syndrome and monitor the treatment response.

                  1655.     Evaluation of Sensitivity Increase by T1 and T2 Contrast Agents in 19F MRI of PF15C

Thomas Christian Basse-Luesebrink¹, Thomas Kampf², Christian Herbert Ziener², Gert Klug², Wolfgang Rudolf Bauer², Peter Michael Jakob², Daniel Haddad¹

¹Research Center Magnetic Resonance Bavaria, Wuerzburg, Germany; ²University of Wuerzburg, Wuerzburg, Germany

In this study the influence of T1 and T2 contrast agents in 19F MRI on SNR per measurement time is investigated.  Therfore Turbo Spin Echo (RARE) and Fast Gradient Echo (FLASH) sequences are applied to different cases. As a result this study suggests that the usage of T1 contrast agents is not generally preferable.

                  1656.     On the Optimal Field Strength for Detection of Targeted Gd-Based Contrast Agents in Molecular
                                MR Imaging

Olivier Girard¹, Philippe Robert², Luc Darrasse¹

¹Univ Paris-Sud, CNRS, Orsay, France; ²Guerbet Recherche, Roissy Charles-De-Gaulle, France

Molecular MR Imaging using targeted Contrast Agents (tCA) is a promising tool for early diagnosis purpose. For macromolecular Gd-based CA (e.g. a tCA bound to its target) the T1-relaxivity tends to decrease rapidly above 20-30 MHz raising the question of their efficiency at high field. We investigate theoretically and experimentally the Contrast-to-Noise Ratio (CNR) obtained with a paramagnetic tCA as a function of the field strength, accounting for NMRD profiles derived from experimental data. Results indicate that  Molecular MR Imaging with Gd-based tCA do not benefit from high field strength and should be performed  around 1-1.5 T to optimise the CNR and differentiate bound and free pools of tCA.

                 1657.     Early Stage Investigations of USPIO-Induced Signal Changes After Focal Cerebral Ischemia in Mice

Virginie Desestret¹, Jean-Christophe Brisset¹, Emilie Devillard¹, Samir Moucharrafie¹, Serge Nataf², J. Honnorat², Norbert Nighoghossian¹, Yves Berthezene¹, Marlène Wiart¹

¹Creatis-LRMN (UMR CNRS 5220, Inserm U630), Bron, France; ²Inserm U842, Bron, France

The interpretation of USPIO-induced MR signal changes at the early stages of focal cerebral ischemia remains controversial. Cerebral infarctions were induced in 40 mice using permanent middle cerebral artery occlusion (pMCAO) by electrocoagulation. Ferumoxtran-10 USPIO (Guerbet, France) were injected i.v. 5h post-injury. On gradient echo T1- weighted imaging, four areas of signal changes after USPIO injection were observed at 6 and 24h post ischemia. On histology, iron staining was mostly associated to the vascular and the cerebrospinal fluid compartments. These results strongly suggest that we must consider time-window dependent results interpretation of USPIO-related signal changes in experimental stroke models.

Perfusion & Permeability Methodology

                  1704.     Is It Possible to Measure Water Exchange Using Conventional DCE-MRI?

David L. Buckley¹, Lucy E. Kershaw¹, Greg J. Stanisz², ³

¹University of Manchester, Manchester, UK; ²University of Toronto, Toronto, Canada; ³Sunnybrook Health Sciences Centre, Toronto, Canada

The importance of cellular-interstitial (transcytolemmal) water exchange (WX) in the measurement of contrast agent concentration using MRI is unclear. Early studies suggested that its influence could be avoided by appropriate choice of imaging sequence. More recent papers stress the importance of measuring WX (e.g. using a “shutter-speed” model). We studied human muscle and analyzed DCE-MRI data with a series of models encompassing the full range of WX scenarios. We concluded that the effect on our data was small and that conventional DCE-MRI is unsuitable for WX measurement.

                  1705.     Impact of AIF Errors on DCE-MRI Pharmacokinetic Parameters: Comparison of a High Temporal
                                 Resolution AIF and a Biexponential Description

Hai-Ling Margaret Cheng¹, ²

¹The Hospital for Sick Children, Toronto, Canada; ²University of Toronto, Toronto, Canada

A rapidly sampled arterial input function (AIF) is known to be essential for accurate DCE-MRI quantification. However, AIF measurement is subject to errors, particularly in the initial bolus phase, and it is unknown how these affect pharmacokinetic parameter accuracy. Simulations are performed to investigate the impact of AIF errors and to compare with a more slowly but carefully acquired biexponential AIF. Results show that a bolus amplitude error in a rapidly sampled AIF introduces significant error in the transfer constant and blood volume. When the bolus amplitude error exceeds 25%, the biexponential AIF provides more accurate parameter estimates.

                  1706.     Sampling Frequency Dependent Identifiability of Pharmacokinetic Parameters for Small and Large
                                 Molecular Contrast Agents

Karolien Jaspers¹, ², Mark J. Post², Tim Leiner¹, ², Walter H. Backes¹, ²

¹Maastricht University Hospital, Maastricht, Netherlands; ²Maastricht University, Maastricht, Netherlands

The influence of the sampling rate and SNR on the identifiability of pharmacokinetic parameters, as defined in the Kety model, was investigated using Monte Carlo simulations, and the implications on designing a DCE-MRI protocol were discussed. In vivo measurements were used for construction of the arterial input function and noise level determinationThe precision and accuracy in parameter estimation at the lower frequencies was better for large molecular than for small molecular contrast agents This would allow lower sampling rates for large molecular contrast agents, and concomitant increase of spatial resolution, provided that sufficient SNR is obtained.

                  1707.     IntraVoxel Partially Coherent Motion (IVPCM) Technique: Application on Skeletal Muscle Microvasculature

Dimitrios C. Karampinos¹, ², Kevin F. King², Bradley P. Sutton¹, John G. Georgiadis¹

¹University of Illinois at Urbana-Champaign, Urbana, Illinois, USA; ²General Electric Healthcare, Waukesha, Wisconsin, USA

A reformulation of the IntraVoxel Incoherent Motion (IVIM) technique, based on the low b-value diffusion-weighted imaging regime, is proposed to characterize microcirculation of tissues perfused with partially coherent blood flow. The new methodology, termed IntraVoxel Partially Coherent Motion (IVPCM) technique, is suitable for tissues with oriented microvasculature and is currently applied to analyze certain geometrical parameters of the microvasculature of human calf muscle at rest. The proposed methodology has the unique capability of characterizing the anisotropy of microcirculation in vivo, in a manner analogous to the capability of high b-value diffusion to characterize the anisotropy of neural tracts.

                  1708.     The Effect of Blood Inflow and B₁ Inhomogeneity on the Arterial Input Function in 3-D DCE-MRI

Caleb Roberts¹, Ross A. Little¹, Sha Zhao¹, Yvonne Watson¹, David L. Buckley¹, Geoff J. Parker¹

¹The University of Manchester, Manchester, UK

An accurate individual measurement of the arterial input function (AIF) is required to achieve reliable estimates of microvascular characteristics in 3D dynamic contrast-enhanced MRI studies. Specific errors due to blood inflow and B₁ inhomogeneity mean that the AIF measurement is challenging and location choice within the imaging slab is crucial. Using a combination of computer simulations, a flow phantom (including a “gold-standard” AIF), B₁ mapping and clinical in vivo measurements we quantify these errors and demonstrate that accurate measurements of the AIF are achievable under certain conditions.

                  1709.     Portal Delay Estimation from DCE-MRI Liver Tissue Data : Feasibility and Effect on Vascular Parameter
                                Estimates

Matthew R. Orton¹, James A. d'Arcy¹, David J. Collins¹, David Atkinson², David J. Hawkes², Martin O. Leach¹

¹Institute of Cancer Research, Sutton, UK; ²Univeristy College London, London, UK

Using a fixed input function for kinetic modelling of DCE-MRI data is a simple and robust technique, and in many cases gives more repeatable estimates than using a measured input function.  For DCE-MRI liver data a more complex dual input function is needed, and accurate determination of the delay between the arterial and portal components is critical for obtaining accurate vascular parameter estimates.  This abstract presents a technique for estimating the portal delay using only tissue data, and assesses the impact of this approach on estimates of the arterial-portal partition term (also known as the Hepatic Portal Index).

Guido Correia Carreira¹, Dirk Beyersdorff¹, Lutz Lüdemann¹

¹Charité, Berlin, Germany

Aim of this work is to assess systematically possible over- and underestimation of vascular permeability, when using compartment models to fit DCE-MRI data of normal and tumor tissue in prostate. To this end, numerical diffusion simulations for several tissue models, describing the CM distribution process at the microscopic scale, are compared with a standard PS-limited (low permeability) two-compartment model. The results show a systematic underestimation of vessel permeabilty by the two compartment model of at least 3% to 7% up to 20% depending on the diffusion coefficient of tissue.

                  1711.     The Effect of RF Field Non-Uniformity on Tracer Quantification in DCE MRI of the Pelvis at 3 T

Cornelis A.T. van den Berg¹, Reto Treier², Jan G. Korporaal¹, Greetje Groenendaal¹, Fredy Visser¹, Marco van Vulpen¹, Uulke A. van der Heide¹

¹University Medical Center Utrecht, Utrecht, Netherlands; ²University of Zurich and ETHZ, Zurich, Switzerland

In DCE MRI of the pelvis the variable flip angle method is popular method to measure T1 changes allowing quantification of the bolus passage. In this study we investigated how flip angle deviations corrupt the tracer concentration quantification at 3 T. Using a B1 mapping method we found that RF field effects result in a highly inhomogeneous flip angle pattern.  Flip angles in the prostate deviate 10 to 15%, flip angles in the right Iliac Externa can be 30  to 40 % lower. This will result in a large overestimation of the AIF tracer concentration as demonstrated by simulations and measurements.

                  1712.     WITHDRAWN

Hyperpolarized ¹³C & Other Nuclei

Simon Hu¹, ², Albert P. Chen¹, Matthew L. Zierhut¹, ², Robert Bok¹, ², Mark Van Criekinge¹, Yi-Fen Yen³, Marie A. Schroeder⁴, Ralph E. Hurd³, Sarah J. Nelson¹, ², John Kurhanewicz¹, ², Daniel B. Vigneron¹, ²

¹University of California at San Francisco, San Francisco, California , USA; ²UCSF & UCB Joint Graduate Group in Bioengineering, San Francisco, California , USA; ³GE Healthcare, Menlo Park, California , USA; ⁴University of Oxford, Oxford, UK

Previous studies have demonstrated the ability to detect uptake and metabolism of hyperpolarized ¹³C  pyruvate through non-localized dynamic MRS spectra in normal rats. This study was designed to obtain and characterize hyperpolarized ¹³C dynamic spectra localized to the liver and to investigate differences between non-fasted and fasted rats. Following fasting, we observed significantly altered ¹³C-alanine levels in the liver, which agrees with prior biochemical studies showing elevated levels of alanine aminotransferase during fasting. This study demonstrated that hyperpolarized ¹³C MR can detect significant changes in liver metabolic states in vivo and may be valuable for liver disease state studies.

                  1746.     Pulse Sequence for Dynamic Volumetric Imaging of Hyperpolarized Metabolic Products

Charles H. Cunningham¹, Albert P. Chen², Michael Lustig³, Janine Lupo², Duan Xu², John Kurhanewicz², Ralph E. Hurd⁴, John M. Pauly³, Sarah J. Nelson², Daniel B. Vigneron²

¹Sunnybrook Health Sciences Centre, Toronto, Canada; ²UCSF, San Francisco, California , USA; ³Stanford University, Stanford, California , USA; ⁴ASL, GE Healthcare, Menlo Park, California , USA

A spectral-spatial excitation pulse was designed to excite a single line of the spectrum resulting from metabolism of hyperpolarized carbon-13.  The pulse was implemented along with an echo-planar readout trajectory to give volumetric coverage every 3.5 seconds.  The pulse sequence was applied to measure the different lactate dynamics in different tissues in a normal rat model and a mouse model of prostate cancer.  The results suggest that significantly different dynamic curves can be observed in tumour vs. non-cancerous tissue.

                  1747.     Apparent T₂ of ¹³C-Labeled Metabolites In Vivo

Yi-Fen Yen¹, Patrick Le Roux², Robert Bok³, Jim Tropp¹, Albert Chen³, Vickie Zhang³, Matthew Zierhut³, ⁴, Mark Albers³, ⁴, Ilwoo Park³, ⁴, Sarah Nelson⁵, Dan Vigneron³, John Kurhanewicz³, Ralph Hurd¹

¹GE Healthcare, Menlo Park, California , USA; ²GE Healthcare, France; ³University of California, San Francisco, California , USA; ⁴Berkeley, California , USA; ⁵UCSF, San Francisco, California , USA

Apparent T2 relaxation time of 13C-labeled metabolites was measured for the first time in animals, following an injection of hyperpolarized 13C-pyruvate solution. Pyruvate and its metabolic products such as lactate, alanine and bicarbonate, were observed in the T2 measurements. A CPMG sequence was employed to acquire spectra at every spin echo for total of 8-10 seconds. The T2-decay curves were best described by multiple T2 components. The in vivo T2 results presented here are valuable for future sequence designs in13C metabolic imaging applications. The T2 information may also have diagnostic value as T2 may vary between healthy and diseased tissues.

                  1748.     First Studies with Hyperpolarized [2-¹³C]pyruvate in the Rat Brain

Isabelle Iltis¹, Dinesh Kumar Deelchand¹, Malgorzata Marjanska¹, Christopher Nelson¹, Kamil Ugurbil¹, Pierre-Gilles Henry¹

¹Center for Magnetic Resonance Research, Minneapolis, Minnesota, USA

In this work, we successfully hyperpolarized [2-¹³C]pyruvate and measured hyperpolarized ¹³C signals in the rat brain in vivo following i.v. injection of the molecule. Resonances from [2-¹³C]pyruvate, [2-¹³C]pyruvate hydrate and [2-¹³C]lactate, but not from TCA cycle intermediates, were observed. The T₁ relaxation time of [2-¹³C]pyruvate was on the order of 15 s in vivo.

                  1749.     Combined Reconstruction of Rosette Sampled Data for Hyperpolarised ¹³C Metabolic Imaging

Rolf F. Schulte¹, Florian Wiesinger¹, Ileana Hancu²

¹GE Global Research, Munich, Germany; ²GE Global Research, Niskayuna, New York, USA

Metabolic imaging with hyperpolarised ¹³C requires rapid acquisition sequences in order to encode both spectrally and spatially with a high resolution. The rosette trajectory is one of the most efficient acquisition sequences with a high spectral selectivity, however at the cost of a high sensitivity towards B₀ inhomogeneities. Including as much physical information as possible improves the robustness considerably. In this work, we combine coil sensitivities, B₀ correction and chemical shifts in one reconstruction. The encoding model is extended by this information and solved with conjugate gradient iterative reconstruction. Simulations, in-vivo and in-vitro scans show promising results.

                  1750.     Parallel Imaging with GRAPPA CSI for Hyperpolarised ¹³C Metabolic Imaging

Rolf F. Schulte¹, Arjun Arunachalam², Kenneth M. Fish³, David B. Whitt³, Florian Wiesinger¹, James Tropp⁴, Ileana Hancu²

¹GE Global Research, Munich, Germany; ²GE Global Research, Niskayuna, New York, USA; ³GE Global Research, Niskayuna, USA; ⁴GE Healthcare, Menlo Park, California , USA

Metabolic imaging of hyperpolarised ¹³C₁-pyruvate was combined with parallel image CSI acquisition and GRAPPA reconstruction. A custom-built, four-channel rat-coil was used for signal acquisition after injecting hyperpolarised ¹³C₁-pyruvate in a rat. Data was acquired with a fully-sampled FID CSI sequence with a spatial and spectral resolution of 24x24 and 256, respectively. The spectral dimension was reconstructed by chemical-shift modelling (matrix inversion) to the three resonances (pyruvate, alanine/pyruvate-ester and lactate). One GRAPPA kernel was fitted from the k-space centre for the three resonances The data was artificially undersampled and then GRAPPA interpolated to yield the same metabolic maps as from the fully sampled data.

                  1751.     Sensitivity Mapping for Parallel Imaging of Hyperpolarized ¹³C Compounds

Daniel Blezek¹, ², Arjun Arunachalam¹, David Whitt¹, Kenneth Fish¹, Ileana Hancu¹

¹GE Research, Niskayuna, New York, USA; ²Mayo Clinic, Rochester, Minnesota, USA

Coil sensitivity profiles for hyperpolarized ¹³C compounds may be acquired in phantoms prior to actual scanning.  Low coil loading permits profiles to be measured once for rigid geometry coils.  We have developed a method to register pre-acquired sensitivity profiles to ¹³C-pyruvate acquisition using fudical spheres rigidly affixed to the coil.  A ¹³C-acetate and ²³Na phantom were constructed and imaged.  Using a custom 4 channel rat-sized coil, we compare ¹³C-pyruvate rat images reconstructed from fully sampled signals to SENSE recon using self-calibrated profiles(R=1.5), pre-acquired ¹³C-acetate(R=2) and ²³Na profiles(R=2).  Parallel imaging results compare favorably with full reconstruction.

                  1752.     Accelerated Imaging of Hyperpolarized Carbon-13 Compounds

Arjun Arunachalam¹, David Whitt¹, Kenneth Fish¹, Randy Giaquinto¹, Ileana Hancu¹

¹GE Global Research Center, Niskayuna, New York, USA

The long scan time characteristic of traditional chemical shift imaging sequences (CSI) is a major limitation for metabolic imaging of hyperpolarized compounds. In this work, the ability to accelerate the spatial encoding process during a CSI scan of hyperpolarized compounds is demonstrated in vivo through parallel imaging. A hardware set up designed to acquire 13C signal data from multiple receivers simultaneously is presented and the ability to perform parallel imaging is demonstrated in vivo. The CSI data from the accelerated scans is reconstructed using self-calibrated SENSE, by using coil sensitivity maps obtained from the central region of k-space.

                  1753.     Hyperpolarized ¹³C MRI with a Triple-Frequency RF Coil

S. James Wood¹, Brian K. Rutt¹, Joseph Piel², David B. Whitt², Kenneth M. Fish², W. Thomas Dixon², Ileana Hancu²

¹Robarts Research Institute, the University of Western Ontario,, London, Canada; ²GE Global Research Center, Niskayuna, New York, USA

Imaging ¹³C hyperpolarized compounds requires new MR hardware capabilities. ¹H imaging is required for anatomical localization and shimming. Accurate flip angle calibration for ¹³C scans can be performed using natural abundance ²³Na signal. We demonstrate here a triple-frequency RF resonant structure, which mitigates signal loss of previous designs and eliminates the need for multiple coils. In vivo, hyperpolarized ¹³C experiments are demonstrated where images are acquired sequentially at ¹H, ²³Na, and ¹³C frequencies without moving the coil or specimen.  Proton, sodium, lactate, and pyruvate images are acquired; as expected, high ²³Na and ¹³C signals are visible over the kidneys.

                  1754.     Time-Resolved Metabolic Imaging in the Rat After Injection of Hyperpolarized ¹³C-1-Pyruvate at 3 Tesla

Dirk Mayer¹, ², Yi-Fen Yen³, Yakir S. Levin¹, James Tropp³, Adolf Pfefferbaum, ¹², Ralph E. Hurd³, Daniel M. Spielman¹

¹Stanford University, Stanford, California , USA; ²SRI International, Menlo Park, California , USA; ³GE Healthcare, California , USA

This work reports on the development of rapid time-resolved ¹³C spiral chemical shift imaging for the assessment of localized metabolic parameters. The sequence was tested at 3 Tesla on a phantom to measure the longitudinal relaxation constant of hyperpolarized ¹³C-1-pyruvate. In the in vivo application, the time course of pyruvate, lactate, and alanine were measured in the kidneys and the liver of a rat after tail vein injection 1 ml of hyperpolarized ¹³C-1-pyruvate.

                  1755.     Ultra-Fast in Vivo Metabolic Imaging in the Rat After Injection of Hyperpolarized ¹³C-1-Pyruvate at 3 Tesla

Dirk Mayer¹, ², Yi-Fen Yen³, Yakir S. Levin¹, James Tropp³, Adolf Pfefferbaum, ¹², Ralph E. Hurd³, Daniel M. Spielman¹

¹Stanford University, Stanford, California , USA; ²SRI International, Menlo Park, California , USA; ³GE Healthcare, California , USA

A fast spiral chemical shift imaging sequence has been developed for application in hyperpolarized ¹³C imaging. The sequence exploits sparse spectra and prior knowledge of resonance frequencies to reduce the measurement time by undersampling the data in the spectral domain. Multiple data sets having only frequency components within a certain bandwidth are reconstructed “in-focus” while others are severely blurred (“spectral tomosynthesis”). The sequence was applied to sub-second metabolic imaging of the rat in vivo after injection of hyperpolarized ¹³C-1-pyruvate.

                  1756.     Detection and Imaging of Hyperpolarized 6-Lithium in the Rat Brain in Vivo

Ruud Bernardus van Heeswijk¹, Kai Uffmann¹, Arnaud Comment¹, Fiodar Kurdzesau¹, ², Chiara Perazzolo¹, Cristina Cudalbu¹, Sami Jannin¹, J.A. Konter³, Patrick Hautle², Ben van den Brandt², Jacques J. van der Klink¹, Gil Navon⁴, Rolf Gruetter¹, ⁵

¹Ecole Polytechnique Fédérale de Lausanne, Lausanne, Switzerland; ²Paul Scherrer Institute, Villigen, Switzerland; ³Paul  Scherrer Institute, Villigen, Switzerland; ⁴Tel Aviv University, Tel Aviv, Israel; ⁵Universities of Lausanne and Geneva, Switzerland

This study focused on the feasibility and characterization of the usage of in vivo hyperpolarized lithium-6. Therefore, hyperpolarized lithium-6 chloride was infused in the rat femoral vein. Pulse-acquire as well as fast spectroscopic imaging were used to study the signal decay and spatial distribution of the lithium in the rat head. The relaxation time appeared to change over time from an initial 32 to 94 s and was detectable up to five minutes after dissolution. The spectroscopic imaging demonstrated that lithium appears to distribute in the brain.

                  1757.     A Simple, Low-Cost Device for Producing Hyperpolarized Heteronuclear Contrast Agents Using
                                Parahydrogen-Induced Polarization

Stephen Kadlecek¹, Vahid Vadhat¹, Robert V. Cadman¹, Kiarash Emami¹, John MacDuffie Woodburn¹, Richard A. Guyer¹, Masaru Ishii, ¹², Jiangsheng Yu¹, Hans Hyonchang Kim¹, Warren Gefter¹, Rahim R. Rizi¹

¹University of Pennsylvania, Philadelphia, Pennsylvania, USA; ²Johns Hopkins University, Baltimore, Maryland, USA

We present the essential design details for a device which produces hyperpolarized, 13C-containing solutions using the spin-order derived from parahydrogen.  The device provides flexible conditions for changing the target molecule and optimizing hydrogenation chemistry.  It is based on a small jet reactor, which is housed in a low-field, transmit-only NMR apparatus.  Preliminary results hyperpolarizing several 13C-labeled molecules is presented.

                  1758.     A Mobile DNP Polarizer for Clinical Applications

Kerstin Münnemann¹, Christian Bauer², Jörg Schmiedeskamp², Hans Werner Spiess², Wolfgang G. Schreiber¹, Dariush Hinderberger²

¹Mainz University Medical School, Mainz, Germany; ²Max-Planck-Institute for Polymer Research, Mainz, Germany

A mobile and low cost apparatus for DNP hyperpolarization is presented and its performance is demonstrated at room temperature and with three radical species. DNP enhancement factors for a triarylmethyl-based radical (TAM), TEMPOL and a polyelectrolyte spin-labeled with nitroxide radicals are measured as a function of mi-crowave power and the projected maximum enhancement factors are compared. The spin-labeled polyelectro-lyte shows the best DNP efficiency and 1H DNP enhancement factors up to -50 were observed. The results demonstrate that we are able to perform reliable DNP experiments in the proposed mobile apparatus and to ob-tain reasonably high 1H-DNP enhancements.

                  1759.     Hyperpolarization of Hetero Nuclei Via Adiabatic Field Cycling of Parahydrogenated Molecules

Bob C. Hamans¹, Anna Andreychenko², Sybren S. Wijmenga², Arend Heerschap¹, Marco Tessari²

¹Radboud University Nijmegen Medical Centre, Nijmegen, Netherlands; ²Radboud University Nijmegen, Nijmegen, Netherlands

Para-hydrogen induced polarization (PHIP) makes use of the macroscopic spin order contained in para-hydrogen gas (pH2) to produce hyperpolarization. Field cycling following a para-hydrogenation reaction allows for polarization transfer from the pH2 protons to other protons and hetero nuclei (e.g. 13C, 31P) of the hydrogenated substrates. A complete numerical simulation of this field cycling experiment is of great help in the selection of the optimal biologically interesting para-hydrogenatable substrates as well as for determining optimal field cycling conditions.

                  1760.     New Hyperpolarized Agents from Para-Hydrogenation of ¹³C-Labelled Butynoic Acid and Methylbutynoate

Alessandra Viale¹, Daniela Santelia¹, Francesca Reineri¹, Roberta Napolitano¹, Roberto Gobetto¹, Silvio Aime¹

¹University of Torino, Torino, Italy

Para-hydrogenations of butynoic acid and its methyl ester (both protonated and deuterated) yield ¹³C signal enhancements in the reaction products, useful for ¹³C-MRI. Since the hydrogenation yield is higher for the ester and it is well soluble in water, it has been used to produce ¹³C MR images.

                  1761.     Method for Measuring Spin Relaxation During Production of Hyperpolarized 13C

Robert V. Cadman¹, Stephen Kadlecek¹, Kiarash Emami¹, Richard A. Guyer¹, Masaru Ishii, ¹², Hans Hyonchang Kim¹, Vahid Vadhat¹, John MacDuffie Woodburn¹, Jiangsheng Yu¹, Rahim R. Rizi¹

¹University of Pennsylvania, Philadelphia, Pennsylvania, USA; ²Johns Hopkins University, Baltimore, Maryland, USA

We perform a numerical calculation of the effect of relaxation during a pulse sequence used to transfer spin order from protons in a state with total spin zero to a heteronucleus.  We propose that by measuring polarization resulting from a variety of pulse sequences, the relaxation time constants may be determined.  Knowledge of relaxation time constants would allow for better optimization of the pulse sequences.

                  1762.     Sensitivity Enhancement of Hyperpolarized Nuclei Through Polarization Transfer

Stephen Kadlecek¹, Kiarash Emami¹, Vahid Vadhat¹, Jiangsheng Yu¹, Richard A. Guyer¹, Robert V. Cadman¹, Masaru Ishii, ¹², John MacDuffie Woodburn¹, Hans H. Kim¹, Warren Gefter¹, Rahim R. Rizi¹

¹University of Pennsylvania, Philadelphia, Pennsylvania, USA; ²Johns Hopkins University, Baltimore, Maryland, USA

We demonstrate a technique for enhancement of hyperpolarized heteronuclear sensitivity by transferring polarization to a vicinal, higher-gyromagnetic ratio nucleus.  Phantom imaging results provide evidence that significant signal-to-noise gains can be achieved without sacrificing the advantages of a negligible background when imaging the heteronucleus directly.  Efficient transfer can be achieved despite the non-ideal nature (low-power pulse, unavoidable between-pulse delays) of a clinical MRI environment.

                  1763.     Mimicking the Role of NADH and FADH2 with Homogenous Hydrogenation Catalysts and Parahydrogen
                                Providing 13C-Hyperpolarized Biochemical Intermediates for 13C-MRI or 13C-MRS

Joachim Bargon¹, ², Ute Bommerich³, Matthias Stephan¹, Rahim R. Rizi²

¹University of Bonn, Bonn, Germany; ²University of Pennsylvania Medical Center, Philadelphia, Pennsylvania, USA; ³Leibniz Institute for Neurobiology, Magdeburg, Germany

The role of both enzyme-mediated biochemical redox systems, flavin adenine dinucleotide (FAD), the precursor molecule to 1,5-dihydro-FAD (FADH2), and nicotinamide adenine dinucleotide (NAD+) and its reduced form, NADH, can be mimicked by water-soluble homogeneous hydrogenation catalysts, which convert the unsaturated forms into their saturated derivatives. In connection with parahydrogen both proton and 13C-hyperpolarization result, which boosts the sensitivity of 13C-MRS and -MRI by some orders of magnitude. In this fashion various intermediates, for example of the degradation and synthesis of fatty acids and amino acids have been hyperpolarized successfully. Characteristic examples include glutamate, amino butyrate, succinate, fumarate, L-DOPA, etc.

                  1764.     Investigating the Metabolism of Glucose: An Alternative to 13C-Hyperpolarized Pyruvate

Joachim Bargon¹, ², Rahim R. Rizi, ¹²

¹; ²University of Pennsylvania Medical Center, Philadelphia, Pennsylvania, USA

Water-soluble chiral homogeneous Rh(I)-hydrogenation catalysts containing ligands derived from the chiral pool of natural sugars hydrogenate glucose-derived unsaturated precursors to yield hyperpolarized derivatives boosting the sensitivity of 13C-MRI or 13C-MRS, providing an alternative to hyperpolarized pyruvate. The 1H-PHIP spectra reveal the formation of intermediate complexes consisting of the catalyst and the unsaturated substrate. The number of possible stereoisomers of these intermediates depends on the symmetry of the ligands used. The spectral analysis of these intermediates yields information about the enantiomeric purity of the hydrogenation product formed. For hyperpolarizing glucose this is of special importance since only D-glucose can be metabolized.

                  1765.     Providing 13C-Hyperpolarized Nicotine Derivatives for Use in 13C-MRI or 13C-MRS

Joachim Bargon¹, Ute Bommerich², Achim Koch³, Rahim R. Rizi⁴, Meike Roth³, Jorg Schmiedeskamp³, Hans W. Spiess³

¹University of Bonn, Bonn, Germany; ²Leibniz Institute for Neurobiology, Magdeburg, Germany; ³Max-Planck Institute for Polymers, Mainz, Germany; ⁴University of Pennsylvania Medical Center, Philadelphia, Pennsylvania, USA

13C-hyperpolarized alkylated nicotines mimic neurotransmitters like acetylcholine, but unlike those they readily and hence quickly penetrate the blood-brain barrier. Nicotine derivatives are considered as medications to treat Parkinson's disease, Alzheimer's disease, schizophrenia, depression, anxiety or pain. On the other hand, nicotine serves an important role in provi-ding pleasure and relaxation. Due to its addictive properties, quitting smoking is difficult, even though the consequences of smoking health-wise are severe. 13C-hyperpolarized nicotines are a prerequisite applying 13C-MRI or 13C-MRS due to the low sensitivities of these methods. The required unsaturated precursors are readily synthesized. SIB-1509Y is a drug itself.

                  1766.     Observation of Anomalously Long-Lived Hyperpolarized C13 States in  Parahydrogen-Induced
                                 Polarization<

Aaron K. Grant¹, Elena Vinogradov¹

¹Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, Massachusetts, USA

Hyperpolarized contrast media based on nuclei such as 13C yield dramatic enhancements of signal-to-noise ratio in MRI. The comparatively short T1 relaxation time of the available agents presents a challenge to applications of hyperpolarization.  Recent observations have shown that certain quantum-mechanical spin states possess relaxation times significantly longer than T1 when they are stored under low field conditions.  Measurements in parahydrogen-induced polarization have demonstrated significant lifetime enhancements in homonuclear proton systems, but large enhancements have not been observed in heteronuclear systems.   Here we present observations in a heteronuclear system where a carbon in a ‘CH’ group possesses an anomalously long lifetime.

                 1767.     Conservation of Hyperpolarized Long-Lived States in Low Field: Theory and Experiment

Elena Vinogradov¹, Aaron K. Grant¹

¹Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, Massachusetts, USA

Hyperpolarized liquid-phase contrast media dramatically improve the sensitivity of MRI.  Although the signal enhancements achieved with hyperpolarization are dramatic, they are also short-lived owing to the comparatively short T1 relaxation times of the available agents.  Recent observations have shown that certain quantum-mechanical states can have lifetimes significantly longer than T1.  Here we present results of theoretical and experimental work on long-lived states in three- and four-spin systems at low field.   We document a long-lived state in hyperpolarized ethyl acrylate, and describe how a ‘bottleneck’ in the relaxation process accounts for the observed time dependence of NMR spectra from this compound.

                  1768.     MRI of Hyperpolarized 3He at 3T

Kevin Teh¹, Nicola de Zanche², Salma Ajraoui¹, Matthew Clemence, Klaas Pruessmann², Jim M. Wild¹

¹University of Sheffield, Sheffield, UK; ²University and ETH Zürich, Zurich, Switzerland

The optimum B0 field strength for hyperpolarised (HP) MRI experiments is the subject of some discussion. The non-Boltzmann polarisation (achieved here with laser optical pumping) makes the magnetisation independent of B0, and low field MRI becomes a realistic possibility. Furthermore the field inhomogeneity at high B0 in the lungs makes the use of higher fields for lung imaging more challenging. Nevertheless, MRI manufacturers are shifting their multinuclear engineering focus on to higher B0 systems and engineering quality as well as electromagnetic physics will ultimately determine SNR in practice. The objective of this work was to investigate the feasibility of HP 3He MRI at 3T on a whole body system. The engineering aspects of upgrading our 3T system for 3He transmit receive are presented. Preliminary imaging results and SNR comparisons are made with data acquired at 1.5T with non-lossy 3He samples and coils of the same size and geometry.

                  1769.     Radiation Damping as a Diagnostic Tool for 3He Polarimetry in Optical Pumping Cells

Steven R. Parnell¹, ², Steven Boag², Max Skoda², Chris D. Frost², Jim M. Wild¹

¹University of Sheffield, Sheffield, UK; ²Rutherford Appleton Laboratories, UK

For hyperpolarised 3He, accurate measurement of the polarisation is necessary. In-situ polarimetry of gas in the cell typically relies on small tip angle NMR from a coil close to the cell. In this work we investigate the effects of the high magnetisation on the observed FID from the optical pumping cell in both the high and low energy states and the effect of radiation damping on the observed total magnetisation.

                  1770.     Pulmonary Oxygen Mapping with ³He MRI at Very-Low-Field

Ross William Mair¹, Rachel Nora Scheidegger¹, ², Leo Lee Tsai, ²³, Matthew Scott Rosen¹, ⁴, Ronald Lee Walsworth¹, ⁴

¹Harvard-Smithsonian Center for Astrophysics, Cambridge, Massachusetts, USA; ²Harvard-MIT Division of Health Sciences and Technology, Cambridge, Massachusetts, USA; ³Harvard Medical School, Boston, Massachusetts, USA; ⁴Harvard University, Cambridge, Massachusetts, USA

The partial pressure of oxygen in the lung may be derived from the attenuation of the ³He MRI signal, if the RF flip-angle is precisely known. In clinical MRI systems this measurement is non-trivial, due to coil-loading effects of different human subjects and the B₁ inhomogeneity of RF coils used at high field. Therefore, flip-angle calibrations are incorporated into every measurement. We employ an open-access, very-low-field human MRI system to study posture-dependent effects on pulmonary function.  Operation of this system at 210 kHz has resulted in a simplification of the pulmonary oxygen measurement technique in comparison to high-field methods.

                  1771.     High Production of Hyperpolarized Helium-3: Commercial Prototype

Iulian C. Ruset¹, ², David Watt¹, Jan Distelbrink¹, Adam Straub¹, John Brackett¹, Edward J. Kotkowski², Chad D. Everbeck², Korac MacArthur¹, Alex J. Johnston², Peter L. Conti², F.W. W. Hersman¹, ²

¹Xemed LLC, Durham, New Hampshire, USA; ²University of New Hampshire, Durham, New Hampshire, USA

We present a commercial prototype for the high-production of hyperpolarized helium-3. The system is based on hybrid-alkali optical-pumping using kilowatt laser power. The polarizing cell is a large eight liter cylinder placed inside a pressure aluminum vessel which minimizes the stress on the glass. A gas line designed to minimize polarization losses and alkali diffusion allows for helium removal and renewal. The system is envisioned to produce 100 liters/day at 70% hyperpolarized helium-3. Preliminary tests confirmed the feasibility of the system. Polarization was limited to few percent by the low relaxation time of the Pyrex cell used. Improvements are underway.

                  1772.     Accelerated 3D Imaging of Oxygen Partial Pressure Using Projection Acquisition and Constrained
                                 Reconstruction

Rafael Luis O'Halloran¹, James Hartman Holmes¹, Sean Bedillion Fain¹

¹University of Wisconsin, Madison, Wisconsin, USA

A 3D stack-of-stars acquisition for oxygen partial pressure and flip angle parametric mapping in hyperpolarized helium-3 MRI is presented, validated in a phantom and demonstrated in a healthy volunteer. The acquisition provided 10 axial slices at 16 distinct time-points within a 16 s breath-hold. Measured oxygen pressures matched previously observed values in human lung and demonstrated the known anterior/posterior dependence in the supine position.

                  1773.     Evaluation of Dynamic Lung Function Using Non-Equilibrium Xenon Uptake Spectroscopy (NEXUS)

Kai Ruppert¹, Jaime F. Mata², Talissa Altes¹, James R. Brookeman², Klaus D. Hagspiel², John P. Mugler III²

¹The Children's Hospital of Philadelphia, Philadelphia, Pennsylvania, USA; ²University of Virginia, Charlottesville, Virginia , USA

Non-Equilibrium Xenon Uptake Spectroscopy (NEXUS), an extension of a commonly-used hyperpolarized xenon-129 MRS pulse sequence, might provide unique insights into the lung-function dynamics of a freely breathing subject that far exceed those of conventional spirometry. For instance, the gas transit time to travel to the alveolar gas-exchange sites might become an indicator of small airway disease. Other applications for NEXUS might include a free-breathing exam in children who cannot perform forced respiratory maneuvers. Also, with a temporal resolution of 30ms or better, which is much greater than that of any existing technique, not even a breath hold would be required.

                  1774.     Temperature-Sensitive Imaging by Means of Exchangeable Functionalized 129Xe

Monica A. Smith¹, ², Leif Schröder¹, ², Tyler Meldrum¹, ², Thomas J. Lowery¹, ², Alexander Pines¹, ², David E. Wemmer¹, ²

¹University of California, Berkeley, Berkeley, California , USA; ²Lawrence Berkeley National Laboratory, Berkeley, California , USA

Recent advances using functionalized 129Xe biosensors and hyperpolarized chemical exchange saturation transfer (HYPER-CEST) have overcome some of the intrinsic limitations when using 129Xe as a contrast agent, but additional gains in sensitivity are required for use of these methods in vivo.  Here, the effect of increasing the temperature of the system was investigated.  It was shown that HYPER-CEST contrast increased when the temperature was increased from 26 C to 32 C.  Preliminary data suggests that additional contrast is achievable at body temperature, which holds promise for in vivo applications.

                  1775.     Absolute Quantification of Pulmonary Perfusion Using Intravenous Injection of Hyperpolarized ¹²⁹Xe

Harald E. Möller¹, Bastiaan Driehuys², James Pollaro², Laurence W. Hedlund²

¹Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany; ²Center for In Vivo Microscopy, Durham, North Carolina, USA

Recently, a means of lung perfusion imaging based on intravenous injection of hyperpolarized ¹²⁹Xe dissolved in saline was proposed. Combination of such images with information obtained from serial gas-phase spectroscopy allows quantification of global lung perfusion. This information that can be used for scaling the image to obtain maps of regional lung perfusion in absolute units.

                 1776.     Hyperpolarized Xenon at 10 Liters Per Hour for Diagnostic MRI

Jan Distelbrink¹, Jeff Ketel¹, David Watt¹, Adam Straub¹, Iulian C. Ruset¹, ², Walt Porter¹, Stephen Ketel¹, John Brackett¹, Aaron Hope¹, Korac MacArthur¹, Silviu Covrig², F.W. W. Hersman¹, ²

¹Xemed LLC, Durham, New Hampshire, USA; ²University of New Hampshire, Durham, New Hampshire, USA

We present a compact system for high polarization, high volume (50% at 10 liters/hr) production of hyperpolarized xenon. Preparations to build production units have started. The system is based on the existing 1 liter/hr Xemed XeBox-B polarizer. Its 2” diameter glass column was replaced by a 6” square copper column for increased gas flow and dissipation of heat from the 800 Watt wavelength narrowed CW laser. An on-board computer controls the polarizer. Automated operation is facilitated by a web-based interface. Compliance with FDA 21 CFR 210 and 21 CFR 211 is anticipated.

Diffusion: Artifacts, Phantoms, QA, Reproducibility

Jesper Frandsen¹, Leif Østergaard¹, Eva B. Vedel Jensen²

¹Aarhus University Hospital, Aarhus C, Denmark; ²Aarhus University, Aarhus C, Denmark

From Diffusion Tensor Imaging (DTI) scalar indices such as apparent diffusion coefficient (ADC) and fractional anisotropy (FA) can be extracted. This indices are often used for comparing groups of healthy controls to groups with neurodegenerative deceases. In this study we have performed DTI on the same subject 20 times within few days to examine the reproducibility of these indices on a 3.0T scanner.

                  1819.     Quantitative Evaluation of  Diffusion Weighted-MRI Phantoms

Wilhelmus LPM Pullens¹, ², Alard Roebroeck³, Rainer Goebel³

¹Maastricht University, Maastricht, Netherlands; ²Brain Innovation BV, Maastricht, Netherlands; ³Maastricht University, Netherlands

This paper presents phantoms for qualitative and quantitative validation of diffusion weighted MRI and fiber tracking. The aim is to construct a phantom with properties (T2 and anisotropy) similar to human white matter , combined with the possibility to create different geometries, to be measured with clinical protocols. Phantoms show higher perpendicular versus parallel ADC, as well as exponential signal decay when b-value is increased. FA is in a similar range as FA in human white matter. The phantom can be used for validation of fiber tracking; the construction method is suitable for crossing or kissing geometries at various angles.

                  1820.     Effects of Motion on Clinical Diffusion Tensor Imaging

Rob Hendrikus Tijssen¹, Jacobus F. Jansen², Walter H. Backes²

¹Eindhoven University of Technology, Eindhoven, Netherlands; ²Maastricht University Hospital, Maastricht, Netherlands

In DTI, the gradient sampling scheme affects the propagation of image noise. Various optimized schemes have been proposed. However, these schemes are not immediately available in a clinical setting. Additionally, subject motion may have a considerable effect. This work assesses six clinically available gradient schemes by in vivo analyses and computer simulations. Our data show that, like image noise, motion effects are dependent on the sampling scheme. Whereas schemes with medium and high angular resolution showed similar results, gradient schemes with only six gradient directions demonstrated a considerably lower reproducibility and incur positive bias in the presence of motion.

                  1821.     Interpolation and Regularization of Diffusion Tensors Along Geodesics

Jaime E. Cisternas¹, Marcelo Gálvez², Gonzalo Rojas², Takeshi Asahi²

¹Universidad de los Andes, Santiago, Chile; ²Universidad de Chile, Santiago, Chile

The processing of diffusion tensor imaging (DTI) and tractography involve a number of steps, such as the registration, the estimation and the smoothing of tensor fields. These procedures implicitly rely on a measure to compare two or more tensors as well as on a continuous transformation of tensors.Several strategies for measuring distances between tensors have been proposed, some of them inspired from sophisticated mathematics and some others aiming at fast computations. Here we propose a simple theoretical framework to separate the problem into two parts: the first considers the shape of tensors, and the second their orientation. We illustrate the application of the interpolation and the regularization methods built from such a distance using a synthetic example and a DTI scan from a fiber phantom.

                  1822.     Towards a Diffusion Standard Ruler: Rigid Diffusion Phantom

Koji Sakai¹, ², Takashi Azuma¹, Susum Mori², Koji Koyamada¹, Sadami Tsutsumi¹

¹Kyoto University, Kyoto, Japan; ²Johns Hopkins University, Baltimore, Maryland, USA

Anisotropy phantom for DTI studies is much needed for calibration of multi-center studies. Ideally, we can control the degree of anisotropy and the phantom is stable and mass-producible. In this paper, we introduce a DTI phantom based on channeled silicon plates. We demonstrate that the water diffusivity and anisotropy can be controlled by changing the channel depth and the anisotropy can be as high as 0.8. Interestingly, we observed axial diffusivity higher than free diffusion, posing us an interesting question about the interpretation of DTI results. Quality control for the production is a current issue to make this phantom available.

                 1823.     The Impact of Robust Tensor Estimation on Voxel-Wise Analysis of DTI Data

Daniel Jon Peterson¹, Bennett A. Landman², Laurie E. Cutting¹, ²

¹Kennedy Krieger Institute, Baltimore, Maryland, USA; ²Johns Hopkins School of Medicine, Baltimore, Maryland, USA

Methods of robust diffusion tensor estimation, like RESTORE, have been shown to decrease the effect of artifacts in a single subject. Here we demonstrate the impact of RESTORE on a voxel-wise group analysis of DTI data, using Cerebellar Ataxia as a sample condition. In a voxel-wise statistical comparison of the difference in FA between controls and patients, RESTORE was found to decrease the number of small areas of significance that are likely to be spurious. Additionally, a high number of outliers were found in regions vulnerable to artifacts, and RESTORE was shown to alter FA in a spatially heterogeneous manner

                  1824.     How Reliable Are Diffusion Tensor Spectroscopy Measures of Metabolite Diffusion?

Jacob Ellegood¹, Chris C. Hanstock¹, Christian Beaulieu¹

¹University of Alberta, Edmonton, Canada

Diffusion tensor spectroscopy (DTS) of metabolites in human brain could provide novel information, but it is challenging given low apparent diffusion coefficients (ADC) and poor signal-to-noise.  Too much variability in measured ADC with diffusion gradient direction could artificially inflate fractional anisotropy (FA) values, particularly at low ADC, and make isotropic regions appear anisotropic. More accurate measurements could result from the use of higher b values (5000 s/mm2) than that more commonly used in DTS (1500-2000 s/mm2). The sensitivity of FA was shown to be true for a range of isotropic alcohols with low ADC and in human occipital gray matter.

                  1825.     Non-Linearity in Diffusion-Gradient Induced Eddy-Current Fields in a Head Only 3T Scanner

Pablo Velasco¹, Souheil J. Inati¹

¹NYU, New York, New York, USA

We investigated the linearity of the eddy-current induced fields in diffusion-weighted imaging in a head-only 3T scanner.  We find that the short-time eddy-current fields scale with the amplitude of the diffusion gradients, but the long-time eddy-current induced fields do not.  This non-linearity implies that a full characterization of the spatio-temporal distribution of the field is needed for optimal image quality in DWI.

                  1826.     Employing Bootstrapping Methods to Examine the Need for Pulse Triggering in Diffusion-Weighted
                                 Imaging

Zoltan Nagy¹, Chloe Hutton¹, Daniel C. Alexander¹, Ralf Deichmann¹, ², Nikolaus Weiskopf¹

¹University College London, London, UK; ²University of Frankfurt, Frankfurt, Germany

Employing pulse triggering is in wide-spread use for diffusion weighted imaging since the demonstration that this reduces image variance. However, gradient systems have improved in speed/strength and parallel imaging methods have further reduced the acquisition time. We employed bootstrapping statistical methods to establish whether pulse triggering reduces the variance in diffusion-weighted datasets acquired using recent gradient systems. To this effect we found no significant advantage resulting from the use of pulse triggering. This offers the possibility for a significant reduction in acquisition time or the chance for multiple acquisitions but should be determined individually for each scanner/site/patient group.

                  1827.     Robust Diffusion Tensor Estimation by Maximizing Rician Likelihood

Bennett Allan Landman¹, Pierre-Louis Bazin¹, Jerry L. Prince¹

¹Johns Hopkins University, Baltimore, Maryland, USA

Maximum likelihood tensor estimation based on a full treatment of Rician distributed MR intensities (DTEMRL) has been recently shown to significantly improve DTI reliability in clinical and high SNR applications. However, DTEMRL suffers from low reliability in low SNR applications. Here, we present a robust maximum a posteriori generalization (rDTEMRL) of the DTEMRL technique by inclusion of noise level Bayesian priors and a robust likelihood function.  In simulation, the new method is shown to offer superior reliability and robust performance in the presence of artifact. In an empirical study, rDTERML improves the consistency of diffusion tensor estimates.

                  1828.     Dealing with Artifacts Induced by Spike Noise in Diffusion Tensor Imaging

Jonathan S. Jackson¹, Elvina M. Chu¹, Mara Cercignani², Maria A. Ron¹, Claudia A. Wheeler-Kingshott¹

¹UCL Institute of Neurology, London, UK; ²Fondazione Santa Lucia, Rome, Italy

DTI data can be affected by many types of artifact. In a case where one out of 68 diffusion weighted images was affected by spike noise, the FA map showed an artifactual 5% variation in white matter, which is of the order of changes induced by pathology. The effect of this artifact can be removed through a combination of manual editing and the RESTORE algorithm proposed by Chang et al. (2005).

                  1829.     Comparison of EPI Distortion Correction Methods in Diffusion Tensor MRI

Minjie Wu¹, ², Alan S. Barnett¹, Stefano Marenco¹, Lindsay Walker¹, Herve Lemaitre¹, Carlo Pierpaoli¹

¹National Institutes of Health, Bethesda, Maryland, USA; ²University of Pittsburgh, Pittsburgh, Pennsylvania, USA

EPI distortions in diffusion-weighted images (DWIs) degrade the quality of brain DT-MRI. In this study, we evaluate the performance of two methods for correcting EPI distortions: the standard method based on B0 field mapping (B0M) and a newly proposed b-spline image registration method (BSP). We found that both methods can effectively reduce the EPI geometric distortion improving the anatomical accuracy of quantities derived from the diffusion tensor. However, the BSP algorithm consistently provides better correction for rostral brain regions while the B0M method performs better at the base of the brain, including temporal lobes, brainstem, and cerebellum.

                  1830.     Robust DTI Noise Level Estimation Improves RESTORE Tensor Estimation

Bennett Allan Landman¹, Pierre-Louis Bazin¹, Jerry L. Prince¹

¹Johns Hopkins University, Baltimore, Maryland, USA

We present a noise level estimation method that specifically addresses the challenges of diffusion tensor imaging (DTI) data. Our method utilizes two repeated DTI acquisitions, which are commonly acquired in clinical DTI protocols, to identify the spatially varying noise profile. The novel noise level estimates significantly improve accuracy of the RESTORE tensor estimation method when compared with conventional noise level estimation approaches. This method is robust to background signal suppression, spatial noise correlation, and coil sensitivity variability.

Perfusion & Permeability: Dynamic Contrast Enhanced Methods

Michael Ingrisch¹, Steven Sourbron¹, Karin Herrmann¹, Maximilian Reiser¹, Michael Peller¹

¹Klinikum Großhadern, Munich, Germany

Quantification of tissue perfusion and permeability parameters from DCE-MRI bolus-tracking data requires a measurement of the tracer concentration, which can be obtained from a measurement of pre-contrast tissue relaxation rate. This study compares the effect of T1-quantification on the perfusion parameters measured with DCE-MRI in brain metastases, to a simpler approach based on relative signal enhancement. Contrary to expectations, values of perfusion and permeability parameters are increased after T1-quantification. Simulations confirm that T1-quantification may introduce additional errors if the actual flip angle is not precisely known.

                  1883.     Measuring Cerebral Blood Flow and Blood-Brain-Barrier Leakage with DCE-MRI at 3T

Steven Sourbron¹, Michael Ingrisch¹, Axel Siefert¹, Maximilian F. Reiser¹, Karin Herrmann¹

¹Ludwig-Maximilian-University Munich, Munich, Germany

DSC-MRI measurement of perfusion and permeability in the brain suffers from serious quantification issues. Here we propose an alternative in the form of an optimized DCE-MRI protocol at 3T, and evaluate it using measurements in normal tissue and in brain tumors. Maps of CBF approach the quality of DSC-MRI images. AIF selection is straightforward, CBF in Grey Matter agrees with gold-standard values, and Extraction Flow can be quantified from the same data. We conclude that DCE-MRI at 3T provides a viable alternative to DSC-MRI for the quantification of cerebral perfusion and permeability in a wide range of applications.

                  1884.     Correction of Partial Volume Effects in Plasma Time Curve for Tracer Kinetic Analysis in DCE-MRI

RKS Rathore¹, A Singh¹, RK Gupta², M Haris², SK Verma¹, A Purwar¹, G Bayu¹, MK Sarma¹, J Singh¹, S Agarwal¹, DKS Rathore¹

¹Indian Institute of Technoloky, Kanpur, Kanpur, India; ²SGPGIMS, Lucknow, India

Tracer kinetic analysis of DCE-MRI data requires the plasma concentration time curve. The use of individually measured plasma curve (if accurately measured) has advantages over standard plasma curve as it improves accuracy of measurements. In a brain DCE-MRI data the partial volume effect (PVE) is the main problem in the measurement of accurate plasma curve. The automated AIF extraction method proposed here is similar to that described by Rijpkema et al., 2001 and Parker et al., 2006, but it additionally corrects for the PVE, and also does an automatic estimation of bolus arrival time (BAT) using PL model.

                  1885.     Comparison of Arterial Input Functions Obtained from Unlabeled- And ¹⁴C-Labeled-Gadolinium-
                                 Diethylenetriaminepentaacetic Acid and Its Application in MRI Estimation of Blood-To-Brain Influx
                                 and Cerebral Microvascular Blood Space

Kishor Karki¹, ², Tavarekere N. Nagaraja¹, James R. Ewing¹, ², Joseph D. Fenstermacher¹, Robert A. Knight¹, ²

¹Henry Ford Hospital, Detroit, Michigan, USA; ²Oakland University, Rochester, Michigan, USA

The arterial input function (AIF) of contrast agent (CA) is required to calculate blood-to-brain transfer constant (K_i) and cerebral microvascular blood volume (v_D). We compared AIFs from MRI and quantitative autoradiography studies using identically prepared CAs viz., unlabeled- and ¹⁴C-labeled- gadolinium-diethylenetriaminepentaacetic acid (Gd-DTPA and Gd-[¹⁴C]DTPA, respectively) in a rat model of transient focal cerebral ischemia to test our hypothesis that the reconstructed AIF generated from Gd-[¹⁴C]DTPA can be used to correct MRI-derived estimates of K_i and v_D. For the regions with a leaky BBB, the reconstructed MRI-AIF produced significantly lower values of K_i and v_D than did the original.

                  1886.     An MRI Estimate of Vascular Permeability in 9L Cerebral Tumor Agrees with Those of Quantitative
                                 Autoradiography

James Russell Ewing¹, ², Tavarekere N. Nagaraja¹, ramesh paudyal³, Hassan Bagher-Ebadian¹, ⁴, Knight Robert¹, Ledbetter Karyn¹, Joseph D. Fenstermacher¹

¹Henry Ford Hospital, Detroit, Michigan, USA; ²Oakland University, Rochester, Michigan, USA; ³oakland University, rochester, Michigan, USA; ⁴Amir-Kabir University of Technology, Tehran, Iran

In 15 rats implanted with 9L tumor, the vascular transfer constant (K1 or Ktrans) was assessed by MRI means, and then, in a terminal experiment, by quantitative autoradiography (QAR).  Essentially the same indicator, albumin, tagged with Gd in the MRI, and with 125I in the QAR, was used.  The K1’s estimated by the two techniques were in agreement.

                  1887.     Vascular Permeability to Gd-DOTA and USPIO in C6 and RG2 Rat Glioma Models

Marine Beaumont¹, ², Benjamin Lemasson, ¹³, Régine Farion¹, ², Christoph Segebarth¹, ², Chantal Remy¹, ², Emmanuel Luc Barbier¹, ²

¹INSERM U836, Grenoble, France; ²Université Joseph Fourier, Grenoble Institut des Neurosciences, UMR_S836, Grenoble, France; ³Oncodesign Biotechnology, Dijon, France

DCE-MRI provides insight into vascular permeability of tumors. Typically, a contrast agent such as Gd-DOTA is used and data are analyzed using classical pharmacokinetic models to obtain microvascular parameters. For further characterization, intravascular contrast agent (USPIO) can be used to determine blood volume and vessel size index under stationary conditions. In view of eventually combining these two approaches into a single imaging protocol, we evaluated whether the USPIO remains intravascular during a DCE-MRI protocol. DCE-MRI experiments using Dotarem® and Sinerem® were therefore performed on two brain tumor models. The presence of macrophages, potential transporters of iron, was equally investigated.

                  1888.     Quantitative Permeability MRI in Acute Ischemic Stroke: How Long Do We Need to Scan?

Logi Vidarsson¹, Fang Liu¹, Brandy Moran¹, David Mikulis², Andrea Kassner¹

¹The Hospital for Sick Children and the University of Toronto, Toronto, Canada; ²The Toronto Western Hospital, Toronto, Canada

Dynamic T1-weighted MR permeability imaging with subsequent pharmacokinetic modeling, provides valuable information on blood-brain-barrier (BBB) integrity in acute ischemic stroke (AIS) and can lend insight into which AIS patients later develop hemmorrhage. However, the relatively long scan duration (5 min) is problematic in critically ill AIS patients. In this work we examine the effects of reducing scan time on quantitative permeability (KPS). Our results indicate that the uncertainty in permeability estimates increases as less and less data is used for analysis.

                  1889.     Which Voxels Should Be Analysed in DCE-MRI Studies of Anti-Vascular/angiogenic Compounds?

Chris J. Rose¹, James P. O'Connor¹, Brandon Whitcher², Geoff J. Parker¹

¹The University of Manchester, Manchester, UK; ²GlaxoSmithKline, London, UK

This abstract describes a problem with the way that average K^trans is often computed in the context of studies of anti-vascular/angiogenic compounds, provides a simple geometrical explanation, proposes a solution and compares the statistical power of tests based on the two methods. Tumours often have an enhancing rim, thought to correspond to increased angiogenic activity. In such cases, we found that a given statistical power can be achieved using a smaller sample size when computing a voxel-wise average K^trans from the enhancing region only, compared to computing the voxel-wise average K^trans from all tumour voxels.

                  1890.     Quantifying Spatial Heterogeneity in Dynamic Contrast-Enhanced MRI Parameter Maps

Chris J. Rose¹, Sam Mills¹, James P. O'Connor¹, ², Gio Buonaccorsi¹, Caleb Roberts¹, Yvon Watson¹, Sha Zhao¹, Brandon Whitcher³, Gordon Jayson², Alan Jackson¹, Geoff J. Parker¹

¹The University of Manchester, Manchester, UK; ²Christie Hospital NHS Trust, Manchester, UK; ³GlaxoSmithKline, London, UK

In DCE-MRI-based drug trials, each tumour voxel is summarised using model-free or tracer kinetic model parameters. The tumour is summarised by average parameter value, but this fails to capture heterogeneity information. Heterogeneity has been described using histograms (for example), but previous methods discard all spatial information. This may be important to the correct interpretation of the structure of DCE-MRI parameter maps. We have developed statistics that consider parameters’ values and spatial locations. We show that these are sensitive to known drug effects and may predict the grade of gliomas (which are determined in histology by spatial heterogeneity).

                  1891.     Assessing Drug Effects by Comparing DCE-MRI Parameter Maps Using the Earth Mover's Distance Metric

Chris J. Rose¹, Heather Reynolds¹, James P.B. O'Connor¹, ², Sue Cheung¹, Yvon Watson¹, Gordon Jayson², Brandon Whitcher³, Geoff J. Parker¹

¹The University of Manchester, Manchester, UK; ²Christie Hospital NHS Trust, Manchester, UK; ³GlaxoSmithKline, London, UK

In DCE-MRI drug trials, current practice involves computing parameters such as K^trans at each tumour voxel and then summarising the tumour by the median K^trans value. The change in median values before and after treatment is used as a dissimilarity measure. Drug effects are often observed in particular locations within the tumour. Median K^trans may be insensitive to these heterogeneous local changes as they neglect the spatial location of parameters. This abstract describes how drug effects can be assessed by using the cost of matching pre- and post-treatment parameter maps using the Earth Mover's Distance metric. 

                  1892.     Fusion of Dynamic Contrast Enhanced (DCE) Perfusion Metrics with DTI Metrics Results in Better
                                Assessment of Corticospinal Tract Infiltration in Malignant Gliomas

Rakesh K. Gupta¹, Mohammad Haris¹, Mazhar Husain², Sanjay Verma³, Anup Singh³, Abhishek Yadav¹, Nuzhat Husain², Ram Kishore Singh Rathore³

¹Sanjay Gandhi Postgraduate Institute of Medical Sciences, Lucknow, India; ²CSM Medical University, Lucknow, India; ³Indian Institute of Technology, Kanpur, India

Edema surrounding the malignant tumors is a morphological sign of tumor infiltration in the adjacent white matter. Diffusion tensor imaging has been used to differentiate the infiltrative white matter tract from the normal white matter tract. The Malignant gliomas have shown high value of cerebral blood volume (CBV), cerebral blood flow (CBF), permeability (ktrans) and leakage (ve). In the current study, we fused DTI indices maps with perfusion maps with a hypothesis that the perfusion maps when combined with DTI maps should able to separate the edematous region from infiltrating fibers which may not be possible alone on DTI.

                  1893.     DCE-MRI for Assessment of Effects of Anti-Angiogenic Therapy: Comparison of the Transfer Constant
                                (Ktrans) to Blood Permeability Derived by a Distributed Parameter Model

Choon Hua Thng¹, Tong San Koh², Septian Hartono¹, ², Helmut Rumpel³, James Boon Kheng Khoo¹, Albert Su Chong Low³, Ai Bee Ong⁴, Norita Sukri⁴, Bee Choo Tai⁵, Ross Soo⁴, Rod A. Humerickhouse⁶, Boon Cher Goh⁴

¹National Cancer Centre, Singapore, Singapore; ²Nanyang Technological University, Singapore; ³Singapore General Hospital, Singapore; ⁴National University Hospital, Singapore; ⁵National University of Singapore, Singapore; ⁶Abbott Laboratories, North Chicago, Illinois, USA

Dynamic contrast enhanced MRI (DCE-MRI) with tracer kinetic modeling has been proposed as a biomarker of angiogenesis imaging. Generalized kinetic (GK) model  and uptake integral approach are commonly used DCE-MRI models whose representative parameters are Ktrans and initial area under the signal-time curve (IAUC), respectively. The distributed parameter (DP) model is a DCE-MRI model that enables derivation of blood flow and capillary permeability-surface area product (PS) independently. We aim to study the DP model as an alternative method of angiogenesis assessment and correlate the above parameters to drug exposure and patient outcome in a Phase I anti-angiogenic trial.

                  1894.     Effect of Image Acquisition Protocol on Vascular Parameter Estimates from DCE-MRI Liver Data

Matthew R. Orton¹, David J. Collins¹, David J. Hawkes², David Atkinson², Martin O. Leach¹

¹Institute of Cancer Research, Sutton, UK; ²Univeristy College London, London, UK

Since the liver is a highly vascularised organ, accurate pharmaco-kinetic models for DCE-MRI liver data must include a plasma fraction.  The temporal resolution of the acquired data must therefore be fast enough to allow this feature to be observed.  This work assesses the impact on the vascular parameter estimates of acquiring two volumes per breath-hold in a liver imaging protocol instead of one.

Imaging in Stroke: Clinical Studies

                 1948.     Biomarkers of Cerebral Microvascular Angiopathy in Healthy Subjects at Risk of Stroke

Alan Jackson¹, Johann Selvarajah², Marietta Scott³, Sharon Hulme³, Rachel Georgiou³, Nancy Rothwell³, Pippa Tyrell⁴

¹University of Manchester, Withington, UK; ²Greater Manchester Neuroscience Centre, Salford, UK; ³University of Manchester, Manchester, UK; ⁴Greater Manchester Neuroscience Centre, Manchester, UK

We have examined 2 imagign biomarkers of cerebral microvascular disease in normal elderly subjects at risk of stroke. Both dilated Virchow Robin Spaces and Arterial-CSF pulsewave propogation rates were significantly different in subjects at risk and subjects not at risk whilst white matter lesion distribution was not significantly different.

                  1949.     Detection of Crossed-Cerebellar Hypoperfusion in Acute Stroke Using Perfusion-Weighted MRI

Jonathan T. Kleinman¹, Doris D.M. Lin¹, Robert J. Wityk¹, Rebecca F. Gottesman¹, Andrew W. Lee¹, Peter B. Barker¹

¹Johns Hopkins University, Baltimore, Maryland, USA

Crossed cerebellar diaschisis (CCD) is commonly reported in the nuclear medicine literature in studies of acute stroke. This abstract reports on the incidence of CCD (hypoperfusion) in acute stroke as detected by MR perfusion imaging.

                  1950.     Correlation of Acute Perfusion Lesion Volumes with Neurological Deficits Depends on Deconvolution
                                Algorithm

Ona Wu¹, Jie Lu¹, Vicky J. Tiglias¹, Christian A. Holt², Thomas Benner¹, William A. Copen², E Murat Arsava¹, Hakan Ay¹, Pamela W. Schaefer², R Gilberto Gonzalez², Lee H. Schwamm², Aneesh B. Singhal², A Gregory Sorensen¹

¹Massachusetts General Hospital, Charlestown, Massachusetts, USA; ²Massachusetts General Hospital, Boston, Massachusetts, USA

A consecutive series of acute stroke patients who received PWI<12h of stroke onset was retrospectively studied (N=144) to determine whether the extent of correlation of PWI maps with neurologic deficits depends on deconvolution algorithm. Lesion volumes on perfusion maps calculated using truncated standard singular value decompositions (sSVD) was compared with PWI maps generated using a delay-insensitive technique (oSVD). oSVD was found to produce maps that correlated better with acute and follow-up neurological deficits than sSVD. oSVD produced maps that also better corresponded with follow-up lesion volumes in patients not given thrombolytic therapy.

                  1951.     Acute Middle Cerebral Artery Stroke - Comparison of Haemodynamic Timing Properties Assessed
                                 by Dynamic Contrast-Enhanced Perfusion Measurement and Arterial Spin Labeling

Johannes Gregori¹, ², Achim Gass¹, ³

¹Klinikum Mannheim, Universität Heidelberg, Mannheim, Germany; ²mediri GmbH, Heidelberg, Germany; ³Universitätsspital Basel, Basel, Switzerland

Dynamic Susceptibility Contrast (DSC) imaging and Arterial Spin Labeling (ASL) yield different basic timing properties – e. g. time to peak (TTP) in DSC and bolus arrival time (BAT) in ASL - which give different measures of the same underlying haemodynamic mechanisms. While the temporal resolution of DSC is limited to 1-2 seconds, ASL time series can in principle be acquired in arbitrarily small intervals.Reliable assessment of haemodynamics is crucial in acute stroke - especially penumbral areas are of major interest. In this work we present a comparison of BAT and TTP maps for cases of acute middle cerebral artery stroke. Focus is layd on arterial inflow delay differences on a sub-second scale which are difficult to assess by DSC imaging.

                  1952.     Does Arterial Spin Labeling Have a Role in Stroke Imaging? Preliminary Results from a 180 First Time
                                Stroke Patients Study

Esben Thade Petersen¹, Amandine Cheze¹, Violet Chua¹, N V. Ramani¹, Robert Ngo Gan¹, Xavier Golay, ¹²

¹National Neuroscience Institute, Singapore, Singapore; ²Singapore Bioimaging Consortium, Singapore, Singapore

Arterial Spin Labeling suffers from an inherently low signal-to-noise ratio which necessitates averaging and therefore relative long scan times, making the technique prone to motion artifacts. In addition, the white matter signal is at the sensitivity limit and it is often debatable whether ASL can contribute to stroke imaging at all. In this work, we compare information obtained with standard gadolinium-based perfusion methods to that obtainable using arterial spin labeling techniques. Comparable CBF and timing information was demonstrated with both techniques, while ASL delivered additional information on collateral perfusion which plays an important role in patients with cerebral artery occlusion.

                  1953.     Added Combined Value of Pulse Arterial Spin Labelling (PASL) and Susceptibility Weighted Imaging (SWI)
                                in Cerebral Vascular Occlusion, Stroke and Recovery

Magalie Viallon¹, Andrea Federspiel², Stephen Altrichter¹, Aphrodite Syrogiannopoulou¹, Roman Sztajzel¹, Maria-Isabel Vargas¹, Jacqueline Delavelle¹, Karl Olof Lövblad¹

¹Hopital Universitaire de Genève, GENEVA, Switzerland; ²University Hospital of Psychiatry, BERN, Switzerland

Improving diagnosis, prognosis and management patient with cerebral vascular diseases means determining degree of tissue injury, location of occlusion and ischemic lesions, presence or not of micro-hemorrhages, of collateral luxury perfusion or not, etc. We believe that advanced new technique like ASL and SWI can bring new determinant informations to hightligh the hemodynamic compromise and improve patient management, with ASL being an helpful technique to follow reliably recovery. We try to address on 30 patients how ASL and SWI added to a clinical stroke protocol could lead to improved diagnosis, prognosis and patient management.

                  1954.     Which Physiological Parameters Determine Outcome in the Acute and the Subacute Phases of Acute
                                 Ischemic Stroke?

Kristjana Yr Jonsdottir¹, Leif Ostergaard¹, Kim Mouridsen¹

¹Aarhus University Hospital, Aarhus, Denmark

Patients presenting with acute ischemic stroke within 3 hours of symptom onset are routinely treated with rt-PA. However, chronic tissue damage may not be avoided by acute recanalization alone. We use predictive algorithms to characterize the temporal evolution of infarct risk in acute stroke, and assess the extent to which perfusion and diffusion MRI parameters predict subsequent infarct. Perfusion parameters are primarily dominant for the risk of tissue damage in acute phases, while diffusion determines final outcome in sub-acute phases. We speculate that predictive algorithms may elucidate the diagnostic significance of imaging findings at various time intervals after stroke onset.

                  1955.     Clinical Utility of Parametric Perfusion Estimates in Prediction of Final Outcome in Acute Stroke

Kim Mouridsen¹, Kristjana Yr Jonsdottir¹, Sune Jespersen¹, Leif Ostergaard¹

¹Aarhus University Hospital, Aarhus, Denmark

DSC-MRI parameters such as cerebral blood flow (CBF) and mean transit time (MTT) are important diagnostic maps, e.g. in acute stroke where they are used to identify ischemic regions. These parameters are commonly estimated using standard SVD (sSVD) or block-circulant SVD (oSVD). Recently, a parametric Bayesian approach based on a physiological model of the microvasculature has been suggested which additionally allows computation of oxygen extraction fraction (OEF). Here we use voxel-wise predictive algorithms to demonstrate that the parametric model leads to significantly improved prediction of final infarct size. Moreover, the highest performance is observed by inclusion of the OEF.

                  1956.     DSC-MRI Errors Due to Bolus Delay and Dispersion in Sub Acute Stroke Patients: Implications for
                                Extending the Therapeutic Time Window

Lisa Willats¹, Alan Connelly¹, ², Henry Ma, ²³, Geoffrey Donnan², ³, Fernando Calamante¹, ²

¹Brain Research Institute, Melbourne, Australia; ²University of Melbourne, Melbourne, Australia; ³National Stroke Research Institute, Melbourne, Australia

DSC-MRI perfusion measurements in stroke patients with cerebrovascular abnormalities are often erroneous because of bolus delay/dispersion.  Treatment of ischaemic stroke patients using thrombolytic agents such as tissue-type plasminogen activator (tPA) is currently restricted to a logistically demanding time window of less than 3 hours from symptom onset. Extension of this time window in individual patients requires reliable perfusion information. We show that the presence of dispersion causes an overestimation of the perfusion abnormality using standard and delay insensitive deconvolution analyses in a group of sub acute patients, and present a methodology to minimise dispersion errors in the affected patients.

                  1957.     Predicting Infarct Growth with Multi-Parametric Modeling in Acute Ischemic Stroke

Michael S. Bristow¹, Brett W. Poulin¹, Jessica E. Simon¹, Michael D. Hill¹, Jayme C. Kosior², ³, Shelagh B. Coutts¹, ⁴, Richard Frayne, ¹³, J Ross Mitchell, ¹³, Andrew M. Demchuk¹, ⁴

¹University of Calgary, Calgary, Canada; ²Schulich School of Engineering, Calgary, Canada; ³Seaman Family MR Research Centre, Calgary, Canada; ⁴Hotchkiss Brain Institute, Calgary, Canada

The purpose of this study was to determine if different multi-parametric models optimally distinguish infarct growth from each of the physiologically distinct core infarct and region of benign oligemia in acute ischemic stroke. We used logistic regression analysis to build models using diffusion- and perfusion-weighted MR imaging parameters that differentiate among these regions in 13 acute ischemic stroke patients. Interaction among MR parameters was also tested. We found that indeed different parameters, and consequently different models, optimally distinguish infarct growth from core infarct as compared to those that optimally distinguish infarct growth from benign oligemia.

                  1958.     Semi-Automated Topographical Scoring for MR Imaging of Ischemic Stroke

Robert Karl Kosior¹, ², Nikolai Steffenhagen², Jayme Cameron Kosior¹, ², Andrew M. Demchuk², ³, Richard Frayne², ³

¹University of Calgary, Calgary, Canada; ²Foothills Med Ctr, Calgary Health Region, Calgary, Canada; ³Hotchkiss Brain Institute, University of Calgary, Calgary, Canada

The Alberta Stroke Program Early CT Score (ASPECTS) is a topographical stroke scoring system based on the regional occupancy of identifiable brain infarct lesions on CT images. Infarct can also be detected by MR diffusion-weighted imaging (DWI). ASPECTS may be extended to MR and such an approach may have advantages over CT. We assessed ASPECTS-based MR topographical scoring (auto-MR-TS) using a digital brain atlas for 30 ischemic stroke patients. There were small but significant differences between auto-MR-TS and ASPECTS. Auto-MR-TS is objective and reproducible and may provide greater accuracy through greater lesion sensitivity.

                  1959.     Can Diffusion Tensor Imaging Detect the Degree of Neuronal Cell Membrane Damage in Stroke Patients? :
                                A Patient Study

Koji Sakai¹, ², Kei Yamada³, Susum Mori², Tsunehiko Nishimura³

¹Kyoto University, Kyoto, Japan; ²Johns Hopkins University, USA; ³Kyoto Prefectural University of Medicine, Japan

We attempted to further extend our hypothesis to real life brain infarctions involving the tightly packed white matter bundle. We performed retrospective measurements of diffusion tensor coefficients on patient with hyperacute/acute stage infarction involving posterior limb of internal capsule. From the results, we will consider the changes of diffusion tensor coefficients as the function and structure changing discriminants in hyperacute/acute brain infarction at different stage.

                  1960.     Elevations of Diffusion Anisotropy Are Associated with Hyper-Acute Stroke: A Serial Imaging Study

Yusuf A. Bhagat¹, Muhammad S. Hussain¹, Robert W. Stobbe¹, Kenneth S. Butcher¹, Derek J. Emery¹, Ashfaq Shuaib¹, Muzzafar M. Siddiqui¹, Perkash Maheshwari¹, Fawaz Al-Hussain¹, Christian Beaulieu¹

¹University of Alberta, Edmonton, Canada

DTI studies of human ischemic stroke ≤24 hours of symptom onset have reported variable findings of changes in diffusion anisotropy. To address these heterogeneous results, we characterized longitudinal changes of fractional anisotropy (FA) by analyzing discrete ischemic deep and subcortical white matter, and deep and cortical gray matter regions during the hyperacute (2.5-7h) and acute (21.5-29h) phases of stroke onset in 13 patients. Overall, 9/13 patients scanned ≤7h showed elevated FA in at least one of the four tissues, and within the same cohort, 11/13 patients showed reduced FA in one of the four tissues 21.5-29h after stroke.

                  1961.     Comparing Mean and Directional Diffusivity in Human Ischemic Stroke

Chin-I Chen¹, ², Tammie L.S. Benzinger², Peng Sun², Tzy-Haw Wu, ²³, Agus Priatna⁴, Chung-Yi Hsu⁵, Sheng-Kwei Song²

¹Taipei Medical University - Wanfang Municipal Hospital, Taipei, Taiwan; ²Washington University, School of Medicine, St Louis, Missouri, USA; ³National Taiwan University Hospital, Taipei, Taiwan; ⁴Siemens Medical Solutions, USA, Inc, USA; ⁵Graduate Institute of Neuroscience, Taipei Medical University, Taipei, Taiwan

We present a study which compared DWI and DTI study in ischemic stroke with four different parameters: apparent diffusion coefficient (ADC), scaled relative anisotropy (sRA), axial diffusivity (AD) and the radial diffusivity (RD). We found the ADC and AD are significantly decreased in acute/subacute cases. RA and RD show the same trend but not statistically significant. In conclusion, axial diffusivity is comparable to ADC. DTI could potentially provide a quantitative biomarker to evaluate the ischemic stroke and predict outcome.

                  1962.     Neural Predictors of Immediate and Delayed Intense Naming Training Success in Chronic Aphasia

Harald Kugel¹, Ricarda Menke¹, Michael Deppe¹, Annette Baumgaertner², Hagen Schiffbauer¹, Marion Thomas¹, Kira Kramer¹, Hubertus Lohmann¹, Walter Heindel¹, Stefan Knecht¹, Caterina Breitenstein¹

¹University of Muenster, Muenster, Germany; ²Univerity Medical Center Hamburg-Eppendorf, Hamburg, Germany

Intense training facilitates the reacquisition of language in the chronic stage after stroke. We investigated with functional MRI, which brain regions are involved in immediate versus long-term success of intense naming training in patients with single left hemisphere strokes involving both Broca's and Wernicke's areas. All patients significantly improved after training. Short-term training success was predicted by increased activity in hippocampi, (pre)motor and attention areas, and right occipital lobe. Long-term training success, however, was predicted by increased activity in the left superior temporal lobe perilesionally to Wernicke's area and in the right-sided Wernicke's homologue.

                  1963.     Cerebrovascular Reserve Impairment is Associated with Recurrent Events in Patients with Carotid
                                Artery Occlusion

Stephen Goode¹, Sunil Munshi², Shane MacSweeney², Dorothee Auer²

¹Nottingham University Hospitals (NUH), Nottingham, UK; ²NUH, UK

Cerebrovascular reserve (CVR) measurements have been shown to be useful in predicting haemodynamic impairment and risk of future ischaemic events in patients with carotid artery disease. The aim of this study was to investigate the CVR of patients with carotid artery occlusion in relation to the presence of recurrent symptoms using hypercapnia fMRI. 13 patients with symptomatic carotid artery occlusion were scanned(1.5T). The degree of side-to-side asymmetry in the brain was calculated using the Asymmetry Index for all patients. In patients with recurrent symptomatic episodes we found a significantly increased asymmetry with ~18% ipsilateral reduced CVR.

                  1964.     Q-Ball MR Imaging of Longitudinal Brain Rewiring During Functional Recovery After Ischemic Stroke

Cristina Granziera¹, Thomas Benner², Dave Tuch², Aneesh Singhal³, Walter Koroshetz³, Gunnar Krueger, Gregory Alma Sorensen²

¹CHUV, Lausanne, Switzerland; ²Martinos Center-MGH-Harvard medical school, Charlestown, Massachusetts, USA; ³MGH, Boston, Massachusetts, USA

Investigating connectivity changes underlying post-stroke recovery is a challenging research’s field, with a potential future impact in treatment options and rehabilitation. We applied Q-ball MRI to monitor white matter plasticity in patients recovering from small MCA strokes. Our results showed a decrease in fiber trajectories at 1 month after stroke followed by a relative increase at 6 months in functional tracts involved in patients’ symptoms. This phenomenon could be due to masking/unmasking of the dominant ODF (edema, cellular/connectivity necrosis or damage) or to initial axonal degeneration followed by regenerative phenomena. Q-ball MRI appears a valuable method to  establish new therapeutic targets in stroke rehabilitation. Larger cohort studies are needed to confirm these preliminary findings.

Clinical Imaging with DTI & fMRI

Takehiko Takagi¹, ², Masaya Nakamura¹, Masayuki Yamada², Keigo Hikishima³, Suketaka Momoshima¹, Kanehiro Fujiyoshi¹, Hirotaka James Okano¹, Yoshiaki Toyama¹, Hideyuki Okano¹

¹Keio University School of Medicine, Shinjuku-ku, Japan; ²Fujita Health University School of Health Sciences, Toyoake, Japan; ³Keio University Center of Integrated Medical Research, Shinjuku-ku, Japan

We tried to determine whether noninvasive diffusion tensor tractography (DTT) could be used to track the peripheral nerve and whether recovery from contusion would show up as in ability to track fibers distal to the lesion site, correlating with histological and functional recovery. Diffusion tensor MRI data and fiber tracking were analyzed for peripheral nerve injury. We also measured fractional anisotropy (FA) values, histological and functional parameters. FA values reflected histological and functional changes, demonstrating the possible contribution of DTT to the evaluation of the some clinical events for peripheral nerve degeneration and regeneration.

                  1966.     Functional MRI Detection of Acute and Chronic Brain Plasticity Following Median and Ulnar Nerve
                               Transection Using Direct Nerve Stimulation at 9.4T

Rupeng Li¹, Seth R. Jones¹, Christopher Pawela¹, Daniel Lee Shefchik¹