TRADITIONAL POSTERS |
Hall D Wednesday 13:30-15:30 yuichi yamashita1, Takao Yamamoto, Syuhei Takemoto, Ayako Ninomiya, Ikuo Aoki, Masao Yui 1Toshiba Medical Systems Corporation, Tokyo, Japan Adamkiewicz artery is am important artery to identify before the surgery in patients with aortic disease. Present studies reported that CTA and CE MRA show visualization of the Adamkiewicz artery. An arterial spin labeling, time-spatial labeling inversion pulse (time-SLIP) allows visualization of marked blood vessels without use of contrast medium. Combination of time-SLIP with 3D balanced SSFP (bSSFP) allows depiction of marked blood traveled within an inversion time (TI) range with good contrast. In this study, we have investigated the technique to depict the Adamkiewicz artery by adjusting TI and the position of a free-hand tag to mark the vessel. Isabelle Parienty1 1Bois de Verrières Medical Imaging Center, Antony, France Renal Artery Stenosis (RAS) is the major cause of renovascular hypertension and an important cause of chronic renal insufficiency and end-stage renal disease. Until recently, Contrast Enhanced MRA (CE-MRA) was considered a non-invasive and totally safe technique for the evaluation of RAS. The recent link between gadolinium and nephrogenic systemic fibrosis (NSF) in patient with renal insufficiency ranked down CE-MRA to the least invasive technique with a considerable risk for patients with RAS. In this work we evaluated the clinical usefulness of a totally safe new commercially available non-enhanced MRA technique, Time-SLIP, in the screening of RAS.
919.
Automatic Detection of Systolic and Diastolic Phase for NATIVE
Alto Stemmer1, Peter Schmitt1, Berthold Kiefer1 1Siemens Medical Solutions, Erlangen, Germany Blood flow synchronized 3D turbo spin echo imaging can be used for non contrast enhanced peripheral MRA. The optimal trigger times are often determined in a 2D scout scan, which acquires multiple single shot images in various cardiac phases. Here we introduce a technique that evaluates the scout images by means of image processing methods. 920. Non-Contrast Pulmonary Vein Angiography Using Off-Resonance RF Excitation Reza Nezafat1, Christian Stoeck1, Prity Bengani2, Dana C. Peters1, Thomas Hauser1, Neil M. Rofsky1, Warren J. Manning1 1Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, Massachusetts, USA; 2Boston Univeristy, Boston, Massachusetts, USA Atrial fibrillation is a disorder afflicting over 2 million Americans with higher likelihood occurrence with increasing age. Radiofrequency pulmonary veins isolation is commonly used to treat patients with AF. For these patients evaluation of PV anatomy is essential in pre-procedural planning and post-procedural assessment of this treatment. Today, contrast-enhanced MR angiography is commonly used to evaluate PVs in RF ablation. In today’s clinical practice, PV image acquisition is performed during first pass injection of gadolinium contrast or with computed tomography and iodinated contrast. However, contrast enhanced approaches have adverse safety implications for patients with renal insufficiency. In this study, we investigated the use of off-resonance RF excitation demonstrate the blood in the pulmonary veins without exogenous contrast media. Yasuo Amano1, Katsuya Takahama1, Shinichiro Kumita1 1Nippon Medical School, Tokyo, Japan The purpose of this study was to evaluate the usefulness of free-breathing non-contrast-enhanced 3D SSFP MR angiography for the detection of the thoracic aortic diseases and simultaneous visualization of internal thoracic and coronary arteries. The 3D SSFP imaging was combined with cardiac and navigator-gating techniques, and T2-prepared and fat-suppression pulses. Coronal contrast-enhanced 3D MR angiography was performed for comparison. High correlation was observed for diameter of the thoracic aorta between the two MR angiography techniques. Accurate diagnoses of the diseases were made in most patients. Internal thoracic and coronary arteries were visualized better by free-breathing non-contrast-enhanced 3D MR angiography. 922. STAR and STARBURST for Combined Flow Dependent and Flow Independent Carotid MR Angiography Ioannis Koktzoglou1, 2, Robert R. Edelman1, 2 1Evanston Northwestern Healthcare, Evanston, Illinois, USA; 2Feinberg School of Medicine, Northwestern University, Chicago, Illinois, USA In the evaluation of extracranial carotid artery disease, non-contrast MR angiography must be improved to compete with contrast-enhanced examinations. We present the combination STAR and STARBURST, flow dependent and flow independent sequences capable of generating high quality carotid MR angiograms without the need for contrast material. Isabelle Parienty1 1Bois de Verrieres Medical Imaging Center, Antony, France CT Angiography and contrast-enhanced MR Angiography (CE-MRA) techniques have become widely established and are nowadays used in routine clinical exploration of peripheral, lower extremities and renal arteries. However, in-patient with renal dysfunction, there is a growing amount of evidence against the systematic use of gadolinium-based CE-MRA as it has been linked to the development of nephrogenic systemic fibrosis. In this work, we investigated the reliability of the commercially available non-contrast enhanced MRA technique: Fresh blood imaging (FBI) in our clinical setting to explore patients with obstructive vascular disease. Randall B. Stafford1, 2, Mohammad Sabati1, 2, Michael J. Haakstad2, Houman Mahallati1, 2, Richard Frayne1, 2 1University of Calgary, Calgary, Canada; 2Foothills Medical Centre, Calgary Health Region, Calgary, Canada Huang et al. showed that the balanced steady-state free precession Dixon method provides good fat-water separation. Separation is achieved by adjusting the centre frequency to collect two images where fat and water are in-phase and opposed-phase, respectively. Adding these two images results in a fat-suppressed water-only image. We hypothesize that this technique can perform fast 3D non-contrast enhanced (NCE) MR angiography (MRA). We collected 3D image volumes in the legs of five healthy volunteers using this method. Good vessel conspicuity was found in our water-only image volumes. We conclude that this technique has the potential for non-contrast-enhanced MR angiography. Zheng Chang1, Qing-San Xiang2, Jim Ji3 1Duke University, Durham, USA; 2University of British Columbia, Vancouver, Canada; 3Texas A&M University, College Station, Texas, USA The efficient multiple acquisition technique using shared spatial information by Skipped Phase Encoding and Edge Deghosting (SPEED) has been successfully demonstrated in water-fat imaging with double acquisitions. In this work, the technique is further used to accelerate phase-contrast (PC) MRA. Multiple acquisitions in PC-MRA share similar spatial information. By making use of the shared information, PC-MRA is accelerated by SPEED with factors greater than that achievable for a single acquisition. In this work, a PC-MRA study with two acquisitions is accelerated by a factor of nearly 2, reducing scan time to that of about one acquisition. 926. Axial 2D TOF-Venography with Continuously Moving Table Acquisitions Sandra Huff1, Dominik Paul1, Michael Markl1, Ute Ludwig1 1University Hospital Freiburg, Freiburg, Germany Time-of-Flight MR-Angiography (TOF-MRA) allows for vessel imaging without the application of contrast agents. A difference method for TOF-MRA of an extended Field of View is presented while the patient table is continuously moved during data acquisition. The peripheral veins from the bifurcation to the feet were imaged using the proposed acquisition technique. Furthermore an optimal flip angle was determined resulting in a high venous signal for the chosen sequence parameters. 927. 3D Dark Blood MR Angiography of the Thoracic Vessels Yiu-Cho Chung1, Stephen Cook, Jaeseok Park2, Marshall Winner, Renate Jerecic, Orlando Simonetti 1Siemens Medical Solutions USA, Inc., Columbus, USA; 2Siemens Medical Solutions, Germany Followup of patients with aortopathy and congenital heart diseases post-surgery is difficult with current 3D MR angiographic techniques because of their sensitivity to susceptibility artifacts from metallic stents commonly found in these patients. The conventional 2D turbo spin echo techniques suffer from partial volume effects, long scan time and difficult slice positioning in patients with congenital diseases. We propose a new 3D turbo spin echo technique that provides isotropic, high resolution dark blood thoracic MR angiography. The technique has reduced susceptibility artifacts and was found to improve visualization of vessel patency in patients with aortic stents. 928. Non-Contrast, Free Breathing 3D SSFP MR Angiography of Pulmonary Veins: Initial Experience Aparna Singhal1, Anderanik Tomasian1, ALex Sassani2, Vibhas Deshpande2, Gerhard Laub2, J. Paul Finn2, Stefan Ruehm2, Mayil S. Krishnam3 1UCLA, Los Angeles, USA; 2UCLA, USA; 3UCLA, Los ANgeles, USA We sought to investigate the feasibility of non-contrast, free breathing 3D-SSFP MR Angiography for pulmonary veins evaluation and to correlate with conventional contrast-enhanced MRA. 3D-SSFP MRA provides sufficient vascular delineation and SNR, CNR to support confident evaluation of pulmonary veins. 929. A Method for Removing Fluid-Bearing Voxels from STARBURST MR Angiographic Images Ioannis Koktzoglou1, 2, Robert R. Edelman1, 2 1Evanston Northwestern Healthcare, Evanston, Illinois, USA; 2Feinberg School of Medicine, Northwestern University, Chicago, Illinois, USA STARBURST is a newly proposed method for creating high quality flow-independent angiograms. When acquired with a trueFISP image readout, the STARBURST technique enhances signals from fluids that may obscure view of the angiogram. We present a method for identifying and automatically removing fluid-bearing voxels. 930. Fast Spin-Labeled Projectional Carotid MR Angiography Ioannis Koktzoglou1, 2, Robert R. Edelman1, 2 1Evanston Northwestern Healthcare, Evanston, Illinois, USA; 2Feinberg School of Medicine, Northwestern University, Chicago, Illinois, USA Non-contrast methods for angiography may reduce procedural cost and risk of adverse reactions to contrast material. We present an arterial spin-labeled technique that allows for projectional carotid MR angiography in twenty seconds. 931. Investigation of Background Suppression Strategies in Interactive Fresh Blood Imaging Pauline Wong1, Martin John Graves1, David John Lomas1 1University of Cambridge and Addenbrooke's Hospital, Cambridge, UK Background suppression in subtraction MR angiography can be a problem when the TR of two acquisitions are different, causing unequal tissue signal which results in an angiogram with high background signal. This is encountered in interactive FBI where triggering occurs alternately in systole and diastole at short TR. Five tissue suppression strategies are investigated to improve the flow contrast in this case: Magnetisation Transfer (MT) prep, Projection dephaser, Inversion Recovery prep, Variable Refocusing Flip Train and single-triggered Fixed TR. Preliminary results suggest that MT, IR, variable flip and Fixed TR can be optimised for background suppression in interactive subtraction angiography.
932.
“Inflow“ Renal MR-Angiography with Steady-State Free-Precession and
Slab-Selective Spin Inversion: Marcus Katoh1, Jan Weidner2, Arno Buecker1, Matthias Stuber3, Rolf W. Gunther2, Elmar Spuentrup2 1University Hospital Saarland, Homburg, Germany; 2RWTH Aachen University Hospital, Aachen, Germany; 3Johns Hopkins University Medical School, Baltimore, Maryland, USA The reliability of a previously introduced inversion-prepared renal MRA technique was evaluated in patients. 103 renal arteries in 45 patients were investigated on a 1.5 T whole-body MR-system using a free-breathing navigator-gated and cardiac-triggered 3D steady-state free-precession sequence with a preceding slab-selective inversion pulse (inflow-MRA). For comparison contrast-enhanced MRA (ce-MRA) was performed. Images were evaluated in terms of artifacts and visualization of the renal arteries. In addition, the stenosis grade was assessed. Inflow-MRA yielded sonsistently good image quality and showed excellent correlation to ce-MRA with respect to the assessment of the stenosis grade without the need for contrast medium application or breath-hold.
933.
Non Contrast Enhanced MRA of the Lower Extremities Using an ECG-Gated
Variable Flip Angle 3D Fast Ruth P. Lim1, Andrew D. Hardie1, Elizabeth M. Hecht1, Danny C. Kim1, Jian Xu2, Pippa Storey1, Thomas P. Mulholland1, Sooah Kim1, James S. Babb1, Vivian S. Lee1 1NYU Medical Center, New York, USA; 2Siemens Medical Solutions, New York, USA The association of Nephrogenic Systemic Fibrosis with gadolinium chelates has increased interest in the development of clinically viable non contrast MRA techniques. We describe our initial clinical experience with an ECG-gated variable flip angle fast spin echo technique to image the lower limb vasculature. This has the advantage of shorter interecho spacing and a shorter acquisition window compared with a previously described ECG-gated HASTE MRA technique, with improved spatial resolution and timing of systolic and diastolic acquisitions. The technique is feasible with a high negative predictive value, however further optimization includes addressing B1 inhomogeneity. 934. Comparison of Image Reconstruction Algorithms for the Depiction of Vessel Anatomy in PC VIPR Datasets Ashley Gould Anderson1, Kevin M. Johnson1, Jelena Bock2, Michael Markl2, Oliver Wieben1 1University of Wisconsin, Madison, Wisconsin, USA; 2University Hospital Freiburg, Freiburg, Germany This study compares eight algorithms for the generation of angiograms from cine phase contrast VIPR imaging, a truly 3D radial trajectory with radial undersampling. The motivation for this investigation is based on the renewed interest in non-contrast enhanced MRA methods and the need for accurate depiction of vessel anatomy and boundaries for clinical evaluations and for the direct derivation of hemodynamic parameters such as wall shear stress and trans-stenotic pressure gradients from the velocity vector fields. |
Hall D Wednesday 13:30-15:30 Peter D. Gatehouse1, Jennifer Keegan, Ricardo Wage, David N. Firmin 1Royal Brompton Hospital, London, UK Aiming to avoid the need for stabilisation cycles, real-time cine bSSFP imaging was modified to slide the image slice continuously along the long axis of the heart. Cardiac function measurements and blood-myocardium contrast were compared against real-time as normally used with stabilisation on alternate cycles. The sliding-slice continuous cine enables more rapid volume acquisition at reduced blood-to-myocardium contrast.
977.
Evaluation of Right Ventricular Function and Pulmonary Perfusion in
Ross Procedure Patients Using Anna Lakoma1, David Tuite1, John Sheehan1, Peter Weale, James Carr1 1Northwestern University, Chicago, Illinois, USA The purpose of this retrospective study was to evaluate whether valve dynamics and pulmonary perfusion parameters as measured by MRI can predict right heart dysfunction in patients post Ross procedure. 978. TGRAPPA Accelerated Free Breathing Real-Time Cine Cardiac Imaging with a 32-Channel Coil Neil Isaac1, Sven Zuehlsdorff2, Peter Weale2, Sonia Nielles-Vallespin3, Renate Jerecic2, Harold Litt1 1University of Pennsylvania, Philadelphia, Pennsylvania, USA; 2Siemens Medical Solutions, Chicago, Illinois, USA; 3Siemens Medical Solutions, Erlangen, Germany A comparison of TGRAPPA accelerated real time cine images of the heart using a 32 channel phased array coil, in comparison with traditional segmented k-space acquisitions using a 32 channel coil, as well as traditional 8 channel phased array coil. 979. Increased Diastolic Pressure Gradients Are Measured During Dobutamine Stress Tests June Cheng Baron1, Ian Paterson1, Mark Haykowsky1, John Mackey1, Richard Thompson1 1University of Alberta, Edmonton, Canada Measurement of blood pressure gradients (ÄP) between the atrium and ventricle during diastole provides insight into the ability of the ventricle to facilitate filling by rapid relaxation. We examine, for the first time, changes in ÄP during a dobutamine stress test to characterize stress-related increases in these gradients. In 8 subjects, heart rates increase from 58-77 (rest) to 106-160bpm at peak stress, with corresponding increases in filling pressure gradients from 2.4±0.5mmHg (rest) to 4.8±1.8mmHg at peak stress. We attribute increases in ÄP to increased diastolic suction and not pre-load, since end-diastolic volumes drop with stress by 35±16ml. 980. Volumetric Cardiac Quantification Using Three Dimensional Dual Phase Whole Heart MRI Sergio Andres Uribe1, Tarinee Tangcharoen1, Victoria Parish1, Ivo Wolf2, Reza Razavi1, Gerald Greil1, Tobias Schaeffter1 1Kings College London, London, UK; 2Deutsches Krebsforschungszentrum, Heidelberg, Germany In this abstract we propose a new acquisition that allows precise cardiac volume quantification. The technique acquires isotropic data sets of the whole heart during systole and diastole in a single free breathing scan using independent navigators for each cardiac phase. Quantification of LV and RV cardiac volumes is performed using a semi-automatic segmentation tool. A comparison between the standard Simpson approach and the proposed method shows a good agreement for all volunteers and patients. 981. Inflow Quantification in 3D Cardiac MR: Implications for Whole Heart Coronary Imaging and 3D Cine Reza Nezafat1, Daniel A. Herzka2, Christian Stehning3, Dana C. Peters1, Anne Riley4, Kay Nehrke3, Warren J. Manning1 1Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, Massachusetts, USA; 2Clinical Sites Research Program, Philips Research North America, Briarcliff Manor, New York, USA; 3Tomographic Imaging, Philips Research Europe, Hamburg, Germany; 4Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, USA Large volume 3D cardiac MR imaging (CMR) is becoming more appealing with applications both in 3D cardiac cine and whole heart coronary imaging. The transition from single 2D slices or small 3D slabs to thicker 3D slabs theoretiecally could result in an increased SNR and better image registration. However, for CMR there is an associated penalty that results from saturation of blood flowing into the imaging volume which can affect blood SNR and blood-myocardium CNR. In this study, we investigated inflow enhancement (or lack thereof) in 3D single-phase imaging (for coronary applications) as well as cardiac cines (for function evaluation). Chun Xu1, Lawrence Dougherty2, Gamaliel Isaac2, Aaron Blom1, Joseph H. Gorman1, Robert C. Gorman1, James J. Pilla1 1University of Pennsylvania, Glennolden, Pennsylvania, USA; 2University of Pennsylvania, Philadelphia, Pennsylvania, USA The aim of this work is to investigate the use of a 3-D MR tissue-tagging imaging sequence combined with a motion estimation approach using a 3-D Optical Flow Method (OFM) for the quantification of left ventricular (LV) mechanics. Utilizing imaging and OFM parameters optimized in a previous simulation study, this method was applied to a healthy volunteer, from which 3-D displacement fields were derived and used to quantify LV contractile functions. Regional heterogeneity of 3-D displacement and strain was characterized in high spatial resolution through out the human LV, with significantly shorter acquisition and analysis time compared with other methods
983.
Association Between Elevated Fasting Glucose, in the Absence of
Diabetes, and Increased Left Michael L. Chuang1, Philimon Gona2, 3, Carol J. Salton1, Noriko Oyama1, Susan J. Blease2, Daniel Levy2, Christopher J. O'Donnell2, Warren J. Manning1 1Beth Israel Deaconess Medical Center, Boston, Massachusetts, USA; 2the NHLBI's Framingham Heart Study, Framingham, Massachusetts, USA; 3Boston University, Boston, Massachusetts, USA Elevated fasting glucose, in the absence of diabetes, is associated with increased left ventricular (LV) mass in adult women but not men. There were no differences in LV geometry or ejection fraction between subjects with normal or elevated fasting glucose for either women or men.
984.
Automated Segmentation of the Left Ventricle Using Myocardial
Effusion Threshold Reduction and Noel Christopher Codella1, Jonathan W. Weinsaft1, Matthew D. Cham1, Matt Janik1, Martin Prince1, Yi Wang1 1Weill Medical College of Cornell University, New York, New York, USA An automated left ventricular segmentation algorithm for cine balanced SSFP images that accounts for partial voxel effects through linear interpolation between the signals of blood and myocardium is presented. 985. Motion-Corrected Strain Calculation from Long-Axis MRI Strain-Encoded (SENC) Images Ahmed Amr Harouni1, Khaled Z. Abd-Elmoniem1, Monda L. Shehata1, David A. Bluemke1, Nael F. Osman1 1Johns Hopkins University, Baltimore, Maryland, USA Strain measured from Strain-Encoded (SENC) images are currently extracted from fixed regions, which introduce large error especially in the LV freewall. In this work, we developed a novel algorithm to automatically detect and track the motion of heart wall in long-axis SENC images. We showed that strain values calculated from our algorithm are acceptable and close to strain values calculated from the manual segmentation. Daniel B. Ennis1, J Andrew Derbyshire2 1Stanford University, Stanford, California , USA; 2National Institutes of Health, Bethesda, Maryland, USA We first define a simplified 4x4 matrix expression that describes the mathematics of an MRI displacement encoding experiment using homogenous coordinates. This expression is then used to characterize the effect of imaging noise on the strain calculated from simulated displacement measurements. It is shown that the error in estimating the strain is non-linearly related to both the image SNR and the imaging resolution. The highest strain errors occur for combined low SNR and high image resolution. Increases in imaging resolution (decreasing pixel size) can increase the error in the strain estimate even when SNR is held constant. 987. Tracking Motion in TMRI Data Using Binary Image Processing Techniques Tareq Alrefae1, Mohammed D. Alenezy2, Elena I. Popel2, Mehmet Bilgen3 1Kuwait University, Kuwait City, Kuwait; 2University of Kansas, Lawrence, Kansas, USA; 3Medical University of South Carolina, Charleston, South Carolina, USA Tagged magnetic resonance imaging (tMRI) has found wide-spread applications in various clinical and research areas. The strength of this technique lies in its ability to reveal regional motion of tissue such as that observed in cardiac studies. For purposes of quantification, offline applied algorithms are developed to track the tissue motion of user-selected regions of interest (ROI). Examples of such algorithms include harmonic phase (HARP) techniques and others. To further enrich the library of tracking algorithms, we present an automated method that utilizes binary image processing techniques to follow the tissue motion of user-selected ROI in tMRI data. Tareq Alrefae1, Mehmet Bilgen2 1Kuwait University, Kuwait, Kuwait; 2Medical University of South Carolina, Charleston, USA Application of tMRI combined with sensitive motion estimation techniques, such as harmonic phase (HARP) analysis, has proven to be feasible and diagnostically valuable in evaluating the performance of normal or diseased hearts in live subjects using conventional global and regional measures. In the presence of large motions, however, the current analysis techniques including HARP fail to accurately describe the absolute displacement of the myocardial tissue. This paper addresses this issue and offers a solution - a simple time-dependent model and its time-reversal- providing motion estimates with improved performance even when the tissue is subject to large movements. 989. Improved Cardiac Strain Estimation from DENSE Using Automatic Outlier Rejection Sandeep Narendra Gupta1, Anthony Aletras2, Maureen N. Hood3, Vincent B. Ho3, Ehud J. Schmidt4, Pelin Aksit5 1GE Global Research, Niskayuna, New York, USA; 2NIH, Bethesda, Maryland, USA; 3NNMC and USUHS, Bethesda, Maryland, USA; 4GE Healthcare, Boston, Massachusetts, USA; 5GE Healthcare, Bethesda, Maryland, USA DENSE technique allows quantification of regional myocardial strain. We present here a novel method for processing of DENSE data which automatically detects and rejects pixels with poorly conditioned strain calculation. With this outlier rejection method, a three-fold reduction in the variability of circumferential shortening and radial thickening quantification is demonstrated in human studies.
990.
A Premature Ventricular Contraction Detection Method for Cardiac MR
Acquisitions Using Morphology Liewei Sha1 1GE Healthcare, waukesha, Wisconsin, USA Premature Ventricular Contractions are ectopic heart beats with ventricular origin. The presence of PVCs degrades image quality, if the gated scans do not differentiate PVCs from normal beats. Accurate detection of PVC not only improves image, but also provides auxiliary information to facilitate diagnosis of ventricular arrhythmias. Many PVC detection algorithms have been proposed [1,2], few consider the noise in MR environment,including distortion due to magnetohydrodynamic effect, gradient noise, and radio-frequency interference. This work proposed a method to detect R-wave and PVC, using combined feature of MR noise resistant 2D VCG morphological distance, the R-peak duration and RR interval.
991. Relation of Myocardial Fiber Structure with Cardiac Wall Motion Using DTI and MR Tagging Yin Wu1, 2, Tracy Yee Chow1, 2, Ke Xia Cai1, 2, Ed X. Wu1, 2 1Laboratory of Biomedical Imaging and Signal Processing, The University of Hong Kong, Pokfulam, Hong Kong; 2Dept. of Electrical and Electronic Engineering, The University of Hong Kong, Pokfulam, Hong Kong LV structure is known to be associated with cardiac function, but the relation of myocardial fiber distribution with regional wall motion remains to be elucidated. In this study, both tagging and DTI were performed in adult rats. LV myocardium twist, circumferential strain, and myocardial fiber architecture were investigated. Results show that the myocardial fiber distribution has direct relation with the LV myocardium twist angle and magnitude of circumferential strain. Such integrated functional and structural analysis may provide more information for understanding the fundamental cardiac mechanics and assessment of pathological changes. 992. Model-Based Estimation of 3D Myocardial Motion Based on Cine DENSE MRI Patrick Helm1, Frederick Epstein1 1University of Virginia, Charlottesville, Virginia , USA Multi-slice cine DENSE can acquire displacement-encoded images representing the 3D motion of the entire left ventricle with high spatial resolution and accuracy relative to conventional myocardial tagging, but still with finite signal-to-noise ratio and discrete spatiotemporal sampling. We developed a technique where a patient-specific cardiac model was employed to improve DENSE estimates of the continuous spatiotemporal displacement field throughout the 3D left ventricle. The model-based reconstruction of cardiac motion agreed closely with tagging when undersampled to similar spatial resolution. The higher spatial resolution of DENSE enabled the use of fewer model constraints and potentially improved estimation of 3D motion. 993. Epicardial Contour Extraction by LV-Metric and Active Contour Model Using Cardiac MRI Hae-Yeoun Lee1, Yi Wang1 1Cornell University, New York, USA In this study, an automatic left ventricle (LV) segmentation algorithm using short-axis cine cardiac MRI is presented. To segment LV and detect endocardial contour, we compensate the coil sensitivity of magnitude images and apply region-growing scheme, termed LV-METRIC. To detect epicardial contour and segment myocardium (MC), we generate a circular map by polar mapping, extract and filter edge information using segmented LV. After defining the external force with prior knowledge, we apply guided active contour model to find epicardial contour, where initial contour is the endocardial contour from LV and only moves to radius direction on the circular map. 994. One Touch Imaging for Improved Cardiac Workflow Robert D. Darrow1, Vivek Vaidya2, Ambey Govenkar3, Rakesh Mullick2, Thomas K.F. Foo4 1GE Global Research, Niskayuna, USA; 2GE Global Research, Bangalore, India; 3Extenprise, Inc., Pune, India; 4GE Global Research, Niskayuna, USA Cardiac examinations demand highly proficient technologists or operators to complete clinical studies in an efficacious and time-efficient manner. One Touch imaging minimizes operator impact on the quality of the cardiac examination and produces highly repeatable examinations. In addition, by reducing the number of steps needed to identify the cardiac scan planes to a single button push, a simplified cardiac exam is demonstrated with substantial timesavings. We demonstrate a seamlessly integrated image acquisition/feature recognition method that produces short-axis cardiac CINE images with a single button push in less than 10 minutes, including breath-holding time. |
Hall D Wednesday 13:30-15:30 Arteries: Overview of Phase III Clinical Trials Gilles Soulez1, Siegfried Thurnher2, Nicoletta Anzalone3, Gianpaolo Pirovano4 1University of Montreal, Montreal, Canada; 2University of Vienna, Vienna, Austria; 3San Raffaele Hospital, Milan, Italy; 4Bracco Diagnostics Inc., Princeton, New Jersey, USA Results of clinical trial evaluating Gd-BOPTA for CE-MRA are presented. In the carotid arteries, blinded readers noted significantly (p<0.001) increased specificity and accuracy compared to 2D-TOF MRA. The sensitivity, specificity, and accuracy of CE-MRA for renal artery stenoses ranged between 60–84%, 89–95%, and 80–87%, respectively. For peripheral MRA, the sensitivity (54–81%), specificity (90%–95%), and accuracy (85%–88%) of CE-MRA for detection of significant iliofemoral disease was significantly (p<0.001) better than TOF MRA (33–63%, 74–89% and 68–77%, respectively). Gd-BOPTA was safe and accurate for CE-MRI at a dose of 0.1 mmol/kg. 1033. Low Dose, Contrast-Enhanced MR Angiography of the Lower Extremities at 3.0 Tesla Reza Habibi1, Derek G. Lohan1, Mayil S. Krishnam1, Fatemeh Barkhordarian1, Mehdi Jalili1, Roya Saleh1, Stefan G. Ruehm1, John Paul Finn1 1UCLA, Los Angeles, USA Even though the exact correlation between the gadolinium dosage and occurrence of nephrogenic systemic fibrosis (NSF) in patients with low GFR remains to be established, it seems reasonable to minimize dose in susceptible patient groups, awaiting further clarification. The purpose of this study was to assess the feasibility of low-dose CE-MRA of the lower extremities at 3.0 Tesla and to compare the image quality with high-dose and intermediate-dose protocols. The results of our study showed that, taking advantage of the higher SNR at 3.0T, contrast dose for lower extremity MRA can be reduced several fold without any compromise in image quality. 1034. Targeted Contrast Enhancement Using Linear System Theory Daniel Kopeinigg1, 2, Dominik Fleischmann1, Matus Straka1, Rudolf Stollberger2, Roland Bammer1 1Stanford University, Stanford, California , USA; 2Medical University Graz, Graz, Austria Correct timing of the contrast bolus arrival is a crucial part in CE-MRA. The inability to play continuous waveforms in clinically used power injectors is also an issue in achieving a constant vessel target enhancement. FORWARD was developed to account for this limitation, as well as of physiological limitations like maximum injections flows, among others. In the FORWARD process, characteristics of the test bolus are used to tailor the injection profile to create a more constant vessel enhancement. The possibility to analyze the contrast agent bolus over time and the corresponding dispersion leads to new optimization methods for k-space sampling.
1035.
Three-Dimensional High Spatial Resolution Magnetic Resonance
Angiography of the Supra-Aortic Arteries Anderanik Tomasian1, Noriko Salamon, Derek Lohan, Mayil Krishnam, J. Pablo Villablanca, J. Paul Finn 1UCLA, Los Angeles, USA The purpose of our study was to prospectively establish the non-inferiority of diminished dose regimens compared to higher dose regimens as reflected in the diagnostic image quality of high-spatial-resolution three-dimensional MR angiography of the supra-aortic arteries at 3.0 Tesla.High spatial resolution CE-MRA of the supra-aortic arteries at 3.0T can be performed with a gadolinium dose at least as low as 0.05 mmol/kg, without compromising image quality compared to 0.1 mmol/kg and 0.15 mmol/kg. Although further work is warranted, these initial results suggest that adoption of low-dose protocols in clinical practice may diminish sensitivity to contrast dose-dependent complications and result in cost savings. |
Hall D Wednesday 13:30-15:30 Carl Jason Snyder1, Lance DelaBarre1, Greg Metzger1, Can Akgun1, Patrick Bolan1, Kamil Ugurbil1, Thomas Vaughan1 1University of Minnesota, Minneapolis, USA It is known that using separate volume transmit coil with local phased array receivers has benefits over using either coil separately. Unfortunately, most body imaging at 7T currently uses surface transceive arrays. We have designed an 8 channel TEM transmit array to be used in conjunction with a 16-channel loop receiver array and a B1+ shimming algorithm for body imaging at 7T. This combination provides a relatively homogeneous volume over a localized region in the body. 1066. Rotating Transmission Line Elements for Optimized Parallel Imaging Zhangwei Wang1, Michael Craig1, Ye Li2, Qing X. Yang1, Christopher M. Collins1, Wei Chen2 1Penn State University, Hershey, Pennsylvania, USA; 2University of Minnesota, Minneapolis, Minnesota, USA For transmit arrays, independent amplitude and phase control can be used in a variety of ways achieve a homogeneous excitation distribution. However, for a receive array manipulation of sensitivity distributions and the g-factor require variation of the coil geometry. Stemming from work with stripline transmission line elements, we propose a new type of transmission line element which can be used to optimize the field distribution and g-factor easily by adjusting the orientation of independent coils. 1067. Shielded Quadrature Coil Elements for NMR Phased-Arrays Matthias F. Mueller1, 2, Titus Lanz2, Tobias Wichmann2, 3, Felix A. Breuer3, Mark A. Griswold4, Peter M. Jakob1 1University of Wuerzburg, Wuerzburg, Germany; 2Rapid Biomedical GmbH, Rimpar, Germany; 3Research Center Magnetic Resonance Bavaria (MRB), Wuerzburg, Germany; 4University Hospitals of Cleveland, Cleveland, Ohio, USA Recently MR systems with 128 independent receiver channels have been developed. With these, phased-arrays with up to 128 coil elements have been introduced. While these arrays have provided impressive results, the minimal size of elements is limited if maintaining sample noise dominance is required. Thus, only a fraction of the receive channels can be used to full potential for imaging of more localized region-of-interests. To this end, the concept of shielded quadrature array coil elements is introduced which utilizes additional, perpendicular coil elements to improve signal-to-noise ratio and encoding capability for parallel imaging. 1068. A Novel Decoupling Technique for Non-Overlapped Microstrip Array Coil at 7T MR Imaging Zhentian Xie1, Xiaoliang Zhang1, 2 1UC San Francisco, San Francisco, California , USA; 2UC Berkeley Joint Graduate Group in Bioengineering, San Francisco & Berkeley, California , USA . In this work, a new magnetic wall decoupling method for designing non-overlapped coil arrays is introduced by using microstrip lines. The micsrotrip lines are the magnetic-wall generators, laced between two adjacent resonant coil elements. There is no physical connection between the decoupling microstrip line and coil elements. Theoretically this method is not frequency-sensitive, therefore the decoupling performance can be maintained in a broad frequency range. Bench test and preliminary imaging results are shown, using the proposed coil array at 7T. The MR images from each resonant element have no signal is observed from other coil, indicating the effectiveness of the proposed decoupling technique at ultrahigh field of 7T. The proposed design provides a robust approach to design of parallel imaging arrays at ultrahigh fields. 1069. 1.5-Tesla Integrated Dual Mode Hand/Wrist Array Coil for Imaging Jacob Weaver1, Joseph Herczak1, Tsinghua Zheng1, Xiaoyu Yang1, 2, Hiroyuki Fujita1, 2 1Quality Electrodynamics, LLC., Mayfield Village, USA; 2Case Western Reserve University, Cleveland, USA A six-channel dual-mode array coil for hand/wrist MRI imaging at 1.5 Tesla was constructed and tested. The coil offers two variable FOV modes, a full mode with an S-I coverage of 20cm for the entire hand/wrist imaging and a magnifying FOV mode covering 10cm in S-I for high resolution wrist only imaging. The high resolution mode shows significant SNR gains in the wrist region compared to a commercially available QD wrist coil. 1070. 8 Channel Multi-Coil to Image Both Hands and Both Wrists Symmetrically for RA Diagnosis Akira Nabetani1, Harushi Mori2, Atsushi Nozaki1 1GE Yokogawa Medical Systems, Ltd., Hino, Japan; 2University of Tokyo, Bunkyo-ku, Japan 8 channel muti-coil to image both hands and both wrists symmetrically for Rheumatoid Arthritis (RA) diagnosis was designed and developed. At first patientfs posture for symmetrical imaging of both hands and both wrists in comfortable manner was considered. Then holding fixture was designed to keep such posture stably and comfortably. 8 coil element size and layout was determined for the B1 sensitivity of the multi-coil to cover the area of both hands and both wrists as homogeneously as possible. The developed holding fixture had stable and comfortable holding capability so there were no apparent motion artifacts in the images.
1071.
Twelve-Channel Receive-Only MRI Breast Coil at 3T
John A.J. de Groot1, Dennis W.J. Klomp2, Evert G.J. Beerens1 1Machnet BV, Maarn, Netherlands; 2Radboud UMC Nijmegen, Nijmegen, Netherlands In this study we show simulation results of a twelve-channel MRI breast coil at 3T and measurements with phantoms and volunteers. The system consists of an existing 4-channel breast coil and a newly designed 8-channel add-on coil. The goal was to increase acquisition speed without compromising SNR compared to the existing breast coil. High reduction factors can be achieved by keeping coil couplings low and maintain good g-factor maps. We show that the system SNR is improved significantly and that high reduction factors can be applied. Graham C. Wiggins1, Jonathan Polimeni1, Andreas Potthast2, Thomas Witzel1, Lawrence L. Wald1, 3 1Massachusetts General Hospital, Charlestown, Massachusetts, USA; 2Siemens Medical Solutions,Inc, Charlestown, Massachusetts, USA; 3Harvard-MIT Division of Health Sciences and Technology, Cambridge, Massachusetts, USA When using arrays with large numbers of small coils there is commonly a concern about the sensitivity provided by the small elements at distances farfrom the array. It is important to understand how the SNR in the center of the sample increases as a function of the number and distribution of coil elements around the head, whether in fact the SNR may decrease beyond a certain number of coils, and therefore how much benefit is gained from costly and complicated designs with 96 channels or more. We find particular benefit from the coils arranged over the dome of the head in our high-N arrays even for central SNR. Nicola De Zanche1, Jurek Antonin Nordmeyer-Massner1, David Otto Brunner1, Klaas Paul Pruessmann1 1University and ETH Zurich, Zurich, Switzerland Overlap is commonly used to minimize mutual inductance, and thereby coupling, among nearest neighbors in a coil array. In parallel imaging, however, it has been shown that non-overlapped coils provide more favorable sensitivity profiles than their overlapped counterparts. We provide data that demonstrate a clear advantage of non-overlapped over overlapped coils also in terms of noise correlation among channels when effective preamplifier decoupling provided by standard techniques is present. Riccardo Lattanzi1, 2, Aaron K. Grant2, 3, Daniel K. Sodickson4 1Harvard-MIT, Cambridge, Massachusetts, USA; 2Beth Israel Deaconess Medical Center, Boston, Massachusetts, USA; 3Harvard Medical School, Boston, Massachusetts, USA; 4New York University Medical Center, New York, New York, USA In this work we compare the SNR and current patterns of realistic coil models with the ultimate intrinsic SNR and the corresponding ideal current patterns for a homogeneous cylindrical sample. Calculations were repeated at different magnetic field strengths and for various acceleration factors, using coil arrays with increasing numbers of elements, in order to investigate how rapidly the ultimate behavior was approached in each case. In the center of the sample, 128-element arrays reached approximately 90% of the optimum SNR at 1.5T, whereas performance was lower at higher field strength due to increasing complexity in the ideal current patterns. 1075. Empirical Validation of 3T 96 Channel G-Factors and Comparison to Ultimate G-Factor Graham C. Wiggins1, Jonathan R. Polimeni1, Andreas Potthast2, Thomas Witzel1, Lawrence L. Wald1, 3 1Massachusetts General Hospital, Charlestown, Massachusetts, USA; 2Siemens Medical Solution, Inc, Charlestown, Massachusetts, USA; 3Harvard-MIT Divisions Of Health Sciences and Technology, Cambridge, Massachusetts, USA Accelerated imaging performance is one of the primary motivators for the development of receive arrays with higher and higher numbers of elements, typically characterized by the SENSE G-factor Metrics such as the maximum or the mean G-factor for a particular acceleration rate are frequently given as a key characteristic of a particular coil array. In practice, there are many factors which can influence the calculated G-factor map, including the field of view relative to object size (which determines the degree of aliasing), noise level and degree of smoothing in the sensitivity profiles, and whether regularization is used. We use time series data to validate analytical G-factor values for a 3T 96 Channel coil and compare these to the Ultimate G-factor. The 96 Channel coil appears to approach the ultimate G-factor for accelerations up to R=6 1076. A 1.5T 32-Channel Cardiac Array Coil for Coronary and Whole Heart MRI Hiroyuki Fujita1, 2, Tsinghua Zheng1, Xiaoyu Yang1, 2, Joseph Herczak1, Jacob Weaver1, Kazuya Okamoto3, Takahiro Ishihara3 1Quality Electrodynamics, LLC., Mayfield Village, USA; 2Case Western Reserve University, Cleveland, USA; 3Toshiba Medical Systems Corporation, Otawara, Japan MRI provides unparalleled soft-tissue contrast, and the temporal resolution has been much improved because of the rapid developments of high channel receiver count MRI hardware system and sophisticated parallel imaging applications. These state-of-the-art parallel imaging technologies enable effective real-time MR cardiac imaging. In this work, the authors constructed and tested a 1.5T 32-channel cardiac array, which is capable of higher acceleration factors to advance coronary and whole heart MRI. 1077. A 3D Parallel Imaging Capable Transmit and 15-Channel Receive Array Knee Coil at 3T Matthew Finnerty1, Joseph Herczak1, Tsinghua Zheng1, Jacob Weaver1, Xiaoyu Yang1, 2, Hiroyuki Fujita1, 2 1Quality Electrodynamics, LLC., Mayfield Village, USA; 2Case Western Reserve University, Cleveland, USA While knee imaging at 3T offers advantages over 1.5T in regards to greater spatial resolution and higher SNR, current commercial knee coils are susceptible to pulsation/flow artifacts since they only have one row of elements in the S-I direction. To address these challenges, a 15-channel knee coil with local 12-rung birdcage transmitter has been developed that utilizes three rows of elements in the S-I direction, which allows for 3D parallel imaging while providing high SNR and good uniformity with a larger coil inner diameter. 1078. SENSE Optimized Sixteen Element Receive Array for Cervical Spinal Cord Imaging at 3T Jerzy Bodurka1, Patrick Ledden2, Peter Bandettini1 1National Institute of Mental Health, NIH, Bethesda, Maryland, USA; 2Nova Medical Inc., Wilmington, Massachusetts, USA High quality and high resolution anatomical and functional imaging the human spinal cord remains a significant challenge in MRI. The benefits of parallel imaging with surface coil arrays have been clearly shown particularly in the brain imaging for both anatomical and fMRI studies. Here we demonstrate a custom design and build of a sixteen element receive-only surface coil array for spinal cord MRI imaging at 3 Tesla. 1079. Coil Design for Highly Accelerated 2D SENSE MRA of the Lower Legs Casey P. Johnson1, Clifton R. Haider1, Phillip J. Rossman1, Thomas C. Hulshizer1, Eric A. Borisch1, Stephen J. Riederer1 1Mayo Clinic, Rochester, Minnesota, USA Typical acceleration directions for 2D SENSE of the legs are left-right (L/R) and anterior-posterior (A/P). However, the fields of view along these directions can differ by a factor of three. In this work we describe an eight-element array in which the A/P elements are designed to have reduced depth of response compared to the L/R elements to account for the FOV differences. Compared to an eight-element array with uniformly-sized elements, the new design typically provides a 25% reduction in g-factors for acceleration factors >7. 2D SENSE CE-MRA of the lower legs at these high accelerations is of high quality. 1080. SNR and Parallel Imaging Performance of a 32-Channel Array for Human Brain Imaging at 7 Tesla Jacco A. de Zwart1, Peter van Gelderen1, Shumin Wang1, Jeff H. Duyn1 1National Institutes of Health, Bethesda, Maryland, USA Multi-channel coil arrays have led to substantial sensitivity and resolution improvements in human brain MRI. These benefits generally increase with the number of coil elements but level off at large element counts. An increased number of array elements also increases parallel imaging (PI) performance. Data from a 32-channel array at 7T were compared with a simulation and with numerically derived lower element count arrays to investigate these two aspects. SNR improves about a factor of 4 on average over the brain compared to one channel. Up to ~6-fold acceleration can be achieved with a mean PI penalty of <~20%. |
Hall D Wednesday 13:30-15:30 Jarek Wosik1, Krzysztof Nesteruk2, Maged R. Kamel1, Flora Ip1, L.-M Xie1, Alexander C. Wright3, Felix W. Wehrli3 1University of Houston, Houston, USA; 2Institute of Physiscs Polish Academy of Sciences, Warsaw, Poland; 3University of Pennsylvania Medical Center, Philadelphia, USA We report on the design and fabrication of a novel planar superconducting 3.0 T coil for MRI. We have designed and fabricated a high Q coil which is, in principle, intrinsically isolated from the entire body scanner transmit coil. The coil was made out of YBa2Cu3O7-x thin film deposited on a low-loss dielectric substrate. The design consists of two connected double-sided split-gap resonators. Bench measurements as well as numerical simulations of the coils showed more than -40 dB isolation from the transmit coil. In a uniform rf field, this coil should not require special detuning circuitry.
1108.
Optimization of SNR Via Cryogenically Cooled Radiofrequency
Coils in Hyperpolarized Noble Gas MR Warren Berger1, 2, Willaim Dominguez – Viqueira1, 2, Juan Parra – Robles1, Giles Santyr1, 2 1J.P. Robarts Research Institute, London, Canada; 2University of Western Ontario, London, Canada Operating at the low magnetic field strength provided by hyperpolarized noble gases allows significant increases in SNR to be achieved by decreasing the coil’s operating temperature. A receiver coil was constructed and used to image a Xe-129 phantom and rat ventilated with Xe-129. Experiments were performed at room temperature (300K) and at 118K (using liquid nitrogen as a coolant). The gain in SNR by operating at 118 K was measured to be 1.55 and 1.41 in the phantom and rat respectively (corresponding to theory) and are only possible because of the low noise from the sample provided by using Xe-129. 1109. HTS Volume Coil with Improved Imaging Volume S. Y. Chong1, J. G. Liu1, X. G. Zhao2, Keddy Chandran3, S. M. Yeung1, Q. Y. Ma1, E. S. Yang2 1Time Medical System, Hong Kong, Hong Kong; 2University of Hong Kong, Pokfulam, Hong Kong; 3Duke University, USA In this abstract, a multi-turn 5.5¡¨ large HTS solenoid volume coil & compact cryostat prototype with improved field-of-view (FOV) & high SNR is presented. Although HTS thin-film phased array coil has been demonstrated to increase imaging volume, coupling & isolation between multi-coil, tuning & matching element require complex operation. The result shows the possibility of future large HTS volume coil with compact cryostat for human neuro-imaging with increased FOV & SNR. Mona Julia Katharina Klein1, 2, Takahito Ono3, Masayoshi Esashi4, Jan Gerrit Korvink2 1CSEM, Neuchâtel, Nebraska, Switzerland; 2IMTEK, University of Freiburg, Freiburg, Germany; 3Tohoku University, Sendai, Japan; 4Tohoku University, Tohoku, Japan The objective of this work is the development of a MEMS-compatible process for the parallel fabrication of solenoidal micro-coils for MR microscopy of biological tissues, e.g. cells. The micro-coil is embedded within two bonded wafers and has a hollow core which enables direct micro-fluidic integration. Signal loss caused by connection to external circuitry may be minimized in this design due to the planar wafer geometry, enabling the use of well-established device bonding techniques. In this paper, we particularly focus on the mold metallization process. 1111. Noise Parameter Extraction in the Design of Low Noise Amplifiers (LNA) for MRI Barbara L. Beck1 1University of Florida, Gainesville, Florida, USA The first stage of amplification for the very small MRI signal is the Low Noise Amplifier (LNA). An engineering method commonly applied to LNA design in communications utilizes matching components that transform the source impedance to the optimum reflection coefficient of the transistor. Key parameters necessary to perform this method are listed on manufacturer data sheets, but typically for frequencies much higher than those used in MRI. This paper describes a technique to extract these key parameters from simulation, utilizes them to design an optimized LNA, and compares the LNA to one that has been configured for phased array operation. Nelly A. Volland1, Ioannis Constantinidis2, Thomas H. Mareci2, Nicholas E. Simpson2 1University of Florida, Gainesville, Florida, USA; 2University of Florida, Gainesville, USA Introduction: Development of an implantable inductively-coupled coil system for non-invasive in vivo monitoring of a bioartificial pancreas (tissue engineered construct) at 11.1T is discussed. Methods: System (implantable coil inductively coupled to external coil) was simulated and constructed. Coating and integration of implantable coil with the construct were pursued to test system in vitro. Results: Implantable coil was coated with 1-mm PDMS and cell-free and cell-containing beads were imaged. Sensitivity was shown to be over 10 times higher than sensitivity of previous studies. Conclusions: Implantable inductively coupled coil system was successfully developed and used to image bioartificial pancreas in vitro. 1113. A Dual-Tuned Quadrature Microstrip Volume Coil for 13C/1H MRI/S at 7T Zhentian Xie1, Duan Xu1, Douglas A. Kelley2, Dan B. Vigneron1, 3, Xiaoliang Zhang1, 3 1UC San Francisco, San Francisco, California , USA; 2GE Healthcare, San Francisco, California , USA; 3UCSF/UC Berkeley Joint Graduate Group in Bioengineering, San Francisco & Berkeley, California , USA We present an 8-element by 8-element dual-tuned microstrip volume coil for in vivo 13C/1H MRI/S animal studies at 7T. The design strategy of using a mix of half-wavelength resonators and ƒ8¡51/4 wavelength resonators was employed to the 300MHz and 75MHz operating frequencies simultaneously.Preliminary results from a corn oil phantom were acquired using the proposed coil at 7T. The results show that dual-tuned microstrip volume coil is feasible and efficient, provided that sufficient radial separation is maintained between the strip conductor and the ground to provide sufficient coupling between corresponding elements. The proposed design may provide a simple approach to dual-tuned volume coil design for in vivo multinuclear MR at ultrahigh fields. 1114. Dual Tuned Helmholtz Coil for Breast Cancer Imaging Anderson Nnewihe1, 2, Ernesto Staroswiecki1, Neal Bangerter1, Brian Hargreaves1 1Stanford University, Stanford, California , USA Clinical breast exams using x-ray mammography have limited specificity and sensitivity for detecting breast tumors. Additional imaging modalities such as sodium (23Na) MR have the potential to increase the sensitivity and specificity of breast cancer detection because 23Na MR can reflect the disruption of the membrane Na-K pump associated with cancer. The purpose of this study was to investigate the difference between a dual-tuned Helmholtz coil and a dual-tuned surface coil in providing high SNR and high quality B1 homogeneity for multinuclear imaging as well as perform preliminary breast imaging using the dual-tuned Helmholtz coil.
1115.
Circularly Polarized Surface Coil with a Single Port
Yoshihisa Soutome1, Hideta Habara1, Yoshitaka Bito1 1Hitachi Ltd., Kokubunji, Japan A circularly polarized surface coil with a single port has been designed and fabricated. The designed coil was constructed with two series connected loop-coils with an overlap, and its capacitors were designed to have a phase difference of 90o between the two loop currents. A phantom image of the designed coil was very similar to that of a QD surface coil. The SNR profile of the phantom image of the designed coil was identical to that of the QD surface coil. These results indicate that our coil can receive a circularly polarized B1 field without having a QD hybrid like that of the QD surface coil. 1116. Wireless Control of an Implantable Coil System for MRI/S Barbara L. Beck1, Brian S. Letzen1, Rizwan Bashirullah1, Thomas H. Mareci1 1University of Florida, Gainesville, Florida, USA A proposed solution for monitoring a bioartificial pancreas is a microfabricated wirelessly-controlled multiple-frequency MRI/S coil system. This coil system consists of an external surface coil inductively coupled to an internal monitoring coil. The design incorporates a switchable capacitor/varactor array driven by a microcontroller based on information received wirelessly. The focus of this work was to build a proof-of-concept prototype before developing the entire micro fabricated integrated chip system.
1117.
An Efficient Switched Double-Frequency Birdcage Coil for 3He
and 1H Imaging
Jian-Xiong Wang1, Eddy B. Boskamp2, Giles E. Santyr3, Brian K. Rutt3 1GE Healthcare, London, Canada; 2GE Healthcare, Waukesha, USA; 3University of Western Ontario, London, Canada This work presented a switched double-frequency birdcage coil aiming for the development of double-frequency clinic body coil. This new technique doesn’t have the coupling problem and it’s easy to be implemented into existing clinic system. 1118. Optimization of Internal MRI Coils Using Ultimate Intrinsic SNR Yigitcan Eryaman1, Yusuf Öner2, Ergin Atalar1 1Bilkent University, Ankara, Turkey; 2Gazi University, Ankara, Turkey In this work an optimization method for internal MRI coils is explained. Ultimate intrinsic SNR which is the maximum value of the intrinsic SNR that can be obtained by internal MRI coils is used as a measure of performance. As a sample optimization endorectal MRI coils are optimized. The optimized coil is tested in both phantom and patient studies for prostate imaging. Similar methods can be used to optimize different type of internal MRI coils. |
Computational Electromagnetics Hall D Wednesday 13:30-15:30 Method of Moments Christian Findeklee1, Ulrich Katscher1 1Philips Research Europe, Hamburg, Germany The distribution of the electric conductivity s; and permittivity e; of human tissue can be used as diagnostic parameter and for a prediction of local SAR during MR measurements. Electric Properties Tomography (EPT) offers a way to determine these electric properties via MRI. In this work, equivalent volume currents in the Method of Moments were used for the iterative 3D EPT reconstruction.
1179.
Electric Properties Tomography: Calculation of the In-Vivo
Electric Field Within Realistic Computation Cornelis A.T. Van den Berg1, Mikel Boute1, Bob Van den Bergen1, Jan J.W. Lagendijk1, Ulrich Katscher2 1University Medical Center Utrecht, Utrecht, Netherlands; 2Philips Research Europe-Hamburg, Hamburg, Germany Electric Properties Tomography (EPT) is a method to determine the dielectric properties during an MRI scan. It can be used as a diagnostic parameter or more precise local SAR assessment. Using an iterative computation scheme the dielectric properties are determined from the curl of the measured transverse magnetic field and a calculated axial electric field distribution. Conventional methods such as FDTD solvers are too slow to calculate the electric field with millimetre resolution. We successfully tested a method called quasi-stationary zooming that reduced the computation time with a factor of 20 without serious degrading the EPT results. 1180. Empirical and Computed B0 Perturbations Induced by Metallic Implants Kevin M. Koch1, R Scott Hinks1 1GE Healthcare, Waukesha, Wisconsin, USA Metallic implants used in arthroplasty procedures can induce severe spatial B0 perturbations and significantly distort conventional MRI. Information on the nature of these field perturbations has previously been difficult to assess. Here, it is demonstrated that FFT-based computations of approximated solutions to Maxwell’s equations can be used to rapidly and accurately predict high-resolution B0 distributions induced by metallic implants. A novel method is also presented that allows for MR-based collections of far-off resonance (>8kHz) contour B0 maps. The presented techniques are utilized to determine the previously unpublished magnetic susceptibility of a commonly used cobalt-chromium metal alloy. 1181. A Physical Shepp-Logan Head Phantom Julian R. Maclaren1, Philip Bones1, Rick P. Millane1, Richard Watts1 1University of Canterbury, Christchurch, New Zealand An MR-compatible Shepp-Logan head phantom has been constructed from layers of polycarbonate. When imaged using the correct slice thickness, the partial volume effect generates the required contrast values for each region. This phantom can be used to validate simulations conducted using the well-known numerical version. The physical phantom has been combined with a computer controlled moving platform to allow the effect of motion on various pulse sequences to be studied. This abstract compares data obtained from the physical phantom with data generated using the mathematical model of the Shepp-Logan head phantom. 1182. B1 Field Optimization for Microstrip Transmission Line Volume Coil at High Field Chunsheng Wang1, Xiaoliang Zhang1 1University of California San Francisco, San Francisco, USA B1 field distributions within MTL head coils with different substrate thickness were investigated by finite difference time domain method. Although the B1 field homogeneity can be improved by increasing substrate thickness of MTL coils for unloaded case, the B1 field homogeneity of MTL coils after loading can be optimized by selecting appropriate substrate thickness by making use of ¡®dielectric resonance¡¯ effect. 1183. Numerical Model of a Dielectric Resonator for High Field MRI Bu Sik Park1, Andrew G. Webb2, Michael B. Smith3, Christopher M. Collins3 1Penn State University, College of Medicine, Hershey, Pennsylvania, USA; 2Penn State University, University Park, Pennsylvania, USA; 3Pennsylvania State University, Hershey, Pennsylvania, USA Based on previous research, we have designed a dielectric resonator of barium strontium titanate (having a dielectric constant of 323) with a diameter of 28 mm for high field MRI and performed full Maxwell numerical calculations of the electromagnetic fields to evaluate the resonator. The cylindrical dielectric resonator has a region of low loss and high sensitivity along its central axis, and should provide lower sample loss than (for example) a solenoid design, which is found to produce a significant electric field in the sample due to the electrical potential on the wires. 1184. A New Analytical Approach to RF Coils at High Frequency: the Spiral Coil Example Xin Chen1, Xingxian Shou1, Hiroyuki Fujita, 12, Victor Taracila3, Robert W. Brown1 1Case Western Reserve University, Cleveland, Ohio, USA; 2Quality Electrodynamics, LLC., Mayfield Village, Ohio, USA; 3GE Healthcare, Aurora, Ohio, USA RF field inhomogeneity continues to be a major challenge in today’s high-field MRI. The RF wavelength is significantly shortened in biological samples due to a high dielectric constant, leading to field nonuniformities. We present a new analytical approach that is robust for optimization studies. As an example, we analyze a spiral coil, with special attention paid to the interplay between transverse and axial wavelength effects. The analytical results are consistent with previously published experiments and preliminary numerical simulations. An optimized spiral coil leads to improved field homogeneity, relative to the birdcage coil, up to (at least) 400MHz. Bu Sik Park1, Christopher M. Collins2 1Penn State University , Hershey, Pennsylvania, USA; 2Penn State University, Hershey, Pennsylvania, USA Although it is generally believed that almost all of the sample loss for the Birdcage Coil is magnetically-induced, there have been some suggestions that conservative E-fields may play a significant role, including suggestions that low-pass (LP) and high-pass (HP) BCs have significantly different E-field distributions (and may couple to the human subject very differently) due to their differing capacitor placement. Here we apply a recently-developed method to evaluate the conservative and magnetically-induced E-fields, as well as the sample power loss for a birdcage coil loaded with the head. 1186. Investigation of Mutual Inductance Coupling and Capacitive Decoupling of N-Element Array System Lian Xue1, L.-M Xie1, Maged R. Kamel1, Flora Ip1, Jarek Wosik1, Alexander C. Wright2, Felix W. Wehrli2 1University of Houston, Houston, USA; 2University of Pennsylvania Medical Center, Philadelphia, USA We report on the analyzes of the inductive coupling and capacitive decoupling of N-element array resonant system. A two-state rf current direction model for magnetically coupled coils was developed to calculate number of possible resonant modes in N-element array. We have obtained an analytical solution for number of possible frequency modes in a symmetric resonant linear and circular system of N-element array. Decoupling was achieved by using a set of capacitors to reduce the frequencies of higher modes to the lowest mode frequency. Examples of capacitive-only decoupling of 2x1, 2x2 and 4x1 arrays are shown and mechanism of such decoupling is analyzed. 1187. Exchange of Surface Coil Antenna Patterns Due to Gyrotropism Cornelius von Morze1, 2, Albert Pofu Chen1, Duan Xu1, Daniel B. Vigneron1, 2, James Tropp3 1UCSF, San Francisco, USA; 2UCSF/UCB Joint Graduate Group in Bioengineering, San Francisco, USA; 3GE Healthcare, Fremont, USA A recent formulation of NMR reciprocity that accounts for the gyrotropism of spins was tested through experiments on oppositely polarized 3.0T MRI scanners. The transmit and receive antenna patterns of a surface coil are distinct, and furthermore exchange on field reversal. The receive patterns of a bilaterally symmetric dual array were first measured through a receive-only experiment. Following the proper pattern exchanges and multiplications, these results predicted the distinct appearance of images obtained by operating the coil in transmit-receive mode on oppositely polarized systems. 1188. Fast, Accurate Calculation of Maximum Local N-Gram Specific Absorption Rate Adam Charles Zelinski1, Vivek K. Goyal1, Elfar Adalsteinsson1, 2, Lawrence L. Wald, 23 1MIT, Cambridge, Massachusetts, USA; 2Harvard-MIT Division of Health Sciences & Technology, Cambridge, Massachusetts, USA; 3Harvard Medical School, MGH, Charlestown, Massachusetts, USA SAR is a major concern in MRI; it is crucial to be able to quickly and accurately calculate maximum N-gram local SAR due to a pulse. In contrast with low-resolution models that lack many compartments and have overly-smooth tissue-boundaries, high-resolution models permit accurate local SAR calculations. Unfortunately, the fine resolution of such models makes calculating average N-gram SAR at each location a computationally intractable problem (straightforward region-growth methods may no longer be used). We pose a rapid algorithm that is equal in speed to the fixed-cube method, yet always averages over a N-gram region, ensuring accuracy of local SAR estimates. 1189. Phased Array Evaluation Using a Human Body Model at 3T Selaka Bandara Bulumulla1, Christopher J. Hardy1 1GE Global Research, Niskayuna, USA We use an anatomically accurate human body model with distinct electrical properties for different tissue types for evaluating the SNR of a 16-element phased array used for body imaging at 3T. Using numerical methods, we estimate the electromagnetic fields in the model and calculate an intrinsic SNR metric to evaluate the imaging performance of the array. The results highlight many of the artifacts common in 3T imaging, including left-right shading arising from wavelength effects. This method can be used to evaluate the imaging performance of a given array, and to optimize array configurations at various field strengths.
1190.
Lower RF Power Absorption Combined with Uniform Excitation
Using Multi-Port B1 Shimming: Numerical Lin Tang1, YiK-Kiong Hue2, Tamer S. Ibrahim, 12 1University of Oklahoma, Norman, Oklahoma, USA; 2University of Pittsburgh, Pittsburgh, Pennsylvania, USA We show using the finite difference time domain method that improving the B1+ field homogeneity over 3-D regions with volumes as large as the whole head while simultaneously reducing the total RF absorption by human head to values lower than that associated with quadrature excitation can be achieved at 7T. The concept of B1 shimming with reduced RF power absorption was validated using a 7T whole-body system with a Tx array. 1191. Estimation of Local SAR Using B1 Mapping Ulrich Katscher1, Tobias Voigt2, Christian Findeklee1, Kay Nehrke1, Steffen Weiss1, Olaf Doessel2 1Philips Research Europe - Hamburg, Hamburg, Germany; 2University Karlsruhe, Karlruhe, Germany The local specific energy absorption rate (SAR) is one of the major issues for MRI at 3T and above. For an exact determination of local SAR, the spatial distribution of the electric field and electric conductivity throughout the patient are required. Alternatively, according to Maxwell's equations, the required quantities can be deduced by the magnetic field of the RF coil. In this study, the local SAR is estimated via suitable post-processing of a standard B1 map. Simulations show, that the information content of a B1 map might be sufficient to estimate local SAR close to the correct local SAR. 1192. A BiCG Solution Based Quasi-Static Finite-Difference Scheme for Induced Field Evaluation in MRI Hua Wang1, Feng Liu1, Adnan Trakic1, Stuart Crozier1 1The University of Queensland, Brisbane, Australia This study presents a biconjugate gradient method (BiCG) that can significantly improve the performance of the quasi-static finite-difference scheme, which has been widely used to model field induction phenomena in voxel phantoms. The proposed BiCG method offers remarkable computational advantages in terms of convergence performance and memory consumption over the conventional iterative, successive over-relaxation algorithm. The wide application capability and computational performance of the BiCG method is demonstrated by modelling the exposures of MRI healthcare workers to fields produced by pulsed field gradients, which is presently an important topic of research in light of the Physical Agents Directive 2004/40/EC. 1193. A Complete Modeling System with Experimental Validation for Calculating the Transmit and ReceiveFields, Total Power Deposition, Input Impedance, and Coupling Between Coil Elements YiK-Kiong Hue1, Tiejun Zhao, Yongxian Qian1, Fernando E. Boada1, Tamer S. Ibrahim1 1University of Pittsburgh, Pittsburgh, Pennsylvania, USA A complete finite-difference time-domain (FD-TD) model of a highly coupled RF coil is provided at 7T. Experimental results that validate all of the coil’s (while loaded) electromagnetic characteristics (calculated using the FD-TD method) including coupling, excite and receive field, input impedance, and total RF power deposition are presented. 1194. Time-Domain Finite-Difference/Finite-Element Hybrid Simulations of High-Field RF Coils Shumin Wang1, Jeff H. Duyn1 1LFMI/NINDS/NIH, Bethesda, Maryland, USA A Time-Domain Finite-Difference/Finite-Element (TD-FD/FE) hybrid method was developed to simulate high-field RF coils with realistic modeling of both the human body and the RF coils. The FEM is only applied to model curved coil structures while the human body is simulated by the FDTD method. Since both methods are applied simultaneously in a same simulation, the interaction between the human body and RF coils are fully accounted for with well balanced accuracy and efficiency. Numerical results demonstrate the validity, accuracy and capability of this approach. Bob van den Bergen1, Chris C. Stolk2, Jan Bouwe van den Berg3, Cornelis A. T. van den Berg1 1University Medical Center Utrecht, Utrecht, Netherlands; 2University of Twente, Enschede, Netherlands; 3VU University Amsterdam, Amsterdam, Netherlands RF shimming and parallel transmission have shown to be able to improve the image quality and reduce the SAR for high field MR imaging. However, both methods require information about the electromagnetic fields, of which only the magnetic part can be measured. FDTD calculations have become standard as alternative to measurements, although these calculations are time-consuming and therefore unsuitable for on-line use. We present a new method that can calculate the electric and magnetic field in the body in one or two minutes, thereby increasing the on-line feasibility of both RF shimming and parallel transmission drastically. Can Akgun1, Carl J. Snyder1, Steen Moeller1, Pat John Bolan1, Thomas Vaughan1, Kamil Ugurbil1, Pierre-Francois Van de Moortele1, Greg Metzger1 1University of Minnesota, Minneapolis, Minnesota, USA In an era where large arrays and complex geometries are being proposed for UHF to maximize parallel imaging performance, B1+ shimming flexibility, and receive sensitivity, the ability to predict transmit and receive characteristics in vivo becomes imperative. In support of our 7T prostate imaging, we measured in vivo B1+ maps and g-factor maps with 8- and 16-channel phased array coils and found the results were correctly predicted by simulations using Remcom’s XFDTD software. These findings build confidence in our ability to evaluate SAR for the development of future optimization procedures, as well as guiding future RF coil development. |
Parallel RF Transmission Methods Hall D Wednesday 13:30-15:30 William Allyn Grissom1, Dan Xu2, Jeffrey A. Fessler1, Douglas C. Noll1 1University of Michigan, Ann Arbor, USA; 2GE Healthcare, Waukesha, USA When applied to parallel excitation, the optimal control pulse design method has some nice advantages over other methods. However, the method requires long computation times, limiting its practical use. We introduce a fast optimal control method that reduces computation time by linearizing the Bloch equation about an underlying pulse, and designs perturbations to the pulse using the linear conjugate gradient algorithm. The new method also has reduced memory requirements compared to conventional optimal control. 1302. Sparsity-Enforced Coil Array Mode Compression for Parallel Transmission Adam Charles Zelinski1, Vijayanand Alagappan2, Vivek K. Goyal1, Elfar Adalsteinsson1, 3, Lawrence L. Wald2, 3 1MIT, Cambridge, Massachusetts, USA; 2Harvard Medical School, MGH, Longwood, Massachusetts, USA; 3Harvard-MIT Division of Health Sciences & Technology, Cambridge, Massachusetts, USA Forming linear combinations of the elements of a parallel transmission array can transform its spatial modes into a different basis set, potentially capturing TX efficiency and acceleration capabilities in a subset of the modes. Available TX channels are then applied to only a subset of array modes, those that best contribute to TX efficiency and encoding. Here we propose a target-dependent sparsity-enforced subset selection (SESS) algorithm that explicitly accounts for the desired excitation when choosing the mode subset. Brute-force search shows that SESS actually determines the best of all 12,871 possible subsets during an experiment using a 16-mode, 8-channel system. 1303. Sparsity-Enforced Joint Spiral Trajectory & RF Excitation Pulse Design Adam Charles Zelinski1, Vivek K. Goyal1, Lawrence L. Wald2, 3, Elfar Adalsteinsson1, 3 1MIT, Cambridge, Massachusetts, USA; 2Harvard Medical School, MGH, Charlestown, Massachusetts, USA; 3Harvard-MIT Division of Health Sciences & Technology, Cambridge, Massachusetts, USA A fast sparsity-enforcement algorithm that applies to both single-channel and parallel transmission systems is proposed that generates sparse, quickly-traversable k-space trajectories & a corresponding RF excitation pulses when provided with user-defined target patterns. L1-regularization is used to tractably search over a number of possible candidate k-space trajectory segments and determine a small subset of these segments, along with an RF pulse, that best form the desired excitation. A trajectory algorithm then connects the segments in the small subset into an overall smooth trajectory. Here, we provide the method with candidate rings and generate trajectories that significantly outperform conventionally-accelerated spirals.
1304.
Magnitude Least Squares Optimization for Parallel RF
Excitation Design Demonstrated at 7 Tesla Kawin Setsompop1, Vijayanand Alagappan2, Borjan Gagoski1, Lawrence Wald3, Elfar Adalsteinsson1 1Massachusetts Institute of Technology, Cambridge, Massachusetts, USA; 2A. A. Martinos Center for Biomedical Imaging, MGH, Charlestown, Massachusetts, USA; 3A. A. Martinos Center for Biomedical Imaging, Department of Radiology, MGH, Charlestown, Massachusetts, USA Parallel RF excitations are often designed as a least-squares (LS) optimized approximation to a target magnitude and phase profile. However, adherence to the target phase profile is usually not important as long as the excitation phase is slowly varying compared to the voxel dimension. We describe and validate an MLS method with slice selective spoke and 4X spiral square target excitations, using an 8-channel transmit array on a 7T human MRI scanner. The method resulted in significant benefits, particularly for the spoke, where a 34% drop in root magnitude mean square error and 49% drop in RF power were observed.
1305.
Slice-Selective RF Pulses for in Vivo B1+ Inhomogeneity
Mitigation at 7 Tesla Using Parallell RF Excitation Kawin Setsompop1, Vijayanand Alagappan2, Borjan Gagoski1, Thomas Witzel2, 3, Jonathan Polimeni2, Andreas Potthast4, Ulrich Fontius5, Franz Schmitt5, Lawrence Wald, 36, Elfar Adalsteinsson1, 3 1Massachusetts Institute of Technology, Cambridge, Massachusetts, USA; 2A. A. Martinos Center for Biomedical Imaging, MGH, Charlestown, Massachusetts, USA; 3Harvard-MIT Division of Health Sciences and Technology, Cambridge, Massachusetts, USA; 4Siemens Medical Solutions, Charlestown, Massachusetts, USA; 5Siemens Medical Solutions, Erlangen, Germany; 6A. A. Martinos Center for Biomedical Imaging, Department of Radiology, MGH, Charlestown, Massachusetts, USA At 7T, B1+ inhomogeneity for human brain imaging causes spatially inhomogeneous flip-angle with detrimental non-uniformity for both SNR and CNR. In this work we applied 8-channel parallel excitation waveforms in the low-flip-angle regime with echo-volumnar k-space trajectories that deposit slice-selective sinc “spokes” in kz, whose complex amplitudes are modulated in (kx,ky) to mitigate in-plane B1+ inhomogeneity. We designed parallel RF spoke pulses with magnitude least squares criteria, spoke location optimization, and B0 field map incorporation that demonstrated excellent B1+ mitigation for brain imaging on three human subjects at 7T. 1306. High-Flip-Angle Slice-Selective Parallel RF Excitation with 8 Channels at 7 Tesla Kawin Setsompop1, Vijayanand Alagappan2, Andreas Potthast3, Ulrich Fontius4, Lawrence Wald5, Elfar Adalsteinsson1 1Massachusetts Institute of Technology, Cambridge, Massachusetts, USA; 2A. A. Martinos Center for Biomedical Imaging, MGH, Charlestown, Massachusetts, USA; 3Siemens Medical Solutions, Charlestown, Massachusetts, USA; 4Siemens Medical Solutions, Erlangen, Germany; 5A. A. Martinos Center for Biomedical Imaging, Department of Radiology, MGH, Charlestown, Massachusetts, USA At high magnetic field, B1+ inhomogeneity causes undesired non-uniformity in SNR and contrast. Parallel RF “spoke”-based designs have been shown to correct this problem and produce highly uniform in-plane magnetization with excellent slice selection with relatively short excitation durations. However, at large flip angles the excitation k-space based design method fails. Recently, several large-flip-angle excitation designs have been proposed. In this work we propose and successfully demonstrate a large-flip parallel excitation design for 90º and 180º spin-echo pulses for 8 channels at 7T in the presence of B1+ variation similar to that of human brain. 1307. In-Vivo Comparison of B1 Shimming and Spatially Tailored Parallel Excitation at 7T Vijayanand Alagappan1, 2, Kawin Setsompop3, Andreas Potthast4, Ulrich J. Fontius5, Franz Schmitt5, Elfar Adalsteinsson3, Lawrence L. Wald1 1Athinoula A. Martinos Center for Biomedical Imaging, Charlestown, Massachusetts, USA; 2Tufts University, Medford, Massachusetts, USA; 3MIT, Cambridge, Massachusetts, USA; 4Siemens Medical Solutions, Charlestown, USA; 5Siemens Medical Solutions, Erlangen, Germany The inhomogeneous transmit B1 pattern in conventional volume coils at higher field strengths has motivated the development of new RF excitation strategies with multiple transmit channels. The array approach allows the transmit B1+ field inhomogeneity to be addressed by B1 shimming or by using accelerated spatially tailored 2D or 3D RF pulses. In this work we have compared the achievable flip angle homogeneity of the 2 techniques on a slice selective excitation using 8 transmit channels. 1308. RF Encoding Via Non-Linear, Iteratively Determined Encoding Functions Ulrich Katscher1, Jonathan Lisinski2, Peter Boernert1 1Philips Research Europe - Hamburg, Hamburg, Germany; 2University of Applied Sciences, Hamburg, Germany For spatial MRI signal encoding, RF encoding via B1-gradients are an alternative to the standard Bo-gradients. RF encoding might omit all Bo-gradients, allowing for MR scanning almost free of acoustic noise. This study investigates RF encoding using non-linear encoding functions using an eight-channel transmit array for phantom imaging. Encoding functions are determined iteratively by randomized superposition of the array element sensitivities. The possibility of pure RF phase encoding (via constant B1 amplitude encoding functions) is investigated to allow future steady state-like applications. Furthermore, RF encoding was combined with 1D Fourier encoding as additional alternative for sequence acceleration. Sukhoon Oh1, Sylvie Garrett2, Christopher M. Collins1 1The Pennsylvania State University, Hershey, USA; 2The Pennsylvania State University, State College, USA In this numerical study, an array-optimized composite RF pulse is designed to simultaneously improve the excitation homogeneity and reduce the peak SAR within the human body at 3T. The composite RF pulse can achieve both significantly better excitation homogeneity and significantly lower SAR levels than are possible with RF shimming alone. 1310. Transmit-PILS RF Pulse Design for Small-Tip-Angle Parallel Excitation William Allyn Grissom1, Jeffrey A. Fessler1, Douglas C. Noll1 1University of Michigan, Ann Arbor, USA Existing pulse design methods for small-tip-angle parallel excitation either design pulses iteratively and can have high computational cost, or design pulses non-iteratively and are restricted to echo-planar trajectories. We introduce a method that is non-iterative and applicable to general trajectories. Compared to iterative methods, transmit-PILS achieves lower computational cost by dividing the pulse design problem for R coils into R sub-problems that are solved individually using conjugate phase pulse design. 1311. 3D RF Shimming Using Multi-Frequency Excitation Ulrich Katscher1, Holger Eggers1, Ingmar Graesslin1, Giel Mens2, Peter Boernert1 1Philips Research Europe - Hamburg, Hamburg, Germany; 2Philips Medical Systems, Best, Netherlands For RF shimming, amplitudes and phases of a transmit RF coil array are optimized to obtain a homogeneous B1 profile. In this study, the different Tx elements are driven with different frequencies (Multi-Frequency Excitation, "MULTIFEX"). Thus, the different elements relate to different slices due to the underlying slice selection gradient. MULTIFEX enables the simultaneous excitation of different slices with different B1 profiles. It might also be used for saturation pulses, refocusing pulses, 3D RF encoding, and 3D local excitation. The study demonstrates the feasibility of MULTIFEX experimentally using a whole-body, eight-channel Tx system at 3T. 1312. Toeplitz-Based Acceleration of RF Pulse Design for Parallel Excitation Daehyun Yoon1, William A. Grissom2, Jeffrey A. Fessler2, Douglas C. Noll2 1University of Michigan, Ann Arbor, USA; 2University of Michigan, Ann Arbor, Michigan, USA Recently, an iterative algorithm for designing the RF pulses in the spatial domain for parallel excitation using a Conjugate Gradient(CG) algorithm together with Non-Uniform FFT(NUFFT) operations was proposed. This abstract presents a method based on Toeplitz matrices to further accelerate the computation time of the aforementioned method. This method eliminates the need for interpolation and in certain circumstances, can reduce the dimensionality of the minimization problem. 1313. Parallel Excitation Experiments Using Measured K-Space Trajectories for Pulse Calculation Peter Ullmann1, Martin Haas2, Franciszek Hennel1, Markus Wick1, Jerôme Voiron1, Maxim Zaitsev2, Jürgen Hennig2, Wolfgang Ruhm1 1Bruker BioSpin MRI GmbH, Ettlingen, Germany; 2University Hospital Freiburg, Freiburg, Germany Parallel Excitation has proven to be a powerful tool to reduce the typically long durations of spatially selective pulses and the resultant degradation of the excitation patterns due to off-resonance or B0-inhomogeneity effects. However, the effects of some experimental imperfections such as k-space trajectory deviations, e.g. by eddy currents, are not reduced by Parallel Excitation. In this study it is demonstrated that using measured k-space trajectories for the calculation of Parallel Excitation pulses is an effective and convenient method of reducing errors in the excitation patterns caused by such experimental imperfections of the gradient system. 1314. Automated Pulse Sequence Design for Minimal SAR in Parallel Transmission Yinan Liu1, Jim Ji1 1Texas A&M University, College Station, Texas, USA Managing and reducing SAR is a critical issue for parallel transmission. In this work, an automated pulse sequence design method is presented to achieve minimal SAR for a given excitation pattern without extending the pulse duration. The Lagrange dual method is used to iteratively optimize the RF pulses and corresponding k-trajectory. The method is verified using computer simulations of a 4-channel parallel transmission system.
1315.
Specific Absorption Rate Studies of the Parallel Transmission
of Inner-Volume Selective Excitations Adam Charles Zelinski1, Leonardo M. Angelone2, Vivek K. Goyal1, Giorgio Bonmassar2, Elfar Adalsteinsson1, Lawrence L. Wald, 12 1MIT, Cambridge, Massachusetts, USA; 2MGH, Charlestown, Massachusetts, USA SAR is a major concern in the parallel transmission (pTX) of spatially-tailored 2D and 3D excitation pulses due to E field superposition that occurs when driving multiple channels concurrently and the inefficiency of producing excitations via regional cancellation. In this work, we model average and local SAR using FDTD-simulated field data of an 8-channel pTX array and a human head model. Mean and local SAR are analyzed for 2D spiral-trajectory, inner-volume, spatially-selective RF excitation pulses as a function target flip angle, position, size, smoothness, orientation, and trajectory acceleration factor. 1316. Equiripple Design of Multidimensional RF Pulses Via Convex Optimization Daeho Lee1, Michael Lustig1, John Mark Pauly1 1Stanford University, Stanford, USA We propose a new approach to designing RF pulses which achieves equiripple excitation profile. It formulates the RF design problem as a convex optimization problem. Thus, other physical constraints (e.g. peak RF) and regularization terms (e.g. RF power) can be easily embedded into the problem if they are represented in the form of convex functions. This approach can be extended to parallel RF transmission.
1317.
Improved Accuracy of Actual Flip Angle Imaging (AFI)
Kay Nehrke1, Peter Börnert1, Ulrich Katscher1 1Philips Research Europe, Hamburg, Germany Simulations based on configuration theory have been performed to study the accuracy of the AFI technique depending on different sequence and tissue parameters. The simulation show that the spoil properties, and hence, the accuracy of the technique may be controlled by the RF spoil phase shift increment along with the diffusion sensitivity resulting from spoiling gradients. For a favorable spoil phase, moderate gradient spoiling is already sufficient to allow a good performance of the technique for a reasonable range of tissue parameters. Phantom and in-vivo experiments showed an excellent accordance with the simulations. 1318. Multidimensional Spatial Encoding by Parallel ExcitationWolfgang Ruhm1, Peter Ullmann1 1Bruker BioSpin MRI GmbH, Ettlingen, Germany A method of MRI and spectroscopic imaging is described which allows multidimensional spatial encoding during the excitation period. It is called SPEEDI (Spatial Encoding by Excitation with Multidimensional RF Pulses) and uses multidimensional RF pulses not only for spatial modulation of the transverse magnetization magnitude, but also for generating a targeted spatial distribution of the transverse magnetization phase. Repeated excitation with different phase distributions according to a spatial encoding scheme allows measuring multidimensional images with different contrasts or spatial distributions of spectral information. SPEEDI has been implemented using Parallel Excitation and its successful experimental verification is reported. Martin Haas1, Peter Ullmann2, Jürgen Hennig1, Maxim Zaitsev1 1University Hospital Freiburg, Freiburg, Germany; 2Bruker BioSpin MRI GmbH, Ettlingen, Germany A new method for designing two-dimensional spatially selective 180° pulsesis presented, which is based on the observation that the design equation for the small tip angle regime also holds when the spin population is aligned close to both the +z and -z directions. In combination with one Bloch integration, the gap between these regimes is bridged.We demonstrate that these new pulses can perform both selective inversion and refocusing in an arbitrarily shaped target area. Similarly to transmit SENSE, STM pulses can be accelerated and RF transmit profiles of multiple coils and B0 inhomogeneities can be taken into account. |
Hall D Wednesday 13:30-15:30 1350. Advantages of Long-Axis PROPELLER EPI Via K-Space Weighting: Comparison of Point Spread Function with Short-Axis PROPELLER EPI Tzu-Chao Chuang1, Teng-Yi Huang2, Fu-Nien Wang3, Hsaio-Wen Chung1 1National Taiwan University, Taipei, Taiwan; 2National Taiwan University of Science and Technology, Taipei, Taiwan; 3National Tsing Hua University, Hsinchu, Taiwan Due to the largely shortened echo spacing, short-axis PROPELLER EPI (SAP-EPI) reduces the image blurring without the k-space weighting scheme, which has been previously proposed for de-blurring in long-axis PROPELLER EPI (LAP-EPI) by emphasizing data suffering less from off-resonance field during the combination of blades. Results in this simulation study revealed that the k-space weighting scheme dramatically improved sharpness on LAP-EPI by suppressing side lobes of point spread function in magnitude, while little influence was found for SAP-EPI. As a result, LAP-EPI using a proper weighting function has potentials to obtain images without blurring at the cost of SNR.
1351.
PROPELLER EPI Using Asymmetric Blades: A Preliminary Study on
Point Spread Function
Fu-Nien Wang1, Tzu-Chao Chuang2, Teng-Yi Huang3 1National Tsing Hua University, Hsinchu, Taiwan; 2National Taiwan University, Taipei, Taiwan; 3National Taiwan University of Science and Technology, Taipei, Taiwan Blades with asymmetric number of positive and negative phase encoding lines are implemented to long-axis PROPELLER EPI. Simulations and in vivo experiments are performed. The point spread function investigated in this study validate that the proposed method could reduce the image blurring effect introduced by distorted EPI blade. 1352. 3D SAP-EPI for Self-Navigated T1w Spoiled Gradient Echo Imaging Samantha J. Holdsworth1, Stefan Skare1, Anders Nordell2, Rexford D. Newbould1, Roland Bammer1 1Stanford University, Palo Alto, California , USA; 2Karolinska University Hospital, Stockholm, Sweden 3D T1w Spoiled Gradient Echo (SPGR) imaging is commonly used in the clinical care due to its high gray-white matter contrast and isotropic resolution. While the short TR makes the total scan time reasonable (~3-10 mins), it is prone to motion-induced ghosting. Short-axis readout propeller EPI (SAP-EPI) has been suggested as an approach to efficient self-navigated imaging. A 3D GRE version of the SAP-EPI sequence is implemented here, by acquiring 3D ‘bricks’ instead of blades. We demonstrate motion and distortion-corrected phantom data from controlled motion experiments. Initial human images acquired in a sub-minute scan time are presented.
1353.
View-Sharing PROPELLER with Pixel-Based Optimal Blade
Selection (POBS) for Dynamic Contrast- Tzu-Chao Chuang1, Fu-Nien Wang2, Teng-Yi Huang3, Chih-Ching Wu1, Hsaio-Wen Chung1 1National Taiwan University, Taipei, Taiwan; 2National Tsing Hua University, Hsinchu, Taiwan; 3National Taiwan University of Science and Technology, Taipei, Taiwan A novel method termed pixel-based optimal blade selection (POBS) to minimize reconstruction artifacts in dynamic contrast-enhanced view-sharing PROPELLER imaging due to under-sampled outer k-space data. POBS is based on an iterative process to search for one set of blades exhibiting the closest image contrast with the target blade, on a pixel-by-pixel basis by least square error estimation. Results from computer simulations and animal experiments show that POBS yields errors less than 15% with five-fold acceleration as compared with the fully sampled images, with reduction of reconstruction artifacts visually perceivable. 1354. CNR Optimization in Variable Pitch PROPELLER MRA Nicholas R. Zwart1, James G. Pipe1 1Barrow Neurological Institute, Phoenix, USA The optimization of inflow enhancement in Variable Pitch PROPELLER relies on the fine tuning of the RF-chirp pulse design, refocussing of slice selective gradient and the order in which blade-slabs are overlapped. These mechanisms provide eight distinctive parameters that are optimized in terms of an image quality metric that is constrained through the experimental results using a custom flow phantom. 1355. BLADE-VAT for Geometric Distortion Correction Guobin Li1, Bida Zhang1, Hua Guo, Qiang Zhang1 1Siemens Mindit Magnetic Resonance Ltd., Shenzhen, People's Republic of China The View angle tilting (VAT) corrects for in-plane geometric distortion from field inhomogeneity, but may introduce serious image blurring. In this abstract,a novel method, called BLADE-VAT, which employs low resolution acquisition in readout direction, provides high efficient reduction of such blurring. 1356. Accelerating PROPELLER-MRI by Means of Under-Sampling and Iterative Image ReconstructionUsing the Non-Uniform Fast Fourier Transform Ashish A. Tamhane1, Mark A. Anastasio1, Konstantinos Arfanakis1 1Illinois Institute of Technology, Chicago, Illinois, USA PROPELLER-MRI is a multi-shot fast spin-echo (FSE) data acquisition technique with reduced sensitivity to motion and B0 inhomogeneities. However, PROPELLER acquisitions are at least 50% slower than conventional FSE. Under-sampling k-space in PROPELLER by reducing the number of blades per acquisition, while increasing the distance between lines, can significantly reduce imaging time, but leads to severe image artifacts when reconstructing images using conventional gridding. In this study, we investigated the trade-off between imaging time, artifacts and noise, for under-sampled PROPELLER acquisitions with iterative reconstruction using the non-uniform fast Fourier transform. |
Hall D Wednesday 13:30-15:30 1380. k-T Accelerated IDEAL for Robust Water-Fat Dynamic Imaging Jeffrey Tsao1, Yun Jiang1 1Novartis Institutes for BioMedical Research, Cambridge, Massachusetts, USA IDEAL is an attractive water-fat separation method. A drawback is the prolonged scan time due to the need to acquire at least 3 echo times. Dynamic imaging with IDEAL can be combined with k-t acceleration to reduce scan time below the nominal time for a single image. In this work, we demonstrated net acceleration factors of 4.3x and 6.3x, resulting in scan times of 70% or 47% compared to a single-average scan. The combination of IDEAL and k-t acceleration is favorable since the IDEAL echo times present an extra data dimension, which can be used for signal packing and acceleration. 1381. Water-Fat Separation with IDEAL, Undersampled Radial Acquisition, and Off-Resonance Deblurring Yun Jiang1, Jeffrey Tsao1 1Novartis Institutes for BioMedical Research, Cambridge, Massachusetts, USA In this work, we combined radial acquisition with IDEAL water-fat separation and off-resonance correction. Phantom and in vivo data demonstrate that it is a robust method to accelerate fat-water separation. 1382. Water-Fat Decomposition with Regularized Field Map Wonseok Huh1, Jeffrey A. Fessler1, Alexey A. Samsonov2 1University of Michigan, Ann Arbor, Michigan, USA; 2University of Wisconsin-Madison, Madison, Wisconsin, USA We proposed a method for water-fat decomposition with regularized field map estimation. In this work, we describe a penalized-likelihood method for jointly estimating water, fat, and the field map. The method uses the a priori knowledge that field maps are usually smooth by including a regularization term in the cost function. The fat and water components can be solved for analytically as a function of the field map. By providing a field map estimate that is inherently smooth even in regions of low signal intensity, the method may yield improved water and fat images. 1383. Simultaneous Water-Fat and T2* Mapping with 3-Point Acquisitions Qing-San Xiang1 1University of British Columbia, Vancouver, Canada Water and fat images are obtainable using either gradient-echo or spin-echo sequences with incremental echo shifts. Similar sequences can also be used for T2* mapping. This work demonstrates simultaneous mapping of both water-fat and T2* with as few as 3 acquisitions. Three complex images were acquired with water and fat vectors at (90, 270, 450) degrees, forming Three-Orthogonal-Phase (TOP) sampling. Orthogonality between water and fat allows decoupling of T2* and chemical shift effects. T2* map is readily computed from three otherwise identical magnitude images while water-fat images are obtained after phase correction with Regional Iterative Phasor Extraction (RIPE). 1384. Water Fat Shift Displacement Artifact Correction in Two Point Dixon Imaging Olof Dahlqvist Leinhard1, 2, Andreas Johansson1, Peter Lundberg1, 2 1Linköping University, Linköping, Sweden; 2Center for Medical Image Science and Visualization (CMIV), Linköping, Sweden The chemical shift artifact in a two point Dixon acquisition can be removed by an iterative solver taking advantage of the known fat water shift. A simple implementation has been made and the initial results are presented with promising results. 1385. Multislice Water-Fat Imaging with Simultaneous Echo Refocusing (SER-Dixon) Kuan J. Lee1 1University of Sheffield, Sheffield, UK A new multislice water-fat imaging sequence is introduced which combines Simultaneous Echo Refocusing (SER) with Partially-Opposed-Phase (POP) acquisition. By slightly increasing the interval between echoes from one slice, echoes from another slice may be fitted in between. 1386. TIDE BSSFP as a Fat-Suppression Preparation Yin-Cheng Kris Huang1, Hsiao-Wen Chung2, Teng-Yi Huang3 1National Taiwan University , Taipei City, Taiwan; 2National Taiwan University, Taipei City, Taiwan; 3National Taiwan University of Science and Technology, Taipei City, Taiwan The Transition into Driven Equilibrium balanced steady-state free precession (TIDE bSSFP) sequence has been shown to carry an intrinsic suppression of off-resonance signals after experiencing a certain number of RF pulses, just with an appropriate choice of the timing passing through the k-space center. The stopbands are wide in comparison to other fat-suppression schemes such as CHESS. Therefore, it should be possible to incorporate TIDE bSSFP fat suppression into other routine readout module, by truncating the TIDE bSSFP up to the null point for off-resonance signals. In regions where good shimming is harder to achieve, the wide stopbands of TIDE may be helpful. 1387. RF Phase-Cycling Water-Saturated 3D B-SSFP for Fast Abdominal Fat Imaging Yao Ding1, Anqi Zhou2, Roderick W. McColl1, Paul T. Weatherall1, Qi Peng2 1UT Southwestern Medical Center at Dallas, Dallas, Texas, USA; 2UT Health Science Center at San Antonio, San Antonio, Texas, USA A fast water-suppressed b-SSFP sequence was reported previously for abdominal fat imaging using a manual f0 shift of -220Hz (at 1.5T) to fat resonant frequency peak to avoid banding artifacts on fat. We propose here a clinically available, automatic phase-cycling b-SSFP scheme to avoid manual f0 shift difficulties. Theoretical analysis of the new phase-cycling approach is provided and three subjects were scanned using both techniques. No image quality difference was noticed, and fat quantification on those images show little differences (<0.5%). Therefore, the proposed approach can replace the previously reported manual f0 shift approach for fast abdominal fat imaging. 1388. Comparison of Fat-Water Separation by 2D RF Pulse and Dixon Method in Balanced Steady-StateFree Precession Jing Yuan1, Bruno Madore1, Lawrence P. Panych1 1Department of Radiology, Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts, USA 2drf pulse excitation is presented as a novel method for fat-water separation in b-SSFP imaging. It has the advantages of only single acquisition and no image post-processing. This study compares fat-water separation by 2-point Dixon method and 2drf pulse in b-SSFP sequence. The results show that 2drf pulse has comparable separation performance to Dixon method. Pros and cons by these two fat-water separation are discussed. |
Hall D Wednesday 13:30-15:30 Torsten Rohlfing1, Martin H. Rademacher1, Adolf Pfefferbaum1 1SRI International, Menlo Park, California , USA Patient motion between passes of an interleaved image acquisition makes the resulting image stack unusable for 3D image analysis. This paper introduces an algorithm for retrospective correction of characteristic interleaved motion artifacts. The correction is cast as an image reconstruction problem, which is solved by an inverse interpolation algorithm. 1463. Spectral Selective Pencil-Beam Navigator for Abdominal Imaging Max Köhler1, Chrit Moonen2, Mario Ries2 1Philips Medical Systems, Finland; 2IMF, CNRS / Univ. Bordeaux2, Bordeaux, France A spectral selective pencil-beam navigator is presented, which is based on a 121-binomial pulse train of 2D-selective pulses. The beam allows robust real-time slice tracking of abdominal organs without interference with water-selective MR-imaging. 1464. Parallel Imaging PROPELLER with Across-Blade Calibration: In Vivo Results Philip James Beatty1, Zhiqiang Li2, Zachary William Slavens3, Anja C. S. Brau1 1GE Healthcare, Menlo Park, California , USA; 2GE Healthcare, Phoenix, Arizona , USA; 3GE Healthcare, Waukesha, Wisconsin, USA Parallel imaging PROPELLER with across-blade calibration supplements intra-blade calibration with an external "calibration blade". This allows the number of calibration phase encodes acquired with each accelerated blade to be minimized, increasing the net acceleration factor. Using in vivo acquisitions performed during subject motion, it is demonstrated that the across-blade approach is resilient to motion. This study used an acceleration factor of three with two additional calibration phase encodes on each accelerated blade; combined with an echo-train length of 24, this resulting in a net acceleration factor of 2.67 per accelerated blade. 1465. Decoupling Motion Navigation from Imaging Using Spatial-Spectral RF Pulses Andre J. W. van der Kouwe1, Thomas Benner1, Lawrence L. Wald1 1Massachusetts General Hospital, Charlestown, Massachusetts, USA Cloverleaf navigators can be used for real-time motion correction of anatomy that moves as a single rigid body. The technique requires that the signal for imaging and navigation come from the same slab and that the magnetization be kept in steady-state. These requirements may be relaxed by separately exciting fat for the navigation and water for imaging. The utility of cloverleaf navigator motion correction may thus be expanded to include imaging techniques such as clinical 2D fast spin echo and DTI. The principle of decoupling imaging and navigation is shown for translational motions in a phantom and human volunteer. 1466.
Navigator Gated Imaging with Phase Based Edge Detection
Kenichi Kanda1, Yuji Iwadate1, Aki Yamazaki1, Tetsuji Tsukamoto1, Susumu Kosugi1 1GE Yokogawa Medical Systems, Hino, Japan In navigator echo technique, accurate position detection of the diaphragm is essential. As most algorithms rely on the magnitude of the navigator spatial profile, they have some difficulties in the accurate detection when the navigator tracker is partially saturated by the RF pulse of the imaging sequence. We propose a new navigator algorithm that uses the phase profile of navigator echo data to improve the accuracy of the position detection even with the saturation effect. Our results show that the phase based algorithm is feasible and can be used in the navigator gated imaging to reduce motion artifacts. 1467. Head Movement Correction for MRI with a Single Camera Lei Qin1, 2, Peter van Gelderen1, Jacco de Zwart1, Fenghua Jin2, Yang Tao2, Jeff H. Duyn1 1AMRI, LFMI,NINDS,National Institutes of Health, Bethesda, Maryland, USA; 2University of Maryland, College Park, Maryland, USA We propose a new motion correction method for MRI using single camera that does not require calibration. A short training scan is required to relate the camera images to motion parameters derived from concurrently acquired EPI data. With this information, motion during an actual MRI scan with arbitrary pulse sequence is estimated by matching each newly captured camera image with one or more images from the training data. The average motion parameters of these training images served as motion estimate for that new image. Initial results show an overall accuracy of better than 0.3 mm translation and 0.3° rotation. 1468. Navigator Echo Analysis Hybridizing Magnitude and Phase Edge Detection Yuji Iwadate1, Kenichi Kanda1, Aki Yamazaki1, Susumu Kosugi1, Tetsuji Tsukamoto1 1GE Yokogawa Medical Systems, Hino, Japan Applicability of navigator echo is limited to the series where the main sequence FOV is distant from the navigator position, due to interference of RF pulses. We developed a hybrid algorithm utilizing both magnitude and phase information to detect the diaphragm position robustly. The proposed method could detect the diaphragm position correctly without RF interference affection. This method can be used to improve the abdominal image quality of the patients who have difficulty in breath hold. 1469. Spiral Projection Imaging Motion Correction Using Lines of Intersection Ken O. Johnson1, James G. Pipe1 1Barrow Neurological Institute, Phoenix, Arizona , USA In Spiral Projection Imaging, planes of data are collected to fill a sphere in k-space. Every plane intersects every other plane on a line. Therefore, any two planes can be compared, to determine which line from each plane is common between them. Planar orientations can then be deduced, that are consistent with the calculated lines of intersection. 1470. In-Plane Estimation Based Motion Correction for Spiral Projection Imaging Ryan K. Robison1, Eric Aboussouan1, James G. Pipe1 1Barrow Neurological Institute, Phoenix, USA Prior studies have asserted the potential of SPI (Spiral Projection Imaging) to accomplish 3D self-navigated motion estimation and correction. The goal of this study was to verify the capability of SPI to correct for three degrees of rotational motion. Through the application of an in-plane estimation based correction algorithm to simulated SPI data, it was determined that SPI is capable of producing a motion corrected 3D data set. Juergen Rahmer1, Jochen Keupp1, Shelton D. Caruthers2, 3 1Philips Research Europe, Hamburg, Germany; 2Washington University, St. Louis, Missouri, USA; 3Philips Medical Systems, Andover, Massachusetts, USA MR detection of low concentrations of imaging agents labeled with 19F often requires long signal averaging times and therefore motion compensation is desirable. Simultaneous acquisition of 1H and 19F signal allows self-navigated motion tracking using the stronger 1H signal. In this work, 3D isotropic radial imaging has been combined with a profile acquisition order based on the 2D golden means. Thereby, the intrinsic robustness of radial sequences against motion and undersampling can be combined with a flexible, motion-adaptive temporal frame rate. The technique is demonstrated in 19F/1H in-vivo scans with 1D translational motion and phantom scans with 3D rigid-body motion. 1472. k-Space Sampling for Motion Correction with Parallel Imaging Techniques Philip Martyn Robson1, Aaron K. Grant1, David C. Alsop1 1Beth Israel Deaconess Medical Center, Boston, Massachusetts, USA Motion related artifacts are a common cause of image corruption in MRI. For many protocols, even speeding image acquisition with parallel imaging is not sufficient to freeze motion. Rigid body motion can be detected with a variety of MRI and non MRI methods, and potentially used to correct for motion during reconstruction. However, the ability to faithfully reconstruct a motion corrected image depends upon the chosen k-space trajectory, the number of receive coils employed, and the time dependence of the motion. We implemented a general least squares reconstruction algorithm for motion correction using all redundant information in the receive coils. In simulated data, occasional, sudden translation and rotation can be fully corrected if a suitable k-space acquisition order is employed. 1473. Automatic Motion Correction of FSE Wrist Exams: Towards Full Clinical Implementation Armando Manduca1, David S. Lake1, Heidi A. Ward1, Jie Liang1, Lauren C. Edge1, Scott O. Stiving1, Natalia Khaylova1 1Mayo Clinic, Rochester, Minnesota, USA Fully automatic translational motion correction of high resolution FSE wrist exams, based on optimization of an image quality measure, was tested on all FSE wrist exams acquired on a clinical scanner over a 3 month period. Motion correction was automatically performed on all exams in an average time of 1 min 51 sec. 44% of the 160 series acquired were judged to have some sign of motion, and 91% of these were improved to some degree. A separate study in which the corrected exams are returned to the scanner and evaluated by radiologists is now under way. 1474. Extraction of Abdominal Motion for Molecular Imaging Frederik J. de Bruijn1, Munibe Simsek-Yildirim, 12, Aart J. Nederveen3, Muhammed Yildirim1, 4, Andre M.J. Sprengers3, Jaap Stoker3, Rolf M.J.N. Lamerichs1 1Philips Research, Eindhoven, Netherlands; 2Gebze Institute of Technology, Gebze, Turkey; 3Academic Medical Center, Amsterdam, Netherlands; 4Eindhoven University of Technology, Eindhoven, Netherlands MRI of the human bowel is generally complicated due to its motility and due to the intrinsically weak and often ambiguous tissue contrast. We present a method to extract colon motion from volumetric time sequences without the need for extensive bowel preparation. The method combines the use of chemical-shift induced tissue features with a recursive spatiotemporal method to generate spatially consistent motion patterns without the use of prior tissue-segmentation and -modelling. Its application is discussed to motion-corrected signal accumulation of 19F-based targeted contrast agents. 1475.
Motion Insensitive 3D T2 and T1-Weighted Imaging with a
Real-Time, Image-Based PROspective MOtion Ajit Shankaranarayanan1, Eric Han1, Cooper Roddey2, Nate White2, Reed Busse3, Joshua Kuperman2, Juan Santos4, Dan Rettmann5, Ehud Schmidt6, Anders Dale2 1GE Healthcare, Menlo Park, California , USA; 2UCSD, San Diego, California , USA; 3GE Healthcare, Madison, Wisconsin, USA; 4Stanford University, Palo Alto, California , USA; 5GE Healthcare, Rochester, USA; 6GE Healthcare, Boston, Massachusetts, USA A complete motion correction solution for 3D imaging has been presented here. The real time image-based 3D PROspective MOtion correction method (3D PROMO) previously presented has been integrated with a 3D T2-weighted imaging sequence (3D FSE acquisition) as well as a 3D T1 weighted imaging sequence (3D IR-SPGR acquisition.). In addition, an automated re-acquisition of left over motion-corrupted k-space views to further increase the robustness of PROMO has been described. In vivo scans demonstrate the efficacy of both 3D T2w and T1w PROMO in presence of motion. John Cooper Roddey1, Ajit Shankaranarayanan, Eric Han, Nathan White, Anders M. Dale 1University of California, San Diego, La Jolla, California , USA A region-of-interest (brain) in navigator images is automatically specified and tracked. Regions within the navigator FOV that move non-rigidly relative to the ROI are effectively ignored during tracking. Motion insensitive 3D MR images are acquired. |
Image Analysis - Parameter Mapping Hall D Wednesday 13:30-15:30 Weighted Imaging Study Andreas Ebel1, 2, Lara Stables3, Gail A. Kang4, Graham Glass, 34, Rachel Millin1, 2, Daniel McCoy1, 3, Peter Lorenzen1, 3, Yu Zhang1, 3, Wang Zhan1, 3, Michael W. Weiner1, 3, William Marks, 34, Norbert Schuff1, 3 1Center for Imaging of Neurodegenerative Diseases, San Francisco, California , USA; 2Northern California Institute for Research and Education, San Francisco, California , USA; 3University of California, San Francisco, California , USA; 4VA Medical Center, San Francisco, California , USA Susceptibility-weighted imaging (SWI) was used to assess abnormalities in the nigrostriatal system in Parkinson’s disease (PD). Abnormal SWI phase values, consistent with increased iron content in brain tissue, were observed in the caudate nucleus, putamen, and the subthalamic nucleus. Furthermore, the regional dispersion of SWI phase, measured in terms of entropy, strongly correlated with PD severity. Taken together, these results imply that SWI has potential for diagnosing PD, assessing disease severity, and tracking response to disease modifying interventions.
1512.
Phase-Based Quantification of Magnetic Susceptibility in
Healthy Volunteers and Patients with Kathryn E. Hammond1, Janine M. Lupo, Suchandrima Banerjee, Doug A. Kelley, Dan B. Vigneron, 1, Susan M. Chang, Pratik Mukherjee, Sarah J. Nelson, 1 1University of California San Francisco (UCSF) / Berkeley Joint Graduate Group in Bioengineering, San Francisco, California , USA This study presents a technique for quantifying magnetic susceptibilities c by post-processing the phase of gradient echo images acquired at 7T at a spatial resolution of 0.35x0.35x1mm. The c of hemorrhages and calcifications was calculated by comparing phase images of the B0 field to theoretical models of c effects. Model c-effects were also used to automatically separate hemorrhage from veins. Detection and characterization of c-shifted compounds is clinically relevant as a surrogate for brain injury following trauma or secondary to therapeutic radiation and may have an enormous impact on studies of iron deposition in neurodegenerative disorders. 1513. High Resolution Positive Contrast Via Post-Processing from Conventional 3D Imaging Hannes Dahnke1, Wei Liu2, Richard Bowtell3, Joseph A. Frank4 1Philips Research Europe, Hamburg, Germany; 2Philips Research North America, Briarcliff, USA; 3University of Nottingham, Nottingham, UK; 4National Institutes of Health, Bethesda, USA We demonstrate a method to calculate positive contrast from a conventional gradient echo image, without loss in resolution. The method combines the high resolution, up to now only achieved by means of dedicated positive contrast sequences, with the advantages of a post-processing method that requires no extra measurement time and less a-priori knowledge about the induced susceptibility. We show that it allows for the selective imaging of SPIO labeled cells at higher CNR than in the original image. First 7T data demonstrate the selective imaging of blood vessels by means of this method. 1514. Estimation of Sparse Magnetic Susceptibility Distributions from MRI Using Non-Linear Regularization Bryan Kressler1, 2, Ludovic de Rochefort2, Pascal Spincemaille2, Tian Liu1, 2, Yi Wang1, 2 1Cornell University, Ithaca, USA; 2Weill-Cornell Medical College, New York, USA Non-linear regularization techniques are used to estimate sparse magnetic susceptibility distributions from MRI field maps. Use of regularization terms that promote sparsity of susceptibilities or sparsity of boundaries between regions of different susceptibilities leads to reduced streaking artifacts, improved boundary delineation, and more accurate estimation of susceptibilities than conventional Tikhonov regularization. The techniques are demonstrated on gadolinium phantoms and agar gels containing superparamagnetic iron oxide particles, and yield estimates of the susceptibility of every voxel in the image. 1515. Estimation of Field Maps with Susceptibility Gradient Mapping Sunbok Lee1, Qun Zhao1 1University of Georgia, Athens, Georgia, USA In this study a new method based on the Susceptibility Gradient Mapping (SGM) is proposed for estimating a field map without a phase unwrapping procedure. The shift theorem of Fourier Transform shows that the SGM map and field map are equivalent with a proportional constant. Compared with traditional methods of acquiring a field map that needs two phase maps at two different TEs and requires a reliable phase unwrapping algorithm, only one phase map acquired with a single TE is required by the SGM method to generate a field map. 1516. Visceral Fat Assessment with 3D Dual Echo Dixon Technique Kajoli Banerjee Krishnan1, Rakesh Mullick1, Uday Patil1, Ajay Narayanan1, Anthony T. Vu2, Patrice Hervo3 1GE Global Research, Bangalore, India; 2GE Healthcare, Wisconsin, USA; 3GE Healthcare, Buc, France We have devised an acquisition-based threshold to automatically assess visceral fat fraction from fat-only MEDAL images. The threshold works on a slice-by-slice basis and has been applied to seven clinical datasets representing varying age, gender and body weight. The results are compared with manually drawn visceral masks on water-only MEDAL images that are used to demarcate the subcutaneous layer from the visceral region on the fat-only image. The automatic algorithm provides a systematic estimate of visceral fat within a fraction of 0.1 of the manually assisted method comparable to the range of values reported in the literature. 1517. Estimation of Water/fat Images, B0 Field Map and T2* Map Using VARPRO Diego Hernando1, Peter Kellman2, Justin Haldar1, Zhi-Pei Liang1 1University of Illinois at Urbana-Champaign, Urbana, Illinois, USA; 2National Institutes of Health, Bethesda, Maryland, USA T2* estimation in Dixon imaging is important for obtaining accurate water/fat intensity estimates when the relaxation effect cannot be neglected. Moreover, the T2* map can have diagnostic value of its own. Here we present a method for estimating B0- and T2*-maps along with water/fat images from Dixon acquisitions, by extending a recently proposed variable projection method. This method provides accurate estimates regardless of the nonconvexity of the corresponding estimation problem. Furthermore, an efficient approximate algorithm is derived based on Cramer-Rao bound analysis. The performance of the proposed methods has been validated using cardiac imaging data. 1518. Fully-Automated Abdominal Fat Quantification on Water-Saturated MRI Anqi Zhou1, Yao Ding2, Roderick W. McColl2, Paul T. Weatherall2, Qi Peng1 1UT Health Science Center at San Antonio, San Antonio, Texas, USA; 2UT Southwestern Medical Center at Dallas, Dallas, Texas, USA A fully automated abdominal fat quantification method on water-saturated MR images is presented here. This method combines automatic SAT and VAT segmentation with an automated gray-scale thresholding method for fat voxel determination. 64 slices MR images obtained from 8 subjects were tested using this fully-automated method. A slower semi-automated approach was also used as a reference technique. SAT and VAT quantified using the automated method were slightly smaller (4.4% and 2.7%, respectively) than those obtained by the reference method. About 95% of the slices were quantified with excellent accuracy. We conclude that fully-automated abdominal fat quantification is feasible. 1519. Intensity Inhomogeneity Correction in Two Point Dixon Imaging Olof Dahlqvist Leinhard1, 2, Andreas Johansson1, Joakim Rydell1, 2, Magnus Borga1, 2, Peter Lundberg1, 2 1Linköping University, Linköping, Sweden; 2Center for Medical Image Science and Visualization (CMIV), Linköping, Sweden A simple and effective approach for correcting intensity inhomogeneity in fat volumes reconstructed from a two point Dixon protocol is to identify pure adipose tissue values and interpolate a correction field with normalized convolution from these. An implementation has been made and tested on 50 acquisitions from 20 different subjects showing high stability. 1520. Phase-Correction Algorithms for Fat-Water Separation with Dual-Echo Sequences Maria A. Schmidt1 1St George's Hospital NHS Trust, London, UK Dual-echo sequences for two-point fat-water separation often result in images characterised by phase changes along the readout direction due to asymmetric echoes and eddy currents. This work compares phase-correction algorithms in test objects and volunteers, using dual-echo sequences. Pre-processing to remove linear phase drifts and different quality maps to guide the region-growing process were investigated. Algorithms starting from low resolution images and stepping up towards the original resolution were found to provide the best performance. Phase shifts up to 4.1º/pixel were found, and for the largest shifts a linear correction of the phase map was required for correct fat-water separation. Joakim Rydell1, 2, Andreas Johansson1, Olof Dahlqvist Leinhard1, 2, Hans Knutsson1, 2, Gunnar Farnebäck1, Peter Lundberg1, 2, Magnus Borga1, 2 1Linköping University, Linköping, Sweden; 2Center for Medical Image Science and Visualization (CMIV), Linköping, Sweden Three dimensional phase sensitive reconstruction on two point Dixon volumes has been implemented with use of the inverse gradient. The results has been compared with the inverse gradient method in two dimensions as well as with the well established region growing method proposed by Ma. The inverse gradient method in 3D is able to unwrap the phase field in uncertain regions where the region growing method and the inverse gradient method in 2D come to a stop. 1522. R2* Measurement Errors at Ultrahigh Field in the Presence of Nonlinear B0 Inhomogeneities Xiangyu Yang1, Petra Schmalbrock1, Steffen Sammet1, Michael V. Knopp1 1The Ohio State University, Columbus, Ohio, USA Accurate measurement of R2* was hindered by B0-inhomogeneity-induced errors in 2D MRI. While such inhomogeneities can be neglected or corrected as a linear gradient at low field, nonlinearity must be taken into consideration at ultrahigh field. In this study, we demonstrated that all three commonly-used relaxation models (exponential, exponential-plus-constant, and sinc-modulated exponential) generate severe biases in R2* estimation in the presence of quadratic B0 inhomogeneity. Phantom study results suggest that such biases can be largely compensated by a quadratic correction using measured B0 map, even when the actual B0 inhomogeneity is not strictly quadratic. 1523. Comparison of ASSETx1 and Corrected Power Images for T2 Quantification Andrew James Patterson1, Martin John Graves1, Tjun Tang1, Jonathan H. Gillard1 1University of Cambridge and Addenbrooke's Hospital, Cambridge, UK The accuracy of T2 quantification is limited by the decay of the signal into the background noise. In low SNR conditions an optimal multi-coil B1 reconstruction method decreases the mean background noise in comparison to the standard root-sum-of-squares (RSS) reconstruction. We compared the RSS and optimal reconstruction method with a previously published power corrected algorithm. Quantitative T2 measurements obtained in a phantom demonstrated that the optimal reconstruction and the power method are both effective and comparable at improving accuracy and precision. Berengere M. Aubert-Broche1, Christophe Grova1, Bruce G. Pike1, Louis D. Collins1 1MNI, Montreal, Canada The aims of the present study were to use rapid volumetric relaxometry methods to (1) quantify T1 and T2 values in specific cortical grey matter (GM) regions for healthy volunteers at 1.5T and (2) aggregate regions showing statistically dependent relaxation times to define cortical areas. To study the spatial statistical dependencies seen in cortical GM relaxation times, the method combines standardized anatomic segmentation, correspondence analysis and ascending hierarchical clustering. Groups of structures that are related statistically are identified and correspond to large cortical regions that have been previously defined anatomically. 1525. Determining Uncertainty in Estimates of Relaxation Time (T2) and Proton Density (S0) Derived fromT2 –weighted MRI Using Bootstrap Method Ruiwang Huang1 1Research Center Juelich, Juelich, Germany T2–weighted spin echo MRI permits determination of the transverse relaxation time (T2) and proton density (S0) within image voxels. Several MR sequences for T2 mapping have been proposed, but little effort has been made to estimate the uncertainty of T2–map. Here, we use the bootstrap approach to investigate uncertainties of (S0, T2) from real T2-weighted MR human brain images and to examine the influence of voxel size on the reliability of (S0, T2) derived from fitting a mono–exponential function.
1526.
Polar Decomposition Radio-Frequency Current Density Imaging
with Dual-Unwrapping: Simulation and Weijing Ma1, 2, Dinghui Wang1, Tim P. DeMonte2, Adrian I. Nachman1, Michael L. G Joy1 1University of Toronto, Toronto, Canada; 2Field Metrica Inc., Toronto, Canada Polar Decomposition Radio-frequency Current Density Imaging (PD-RFCDI) is an imaging technique that non-invasively measures current density components inside a sample using a MRI imager. However, due to the requirement of PD-RFCDI to avoid phase wraps, this method was noisier than the rotary echo method and was not implemented on clinical MRI scanners. This work proposes a novel dual-unwrapping technique that removes these constraints on PD-RFCDI. Both the simulation and experiment were used to verify this algorithm. Dinghui Wang1, 2, Weiing Ma1, Tim P. DeMonte2, Adrian I. Nachman1, Michael L. G Joy1 1University of Toronto, Toronto, Canada; 2Field Merica Inc., Toronto, Canada Previously implemented radio frequency current density imaging (RF-CDI) techniques can evaluate the component of the injected Larmor frequency current density in the direction of the static magnetic field only under strict assumption which can be easily violated in biomedical applications. The goal of this work is to experimentally verify a new reconstruction method based on one 180-degree rotation. The results show that with this method one component of the current density can be fully recovered by the two sample orientations. 1528. A Robust Method for Estimation of Regional Pulmonary Parameters in the Presence of Noise Richard A. Guyer1, Michael D. Hellman2, Kiarash Emami1, Stephen Kadlecek1, Robert V. Cadman1, Jiangsheng Yu1, Vahid Vadhat1, Masaru Ishii1, 3, John MacDuffie Woodburn1, Hans Hyonchang Kim1, Rahim R. Rizi1 1University of Pennsylvania, Philadelphia, Pennsylvania, USA; 2Jefferson Medical College of Thomas Jefferson University, Philadelphia, Pennsylvania, USA; 3Johns Hopkins University, Baltimore, Maryland, USA Principal component analysis (PCA) permits data points in a series of hyperpolarized gas MR images to be clustered based upon their similar signal kinetics rather than their spatial proximity in the images. This allows points to be grouped and averaged to control noise without degrading the resolution of the image. As a result, more reliable estimations of regional lung parameters, such as the alveolar partial pressure of oxygen and the oxygen depletion rate, can be obtained. This is demonstrated with a synthetic dataset derived from images of a rabbit’s lungs. 1529. Noise Statistics, Discretization Artifacts, and SNR Measurements at Very Low Noise Levels Olaf Dietrich1, José G. Raya1, Maximilian F. Reiser1 1Ludwig Maximilian University of Munich, Munich, Germany The well-known noise distributions of MRI data (Rayleigh, Rician, or non-central chi-distribution) describe the probability density of real-valued (i.e. floating-point) intensity signals; however, MR image data is typically discretized to integers before visualization or archiving in the DICOM format. Depending on the scaling factors used for the discretization and the signal-to-noise ratio (SNR), very low noise levels with substantial discretization artifacts can occur. The purpose of this study was to analyze the consequences of such discretization artifacts and to suggest an improved method for noise and SNR measurements in the presence of very low noise levels. 1530. A Method for Automatic Estimation of Noise Variance and SNR Philip James Beatty1 1GE Healthcare, Menlo Park, California , USA This work proposes a straightforward method to automatically estimate the noise variance of each receive channel directly from parallel imaging calibration data. The proposed method can be used to identify low SNR channels that do not contribute in a meaningful way to the image reconstruction. By using only data acquired during image acquisition, the need for a separate scan to determine noise variance is eliminated and there is no possibility of incorrect registration between a noise variance acquisition and the image acquisition. |
Hall D Wednesday 13:30-15:30 Åsa Carlsson1, 2, Maria Ljungberg1, 2, Göran Starck1, 2, Eva Forssell-Aronsson1, 2 1Göteborg University, Gothenburg, Sweden; 2Sahlgrenska University Hospital, Gothenburg, Sweden The spurious echo artifact is often encountered in in vivo MRS and is once again actualized with application of MRS in more susceptibility influenced regions. It originates from signals excited in regions outside the VOI that refocus into an echo during signal acquisition. Here k-space formalism was used as a new approach to analyze the spurious echo artifact of the PRESS sequence, and in vitro measurements were performed for verifications and characterisations. Different coherences were separately refocused and studied. The FID of the second refocusing pulse was most probable to cause an artifact and also caused the most pronounced.
1552.
Exploiting the Chemical Shift Displacement Effect in the
Detection of Glutamate and Glutamine (Glx) Atiyah Yahya1, 2, B. Gino Fallone1, 2 1Cross Cancer Institute, Edmonton, Canada; 2University of Alberta, Edmonton, Canada The observation of the collective levels of glutamate/glutamine (Glx) by proton magnetic resonance spectroscopy is of value in the study of cancer, neurodegenerative and psychiatric disorders. A PRESS sequence is presented in this work that can yield high signal, comparable to that which can be achieved by a one-pulse acquire sequence, from the C2 protons of Glx at long echo times which is favourable for the decay of macromolecule signal. The efficacy of the sequence was verified on phantoms and in vivo on brain. Changho Choi1 1University of Texas Southwestern Medical Center, Dallas, USA A preliminary result of a numerical study for suppression of myo-inositol (mI) for detection of the glycine (Gly) singlet at 3.55 ppm for 1.5T is reported. Multiple refocusing was investigated with density-matrix simulation to calculate the responses of the strongly-coupled resonances of mI. The study indicated that suppression of mI enhances with increasing number of 180° pulses, N. It is predicted that, for an mI-to-Gly concentration ratio of 8, ignoring T1 and T2 effects, mI-to-Gly peak amplitude ratio is 39%, 35%, 31%, 12% and 4% at TE = 108, 146, 176, 239, and 298 ms, for N = 3, 4, 5, 6, and 7, respectively. 1554. Glutamate and Glutamine Discrimination Via Constant Echo Time Difference SpectroscopyJeff Snyder1, Richard B. Thompson1, Alan H. Wilman1 1University of Alberta, Edmonton, Canada This study investigates the discrimination of glutamate and glutamine using a simple constant echo time subtraction spectroscopy method based on PRESS. The technique relies on coupled spin signal variation differences between glutamate and glutamine along a constant TE line. The theoretical simulations calculated 96% of the glutamate/glutamine signal in the 2.06-2.15 ppm was due to glutamate. Phantom and volunteer measurements were performed to support the simulations and validate the technique. The results illustrate a large signal remaining in the above noted ppm range, with other resonances suppressed. 1555. 13C MRS Study of Human Ornithine Transcarbamylase Deficiency Andrea Lynne Gropman1, Napapon Sailasuta2, Lawrence Robertson3, Kent Harris4, Peter S. Allen5, Brian D. Ross 1Children's National Medical Center, Washington, USA; 2Huntington Medical Research Institutes, Pasadena, California , USA; 3Spin Dynamics, South Pasadena, USA; 4Huntington Medical Research Institutes, Pasadena, USA; 5University of Alberta, Edmonton, Canada Ornithine transcarbamylase deficiency leads to hyperammonemic encephalopathy. 1H MRS shows elevations in glutamine and decreases in myoinositol and choline. Specificity of 1H MRS suffers from complex peak patterns due to J-coupling. 13C MRS can quantitate distinct signals from Glu and Gln. We imaged three subjects with partial OTCD and compared to controls using natural abundance 13C MRS followed by intravenous low dose 99% 1-13C glucose infusion. 1-13C glucose appeared in the 13C brain spectra; enrichment of C1 – thru C5 glutamate and glutamine occurred. Excess enrichment in glutamine C4 and C2 in subjects was consistent with abnormality in glial glutamine synthesis.
1556.
13C Isotopomer Metabolic Modeling: Automatic Generation of
the Mathematical Model
Julien Valette1, Alexander Shestov1, Kamil Ugurbil1, Pierre-Gilles Henry1 1CMRR, University of Minnesota, Minneapolis, Minnesota, USA Because they describe 13C spectra fine structure, 13C isotopomer metabolic models take full advantage of the biochemical information that can be measured by 13C NMR during an infusion of labelled substrate, allowing the measurement of metabolic fluxes with an increased accuracy. In this work, a concept related to isotopomers, dubbed “probabimer”, which allows reducing the number of equations compared to isotopomer models, is briefly introduced. Then a program is presented that automatically generates and simplifies the differential equations describing the time evolution of probabimers. This program is finally exemplified on the neuron-glia model. Vladimir Mlynárik1, Cristina Cudalbu1, Lijing Xin1, Rolf Gruetter1, 2 1Ecole Polytechnique Fédérale de Lausanne, Lausanne, Switzerland; 2Universities of Lausanne and Geneva, Switzerland Ultra-short-echo time spectra of rat brain were obtained at 14.1 Tesla. Metabolite concentrations were calculated using LCModel and were compared with those obtained at 9.4 T. It was found that quantitation of metabolites, which are poorly characterized in proton MR spectra in vivo (such as alanine, glycerophosphocholine, GABA, glycine and NAAG), was improved.
1558.
Determination of the Glutamate-Glutamine Cycling Flux Using
Two-Compartment Dynamic Metabolic Jun Shen1, Douglas Rothman, Kevin Behar 1NIMH, Bethesda, Maryland, USA 13C MRS combined with the infusion of [1-13C]glucose has been used to measure the rate of the glutamate-glutamine cycling. Here we evaluated the accuracy and precision of 13C MRS measurement of this cycling rate using Monte Carlo analysis of a previously published two-compartment metabolic model of the glutamate-glutamine cycle which takes into account astroglial dilution. The Monte Carlo analysis confirmed that the neuronal TCA cycle and glutamate-glutamine cycle fluxes can be determined reliably from the 13C enrichment time courses measured with 13C MRS during infusion of [1-13C]glucose. Alexander Gussew1, Reinhard Rzanny1, Hans Christoph Scholle2, Werner Alois Kaiser1, Jürgen Rainer Reichenbach1 1Friedrich-Schiller University of Jena, Jena, Germany; 2Friedrich-Schiller-University of Jena, Jena, Germany Cortical alternations of glutamate concentrations induced by cyclic peripheral painful heat stimulations at the left forearm were measured by means of time resolved in vivo 1H-fMRS in the left insular cortex of 3 volunteers. Synchronization between the stimulation unit and the MR-scanner allowed acquisition of spectra at two different time points (stimulus and rest) of the stimulation cycle. For all volunteers spectra acquired during the stimulus showed increases of glutamate concentration (up to 17%) compared to the reference spectrum without stimulation. However, observed differences of glutamate concentration during the resting period among the volunteers indicate interindividual variability. 1560. In Vivo 13C Spectroscopy of Human Brain on a Clinical 3T Scanner Using [2-13C]Glucose Infusion Shizhe Steve Li1, Yang Zhang1, Jehoon Yang1, Maria Ferraris Araneta1, Robert B. Innis1, Jun Shen1 1NIH, Bethesda, Maryland, USA Contaminatiion from subcutaneous lipid signals and strong RF decoupling power requirement are two major technical difficulties in human brain 13C spectroscopy when [1-13C]glucose is used. In a recent study of monkey brain, we proposed a novel strategy of using [2-13C]glucose infusion and detecting 13C label incorporation into the caboxylic/amide spectral region to overcome those difficulties. Here we report results of human brain 13C spectroscopy using [2-13C]glucose infusion on a 3 Tesla clinical scanner. With pseudo noise decoupling, we were able to acquire 13C spectra of carboxylic/amide carbons with reduced average RF power deposition and without lipid contamination. 1561. A Strategy for 13C MRS Study of Human Frontal Lobes Napapon Sailasuta1, Larry Robertson2, Kent Harris1, Brian Ross1 1Huntington Medical Research Institutes, Pasadena, California , USA; 2Spin Dynamics, South Pasadena, California , USA Evaluation of several low power noise proton decoupled carbon detect MRS method for glutamatergic neurotransmission of the human frontal lobes. Stefan Kirsch1, William Edmund Hull1 1German Cancer Research Center (DKFZ), Heidelberg, Germany In contrast to the decay of a conventional pulse-acquire FID, the time-domain signal from of a CRAZED-type pulse sequence for coherence order n > 1 increases from initially zero to a maximum, followed by an exponential decay. We show that the finite duration of the coherence rephasing gradient and any subsequent recovery delays lead to loss of the initial portion of the time-domain signal. This so-called “time-zero truncation artefact” results in an altered frequency-domain lineshape with incorrect integral and linewidth. Correct analysis requires time-domain fitting of right-shifted data. 1563. In Vivo Human Whole Cerebellum MRS Under Severe Field Inhomogeneity with IDQC Method Tianliang Gu1, 2, Zhong Chen1, Xiaoxu Liu1, Ling-chih Lin1, Jianhui Zhong1 1University of Rochester, Rochester, New York, USA An intermolecular double quantum coherence (iDQC) 2D MRS pulse sequence was created and implemented on a 3T scanner for acquisition of human brain 1H spectra in regions susceptible to field inhomogeneity such as the cerebellum. High resolution 1D MRS could be obtained in a few minutes with iDQC over the whole cerebellum, whereas conventional single voxel 1D MRS was working successfully only when regions much smaller were covered. NAA/Cr and Cho/Cr measured with both methods in 5 healthy subjects agreed well with each other. 1564. Two-Dimensional Zero-Quantum Coherence 1H NMR Spectroscopy of Glutamate and Glutamine Sarah Rebecca Snyder1, Stefan Kirsch1, Katharina Kraus1, Peter Bachert1 1German Cancer Research Center, Heidelberg, Germany Strong signal overlap of glutamate and glutamine creates a spectroscopic imaging challenge. To more precisely differentiate the two, we recorded series of ¹H NMR stimulated-echo acquisition mode (STEAM) spectra with varying mixing time intervals from Glu and Gln model solutions on a 3 Tesla whole-body MR tomograph. For scalar-coupled molecules, this leads to an intrinsic phase modulation frequency of zero-quantum coherences (ZQC). The outer peaks of the α -CH triplet of both glutamate and glutamine could be clearly identified due to their chemical shifts and ZQC modulation. This provides an additional parameter for the identification of glutamate and glutamine.
1565.
Double Quantum Spectroscopy Using Phase Rotation at 7T
Saadallah Ramadan1, Eva M. Ratai2, Lawrence L. Wald2, Graham C. Wiggins2, Carolyn E. Mountford3 1Brigham and Women's Hospital, Boston, Massachusetts, USA; 2Massachusetts General Hospital, Boston, Massachusetts, USA; 3Brigham and Womens Hospital, Boston, Massachusetts, USA 1566. Sensitivity of Single Shot Multiple J-Coupled Metabolite Detection Using Dual Sel-MQC (D-Sel-MQC) Editing Gerd Melkus1, Philipp Mörchel1, Markus Kotas1, Michael Flentje1, Peter Michael Jakob1 1University of Würzburg, Würzburg, Germany The spectral editing technique Sel-MQC (Selective Multiple Quantum Coherence) is extended by an additional spectral selective pulse to increase the number of detectable J-coupled resonances in a single scan. The first frequency selective pulse is placed on the 1.1-1.7 ppm region to excite lactate and alanin, the second frequency selective pulse is placed on the 1.8-2.4 ppm region to excite g-aminobutyric acid, glutamate and glutamine. The timing of the sequence is defined by J of the edited resonances. Double-quantum coherences are filtered using gradients. The sensitivity of the one scan editing sequence is evaluated using simulations and phantom experiments. 1567. Shortening of Metabolite Relaxation Times of Prefrontal Cortex in Mild Cognitive Impairment Antonio Napolitano1, 2, Federico Giove, 23, Gianfranco Spalletta4, Dorothee P. Auer1, Bruno Maraviglia2, 3 1University of Nottingham, Nottingham, UK; 2La Sapienza University, Rome, Italy; 3Enrico Fermi Center, Rome, Italy; 4Santa Lucia Foundation, Rome, Italy A spectroscopy study was performed to measure the T2 values of NAA, tCrin Alzheimer’s patients and in controls. The amplitudes of NAA, tCr and Cho from 16 spectra with different TE were fitted and the T2 values of control subjects were compared with those of Alzheimer’s patients. We found a 13% decrease in T2 of NAA (p<0.035), a 9% decrease in T2 of tCr (p<0.01) and a 15% decrease in T2 of Cho (p<0.033). Such decreases we found in the metabolite relaxation times can be correlated to the presence of elevated iron levels associated with neurodegenerative disorders such as Alzheimer’s disease.
1568.
Bone Marrow 1D and 2D Correlation MR Spectroscopy at 7T
Saadallah Ramadan1, Eva M. Ratai2, Lawrence L. Wald2, Graham C. Wiggins2, Carolyn E. Mountford3 1Brigham and Women's Hospital, Boston, Massachusetts, USA; 2Massachusetts General Hospital, Boston, Massachusetts, USA; 3Brigham and Womens Hospital, Boston, Massachusetts, USA When compared to low field MRS, high field MRS offers additional benefits, including spectral dispersion. However, benefits come with more technological challenges from an RF, Bo and pulse sequence design perspective. An attempt to implement 1D and 2D spectroscopic correlation experiments on a 7T whole body magnet is carried out, with application to tibial bone marrow. The resulting spectra are analyzed and compared to existing spectra at lower fields.
1569.
1D and 2D Correlation Spectroscopy of Muscle at 7T
Saadallah Ramadan1, Eva M. Ratai2, Lawrence L. Wald2, Graham C. Wiggins2, Carolyn E. Mountford3 1Brigham and Women's Hospital, Boston, Massachusetts, USA; 2Massachusetts General Hospital, Boston, Massachusetts, USA; 3Brigham and Womens Hospital, Boston, Massachusetts, USA 1570. 1H MRS in the Human Calf with a Spatially Selective RF Surface Coil Maria Alfonsetti1, Claudia Testa2, Stefano Iotti2, Emil Malucelli2, Valeria Clementi2, Bruno Barbiroli2, Giuseppe Placidi1, Antonello Sotgiu1, Marcello Alecci1 1Università di L'Aquila, L'Aquila, Italy; 2Università di Bologna, Bologna, Italy In this work we describe the use of transverse field RF coils to improve proton MRS SNR and spatial selectivity at 1.5T in the resting calf of a healthy volunteer. The RF coil of diameter 10 cm allows in the calf muscle an improved 1H-PRESS SNR along the A/P direction within a given depth from the surface. These features should be of benefit in clinical MRS studies of muscle metabolism and in brain metabolites quantification. 1571. An Integrated Analysis Platform for Magnetic Resonance Spectroscopy Jordan Hovdebo1, 2, Herve Momo-Jeufack2, Michael West2, Lawrence Ryner1, 2 1National Research Council Institute for Biodiagnostics, Winnipeg, Canada; 2University of Manitoba, Winnipeg, Canada We present a new system for individual and population analyses of magnetic resonance spectroscopy data. The system consists of two parts: a database for storage, and a client application for accessing the information and performing analyses. Combined, these two parts give an analysis platform capable of managing the large volumes of information generated and sufficiently flexible to accommodate several sources of clinically-relevant data. 1572. On the Importance of Quality Assurance in Spectroscopic Imaging of Prostate Mahir Sinan Ozdemir1, Yves De Deene2, Geert Villeirs2, Eric Achten2, Ignace Lemahieu1 1Gent University, Gent, Belgium; 2Gent University Hospital, Gent, Belgium Spectroscopic imaging of prostate (SI) aided by endorectal coils (ER) has gained considerable interest recently as it provides high SNR and spatial resolution and allows one to detect signal changes, in particular, of the metabolites such as citrate (Cit) and choline (Cho). Although the diagnostic statements are based on the obtained data and the quality of the data and accuracy of the methods used are important for a correct diagnosis, a routine quality assurance of prostate spectroscopy is lacking. In this contribution, we duplicated a typical clinical prostate SI as performed commonly by using a homogenous phantom to assess the accuracy of ratio-based results. 1573.
Benign and Atypical Meningioma Metabolic Differences by
HR-MAS Molecular Profiling
Daniel Monleon1, Jose Manuel Morales2, Jose Gonzalez, Fernando Talamantes, Concha Lopez-Gines2, Miguel Cerda-Nicolas2, Bernardo Celda2, 3 1Fundacion Investigacion Hospital Clinico Universitario Valencia, Valencia, Spain; 2University of Valencia, Valencia, Spain; 3Centro de Investigación Biomédica en Red de Bioingeniería, Biomateriales y Nanomedicina (CIBER-BBN), Spain Meningiomas are neoplasms that arise from the leptomeningeal covering of the brain and spinal cord, accounting for 15%–20% of CNS tumors. Although the majority of these tumors are histologically benign, some meningiomas show signs of malignancy such as marked vascularity, mitotic figures, nuclear pleomorphism, focal necrosis, or infiltration to the adjacent brain. The WHO classifies meningiomas into three histologic grades: benign, atypical, and anaplastic in accordance with the clinical prognosis. Atypical and anaplastic meningiomas tend to recur. In this communication, we show differences between benign and atypical meningiomas in HRMAS molecular profiles of meningioma biopsies. 1574. Three-Dimensional Isotropic Filter Design with Arbitrary Pass-Band and Stop-Band Specifications Joonsung Lee1, Elfar Adalsteinsson1 1Massachusetts Institute of Technology, Cambridge, Massachusetts, USA Variable-density sampling of k-space with accompanying proportional filtering, offers improved spatial impulse response without SNR tradeoffs . Prior work relied on 1-D filter specifications, a widely investigated topic, while the general three-dimensional case is much less studied. While three dimensional filters can be derived by simple transformations from 1D, optimality of the filter will in general be lost. We present here an algorithm for the direct design of an optimal, spherically symmetric, three-dimensional filter with arbitrary extent in k-domain, and pass-band (i.e. voxel size) and stop-band (side lobe suppression) specified in the image domain. 1575. Optimal Phased Array Combination for Spectroscopy Mark Bydder1, Gavin Hamilton1 1University of California San Diego, San Diego, California , USA A method is described for making a weighted linear combination of the spectra acquired by a phased array coil. Unlike previous combination methods, no reference points in the data are chosen to represent the coil weights. Instead all the data points are used, which results in more reliable estimation. |
NMR of Cells, Body Fluids & Other Hall D Wednesday 13:30-15:30 Pratima Tripathi1, Ashish Gupta1, Raja Roy1, S K. Mandal1, Rajan Saxena2, S K. Yachha2, Chunni Lal Khetrapal1 1Centre of Biomedical Magnetic Resonance, Lucknow, India; 2Sanjay Ghandhi Post-graduate Institute of Medical Sciences, Lucknow, India 1H NMR spectroscopy has been used to study the metabolic profile of serum in Fulminant Hepatic Failure (FHF) patients and liver grafts (LG). Both in FHF and LG concentrations of alanine, lysine, glutamine, tyrosine, histidine and phenylalanine were significantly elevated in fatal cases when compared to recovered cases. Additionally in LG lactate, methionine and citrate were significantly more in fatal cases compared to successful cases. Principal Component Analysis showed lysine, glutamine, tyrosine and histidine to be the early discriminants for predicting the status of both FHF and LG.
1628.
Metabolic Profiling of Renal Cell Carcinoma Studied by
High-Resolution 1H NMR Spectroscopy on Human Hongchang Gao1, Baijun Dong2, Xia Liu1, Yiran Huang2, Donghai Lin1 1Shanghai Institute of Materia Medica, the Chinese Academy of Sciences, Shanghai, People's Republic of China; 2Renji Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, People's Republic of China 1H NMR-based metabonomics was applied in human serum samples to characterize the metabolic profiling of the renal cell carcinoma (RCC) tumor. The results showed that (a) the early RCC patients can be well distinguished from the healthy humans, and (b) the RCC patients after nephrectomy can be discriminated well from those before nephrectomy. Compared to healthy human serum, the RCC serum displayed higher levels of lipid (mainly very low-density lipoproteins), lactate, alanine, pyruvate, glycerol, and unsaturated lipid, but declined levels of acetoacetate, choline, and glucose/glycogen. Such systemic changes of metabolite concentrations were reversed after nephrectomy. 1629. Study on Vanadyl Sulfate Toxicity Using NMR-Based Metabonomics Jiyang Dong1, Xiaoxia Dai1, Hongting Cao1, Zhong Chen1 1Xiamen University, Xiamen, People's Republic of China This study investigated the toxicity effects of VOSO4 oral administration in male Wistar rats using 1H NMR-based metabonomic analysis and clinical chemistry analysis. Urine was collected and their 1H NMR spectra were acquired, and then subjected to multi-variable statistical analysis. Compared to control groups, urinary excretion of lactate, TMAO, creatinine, taurine and hippurate increased following VOSO4 dosing, with concomitant decrease in the level of acetate and succinate. The dosed groups can be readily discriminated from the controls by principle component analysis. All of the results show that VOSO4 can affect energy metabolism process and induce hepatic and renal lesion. 1630. Phospholipid Composition of Postmortem Schizophrenic Brain by 31P NMR Spectroscopy Richard A. Komoroski1, 2, John M. Pearce1, 2, Robert E. Mrak2 1University of Cincinnati, Cincinnati, Ohio, USA; 2University of Arkansas for Medical Sciences, Little Rock, Arkansas, USA Cell membrane abnormalities due to changes in phospholipid (PL) composition and metabolism have been implicated in schizophrenia pathogenesis. High resolution 31P NMR spectroscopy was used to characterize PLs of gray matter in three regions of postmortem brain for 20 schizophrenic and 20 control subjects. There were no statistically significant differences (p<0.05) between the groups for any of the individual PLs. The sum of all phosphatidylethanolamine headgroups was significantly lower (p=0.047) for schizophrenics than for controls in frontal cortex. The present results are minimally correlated with previous results for aqueous PL metabolites on these same samples. Alina Nicolescu1, Tedros Bezabeh2, Brion Dolenko2, Lorena Ivona Stefan3, Coziana Ciurtin4, Eugenia Kovacs5, Ian C. P. Smith2, Calin Deleanu6 1 “Petru Poni” Institute of Macromolecular Chemistry, Iasi, Romania; 2Institute for Biodiagnostics, National Research Council, Winnipeg, Canada; 3Craiova Clinical Hospital, Craiova, Romania; 4 “Dr. I. Cantacuzino” Clinical Hospitat, Bucharest, Romania; 5 “Carol Davila” University of Medicine and Pharmacy, Bucharest, Romania; 6Institute of Organic Chemistry, Bucharest, Romania While 1H NMR spectroscopy of urine can detect glucose in patients with type II diabetes mellitus, such information may not help much for the early diagnosis of diabetes. Changes in other metabolites present in urine may precede glucosuria and 1H NMR spectroscopy can play a role in the assessment of such changes. In patients already exhibiting glucosoria, the spectrum is dominated by signals from glucose. In such instances, conventional methods of analysis are not helpful. In this study, the use of a statistical classification strategy (SCS) has resulted in high accuracy in classifying urine spectra of normal controls from those of diabetic patients using the non-glucose spectral regions.
1632.
High Resolution Nmr Based Analysis of Serum Lipids in
Duchenne Muscular Dystrophy Patients and Its Niraj Kumar Srivastava1, Sunil Pradhan1, Balraj Mittal1, Raja Roy2 1Sanjay Ghandhi Post-graduate Institute of Medical Sciences, Lucknow, India; 2Center of Biomedical Magnetic Resonance, Lucknow, India Lipid extract of serum of thirty-one clinically, genetically and biochemically proven Duchenne Muscular Dystrophy patients (DMD) was studied by 1H NMR spectroscopy. Triglycerides, phospholipids, free-cholesterol, cholesterol-esters and total cholesterol were 2 to 3 times higher in DMD patients than controls. Ratio of free-cholesterol to cholesterol-esters was also significantly higher. Ten cases of DMD that showed negative gene-deletion-test due to point-mutation, demonstrated similar lipid profile like proven-cases in NMR analysis. Among all these variables, there was a consistent increase in the quantity of phospholipids in DMD patients when compared to controls. These observations indicate the potential of NMR to discriminate DMD. 1633. Multiple-Oocyte in Vivo Localized MR Spectroscopy Verena Hoerr1, Armin Purea1, Cornelius Faber1 1University of Wuerzburg, Wuerzburg, Germany A novel approach to intracellular MR spectroscopy, which combines solvent-localized (SOLO) NMR with chemical shift reagents, allowed for recording of intracellular MR spectra from a sample with 100 Xenopus leavis oocytes. Upon selective excitation of either intra or extracellular water signal, only intra or extracellular components were observed, respectively. SOLO afforded a 30-fold sensitivity gain compared to PRESS. Since only a single-axis gradient is required, the method was also implemented on a high-resolution NMR-system with cryo-probe. The method is readily applicable on most NMR spectrometers, making novel applications in pharmaceutics and developmental biology possible. 1634. Development of the Layered Dumbbell Shape Phantom for MRS Quality Assurance Dong-Cheol Woo1, Sang-Soo Kim1, Hyang-Shuk Rhim1, Hyun-Man Baik2, Orhan Nalcioglu2, Geon-Ho Jahng3, Bo-Young Choe1 1The Catholic University of Korea, Seoul, Republic of Korea; 2University of California, Irvine, Irvine, USA; 3East-West Neo Medical Center, Kyung-Hee University, Seoul, Republic of Korea The purpose of this study was to develop a standard MRS phantom, to establish a QA protocol with principal factors and to perform QA of an MR system. For the QA of MRS, seven principal factors were selected and measured. The form of the phantom was made of acrylic resin and the layered dumbbell shape vial was filled with various metabolite and all experiments were performed on two 3T systems. All MRS QA factors of each MRI/MRS system were calibrated. The follow-up study showed that the QA factor values in September 2007 were superior to those obtained in June 2007. 1635. Silicon-Cylinder Nano Liter NMR-Probe for Biomedical Analysis Systems Stefan Leidich1, Thomas Riemer2 1Chemnitz University of Technology, Chemnitz, Germany; 2University of Leipzig, Leipzig, Germany We present a new NMR-micro-probe designed for the analysis of sample volumes in the range of 10 nl to 50 nl. It is based on spiral micro-coils in a Helmholtz configuration which is housed in a silicon cylinder for susceptibility matching purpose.
1636.
Development of a Diagnostic Test Based on Nanoscale Distance
Measurements Between the Fatty Dariush Hinderberger1, Marcos Gelos2 1Max Planck Institute for Polymer Research, Mainz, Germany; 2Augusta-Kranken-Anstalt, Bochum, Germany We characterize the affinity of the fatty acid binding sites by exposing HSA in different ratios to paramagnetic fatty acids spin labels. We employ double-electron-electron resonance (DEER), a modern method of pulse EPR that measures dipolar couplings between individual spin labels. Through the known r-3-dependence of the dipolar coupling the distances r between the spin labels are obtained. Data analysis is so far in an experimental stage but we aim at standardizing the method and developing it into a diagnostic tool by expanding the studies from simple HSA in solution to actual blood samples from patients with different disease backgrounds. |
Hall D Wednesday 13:30-15:30 1673. Measuring Dendritic Cell Migration to Lymph Nodes Jonatan Ashchar Snir1, Greg A. Dekaban2, Peta O’Connell2, Paula J. Foster1 1Robarts Research Institute, London, Canada; 2Robarts research Institute, London, Canada The immune system plays an important role in preventing cancer. Dendritic Cells (DC) are considered effective therapeutic candidates for treatment of cancer; however, many fundamental questions about how to best prepare and administer them are unanswered. At present, it is not possible to quantitate the delivery and migration of DC-based vaccines without biopsy or excision of the lymphoid tissue. Our study shows that DC can be labeled with Feridex and imaged using a 1.5T GE MR scanner. Further, the lymph node volume, signal void volume and fractional signal loss measured correlate with the delivery and migration of in vitro-generated DC.
1674.
Magnetically Labeled Neural Stem Cells Exhibit Differential
Gene Expression of Zinc (Finger) Binding Dorota Anna Kedziorek1, Piotr Walczak1, Chunfa C. Jie1, Naser Muja1, Jesus Ruiz-Cabello2, Assaf A. Gilad1, Jeff W.M. Bulte1 1Johns Hopkins University, Baltimore, USA; 2Universidad Complutense de Madrid, Madrid, Spain We assessed the time-dependent changes in gene expression following labeling of neural stem cells with Feridex and poly-L-lysine.A pronounced difference in the time course of the expression of zinc (finger) binding proteins was found. These specific regulatory proteins are responsible for maintaining cellular homeostasis indirectly through DNA and RNA binding interference. Our results are relevant for further clinical implementation of MRI cell tracking.
1675.
Developing Cellular MPI: Initial Experience Jeff W. M. Bulte1, 2, Bernhard Gleich3, Jurgen Weizenecker3, Segun Bernard1, Piotr Walczak1, Dennis E. Markov4, Hans C. J. Aerts5, Jorgen Borgert3, Hans Boeve4 1The Johns Hopkins University School of Medicine, Baltimore, USA; 2The Johns Hopkins University Whiting School of Engineering, Baltimore, USA; 3Philips Research Europe, Hamburg, Germany; 4Philips Research Europe, Eindhoven, Netherlands; 5Philips Medical Systems, Best, Netherlands Using clinical formulations of SPIO particles used as MRI contrast agents, we investigated the potential of magnetic particle imaging (MPI) for quantitative stem cell tracking. We observed a linear correlation between MPI signal and iron content, representative for the total cell number over a wide range of concentrations, independent of the particle state as free or intracellular entity. Unlike in MRI, Resovist had a 4–fold higher efficacy per unit Fe than Feridex for two different stem cell types tested. MPI has potential for non-invasive quantitative cell tracking and deserves further exploration with or without the use of MRI in parallel.
1676.
Non-Invasive Assessment of Chronic Cardiac Allograft
Rejection in a Rat Model with in Vivo MR
Qing Ye1, Yijen L. Wu1, Lesley M. Foley1, T. Kevin Hitchens1, Haosen Zhang1, Haval Shirwan2, Chien Ho1 1Carnegie Mellon University, Pittsburgh, Pennsylvania, USA; 2University of Lousville, Louisville, Kentucky, USA Chronic cardiac allograft rejection (CCAR) constitutes a major obstacle to the long-term success of heart transplantation. Studies indicate the involvement of immune cells in the development of CCAR. In this study, we used a chronic rejection working heart rat model. Immune cells (mainly macrophages) were labeled in situ with micrometer-sized paramagnetic iron oxide (MPIO) particles, and their accumulation in the rejecting allograft was monitored non-invasively with MRI. Our data showed that the distribution of macrophage can be non-invasively assessed using in vivo MRI, and the MRI results were well correlated with ex vivo MR microscopy and the pathology changes. 1677. Does Prussian Blue Staining Correlate with Dextran Staining of Ferumoxides Labeled Cells? Aneeka Chaudhry1, Joseph A. Frank1, Edyta Pawelczyk1 1National Institutes of Health, Bethesda, USA Dextran coated iron oxide nanoparticles (SPION) are used to magnetically label cells for cellular MRI. Prussian blue (PB) and/or dextran staining has been used to identify SPION labeled cells in tissue on histology. However, no comparisons have been made on the detection efficiency of SPION in cells. Ferumoxides (FE) labeled cells were followed over 14 days for dextran positive cells compared to PB + cells and iron content. Results demonstrated that there was greater percentage of PB +cells compared to dextran positive cells at 7 and 14 days post FE labeling. 1678. MRI of Neural Stem Cell Migration in the Adult Mouse Brain Jeffrey You Ling Shyu1, Brian J. Nieman1, Daniel H. Turnbull1 1New York University School of Medicine, New York, New York, USA Iron-oxide particles, a favored agent for cellular MRI, can be used to label endogenous neural stem cells (NSCs) in situ in rodents. Previous studies required the injection of large numbers of micron-sized particles of iron oxide (MPIOs), creating a large hypointensity artifact that restricts the study of NSC response to injury. In this work, MPIOs were injected directly into the subventricular zone of the adult mouse brain; a much smaller volume of iron-oxide particles was needed to observe migration along the rostral migratory stream. This method may greatly aid the study of endogenous NSC responses to brain injury and disease.
1679.
Quantitative Cellular Tracking MRI of Human Mesenchymal Stem
Cells Home to the Metastatic Breast Ho-Taek Song1, 2, Elain K. Jordan1, Bobbi K. Lewis1, Jae-Hyun Lee, Young-wook Jun, Jinwoo Choen, Joseph A. Frank1 1Clinical Center, National Institite of Health, Bethesda, USA; 2College of Medicine, Yonsei University, Seoul, Republic of Korea In this study we demonstrate that the MR T2* map histogram can be used to monitor the homing of magnetically labeled human mesenchymal stem cells (hMSCs) to sites of metastatic breast cancer in rat brains at 3T. hMSCs were labeled with ferumoxides-protamine sulfate, micron sized iron oxide particles and engineered manganese iron oxide nanoparticles. Quantitative T2* maps revealed no difference among the labeled MSC to home to sites of brain metastatsis as confirmed by immunohistochemical stainning for MSCs< 1680. In Vivo Serial Monitoring of Implanted Chemically Biotinylated Cells by MRI Po-Wah So1, Tammy Kalber1, Jimmy D. Bell1 1Hammersmith Hospital, Imperial College London, London, UK In vivo monitoring of specific cell populations is essential for the development of cell based therapies. Recently, chemical biotinylation of cells and their subsequent ‘tagging’ by antibodies to biotin conjugated to SPIOs has been proposed as a ready and rapid method of labeling cells for MRI tracking in vivo. In this study we have demonstrated the feasibility of cells labeled in this manner for serial MRI monitoring over a period of 3 days after implantation. 1681. Direct Incubation of PBMCs with Ferumoxides Labels Monocytes and B Cells John S. Pratt1, Anna Moore1, Alma Gregory Sorensen1 1Massachusetts General Hospital, Charlestown, Massachusetts, USA The purpose of this study was determine the feasibility of labeling peripheral blood mononuclear cells (PBMCs) by direct incubation with ferumoxides, and to evaluate the types of leukocytes labeled by this method for the eventual development of MRI-based clinical monitoring of leukocytes in patients. Human peripheral blood mononuclear cells are readily labeled by direct incubation with clinically approved ferumoxides. Labeling was sufficient for detection of cells by 4.7T MRI. Detailed analysis of the cell types labeled revealed B cells were primary lymphocytes labeled. Application of this methodology may contribute new knowledge to the poorly understood role of B cells in leukemia and other diseases, and further clinical translation of MRI-based cellular imaging. 1682. Quantification of Magnetically Labeled Cells with an SPIO Labeled Tumor Model in Rats wei liu1, E Kay Jordan2, Hannes Dahnke3, Joseph A. Frank2 1Philips Research North America, Briarcliff Manor, New York, USA; 2National Institutes of Health, Bethesda, Maryland, USA; 3Philips Research Europe, Hamburg, Germany Quantifying the number of labeled stem cells in target tissues in experimental models is of great importance to optimize dose and timing of cellular therapy. In this study, we investigated the quantitative relationship between labeled cells and tumor R2* relaxation rate at different stages of the tumor development with an SPIO labeled tumor model in rats. The individual linear relationship from each tumor was consistent with the gross linear correlation between tumor R2* and number of SPIO labeled cells/mm3. The detection limit for SPIO labeled tumor cells was estimated at about 316 cells/mm3 in this model. 1683. Tracking Magnetotactic Bacteria Using an MRI System for Future In-Vivo Therapeutic Targeting Ouajdi Felfoul1, Sylvain Martel1 1École Polytechnique de Montréal (EPM), Montreal, Canada Magnetotactic bacteria (MTB) are flagellated microorganisms that can be controlled using an external magnetic field. MTB synthesize intracellular magnetite nanoparticles, called magnetosomes, assembled in a chain acting as a compass to influence through magnetotaxis the swimming direction of the bacteria. The magnetosomes that disturb the local magnetic field could allow MTB tracking deep in the human body using an MRI system. MR-images of MTB are studied in order to validate the possibility of tracking MTB-based drug delivery operations using a clinical MR scanner. MTB affect more the T2 than the T1-relaxation rate and can be though as a negative contrast agent.
1684.
Superparamagnetic Iron Oxide and Protamine Sulfate Do Not
Affect the Viability and Multi- Jixian Deng1, Bo Xiang1, Tonghua Yang1, Lei Wang1, Gang Li1, Jian Wang1, Tarek Kashour1, Alan Menkis2, Boguslaw Tomanek1, Roxanne Deslauriers1, Rakesh Arora2, Ganghong Tian1 1National Research Council Canada, Winnipeg, Canada; 2St. Boniface General Hospital, Winnipeg, Canada This study was designed to assess the effects of superparamagnetic iron oxide (SPIO) and protamine sulfate (PS) on the viability and multi-transdifferentiation capacity of adult stem cells. To achieve the objective, adipose-derived stem cells (ADSCs) were cultured in a medium containing different concentrations of SPIO and PS. Then, the ADSCs were subjected to three types of transdifferentiation inductions (adipogenic, osteogenic, and myogenic induction). We found that SPIO and PS did not affect the viability and multipotent transdifferentiation capacity of the ADSCs. We therefore conclude that SPIO is a safe MR contrast reagent for in vivo tracking of the adult stem cells and PS is a safe SPIO-loading reagent.
1685.
USPIO-Enhanced MR Imaging of Macrophage Recruitment Into the
Rat Brain Following MCP-1 Andrew Peter Brown1, 2, Irene Tracey1, Daniel C. Anthony1, Sebastien Serres1, Yanyan Jiang1, Nicola R. Sibson1 1University of Oxford, Oxford, UK; 2Glaxosmithkline, Hammersmith Hospital, UK USPIOs can be visualised as a distinct reduction in signal intensity on T2* weighted MR images and can provide a method of tracking macrophage recruitment to sites of neuroinflammation. There are issues regarding the distinction between iron-laden recruited macrophages and the entry of free iron across a permeable blood brain barrier (BBB) in disease cases. In the current study we have shown that USPIO-enhanced MRI can provide a technique to determine the relationship between the number of recruited macrophages to an area of CNS tissue damage and signal change following microinjection of MCP-1 into the rodent striatum. 1686. Biocompatible Chitosan Nanoparticles Encapsulating Iron Oxide with a MRI High-Relaxivity Hsiao-Yun Kuo1, 2, Jiun-Jie Wang, Rei-Tsan Tsai, I-Ru Chang1, KoonKuan Ng, Chao-Lin Liu3, Chia-Rui Shen1 1Chang Gung University, Kweishan, Taiwan; 2 Chang Gung Memorial Hospital, Kweishan, Taiwan; 3Min Chi University of Technology, Taishan, Taiwan Chitosan encapsulating superparamagnetic iron oxide maintains the desirable high relaxivity and less cytotoxicity 1687. T1 and T2 Quantification of Free USPIO and USPIO-Labeled Macrophages at 4.7T and 7T Jean-Christophe Brisset1, Virginie Desestret1, Sebastien Marcellino2, Emilie Devillard1, Florence Lagarde2, Norbert Nighoghossian1, Yves Berthezene1, Marlene Wiart1 1Université de lyon, CREATIS-LRMN, UMR CNRS 5220, Inserm U630, Lyon, France; 2Université de lyon, Laboratoire des Sciences Analytiques, UMR 5180 CNRS-UCBL, Lyon, France USPIO-enhanced MRI is a promising tool for the experimental study of inflammation. The aim of this study was to quantify the effects of compartmentalization of USPIO on T1 and T2 relaxivities at 4.7T and 7T, using phantoms of free USPIO versus USPIO-labeled macrophages. Compartmentalization of USPIO resulted in a decrease of T1 and T2 effects of the iron particles. T1 and T2 relaxivities of USPIO-labeled cells were lower at 7T than at 4.7T. Taken together, these results suggest that working at 4.7T may be a good compromise between sufficient spatial resolution achievable to image small animals and USPIO relaxivity properties
1688.
19F MRI Detection of Acute Cardiac Allograft
Rejection with in Situ Perfluorocarbon Labeling of Immune T Kevin Hitchens1, 2, Qing Ye1, Haosen Zhang1, Lesley M. Foley1, Jelena M. Janjic1, Eric T. Ahrens1, Chien Ho1 1Carnegie Mellon University, Pittsburgh, Pennsylvania, USA The gold standard for diagnosing and staging rejection after organ transplantation is biopsy, which is not only invasive but also prone to sampling errors. Cellular imaging of the immune cell accumulation in the rejecting graft may provide a better approach. We present a method to detect rejection in a rodent cardiac transplantation model via in situ labeling of immune cells with a perfluorocarbon nanoemulsion. The accumulation of 19F labeled cells can be located and quantified, unambiguously, by 19F MRI, then the labeled cells can then be placed in their correct anatomical context with a conventional 1H image. Mayumi Yamada1, Paul T. Gurney1, William Stein1, Pratima Kundu1, Alan Smith2, Robert C. Robbins1, Phillip C. Yang1 1Stanford University, Stanford, California , USA; 2Cognate Bio Services, Inc., Sunnyvale, California , USA Manganese chloride can demonstrate the potential to detect cellular viability and biological property and it can be used for non invasive biological evaluation of transplanted human stromal stem cells within injured myocardium.
1690.
In Vivo Mitochondrial Labeling Using Mito-Carboxy Proxyl
(Mito-CP) Enhanced Magnetic Resonance Douglas Edward Prah1, Eric S. Paulson1, Melissa Lynn Wagner-Schuman1, J Zielonka1, Marcos Lopez1, Miguel J. Hardy1, Joy Joseph1, B Kalyanaraman1, Kathleen M. Schmainda1 1Medical College of Wisconsin, Milwaukee, Wisconsin, USA Due to their single unpaired electron, nitroxides could potentially be used as MR contrast agents. Recent evidence suggests that the nitroxide, Mito-Carboxy Proxyl (Mito-CP), preferentially targets mitochondria. MR contrast agents specific to mitochondria would provide another avenue to study mitochondrial diseases or pathologies that exhibit abnormal metabolic and/or mitotic activity (e.g., tumors). We demonstrate here, in an in vivo study, the uptake of Mito-CP into rat breast tumor. 1691. SPIO Labeled Cells: Magnetic Resonance Source Quantification by Inverting the Dipole Field Ludovic de Rochefort1, Maire Poirier-Quinot2, Bryan Kressler1, Claire Wilhelm3, Florence Gazeau3, Jean-Christophe Ginefri2, Luc Darrasse2, Yi Wang1 1Weill Medical College of Cornell University, New York, New York, USA; 2Unité de Recherche en Résonance Magnétique Médicale, U2R2M, Université Paris-Sud, CNRS UMR8081, Le Kremlin-Bicêtre, France; 3Laboratoire Matières et Systèmes complexes, Université Paris 7, CNRS UMR7057, Paris, France The local dipole field resulting from the strong magnetic moment surrounding superparamegnetic iron oxides creates signal voids (T2* effects) very well observed with gradient-echo imaging. Here, we propose the MR-SQUID technique which combines field mapping and data processing to allow quantification of the magnetic source causing the field shifts. This quantitative 'positive contrast' technique is applied on SPIO-loaded cells. Measured mass of iron per cell agrees with calibrated quantities down to the pg level. 1692. In Vitro / in Vivo MRI Visualization of PFC / VSOP Double-Labeled Mesenchymal Stem Cells Thomas Kampf1, Thomas Christian Basse-Luesebrink2, Meike Weber1, Andrea Heymer1, Regina Ebert1, Christian Stoetzel1, Uwe Gbureck1, Ulrich Noeth1, Peter Michael Jakob1, Wolfgang Rudolf Bauer3, Daniel Haddad2 1University of Wuerzburg, Wuerzburg, Germany; 2Research Center Magnetic Resonance Bavaria, Wuerzburg, Germany; 3Universtity of Wuerzburg, Wuerzburg, Germany In this work we introduce an new double lable scheme for non ambigouse monitoring of iron oxide labeled cells. The superparamagnetic iron oxide particles can be used as MRI markers in high resolution T2* weighted 1H-MR images. To distinguish the iron oxide labeled cells from other hypointensities in the MR image cells are additionally labeled with PF15C nanoparticles. In lower resolved 19F-MR images the position of the double labeled cells can be verified. 1693. Rat Glioma C6 Cells Labeled with a Fluorinated Gd-GlyMe-DOTA-Complex Markus Plaumann1, Jan Willmann1, Ekkehard Küstermann1, Uwe Schüssler1, Dieter Leibfritz1 1University of Bremen, Bremen, Germany Fluorine is of interest in medical chemistry and diagnostics, because of its very low natural abundance in living organisms and high MR sensitivity. Fluorinated contrast agents allow to record 19F-images without any background signals. Six different 3,5-Bis(trifluoromethyl)benzyl derivatives were selected as model compounds to compare Gd-DTPA- and Gd-GlyMe-DOTA-complexes with respect to their T1-times. Furthermore the lipophilic properties of these Gd-complexes and also of different M3+ ions complexes of 3,5-Bis(trifluoromethyl)phenylhydrazine-DTPA, -GlyMeDOTA and -GlyMeAsp-DOTA were determined. MRI experiments and ICP-MS measurements proof the possibility to label rat glioma C6 cells with 3,5-Bis(trifluoromethyl)phenylhydrazine-Gd-GlyMeDOTA. Yidong Yang1, 2, Ben Waghorn1, 2, Yuhui Yang1, Brianna Klein1, Nathan Yanasak1, William David Hill1, Tom C-C. Hu1, 2 1Medical College of Georgia, Augusta, Georgia, USA; 2Georgia Institute of Technology, Atlanta, Georgia, USA MRI could serve as a powerful tool to clarify the role of mesenchymal stem cells (MSCs) in myocardial injury and the subsequent recovery process. In this study, MSCs were labeled in vitro with micrometer-sized particles of iron oxide (MPIOs) with an efficiency of 98% plus and introduced to mice via transplantation. Region of attenuated signal and dark spots in MRI were observed in the myocardial infarction site. Further work to quantify the trafficking of MSCs is needed. |
Electron Spin Resonance/Oximetry Hall D Wednesday 13:30-15:30 Sankaran Subramanian1, Shingo Matsumoto, Sonny Batra, Nallathamby Devasahayam, Fuminori Hyodo, Murali C. Krishna 1National Institutes of Health, Bethesda, USA Time-domain radiofrequency EPR imaging has been developed to study the distribution of free radicals in vivo and to quantitatively assess the spatial distribution of oxygen non-invasively. Quantitative oxymetry is important in the treatment of tumor by radiation. Oxygen estimates are based on the spatially resolved spectral characteristics (local line width) of the free radical. Two modalities, namely, (a) the 90-t 80 spin echo and filtered back projection that can provide T2-weighted images, and (b) the single point imaging (with pure phase-encoding and Fourier reconstruction) that can provide a series of T2*-weighted images, are examined for their relative merits.
1726.
3D Superresolution EPR Imaging of Nitroxyl Radicals in Mice
Yusuke Ikebata1, Hideo Sato-Akaba1, Hirotada Fujii2, Hiroshi Hirata1 1Yamagata University, Yonezawa, Japan; 2Sapporo Medical University, Sapporo, Japan We report the superresolution technique of three-dimensional (3D) continuous-wave (CW) EPR imaging of live mice. This post-processing technique could improve the spatial resolution of 3D EPR images. We used anesthetized ICR mice, and administered 0.2-0.25 ml solution of 200 mM 3-hydroxymethyl-2,2,5,5-tetramethylpyrrolidine-1-oxyl (hydroxymethyl-PROXYL) spin probes intraperitoneally. We show a superresolution EPR image that was generated from a blurred image of nitroxyl probes in mouse head.
1727.
EPR Imaging of Short-Lifetime Nitroxyl Radicals in Mouse Head
Hideo Sato-Akaba1, Hirotada Fujii2, Hiroshi Hirata1 1Yamagata University, Yonezawa, Japan; 2Sapporo Medical University, Sapporo, Japan Three-dimensional (3D) EPR imaging of 4-Hydroxy-2,2,6,6-tetramethylpiperidine-d17-1-15N-1-oxyl in mouse head was reported. To enable EPR imaging of nitroxyl spin probes that have a half-life of a few minutes possible, the time of field scanning was decreased to 0.5 s in animal experiments. The total acquisition time of 46 projections for a 3D image was achieved at 30 s. Applicability of 3D continuous-wave (CW) EPR imaging to short-lifetime nitroxyl spin probes in animals was demonstrated. 1728. Nanoencapsulation of Perfluorinated Trityl Radicals and Evaluation as Sensors for EPR Oximetry Nicolas Charlier1, Benoît Driesschaert1, Jacqueline Marchand1, Véronique Préat1, Bernard Gallez1 1Université catholique de Louvain, Brussels, Belgium Trityls radicals are interesting probes for EPR spectroscopy/imaging as they possess a narrow EPR Line Width (LW) (high signal-to-noise ratio). However, they have a low sensitivity to changes in oxygenation. It is possible to increase the sensitivity by using lipophilic solvent, such as perfluocarbon solvents as they have high oxygen solubility and biocompatibility. We developed a nanoemulsion containing a newly synthesized perfluorinated trityl. X-band measurements of LW were carried out at 21% and 0% and showed LW variation of 2.3G compared to 0.1G for hydrophilic trityl. In vivo results showed that emulsion is higly responsive to subtle pO2 variations. 1729. Molecular Electron Paramagnetic Resonance Imaging of Melanin in Melanomas: A Proof-Of-Concept Nicolas s. Charlier1, Emilia Vanea1, Julie De Wever1, Mustapha Dinguizli1, Olivier Feron1, Jean-François Baurain1, Bernard Gallez1 1Université catholique de Louvain, Brussels, Belgium Taking into account the abundance of the naturally occurring free radicals (melanin pigments) in proliferating melanocytes and their localization pattern, we hypothesized that EPR Imaging (EPRI) could be a unique tool to map melanomas with high sensitivity and high resolution. The potential of EPR to image melanoma samples was demonstrated in vitro in animal and human samples. Using EPR systems operating at low frequency, we were also able to record in vivo EPR spectra and images from the melanin present in a subcutaneous melanoma implanted in a mouse.
1730.
Relationship Between Oxygenation Status and Tumor
Metabolites. Noninvasive Evidence for Aerobic Shingo Matsumoto1, Fuminori Hyodo1, Jeeva Munasinghe2, Sonny Batra1, Sankaran Suburamanian1, Nallathamby Devasahayam1, James B. Mitchell1, Murali C. Krishna1 1National Cancer Institute, NIH, Bethesda, Maryland, USA; 2National Institute of Neurological Disorder and Stroke, NIH, Bethesda, Maryland, USA Tumor hypoxia is an important prognostic factor in radiotherapy, and non-invasive methods to assess tumor hypoxia such as Electron Paramagnetic Resonance Imaging (EPRI) can provide useful a priori information to guide therapy. To investigate the correlation between oxygenation status and tumor metabolites detectable by MR spectroscopy (MRS), we developed a pulsed EPRI 3D oxygen imaging technique that operates at 300 MHz enables to be sequentially combined with 7T MRI/MRS using a common RF coil without disturbing the object. This combined system clearly shown the oxygenation dependent MRS metabolic change in tumor, especially aerobic glycolysis (lactate production) known as Warburg effect. 1731. Surface Coil for EPR Irradiation to Reduce SAR in Fixed-Field PEDRI Keerthi Shet1, 2, Sergey Petryakov3, Jay L. Zweier3 1Ohio State University, Columbus, Ohio, USA; 2Davis Heart and Lung Research Institute, Columbus , Ohio, USA; 3Davis Heart and Lung Research Institute, Columbus, Ohio, USA Proton Electron Double Resonance Imaging (PEDRI) is a combination of proton nuclear magnetic resonance (NMR) imaging and electron paramagnetic resonance (EPR). The technique involves irradiating a paramagnetic sample at its electron spin resonance and this is followed by NMR signal acquisition to produce an Overhauser enhanced NMR signal. The EPR frequency at about 660 times the NMR frequency and may cause sample overheating if irradiated for longer periods of time. Proposed here is a study that uses surface coils for EPR irradiation instead of whole body EPR coils when the organ of interest is closer to the surface of the sample. 1732. Comparison of Methods to Measure the PO2 Mustapha Dinguizli1, Nelson Beghein1, Bernard Gallez1 1Université Catholique de Louvain, Brussels, Belgium To clarify the source of discrepancy between pO2 vreported for SC tissue, we used sequential monitoring of pO2 using EPR oximetry and OxyLite probes. Initial pO2 measurements by EPR show low values in subcutaneous area (10 mm Hg). After insertion of the OxyLite probes, the pO2 recorded using this system indicate significant larger values (40 mm Hg). Immediately after the OxyLite measurements, the second EPR measurements indicate systematically pO2 increased compared to the first measurement. While comparison of oximetry methods is generally valuable in deep tissues, the use of invasive methods near the skin potentially disturb the local pO2 1733. Spin Density Distribution in Foodstuff After Heat Treatment or Irradiation Philippe P. Levêque1, Quentin Godechal1, Bernard Gallez1 1Université catholique de Louvain, Brussels, Belgium Irradiated food can be detected by EPR spectroscopy according to current European Standards. When spatial distribution of spin densities is needed, EPR imaging can be used to locate radicals in biomaterial.We applied on going developments of this technique to study spin density distribution in irradiated frog-leg and roasted coffee bean, as an illustration of the instrumentation capabilities.In irradiated frog leg, spin density is concentrated in bone tissue, whereas in roasted coffee bean it follows a gradient from the center to the edges.EPR imaging offers unique capabilities to monitor the fate of free radicals in biological sample. |
Hall D Wednesday 13:30-15:30 1787. Change in Water Diffusion Properties with Altered Muscle Architecture Anneriet M. Heemskerk1, 2, Tuhin K. Sinha1, Kevin J. Wilson1, Bruce M. Damon1 1Vanderbilt University, Nashville, Tennessee, USA For skeletal muscle l2 and l3 are perpendicular to the fiber direction and previous studies suggest that they have different structural origins. The goal of this study was to determine whether and how the DTI indices change upon muscle lengthening. Ten subjects were studied with their foot in two different positions and twice in the same position. The latter showed reproducible DTI values. However, the second eigenvalue decreased with increased foot rotation, whereas the other indices remain constant. This confirms that the λ 2 and l3 have different structural origins. Els Fieremans1, Chi Chen1, Yves De Deene2, Eric Achten3, Ignace Lemahieu1 1Ghent University, Ghent, Belgium; 2Ghent University Hospital, Ghent, Belgium; 3Ghent University Hospital, Belgium Anisotropic fiber phantoms have been proposed for the validation of DTI on clinical MR-scanners and to test fiber tracking algorithms, particularly in the case of fiber crossings. Several fiber materials have been used: rayon, Dyneema®, hemp, linen, acrylic fibers, rayon, acrylic fiber, … Choosing the appropriate fiber material requires insight in the factors influencing the diffusion and the signal-to-noise ratio. This study enumerates the different parameters (fiber diameter, fiber density, surface relaxivity and susceptibility differences) and their impact in the resulting diffusion properties and T2. Burkhard Mädler1, 2, Shannon H. Kolind2, Alex L. MacKay2 1Philips Medical Systems, Vancouver, Canada; 2University of British Columbia, Vancouver, Canada We compare T2-relaxation data and diffusion tensor data from normal human brain. The relationships between myelin water fraction (MWF) and various diffusion tensor measures in white and grey matter structures in the brain were examined in 6 normal subjects at 3.0T. We found some degree of linear correlation between these measurements, but by using ROI-based analysis we also observed several structures which seemed to deviate significantly from a linear relationship. From all investigated relationships between various diffusion tensor measures and myelin-water content, fractional anisotropy (FA) and perpendicular diffusivity (ADCperp) yielded the highest correlation coefficients with MWF. However, diffusion anisotropy was also significantly influenced by factors other than myelin water content. The less operator dependent voxel-based analysis between myelin-water and diffusional anisotropy measures is proposed as an innovative alternative to ROI-based analysis. 1790. Investigation of Bi-Exponential Diffusion in Treated Brain Tumors Benjamin Alan Hoff1, Denis Le Bihan2, Alnawaz Rehemtulla1, Thomas L. Chenevert1, Brian D. Ross1, Craig J. Galban1 1University of Michigan, Ann Arbor, Michigan, USA; 2NeuroSpin, CEA, Saclay, France The purpose of this study was to investigate bi-exponential diffusion of water in a brain tumor following therapeutic treatment. Diffusion-weighted images at 17 different b-values were acquired in a 9L glioma rat model. Both fast and slow diffusion coefficients were found to peak significantly by day 6 of treatment, whereas the fast diffusion fraction in the bi-exponential fit peaked at day 9. We found that the fast diffusion coefficient was the most sensitive to therapy, indicating that the therapeutic response not only involves a shift in the relative fractional volume, but also the actual diffusion rates. Markus Nilsson1, Jimmt Lätt1, Emil Nordh1, Ronnie Wirestam1, Freddy Ståhlberg1, Sara Brockstedt1 1Lund University, Lund, Sweden The signal-vs-b curve was measured perpendicular to the corticospinal tract in vivo to determine whether effects of restricted diffusion could be observed when varying the diffusion time between 64 and 256 ms and sampling b-values up to 86 000 s/mm2. Restricted diffusion with exchange was simulated for equally sized cylinders. The measurements showed no effects of restriction perpendicular to the tract. Comparisons of the signal-vs-b curves from the measurements with the simulations further indicated that no effects of restricted diffusion could be observed perpendicular to the tract. 1792. Probing Intracellular Compartments in Normal Brain and Brain Tumor Using Short Diffusion Times Douglas Edward Prah1, Eric Scott Paulson1, Melissa L. Wagner-Schuman1, Kathleen M. Schmainda1 1Medical College of Wisconsin, Milwaukee, Wisconsin, USA Standard imaging techniques inadequately detect tumor cell invasion. Often the failure of surgical resection and radiation therapy, the primary treatments for gliomas, are attributed to tumor invasion. Treatment outcome may improve if more localized therapies are prescribed to areas of invasion. Results from this study demonstrate that diffusion data acquired at short diffusion times, may be more sensitive to the tissue/tumor intracellular environment, and therefore may be able to detect invading tumor cells. 1793. Diffusional Restrictivity: Looking at the Slow Water Mariana Lazar1, Jens H. Jensen1, Joseph A. Helpern1 1New York University School of Medicine, New York, USA In this abstract we introduce a new metric, the diffusional restrictivity (RD), which primarily depicts contributions from shorter diffusion paths. As a result, RD is more sensitive to restricted water compartments and smaller scale structures and less prone to partial volume averaging effects from the fast water compartments (such as CSF contamination). Thus, it should provide a more robust marker of the tissue microstructure. RD can be decomposed into Gaussian and non-Gaussian components. Restrictivity maps for human brain diffusion data show significant white matter/gray matter contrast with higher restrictivity being observed in white matter. 1794. A Realistic DTI Simulation Environment Gregory T. Balls1, Lawrence R. Frank1 1University of California, San Diego, La Jolla, California , USA Simulations of diffusion in neural tissues have traditionally been limited to analytical solutions or to grid-based solvers unable to simulate realistic complex neural tissues. We have built a simulation environment that combines a Monte Carlo Brownian dynamics simulator with a MR signal generator. The Monte Carlo simulator is capable of simulating diffusion in arbitrarily complex polygonal geometries and modeling chemical reactions and relaxation. The signal generator is flexible enough to handle a variety of pulse sequences. We demonstrate results for aligned and crossing straight fibers, varying fiber packing density and permeability. 1795. Multiexponential Diffusion Measurements of Varying Densities of Astrocytoma Cells Pamela R. Jackson1, Roland G. Henry2, Tracy R. McKnight2 1University of California, Berkeley and San Francisco, San Francisco, California , USA; 2University of California, San Francisco, San Francisco, California , USA To better understand how cell density affects diffusion, we suspended different densities of cells (0%, 25%, 50%, 75%, and 100%) in agarose and measured diffusion using 32 b values. Monoexponential, biexponential, and non negative least squares algorithms were used to fit the data, and the Spearman Rank test was used to test correlation . The monoexponential decrease in diffusion with increasing cell density was dominated the fast diffusion component. The faster diffusion component appeared to be associated with agar presence and the slower component may have been associated with the cells. Jimmy Lätt1, Markus Nilsson1, Danielle van Westen1, Ronnie Wirestam1, Fredrik Örtendahl1, Freddy Ståhlberg1, Stig Holtås1, Sara Brockstedt1 1Clinical Sciences Lund University, Lund, Sweden Diffusion weighted imaging (DWI) enables visualization of ischemic stroke in the acute phase. However, the underlying reasons for the altered signal characteristics of the ischemic lesion are still under debate. In this preliminary study, measurements at different diffusion times were performed, potentially providing information about the tissue microstructure, with respect to restricted diffusion and exchange mechanisms. 1797. Diffusion-Weighted Signal in White Matter: What is Behind the B-Factor Dependence? Valerij G. Kiselev1, Kamil A. Il'yasov2 1University Hospital Freiburg, Freiburg, Germany; 2Kazan State University, Kazan, Russian Federation Diffusion weighted signal can be characterised via an infinite set of coefficients of the expansion in b-factor. In practice, the number of detectable coefficients is restricted by the noise. This description is examined in the corpus callosum. It is shown that only two coefficient can be found using b-factors up to 2500 mm^2/s. Those are the diffusion coefficient and the kurtosis excess. This means that the contributions of the intra- and extra-axonal compartments cannot be recognised in the total signal. The discussion focuses on constraints this result imposes on the currently unknown diffusion parameters in extra-axonal space. |
Crossing Fibers & Non-Tensor Approaches Hall D Wednesday 13:30-15:30 1856. A Tensor Based Approach to Crossing Fibers Using the Wild Bootstrap Ai Wern Chung1, Thomas Richard Barrick2, Chris A. Clark1 1UCL, Institute of Child Health, London, UK; 2Saint George's, University of London, London, UK Multiple-fiber reconstruction algorithms are often mathematically and computationally intensive. We present a simple method for detecting crossing structures by using k-means clustering to group principal eigenvector distributions provided by the Wild bootstrap diffusion tensor model into two populations. Using this Wild-tensor approach, we were able to produce results in the frontal pericallosal region comparable to the Q-Ball method on a 60-directional dataset at b = 0 - 3000s mm-2. Our method shows potential by providing plausible trajectories. Furthermore, the availability of bootstrapped data enables calculation of a variety of indices within and between cluster groups. 1857. Random Effects Modelling of Crossing Fibre Voxels in Diffusion MRI Martin D. King1, David G. Gadian1, Christopher A. Clark1 1UCL Institute of Child Health, London, UK Synopsis Random effect (RE) models were applied to the crossing-fibre problem that arises in the analysis of multiple-directions diffusion-weighted MR data. Among these models is the Besag-York-Mollie model, which includes both exchangeable and spatial RE terms. The analyses were performed using a Markov chain Monte Carlo algorithm (Gibbs sampling). Two crossing fibre brain regions were used to show that RE modelling provides a useful alternative to current methods documented in the MR tractography literature.
1858.
Multi-Model Diffusion Image Analysis: A Correlative Study
Between Diffusion Spectrum Imaging Wen-Yang Chiang1, Li-Wei Kuo2, Van Jay Wedeen3, Wen-Yih Isaac Tseng1 1National Taiwan University College of Medicine, Taipei, Taiwan; 2Department of Electrical Engineering, National Taiwan University, Taipei, Taiwan; 3MGH Martinos Center for Biomedical Imaging, Charlestown, Massachusetts, USA Since there are dramatic differences between each diffusion-based imaging modality, it is challenging to compare different diffusion indices from different imaging modalities. In this study, we used DSI data to analyze structural lateralization of arcuate fasciculus and compared the result which used DTI reconstruction method. The experiment results showed that both diffusion indices and lateralization index are highly correlated between DTI and DSI which means that the structural analysis, even with different diffusion imaging modalities, is comparable. This fundamental study, which might be the first one that compares diffusion indices and lateralization between different diffusion imaging modalities, may facilitate inter-sites collaborations of diffusion MR researches. 1859. Improved Detection of Non-Gaussian Diffusion with Nonlinear Fitting at High B-Values Ryan Fobel1, Greg Stanisz2 1University of Toronto, Toronto, Canada; 2Sunnybrook HSC, Toronto, Canada Higher-order diffusion models, including those based on the Spherical Harmonics and Generalized Diffusion Tensors, are typically fit using linear least-squares methods. This leads to limited performance at high b-values. A nonlinear fitting algorithm which compensates for magnitude bias results in more accurate fitting at b-values of up to 3000 s/mm2. The performance of model selection (ie. classification of voxels as having either isotropic, non-isotropic Gaussian or non-Gaussian diffusion) is significantly improved with this new fitting method. Experimental results show that more than twice as many voxels exhibiting non-Gaussian diffusion are detected at b=2000 s/mm2 relative to b=1000 s/mm2. Flavio Dell'Acqua1, 2, Paola Scifo1, Giovanna Rizzo, 12, Giuseppe Scotti1, Ferruccio Fazio1, 2 1San Raffaele Scientific Institute, Milan, Italy; 2IBFM-CNR University of Milano-Bicocca, Milan, Italy The presence of partial volume effect between white matter and isotropic tissues can increase the number of instabilities and spurious fiber orientations in spherical deconvolution methods. In this work, we proposed a modified damped version of the Richardson-Lucy algorithm to improve results also in these regions. Simulation and in-vivo results are shown. 1861. A Regularised Two-Tensor Model Fit to Low Angular Resolution Diffusion Images Stamatios N. Sotiropoulos1, Christopher R. Tench1, Li Bai1, Paul S. Morgan1, Dorothee P. Auer1, Cris S. Constantinescu1 1University of Nottingham, Nottingham, UK Diffusion Tensor Imaging (DTI) is prone to partial volume artefacts. Given that routine clinical use of high angular resolution diffusion imaging is still tentative, a regularised two-tensor model to resolve fibre crossings from conventional DTI datasets is presented. The model utilizes the planar diffusion profile depicted by DTI in regions of crossing fibres. Relaxation labelling is then applied to regularise and preserve slow changes in the two resolved orientations across a neighbourhood. Orientational, anisotropy, and diffusivity information can be resolved in regions of two fibre crossings using full brain coverage scans acquired in less than 6 minutes. Cristina Granziera1, Thomas Benner2, Ashok Kumar2, Dave Tuch2, Aneesh Singhal3, Walter Koroshetz3, Gregory Alma Sorensen2 1CHUV, Lausanne, Switzerland; 2Martinos Center-MGH-Harvard medical school, Charlestown, Massachusetts, USA; 3MGH, Boston, Massachusetts, USA Diffusion tensor imaging (DTI) has been advocated as a valuable non-invasive tool to monitor post-stroke recovery. However, its inaccuracy in regions of fibers crossing/kissing could limit its application in studying the complex white matter plasticity characterizing functional recovery after stroke. We compared DTI to q-ball imaging in the study of white matter remodeling during stroke recovery of five patients with small ischemic MCA strokes. Qball MRI appeared to be more accurate than DTI in monitoring post-stroke plasticity because : 1. it showed more fiber trajectories at all time points ; it better unraveled longitudinal qualitative and quantitative modulations of fiber tracts ; it disentangled some fibers which appeared to "cross" the compact structure of the cortico-spinal tract. 1863. Real-Time MR Diffusion Tensor and Q-Ball Imaging Using Kalman Filtering Fabrice Poupon1, 2, Alexis Roche1, 2, Jean-François Mangin1, 2, Cyril Poupon1, 2 1CEA,NeuroSpin, Saclay, France; 2IFR49, Saclay, France Magnetic resonance diffusion imaging has become an established technique for inferring structural anisotropy of tissues and mapping the white matter connectivity of human brain. We developed a method to process diffusion tensor and Q-ball imaging in real-time. The basic idea is to use a Kalman filtering framework to fit either the linear tensor or Q-ball model incrementally, thus updating the model estimate after the acquisition of any new diffusion-weighted volume during ongoing scans. 1864. Comparison Between Q-Ball Reconstructions Using Radial Basis Function and Spherical Harmonic Basis Chun-Hung Yeh1, Kuan-Hung Cho2, Ching-Po Lin1, 3 1Department of Biomedical Imaging and Radiological Sciences, National Yang-Ming University, Taipei, Taiwan; 2Interdisciplinary MRI/MRS Lab, Department of Electrical Engineering, National Taiwan University, Taipei, Taiwan; 3Institute of Neuroscience, National Yang-Ming University, Taipei, Taiwan A systematical comparison between q-ball imaging (QBI) reconstruction using radial basis function (RBF) and spherical harmonic basis (SHB) is performed in this study. For both reconstruction methods, the accuracy in defining fiber orientations is assessed using a diffusion phantom model of 45¢X fiber crossing. The results show that both RBF QBI and SHB QBI are capable of resolving crossing fibers, while the separation angles may be underestimated using an insufficient b-value. We propose that b-value is the predominant factor to determine the intrinsic angular resolution in QBI. 1865. Multi-Shelled Q-Ball Imaging: Moment-Based Orientation Distribution Function Eizou Umezawa1, Mayo Yoshikawa2, Kana Ohno3, Emi Yoshikawa4, Kojiro Yamaguchi1 1Fujita Health University, Toyoake, Japan; 2NGK Insulators, Ltd., Nagoya, Japan; 3Gifu Public Health Center, Kakamigahara, Japan; 4Kyushu Koseinenkin Hospital, Kitakyusyu, Japan We propose a method that calculates an orientation distribution function (ODF) based on the moments of the probability density function of diffusion displacement. The method is based on q-ball imaging and an idea of "q-space imaging using small magnetic field gradient" that have been proposed in our previous work. The ability of the ODF to identify fiber crossing is also investigated by numerical simulation. The results imply the possibility that the moment-based ODF can reduce the number of data acquisition measurements in keeping higher ability to identify fiber crossing. 1866. Splines on the Sphere Q-Ball Imaging Nader Metwalli1, 2, John D. Carew, 3, Xiaoping P. Hu1 1Georgia Institute of Technology / Emory University, Atlanta, Georgia, USA; 2Cairo University, Cairo, Egypt; 3Georgia Institute of Technology, Atlanta, Georgia, USA We present splines on the sphere as a method of modeling raw high angular resolution diffusion imaging (HARDI) data with the added feature of smoothing of the diffusion data in the 3D diffusion space. With an extension to Q-Ball imaging (QBI), Splines on the Sphere QBI (S2QBI) resultant orientation distribution functions (ODFs) are of better quality than their QBI counterparts without compromising angular directionality which will be beneficial in QBI-based fiber tracking applications. |
Arterial Spin Labelling Methods Hall D Wednesday 13:30-15:30 1916. Comparison of Continuous Arterial Spin Labeling Perfusion MRI at 7T and 3T S. Lalith Talagala1, T-Q Li1, H. Merkle1, S. Wang1, J. Bodurka1, P. van Gelderen1, J Duyn1 1National Institutes of Health, Bethesda, Maryland, USA This work compares continuous ASL (CASL) perfusion data at 7T and 3T on the same subjects using the same protocol. Good quality 7T CASL data, which compared very well with 3T data, were obtained from all subjects. The measured perfusion signal (range 1.3-1.6%) was similar at the two field strengths. The intrinsic SNR of the difference signal was 55-95% better at 7T compared to 3T. However, the temporal SNR of the difference signal (range 0.7-1.6) was similar at both fields. Results indicate that 7T CASL data may be further improved by optimization of labeling efficiency and reduction of signal fluctuations. Robert Trampel1, Thies H. Jochimsen2, Toralf Mildner2, Carol Docherty2, Harald E. Möller2, Robert Turner2 1Max Planck Institute for Human Cognitive and Brain Sciences , Leipzig, Germany; 2Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany Perfusion imaging using magnetically labeled water is capable of measuring cerebral blood flow. As the signal-to-noise ratio increases with the main magnetic field strength B0 perfusion measurements at 7 T are expected to be more sensitive. For the quantification of perfusion, however, a reliable determination of the inversion efficiency α is required. A numerical method for predicting α under various conditions is demonstrated. At 7 T the specific absorption rate (SAR) limits the application of radio-frequency (RF) pulses. Therefore, we especially investigated whether it is possible to achieve sufficient inversion efficiencies in CASL experiments at high B0 and low B1 field strengths. 1918. Human Brain Perfusion MRI at 7T Using a Segmented True FISP ASL Method Elan J. Grossman1, Ke Zhang1, Jing An2, Yulin Ge1, Jian Xu3, Qun Chen1 1New York University, New York, New York, USA; 2Siemens Medical China, Beijing, People's Republic of China; 3Siemens Medical USA, Malvern, Pennsylvania, USA Ultra-high field (7T) MR imaging should benefit ASL methods due to increased SNR and longer T1 relaxation times. Here we assess the use of true FISP ASL as a feasible alternative to EPI-based methods of perfusion imaging at 7T. Preliminary results show that this sequence is capable of measuring local tissue perfusion in human brain at 7T with high spatial resolution and without the distortion artifacts commonly identified with EPI-based methods of ASL, although banding artefacts associated with true FISP imaging can sometimes degrade image quality. 1919. Hippocampal Blood Flow and Vascular Reserve: TrueFisp ASL at 3T Henry Rusinek1, Miroslaw Brys, Remigiusz Switalski, Francois Haas, Kellyanne Mcgorty, Mony de Leon, Qun Chen 1NYU School of Medicine, New York, New York, USA We have developed and tested a protocol based on segmented true FISP to measure regional cerebral blood flow (rCBF) in the hippocampal region of 12 young normal volunteers. The measurements were repeated during the carbon dioxide challenge achieved with a rebreathing technique. Baseline rCBF were 79.4±16.7 ml/100g/min in the right and 84.3±21.9 in the left hippocampus. During CO2 challenge, rCBF increased to 96.2±19.2 in the right and 98.5±22.9 in the left hippocampus. The technique is relatively quick and easily repeatable, which may be of particular importance in research on Alzheimer’s disease, epilepsy, and other neurological disorders of medial temporal lobe. 1920. Effects of CBV and Capillary Permeability on ASL Signal Changwei Wesley Wu1, 2, Jyh-Horng Chen2, Yihong Yang1 1National Institute on Drug Abuse, Baltimore, Maryland, USA; 2National Taiwan University, Taipei, Taiwan Although the arterial spin labeling (ASL) has been utilized to measure cerebral blood flow (CBF) for decades, none of the existing models considered the effects of cerebral blood volume (CBV) change on CBF quantification. In this study, the contributions of CBV and capillary permeability on the pulsed ASL signal were investigated using a two-compartment (2-comp) model. Results show that these physiological effects might induce up to 8% ASL signal change. CBV effect might have more influence on the estimated perfusion in animal models, while permeability might have more effect on human subjects.
1921.
Reliability and Reproducibility of Perfusion MRI in Elderly
Control Subjects
Li Jiang1, 2, Mina Kim1, 2, BettyAnn Chodkowski1, Manus J. Donahue1, 2, James J. Pekar1, 2, Peter C.M. Van Zijl1, 2, Marilyn Albert2 1Kennedy Krieger Institute, Baltimore, Maryland, USA; 2Johns Hopkins Medical Institution, Baltimore, Maryland, USA Intraclass correlation coefficients (ICC) and within-subject variation coefficients (WSC) were used to estimate reliability and reproducibility of pulsed arterial spin labeling (ASL) MRI. Twelve elderly healthy subjects were scanned four times during one year. The first two scans were processed by two raters, the results of which showed high reliability (ICC>0.96) and reproducibility (WSC <0.01). However, over the total period of one year, reliability reduced (ICC = 0.59~0.68) while the reproducibility still remained high (WSC < 0.08). It was also found that the coefficient of random noise increased to greater than 0.24, which limited the reliability of perfusion measurements. Sina Aslan1, Jinsoo Uh1, Perry Mihalakos1, Binu Thomas1, Carol Tamminga1, Hanzhang Lu1 1UT Southwestern Medical Center, Dallas, Texas, USA CBF and CBV are important physiological parameters associated with neuronal activity measured by ASL and VASO MRI, respectively. In this study, we provide a reference aCBV for normal humans during resting state in four major cerebral lobes: temporal>occipital> parietal>frontal, 4.46±0.55 > 3.89±0.76 > 3.07±0.45 > 2.85 ± 0.46, respectively. Additionally, the relationship between CBF and CBV was examined using exponential model, CBV=0.778*CBF0.497. These results suggest CBV values could be used as a reference biomarker to provide prognosis and diagnosis. Furthermore, CBV can be calculated when CBF is known by using exponential model. 1923. Effects of Pulsatile Flow on Arterial Input Function and CBF Quantification in Continuous ASL Tsukasa Nagaoka1, Xiaodong Zhang1, Robbie Champion1, Govind Nair1, Yoji Tanaka1, Thimothy Q. Duong1 1Emory University, Atlanta, Georgia, USA CBF is quantified by arterial input function (AIF) and the regional distribution of the tracer. However, AIF on ASL is calculated assuming an arbitrarily shaped AIF without considering pulsatile flow profiles. In this study, we simulated the effect of different flow profiles on the AIF. This model was applied to the arterial ASL signals obtained at different post-labeling delays with and without diffusion gradients in anesthetized monkey. We estimated the potential errors in CBF quantification due to flow dynamics over wide ranges of physiological and MRI measurement parameters. 1924. Distribution of Arterial Transit Times Investigated by MATISSE Toralf Mildner1, Karsten Müller1, Stefan Hetzer1, Robert Trampel1, Wolfgang Driesel1, Derek V.M. Ott1, D. Yves von Cramon, 12, Harald E. Möller1 1Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany; 2Max Planck Institute for Neurological Research, Cologne, Germany Mapping of Arterial Transit times by Intravascular Signal SElection (MATISSE) is introduced as an approach to map the distribution of arterial transit times, �a, in the human brain. The method is based on a smooth modulation of the efficiency of CASL between subsequent repetitions, and on the application of a short repetition time. Under these conditions, the difference in �a between two brain regions can be measured directly by evaluating the phase shift between the corresponding perfusion-weighted time series. CASL was performed by a separate labeling coil at the neck. The smooth modulation of the efficiency of CASL was achieved by changing the frequency offset of the labeling radiofrequency stepwise from repetition to repetition. It is demonstrated that MATISSE yields whole-brain � a difference maps which are consistent with respect to the vascular territories of the human brain.
1925.
Comparison Between Pseudo-Continuous and Separate Labeling
Coil in Continuous Arterial Spin Hesamoddin Jahanian1, Tiffany M. Love1, Gregory R. Lee2, Luis Hernandez-Garcia1 1University of Michigan, Ann Arbor, USA; 2Case western reserve University, Cleveland, USA Perfusion based functional MRI through Arterial Spin Labeling (ASL) has remarkable potential as a tool to study brain function because of its quantifiable nature and close relation to neuronal activity. Many techniques have been developed to increase the sensitivity of ASL perfusion MRI. Continuous labeling techniques (CASL) offer SNR advantages over pulsed labeling techniques (PASL). In this study, we compare two continuous labeling schemes in terms of their success in improving the SNR and their activation detection efficiency in vivo: flow driven adiabatin inversion using a separate labeling coil (Double coil CASL)and pseudo-continuous inversion (pCASL).
1926.
Simultaneous Quantification of Cerebral Blood Flow and
In-Vivo T2 and T2* of Cerebral Blood Using Guangping Dai1, Jie Lu2, 3, Yasu Egi2, Eng H. Lo2, Young Ro Kim1 1A.A. Martinos Center/MGH, Charlestown, Massachusetts, USA; 2Neuroprotection Research laboratory/MGH, Charlestown, Massachusetts, USA; 3Xuanwu Hospital, Beijing, People's Republic of China A new pulse sequence has been developed for quantifying in vivo intravascular blood oxygenation level dependence (BOLD) signal using continuous arterial spin labeling (CASL). Specifically, we have measured in-vivo T2 or T2* values of rat cerebral blood using the TE dependence of cerebral blood flow (CBF) signals obtained by CASL. With the transition of inhaled gas from air and 100% oxygen, we have demonstrated that our measurement scheme can be used for reliably quantifying the CBF and in vivo T2 and T2* changes of cerebral blood. 1927. Quantitative Perfusion Imaging Using Q2TIPS-FAIR PROPELLER EPI Huan Tan1, Craig A. Hamilton1, Robert A. Kraft1 1Wake Forest University School of Medicine, Winston-Salem, USA Perfusion imaging of the brain is greatly affected by subject motion and off-resonance effects. PROPELLER is an advanced imaging technique that is less sensitive to both motion artifacts and magnetic field inhomogeneity. We have implemented the Q2TIPS-FAIR PROPELLER EPI sequence that produced quantitative cerebral blood flow maps with improved image quality and spatial resolution. 1928. PASL Filtering: A Method of Improving Clinical Perfusion Imaging Huan Tan1, Joseph A. Maldjian1, Jonathan H. Burdette1, Andrew R. Deibler1, Jeffrey M. Pollock1, Robert A. Kraft1 1Wake Forest University School of Medicine, Winston-Salem, USA Arterial spin labeled MR imaging is a powerful method for quantitatively measuring cerebral perfusion. We have implemented ASL as a routine part of the clinical MR evaluation, acquiring 3500 cases over a 10-month period. In a small, but significant number of these cases, the data is uninterpretable, largely due to excess patient motion and system instability. We have implemented a post processing filter which can recover the perfusion signal in many of these previously uninterpretable cases.
1929.
A Systematic Approach to Optimizing Background Suppression
for Arterial Spin Labeling Perfusion Nasim Maleki1, Weiying Dai2, David C. Alsop2 1University of Toronto, Toronto, Canada; 2Beth Israel Deaconess Medical Ctr, Boston, Massachusetts, USA In this paper a systematic approach to optimizing background suppression for pulsed and continuous perfusion imaging is reported. We describe an algorithm to solve a nonlinear optimization problem with bounds on the pulse timings in order to decrease the background static signal for a broad range of T1s. Results of the optimization were added to pseudo-continuous ASL sequence and validated both in phantoms and in brain perfusion imaging in healthy subjects. Greater than 100 fold decrease in background signal intensity was successfully achieved such that the ASL signal change exceeded the residual signal from static tissue. 1930. Comparison of Pulsed Arterial Spin Labeling Sequences Using Different Absolute Quantification Methods Mustafa Cavusoglu1, Kamil Ugurbil1, 2, Kamil Uludag1 1Max-Planck-Institute for biological Cybernetics, Tuebingen, Germany; 2Radiology, Minnesota, USA Three different ASL schemes (Q2TIPS, FAIR-QUIPSSII and PICORE-QUIPSSII with asymmetric BASSI pulses) were compared measuring absolute cerebral blood flow (CBF) in the human brain at 3T. Signal-to-noise ratios (SNR), magnetization difference between control and tag images and three different absolute CBF quantification schemes were evaluated. For identical sequence parameters and voxels, it was found that a) FAIR-QUIPSSII has the highest SNR b) FAIR-QUIPSSII yields slightly higher CBF values c) absolute CBF values depend on the quantification scheme utilized. One possible explanation for the findings is that effective labeling efficiency and physiological noise contamination are different for the tagging procedures used. 1931. Variable-Density Spiral Improves Quality of Multi-Shot Arterial Spin Labeling Perfusion Images Eric Scott Paulson1, Douglas Edward Prah1, Kathleen Marie Schmainda1 1Medical College of Wisconsin, Milwaukee, USA By oversampling the center of k-space with each interleaf, variable-density spirals offer increased immunity against motion as well as facilitate self-navigated phase correction. We demonstrate here that variable-density spiral readouts improve the quality of multi-shot ASL perfusion images in brain. 1932. Quantification of Retinal Blood Flow Using a Pseudo-Continuous Arterial Spin Labeling Technique Nasim Maleki1, Weiying Dai2, 3, David C. Alsop2, 3 1University of Toronto, Toronto, Canada; 2Beth Israel Deaconess Medical Ctr, Boston, Massachusetts, USA; 3Harvard Medical School, Boston, Massachusetts, USA Measurement of perfusion to the retina has implication in the assessment of many important disorders of retinal vascular supply and growth which could lead to blindness. In this paper, we applied a pseudo-continuous arterial spin labeling technique to quantify the blood flow of the retina using multiple delay times, 3 Tesla imaging, and 8-channel phased array head coil on healthy subjects. The results of this study establish the feasibility of imaging and quantification of blood flow to the retina with contrast free MRI in humans and may also provide a unique window for the study of choroidal blood flow control and pathology. |
ENT: Tumors, Anatomy & Functional Hall D Wednesday 13:30-15:30 and During Early Treatment with Chemoradiation Predict Therapeutic Response? Ann D. King1, Hua Zhou1, 2, Brian K H Yu1, David K W Yeung1, Frankie Mo1, Gary M K Tse3, Alex C. Vlantis1, Anil T. Ahuja1 1The Chinese University of Hong Kong, Shatin, Hong Kong; 2First Affiliated Hospital of Zhejiang University, Hangzhou, People's Republic of China; 3Prince of Wales Hospital, Shatin, Hong Kong Diffusion-weighted imaging (DWI) is a MR technique that has been shown to be useful in the characterization of lesions. We investigated the usefulness of this technique in predicting treatment response in 31 patients with squamous cell carcinoma (SCC). ADC values from DWI of SCC decrease early in the course of treatment but the pre treatment ADC value and early change in ADC values does not seem to be able to predict tumor response.
1998.
Squamous Cell Carcinoma of the Head and Neck: Can 1H MRS
Pre-Treatment and During Early Treatment Ann D. King1, David K. Yeung1, Hua Zhou1, 2, Brian K H Yu1, Frankie Mo1, Gary M K Tse3, Alex C. Vlantis1, Anil T. Ahuja1 1The Chinese University of Hong Kong, Shatin, Hong Kong; 2First Affiliated Hospital of Zhejiang University, Hangzhou, People's Republic of China; 3Prince of Wales Hospital, Shatin, Hong Kong In vivo 1H MR spectroscopy has been shown to be useful in the characterization of tumors, including those in the head and neck. Whether this technique may be useful to predict treatment response following chemoradiation treatment is not known. We employed this technique to examine 39 patients with squamous cell carcinoma (SCC) and to compare spectroscopy results with tumor response assessed at 6 weeks after treatment. Our results showed that this technique has limited use to predict treatment response in patients with SCC. 1999. The Evaluation of Radiation-Induced Changes in Water Content of the Parotid Gland Using MRI Antonetta Christina Houweling1, C A.T. van den Berg1, T Dijkema1, J M. Roesink1, C H. J. Terhaard1, C. P. J. Raaijmakers1 1University Medical Center Utrecht, Utrecht, Netherlands The radiation-induced changes in water content of the parotid glands of five healthy subjects and three radiotherapy patients was evaluated with MR sialography and a multi-point Dixon technique. With MR sialography, the major ductal system of the parotid gland was visualized nicely. The water and fat content were successfully visualized using the multi-point Dixon technique. Radiation-induced changes in water content and the ductal system were not observed. However, the fat content appeared to change after radiotherapy. 2000. High Resolution MRI of the Parotid Gland and Duct at 7 Tesla Oliver Kraff1, 2, Jens M. Theysohn1, 2, Stefan Maderwald1, 2, Stefan Kruszona1, Mark E. Ladd1, 2, Elke R. Gizewski1, 2, Susanne C. Ladd1, 2 1Erwin L. Hahn Institute for MRI, Essen, Germany; 2University Hospital Essen, Essen, Germany Most examinations of gland diseases are still performed with X-ray sialography. MR imaging at ultra high fields (7T) offers an increase in SNR and hence spatial resolution as well as new image contrasts which might be able to replace conventional sialography in the future. Our study aimed to optimize different PD and T2 weighted GRE and TSE sequences at 7T for high resolution MRI of the parotid gland and duct. The final imaging protocol at 7T provides a non-invasive examination within 20 minutes as well as excellent image contrast and resolution of the parotid duct and branches. 2001. Motion Corrected 3D High Resolution Larynx Imaging with a Two-Coil Array Joëlle Karine Barral1, Edward J. Damrose1, Dwight George Nishimura1 1Stanford University, Stanford, California , USA Laryngeal cancer is usually treated with radiotherapy and chemotherapy to preserve speech and swallowing abilities. However, if the cartilage has been invaded, this treatment is ineffective and partial or total laryngectomy is required. The latter has a dramatic impact on the quality of life of the patient. If cartilage invasion is detected early, partial laryngectomy, with preservation of speech and swallowing, can be performed in lieu of total laryngectomy. A 3D trajectory is investigated to get coverage of the larynx at high resolution in less than 5 minutes with motion correction, and a two-coil array is proposed, which better fits the larynx geometry. 2002. Real-Time MRI of Swallowing in Upright Position Juan Manuel Santos1, Kim Butts Pauly1, Gerald R. Popelka1, John M. Pauly1 1Stanford University, Stanford, USA Swallowing disorders are difficult to study because they have multiple etiologies and involve the careful coordination of many different oral structures. These oral structures are not readily visible in their natural state, nor are their mechanics completely understood We have adapted the RTHawk real-time system to the SP 0.5 T GE interventional MRI scanner allowing a person to be scanned in an upright position. Swallowing function can then be studied in a natural position with continuous real-time imaging. 2003. Application of Compressed Sensing to 3D Imaging of the Vocal Tract for Speech MRI Yoon-Chul Kim1, Jon-Fredrik Nielsen1, Shrikanth Narayanan1, Dani Byrd2, Krishna S. Nayak1 1Ming Hsieh Department of Electrical Engineering, University of Southern California, Los Angeles, California , USA; 2Department of Linguistics, University of Southern California, Los Angeles, California , USA Three-dimensional (3D) MRI of the vocal tract during speech production has the potential to provide full anatomical information of the vocal tract shaping, but it requires prohibitively long scan time. In order to reduce scan time and achieve reconstructed image quality comparable to the image from fully sampled data, we apply the principle of compressed sensing MRI to 3D imaging of the vocal tract. Sustained unvoiced English fricative consonants were tested to demonstrate the effectiveness of the method with an acceleration factor of 3. 2004. MR Evaluation in Patients with Congenital Dysosmia Hui You1, Feng Feng2, Jian-Feng Liu2, Xue-Yan Wu2, Jian Wang, Zhengyu King 1Peking Union Medical College Hospital, Beijing, People's Republic of China; 2Peking Union Medical College Hospital, People's Republic of China To study the MR imaging findings of congenital dysosmia, 47 patients with congenital dysosmia including 39 patients with Kallmann syndrome and 8 with isolated dysosmia, and 21 age-matched normal volunteers underwent high-resolution MRI examination. All the patients had abnormal findings in olfactory bulbs and tracts. And the pituitary gland was smaller and the stalk was thinner in patients with Kallmann syndrome than volunteers. MR can readily depict the malformation of olfactory bulbs and tracts in congenital dysosmic patients. 2005. Field-Corrected Dynamic Imaging of the Velopharyngeal Musculature During SwallowCharles A. Conway1, Youkyung Song1, David P. Kuehn1, Bradley P. Sutton1 1University of Illinois at Urbana-Champaign, Urbana, USA Real-time dynamic MRI of swallowing is particularly difficult because field inhomogeneity due to air-tissue interfaces changes as the anatomy moves. The aim of this study is to reconstruct dynamic field-corrected images, allowing us to gain temporal resolution by lengthening readouts without paying the penalty of additional image distortion. This method has been applied to a spiral FLASH sequence for imaging the velopharyngeal musculature during swallow with 61 ms temporal resolution. Scans of healthy adults demonstrate the ability of MRI to compete with traditional methods of clinical swallowing evaluation including endoscopy and X-ray fluoroscopy. 2006. Magnetic Resonance Imaging in Orthodontics: Three-Dimensional Localization of Impacted Teeth Olga Tymofiyeva1, Kurt Rottner1, Florian Schmid1, Ernst-Juergen Richter1, Peter Michael Jakob1 1University of Wuerzburg, Wuerzburg, Germany Impacted teeth are teeth that are unable to erupt properly and remain under the gum. They often become infected or damage neighbouring teeth. Information about three-dimensional positioning of impacted teeth is invaluable in orthodontic diagnosis and treatment planning. The purpose of this study was to assess the feasibility of MRI of three-dimensional localization of impacted teeth in children and adults. MRI yielded a clear separation between impacted teeth and the surrounding tissue, and the position and angulation of impacted teeth in all three spatial dimensions could be assessed. 2007. In Vivo Caries Imaging Using Contrast-Enhanced Dental MRI Olga Tymofiyeva1, Kurt Rottner1, Florian Schmid1, Ernst-Juergen Richter1, Peter Michael Jakob1 1University of Wuerzburg, Wuerzburg, Germany For diagnosis of carious lesions and estimation of their extension X-ray imaging has been routinely used in clinical practice. However, the technique only provides two-dimensional projections, which yields only limited information regarding localization and size of the lesion. In this article the first in vivo demonstration of carious lesion visualization and quantification using contrast-enhanced dental MRI is presented. Advantages of this innovative technique for caries diagnostic include the three-dimensionality of lesion visualization and quantification, the possibility to determine relative position of the lesion in relation to the pulp and absence of ionizing radiation. |
Pediatric Brain: Metabolic, Leukodystrophy Hall D Wednesday 13:30-15:30 Stephan Gruber1, Margareta Holub2, Wolfgang Bogner1, Andreas Stadlbauer3, Martin Krssak2, Olaf Bodamer2 1Medical University of Vienna, Vienna, Austria; 2Medical University of Vienna, Austria; 3Landesklinikum St. Poelten, Austria Fabry and Gaucher disease are rare, progressive, inherited disorders of glycosphingolipid metabolism that affect multiple organ systems. Aim of this studies was to investigate central nervous system involvement using MR spectroscopic imaging. Absolute metabolite values revealed no differences between Fabry- and Gaucher- (Type 1) patients and controls. Previous work reported increased Cho in a Gaucher Type 1 patient, which can be explained by using different echo time and inappropriate control group. In this study we found significantly decreasing NAA/Cho in all three goups measured (between 5 and 8% per decade) pointing out the importance of sex- and age-matched controls. 2058. Quantitative Diffusion Tensor Imaging and T1 Relaxometry in Niemann-Pick C Disease christabel ec Lee1, David M. Thomasson1, Nicole M. Yanjanin2, Eva H. Baker2, Forbes D. Porter2 1National Institutes of Health, Bethesda, Maryland, USA; 2National Institutes of Health, Bethesda, Maryland, USA We imaged the brains of 14 patients with Niemann-Pick C disease using diffusion tensor imaging and T1 relaxometry to develop MR based quantitative measurements for the assessment of disease severity. We correlated patients’ symptom-based severity scores with fractional anisotropy (FA), apparent diffusion coefficient (ADC), and T1 relaxation time. For DTI, we found 8 different anatomical locations had statistically significant correlation to the symptom-based severity score, while for T1 relaxation, we found statistically significant correlations at 2 locations. We suggest DTI may be more sensitive than T1 relaxometry as a quantitative measurement for assessment of this disease. 2059. MRS Measurement of Disease Severity in Niemann-Pick Disease Type C Eva H. Baker1, Nicole M. Yanjanin2, Forbes D. Porter2 1Clinical Center, NIH, Bethesda, Maryland, USA; 2National Institute of Child Health and Development, NIH, Bethesda, Maryland, USA Niemann-Pick disease, type C is an autosomal recessive lipid storage disorder involving liver, spleen, lungs, and brain. Onset of symptoms is usually in childhood; rate of progression is variable, and patients can survive into adulthood. Neurologic findings follow a progressive degenerative course. Brain imaging findings are nonspecific; MRI is often normal. Seeking an MR-based measure of disease severity, we examined metabolite levels at 4 locations in the brain, and found statistically significant correlation between NAA+NAAG and symptom-based severity score at 2 locations in the cerebellum. Thus, NAA+NAAG levels could prove useful for monitoring disease progression and response to treatment. 2060. Quantitative T2 Mapping in a Niemann-Pick Type C Mouse Model John Totenhagen1, Katherine Thome1, Christine Howison1, Robert P. Erickson1, Theodore P. Trouard1 1University of Arizona, Tucson, USA Niemann-Pick Type C (NPC) disease is a fatal autosomal recessive neurodegenerative disorder with no current effective treatments. A quantitative measure of the progression and response to treatment of NPC disease is desired for evaluation of NPC therapies. In this work we have investigated quantitative T2 mapping in a mouse model of NPC disease. Mice were imaged with a radial FSE sequence and T2 maps created from individual radial datasets. Differences in T2 values were measured in white matter regions of NPC, transgenic NPC and control mice.
2061.
1H MR Spectroscopy of Creatine Deficiency Syndrome Caused by
a Novel Mutation
Monika Dezortova1, Filip Jiru1, Milan Hajek1, Vera Malinova2, Jan Petrasek1, Milan Jirsa1 1Institute for Clinical and Experimental Medicine, Prague, Czech Republic; 21st Faculty of Medicine, Charles University, Prague, Czech Republic Two siblings (male and female) were repeatedly examined using 1H MR spectroscopic imaging. First, MR spectra showed significantly reduced creatine concentration only in the male, whereas other metabolites were in normal range. Creatine concentration calculated from female’s spectra was near the lower control limit. Initial hypothesis of creatine transporter deficiency was proved when, one year later, creatine supplementation did not improve clinical status of the male and another decrease in creatine was observed in his spectra. Our findings strongly indicated the diagnosis which was confirmed by finding a novel nonsense mutation del c.219C in exon 1 of the SLC6A8 gene. 2062. L-2-Hydroxyglutaric Aciduria: Pattern of MRI Abnormalities in 53 Patients Marjan Elisabeth Steenweg1, N M. Verhoeven1, C Jakobs1, G S. Salomons1, M S. van der Knaap1 1VU University Medical Center, Amsterdam, Netherlands L-2-hydroxyglutaric aciduria is a rare, inherited metabolic disorder. Affected individuals only have neurological manifestations. We studied the brain-MRI pattern associated with the disease. Using a standard protocol for reviewing the MR images we found a consistent pattern of signal abnormalities. The most severe white matter abnormalities were seen in the frontal, subcortical white matter. In addition, bilateral involvement of the globus pallidus, caudate nucleus, putamen, and dentate nucleus was almost invariably seen. Over time the white matter abnormalities showed antero-posterior and centripetal progression and atrophy of the cerebral white matter occurred. 2063. MRI Findings and Neuropathology in a Mouse Model of Maple Syrup Urine Disease Jelena Lazovic1, Kathryn LaNoue2, Gregg Homanics3, William Zinnanti2 1California Institute of Technology, Pasadena, California , USA; 2Pennsylvania State University, Hershey, Pennsylvania, USA; 3University of Pittsburgh, Pittsburgh, Pennsylvania, USA A mouse model of intermediate maple syrup urine disease (iMSUD) is introduced. Elevated to highly elevated T2-values were present in the striatum, thalamus, occipital lobes and cerebellum indicating irreversible injury. Immunohistochemistry revealed intense vacuolation similar to spongiform changes throughout these regions, accompanied by severe loss of neuronal processes. We propose that prominent vacuolation rather than dysmyelination are contributing to increased signal on T2-weighted images in the iMSUD mouse model. The MSUD associated neuropathology was minimized when animals were placed on low branched chain amino acids (BCAA) diet, indicating the importance of BCAA control to prevent MSUD related brain injury.
2064.
Effect of Standard Oral Aminoacid Mixtures on Cerebral
Phenylalanine Content and the Dynamics of Roland Kreis1, Karin Zwygart1, Thomas Lutz2, Chris Boesch1, Joachim Pietz2 1University Berne, Berne, Switzerland; 2University of Heidelberg, Heidelberg, Germany Phenylketonuria is normally treated with dietary protein restriction and supplementation with the other amino acids (AA). It was shown earlier that oral substitution with high loads of large neutral AA can hinder the entry of additional phenylalanine (Phe) into brain for several hours. We now show with MRS that the same short-term effect can be reached with commercial AA powder in realistic dosage and that regular AA supplementation leads to lowering of cerebral steady state Phe levels on the long run because of its modulating influence on the BBB transport, which is described with asymmetric Michaelis Menton kinetics. 2065. Directional Diffusivity as an MR Biomarker of Axonal Injury in Leukodystrophy Tammie LS Benzinger1, Chin-I Chen, 2, Agus Priatna, Junqian Xu, Peng Sun1, Amy Lee, Anne Cross, Sheng-Kwei Song1, Soe Mar 1Washington University in Saint Louis, St. Louis, Missouri, USA; 2Wangang Municipal Hospital, Taipei, Taiwan Luekodystropy represents a broad category of genetic white matter diseases, characterized by varying degrees of myelin and axonal damage, depending upon the subtype. Frequently, these patients present for MR imaging without a known diagnosis. Although leukodystrophies have a very classic appearance on MRI, identifying the correct genetic subtype, which is critical for prognosis, is currently difficult by MRI alone. Here, we use axial and radial directional diffusivity to estimate the degree of axonal versus myelin damage, thus allowing for more precise diagnosis and prognosis. 2066. Multi-Slice Spectroscopic Imaging of Late Infantile Metachromatic Leukodystrophy (MLD) Lars G. Hanson1, Christine I. Dali2, Jens M. Fogh3 1Copenhagen University Hospital, Hvidovre, Denmark; 2Copenhagen University Hospital Rigshospitalet, Copenhagen, Denmark; 3Zymenex A/S, Roskilde, Denmark Late infantile metachromatic leukodystrophy (MLD) is a fatal genetic disease leading to brain degeneration after the first year of life. Since treatments are being developed, it is important to be able to diagnose and monitor the disease in the early stages. Few spectroscopic studies of this rare condition exist and little is known about the progression and spatial distribution of metabolic changes. Four children in different stages of the disease were scanned with multi-slice echo planar spectroscopic imaging. An almost complete loss of white matter NAA signal was found in the later stages of the disease. 2067. Contrast-Enhanced Ex Vivo MR Reveals Inflammatory Zone in X-Linked Adrenoleukodystrohpy Megan L. Blackwell1, W C. Hubbard, M Selig, Bruce R. Rosen1, Ann B. Moser, Florian S. Eichler 1Martinos Center for Biomedical Imaging, Charlestown, Massachusetts, USA We demonstrate with luxol fast blue-enhanced MR microscopy a zone of macrophages and myelin debris at the leading edge of active demyelination. Both immunohistopathology and biochemistry helped to explain the distinct contrasts seen in LFB-enhanced MR images: inflammation on T2*-weighting and NAWM on T1-weighting. Thus, a single contrast agent is able to measure different biochemical parameters.Our findings support the conclusion that enhanced T2* susceptibility in ALD samples is in part due to compartmentalization of the LFB contrast agent. Our technique of contrast-enhanced ex vivo MR microscopy allows for the investigation of lipids and inflammation in the process of demyelination. |
Hall D Wednesday 13:30-15:30 Daniel T. Franc1, Chris Kodl, Bryon Mueller, Ryan Muetzel, Elizabeth Seaquist, Kelvin O. Lim 1University of Minnesota, Minneapolis, USA Tract-specific fractional anisotropy and regional cortical thickness measures were found to be reduced and correlated in patients with Type 1 diabetes. Benjamin Schmitt1, Peter Bachert1 1German Cancer Research Center, Heidelberg, Germany This study demonstrates a possibility to trace the energy metabolism in human brain tissue non invasively by means of proton NMR spectroscopy. We observed that it is possible to detect resolved glucose signals in the brain of a patient with type 1 diabetes on a commercial 3-T MR scanner. This technique does not require specific equipment and can therefore be easily included in clinical routine MR examinations at 3 Tesla.
2102.
Glucose Transport and Neurochemical Profile in the
Hippocampus of STZ-Induced Diabetic Rats Under João MN Duarte1, 2, Rui A. Carvalho1, Rodrigo A. Cunha1, Rolf Gruetter2 1Center for Neuroscience and Cell Biology, Coimbra, Portugal; 2Laboratory for functional and metabolic imaging, Ecole Polytechnique Federale de Lausanne, Lausanne, Switzerland Diabetes mellitus accentuates the age-dependent decline of cognitive function which is related to reduction of hippocampal volume, a brain structure involved in learning and memory processing. In the present study, we measured the neurochemical profile and the kinetics of glucose transport in the hippocampus of streptozotocin (STZ)-induced diabetic rats using non-invasive 1H NMR spectroscopy. Our results show that chronic hyperglycemia does not alter brain glucose transport/content and that glycemia restoration immediately normalizes most neurochemical alterations. 2103. Cerebral H-MRS Correlates of Painless Diabetic Neuropathy Rajiv Gandhi1, Dinesh Selvarajah1, Celia J. Emery1, Paul D. Griffiths2, Solomon Tesfaye1, Iain D. Wilkinson2 1Royal Hallamshire Hospital, Sheffield, UK; 2University of Sheffield, Sheffield, UK Previous proton spectroscopy (H-MRS) studies have demonstrated abnormalities within the thalamus in patients with diabetic neuropathy (DN). It is unclear which parts of the brain’s sensory matrix are involved. A large 130 subject cohort was studied consisting subjects at different stages of DN plus controls without diabetes. Single voxel H-MRS was performed within the thalamus and somatosensory cortex at two TE’s. In subjects with painless DN, lower NAA/Cr was observed in the thalamus at long TE and lower NAA in the sensory cortex at short TE. There is spatial variation in H-MRS abnormalities in subjects with painless DN. 2104. Neurochemical Changes in Hippocampus of Developing Rats During Acute Hypoglycemia Ivan Tkac1, Kathleen Ennis1, Raghavendra Rao1 1University of Minnesota, Minneapolis, Minnesota, USA In vivo 1H NMR spectroscopy at 9.4 T was used to investigate neurochemical changes in hippocampus of 14-day-old rat pups during insulin induced hypoglycemia. Significant changes from homeostasis were observed for Asp, Glc, Gln, Glu, Lac, Cr and PCr. Time courses of individual metabolites revealed different mechanisms supporting energy production. When endogenous Glc resources ware exhausted, amino acids Gln, Glu became primary sources to feed TCA cycle through different anaplerotic pathways. |
Hall D Wednesday 13:30-15:30
2146.
SPHARM Detects Hippocampal Subfield Pathology in Temporal
Lobe Epilepsy
Hosung Kim1, Mark Niethammer2, Boris C. Bernhardt1, Sylvain Bouix2, Neda Bernasconi1, Andrea Bernasconi1 1Montreal Neurological Institute and Hospital, Montreal, Canada; 2Harvard Medical School, Boston, Massachusetts, USA Our purpose was to detect volume changes in the hippocampal subfields of patients with pharmacoresistant temporal lobe epilepsy (TLE) using SPHARM, a surface-based shape analysis method. We studied 95 TLE patients with unilateral hippocampal atrophy on MR volumetry and 46 controls. SPHARM applied to manual hippocampal labels measured a distance at each vertex between groups. In more than 40% of patients, we found a bilateral CA1 inward deformation more marked ipsilateral to the focus. These changes were negatively correlated with disease duration. Predominant CA1 atrophy with relative sparing of other subfields is in agreement with histopathological hippocampal sclerosis. 2147. Abnormal Gyrification of the Cerebral Cortex in Mesial Temporal Lobe Epilepsy Cathy Scanlon1, Lisa Ronan1, Colin Doherty2, Norman Delanty1, Mary Fitzsimons1 1Beaumont Hospital, Dublin 9, Ireland; 2St. James's Hospital, Dublin 6, Ireland The aim of this project is to identify abnormalities of cerebral cortical gyrification in mesial temporal lobe epilepsy using advanced quantitative magnetic resonance image-based techniques. Our results demonstrate a statistically significant reduction in temporal lobe gyrification. The identification of brain morphological abnormalities in TLE may help to further elucidate the pathogenesis of this disease.
2148.
Simultaneous FMRI and Electrophysiology During Intracerebral
Stimulation of Partial Seizures in Rats Dario J. Englot1, Asht M. Mishra1, Michael Purcaro1, Peter Herman1, Fahmeed Hyder1, Hal Blumenfeld1 1Yale University School of Medicine, New Haven, Connecticut, USA fMRI could be a useful tool for studying partial limbic seizures because of its success with other seizure models. However given patient movement during temporal lobe seizures, human fMRI recordings in temporal lobe epilepsy are technically difficult. Therefore, animal models would be extremely valuable. Here we describe an anesthetized rat model where electrical stimulation of the hippocampus induces limbic seizures which mimic complex-partial temporal lobe seizures in humans. We achieved reliable EEG and fMRI data, simultaneously acquired, during intracerebral stimulation. The network of increased BOLD signal changes occurred within expected areas based on separate neurophysiologic studies. 2149. Decrypting Cryptogenic Partial Epilepsy Using White Matter Fractional Anisotropy Analysis Simon Sean Keller1, Thomas Duning2, Christoph Kellinghaus2, Siawoosh Mohammadi2, Hagen Schiffbauer2, E Bernd Ringelstein2, Stefan Knecht2, Michael Deppe2 1University of Liverpool, Liverpool, UK; 2University of Muenster, Germany In patients with electroclinical evidence of cryptogenic partial epilepsy (PE) without brain abnormalities on conventional MRI, we show reduced fractional anisotropy (FA) in multi-lobar regions symmetrically, and asymmetrical temporal lobe FA reduction consistent with the side of the epileptogenic focus. We suggest that diffusion tensor imaging analysis of FA may represent an effective method for lateralising a neuroanatomical abnormality underlying the epileptogenic zone during pre-surgical evaluation of PE.
2150.
Multimodal MRI Reveals Secondarily Generalized Seizure
Related Microstructural Brain Tissue Jacobus F.A. Jansen1, M Eline Kooi1, Marielle C.G. Vlooswijk1, H J.M. Majoie1, Rianne P. Reijs1, Paul A.M. Hofman1, Klaas Nicolay2, Marc C.T.F.M. de Krom1, Albert P. Aldenkamp1, Walter H. Backes1 1Maastricht University Hospital, Maastricht, Netherlands; 2Eindhoven University of Technology, Eindhoven, Netherlands Patients who have suffered a high number of secondarily generalized tonic-clonic seizures (SGTCS) frequently show cognitive comorbidity. We have investigated whether a high number of SGTCS accumulated over life is associated with microstructural and metabolic changes in brain tissue. Sixteen patients with localization-related epilepsy with SGTCS underwent multimodal quantitative MRI (1.5 T), comprising T2 relaxometry, DWI, and MRSI. SGTCS are associated with significant changes in microstructural brain tissue characteristics within the frontal lobes, but not in the temporal lobes. These frontal changes possibly explain the cognitive problems which are often observed in patients with many SGTCS. This knowledge may help in the development of treatment aimed at preventing decline in cognitive abilities. 2151. Voxel Based Morphometry of 3D FLAIR Imaging in Cryptogenic Intractable Focal Epilepsy Kate J. Riney1, 2, Kling C. Chong2, J Helen Cross1, 2, Chris A. Clark1 1University College London Institute of Child Health, London, UK; 2Great Ormond Street Hospital for Children NHS Trust, London, UK Fluid-attenuated inversion-recovery (FLAIR) magnetic resonance imaging has improved the detection of focal cortical dysplasia (FCD) which can underlie intractable focal epilepsy (IFE). Recently methods have been developed to allow the acquisition of 3D FLAIR volumes. Voxel based morphometry (VBM) is an objective method of assessing 3D volumetric imaging, performing a voxel-wise comparison between an individual brain and a control group. In this study we report on a method for performing VBM of 3D FLAIR and validate this technique through reliably detecting the FCD in 7/8 children with IFE and visible FCD. In addition we report that VBM of 3D FLAIR detected abnormality in 4/14 children with cryptogenic IFE and in 2/4 the VBM abnormality was subsequently found to represent FCD not seen on previous image reporting. For the children in whom FCD has been detected with VBM of 3D FLAIR, epilepsy surgery may now become a possible treatment option. 2152. Combined EEG, FMRI, and Cognitive Testing in Childhood Absence Epilepsy Rachel Berman1, Michiro Negishi1, Marisa Spann1, Mi Hae Chung1, Michael Joseph Purcaro1, Edward J. Novotny1, R.Todd Constable1, Hal Blumenfeld1 1Yale University, New Haven, Connecticut, USA Patients with childhood absence epilepsy (CAE) have repeated 5-10 s episodes of unresponsiveness, associated with a “spike-wave” discharge on electroencephalography. The fundamental mechanisms of altered brain function and impaired attention in CAE are not known. Our central hypothesis is that SWD cause dysfunction in specific networks important for attention, including the medial frontal cortex and thalamus. If confirmed, this may lead to regional therapies targeted at improving impaired attention in patients with CAE. We will, therefore, map abnormal function in CAE, between and during seizures, using simultaneous EEG-fMRI and test attentional vigilance with the continuous performance task (CPT). 2153. Short and Long Term MRI Abnormalities After Experimental Febrile Seizures Jacobus F.A. Jansen1, Evi M.P. Lemmens1, Gustav J. Strijkers2, Jeanine J. Prompers2, M Eline Kooi1, Klaas Nicolay2, Govert Hoogland1, Walter H. Backes1 1Maastricht University Hospital, Maastricht, Netherlands; 2Eindhoven University of Technology, Eindhoven, Netherlands Experimental febrile seizures (FS) are known to increase the risk for eventual temporal lobe epilepsy. Quantitative multimodal MRI may provide early markers of microstructural and metabolic changes. FS were induced in 9-day old rats by hyperthermia exposure. Hippocampal volumetry and 1H-MRS, and cerebral T2 relaxometry and DTI was applied in FS rats and controls (24h and 8w later). At 9 weeks histological analysis was performed. Hippocampal T2 elevations at 24 hours appeared to be transient. DTI abnormalities detected in the amygdala at 24 hours persisted up to 8 weeks. Hippocampal volume was not affected. Histological analysis showed increased hippocampal (fiber density and anisotropy) and amygdalar changes. Quantitative MRI is able to detect both transient and long-term FS induced abnormalities in the limbic system that reflect microstructural alterations. These long-term findings support the concept that early-life FS are related to epilepsy.
2154.
MRI Measurements After Status Epilepticus Can Be Used as an
Early Biomarker for the Hippocampal Injury ManKin Choy1, Rodney C. Scott1, David L. Thomas2, Ken K. Cheung2, David G. Gadian1, Mark F. Lythgoe, 12 1UCL-Institute of Child Health, London, UK; 2UCL, London, UK Status epilepticus (SE) can cause hippocampal injury, with subsequent cognitive decline and development of temporal lobe epilepsy (TLE). We investigated these events using CBF, ADC and T2 in the lithium-pilocarpine rat model of SE and identified time-dependent MR changes that peaked on day 2. A strong relationship was identified between these peak changes and subsequent hippocampal volumes measured on day 21, which suggests that MRI measurements on day 2 after SE can be used as an early biomarker for the associated progressive hippocampal injury.
2155.
Perfusion MRI Changes During Status Epilepticus in the Rat
Pilocarpine Model
ManKin Choy1, Rodney C. Scott1, David L. Thomas2, Jack A. Wells2, Edward Proctor1, David G. Gadian1, Mark F. Lythgoe, 12 1UCL-Institute of Child Health, London, UK; 2UCL, London, UK There is a known relationship between convulsive status epilepticus (SE) and hippocampal injury, although the reasons for particular hippocampal injury remain uncertain. Potential mechanisms of injury include excitotoxicity and ischaemia. During the early phase of seizures, CBF increases in the cortex to meet energy demand, but it remains uncertain whether similar changes occur in the hippocampus. We have used perfusion MRI to investigate SE and have identified a limited CBF response in the hippocampus during pilocarpine-induced SE, which may in part account for the selective vulnerability. 2156. A New Method to Quantify Language Lateralization in FMRI Using Whole Brain Analysis Stephen E. Jones1, Shamseldeen Younes Mahmoud1, Mark J. Lowe1, Michael D. Phillips1 1CCF, Cleveland, Ohio, USA We present a new method to quantify language lateralization from fMRI studies, which uses a whole brain histogram approach. Advantages include no need to a priori threshold, and higher sensitivity for small activations.
2157.
TLE is Associated with Reduced Folding of the Temporal
Neocortex
Boris Christian Bernhardt1, Hosung Kim1, Neda Bernasconi1, Andrea Bernasconi1 1Montreal Neurological Institute and Hospital, Montreal, Canada Our purpose was to quantify the neocortical gyrification in temporal lobe epilepsy (TLE), and to investigate its relationship to underlying morphological parameters using MRI. We extracted skeletons of the temporal neocortex of 92 TLE patients and 33 controls. On the skeletons, we measured the fractal dimension (FD), total sulcal sulcal length (TL), and average sulcal depth. In the lateral temporal cortex, we observed a bilateral decrease of FD and TL in TLE with a left-sided seizure focus. FD was positively correlated to TL, but not to sulcal depth. A ”simplified” lateral temporal lobe may indicate a neurodevelopmental disruption in TLE. |
Innovative Imaging of the Normal Brain Hall D Wednesday 13:30-15:30 2215. The Normal Anatomy of the Substantia Nigra Using High Field 9T MRI: Correlation of Postmortem Images and Histopathology Mario Alcides Miranda1, Luke A. Massey1, Harry G. Parkes1, So Po-Wah2, John S. Thornton1, Janis Holton1, Laura Mancini3, P Cheshire1, K Strand1, Andrew J. Lees1, Tamas Revesz1, Tarek A. Yousry1 1Institute of Neurology, UCL, London, UK; 2Imperial College School of Medicine, MRC Clinical Sciences Centre, Hammersmith Hospital, London, UK; 3Lysholm Department of Neuroradiology, London, UK High resolution T2W MRI images at 9.4T of the midbrain targeted on the Substantia Nigra (SN) have been compared with histopathology samples. The detailed micro environment of the SN and the superior midbrain is described. The SN is seen as two layered band of tissue with a hypointense antero-lateral segment (par reticulata) and a hyperintense postero-medial component (pars compacta). This is probable the first MRI description of both segments of the SN (pars reticulata and compacta) that has been achieved. 2216. Susceptibility-Weighted Imaging (SWI) Delineates the Whole Length of the Optic Radiation Nobuyuki Mori1, Yukio Miki1, Chikara Maeda1, Seiko Kasahara1, Shin-ichi Urayama1, Nobukatsu Sawamoto1, Hidenao Fukuyama1, Kaori Togashi1 1Kyoto University Graduate School of Medicine, Kyoto, Japan This study examined whether low signal-intensity band (LSB) lateral to the lateral ventricles on 3T SWI represented the optic radiation (OR) by measuring profiles of LSB in 4 anatomical planes. Seven healthy volunteers entered. Comparison was also made with diffusion tensor tractography (DTT) of the OR. Measured profiles of the LSBs were consistent with the anatomical literature regarding the OR; even Meyerfs loops were clearly demarcated on SWI. DTT overlapped all LSBs posterior to the geniculate body, while at Meyerfs loop overlapped incompletely. Visualization of the whole length of the OR on SWI may have clinical relevance. 2217. Quantitative Cytoarchitectural Mapping of the Parahippocampal Region Neda Bernasconi1, 2, Thomas Mansi2, Gheorghe Postelnicu1, Jean C. Augustinack1, Bruce Fischl1 1Athinoula Martinos Center, Massachusetts General Hospital, Harvard Medical School, Boston, USA; 2Montreal Neurological Institute, Montreal, Canada Our purpose was to delineate cytoarchitectonic borders of parahippocampal region structures based on their laminar cortical pattern on ultra-high resolution MRI at 7T. We obtained ex-vivo 100 ?m isotropic images of the temporal lobe in three neurologically healthy subjects using a multiecho multi flip-angle sequence. The cortex was extracted by solving the Laplace Equation. This process generated streamlines between GM/CSF and GM/WM boundaries, and allowed computing laminar distribution of grey level intensities across the cortex. Applying multivariate statistics on laminar intensity profiles allowed distinguishing cytoarchitectonic borders between parahippocampal subregions that are currently invisible to conventional means of analysis. 2218. High-Resolution in Vivo MR Imaging of the Human Hippocampus at 7 Tesla Zang-Hee Cho1, Kyoung Nam Kim1, Young-Bo Kim1, Kyongtae T. Bae2, Chan-Hong Moon2 1Gachon University of Medicine and Science, Incheon, Republic of Korea; 2University of Pittsburgh, Pittsburgh, Pennsylvania, USA Volumetric atrophy of the hippocampus is associated with a wide range of neurodegenerative diseases (ND), such as Alzheimer¡¯s disease (AD), and has been thoroughly evaluated at 1.5 or 3 T magnetic field MR imaging. Neuronal deaths precede atrophy of whole hippocampus and affect its subdivisions regionally at different rates. Recent literature concerning in vivo MR imaging of the brain cortex at 7 T proposed that hippocampus layer structures may be discernible with imaging at ~200 μm pixel resolution. Thus, the purpose of our study was to demonstrate the feasibility of high resolution in vivo 7 T MR imaging of the detailed layers of the hippocampus, which may be used as a biomarker for the detection of early stage ND. 2219. Cortical Architecture of the Human HippocampusTie-Qiang Li1, Bing Yao1, Peter van Gelderen1, Hellmut Merkle1, Alan Koresky1, Jeff Duyn1 1NINDS, Bethesda, Maryland, USA High resolution T2*-weighted MRI was used to study the hippocampus of the human brain at 7T. Both the magnitude and phase information were analyzed to provide complementary information about the details of the cortical architecture in and around the hippocampus. The results provide convincing evidence that entorhinal cortical layer structures can be visualized in vivo using T2*-weighted MRI at 7T. 2220. In Vivo Investigation of Cortical Layers in Area V1 by High Resolution MR Imaging Ana-Maria Oros-Peusquens1, Sandro Romanzetti1, N. Jon Shah1, Robert Turner2 1Institute of Medicine, Research Centre Juelich, Juelich, Germany; 2Max-Planck-Institute for Human Cognitive and Brain Sciences, Leipzig, Germany We present a method which provides reliable 3D characterisation of Gennari’s stripe in vivo, using an affordable amount of measurement time at a widely available field strength of 3T. Images of six volunteers (4 male and 2 female, mean age 32 years) were acquired using a magnetisation-prepared 3D TSE sequence with nearly isotropic resolution of 0.4x0.4x0.5mm3. Between 3 and 5 separate volumes are reconstructed for each volunteer and averaged off-line. Besides optimisation of the white-grey matter contrast with the magnetisation prepared TSE, coregistration, complex averaging and parallel imaging turn out to be very important in the detection of myelination patterns in vivo 2221. Assessing Spatial Correspondence Between Subtle Cortical Dysplasias and Sulcal Pattern Pierre Besson1, Andrea Bernasconi1 1McGill University, Montreal, Canada We examined the spatial correspondence between small focal cortical dysplasia (FCD) and sulci on MRI in 43 patients. 21/43 (48%) patients had small FCD. 18/21 (86%) FCD had been overlooked. Sulci were identified automatically and visualized with FCD in 3D. We measured the depth of the sulci related to FCD and corresponding sulci in 21 controls. 18/21 (86 %) small FCD were located at the bottom of a sulcus. Sulci related to FCD were 23% deeper than in controls. Small FCD are located at the bottom of a deep sulcus. Sulcal morphology should be scrutinized when evaluating epileptics with FCD.
2222.
Quantitative Histological Validation of Fiber Orientation
Distributions Based on High-Angular Resolution Nate White1, Trygve B. Leergaard2, Ingeborg Bolstad2, Jan G. Bjaalie2, Helen D’Arceuil3, Alex de Crespigny3, 4, Anders Martin Dale5, 6 1University of California, San Diego, USA; 2University of Oslo, Norway; 3Massachusetts General Hospital, Harvard Medical School, USA; 4Oxford University, UK; 5UCSD, La Jolla, California , USA; 6Massachusetts General Hospital, Harvard Medical School, Charlestown, Massachusetts, USA
Novel diffusion-based magnetic resonance imaging techniques are increasingly applied to measure multi-directional tissue architecture. Although these methods are promising, there is a need for validation against measurements derived directly from the underlying tissue. We here compare standard diffusion tensor estimates and high-angular resolution diffusion-MRI derived fiber orientation distributions (FODs) against “gold standard” histologically-derived FODs of myelinated fibers in fixed rat brain tissue. In regions with unidirectional myeloarchitecture, both methods correlate well with average histological measurements. In regions with in-plane crossing myeloarchitecture, DT estimates correlate poorly, while the high-angular diffusion-based FOD estimates yield average correlations above 85%.
2223. Optimized In-Vivo High-Resolution Monkey DTI in Practice Xiaoxu Liu1, Jianhui Zhong1 1University of Rochester, Rochester, New York, USA DTI coupled with fiber tracking provides a unique tool to study the brain structure and function non-invasively. Non-human primate brain plays important roles in the related studies but fixed brains are often used. In this study, an optimized in-vivo high-resolution monkey DTI scan procedure was proposed and the impact of scan resolution on fiber tracking was analyzed. Using the proposed protocol, we are able to derive the diffusion tensor and track the fiber at relatively high spatial resolution with the DTI data acquired in an acceptable scan time. It is important to collect the real high-resolution data instead of interpolation for accurate tensor calculation and fiber tracking. 2224. T1, T2, and ADC of the Rat Retina at 7T Govind Nair1, Qinag Shen1, Timothy Q. Duong1 1Emory University, Atlanta, Georgia, USA The retina is about 200-300 μm thick, posing a significant challenge for MRI applications. Nonetheless, the highly structured retina with well-defined and multiple layers provides an excellent model to test high-resolution anatomical, physiological and functional MRI techniques. With increasing interests in using MRI to study the thin retina, knowledge of quantitative T1, T2 and ADC would help to systemically optimize structural, physiological and functional MRI contrasts of the retina. The goal of this study is to quantitatively measure T1, T2 and ADC in the rat retinas at very high spatial resolution. 2225. Super-Resolution in Layer-Specific MRI of the Retina Eric R. Muir1, Timothy Q. Duong1, Govind Nair1 1Emory University, Atlanta, Georgia, USA Super-resolution (SR) algorithms are post-processing methods that use sets of spatially shifted images to improve spatial resolution. Super-resolution applied in-plane to MRI has been problematic, and the consensus is that little information is added by SR from shifting the field of view. Physical motion of the object being imaged can be used to obtain the spatial shifts for SR MRI, but this technique has not yet been studied. We apply in-plane SR to in vivo MRI of the rat retina to improve layer-specific resolution by utilizing the spontaneously slow drifting movement of the eye to obtain the spatial shifts. 2226. Noninvasive Detection of Retina Degeneration in Mice Using Diffusion MRI at 11.74T Qing Wang1, Junjie Chen, Huiying Zhang, Shiming Chen, Sheng-Kwei Song1 1Washington University, Saint Louis, Missouri, USA Diffusion weighted images were acquired along three orthogonal directions on the retina in retina degeneration-1 (rd1) and the age matched control mice at 11.74T. Apparent Diffusion Coefficient maps were examined for both rd1 and control mice. The retina thickness of the rd1 mouse was thinner than that of the control due to the degeneration of photoreceptor cells, resulting in a single MR-detected retina layer observed in the rd1 mouse. This layer exhibited significantly higher ADC^ and ADC„U than the corresponding MR-detected inner retina layer of the control mouse. The current findings demonstrate that diffusion MRI can be a noninvasive tool to detect retinal degeneration in mice. 2227. In Vivo Measurements of Ocular Accommodation Eric Frederick1, 2, Kathryn Richdale3, Mirko Hrovat4, Samuel Patz2 1UMass Lowell, Lowell, USA; 2Brigham and Women's Hospital, Boston, Massachusetts, USA; 3Ohio State University, Columbus, Ohio, USA; 4Mirtech Inc, Brockton, Massachusetts, USA We observed ocular accommodation in vivo with a custom 1" surface coil over a taped eye. The contralateral eye was allowed to fixate on two optical stimuli: eye charts located 22cm away and another at optical infinity. This setup reduced motion artifacts associated with blinking and allowed for accurate measurements of the accommodative structures of the eye. The accommodative changes in the ciliary body ring diameter, equatorial lens diameter, and lens thickness were measured. The mean chord lengths of the ciliary body diameter and lens diameter both decreased when presented with the closer stimulus while the lens thickness increased as expected. 2228. Improved Structural MRI of Mouse Brain in Vivo by Combined T1 and T2 Mapping at 9.4 T Lars Kasper1, Jens Frahm1, Thomas Michaelis1, Susann Boretius1 1Biomedizinische NMR Forschungs GmbH am Max-Planck-Institut für biophysikalische Chemie, Göttingen, Germany Apart from high spatial resolution, a sufficient contrast-to-noise ratio is essential for distinguishing neuronal structures. Here, anatomical details of the cerebellum in healthy adult mice could be resolved by their relaxation characteristics in T1 and T2 maps obtained by high-resolution 3D MRI at 9.4 T. Respective data sets were exploited to improve the distinction of cerebellar white matter from the granular cell layer and molecular layer by combining the information of T1 and T2 maps in a suitably synthesized T1+T2 map. Christian T. Farrar1, Carsten D. Ley2, Walid Kamoun2, Young R. Kim1, Bruce R. Rosen1, Rakesh K. Jain2, A. Gregory Sorensen1 1Massachusetts General Hospital, Charlestown, Massachusetts, USA; 2Massachusetts General Hospital, Boston, Massachusetts, USA Relative vessel caliber index (rVCI) measurements may provide an important biomarker for assessing tumor angiogenesis. The aim of this study is to validate the MR methods for determining rVCI with in-vivo 2-photon microscopy data in a U87 mouse brain tumor model. In particular, two models for calculating the rVCI from MR measurements of R2* and R2, acquired before and after MION injection, have been proposed: a linear and a 3/2 power model. For the field strength (9.4T) and MION concentration (20 mg/kg) used in this study we find excellent agreement between the 2-photon data and the linear rVCI model. 2230. Venous Detectability with 9.4-T BOLD 3D Microscopy: Comparison with Two-Photon Microscopy Sung-Hong Park1, 2, Kazuto Masamoto2, Iwao Kanno2, Seong-Gi Kim1 1University of Pittsburgh, Pittsburgh, Pennsylvania, USA; 2National Institute of Radiological Sciences, Inage, Japan Rat brain vasculature was imaged at 9.4 T with blood oxygenation level dependent (BOLD) microscopy without exogenous contrast agent. Quantification in the somatosensory cortex shows a decrease of intracortical venous density with cortical depth, with smaller diameter veins being more numerous in shallow regions, while larger diameter veins drain the deeper cortical regions. Since venous diameters are not directly quantifiable by BOLD, we also measured diameter dependent intracortical venous density by in vivo two photon microscopy. Density comparisons between the two modalities show that veins as small as 15 µm diameter are detectable with 9.4 T BOLD microscopy under our imaging conditions. 2231. Angiography and Venography in a Single SWI Acquisition Samuel Barnes1, Elena Manova1, E. Mark Haacke1 1Wayne State University, Detroit, Michigan, USA We demonstrate that it is possible to collect the entire vasculature of the brain, both arteries and veins, in a single scan using an SWI sequence with a contrast agent or with MTC. Katharina Helm1, Jan Sedlacik1, Alexander Rauscher2, Hans-Joachim Mentzel1, Jürgen R. Reichenbach1 1Friedrich Schiller Universität, Jena, Germany; 2UBC MRI Research Center, Vancouver, Canada Caffeine enhances the BOLD (blood oxygenation level dependency) signal change by acting as an adenosine antagonist and thus changing the cerebral blood flow. Recent study has shown that caffeine has a contrast enhancing effect on cerebral veins using SWI. Aim of this study was to investigate moment and dimension of maximal signal change comparing caffeine-inured and -abstinent volunteers. 2233. Microangiography of the Primate Brain at 7 Tesla Using USPIO Particles Myriam Marianne Chaumeil1, 2, Christopher Wiggins1, Valdis Gudmundsdottir1, Jean-Sebastien Raynaud3, Eric Giacomini1, Yoann Buvat1, philippe hantraye2, vincent lebon1, gaelle louin3 1NeuroSpin, Gif-Sur-Yvette, France; 2MIRCen, Fontenay-aux-Roses, France; 3Guerbet, Roissy CDG, France Utrasmall SuperParamagnetic Iron Oxide (USPIO) particles were used to perform microangiography of the primate brain on a clinical 7 Tesla system. Cumulated doses of 5/15/45/135ìmol/kg were injected and the relationship between dose and contrast was evaluated qualitatively. Microvasculature was shown to be clearly visible at a 45ìmol/kg dose of USPIO, giving potential way for applications of this technique in humans. 2234. High-Resolution Contrast-Enhanced Magnetic Resonance Angiography of the Mouse Circle-Of-Willis Ketan B. Ghaghada1, 2, Gabriel Philip Howles, 12, Yi Qi1, G A. Johnson1, Srinivasan Mukundan1 1Duke University Medical Center, Durham, North Carolina, USA; 2Duke University, Durham, North Carolina, USA In clinical practice, magnetic resonance angiography (MRA) is a valuable tool for non-invasively evaluating neurovascular pathology. In the laboratory, transgenic mouse models for these diseases are available; however, translation of MRA techniques for use with MR microscopy has been limited by spatial resolution and signal-to-noise. In this work we use a surface-conjugated Gadolinium liposomal blood pool contrast agent to achieve adequate signal-to-noise to facilitate high-resolution (50x50x100 µm3) MRA of the mouse brain. Images generated at this resolution are able to demonstrate anatomic detail in the mouse that begins to approach anatomic detail seen in clinical MRA of humans. 2235. Physiological Component in Background Flow Velocity in MR Phase Contrast Measurements Stefan K. Piechnik1, Peter Jezzard2, James V. Byrne2, Paul E. Summers2, 3 1Oxford University , Oxford, UK; 2Oxford University, Oxford, UK; 3University of Modena, Modena, Italy We constructed a simple model to investigate the partial volume of phase encoded velocity components in a large reference region as typical in background correction of MR phase contrast images. We demonstrate that perfect nulling of velocity components from microvascular pools should be treated as an exceptional coincidence of factors, rather than the general case. Even though the magnitude of the effect shown in simulation and experimental data is well under 1mm/s it may affect the measurements of slow flowing compartments, such as CSF. 2236. In Vivo Characterization of White Matter Microvasculature Anisotropy with Diffusion-Weighted MRI Dimitrios C. Karampinos1, 2, Bradley P. Sutton1, John G. Georgiadis1 1University of Illinois at Urbana-Champaign, Urbana, Illinois, USA Cerebral microvasculature is generally thought as being isotropic and totally random. However, there are areas in brain parenchyma where the local network of the microvessels has a preferential orientation, typically aligned with the neural tracts. The aim of the current study is to use of low b-value Diffusion-Weighted Imaging regime to characterize in vivo the cerebral white matter microvasculature anisotropy. Extending the IntraVoxel Incoherent Motion (IVIM) technique, we extract direction-sensitive Cerebral Blood Volume values which are consistent with the DTI results for the orientation of neural tracts.
2237.
Brain Morphology and Vascular Patterning Analysis in Gli2-KO
Mutant Mice Using Contrast Enhanced Cesar Augusto Berrios-Otero1, Youssef Zaim Wadghiri, Alexandra Joyner2, Daniel H. Turnbull1 1Skirball Institute of Biomolecular Medicine, NYU School of Medicine, New York, USA; 2Memorial SloanKettering Cancer Center, New York, USA Vascular cell proliferation and patterning is critical for normal embryonic development and also underlies many disease processes such as tumor angiogenesis. In the current study, we use 3D micro-MRI to analyze brain defects in Gli2 mutant mice, which have previously been shown to have midbrain and cerebellum defects. Furthermore the Gli gene family have been implicated as transducers in the Hedgehog signaling pathway and thought to act upstream of VEGF and Notch pathways, which are important mediators of vascular development. Therefore, we also used contrast-enhanced 3D micro-MRI to analyze the cerebral vasculature of the Gli2 mutants and wild type littermates.
2238.
MRI Study in Rat to Evaluate the Effect of Cyclooxygenase
Inhibition on Blood Brain Barrier Disruption Saeid Taheri1, Eduardo Candelario-Jalil1, Eduardo Estrada1, Gary Rosenberg1, Rohit Sood1 1University of New Mexico, Albuquerque, New Mexico, USA Recombinant human Tumor Necrosis Factor (rhTNF-ƒÑ) is elevated in a variety of neuropathologies including multiple sclerosis and HIV encephalitis. In the context of ischemic stroke, several studies indicate that rhTNF-ƒÑ is involved in BBB disruption and the initiation of inflammation in the brain. There is increasing evidence that cyclooxygenase (COX-1 and 2 isoforms) inhibition limits BBB disruption following ischemic stroke but the mechanisms involved are unknown. In this study, an MRI technique for estimating barrier permeability coefficient, based on a graphical analysis method has been used for investigating the effect of inhibition of COX-1 and 2 isoforms by indomethacin on BBB permeability in a well characterized animal model of rhTNF-ƒÑ induced BBB disruption. The results suggest that MRI based permeability quantification technique would be a useful non-invasive tool to study and investigate the effect of pro-inflammatory cytokines on BBB damage and test the sensitivity of BB blocking effect of drugs. 2239. Oxygenation-Dependent 9.4-T BOLD 3D Microscopy Sung-Hong Park1, 2, Seong-Gi Kim1 1University of Pittsburgh, Pittsburgh, Pennsylvania, USA Blood oxygenation level dependent (BOLD) microscopy was performed for rat brain at 9.4 T with three different levels of inspired oxygen (30, 21, and 15%). The results confirmed that the hypointense pixels in the images were indeed of venous origin, and not magnetic susceptibility effects from sources other than venous dHb (e.g., arterial dHb, iron deposits, air/tissue/bone interfaces). When the systemic arterial oxygen saturation level was ~85%, BOLD microscopy revealed additional intracortical vessels presumed to be arteries. According to previous vascular histology, diameters of these intracortical artery-like vessels range 30-50 µm, in agreement with our computer simulations. 2240. Cortical Thickness Measurements as a Function of Choice of T1-Weighted Volume Rysia Burmicz1, A. Simone Reinders, Paula Dazzan, Andrew Simmons1 1Institute of Psychiatry, Camberwell, UK The impact on cortical thickness measures of using a range of different 3D T1-weighted volumes at 1.5T has been investigated with the aim of determining the sensitivity of these measures to acquisition parameters. Across the different volumes small systematic variations in whole brain cortical thickness measurements were found but data was found to be reasonably consistent. The mean thicknesses of cortical lobes varied consistently across volumes with the exception of the temporal lobes. This may be indicative of the challenges of isolating this region of cortical ribbon.
2241.
Multi-Parameter Mapping of the Human Brain at 1mm Resolution
in Less Than 20 Minutes
Nikolaus Weiskopf1, Gunther Helms2 1University College London, London, UK; 2Goettingen University, Goettingen, Germany A 3D FLASH-based protocol was designed for the human brain to obtain maps of T1, T2*, PD and MT at 1 mm resolution in less than 20 minutes. Spatial distortions were limited by acquisition of multiple gradient echoes with high receiver bandwidth; the associated loss of SNR was compensated by averaging the echoes. MT-maps provided high contrast between GM, WM and CSF without spatial bias and excellent delineation of basal cortical structures. T2*-maps identified iron-containing structures and vessels. Both promise to improve the segmentation of brain structures. 2242. High-Resolution Cross-Relaxation Imaging of the Rat Brain at 3.0T Hunter R. Underhill1, Chun Yuan1, Vasily L. Yarnykh1 1University of Washington, Seattle, USA We sought to determine the feasibility of capturing the in vivo white matter fiber tracts of the rat brain using cross-relaxation MRI at 3.0T. Pulsed Z-spectroscopic data with variable offset frequencies of the off-resonance saturation pulse were acquired with a 3D spoiled GRE pulse sequence. Through a constrained matrix model of pulsed magnetization transfer, we produced k- and f-maps of the in vivo rat brain. These quantitative maps yielded high-quality anatomic detail, which enabled the visualization of distinct white matter fiber tracts. This study provides compelling evidence for using this technique in translational research involving rat models of human disease.
2243.
Detection of the Myelin Water Fraction in 4 Tesla
Longitudinal Relaxation Data by Cross-Regularized Christian Labadie1, 2, Jing-Huei Lee3, Silvia Jarchow4, William D. Rooney5, Charles S. Springer5, Harald E. Möller1 1Max Planck Institute for Human Cognitive and Brain Science, Leipzig, Germany; 2University of Leipzig, Leipzig, Germany; 3University of Cincinnati, Cincinnati, USA; 4University of Münster, Münster, Germany; 5Oregon Health and Science University, Portland, USA The multiexponential analysis of normal brain 1H2O inversion recovery data, geometrically sampled starting at ca. 40 ms, reveals the presence of a small relaxographic peak at T1 = ca. 200 ms. This corresponds to 0.12 fraction of the total white matter water signal. It is attributed to the myelin water fraction (MWF), and suggests that water exchange in the human brain is sufficiently slow to enable the observation of water compartmentalization in longitudinal relaxation. In gray matter, the MWF represents ca. 0.05 of its total water and the peak is positioned at the same T1 value as in white matter. 2244. Measuring Cerebral Metabolism of Oxygen with Multimodal MRI and Near-Infrared Spectroscopy Qiong Zhang1, 2, Ying Wu, 1, Ursula Tuor1, Dave Kirk, 1, Foniok Tadeusz, Kumud Deka, MIchael Smith1, Jeff Dunn, 1 1University of Calgary, Calgary, Canada Cerebral metabolic rate of oxygen (CMRO2) is the measurement of O2 which is consumed in the brain by metabolic processes and is an important indicator of brain function. A non-invasive novel means of accurately measuring CMRO2, using a multimodality MR and near-infrared spectroscopy system, was presented in this study. Masafumi Harada1, Hitoshi Kubo, Naomi Morita, Hiromu Nishitani, Tsuyoshi Matsuda 1University of Tokushima, Tokushima city, Japan The purposes of this study are to evaluate reproducibility of GABA values measured by MEGA-PRESS at a clinical 3T machine and differences of GABA by female genital cycle depending on the brain locations. The reproducibility of GABA was considered to be acceptable under the clinical measurement because the ICC was more than 0.7. Though the GABA concentrations in the follicular state were statistically different depending on the brain location, this regional difference became ambiguous in the luteal state. The GABA concentration in the normal brain would be different depending on the location and influential by the female genital cycle.
2246.
3D Inhomogeneity Correction for MRI
Brian Coe1, Douglas P. Munoz1, Olivier Salvado2 1Queens University, Kingson, Canada; 2e-Health Research Centre - CSIRO, Brisbane, Australia High field magnetic resonance imaging (MRI) can give us beautiful images of the brain but even the best images suffer from intensity inhomogeneity. Neuroscientists face the difficult problem of correcting for intensity inhomogeneity when identifying white matter from grey matter. This delineation is required to mathematically flatten or inflate the cortex for further analysis. Here we present an improved 3D version of a local entropy minimization technique which deals with both local (tissue susceptibility) and global (receiver coil sensitivity) inhomogeneity issues. This new technique also corrects for intersubject variability to assist with an automated grey matter/white matter segmention. Robert E. Watson Jr. 1, Matt A. Bernstein1, Krzysztof R. Gorny1 1Mayo Clinic, Rochester, Minnesota, USA Initial experience of scanning 9 patients with an implanted vagal nerve stimulator (VNS) at 3T is described. All exams were performed with a transmit/receive head coil. Average head SAR, as calculated and displayed by the scanner, was kept below 1.3 W/kg as monitored by an MRI physicist. The VNS was turned off, the patients were instructed to use the squeeze-ball alarm at the onset of heating or any other discomfort. Patient well-being was monitored further by intercom communication between all imaging series. Although no patient discomfort or adverse event was reported, extreme caution should be exercised when generalizing this experience. |
Advanced Imaging of the Spine & Spinal Cord Hall D Wednesday 13:30-15:30 Nashiely Pineda-Alonso1, Michael Benatar2, Robert Smith1, Xiaoping Hu1, John Carew3 1Emory University, Atlanta, USA; 2Emory University. School of Medicine, Atlanta, USA; 3Emory University. School of Public Health, Atlanta, USA 1H-MRS of the spinal cord has been applied to observe the differences between multiple sclerosis patients, spinal tumor patients and healthy subjects, comparing sometimes the concentrations of major metabolites within the spinal cord to other regions of the central nervous system. The purpose of this study was to quantify the intra-subject variability (reliability) of the spectroscopy of the spinal cord at 3T, a useful estimate for determining the detectability of metabolite changes that are measured longitudinally within-subject. Our pooled standard deviation estimates of intra-subject reliability are the first to be reported in the literature. Niousha Foad Ghazni1, Cathy M. Cahill1, Caroline F. Pukall1, Natalie Kozyrev1, Patrick W. Stroman1 1Queen's University, Kingston, Canada To understand how peripheral sensitization affects the nervous system, we must first examine the transmission pathways in the spinal cord and brainstem of healthy individuals. Spinal fMRI studies, using a HASTE sequence, were carried out in 26 healthy individuals. Innocuous and noxious stimuli were applied before and after sensitization using capsaicin. The findings from this study suggest that peripheral sensitization alters what would normally be an innocuous response to that of a painful response, strongly implying pain pathway activation. 2289. Diffusion Tensor Imaging and Quantitative Tractography in Chronic Spinal Cord Injury Benjamin Michael Ellingson1, Shekar N. Kurpad2, John L. Ulmer2, Brian D. Schmit1 1Marquette University, Milwaukee, Wisconsin, USA; 2Medical College of Wisconsin, Milwaukee, Wisconsin, USA This study focuses on diffusion tensor imaging and quantitative tractography in subjects with chronic spinal cord injury. We observed significant differences in fiber tract density, as well as rostral-caudal asymmetry at the lesion site in subjects with chronic lesions. 2290. 3D Myelin Water Imaging of Cervical Spondylotic Myelopathy at 3T Erin Leigh MacMillan1, Armin Curt1, 2, Burkhard Mädler1, David K B Li1, Marcel F. Dvorak1, 2, Alex L. MacKay1 1University of British Columbia, Vancouver, Canada; 2International Collaboration on Repair Discoveries, Vancouver, Canada Cervical spondylotic myelopathy (CSM) is characterized by spinal canal narrowing leading to cord damage, however, conventional MRI cannot elucidate the underlying pathological processes. We present a multi-echo T2 relaxation measurement, which resolves the ratio of myelin-associated water to the total water through analysis of the T2 decay curves, providing a myelin water fraction (MWF). This study presents the first MWF determined for spinal cord white matter, and illustrates the promise of T2 relaxation imaging as a tool to investigate pathological changes in spinal cord white matter with a successful application to CSM. 2291. Longitudinal Dynamic Contrast-Enhanced (DCE-) MRI of Spinal Cord Injury in Mouse Mehmet Bilgen1 1Medical University of South Carolina, Charleston, USA Spinal cord injury (SCI) disrupts blood spinal cord (SC) barrier (BSCB) permeability, which leads to secondary injuries, resulting in permanent neurological deficits. Recent trend indicates that mouse models are gaining popularity in SCI research, despite that the resulting inflammation response is different than rat. Our goal is to demonstrate the feasibility of performing DCE-MRI of SCI and determine vascular changes to explain why the neuropathology of injured SC in mouse is different. We also provide insights on how to develop pharmacokinetic models to quantify the BSCB permeability and explore surrogate bioimaging markers for assessing the functional status of injured SC. 2292. Neuronal and Vascular Changes in Syringomyelia: Investigations Using Longitudinal MRI Mehmet Bilgen1 1Medical University of South Carolina, Charleston, USA Syringomyelia is a heterogeneous collection of neuropathological conditions characterized by abnormal fluid-filled cavities (syringes) within spinal cord (SC). This condition develops in significant portion of patient population with spinal cord injury (SCI). The mechanisms causing the formation and enlargement of syringes following SC trauma are complex and still not fully understood. Using longitudinal MRI, the current study aims to show the evolution of neuronal and vascular changes, describe the onset and advancement of syringomyelia and demonstrate how this condition affects the overall structure and function of the injured SC. 2293. Diffusion Weighted SSFSE with Dixon Fat-Water Separation Ken-Pin Hwang1, Jingfei Ma2, Ananth J. Madhuranthakam3, Eric T. Han4, Wei Sun5, Zachary W. Slavens5, David C. Alsop6, 7 1GE Healthcare, Houston, Texas, USA; 2University of Texas M.D. Anderson Cancer Center, Houston, Texas, USA; 3GE Healthcare, Boston, Massachusetts, USA; 4GE Healthcare, Menlo Park, California , USA; 5GE Healthcare, Waukesha, Wisconsin, USA; 6Beth Israel Deaconess Medical Center, Boston, Massachusetts, USA; 7Harvard Medical School, Boston, Massachusetts, USA High lipid signal in diffusion weighted images could potentially mask lesions near or within fatty tissue. Image-based Dixon methods have demonstrated excellent fat suppression but are not readily compatible with EPI acquisitions. In this work we insert time shifts in a diffusion weighted SSFSE sequence that provides phase information from the start of the echo train. In-phase and out-of-phase images are acquired with this sequence and are passed into a recently developed Dixon algorithm (MEDAL). We demonstrate that undistorted diffusion weighted images with uniform fat suppression can be obtained in challenging areas such as the cervical spine. Seth A. Smith1, 2, Visar Belegu1, 2, Bennett A. Landman2, BettyAnn A. Chodkowski1, 2, Jonathan A.D. Farrell2, Peter C. van Zijl1, 2, John W. McDonald1, 2 1Kennedy Krieger Institute, Baltimore, Maryland, USA; 2Johns Hopkins University, Baltimore, Maryland, USA Spinal cord damage contributes to functional deficits in patients with traumatic spinal cord injury (SCI). Clinical and conventional MRI methods may lack specificity for prognosis and identification of the potential for recovery. Thus, a spinal cord MRI protocol that quantitatively examines the tract-specific structural integrity that can be related to function would be beneficial in the clinic. We present a combined diffusion tensor and magnetization approach that shows aberrant metrics in a patient with SCI compared to healthy controls. DTI tractography and column-specific metrics may be particularly useful in revealing the microstructural evolution of SCI in vivo. 2295. IDEAL ASSIST (Automated Spine Survey Iterative Scan Technique): Metastasis Detection Kenneth Lee Weiss1, Rebecca S. Cornelius2, Elyse E. Lower3, Jane L. Weiss4, Dongmei Sun5, Jun Ying6, Bruce W. Mahoney2, Maria I. Altbach7, Mariano Fernandez-Ulloa2 1University of Cincinnati, Cincinnati, Ohio, USA; 2University of Cincinnati, Cincinnati, USA; 3University of Cincinnati, Cincinati, USA; 4WestImage, Cincinnati, USA; 5Jiaotong University, Beijing, People's Republic of China; 6Univeristy of Cincinnati, Cincinnati, USA; 7University of Arizona, Tucson, USA We prospectively tested IDEAL (Iterative Decomposition of Water and Fat with Echo Asymmetric and Least-squares Estimation) ASSIST (Automated Spine Survey Iterative Scan Technique) in 24 breast cancer patients at 3T. The novel methodology provided high sensitivity, specificity and inter-rater agreement for the detection of spinal metastases, performing better than PET-CT. In all studies, FSE IDEAL provided uniform fat and water separation throughout the entire 70cm FOV. FGRE ASSIST afforded subminute submillimeter in-plane resolution with high contrast between disc and vertebrae in two contiguous breath-hold series. The rapid automated high resolution and contrast MRI technique holds significant promise for metastasis surveillance. Joshua Levitz1, Burkhard Mädler1, Brian Kwon1, David Wilson1 1University of British Columbia, Vancouver, Canada Delayed gadolinium enhancement of cartilage has been used to identify cartilage degeneration in synovial joints. The same underlying mechanisms of degeneration are present in the intervertebral disc as in cartilage. This study looks at the feasibility of using anionic contrast agent uptake in porcine intervertebral discs, in-vitro, to quantify disc degeneration. T1 maps of healthy and chemically degenerated discs after immersion in anionic contrast agent were created. Degenerated discs showed a greater change in T1 times after contrast agent enhancement compared to healthy specimens, showing this to be a feasible technique.
2297.
Non-Subjective Spatial and Temporal Evolution of Damage
Following Spinal Cord Injury: Automated and Revital Nossin-Manor1, 2, Yoram Cohen1 1Tel-Aviv University, Tel-Aviv, Israel; 2The Hospital for Sick Children, Toronto, Canada Quantification of spinal cord damage following trauma is difficult. In the past, we have demonstrated the utility of using high b-value q-space diffusion-weighted MRI to follow the spatial and temporal damage evolution after spinal cord hemi-crush injury. In these studies spinal cords were assigned according to their unilateral trauma severity. The spinal cord were removed and fixed 5 days, 10 days and 6 weeks after hemi-crush and scanned by MRI using heavily diffusion-weighted imaging. Here we preset automated and semi-automated recognition methods to obtain non-subjective spatial and temporal evolution of the damage following spinal cord injury (SCI).
2298.
Exploring the Information Content of Q-Space Diffusion
Weighted Imaging: Application to Multiple Bennett Allan Landman1, Jonathan Andrew David Farrell1, 2, Seth A. Smith2, 3, Peter A. Calabresi3, Peter C. van Zijl, 23, Jerry L. Prince1 1Johns Hopkins University, Baltimore, Maryland, USA; 2Kennedy Krieger Institute, Baltimore, Maryland, USA; 3Johns Hopkins University School of Medicine, Baltimore, Maryland, USA Q-space imaging is an emerging analysis technique for diffusion weighted images that shows promise for assessing tissue microstructure. Using a robust analysis method that accounts for image noise characteristics, we investigated q-space MR findings in the in vivo human spinal cord for a control and two patients with multiple sclerosis (MS). The probability density function (PDF) for diffusion showed substantial heterogeneity, which may be indicative of substructure within the white matter and gray matter (e.g., dorsal root collaterals). We present a compact and intuitive visual representation of the information content in the PDF in healthy tissue and in MS lesions. 2299. Direct Imaging of Spinal Cord Axons in Intact Lamprey by Diffusion-Weighted μ-MRI Alexander C. Wright1, Guixin Zhang1, Suzanne L. Wehrli2, Michael E. Selzer1, Felix W. Wehrli1 1University of Pennsylvania, Philadelphia, Pennsylvania, USA; 2Children's Hospital of Philadelphia, Philadelphia, Pennsylvania, USA Sea lampreys (Petromyzon marinus) are used to model axonal regeneration following spinal cord injury, since axons in these animals spontaneously regenerate. Recently µ-MRI has been used to image axon tracts in excised lamprey spinal cord, as well as regional differences in diffusion anisotropy. This suggests the possibility of in vivo serial imaging of axonal recovery from injury, provided sufficient SNR and diffusion-weighted contrast can be achieved. As a step toward this goal, we here report direct visualization of spinal cord axons in an intact lamprey by diffusion-weighted µ-MRI. 2300. Sub-Acute DTI Predicts Long-Term Behavioral Outcome in the Mouse Models of Spinal Cord Injury Tsang-Wei Tu1, Joong Hee Kim2, Sheng-Kwei Song2 1Washington University School of Engineering, St. Louis, Missouri, USA; 2Washington University School of Medicine, St. Louis, Missouri, USA SynopsisCurrently, there is no effective noninvasive method capable of predicting long-term neurological outcome in spinal cord injury (SCI). In this study, we applied the diffusion tensor imaging (DTI) to a mouse model of SCI acutely (~ 3 hrs after injury) and sub-acutely (7 days pst injury) to evaluate the intact (or survived) white matter after SCI. The determination of residual ventrolateral white matter (VLWM) using DTI derived anisotropy index was demonstrated to predict the long-term behavioral outcome of the injured mice. Our results suggest that anisotropy map derived by DTI measurement, specifically sub-acute phase, has potential to estimate non-invasively the extent of the residual VLWM and predict the long-term behavioral outcomes in SCI. 2301. Correlating the Decreased Axial Diffusivity with Morphological Changes After Axonal Injury Tsang-Wei Tu1, Joong Hee Kim2, Mingqiang Xie2, Sheng-Kwei Song2 1Aerospace and Structural Engineering, Washington University School of Engineering, St. Louis, Missouri, USA; 2Washington University School of Medicine, St. Louis, Missouri, USA SynopsisRecently, Diffusion Tensor Imaging (DTI) has been used to quantify the severity of white matter damage in rodent models of Spinal Cord Injury (SCI). It has been demonstrated that axial diffusivity derived from DTI was a sensitive biomarker of axonal damage in SCI. It is hypothesized that the axonal morphology plays a key role influencing the modulation of axial diffusivity after injury. In this study, the comparison of in vivo DTI and confocal microscopic images were conducted on the Yellow Fluorescent Protein (YFP) mice. Swelling, beading, and fragmentation of YFP+ axons were found in the injured cord paralleling the decreased the axial diffusivity. The present result suggests that the axial diffusivity strongly depends on the morphological changes resulting from the injury to axons. 2302. Brain DTI of Spinal Cord Injured Rats Jaivijay Ramu1, Juan Herrera1, Raymond Grill1, Tobias Bockhorst1, Ponnada Narayana1 1University of Texas, Houston, USA In order to understand the neural basis for the extensive cortical plasticity seen on fMRI in experimental spinal cord injury, brain DTI and immunohistochemical studies were performed on normal and injured rats. These studies revealed fiber plasticity in the internal capsule and cerebral peduncle. The plasticity includes both strengthening of the connectivity of the existing fibers, as inferred by the increased FA values, and increased number of fibers, as inferred from the tractography and GAP43 expression. Juan Jose Herrera1, Jaivijay Ramu1, Alex Li1, Tessy Chacko1, Ponnada A. Narayana1 1Univ. of Texas Health Science Center at Houston, Houston, Texas, USA Fractional anisotropy (FA), based on diffusion tensor imaging (DTI), provides important information about tissue integrity. The value of FA depends on a number of factors that include degree of myelination and axonal morphometry and thus lacks pathologic specificity. There is some literature, mainly based on animal studies, that the individual diffusivities, longitudinal (?l) and radial (?t) improve pathologic specificity over FA. In this study we performed DTI and histology in both spinal cord and brain in experimental animal models. These studies demonstrate in some instances there is no direct correlation between the individual diffusivities and myelin and neurofilament. 2304. Actively Decouled Two Coil System Enables in Vivo DTI of Mouse Cervical Spinal Cord at 4.7 T Joong Hee Kim1, Justin Haldar2, Zhi-Pei Liang2, Sheng-Kwei Song1 1Washington University School of Medicine, St. Louis, Missouri, USA; 2University of Illinois at Urbana Champaign, Urbana Champaign, Illinois, USA In vivo DTI was performed for mouse cervical spinal cord at a 4.7 T magnet. Actively decoupled volume coil (RF excitation) and saddle type surface coil (signal receiver) provided good SNR to perform in vivo DTI within one hour using the conventional spin echo diffusion weighted imaging sequence. The acquired DTI maps revealed anisotropic characteristics of white matter and dorsal gray matter. Also, the coherent, axially elongated axonal fiber tracts can easily be seen with diffusion ellipsoids. The present results showed feasibility of in vivo diffusion observation of mouse cervical spinal cord at 4.7 T with reduced scan time. |
Physiological Characterization in fMR Hall D Wednesday 13:30-15:30 in the Visual Cortex Andrew S. Nencka1, Eric S. Paulson1, Daniel B. Rowe1 1Medical College of Wisconsin, Milwaukee, Wisconsin, USA Dependent upon deoxyhemoglobin concentration, the blood oxygen level dependent (BOLD) signal used in functional MRI is sensitive to both active parenchyma and down-stream draining veins. Some methods have been previously introduced which utilize phase data to identify and discount draining veins in cortical activation computations. In this abstract we consider the phase regressor and complex constant phase statistical methods in data with obvious draining veins to evaluate the efficacy of the statistical methods. In agreement with previous simulations, the constant phase method more robustly biases against draining veins. 2339. Estimating Cerebral Blood Volume with Expanded VASO Slice Coverage Christopher B. Glielmi1, Ronald A. Schuchard2, Xiaoping P. Hu1 1Georgia Institute of Technology / Emory University, Atlanta, Georgia, USA; 2VA Medical Center, Atlanta, Georgia, USA Recent research efforts have focused on quantifiable complements to blood oxygenation level dependent (BOLD) contrast including cerebral blood flow (CBF) measurement with arterial spin labeling (ASL) and cerebral blood volume (CBV) estimation with vascular space occupancy (VASO). The present work combines simultaneous VASO, CBF and BOLD acquisition with a previous CBV estimation model based on singular VASO acquisition. Furthermore, this work expands this model from a single slice to multiple slice acquisition and models effective blood relaxation to account for the fresh inflow of unsaturated blood, a previous confound of CBV quantification based on VASO contrast. Yuji Shen1, Ida Mengyi Pu2, Risto A. Kauppinen3 1University of Birmingham, Birmingham, UK; 2Goldsmiths, University of London, London, UK; 3Dartmouth Medical School, Hanover, USA We investigated partial volume effects of cerebrospinal fluid (CSF) and white matter (WM) on the recently introduced grey matter nulled (GMN) fMRI. In GMN fMRI, grey matter (GM) signal is suppressed by inversion recovery approach to reveal cerebral blood volume (CBV) change to brain activation. However, residual CSF and WM contributions in the inversion recovery scans will contaminate the desired blood signal, causing the true CBV change being underestimated and overestimated, respectively. Residual CSF is a dominant source of error for CBV estimation with GMN fMRI, giving more than 75% underestimation of true CBV change when CSF fraction > 0.1. Yuji Shen1, Ida Mengyi Pu2, Risto A. Kauppinen3 1University of Birmingham, Birmingham, UK; 2Goldsmiths, University of London, London, UK; 3Dartmouth Medical School, Hanover, USA We explored intravascular and extravascular BOLD signals at 3T using grey matter nulled (GMN) and VASO fMRI techniques. Neither GMN nor VASO fMRI showed practical contamination from large draining veins. The activation maps obtained by GMN and VASO fMRI were used to delineate intravascular and extravascular activation regions for the BOLD response. At the spatial resolution of 2x2x5 mm3, the intravascular BOLD signal change is larger than the extravascular one, and the BOLD response outside the CBV increase areas was only 30% of that determined in GMN and VASO positive parenchyma. 2342. Evaluation of Hypercapnic Tasks to Hemodynamically Scale Activation-Induced FMRI-BOLD Signals Nirvish Shah1, Sridhar S. Kannurpatti1, Bharat B. Biswal1 1UMDNJ, Newark, New Jersey, USA Hemodynamic scaling of fMRI responses reduces BOLD signal variation due to vascular sensitivity. Scaling of fMRI-BOLD response to a neural task is currently carried out using vascular reactivity information usually determined by a hypercapnic task such as breath hold (BH), breathing CO2 or hyperoxia. Hemodynamic scaling is accomplished by simply dividing the task induced response signal with the response obtained during a hypercapnic task. However, the optimal level of hypercapnia obtained either during the BH task or breathing CO2 for hemodynamic scaling is not yet estimated. If the hypercapnic response in a corresponding voxel is greater than the neural-task response in the same voxel, then it has a greater chance to get scaled out (loses significance as a result of scaling). To optimize hemodynamic scaling using hypercapnia tasks such as BH or CO2, we have studied different durations of inspirational BH and breathing various levels of CO2 in healthy young individuals. We have estimated the extent of scaling during each condition on the neural task-induced fMRI-BOLD response. 2343. Effects of Oxygen and Carbon Dioxide on BOLD, CBF, CBV and VSI in Anesthetized Rats Jie Lu1, 2, Guangping Dai, Yasu Egi, Shuning Huang, Eng H Lo, Young Ro Kim 1 Massachusetts General Hospital , Charlestown, Massachusetts, USA; 2Xuanwu Hopital,Capital Medical University, Beijing, People's Republic of China Cerebrovascular responses during the inhalation of oxygen (O2) and carbon dioxide (CO2) in normal brain tissue are important for investigating exogenous regulation of cerebral hemodynamics. In this study, we characterized various cerebrovascular parameters in response to 100% O2 and 5% CO2 gas inhalation using multiple MRI techniques in rat models. In particular, using both gradient and spin echo- echo planar imaging (GE-EPI and SE-EPI) pulse sequences, we quantified blood oxygenation level dependence (BOLD), cerebral blood flow (CBF), and blood volume (CBV) changes with the use of arterial spin labeling and an intravascular contrast agent. Otto A. Sanchez1, Elizabeth A. Louie1, Bruce M. Damon1 1Vanderbilt University, Nashville, Tennessee, USA The repeatability of a dual gradient-recalled echo (GRE) MRI sequence that measures change in blood volume and oxygenation in the microvasculature of skeletal muscle is presented. Subjects performed 3 types of isometric dorsiflexion contractions and dual GRE EPI dynamic scans were acquired before, during and after contractions. Signal intensities (δSI) were plotted as a function of time and post-contraction parameters �SI, time to half peak and time to peak were determined. Intraclass correlation coefficient was determined to compare the between-day variations. The method is reliable although the repeatability method depends on the parameter evaluated, the muscle, and the TE. 2345. Mapping Hypercapnic Cerebral Vasoreactivity Using Bold FMRI: Regional Variation and Correlations Uma S. Yezhuvath1, Kelly Lewis-Amezcua1, Rani K. Varghese1, Hanzhang Lu1 1UT Southwestern Medical Center, Dallas, USA BOLD fMRI with 5% CO2 breathing was used to study cerebrovascular reactivity (CVR) in different brain regions in young, healthy controls. Physiological parameters were measured during scan. CVR was computed by regressing measured EtCO2 against BOLD signal. Results indicate heterogeneous vascular reserve across regions (cerebellum had maximal and insula minimal vascular responsiveness). Additionally, we found that regional CVRs are correlated and there appears to be a global factor specific to each person. Based on this, we recommend use of relative CVR instead of absolute CVR in group comparisons. No correlations were found between CVR and measured physiologic parameters or age. 2346. Fast, Calibrated, Quantitative Functional MRI for Single Repeats Using HyperoxiaDaniel Bulte1, Peter Jezzard1 1University of Oxford, Oxford, UK Increased fractions of inspired oxygen enables the modelling of relative changes in CMRO2, as well as estimating the baseline cerebral blood volume. Subjects were imaged using an interleaved, BOLD/ASL sequence while they performed a 2-level graded motor task followed by two short hyperoxic epochs. These data were used to estimate the theoretical maximal BOLD signal and calculate the relative regional increase in CMRO2 for each task. The hyperoxia BOLD data were also used to produce baseline CBV maps. From the measured relative CBF changes, estimates of quantitative blood volume changes in the motor regions with activation were calculated. Molly Gallogly Bright1, 2, Silvina G. Horovitz1, Peter Jezzard2, Jeff H. Duyn1 1National Institutes of Health, Bethesda, Maryland, USA; 2University of Oxford, Oxford, UK Measurements of regional cerebrovascular reactivity to changes in arterial gas tensions using functional MRI have been shown to accurately indicate pathological abnormalities associated with disease. However, these studies have traditionally focused on the magnitude of the BOLD signal change during prolonged breath-holding or carbogen inspiration. Here we show that simple deep breaths, a task suitable for clinical use, cause transient BOLD signal decreases comparable in magnitude to hypercapnia studies. A new fitting technique using linear optimized basis functions allows for whole-brain voxelwise characterization of both magnitude and relative timing of the cerebrovascular response to the breathing task. Bradley J. MacIntosh1, Kyle TS Pattinson1, Daniel Gallichan1, Imran Ahmad1, Karla L. Miller1, David Feinberg2, Peter Jezzard1 1FMRIB Centre, Oxford, UK; 2Advanced MRI Technologies LLC, Sebastopol, California , USA Abstract: Adequate Cerebral Blood Flow is necessary to maintain healthy brain tissue. Arterial Spin Labeling is a technique that provides perfusion-weighted images without the use of a contrast agent. ASL can be used to investigate stroke and cerebrovascular disease, but to date there has been only a limited amount of clinical translation. Ideally, we would want our ASL acquisition to produce whole brain quantitation perfusion estimates, in a reasonable length of time (< than 6 minutes). The present study attempts to address these issues, using a hypercapnia model to test the sensitivity of a 3D GRASE-ASL sequence. Andrea Kassner1, 2, Julien Poublanc3, Jeff Winter2, David Mikulis3, Adrian Crawley3 1University of Toronto, Toronto, Canada; 2The Hospital for Sick Children, Toronto, Canada; 3The Toronto Western Hospital of the UHN, Toronto, Canada Sensitivity of spiral EPI and cartesian EPI with parallel imaging (ASSET) techniques to susceptibility effects, specifically signal dropout, was quantitatively assessed using hypercapnia as a global BOLD stimulus. Five healthy volunteers were imaged at both 1.5 T and 3.0 T, with both EPI techniques. A significant increase in signal dropout existed for spiral EPI compared with ASSET EPI, however, only at 3.0 T. No differences in the magnitude or variance of the BOLD signal response to hypercapnia existed. Results suggest that using ASSET EPI may improve the quality and interpretation of functional MRI in regions of susceptibility differences. |
Functional Brain Networks & Connectivity Hall D Wednesday 13:30-15:30 Toshiharu Nakai1, Epifanio Bagarinao2, Yoshio Tanaka2, Chikako Nakai3, Masafumi Hiraoka1, Kayako Matsuo1 1National Center for Geriatrics & Gerontology, Ohbu, Japan; 2National Institute of Advanced Industrial Science and Technology, Tsukuba, Japan; 3Toyohashi Sozo University, Toyohashi, Japan Dynamic monitoring of t-statistics is potentially useful to monitor and characterize the ongoing the brain activities during switching of several task conditions. The correlation coefficients (CCs) of the t-statistics response function among the motor areas were obtained and schematically labeled on a wire-frame map (SCM). The dominance of the left higher motor areas during sequential movements of the left fingers could be systematically visualized. SCM is a data driven analysis without pre-defined restriction. This method will be useful to conceptually organize the dynamics of functional network including several brain areas.
2398.
Dynamics for Functional Connectivity in 24-Hour Default-Mode
Networks
Hae-Jeong Park1, 2, Bumhee Park1, 2, Joong-il Kim1, 2 1Yonsei University College of Medicine, Seoul, Republic of Korea; 2Yonsei University College of Medicin, Seoul, Republic of Korea The study evaluated the intra-individual variations of connectivity measures of default-mode network using fMRI according to brain states by measuring data during 24 hours, every 3 hours. The results showed that no significant difference between connectivity of different acquisitions times, which implies the long-term representation of the default-mode network. The fMRI of the default-mode network might be used to evaluate connectivity changes between groups. 2399. Deriving Unbiased Seed-Based Functional Connectivity Maps with an Iteration Strategy Wen-Jun Pan1, Hao Lei1 1Wuhan Institute of Physics and Mathematics, Chinese Academy of Science, Wuhan, People's Republic of China Seed-based cross-correlation analysis (sCCA) has been widely used in functional connectivity MRI (fcMRI) studies. However, it is known that, with this method, the detected network may be biased by empirical and often subjective selection of the initial seed. To address this problem, here we propose a novel approach, seed-iteration cross-correlation analysis (siCCA), to examine the brain default network at rest. The method was shown to give results that are not dependent on the choice of the initial seed. 2400. High Spatio-Temporal Resolution Imaging of Resting State Fluctuations at 7T Natalia Petridou1, Andreas Schäfer1, Penny Gowland1, Richard Bowtell1 1University of Nottingham, Nottingham, UK Studies have shown that physiological fluctuations at rest consist of BOLD and non-BOLD related noise. Here we investigate the spatiotemporal nature of these fluctuations at 7T, using a high spatial resolution where physiological and thermal noise are similar in magnitude, multiple TEs, and a short TR as to sample physiological fluctuations adequately. Both magnitude and phase were examined. At so high resolution, magnitude time series fluctuations consisted solely of BOLD related noise at frequencies <0.13Hz. Phase time series fluctuations increased with TE and were dominated by respiration, however, BOLD-like noise at frequencies <0.13Hz was observed.
2401.
The Effect of Respiration Variations on Independent
Component Analysis of Resting State Functional Rasmus Matthias Birn1, Kevin Murphy1, Peter Anthony Bandettini1 1National Institute of Mental Health, Bethesda, Maryland, USA MRI signal changes induced by variations in breathing depth and rate occur at low frequencies (<0.1Hz) and overlap with the default mode network, causing problems for functional connectivity analysis. In this study we investigate the impact of respiration changes on independent component analysis (ICA). We find that ICA can separate respiration-related from default mode network regions, but an independent measure of respiration is important to distinguish artifactual from neuronally relevant components. Furthermore, time series associated with the default mode network component are still significantly correlated with variations in respiration volume per time. 2402. The Effect of Motion on FMRI BOLD Resting State Low-Frequency Fluctuation Jaroslav Hlinka1, Charilaos Alexakis1, Paul S. Morgan1, Dorothee P. Auer1 1University of Nottingham, Nottingham, UK Studies in spontaneous brain activity consistently reveal coherent low-frequency (0.01-0.1 Hz) fluctuations in areas spatially organized into functionally relevant networks (Resting State Networks). The power of these fluctuations has been proposed to reflect altered brain state, while the underlying mechanism is not understood. In our study, we have observed significant increase of low frequency fluctuation power in midazolam sedation in several ROIs including the visual cortex and callosal white matter. An in-depth analysis shows that the apparent increase of LFF power is attributable to the observed increase of motion accompanying sedation. 2403. Reliability of the Resting State Fluctuation Amplitude as a Hemodynamic Scaling Parameter Nirvish S. Shah1, Sridhar S. Kannurpatti1, Bharat B. Biswal1 1UMDNJ, Newark, New Jersey, USA BOLD fMRI responses to a neural task can vary across brain regions, sessions and subjects depending on the underlying cerebro-vasculature and basal conditions. Though hypercapnia, induced through breath holding and CO2 inhalation, has been used for vascular reactivity studies, it is disadvantaged by various concerns such as patient compliance, evoked neural activity and requirement of additional equipment for hemodynamic scaling. Recently, we have observed high correlation between the resting state fluctuation amplitude (RSFA), breath hold response amplitude and CO2 response amplitude and demonstrated the use of RSFA for hemodynamic scaling. In this study, we have used the bootstrap resampling method to validate the use of RSFA as an alternative to hypercapnia-based scaling. Reliability of the correlation between RSFA, breath hold response amplitude and CO2 response amplitude was evaluated by bootstrapping the BOLD-fMRI time series. Stability of the correlation between RSFA, breath hold response amplitude and CO2 response amplitude was evaluated by bootstrapping varying lengths of the BOLD-fMRI time series. 2404. Functional Connectivity of the Motor Cortex: Test-Retest Reliability Neha Ravindra Shah1, 2, Steven R. Roys1, Jiachen Zhuo1, Joel M. Morris2, Rao P. Gullapalli1 1University of Maryland, School of Medicine, Baltimore, Maryland, USA; 2University of Maryland, Baltimore County, Baltimore, Maryland, USA This study examines the spatial consistency of functional connectivity within the primary motor cortex of the resting brain within and across sessions. Seven subjects underwent two functional and three rest scans per session, during three separate sessions. Functional connectivity was determined using seven different sets of seed voxels from the motor cortex as identified from the motor activation scan. The functional connectivity maps obtained before and after filtering of physiological noise were found to be spatially consistent, within and across sessions. Although not significant, physiological filtering removes some of the variability in the results. 2405. Investigating Reproducibility of Working Memory Effective Connectivity Nia Goulden1, Shane McKie1, John Francis William Deakin1, Rebecca Elliott1 1University of Manchester, Manchester, UK Studies carrying out effective connectivity analyses have a wide range of volunteer numbers. A typical number of volunteers for fMRI and effective connectivity analyses is 11 per group. Working memory tasks are likely to have variable results due to different volunteer abilities and strategies for carrying out the task, so 11 may not be a representative number. The aim of this work is to find out whether 11 volunteers is adequate for a study by comparing the results for independent groups of 11 volunteers. 2406. Gender Difference of Resting Rhythms Detected by BOLD-Based FMRI Tzu-Chen Yeh1, Wen-Jui Kuo2, Chou-Ming Cheng1, Jen-Chuen Hsieh1, Low-Tone Ho1 1Taipei Veterans General Hospital, Taipei, Taiwan; 2National Yang-Ming University, Taipei, Taiwan The characteristic brain default network (tripod component) has been detected during resting by BOLD-based 3T fMRI studies. By identifying the default network using informax independent component analysis and automatic sorting using spatial template, resting BOLD-based fMRI studies of 40 normal subjects (gender- and age-matched) demonstrated the gender difference of resting default networks by group random-effect analyses. Men had relatively more extensive default network than women in bilateral posterior cingulate cortices (Brodmann Area 31).
2407.
Functional Network Connectivity in Semantic Object Recall
Task Using Independent Component Analysis Kanchana Jagannathan1, Michal Assaf1, 2, Vince D. Calhoun, 13, Godfrey D. Pearlson1, 2 1Institute of Living, Hartford, USA; 2Yale University, New Haven, USA; 3The MIND Institute , Albuquerque, USA Using features to recall an object is a specific operation in semantic memory processing. The regions involved in the object-recall task were not significantly associated with the performance. We therefore hypothesized that temporal based dependency among independent component analyses ( ICA ) time courses would explain differences in neuronal network connections between good and poor performing healthy individuals. Consistent with our hypothesis, good performers showed higher correlation between language network, right hemisphere network and thalamus.
2408.
Correction of the Low-Frequency Physiological Noise in the
Resting State BOLD FMRI - Effect on the ICA Tuomo Henrik Starck1, Jukka Remes1, 2, Juha Nikkinen1, Osmo Tervonen1, Vesa Kiviniemi1 1Oulu University Hospital, Oulu, Finland; 2Oulu University, Oulu, Finland Lagged estimates of the low-frequency respiratory and cardiac noise fluctuations and RETROICOR type of physiological correction have been shown to explain notable signal variance in the regression analysis of the resting state networks. Default-mode network spatially overlaps with regions significantly affected by physiological low-frequency fluctuations and they occur at similar frequencies, but it is not known how well ICA can separate noise processes from the default-mode network or other resting-state networks. This study showed that in spatial default-mode analysis, ICA accuracy does not improve significantly after physiological noise correction. In time and frequency domain studies correction may be beneficial Stephane, R Simon1, Mohamed, L Seghier, 2, François Lazeyras, Jean-Marie Annoni3, Asaid Khateb3 1Bio-Medical Imaging Center, Geneva University Hospitals, Geneva, Switzerland; 2Wellcome Trust Centre for Neuroimaging, Institute of Neurology, UCL, London, UK; 3Neuropsychology Unit, Geneva University Hospitals, Geneva, Switzerland An original functional ROI-based approach was used to compute hemispheric laterality index obtained from four different linguistic tasks, in a group of healthy volunteers. Different ROIs were derived from the SPM conjunction group analysis based on all the tasks. The present data demonstrate the differential potential of the tasks to elicit clear regional hemispheric dominance. We conclude that the combination of several linguistic tasks, as well as the use of such a ROI-based approach is of great interest for pre-surgical mapping of language related areas. 2410. Motor Sequence Learning: A Functional Integration Study in FMRI David Coynel1, 2, Vincent Perlbarg1, 2, Guillaume Marrelec1, 2, Mélanie Pélégrini-Issac1, 2, Stéphane Lehéricy2, 3, Habib Benali1, 2 1Inserm and Université Pierre et Marie Curie, Faculté de médecine Pitié-Salpêtrière, Paris, France; 2IFR49 - Institut d'Imagerie Neurofonctionnelle, Gif-Sur-Yvette, France; 3Université Pierre et Marie Curie, Paris, France We tested a model of motor sequence learning, according to which the representation of a motor sequence within brain areas shifts from a motor associative to a sensorimotor loop during the learning process. We used fMRI data to investigate the evolution of the motor network during learning. The network involved in the learning task was identified using spatial independent component analysis and hierarchical clustering. Then, the functional interactions within this network were quantified using a hierarchical integration measure. We demonstrated that learning induces changes in both the spatial structure of the motor network and its functional connectivity pattern. 2411. Long-Term Motor Training Induced Changes in Resting State Brain Liangsuo Ma1, Binquan Wang, Jinhu Xiong2 1University of Iowa, Iowa City, USA; 2University of Iowa,, Iowa City, USA Functional activation often only reflects differentiated involvement and temporal relationship of brain regions compared between the task performance and the control states on a transient level. However, lasting changes in metabolic and hemodynamic physiology and synaptic plasticity induced by motor skill learning common to both states have not been elucidated. To address this issue, we explored long-term motor training induced neuronal and physiological changes in normal human subjects during task performance and resting state using MRI and PET. We have observed long-term motor training induced changes in both rCBF, which may reflect more fundamental changes in the brain. 2412. Modulation of Coherent BOLD Fluctuation in Human Visual Cortex with Continuous Brain Stimulation Xiao Liu1, 2, Xiao-Hong Zhu1, Wei Chen1 1University of Minnesota, Minneapolis, Minnesota, USA The present study is to investigate the modulation of coherent BOLD fluctuations within human visual system under two conditions with continuous stimuli. We found that (i) under continuously-stimulated conditions, the activated brain region formed its own coherent network while the other inactivated regions of visual system still remained their temporal correlation as in resting-state; (ii) the coherent network at resting-state covered the widest areas and included some regions not activated by full-field visual stimulus; and (iii) no significant differences in terms of coherence strength were found among those two stimulation conditions and resting-state.
2413.
Studying Coherent BOLD Fluctuation and Functional
Connectivity of Cortico-Thalamic Visual Network in Xiao Liu1, 2, Xiao-Hong Zhu1, Yi Zhang1, Wei Chen1 1University of Minnesota, Minneapolis, Minnesota, USA The present study is to examine the coherent BOLD fluctuation and functional connectivity at resting-state in the anesthetized cat brain with the focus on the visual system. We found that the spontaneous BOLD signals detected in the LGN, PMLS, the areas 18 and 19 were temporally correlated with that in the area 17 (V1). We also observed that such coherence was much stronger at the light anesthesia condition than mild anesthesia condition. These findings not only prove the existence of coherent BOLD fluctuation in the cortico-thalamic visual pathway in the cat brain, but also provided clear evidence for the neural origin of coherent BOLD fluctuation. 2414. Regional Spontaneous Brain Activity in Male ADHD Revealed by Amplitude of Low Frequency Fluctuation Hong Yang1, 2, Qian-Qian Li, Min-Ming Zhang2, Xiang-Yu Long3, Lan-Ting Guo, Chao-Zhe Zhu3, Xiang-Ping Zhou1, Yu-Feng Zang3, Qi-Yong Gong1 1Department of Radiology, West China Hospital of Sichuan University, Chengdu, People's Republic of China; 2First Affiliated Hospital of College of Medical Science, Zhejiang University, Hangzhou, People's Republic of China; 3Beijing Normal University, Beijing, People's Republic of China Using ALFF, to investigate the spontaneous activity in boys with ADHD. Seventeen boys with ADHD were compared with 17 controls. Data were acquired using 3T MRI system and data processing was performed using AFNI software. Significant decreased ALFF was found in the right inferior frontal gyrus, left supplementary motor areas and left anterior cingulate gyrus meanwhile significant increased ALFF was found in the bilateral superior frontal gyrus, left precentral gyrus and right ACG in ADHD patients compared with controls. The decreased spontaneous prefrontal activity is in line with the decreased response control function in prefrontal lobe of boys with ADHD. 2415. Modulation of Functional Connectivity in the Resting Brain by Typical and Atypical Antipsychotic Drugs Yazhuo Kong1, Shane McKie1, Anna Murphy1, Tirthankar Mukherjee1, Serdar Dursun1, Bill Deakin1, Steve Williams1 1University of Manchester, Manchester, UK We investigated the default-mode network in normal volunteers using resting-state BOLD fMRI. 3 groups of subjects all had baseline scans, before receiving either placebo or the anti-psychotics risperidone (D2 antagonist/5HT2A antagonist) or aripiprazole (partial agonist). 3½ hours later they were rescanned. Group ICA was used for the identification of the default mode network nodes. Connectivity maps from anterior and posterior seed regions were generated for all conditions. Pre-drug and placebo maps were consistent with other descriptions of the default-mode network, but there was a significant decrease of the anterior to posterior connection for both drugs compared to pre-drug and placebo.
2416.
Increased Resting State Dorsal Lateral Prefrontal and Medial
Temporal FMRI Activation in Early Xiaowei Song1, Kenneth Rockwood2, Ryan D'Arcy1, Alma Major3, John Fisk, Sultan Darvesh, Elissa Asp2, James Rioux1, Chris Bowen1, Steven Beyea1 1National Research Council Canada, Halifax, Canada; 2QEII Health Sciences Centre, Halifax, Canada; 3Dalhousie University, Halifax, Canada This study investigates the characteristic fMRI fluctuation that exists at rest and how this differs between early AD patients and cognitively healthy adults, particularly in the dorsal lateral prefrontal cortex and the medial temporal lobe. The study suggests that there exist common patterns of neural activity fluctuation during rest. Patients with mild AD have increased resting state neural activity in both regions, suggesting the neural compensation effect in the early stages of AD progression. This study suggests that care needs to be taken with inferring cognitive task fMRI results when referenced to the resting state. 2417. Spontaneous BOLD Signal Changes Mimic Task-Related Activation in Anesthetized Rats Matthew Magnuson1, Shella Keilholz1 1Emory University/Georgia Institute of Technology, Atlanta, Georgia, USA Electrical forepaw stimulation is a commonly used paradigm in rodent functional MRI. In recent experiments signal changes similar in magnitude and duration to those associated with stimulation were observed in scans where no stimulation was administered. This phenomenon was investigated further by comparing two groups of rats on two separate imaging systems under two diffent anesthsias; alpha-chloralose and medetomidine. 2418. Spatial Specificity of Functional Connectivity Maps for Different Frequency Bands in Rats Waqas Majeed1, Matthew Magnuson1, Shella Keilholz1 1Georgia Institute of Technology / Emory University, Atlanta, Georgia, USA This abstract presents some initial resluts which suggest that frequncy band selection for filtering may impact spatial specificity and seed-location dependence of the connectivity maps. 2419. Effect of Different Frequency Bands on Functional Connectivity in the Motor Cortex Network Yin Xu1, Shi-Jiang Li1 1Medical College of Wisconsin, Milwaukee, Wisconsin, USA Spontaneous low-frequency fluctuations (SLF) detected by functional MRI (fMRI) techniques have been utilized to investigate functional connectivity in human brain networks. The effective frequency band of the SLF signal is usually defined as 0~0.1Hz. However, few studies have investigated SLF signal at different narrower frequency bands. In the present study, we investigate the functional connectivity of SLF in the primary motor model and determine the effect of different frequency bands of the SLF signal on functional connectivity maps. 2420. Frequency Specificity of Functional Connectivity in Brain Networks Changwei Wesley Wu1, 2, Hong Gu1, Hanbing Lu1, Jyh-Horng Chen2, Elliot Stein1, Yihong Yang1 1National Institute on Drug Abuse, Baltimore, Maryland, USA; 2National Taiwan University, Taipei, Taiwan The synchronous low-frequency fluctuations in the resting-state fMRI signal have been analyzed in the frequency domain and been evaluated through four brain networks (sensorimotor, visual, default mode, and limbic systems). The results show that strongest connections in these networks falls upon 0.01-0.06 Hz and discrepancies in the spectrograms subsist between brain networks. Our data also showed that long-distance connections seem to be more frequency specific, whereas short-distance connections are distributed in a wider frequency range. Such frequency-specific features in the resting state fMRI signal imply that brain networks have different spectral characteristics at rest. |
fMRI: Signal Variation & Noise Hall D Wednesday 13:30-15:30 Contrast to Noise Ratio BOLD Model Pierre-Francois Van de Moortele1, Kamil Ugurbil, 12, Stephane Lehericy, 13 1University of Minnesota, Minneapolis, Minnesota, USA; 2Max Planck Institute for Biological Cybernetics, Tuebingen, Germany; 3CENIR, INSERM U610, University Paris 6, Paris, France It is generally accepted that BOLD activation peaks at TE=T2*. Thus, matching TE with T2* should improve activation detection. This concept, however, has been challenged by studies where similar fMRI activation detection was obtained at different TE's and it has been shown that a large noise fraction in fMRI series has physiological origins. We propose a new model for BOLD Contrast to Noise Ratio (CNR) which accounts for heterogeneous sources of noise. This model predicts that BOLD CNR varies slowly with TE and that optimal TE can be longer than T2*. Those findings may significantly impact BOLD fMRI acquisition parameters.
2465.
Physiological Noise in Gradient Echo and Spin Echo EPI Using
Multi-Channel Array Coils
Christina Triantafyllou1, 2, Jonathan R. Polimeni2, Lawrence L. Wald2 1Athinoula A. Martinos Imaging Center, McGovern Institute for Brain Research, MIT,, Cambridge, Massachusetts, USA; 2Athinoula A. Martinos Center for Biomedical Imaging, Department of Radiology, MGH , Charlestown, Massachusetts, USA; We evaluate the physiological noise in both GRE and SE single-shot EPI time series by comparing temporal to thermal SNR (tSNR and SNR0). We modulate the SNR0 by using different flip angles as well as single channel and multiple channel coils. Our findings suggest that the relationship between tSNR and SNR0 can be well parameterized by the Krueger model for both SE and GRE time-series and for array data, with 20% difference in lambda for GRE and SE. This suggests that when SNR0 is modulated in this way, the SE physiological noise can be modeled as proportional to the signal. 2466. Retrospective Estimation and Correction of Long-Term Physiological Effects in FMRI Jaemin Shin1, Richard Cameron Craddock2, Xiaoping Hu1 1Georgia Tech / Emory Univ, Atlanta, Georgia, USA; 2Georgia Tech, Atlanta, Georgia, USA Low-frequency fluctuations in the cardiac rate and the depth and the rate of breathing have been identified as a significant source of low-frequency (<0.1 Hz) fluctuations in the BOLD signal. In this study, we estimate these low-frequency physiological fluctuations, pixel-by-pixel from BOLD timecourse by estimating the impulse responses of physiological processes, and correct these signal changes. The correlations between the resting-state fMRI timecourses and 7, and 12 sec delayed respiration volume changes are substantially reduced by this approach. In the grey matter, the fraction of standard deviation (SD) in the original timecourse accounted for by physiological fluctuations was 28.09% for respiration and 12.49% for heart beat. These results indicate that our method is effective for removing physiology-induced low frequency fluctuation, which is significant in fMRI data. 2467. Efficient Correction for Artificial Signal Fluctuations in Resting-State FMRI-Data Andreas Weissenbacher1, Christian Windischberger1, Rupert Lanzenberger, Ewald Moser1 1Medical University of Vienna, Vienna, Austria Spontaneous low frequency fluctuations in the brain that are the basis for functional connectivity analysis are overlaid with various other signals originating from residual motion artifacts, respiration and cardiac action. Here we suggest an automated correction approach using multiple linear regression methodology to compensate the effects of non-neural-related signal changes, including global, ventricle and white matter signal, as well as residual motion. Based on data from 40 healthy subjects we show that the proposed method greatly improves the specificity of functional connectivity analysis results. Erik B. Beall1, Mark J. Lowe1 1Cleveland Clinic, Cleveland, Ohio, USA Respiratory variation is believed to produce significant variation in fMRI timeseries data, and in particular affect resting state connectivity analyses. The RVT (respiratory variation over time) correction is a proposed method of dealing with the problem. However, the RVT delay searching process may produce incorrect lags in the presence of non-white noise. We show detected RVT delay in simulation that are incompatible with injected RVT delay and delays incompatible with the physiologic theory of RVT in fMRI data. Suk-tak Chan1, 2, Timothy G. Reese2, Kenneth K. Kwong2 1The Hong Kong Polytechnic University, Hunghom, Hong Kong; 2Massachusetts General Hospital, Charlestown, Massachusetts, USA Cardiac and respiratory pulsations are the bane of fMRI brain studies. We extended the Simultaneous Echo Refocusing (SER) EPI sequence from its original 2 slice single-shot acquisition into (SER-8) which can acquire 8 slices at one-shot. We demonstrated how SER-8 with a TR of 150ms can be used to remove physiological noise in brain activation. Cardiac gated imaging, collected all at the same cardiac phase, is another approach to reduce cardiac noise and respiratory noise. Our findings showed that physiological noise can be effectively separated from BOLD images acquired with either short TR or T1-corrected cardiac gated SER-8 sequence. 2470. Adaptive FID-Navigators for Respiration Monitoring in Multi-Slice FMRI Applications Tobias Kober1, 2, Wietske van der Zwaag2, 3, Jose Marques2, 3, Reto Meuli4, Rolf Gruetter2, 5, Gunnar Krueger1 1Siemens Medical Solutions-CIBM, Lausanne, Switzerland; 2Ecole Polytechnique Federale de Lausanne, Lausanne, Switzerland; 3University of Lausanne, Lausanne, Switzerland; 4CHUV, Lausanne, Switzerland; 5Universities of Lausanne and Geneva, Switzerland The sensitivity of task-based and resting state fMRI methods is often limited by respiratory and cardiac fluctuations masking the activation-induced BOLD response. Several correction strategies have been proposed, typically applied either during acquisition or post-processing. We propose a more comprehensive implementation that samples respiration-induced B0-changes using the intrinsic MR-signal. Extending previous implementations, respiration is monitored continuously in multi-slice experiments. This allows simultaneous correction of (a) slice dependent global B0-fluctuations during image reconstruction and (b) residual effects through additional regressor for the fMRI analysis. Results proof the feasibility to acquire fully automatic and user-independent fMRI data of highest quality. 2471. Noise Reduction in FMRI GRE EPI Acquisitions Using Real-Time Navigators Daniel Nicolas Splitthoff1, Oliver Speck, 1, Maxim Zaitsev1 1University Hospital Freiburg, Freiburg, Germany As is well known, GRE EPI imaging suffers heavily from susceptibility artifacts. When these effects vary with time, this can influence the results of functional imaging experiments. We here present the first results of the application of a recently introduced real-time prospective shim correction method in a volunteer measurement. It is demonstrated that the method decreases significantly the energy in the spectra of unexplained noise in an fMRI experiment, especially in the lower frequency range, corresponding to artifacts induced by motion. 2472. Minimizing the Effect of Body Motion on EPI Time Series by Digital Tuning Frank Q. Ye1, Qing San Xiang2 1National Institutes of Mental Health, NIH, Bethesda, Maryland, USA; 2University of British Columbia, Vancouver, Canada In fMRI of non-human primates, subjects’ body motion poses a difficult technical challenge. A simple linear model is proposed to approximate the magnet field change induced by the body motion, and a simple digital tuning procedure during image reconstruction is used to restore the motion-affected EPI time series. The method is demonstrated on both phantom and monkey scans. 2473. Preprocessing Pipeline Considerations to Compensate for Paradigm-Related Subject Movement Robert L. Barry1, 2, Joy M. Williams1, L. Martyn Klassen1, Jody C. Culham, Ravi S. Menon1 1Robarts Research Institute, The University of Western Ontario, London, Canada Functional MRI is sensitive to spatial and temporal perturbations in main magnetic field homogeneity within the brain due to movement outside the imaging field of view. To simulate paradigm-related subject movement, a phantom arm was constructed and moved (creating 6 Hz peak-to-peak spatially varying field inhomogeneities in a mid-axial slice) while subjects performed visual activation experiments. The inclusion of a complex phase regressor after navigator correction resulted in a 10.1% increase in average t-statistics, demonstrating the complementary nature of these techniques to further reduce geometric distortions and precondition data to make them more amenable to the benefits of spatial smoothing. Chase R. Figley1, Patrick W. Stroman1 1Queen's University, Kingston, Canada Recent reports suggest that the sensitivity and specificity of functional magnetic resonance imaging of the spinal cord (spinal fMRI) are reduced by motion of the cord itself. However, because this motion is persistent, it has been proposed that modeling the motion-related confounds and including them in a general linear model (GLM) approach is the most effective and efficient way to increase the sensitivity of spinal fMRI to neuronal function. Therefore, we have developed such an approach, using a single-stage GLM, and quantified the improvements in sensitivity and specificity (compared to uncompensated methods) afforded by this new spinal fMRI analysis method.
2475.
Vascular Signal Contributions in MR-Encephalography: Direct
Observation of Intra- And Extravascular Benjamin Zahneisen1, Thimo Grotz1, Jürgen Hennig1 1University Hospital Freiburg, Freiburg, Germany MR-Encephalography is an extremely fast technique to monitor physiological changes by use of simultaneous readout with multiple small RF-coils. A sampling rate far below breathing and ECG frequencies allows for a very accurate determination of the BOLD response. Aim of this work was to separate blood flow related contributions of the BOLD signal from those of tissue and to investigate the influence of blood flow on the physiological noise caused by ECG pulsatility. It is demonstrated that the suppression of intravascular signal reduces the average BOLD amplitude by approximately 40%. Our results show that significant ECG pulsatility persists even when intravascular signal contributions are removed. 2476. Assessment of Signal Variations in EPI, Induced by Breathing and Hardware Instabilities Daniel Nicolas Splitthoff1, Jürgen Hennig1, Maxim Zaitsev1 1University Hospital Freiburg, Freiburg, Germany Many approaches have been suggested for correcting for field inhomogeneities in GRE EPI measurements. To our knowledge, however, none of them has investigated on the typical effects of inhomogeneities induced by physiological noise and hardware instabilities. We here determine the order of typical field changes at 3T in phantom and volunteer measurements over a time frame of 5min and present the means to assess their effects. It will be shown, that for bigger matrix sizes the effect of breathing might not be negligible for fMRI measurements. 2477. The Effect of Physiological Noise on FMRI Phase Time Series Gisela E. Hagberg1, Marta Bianciardi2, Valentina Brainovich1, Antonio Cassarà3, Bruno Maraviglia3 1Santa Lucia Foundation, Rome, Italy; 2NIH, Bethesda, Maryland, USA; 3University “La Sapienza”, Rome, Italy Recently the possibility to use both magnitude and phase time-series for the statistical evaluation of fMRI studies has been proposed, with the prospective of increasing both statistical power and spatial specificity of the activation maps. In order to fully take advantage of the additional information in phase images, knowledge regarding available phase stability is needed. Here we propose and investigate three models for the phase stability at 3T, each model being dependent on the relation between the effect of physiological noise on magnitude and phase variations. |
Ingestion, Drugs & Pain: Effects or fMRI Signed Hall D Wednesday 13:30-15:30 a Spinal FMRI Investigation Patrick W. Stroman1, 2, Chase R. Figley1, Niousha Foad Ghazni1, Natalie Kozyrev1 1Queen's University, Kingston, Canada Neuronal activity in intermediate and ventral regions of the human spinal cord in response to thermal stimulation was investigated. These areas of activity are not known to be associated with sensory responses, yet they are consistently observed with fMRI of the human spinal cord. Thermal stimuli were systematically varied between 15 C and 29 C, as was the order of experiments, during fMRI of the spinal cord and brainstem. Activity in these cervical spinal cord regions depends on the stimulus temperature, and is correlated with activity in brainstem regions involved with the descending modulation of thermal sensation and pain responses.
2509.
Gender Differences in Voluntary Micturition Control - An FMRI
Study
Juergen Baudewig1, Sandra Seseke1, Kai Kallenberg1, Rolf H. Ringert1, Florian Seseke2, Pe Dechent1 1Georg-August-University Goettingen, Goettingen, Germany; 2Martha-Maria Hospital, Halle, Germany As bladder dysfunction affects patients of both sexes, we wanted to compare the activation of micturition-related brain areas in healthy men and women directly to better prepare the use of fMRI in patients. Previously, we characterized the cerebral network involved and reported for the first time reliable BOLD activations in brainstem in women. Here, we focused on healthy men and were able to identify the same cerebral network including well-localized activations in the periaquaeductal grey as well as in the continence and micturition centre in the pons. In general, the network revealed stronger activation in men compared to women. Abdelmalek Benattayallah1, Kate Janse Van Rensburg, Tim Hodgson2, Jon Fulford, Adrian Taylor 1University of Exeter, Exter, UK; 2University of Exeter, UK fMRI studies were undertaken comparing activations areas in somkers with and without exercise, prior to a protocol presenting the images of smoking. Reduction in activation of the orbito-frontal cortex, at the same time as a reduced self-reported craving, may suggest that exercise has the ability to reduce the perceived incentive salience both of the drug and the craving for the drug This is the first study to explore neurobiological mechanisms for how exercise acutely reduces cigarette cravings. 2511. Sildenafil and the Response to Functional and Hypercapnic Activation - A BOLD and VEP Study Christina Kruuse1, 2, Adam Espe Hansen1, Jes Olesen1, Martin Lauritzen1, Henrik BW Larsson1, Egill Rostrup1, 2 1Glostrup Hospital, University of Copenhagen, Denmark; 2Hvidovre Hospital, University of Copenhagen, Denmark Sildenafil (Viagra), a selective inhibitor of the cGMP-hydrolyzing phosphodiesterase5 enzyme (PDE5), can induce migraine in the absence of immediate changes in cerebral blood flow or cerebral artery diameter. Based on its mechanism of action (blocked clearance of second messenger) Sildenafil may be expected to enhance the local or global vascular response to external stimuli. Due to common second messenger systems enhanced neuronal excitability has also been hypothesised. These hypotheses were tested in the visual system by evaluating the BOLD and VEP responses to checkerboard stimulation during Sildenafil administra-tion. The global vascular response was evaluated in relation to CO2 inhalation. 2512. The Effect of Remifentanil Upon the Conscious Control of Breathing Kyle T S Pattinson1, Ricardo G. Governo2, Elizabeth C. Russell1, Bradley J. MacIntosh2, Imran Ahmad1, Stephen D. Mayhew1, Douglas R. Corfield, Irene Tracey1, Richard G. Wise 1University of Oxford, Oxford, UK; 2University of Oxford, UK We investigated the effect of remifentanil, a µ-opioid agonist, upon the neural correlates of a voluntary breathing task, a short breath hold. As remifentanil causes a rise in baseline CO2, we used a series of control measures to disentangle global vascular from neuronally mediated effects. We showed reductions in BOLD responses to the control measures, with significantly greater reductions in the BOLD response to the breath hold in the operculum and dorsolateral prefrontal cortex suggesting depression of neuronal activity. This methodology would be appropriate for investigation of other drugs with profound physiological effects that may confound interpretation of BOLD. 2513. Oral Fat Representation in Human Cerebral Cortex Sally Fathy Eldeghaidy1, Luca Marciani1, Tracey Hollowood2, Kay E. Head1, Joanne Hort2, Jason Stokes3, Francis McGlone4, Johanneke Busch5, Andy Taylor2, Robin C. Spiller1, Penny A. Gowland1, Susan T. Francis1 1University of Nottingham, Nottingham, UK; 2University of Nottingham, Leicestershire, UK; 3Unilever Coporate R&D, Colworth, UK; 4Unilever R&D, Wirral, UK; 5Unilever R&D, Vlaardingen, Netherlands We have mapped the cortical representation of oral fat in iso-viscous fat emulsions using a protocol close to the normal experience of consuming liquid fatty foods. Concurrent recording of EMG during fMRI acquisition allowed identification of the time of swallowing to improve detection of fMRI responses. We have identified for the first time cortical areas correlated with increasing fat concentration using a parametric analysis. Activated areas were (a) taste areas (frontal opercular and anterior insula) (b) areas associated with intraoral somatosensory textural attributes (postcentral gyrus, precentral gyrus, superior parietal cortex), (c) areas representing hedonic properties (dorsolateral prefrontal cortex, thalamus, amygdala). 2514. A High Fat Meal Has No Direct Effect on the Brain BOLD Response in Young Adults Natalie J. Stein1, 2, Joseph J. Carlson1, Sean C. Forbes, Ronald A. Meyer, Robert W. Wiseman1, Jill M. Slade1 1Michigan State University, East Lansing, Michigan, USA The hemodynamic response can be altered by vasoactive substances and can therefore affect the blood oxygen level-dependent (BOLD) response in functional MRI that is associated with neuronal activity. Nutrient composition, e.g. a meal high in fat, has been shown to alter vascular function and may therefore also affect the brain BOLD response. In this study, adding fat to a meal did not affect the brain BOLD response in young, healthy adults. 2515. Expertise Modulates the Perception of Pain in Others Kun-Hsien Chou1, Ya-Wei Cheng, Jean Decety, I-Yun Chen, Ching-Po Lin 1Institute of Biomedical Engineering, National Yang-Ming University, Taipei, Taiwan Perceiving the pain of others activates a large part of the pain matrix in the observer. Since this shared neural representation may lead to empathy or personal distress, regulatory mechanisms must operate in people who inflict painful procedures in their practice with patient populations in order to prevent their distress from impairing their ability to be of assistance. In this fMRI study, physicians who practice acupuncture were compared to naïve participants while observing animated visual stimuli depicting needles being inserted into different body parts. Results indicated that different brain network was activated between control and expert groups Daron Gordon Owen1, 2, Collin Clarke2, Vel Velayutham2, Sugantha Ganapathy2, Frank S. Prato1, 3, Keith S. St. Lawrence1, 2 1Lawson Health Research Institute, London, Canada; 2The University of Western Ontario, London, Canada; 3St. Joseph's Hospital, London, Canada Most imaging studies of pain have involved superficial stimuli of brief duration, whereas clinical pain is generally prolonged and arises from deep tissues. In this study, arterial spin labeling (ASL) was used to collect images of cerebral blood flow during a 15 min intramuscular infusion of hypertonic saline. The largest CBF increases and decreases occurred during initial 5 min of pain, and generally dissipated over time. This study demonstrated that ASL can be used to track changes in regional CBF that occur during a prolonged clinically relevant pain model. |
Hall D Wednesday 13:30-15:30 2540. Stretchable Coil Arrays Enable Knee Imaging at Varying Flexion Angles Jurek Antonin Nordmeyer-Massner1, Nicola De Zanche1, Klaas Paul Pruessmann1 1University and ETH Zurich, Zurich, Switzerland We present an 8-channel wearable knee coil array whose elements stretch to conform to the subject’s anatomy thus optimizing signal detection for varying knee sizes and flexion angles. The array is made by sewing stretchable conductors onto an elastic fabric that mimics form-fitting clothing. The stretchable design is demonstrated by high-resolution images of the knee at different flexion angles, which were obtained in immediate succession without any rearrangement of the coil array. Promising applications in musculoskeletal MRI include such variable-flexion imaging and time-resolved studies of joint dynamics. Both are not possible with conventional rigid knee coil arrays. 2541. Visualization of Collagen Crimp in Tendon Using High-Field MRI Kelsey Mairead Mountain1, Tadeusz Foniok, Jeff F. Dunn, John Robert Matyas 1University of Calgary, Calgary, Canada Crimp is a periodic wave-like pattern that is exhibited by a tendon when illuminated with polarized light. This phenomenon is a molecular property of the constituent collagen fibers, playing an important role in tendon physiology as the crimp pattern tends to disappear as strain increases. High-field MRI can detect collagen organization based on changes in T2 relaxation. In this study, collagen crimp was detected in tendon with MRI, and was validated using polarized light microscopy.
2542.
MRI Study of the Repair Tissue Following ACI in the Defect of
the Human Cartilage Specimens
Zuzana Majdisova1, 2, Pavol Szomolanyi1, 2, Vladimir Juras1, 2, Siegfried Trattnig1 1MR Centre - Highfield MR, Vienna, Austria; 2Institute of Measurement Science, Slovak Academy of Sciences, Bratislava, Slovakia Properties of the repair tissue following Autologous Chondrocyte Implantation in human cartilage specimens were evaluated using MR parameters T1, T2 and apparent diffusion coefficient (ADC). Specimens were implanted and nourished under the skin of nude mice for five and eight weeks. After removing from this model environment the MR microimaging was performed using a 3 Tesla scanner. T1, T2 and ADC values of the repair tissue were compared with the normal cartilage values. Lower relative DR1 and ADC values decreasing after longer maturation process predict higher amount of the proteoglycans and an improvement in the repair tissue quality. Burkhard Mädler1, 2, Derek C. Wilson2, 3 1Philips Medical Systems, Vancouver, Canada; 2University of British Columbia, Vancouver, Canada; 3Vancouver Coastal Health, Vancouver, Canada Quantitative T1-mapping from spoiled gradient recalled echo (SPGR) acquisition with varying flip angles FA allow fast and high resolution T1-quantification in a clinical feasible timeframe. The accuracy of the computed T1-maps from the measured SPGR signal depends strongly on the SNR, number of measurement points and critically from the correct knowledge of FA. Especially at higher field strength calibration of the true FA across the VOI becomes inevitable. Previously introduced methods require additional scan time and if the B1-information was not provided for already performed scans the estimated T1-values might differentiate significantly from the true values. We introduce a novel method of numerically estimating the error of the presumably known FA from the behaviour of the SPGR-signal by means of a orthogonal distance regression (ODR) method. The primer application is for T1-quantification in dGemric (delayed Gd-enhancement of MRI in cartilage) of the knee and hip joints. 2544. Effect of Magnetization Transfer on T1 and T2 Measurements of Articular Cartilage Atsuya Watanabe1, 2, Chris Boesch1, Takayuki Obata3, Toshiyuki Okubo4, Suzanne E. Anderson5 1University of Bern, Bern, Switzerland; 2Teikyo University Chiba Medical Center, Ichihara, Japan; 3National Institute of Radiological Sciences, Chiba, Japan; 4Department of Radiology, Ichihara, Switzerland; 5Department of Diagnostic, Interventional and Paediatric Radiology, Bern, Switzerland The aim of this study was to investigate inaccuracy of T1 and T2 measurements in cartilage when multislice acquisition is used. We especially focused on the effect of MT on T1 and T2 measurements, with consideration of the resulting feasibility of multislice acquisition for clinical evaluation.Based on the results of this study, we conclude that it will be difficult to adapt multislice acquisition for T1 measurement using FSE-IR methodology for clinical evaluation because of strong MT effects. On the other hand, multislice acquisition for T2 measurement using FSE is thought to be applicable for clinical evaluation. 2545. Measurement of Cartilage Water Content Using T1 and T2 Relaxation Time Measurements Jatta E. Kurkijärvi1, Mikko Nissi1, Jukka S. Jurvelin1, 2, Miika T. Nieminen3, 4 1University of Kuopio, Kuopio, Finland; 2Kuopio University Hospital, Kuopio, Finland; 3Oulu University Hospital, Oulu, Finland; 4University of Oulu, Oulu, Finland Increase in water content is characteristically related to degenerative changes of articular cartilage. The ability of T1 and T2 relaxation times to reflect water content was studied at 9.4T in bovine cartilage samples extracted from various joint surfaces. The dependence of MR parameters on proteoglycan content was also investigated. Strong correlations (r=0.8) were observed between water content and T1 or T2. Significant dependencies were observed between MR and compositional parameters, suggesting complex interactions between the constituents. Since T2 is also affected by collagen fibril orientation, T1 relaxation time is proposed as the primary surrogate MRI marker for cartilage water content. 2546. Automatic Segmentation of Articular Cartilage from MRI: A Multi-Contrast and Multi-Dimensional Approach Seungbum Koo1, Brian A. Hargreaves1, Thomas P. Andriacchi1, Garry E. Gold1 1Stanford University, Stanford, California , USA The purpose of this study was to evaluate segmentation of knee articular cartilage automatically from multiple sets of MR images using a SVM, a kernel-based machine learning algorithm. Multiple sets of MR images were taken for a human knee. The bones were automatically pre-prossed. Each pixel in the image could be regarded as a vector with multiple intensity values and a distance (from bone) value. The pixels in the even and odd numbered slices were used for training and testing of a SVM, respectively. The sensitivity and the specificity of the articular cartilage classification result was 93.8% and 99.3%, respectively. 2547. Accuracy of Building a Three-Dimensional Model of a Complex Articular Cartilage Defect from 1.5T MRI Seungbum Koo1, Brian A. Hargreaves1, Garry E. Gold1, Jason L. Dragoo1 1Stanford University, Stanford, California , USA The purpose of this study was to evaluate the accuracy of MRI for measuring a defect with complex shape involving the articular cartilage and underlying bone by comparing with a laser scanning technique. An articular cartilage defect in an intact cadaveric human knee was MR imaged and was made into a three-dimensional model. The knee was dissected and the defect was lasers scanned to create the gold standdard shape of the defect. Average deviation of the surface of the MRI model from the surface of the laser scan model was 0.4±0.4 (SD) mm. 2548. Quantification of Knee Cartilage in Vivo in the MMT Model of Osteoarthritis in Rats Using MRI Shantanu Dhamija1, Zhiyong Xie1, Susan Bove1, Ping Chiao1, Serguei M. Liachenko1 1Pfizer, Inc., Ann Arbor, Michigan, USA In this paper, we present a method for quantification of rat knee cartilage in a surgically induced model of osteoarthritis. Rats were imaged at 7T and cartilage volume and average thickness were quantified using specialized image processing algorithms for semi-automatic cartilage segmentation. A statistical method for the selection of the Region of Interest for increased local sensitivity is presented. This method provides a powerful tool for the preclinical assessment of rat cartilage changes and for drug assessment in osteoarthritis. 2549. MR-Assessment of Load-Influenced Collagen Fiber Orientation of Articular Cartilage Nikita Garnov1, Wilfried Gründer1 1University of Leipzig, Leipzig, Germany Collagen matrix lends the articular cartilage its form and stability and provides special mechanical properties able to resist loading and shear forces under physiological conditions. It has been shown that the collagen fibers are bent under static pressure. Due to the dependence of MR-intensity on collagen fiber orientation relative to B0, the characteristic changing of load-influenced MR-intensity in T2-weighted images can yield quantitative data about the deformation of collagen ultrastructure under load-bearing conditions. Based on these data, the additional biomechanical properties of cartilage tissue by means of MRI can be estimated. 2550. dGEMRIC in the Presence of Metal: Improving T1 Maps at 3.0T Agnes G. d'Entremont1, Shannon H. Kolind, Burkhard Maedler, David R. Wilson, Alexander L. MacKay 1University of British Columbia, Vancouver, Canada An important potential application of dGEMRIC is assessing the effect of surgery on cartilage health. However the presence of metal surgical hardware causes distortion of both images and T1 maps. Two approaches were used to reduce metal artifact: saturation recovery (SR) (versus inversion recovery (IR)), and the Metal Artifact Reduction Sequence (MARS). A gadolinium-doped phantom was imaged with and without metal using SR, IR, and MARS. Both the SR approach and MARS reduced metal artifact, with SR causing the greater improvement. Combinations of these two approaches may reduce artifact enough to map cartilage with dGEMRIC with metal near the joint. 2551. Influence of Knee Positions on T2, T2* and DGEMRIC Mapping in Porcine Knee Cartilage Toshiyuki Shiomi1, Takashi Nishii1, Yoshihiro Sakaguchi2, Norinao Matsumoto2, Yuichiro Matsuoka3, Yasuo Hara4, Norimasa Nakamura1, Yo Myoui1, Hideki Yoshikawa1 1Osaka University Medical School, Suita, Japan; 2Matsumoto medical clinic, Japan; 3Institute of Biomedical Research and Innovation, Japan; 4IVTeC Co. Ltd, Japan Influence of knee positions during MR imaging, such as flexed knee positions, on quantitative MR assessments has been scarcely investigated, even though knee positional change may cause substantial change of intra-articular biomechanical environment or magic angle effect. We examined influence of knee flexed positions on cartilage assessment by T2, T2* and dGEMRIC using cadaver porcine femoral-tibial joints. Our results may suggest that quantitative MR measurements allow intra-articular biomechanical assessment in association with knee flexion, after excluding factors of magic angle effects. 2552. Comparison of T1ρ and T2 Relaxation Times in Articular Cartilage Pavol Szomolányi1, Vladimir Mlynárik2, Siegfried Trattnig1 1Medical University of Vienna, Vienna, Austria; 2Ecole Polytechnique Fédérale de Lausanne, Lausanne, Switzerland T1ρ and T2 relaxation times measured using a Carr-Purcell-Meiboom-Gill technique with short inter-echo delays were compared in articular cartilage in vitro. It was shown that the mean T1ρ obtained with a spin locking amplitude of 500 Hz was close to T2 obtained with an inter-echo delay of 6.2 ms. Relaxation maps were also similar, which suggests that both relaxation parameters provide analogous information about cartilage pathology. 2553. T2 and T2* Relaxation as a Means to Evaluate Cartilage Repair Tissue - Initial Results Goetz Hannes Welsch1, Timothy Hughes2, Sebastian Quirbach1, Stephan Domayer1, Stefan Marlovits1, Siegfried Trattnig1, Tallal Charles Mamisch3 1Medical University of Vienna, Vienna, Austria; 2Siemens Medical Solutions, Erlangen, Germany; 3University of Berne, Berne, Switzerland A reported technique to directly visualize cartilage structure is, among others, quantitative T2-mapping. Underlying reliable results, T2*-mapping with its possible short scan time seems to offer a potential alternative. The goal of the presented feasibility study was to use T2* -in the follow-up of cartilage repair procedures and to compare it to established T2-mapping by a multi-echo spin-echo (SE) technique. The use of T2* allows maps to be created in short time frames and with resolutions that allow a detailed zonal analysis of cartilage. Thus T2* besides T2 might be an additional tool in the evaluation of cartilage ultrastructure. 2554. Precision of T2 Relaxation Times in Healthy and Osteoarthritic Human Tibial Cartilage Annie Horng1, José G. Raya1, Juergen Weber1, Sabine Weckbach1, Kathrin Nixdorf1, Maximilian F. Reiser1, Christian Glaser1 1University Hospital Munich, Campus Großhadern, Munich, Germany T2 relaxation time has high potential as a diagnostic parameter for cartilage. Precision errors of T2 measurements in tibial cartilage were assessed as to date they are only available for patellar cartilage. Precision errors of tibial cartilage T2 were slightly higher than patellar cartilage T2 precision errors, possibly related to smaller tibial cartilage thickness and consecutively increased partial volume effects. T2 precision errors were small compared to reported change in osteoarthritis (up to 180%) suggesting a reasonable discriminatory power of the technique. The data may provide a base for sample size calculations to design longitudinal and cross-sectional trials in osteoarthritis. 2555. A Quantitative Study of Knee Cartilage in Osteoarthritis Using Parallel Imaging MethodJin Zuo1, Xiaojuan Li1, Radu Bolbos1, Thomas Link1, Sharmila Majumdar1 1Univ. of California, San Francisco, San Francisco, California , USA Several quantitative methods have been proposed to study osteoarthritis (OA). However, it is not clear if these measurements can differentiate OA at different stages. In this study, we measured mean cartilage thickness, T1rho and T2 on severe OA patients, mild OA patients as well as healthy volunteers to investigate if these biomarkers will change accordingly with the severity of OA. The measurements were performed with a newly developed parallel imaging method.
2556.
Comparison of Delayed Gadolinium Enhanced MRI of Cartilage (DGEMRIC)
Using Inversion Recovery Tallal C. Mamisch1, M Dudda2, T Hughes3, D Burstein4, Young-Jo Kim5 1University of Bern, Bern, Switzerland; 2Children's Hospital-Boston, Boston, Massachusetts, USA; 3Siemens Medical System, Erlangen, Germany; 4Beth Israel Deaconess Medical Center, Boston, Massachusetts, USA; 5Chidren's Hospital-Boston, Boston, Massachusetts, USA Fast T1 mapping technique for delayed Gadolinium Enhanced MRI of Cartilage (dGEMRIC) was implemented for the hip. The fast T1 mapping technique was validated against inversion recovery technique using phantom and clinical studies. 2557. Inter-Scanner Variability of MRI-Based Tibial Cartilage Volumetric Parameters at 1.5T and 3T Annie Horng1, Sabine Weckbach1, Mike Notohamiprodjo1, Christina Muenkel1, Juergen Weber1, Maximilian F. Reiser1, Christian Glaser1 1University Hospital Munich, Campus Großhadern, Munich, Germany Knowing of slow cartilage loss of 0-5% per year and in view of multicentre studies, inter-scanner variability of MRI-based cartilage volumetric parameters in different scanner types of one manufacturer were evaluated. The results indicate good interscanner agreement at 1.5T and between 1.5T and 3T for all assessed volumetric parameters. In average, differences of 5% have to be taken into account when comparing data from different scanners of one manufacturer. These compare favorably to reported inter-company scanner differences as well as reported differences between scanners of different magnetic field strength. The data may contribute to calculate sample size in multicentric studies. 2558. Accuracy and Precision of Cartilage T2-Relaxation Time Quantification: Impact of the Fit Method Jose G. Raya1, Olaf Dietrich1, Annie Horng1, Jürgen Weber1, Julia Dinges1, Maximilian Felix Reiser1, Christian Glaser1 1Ludwig Maximilian University of Munich, Munich, Germany The low T2 times in articular cartilage and the high resolution necessary to resolve the cartilage layers unavoidably result in low-SNR data for T2 calculation. Noise in MRI has a non-vanishing mean, so that a non-zero signal is present at long TEs. Fitting to an exponential function results in systematic overestimation in T2 (up to 420% for low T2 and SNR). Two new methods for T2 calculation are proposed, which are very accurate and have a precision compatible with the maximal theoretical precision given by the Crámer-Rao lower bound. Impact of the fit on in vivo images has been analyzed. 2559. Resolving Multiple T2 Compartments in Cartilage with MRI David A. Reiter1, Ping-Chang Lin1, Kenneth W. Fishbein1, Richard G. Spencer1 1National Institute on Aging, National Institutes of Health, Baltimore, Maryland, USA Monoexponential fitting of T2 relaxation imaging data has been used extensively to characterize properties of cartilage based on its sensitivity to macromolecular concentration and structure. Due to the macromolecular composition of cartilage which consists of water compartments of differing fractions and mobility, cartilage T2 relaxation is more appropriately described by a multiexponential function. We performed simulations to demonstrate the optimal imaging parameters for accurately and reliably resolving two component T2 relaxation of cartilage. This analysis was applied to intact and enzymatically degraded cartilage in order to demonstrate the diagnostic potential of this analysis. 2560. Dedicated Two-Channel Phased Array Receiver Coils for HR-MRI of the Rat Knee Cartilages at 7T Adrian Rengle1, Radu Bolbos1, 2, Jean-Christophe Goebel3, Mircea Armenean1, Astrid Pinzano-Watrin3, Hervé Saint-Jalmes4, 5, Pierre Gillet3, Olivier Beuf1 1INSA-Lyon, Université Lyon 1, Villeurbanne, France; 2University of California San Francisco, San Francisco, California , USA; 3Université Nancy I, Vandoeuvre, France; 4Faculté de Médecine, Université Rennes 1 , Rennes, France; 5Centre Eugène Marquis, Rennes, France The goal of this work was to develop a dedicated two-channel array coil operating at 300 MHz (7T) for high-resolution MR imaging (HR-MRI) of the rat knee cartilage. A dedicated two-channel array coil with two square elements encompassing knee joint was designed and built. The average gain in Signal-to-Noise Ratio (SNR) compared to a 15 mm reference single loop coil was 2.2. This SNR gain was used to improve spatial resolution of 3D acquisition by decreasing the voxel size from 59x59x156µm3 to 51x51x94µm3 without time penalty. This two-channel array coil is a key element to perform HR-MRI and extract cartilage morphological parameters such as thickness and volume. Finally, a set of two array coils was used to perform the acquisition of both knee joints of a rat at the same time without SNR or resolution degradation compared to single knee joint acquisition. <
2561.
Magnetization Transfer of Cartilage and Cartilage Repair
Tissue at Ultra-High Fields Using SSFP
Oliver Bieri1, Goetz H. Welsch2, Tallal Charles Mamisch3, Oliver Kraff4, Mark Edward Ladd4, Klaus Scheffler1, Siegried Trattnig2 1University of Basel, Basel, Switzerland; 2Medical University of Vienna, Austria; 3University Hospital, Bern, Switzerland; 4University Duisburg-Essen, Germany Assessment of zonal differences in cartilage and its ultra-structure may benefit from improved image quality, contrast, and resolution at ultra-high fields. Here, magnetization transfer (MT) is evaluated for the assessment of cartilage repair. The feasibility of high resolution MT at 7T is demonstrated using MT sensitized steady-state free precession (MT-SSFP). Evaluation of MT for normal and for cartilage repair tissue reveals significant differences between healthy and affected articular cartilage. It is demonstrated that MT SSFP can successfully be used for the assessment of cartilage repair and areas of intact hyaline cartilage at 7T. 2562. 3D Distribution of T1ρ Relaxation Times in the Human Knee Joint Walter Robert Thurmond Witschey II1, Arijitt Borthakur1, Matthew Fenty1, J. Bruce Kneeland2, Matthew A. Sochor, Ravinder Reddy 1University of Pennsylvania, Philadelphia, USA; 2Pennsylvania Hospital, Philadelphia, USA A biomarker for cartilage degeneration T1rho varies in the knee articular cartilage by location (patellar, femoral or tibial cartilage) and by proteoglycan and collage substructure (superficial, middle, and deep zones) using a fully 3D T1rho acquisition technique for axial and coronal slices. Elevated T1rho is shown in a single asymptomatic subject corresponding with focal cartilage loss and bony cyst development. 2563. 3D-FSE-Cube for Rapid Assessment of Cartilage Morphology in the Knee Christina A. Chen1, Richard Kijowski2, Brian A. Hargreaves1, Scott B. Reeder2, Reed F. Busse3, Anja C. S. Brau4, Philip J. Beatty4, Garry E. Gold1 1Stanford University, Stanford, California , USA; 2University of Wisconsin-Madison, Madison, Wisconsin, USA; 3GE Healthcare, Madison, Wisconsin, USA; 4GE Healthcare, Menlo Park, California , USA The contrast provided by fast spin-echo (FSE) has made it the standard in assessing ligaments and menisci. However, two-dimensional FSE is limited by its incapability to measure cartilage volume. In healthy knees at 3.0T, we compared 3D-FSE-Cube to three-dimensional spoiled gradient-echo (SPGR) with IDEAL (Iterative Decomposition of water and fat with Echo Asymmetry and Least squares estimation). 3D-FSE-Cube yields equivalent cartilge volumes, higher fluid SNR, and better visualization of ligaments and menisci in a faster scan time than IDEAL-SPGR. 2564. Regional Difference of Diverse Zones in Human Menisci Investigated by Quantitative T2 Measurements Ping -Huei Tsai1, Teng-Yi Huang2, Hsiao-Wen Chung1, Guo-Shu Huang3 1National Taiwan University, Taipei, Taiwan; 2National Taiwan University of Science and Technology, Taipei, Taiwan; 3Tri-Service General Hospital and National Defense Medical Center, Taipei, Taiwan The purpose of this study is to determine the T2 distributions within different zones of the menisci. Six normal subjects were enrolled in this preliminary study. All of them were imaging in supine position in a 3.0T MR system (achieva, Philips Medical System). Three regions of interest (ROIs) were manually defined including red zone , white zone with radial collagen fibers, and white zone with circular collagen fibers. Differences between any two groups were statistically significant (P < 0.0001). We conclude that quantitative T2 measurements help recognizing the functions of the menisci, with potential in clarifying the relationships among menisci, hyaline cartilage, and osteoarthritis. |
Hall D Wednesday 13:30-15:30 Christina Schraml1, Petros Martirosian1, Nina F. Schwenzer1, Andreas Boss1, Claus D. Claussen1, Fritz Schick1 1University Hospital of Tuebingen, Tuebingen, Germany Arterial-spin-labeling perfusion imaging is a reliable, non-invasive MR technique for the assessment of tissue perfusion working without any contrast media administration. The aim of the study was to test the feasibility of MR perfusion imaging of the pancreas before and after oral stimulation using a FAIR-TrueFISP approach. Perfusion maps of diagnostic quality could be created by means of the extended Bloch equations. Significant stimulation-related changes in organ perfusion could be demonstrated. The method may provide important information in the assessment of acute and chronic inflammatory pancreatic diseases, endocrine and exocrine pancreatic function disorders and for monitoring of pancreatic transplants. 2627. Assessment of the Pancreas with MR Elastography Meng Yin1, Sudhakar K. Venkatesh2, Roger C. Grimm1, Phillip J. Rossman1, Armando Manduca1, Richard L. Ehman1 1Mayo Clinic, Rochester, Minnesota, USA; 2National University of Singapore, Singapore Chronic pancreatitis is a progressive and destructive inflammatory process of the pancreas resulting in necrosis and fibrosis. MR Elastography (MRE) is a potential non-invasive diagnostic tool for detecting emergence of fibrosis in chronic pancreatitis. In this study, we implemented a 7D MRE technique and a 3D LFE inversion algorithm to estimate shear stiffness in normal pancreatic parenchyma. The average shear stiffness of the pancreas tissue was 2.0±0.4 kPa in a series of ten normal volunteers. Preliminary results demonstrate that assessment of the pancreas with MRE technique is feasible, offering the prospect of a new non-invasive method for assessing chronic pancreatitis. 2628. Pancreatic Iron Overload in Thalassemia Patients: T2* MRI Evaluation – Initial Experience Orly Goitein1, Eli Konen1, Eliezer Rachmilewitz2, Tammar Kushnir1 1The Chaim Sheba Medical Center, Tel Hashomer, Israel; 2The Edith Wolfson Medical Center, Holon, Israel Iron deposition in hepatic, myocardial and endocrine gland tissues is a known complication in multitransfused Thalassemia patients. Pancreatic siderosis can affect both endocrine and exocrine functions. In addition, diabetes has been related to excess pancreatic iron deposition, besides increased peripheral insulin resistance. T2* MRI mapping was used to assess, for the first time, pancreatic iron levels of Thalassemia patients compared with normal subjects. Pancreatic T2* values were found to be significantly lower in all Thalassemic patients. Pancreatic data is an important addition to the arsenal of non invasive iron evaluation, allowing better understanding of the full spectrum of iron pathophysiology. 2629. Manganese Enhanced Magnetic Resonance Imaging (MEMRI) of Endogenous Rodent Pancreas Activation Muhammad E. Haque1, 2, Erica Markiewicz1, Suraj Serai3, Matt Zuckerman1, Lara Leoni1, Brian B. Roman1 1University of Chicago, Chicago, Illinois, USA; 2George Mason University, Fairfax, USA; 3University of Illinois, Chicago, Chicago, USA Presently pancreatic endocrine function is assessed using biochemical tests of organ output or indirectly by measuring serum glucose levels. Recently manganese enhanced imaging (MEMRI) has been used to study brain activation and we propose that similar paradigms can be applied to assess the function of the endogenous pancreas. To overcome several difficulties in the application of MEMRI to the pancreas, we used fast magnetization prepared rapid gradient echo (MP-RAGE). Rats were infused with MnCl2 followed by an IV glucose challenge to activate the pancreas. Images were obtained pre and post Mn infusion as well as post glucose stimulation. Mn and Glucose stimulation resulted change in image contrast. |
Hall D Wednesday 13:30-15:30 Foetal R-Waves Pierluigi Di Giovanni1, Alastair R. Allen1, Trevor S. Ahearn1, Thomas W. Redpath1 1Aberdeen University, Aberdeen, UK A snapshot of the foetal cardiovascular system during its contraction phase provides useful information concerning the health of the foetus.The purpose of this work is to investigate the possibility of real-time detection of foetal R-waves; this would allow a foetal ECG gated MRI acquisition. The ECG obtained from a maternal abdomen is a combination of maternal and foetal QRS complex, a signal processing method is needed to extract the foetal R-waves. We implement a Wavelet based method to isolate the singularities of the signal associated with the high frequencies contained in the foetal R-waves.
2694.
3D Visualization of Organs in Chemically Fixed Human Embryos
Acquired with Microscopic Spatial Kazuki Ohya1, Yousuke Otake1, Shinya Handa1, Katsumi Kose1, Shigehito Yamada2, Kohei Shiota2 1University of Tsukuba, Tsukuba, Japan; 2Kyoto University, Kyoto, Japan A mathod to visualize 3D structure of organs in chemically fixed human embryos acquired with microscopic spatial resolutions was developed. At first, a chemically fixed human embryo was imaged using a contrast-optimized T1 weighted 3D gradient echo sequence with a large matrix (256 x 256 x 512). Then an organ to be visualized was extracted using a region growing method and an interactive GUI tool. The extracted 3D region was visualized using a 3D visualization software package (Volume-One). The obtained result demonstrated a promise of our protocol. 2695. Exploring Alternative Contrast in the Uterus at High Magnetic Fields Angela Lynn Styczynski Snyder1, Timo Liimatainen1, Ute Goerke1, Michael Garwood1, Shalom Michaeli1 1Center for Magnectic Resonance Research, University of Minnesota, Minneapolis, Minnesota, USA The contrast in the uterus was generated using a novel rotating frame relaxation method, Relaxations Along a Fictitious Field (RAFF). This rotating frame technique provided unique tissue contrast from the conventional free precession T1 and T2 methods. It is demonstrated that artifact-free RAFF maps could be reliably obtained at a 7 T field strength. |
Hall D Wednesday 13:30-15:30 Federica Pediconi1, Antonella Roselli1, Valeria Dominelli1, Gianpaolo Pirovano2, Miles A. Kirchin3, Carlo Catalano1, Roberto Passariello1 1University of Rome “La Sapienza”, Rome, Italy; 2Bracco Diagnostics Inc., Princeton, New Jersey, USA; 3Bracco Imaging SpA, Milan, Italy Gadobenate dimeglumine and gadopentetate dimeglumine were compared intra-individually in 47 patients referred for contrast-enhanced MR mammography (CE-MRM). The dose of contrast agent (0.1 mmol/kg bodyweight) was identical for the two examinations in each patient. Histopathology data were available for 50 malignant and 28 benign lesions. Significantly more lesions were detected after gadobenate dimeglumine (75/78 [96%] vs. 62/78 [79%]; p=0.0025) with specific benefits noted for the detection of malignant lesions (49/50 [98%] vs. 38/50 [76%]; p=0.0018). The characterization of histologically-confirmed lesions as malignant or benign on CE-MRM was significantly (p=0.0001) better after gadobenate dimeglumine. Federica Pediconi1, Antonella Roselli1, Valeria Dominelli1, Miles A. Kirchin2, Carlo Catalano1, Roberto Passariello1 1University of Rome “La Sapienza”, Rome, Italy; 2Bracco Imaging SpA, Milan, Italy A retrospective evaluation was performed of 238 women with dense breast parenchyma (BI-RADS 3 or 4) who were referred for contrast-enhanced MR Mammography (CE-MRM) after undergoing conventional bilateral x-ray mammography (n=133) and/or ultrasound (n=195). Overall, 89/238 women underwent all three techniques. The diagnostic performance of CE-MRM in terms of sensitivity, specificity, accuracy and positive and negative predictive values was greater on CE-MRM than on x-mammography and ultrasound. Jeon-Hor Chen1, 2, Garima Agrawal1, Orhan Nalcioglu1, Min-Ying Su1 1University of California Irvine, Irvine, USA; 2China Medical University Hospital, Taichung, Taiwan Contrast enhanced benign lesions are the major reasons contributing to the high false positive diagnosis by MRI. The MR morphological features and enhancement kinetics from 16 hyperplasic lesions and 34 DCIS were analyzed based on BI-RADS MRI Lexicon, and compared. There were no clear features that can differentiate between these two groups. The malignant type enhancement kinetics (wash-out or plateau) was seen in 46% hyperplasic lesions, whereas it was seen more frequently in 75% DCIS. DCIS was more likely to present clumped pattern; whereas the pre-malignant lesions were more likely to show punctuate pattern.
2720.
Kinetic Assessment of Breast Lesions Using the Signal
Enhancement Ratio from Rapid Radial Lawrence Dougherty1, Raymond C. Boston, Mark A. Rosen1, Linda White Nunes, Peter J. Moate, Hee Kwon Song1, Mitchell D. Schnall1 1University of Pennsylvania, Philadelphia, Pennsylvania, USA The diagnostic performance was evaluated for DCE-MR breast imaging using the signal enhancement ratio (SER) to assess kinetic response. SER is the ratio of the signal at 60s post-contrast injection to the signal at 350s post injection (intensities relative to pre-contrast signal intensity). Images were acquired using an undersampled radial sequence. Sixty-four (64) enhancing lesions were evaluated using SER. ROIs were placed by two readers and the signal intensity plots were fitted and SER calculated. Using SER<0.8 for a benign classification, the combined readers’ performance was: Sensitivity = 92%, Specificity = 95%, PPV=93%, NPV=93% and the diagnostic accuracy was 93%.
2721.
Kinetic Parameter Estimation Based on Independent Component
Analysis of Breast Dynamic Yi Wang1, Bao Zhang2, Steven Roys2, Rao Gullapalli2 1University of Utah, Salt Lake City, Utah, USA; 2University of Maryland, School of Medicine, Baltimore, Maryland, USA We extent our recent success with the application of independent component analysis (ICA) in the detection of breast cancer to kinetic parameter estimation on spatial components obtained from such analysis. Twenty breast cancer patients¡¯ data were subject to ICA followed by kinetic parameter estimation of the raw data based on the independent components derived from ICA. Kinetic parameters derived from these spatial components were compared with those obtained from radiologist¡¯s assessment of the tumor using an existing CAD system. The sensitivity and specificity of the combined ICA and kinetic parameter estimation was 80 and 100% respectively compared to the radiologist. Ka-Loh Li1, Jessica Gibbs1, Evelyn Proctor1, Savannah C. Partridge2, Xiaoping Zhu1, Laura Esserman1, Nola M. Hylton1 1University of California, San Francisco, San Francisco, USA; 2University of Washington, Seattle, Washington, USA Early prediction of failure is essential in treatment planning for breast cancer patients. We studied 31 consecutive women who had primary invasive ductal carcinoma and preoperatively underwent four cycles of chemotherapy. Patients were divided into 'recurrence-free' (n=18) and 'recurrence' (n=13) groups based on follow-up information. Predictors were developed based on histogram analysis of data from pre-chemotherapeutic three time points contrast-enhanced MRI. A combined use of the predictors has identified 5/13 recurrence and 15/18 recurrence-free patients: None of the recurrence-free patients was misidentified as likely to recur, and none of the recur-patients was misidentified as unlikely to recur. 2723. Silicone-Specific Imaging Using a Single-Echo Dixon Technique Jingfei Ma1 1University of Texas M. D. Anderson Cancer Center, Houston, Texas, USA Previous methods on using the multiple-point Dixon techniques for silicone-specific imaging have to either assume that the frequency separation between W and S is twice the frequency separation between F and S, or to rely on using a preparatory inversion pulse to suppress F, or to require additional data acquisition and extensive processing to directly decompose the three components. In this work, we report that by acquiring a single-echo gradient echo image with the echo time set for water and fat in-phase, a single-echo Dixon algorithm with a flexible echo time can be used to generate excellent silicone specific images both in phantom and in vivo. 2724. Faster Magnetic Resonance Imaging of Breast Implants Using the IDEAL Technique Anna-Marie O'Connell1, Charles A. McKenzie2, Ananth Madhuranthakam1, Ivan Pedrosa1, Vandana Dialani1, Neil M. Rofsky1 1Beth Israel Deaconess Medical Center, Boston, Massachusetts, USA; 2University of Western Ontario, London,, Canada This study evaluates breast implant assessment with Short Tau Inversion Recovery (STIR) sequence modified to allow generation of water-only and silicone-only images with the IDEAL (Iterative-Decomposition of water and fat with Echo Asymmetry and Least-squares estimation) reconstruction technique. STIR techniques in isolation employ chemically selective suppression pulses and are known to be sensitive to magnetic field inhomogeneities. Seventeen women with implants were therefore imaged using the IDEAL-STIR combined technique. All 17 studies were diagnostic. The IDEAL technique took 15 minutes while the standard technique takes more than 35. IDEAL successfully produced fat suppressed water-only and silicone-only images in all 17 cases. No studies needed to be repeated. We conclude that MRI of breast implants using the IDEAL-STIR technique is significantly faster and more reliable than traditional assessment techniques. Christian Geppert1, Evelyn Wenkel2, Gerardo Hermosillo-Valadez3, Werner Bautz2, Berthold Kiefer1, Rolf Janka2 1Siemens Medical Solutions, Erlangen, Germany; 2Radiologisches Institut, Erlangen, Germany; 3Siemens Medical Solutions, Malvern, Pennsylvania, USA Motion artefacts can occur in MR breast imaging even with very cooperative patients and highest care in patient positioning, for example due to progressive relaxation of the pectoral muscles. In this work, the performance of a 3D elastic motion correction to high resolution breast MR data is described. Motion artefacts were measured as apparent skin thickening artefact as well as pectoral muscle subtraction artefact. The mean values of the skin motion artefact were originally 2.7mm (std deviation 1.1mm) and 0.7(0.89) mm after performing the correction. The averaged values for the muscle motion artefact was 4.8(2.95) mm before and 2.04(2.18) mm after the correction. Overall, a complete correction i.e. no remaining measureable artefact was achieved in 23 out of 47 times for skin motion respective 23 out of 48 times for muscle motion artefact.
2726.
Reduced B1-Inhomogeneities in Breast MRI Using Optimized RF
Excitation
Christian Geppert1, Juergen Nistler1, Wolfgang Renz1, Ioannis Panagiotelis1, Thorsten Speckner1 1Siemens Medical Solutions`, Erlangen, Germany B1-inhomogeneities are well known to occur at high field strengths. Only recently this issue has been raised for breast imaging at 3.0T: apart from the basic problem of image homogeneity, severe concern was expressed regarding possible misinterpretations of contrast uptake if the actual excitation angle deviates too strong from the nominal FLASH angle in gradient echo sequences used for dynamic studies. In this work, a novel method to improve the homogeneity that uses optimizations for the feeding of the body coil was tested and compared on volunteers. While different approaches to reduce B1-shading have been suggested previously, such as parallel transmission or the use of B1 saturation pads, this method provides a cost-effective solution using the standard transmit setup of a clinical 3T scanner. 2727. Parallel Imaging of Supine Breast for Recovery of Motion-Corrupted Data Helen Marshall1, Peter Siegler1, Patrick Micheal Devine1, Donald Bruce Plewes1 1Sunnybrook Health Sciences Centre, Toronto, Canada Supine breast imaging alleviates the problem of registering images acquired in a prone position to the supine orientation inherent to breast surgery. Parallel imaging was implemented for supine breast imaging with a gating and k-space reordering technique to compensate for respiratory motion, and used to recover data corrupted by additional non-respiratory motion. The proposed method protects data acquisition from motion events by grouping the even and odd phase encoding steps in time. Parallel imaging allows useable images to be reconstructed from a motion-corrupted dataset, providing that one of the data subsets (even or odd) remains viable. 2728. Dynamic Contrast-Enhanced Breast MRI Using Dual-Resolution 3D Spiral Imaging Brian A. Hargreaves1, Bruce L. Daniel1 1Stanford University, Stanford, California , USA Dynamic contrast-enhanced (DCE) imaging a standard technique for breast MRI, but is limited by both spatial and temporal resolution. Spiral image acquisition is fast and efficient, and can be designed to cover k-space with different densities for low and high spatial frequency information. This allows simultaneous acquisition of low and high spatial resolution images with fast and slower frame rates. We have demonstrated that this combination provides a useful tradeoff for DCE imaging in breast cancer patients, providing low and high resolution 3D images every 12 and 60 seconds. 2729. Accelerated Bilateral DCE 3D Spiral Breast Imaging: Comparison Between TSENSE and TGRAPPAMisung Han1, 2, Bruce L. Daniel1, Brian A. Hargreaves1 1Stanford University, Stanford, California , USA Using a dual-band excitation and 3D spiral imaging, both breast volumes can be efficiently imaged. Acquisition time can be reduced using TSENSE or TGRAPPA, where odd and even phase encode planes are acquired sequentially. Full images are used for calibration, and parallel imaging can then reconstruct accelerated frames. We compared the combination of 3D spiral imaging with TSENSE and TGRAPPA in bilateral DCE breast imaging. Both methods provided diagnostic breast images with 10 second temporal resolution, but TGRAPPA yield less oscillation in the signal intensity curves in tumors due to reduced residual aliasing. 2730. Slab-Phase Modulation Combined with Parallel Imaging in Bilateral Breast Imaging Misung Han1, 2, Brian A. Hargreaves1 1Stanford University, Stanford, California , USA By incorporating phase modulation with a dual-band spectral-spatial pulse, two breast volumes can be imaged simultaneously without the need to phase encode the unexcited space between the two volumes. In this work, we have combined self-calibrated parallel imaging using mSENSE and GRAPPA in the slice direction to the separately excited breast volumes, comparing results with and without slab-phase modulation. Our results show slab-phase modulation improves reconstructed image quality in terms of artifacts and SNR in both mSENSE and GRAPPA reconstructions. 2731. Rapid B0 Shimming with Three Orthogonal Frequency Maps Timo Liimatainen1, Patrick J. Bolan1 1University of Minnesota, Minneapolis, Minnesota, USA A fast image-based shimming method based on phase difference measurement from three orthogonal slices was introduced. This new technique requires only 7 s measurement time for estimation of 1st and 2nd order shim currents. Reliable shim current estimation was achieved also in the presence of discontinuities of frequency and chemical shift (water and fat interface). The results suggest reliable performance of new method in conventionally challenging tissue, such like breast, and will be particularly valuable for shimming in the body where physiological motion reduces conventional shimming performance. Marcus T. Alley1, Brian A. Hargreaves1, Bruce L. Daniel1, Philip J. Beatty2, Anja C. Brau2 1Stanford University, Stanford, California , USA; 2GE Healthcare Technologies, Menlo Park, California , USA For patients needing a Dynamic Contrast-Enhanced (DCE) MR breast exam, bilateral breast imaging is playing an increasingly important clinical role. Our DCE protocol includes a high-resolution 3D acquisition to provide anatomic data for use in diagnosis. Here we present results from a modified version of this acquisition that performs a slice-accelerated bilateral breast study using an autocalibrated parallel imaging technique in a scan time comparable to an equivalent unilateral study. Excellent image quality is obtained in both the native sagittal and reformatted axial planes, with the latter demonstrating the lack of residual aliasing artifacts in the slice direction. 2733. Rapid Parallel Breast MRI with a Custom Coil Array Glen Morrell1, Rock Hadley1, Emilee Minalga1, Eugene Kholmovski1, Matthias Schabel1 1University of Utah, Salt Lake City, USA A prototype breast coil was constructed to allow large reduction factors for parallel breast imaging. The coil consists of four plates, each with eight rectangular coil elements stacked in the z direction in a ladder configuration. Sample images are shown with a reduction factor of 4 for an 8 element subset of the complete coil set. Increased imaging speed made possible by parallel imaging will increase the accuracy of parametric perfusion modeling and allow high resolution depiction of morphology, which has the potential to significantly increase the specificity of breast MRI. 2734. Enhanced Lesion Discrimination in Breast DWI After Contrast Administration Christian Geppert1, Evelyn Wenkel2, Werner Bautz2, Berthold Kiefer1, Rolf Janka2 1Siemens Medical Solutions, Erlangen, Germany; 2Radiologisches Institut, Erlangen, Germany Numerous recent publications describe the use of diffusion-weighted imaging (DWI) for the detection, assessment or treatment response monitoring of breast cancer. While there appears to be good agreement that the additional evaluation of the apparent diffusion coefficient (ADC) can significantly increase the specificity of breast MRI, the applied pulse sequences, b-values and approach (before or after contrast administration, selected slices or complete volumes) differ. In this work, the effect of T1-shortening contrast administration (CA) on the ADC values of fat saturated DW-EPI and its impact on the ability to differentiate benign from malignant breast lesions is investigated. The ADC values of benign lesions were 1.8 (on average, values ranged from 1.4 to 2.6) 10-3 mm2/s before and 1.9(1.5; 2.5) 10-3 mm2/s after contrast administration. Malignant lesions exhibited a mean ADC of 1.0(0.6; 1.2) 10-3 mm2/s prior to and 0.9(0.5; 1.2) 10-3 mm2/s after CA.While there is no significant change for benign lesions the ADC decrease in post-contrast malignant lesions is significant (p<0.05) which in principle would lead to an improved lesion discrimination. 2735. Clinical Application of Diffusion-Weighted Imaging with ASSET Technique for Breast Lesions JIN Guangwei1, An Ningyu, Zeng Qingyu2 1Meitan General Hospital, Beijing, People's Republic of China; 2Meitan General Hospital, People's Republic of China DWI provides unique information about the state of molecular translational motion of water and maily detectes random, thermal motion of hydrone in vivo. The conventional DWI£¨SS-EPI£©used in breast can develope distortion of image. In this study, we investigated the tehnical feasibility of DWI with ASSET (array spatial sensitivity encoding technique) for patients with breast diseases, to study if ASSET-DWI can decrease the distortion with shorter acquisition time and to evaluate the diagnosis value of ASSET-DWI in distinguishing benign and malignant breast lesions. The results indicated ASSET-DWI can be used for breast with decrease of distortion and acquisition time. Either b value being 600s/mm2 or 1000s/mm2£ADC value of ASSET-DWI all can be used to distinguish majority of malignant lesions from benign ones. The diagnostic threshold of ADC value should be matched with the b value used in ASSET-DWI simultaneouly. 2736. Prognostic Imaging in Neoadjuvant Chemotherapy Should Be Cost Effective Brian William Pogue1, Anna N. A. Tosteson2, Marc Schegerin1, Peter A. Kaufman2, Keith D. Paulsen1 1Dartmouth College, Hanover, New Hampshire, USA; 2Dartmouth Medical School, Lebanon, New Hampshire, USA Cost effectiveness modeling was used to analyze if imaging in the course of neoadjuvant chemotherapy for breast cancer treatment should be carried out. The analysis showed that specificity of the imaging test was a secondary factor in knowing if this should be done, and that the primary factor in assessing this was the success of a secondary treatment strategy. The prognostic success of imaging during therapy will allow testing of possible secondary therapies in a clinical trial. 2737. Are Kinetic Parameters Related to Prognostic Indicators in < 2.0 Cm Invasive Ductal Carcinomas? Sanaz Arkani Jansen1, Gregory Karczmar, Akiko Shimauchi, Hiroyuki Abe, Gillian Newstead 1University of Chicago, Chicago, Illinois, USA We separate the effect of lesion size from the assessment of the relationship between MR kinetic parameters and prognostic indicators by evaluating the kinetic characteristics of 71 small T1 (< 2.0 cm) invasive ductal carcinoma (IDC) lesions. The kinetic parameters were classified by estrogen receptor (ER) status, nuclear grade and node invasion. We found that compared with ER positive and grade I and II lesions, ER negative and grade III tumors showed stronger washout. These results suggest that kinetic parameters are related to certain prognostic indicators irrespective of lesion size. Pål Erik Goa1, 2, Mariann Gjervik Heldahl2, Roar Johansen2, Jana Rydland1, Kjell Arne Kvistad1, Steinar Lundgren1, 2, Ingrid Susann Gribbestad2 1St.Olavs University Hospital, Trondheim, Norway; 2Norwegian University of Science and Technology (NTNU), Trondheim, Norway Dynamic contrast-enhanced (DCE) MRI derived parameters correlate to tumor characteristics and treatment response in breast cancer patients. A method for voxel-by-voxel classification of tumor tissue according to the shape of the contrast enhancement curve was implemented, and a color coded overlay illustrating spatial distribution of the different curve types was produced to improve visual inspection of the DCE-MRI data. The volume of tissue classified as having the fastest uptake/washout showed correlation to clinical outcome after 5 years. 2739. Predicting 5 Years Survival for Breast Cancer Patients Using Dynamic Contrast Enhanced MRI Line Rørstad Jensen1, Roar Johansen1, Pål Erik Goa1, 2, Jana Rydland2, Kjell Arne Kvistad2, Tone Frost Bathen1, Steinar Lundgren, 12, Ingrid S. Gribbestad1 1Norwegian University of Science and Technology, Trondheim, Norway; 2St. Olav’s University Hospital, Trondheim, Norway Some of the patients with locally advanced breast cancer fail to respond to neoadjuvant chemotherapy (NAC). Previous studies of dynamic contrast enhanced MRI (DCE-MRI) have shown that correlates to treatment response. The aim of this study was to investigate the feasibility of DCE-MRI in prediction of 5 years survival in breast cancer patients undergoing NAC. The results demonstrate significantly lower values for the DCE-MRI derived parameter RSI in survivors compared to non-survivors.
2740.
DCE-MRI Assessment of Neoadjuvant Chemotherapy and Clinical
Outcome in Locally Advanced Mariann Gjervik Heldahl1, Tone Frost Bathen1, Line Rørstad Jensen1, Pål Erik Goa1, 2, Steinar Lundgren1, 2, Ingrid Susann Gribbestad1 1NTNU, Trondheim, Norway; 2St. Olavs University Hospital, Trondheim, Norway Dynamic contrast enhanced MRI (DCE-MRI) derived parameters can be correlated to effect of breast cancer treatment. In a study of locally advanced breast cancer patients (n=67) receiving neoadjuvant chemotherapy, the DCE-MRI results were analysed for predictive and prognostic information. Mean relative signal intensity (RSI) and area-under-curve (AUC) values showed a significant decrease during NAC in clinical defined responders but not for non-responders. The difference was even more pronounced using DCE-MRI determined response. Based on the RSI-distributions, patients with a skew value higher than the mean skew survived longer than patients with a lower skew value (p=0,031). Min-Ying Su1, Kelly Huynh1, David J.B. Hsiang1, John A. Butler1, Rita S Mehta1, Shadfar Bahri1, Jeon-Hor Chen1, 2, Orhan Nalcioglu1 1University of California Irvine, Irvine, USA; 2China Medical University Hospital, Taichung, Taiwan The predicting accuracy of nodal status in patients after NAC based on MRI-based response and final pathological response was investigated in 48 patients. Her2 negative patients had a lower pCR rate than Her2 positive patients (36% vs. 78%), also had a higher rate with positive nodes (48% vs. 9%). The accuracy of the DCE-MRI results in predicting nodal status was 64% in Her2 – groups; and 87% in Her2 + groups. The prediction based on pathological response of primary tumor was only comparable to that of MRI response. DCE-MRI response had 100% NPV for predicting negative nodes for Her2 + patients, thus they may be spared of axillary node dissection. Jeon-Hor Chen1, 2, Rita S Mehta1, Shadfar Bahri1, Orhan Nalcioglu1, Min-Ying Su1 1University of California Irvine, Irvine, USA; 2China Medical University Hospital, Taichung 40407, Taiwan The MR imaging features and pathological response of 25 inflammatory breast cancers and 12 non-inflammatory breast cancers with skin enhancements on MRI were reported. IBC showed more aggressive disease presentations, with bigger tumor size and a high percentage of non-mass type lesions. Their response to our current neoadjuvant chemotherapy protocol combining AC and taxane-based regimens was similar with 50% patients achieved pathological complete response (pCR) in both groups. The accuracy of MRI in predicting pCR also did not show significant difference (11/16 in IBC vs. 5/6 in non-IBC); with a higher false negative in IBC (5/24) compared to non-IBC (1/12). Byron Feig1, Jeon-Hor Chen1, 2, David J.B. Hsiang1, John A. Butler1, Rita S Mehta1, Shadfar Bahri1, Orhan Nalcioglu1, Min-Ying Su1 1University of California Irvine, Irvine, USA; 2China Medical University Hospital, Taichung, Taiwan We investigated how the MRI-evaluated NAC response affected surgeon’s recommendation for choosing the optimal surgery in 76 patients. Before NAC, consensus recommendation of mastectomy was made for 49 patients, lumpectomy for 23 patients. After NAC, 22 mastectomy candidates remained as mastectomy, with pre-treatment size of 5.6 ± 3.2 cm. Twenty mastectomy candidates were changed to lumpectomy, and they had a smaller pre-treatment size 3.6 ± 1.9 cm. The results indicated that in patients who had more extensive pre-treatment disease, despite of their great response to NAC, the surgeons may still apply an aggressive approach and recommend for mastectomy. 2744. Are Kinetic Parameters Diagnostically Useful for Breast Lesions Exhibiting Nonmass-Like Enhancement? Sanaz Arkani Jansen1, Xiaobing Fan, Gregory Karczmar, Maryellen Giger, Hiroyuki Abe, Gillian Newstead 1University of Chicago, Chicago, Illinois, USA The purpose of this study was to investigate whether the sensitivity and specificity of kinetic parameters can be improved by considering mass and nonmass breast lesions separately. The contrast media uptake and washout kinetics of 112 breast lesions (34 benign, 78 malignant) were analyzed using an empirical mathematical model. Our results suggest that kinetic analysis is not effective for nonmass lesions, while may be quite effective in mass lesions. The diagnostic utility of kinetic analysis of breast lesions, for example in computer aided diagnosis schemes, is likely improved if performed after identifying the lesion as mass-like or nonmass-like enhancement. 2745. MR Imaging Features of Triple Negative Breast Cancer Jeon-Hor Chen1, 2, Hyeon-Man Baek1, Rita S Mehta1, Philip Carpenter1, Orhan Nalcioglu1, Min-Ying Su1 1University of California Irvine, Irvine, USA; 2China Medical University Hospital, Taichung, Taiwan The MR imaging features of 29 TN (triple-negative, ER, PR and HER2 negative) tumors were reported, 25 with pure invasive ductal cancer (IDC), and 4 metaplastic (IDC with squamous or chondroid components). 23/29 lesions were greater than 2 cm, indicating the late diagnosis. The imaging features were consistent with that of IDC; 28/29 were mass lesions. The enhancement kinetics was measured from 22 lesions, and all showed the malignant pattern with wash-out in late phase. Approximately half (14/29) showed enlarged lymph nodes in the axilla. Nine of these patients had MR spectroscopy study, and 7 showing elevated choline. 2746. PET/CT Versus MRI for Detection of Bone Metastases in Patients with Breast Cancer Johan Grankvist1, Rune Vincents Fisker1, Victor Vishwanath Iyer1, Finn Taagehøj Jensen1, Ernst-Torben Wilhelm Fründ1, Carsten Wiberg Simonsen1, Thorkil Christensen1, Marianne Ewertz Kvistgaard1, Elna-Marie Larsson1 1Aalborg Hospital / Aarhus University Hospital, Aalborg, Denmark The project compares MRI with 18-DFG-PET/CT. Breast cancer patients with suspected or known bone metastases, are examined on both modalities. For 27 predefined locations in the spine and pelvic bones, 2 independent groups of doctors assign metastases detected on each modality. The results from both groups were compared for the first 9 patients. 44 metastases were detected on PET/CT. All 44 metastases were found on MRI. In addition 34 lesions were detected on MRI only, probably representing post therapeutic residuals without active tumour. 2747. Frequency Selective Inversion of Lipid Improves Choline Conspicuity in Breast Spectra Martin Lowry1, Ralph Noeske2, Martin D. Pickles1, Lindsay W. Turnbull1 1University of Hull, Hull, UK; 2GE Healthcare, Berlin, Germany The presence of an elevated choline peak in proton spectra appears to be consistent marker of malignancy in many tissues. In most tissues this is relatively easy to detect. In breast tissue, however, the dominant peaks from lipid may considerably reduce the visibility of choline. Several approaches have been used to combat this problem. The method which has gained some acceptance combines T2-weighting and TE-averaging using long echo-times (~150ms) to reduce lipid signal and also removing sideband artefacts. In this study we implemented an additional fat suppression pulse that allows shorter echo times and used enhanced post-processing for frequency referencing. 2748. A Robust Automated Shimming Procedure for Breast MR Spectroscopy Evelyne Balteau1, 2, Geoffrey David Charles-Edwards3 1University of Liege, LIEGE, Belgium; 2University College London, LONDON, UK; 3Guy's & St Thomas' NHS Foundation Trust, LONDON, UK MR spectroscopy offers attractive possibilities for the non invasive in vivo assessment of breast lesion biochemistry, as an aid for diagnosis and an early indicator of disease response to treatment. Effective shimming to achieve a high magnetic field homogeneity is paramount for MRS to achieve sufficient spectral resolution and avoid positional errors. However, few studies aiming breast spectroscopy address this often challenging issue. In this study, an automated shimming procedure is implemented and adapted for optimal breast MRS on a 1.5T clinical scanner. The robustness of the technique is demonstrated by both in vitro and in vivo experiments. 2749. Choline Profiles of Breast Cancer Correlate to Clinical Tumor Characteristics Beathe Sitter1, Tone Frost Bathen1, Maria Tunset Grinde1, Hans Fjøsne2, Jostein Halgunset1, Steinar Lundgren2, Ingrid S. Gribbestad1 1NTNU, Trondheim, Norway; 2St. Olav University Hospital, Trondheim, Norway Choline profiles were obtained from 336 tissue biopsies from breast cancer patients using HR MAS MR spectroscopy. Peak ratios and partial least square regression were used to compare choline profiles to clinical parameters related to tumor characteristics. Significant differences were found for choline peak ratios of ER hormone positive and negative tumors, and for non-involved tissue compared to IDCs. ER status predicted from the choline profiles was significantly correlated to clinically predicted ER status using PLS. The choline profiles displayed a high level of variance, reflecting the complexity of breast tumor biology. 2750. Three Dimensional Spectroscopic Imaging in Breast Cancer at 3Tesla: A Pilot Study Stephan Gruber1, Katja Pinker2, Wolfgang Bogner1, Günther Grabner2, Andreas Stadlbauer3, Thomas Helbich2, Siegfried Trattnig2 1Medical University of Vienna, Vienna, Austria; 2Medical University of Vienna, Austria; 3Landesklinikum St. Poelten, Austria In this study we tested 3D-MR-spectroscopic imaging in breast cancer. Compared to previous studies, performed with single voxel spectroscopy or 2D-MR-spectroscopic imaging, we could cover a larger fraction of the breast. In all patients with biopsy confirmed cancer, Cho was detectable in more than one slice, pointing out the importance of 3D-MRSI for diagnosis of breast cancer. Cho with a relative high SNR may be detected in voxels containing benign tissue, too. Therefore, specificity of this method should be tested in further studies. 2751. Validation of a Novel Spectroscopic Diffuse Optical Tomography System for Breast Imaging with MRI Stephanie M.W.Y. van de Ven1, Sjoerd G. Elias1, Andrea J. Wiethoff2, Marjolein van der Voort3, Anais Leproux3, Tim Nielsen4, Bernhard Brendel4, Leon Bakker3, Martin B. van der Mark3, Willem P.Th.M. Mali1, Peter Luijten1 1University Medical Center Utrecht, Utrecht, Netherlands; 2Philips Medical Systems, Best, Netherlands; 3Philips Research Europe, Eindhoven, Netherlands; 4Philips Research Europe, Hamburg, Germany Diffuse Optical Tomography (DOT) uses near-infrared light to assess the three-dimensional distribution of optical properties of tissue. By combining data acquired at multiple wavelengths (spectroscopy), tissue composition in terms of concentrations of oxy- and deoxyhemoglobin, water and lipid can be determined. In this study, DOT measurements in 24 cystic breast lesions were validated with MRI. Optical images showed excellent spatial agreement with MRI. Spectroscopic analysis of the lesions elucidated their high water content and low total hemoglobin content. This is a promising first step in the validation of a new optical breast imaging tool with functional and molecular imaging potential. 2752. Breast Cancer Characterization with Optical Spectroscopy Integrated Into an MR Breast Biopsy Plate Colin M. Carpenter1, Subhadra Srinivasan1, Brian W. Pogue1, Shudong Jiang1, Keith D. Paulsen1 1Dartmouth College, Hanover, New Hampshire, USA Simultaneous optical spectroscopy and DCE-MR of breast cancer has been studied in simulation and in vivo to characterize breast tumors. The optimization of the patient interface has been investigated, comparing a circular fiber interface and an attachment to a breast biopsy grid. In vivo results of both healthy and abnormal breast exams taken with both approaches are presented. 2753. Diffusion-Weighted Imaging: A Valuable Aid for the Determination of the Margin of Breast Carcinoma Yili Zhang1, Xiaoyan Huang1, Hongwen Du1, Yun Zhang1, Xin Chen2, Peng wang2, Youmin Guo3 1First Hospital of Xi¡¯an Jiaotong University, Xi¡¯an, People's Republic of China; 2Second Hospital of Xi¡¯an Jiaotong University, Xi¡¯an, People's Republic of China; 3Beijing Chao-Yang Hospital, Captial Medical University, Beijing, People's Republic of China In our present study we first applied DWI to compare the ADC value of malignant tumor with that of the peripheral tissue. We found that from the center of tumor to its peripheral tissue, the ADC values gradually increased and 10 mm away from the visible margin of tumor in image outwards, there existed an invisible tumor margin, which we regarded as the so called biological margin. Thus, the excision scope should be at least 10 mm from the visible border so that the excision can be complete.
2754.
DCE-MRI and Dynamic 15O-Water PET/18F-FDG PET for Assessing
Tumor Vascularity, Histology, and Savannah C. Partridge1, Peter R. Eby1, Steven W. White, Robert K. Doot, Brenda F. Kurland, Erin K. Schubert, Lisa K. Dunnwald, Constance D. Lehman, David A. Mankoff 1University of Washington, Seattle, Washington, USA Dynamic contrast-enhanced MRI (DCE-MRI) and positron emission tomography (PET) are two functionally different imaging modalities predictive of response to neoadjuvant therapy in breast cancer. However, the extent of their agreement or discordance for tumor characterization is not well understood. We investigated the combination of DCE-MRI enhancement kinetics, 15O-water PET measures of blood flow, and 18F-FDG PET metabolic measures for characterizing human breast tumors. We found PET and DCE-MRI vascular measures were well correlated, and FDG PET provided complementary metabolic information. Several PET and DCE-MRI measures were significantly associated with tumor grade, ER/PR status, and pathologic response to neoadjuvant treatment.
2755.
Nonrigid Registration for Motion Correction in
Contrast-Enhanced MR Breast Imaging: Comparison Alphonso William Magri1, Andrzej Krol2, Mehmet Unlu1, James Mandel1, Gwen Tillapaugh-Fay2, Wei Lee2, Edward Lipson1, David Feiglin2 1Syracuse University, Syracuse, New York, USA; 2SUNY Upstate Medical University, Syracuse, New York, USA In order to improve diagnostic accuracy for evaluation of breast tissue lesions detected in contrast-enhanced MR imaging one needs to perform nonrigid registration of consecutive time frames. The objective of this study was to compare the quality of registration accomplished by two different approaches: Finite Element Method (FEM) with fiducial skin markers and a Free Form Deformation (FFD) method. We conclude that FFD plus dedicated hardware and software that would reduce execution time is a method of choice for nonrigid registration of consecutive time frames in dynamic MR breast imaging. Karl-Heinz Herrmann1, Alexander Rauscher2, 3, Pascal Baltzer1, Diane Renz1, Jürgen R. Reichenbach1 1Friedrich-Schiller-University, Jena, Germany; 2University of British Columbia, Vacouver, Canada; 3Friedrich-Schiller University , Jena, Germany Simultaneous acquisition of in-phase (ip) and opposed-phase (op) data of a time series allows to create a multitude of new image combinations, which may generate new beneficial or even superior contrasts compared to a standard ip image magnitude subtraction. We are investigating two particular promising new contrasts on in vivo patient data, one acting as a contrast-to-noise booster in comparison to regular magnitude ip-subtraction. The other contrast combines anatomical information, especially Cooper ligaments and tissue boundaries, with a contrast amplification for partially volumed lesion boundaries which might otherwise be difficult to detect. The two contrasts can also be integrated into one image to reduce the number of total images. Erez Eyal1, Daria Badikhi1, Edna Furman-Haran1, Fred Kelcz2, Hadassa Degani1 1Weizmann Institute of Science, Rehovot, Israel; 2University of Wisconsin Hospital and Clinics, USA The purpose of this work is to evaluate a new method for diagnosing breast lesions that integrates principal component analysis with the three time point model based analysis. DCE-MR images of 69 lesions, 38 benign and 31 malignant were analyzed by this method. Two specific eigen-vectors were shown to be sufficient for describing the time dependent patterns of contrast enhancement. They were transformed by rotation to reflect physiological behavior using a correlation with the 3TP method. The reproducibility of the patterns (for a standardized protocol) and the fast image processing makes it an attractive choice for CAD.
2758.
Evaluation of the Diagnostic Accuracy of Computer-Aided
Detection of Breast Cancer Using MRI Jacob Levman1, Petrina Causer1, Ellen Warner1, Don Plewes1, Anne L. Martel1 1Sunnybrook Health Sciences Centre, University of Toronto, Toronto, Canada Breast MRI is likely to continue to play a significant diagnostic clinical role. Researchers are developing new techniques to improve the spatial resolution, temporal resolution and signal to noise ratio of breast MRI. This study evaluates the benefit of increased temporal resolution on our ability to discriminate between malignant and benign lesions using computer-aided detection tools. This study compares two protocols (with high and low temporal acquisitions) by comparing their receiver operating characteristic curve areas. Results show little difference between the two protocols indicating that researchers should focus their efforts on improved spatial resolution and signal to noise ratio. |
Tumor Perfusion & Permeability: Methodology & Applications Hall D Wednesday 13:30-15:30 Analysis Across Anatomical Locations Nina Tunariu1, Jane Taylor1, James Stirling1, James d'Arcy2, David J. Collins2, Simon Walker-Samuel2, Anwar R. Padhani1 1Mount Vernon Cancer Centre, Northwood, Middlesex, UK; 2Institute of Cancer Research, Sutton, Surrey, UK In this study we examine the effects of different population arterial input functions (AIFs) on transfer constant estimates, measurement error, and on model fitting in liver, pelvis and breast tumours. Our analyses indicate that there is an improvement in reproducibility by using “more physiological” (non-Weinmann) AIFs at all anatomical sites. We also noted that tumours vary in the number of pixel failures and goodness of kinetic model fitting by anatomical location. Jan Georg Korporaal1, C. A.T. van den Berg1, Cecile R. Jeukens2, G. Groenendaal1, Marco van Vulpen1, Peter Luijten1, Uulke A. van der Heide1 1University Medical Center Utrecht, Utrecht, Netherlands; 2Elkerliek Hospital, Helmond, Netherlands Differences between DCE-MRI and DCE-CT in prostate cancer may exist, affecting the choice for a perfusion model. A quantitative comparison of the AATH model and Tofts model has been performed in 10 patients on both DCE-CT and DCE-MRI by fitting mean contrast enhancement curves of high-flow regions to both models. The AATH model fitted better to the DCE-CT curves, since arterial first pass peaks (FPP) were only visible on DCE-CT. For the Tofts model the parameters are equal for DCE-MRI and DCE-CT, although the model fails to describe the FPP on DCE-CT. The AATH model also yields similar results for DCE-MRI and DCE-CT, but some caution is required as the transit time shows significant differences. 2784. Short-Term Anti-Angiogenic Therapy Improves Tumor Microvascular Function: Investigation with DynamicContrast Enhanced MRI Ulrike Nöth1, Sebastian Strieth2, Gunnar Brix3, Marc Dellian2, Martin E. Eichhorn2 1Helmholtz Center Munich - German Reseach Center for Environmental Health, Neuherberg, Germany; 2Klinikum Großhadern, University of Munich (LMU), Munich, Germany; 3Federal Office for Radiation Protection, Neuherberg, Germany Anti-angiogenic therapy by blocking VEGF signalling combined with chemotherapy is a novel strategy for clinical cancer treatment. One possible mechanism leading to an improved tumor microvascular function, and therefore, increased drug delivery during chemotherapy could be a “normalization” of tumor vessels after anti-angiogenic therapy. Aim of the study was to study tumor microvascular function in a highly vascularized melanoma in Syrian golden hamsters with dynamic contrast enhanced MRI upon anti-VEGF treatment and compare it to untreated control tumors.
2785.
Quantification of Perfusion Change with Carbogen Breathing in
a Subcutaneous Rat Tumour Model Using John Carr1, Daniel Bradley2, Geoff Parker1, David Buckley1, Jean Tessier2 1University of Manchester, Manchester, UK; 2AstraZeneca, Macclesfield, UK ASL is shown to be a feasible method for measuring F in a subcutaneous tumour model and sensitive enough to quantify changes resulting from carbogen breathing. A positive correlation is shown between T2* change and F change due to carbogen (r = 0.68, p = 0.01) and between ADC and F (r = 0.67, p =0.04). No correlation was found between IAUC and F indicating IAUC measurements are dominated by CA leakage. In summary, a novel application of the ASL technique is demonstrated showing how ASL can be used in tumours outside the brain in animals and potentially in humans. 2786. Mapping of Viable of Tumor Regions Using Gd-DTPA DCE-MRI Minming Huang1, Mutian Zhang1, 2, Jenghwa Chang1, Carl H. Le1, Pat B. Zanzonico1, Humm John1, Jason A. Koutcher1, 3, Cliff C. Ling1 1Memorial Sloan Kettering Cancer Center, New York, New York, USA; 2Columbia University, New York, New York, USA; 3New YOrk, New York, USA As part of our ongoing research program to develop and validate techniques for imaging of tumor hypoxia, we undertook this study using Gd-DTPA DCE-MRI co-registered with 18F-FMISO PET images to map the distribution of viable cells within tumors and thereby optimize the correlation between interstitial pO2 readings and PET-derived FMISO concentrations in tumors. 2787. Evaluation of a Model-Free Approach to Contrast-Enhanced MRI in Lung Tumors Frank Risse1, Tristan Anselm Kuder1, Christian Hintze1, Christian Fink2, Hans-Ulrich Kauczor1, Wolfhard Semmler1 1Deutsches Krebsforschungszentrum, Heidelberg, Germany; 2Medical Faculty Mannheim - University of Heidelberg, Mannheim, Germany The signal after administration of a contrast agent is either analyzed directly or quantitatively using pharmacokinetic (PK) models. A simple approach to analyze contrast-enhanced MR perfusion images for a clinical setting that enables comparisons between follow-up examinations of patients without using a PK-model was evaluated by means of tumor simulations based of a PK-model. The examination of the hemodynamic conditions and their heterogeneity within tumors was feasible with the model-free approach. The calculated initial slope and the maxima are directly related to tumor flow and volume. The normalization with an AIF might enable an intra-individual comparison under therapy. 2788. A Data-Driven Methodology for Cross-Visit Sub-Segmentation of Tumours in DCE-MRI Studies Giovanni Alessandro Buonaccorsi1, James P B O'Connor1, 2, Chris J. Rose1, Caleb Roberts1, Angela Caunce1, Sue Cheung1, Yvon Watson1, Karen Davies1, Lynn Hope2, Alan Jackson1, Gordon C. Jayson2, Geoff J M Parker1 1University of Manchester, Manchester, UK; 2Christie Hospital, Manchester, UK We describe an unsupervised data-driven method of cross-visit image segmentation for DCE-MRI time series data. We employed principal components analysis for dimensionality reduction then performed k-means clustering to generate segmentations that can be applied to 3-D parametric maps resulting from model-based or model-free analyses to derive per-cluster parameter statistics for data from a clinical study: each resulting cluster had a different visit-by-visit pattern of parameter values. On the basis of our results, we propose that our segmentation methodology will provide a useful objective means of following the evolution of tumour sub-compartments in response to treatment over a whole DCE-MRI study. 2789. Tracer Kinetic Analysis of a Simultaneous T1- And T2* Measurement in a Tumor Model Steven Sourbron1, Melanie Heilmann2, 3, Andreas Biffar1, Christine Walczak2, 3, Julien Vautier2, 3, Andreas Volk2, 3, Michael Peller1 1Ludwig-Maximilian-University Munich, Munich, Germany; 2Institut Curie, Orsay, France; 3INSERM, Orsay, France Quantification errors are known to arise with T2* bolus-tracking MRI from the difference in relaxivity between artery and tissue, and tracer leakage effects. Here we assess these errors using a simultaneous measurement of T1 and T2* data in a tumor model. We find that (i) the difference in relaxivities leads to overestimation of flows- and volumes; (ii) T2* relaxivity is reduced during extravasation, which produces an underestimation in the transit times. We conclude that perfusion and permeability cannot be measured from T2* data alone, but that the combination with T1 produces previously unmeasureable parameters with high tissue sensitivity.
2790.
Intraindividual Crossover Studies with Gadobenate Dimeglumine
for Contrast-Enhanced MR Imaging Guenther Schneider1, Federica Pediconi2, Luigi Grazioli3, Michael V. Knopp4 1University Hospital of Saarland, Hamburg, Germany; 2University of Rome “La Sapienza”, Rome, Italy; 3University of Brescia, Brescia, Italy; 4Ohio State University, Columbus, Ohio, USA In MRI crossover studies of the liver, breast, and vasculature, patients received 2 identical MR exams within 2-14 days, one with Gd-BOPTA and the other with comparator. For liver imaging, Gd-BOPTA provided additional diagnostic information and the possibility of using lower overall doses. In breast imaging, Gd-BOPTA depicted significantly (p=0.003) more lesions than Gd-DTPA, and lesions were significantly (p<0.001) more conspicuous, demonstrating the benefit of higher T1-relaxivity for this indication. From the carotid arteries to the peripheral vascular beds, Gd-BOPTA demonstrated superior vessel enhancement and conspicuity, particularly for depiction of smaller vessels. |
Brain Tumor: Spectroscopy, Susceptibility & Functional MRI Hall D Wednesday 13:30-15:30 Spectroscopic Imaging at 3T Ilwoo Park1, 2, Susan M. Chang2, Sarah J. Nelson1, 2 1University of California Berkeley/San Francisco, San Francisco, California , USA; 2University of California, San Francisco, California , USA Lactate may be a valuable predictive marker for treatment outcome. The goal of this study is quantitative and qualitative assessment of lactate prior to and during radiation/chemo treatment in GBM patients using a previously developed lactate-edited 3D MRSI at 3T. Normalized lactate was significantly different between progressing and non-progressing patients, and the location of tumor recurrence coincided with the regions of lactate appearance at earlier exams with no signs of recurrence at the time of progression. It suggests that monitoring lactate may assist in characterizing tumor tissue and predicting the outcome of treatment for GBM patients. 2815. Lactate Production in Human Brain Tumor; Detection by 13C MRS at 3T Jannie P. Wijnen1, Dennis W.J. Klomp1, Albert J.S. Idema1, Bastiaan E. Galan1, Arend Heerschap1 1Radboud University Nijmegen Medical Centre, Nijmegen, Netherlands This study demonstrated the in vivo lactate production in a glioblastoma multiforme in a patient for the first time. We used localized single channel distortionless enhanced polarization transfer (SC-DEPT) 13C MR Spectroscopy at 3T with venous infusion of 1-13C labeled glucose. During an euglycemic clamp lactate labeled at the C3 position was detected in the voxel located in the tumor after approximately 20 minutes of 1-13C-glucose infusion. In the contra lateral voxel no lactate was detected, but much more formation of compounds downstream of lactate. 2816. High Resolution Fast Elliptical SENSE MRSI of Gliomas at 7T Esin Ozturk-Isik1, Wei Bian1, Janine Marie Lupo1, Duan Xu1, Radhika Srinivasan1, Ilwoo Park1, 2, Kate Hammond1, 2, 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 Higher SNR at 7T can be traded for better spatial resolution to reduce the partial voluming effects for MR spectroscopic imaging of patients diagnosed with brain tumors. However, more phase encoding steps would be necessary to achieve higher spatial resolution using conventional PRESS MRSI localization. In addition, MR spectroscopic imaging at 7T requires a longer scan time due to the longer repetition times to avoid signal saturation. In this study, we propose high resolution MR spectroscopic imaging of glioma patients at 7T with a scan time of 6 minutes using elliptical SENSE technique. 2817. In Vivo and Ex Vivo 1H MRS Metabolite Profiles of Gliomas Franklyn A. Howe1, Kirstie S. Opstad1, B. Anthony Bell1, John R. Griffiths2 1St George's, University of London, London, UK; 2Cancer Research UK Cambridge Reseach Institute, Cambridge, UK Knowledge of glioma grade aids treatment selection and patient prognosis. Metabolite ratios determined by 1H MRS are still used as measures of tumour grade, as an alternative to invasive biopsy. We have used in vivo and ex vivo 1H MRS and histopathology to quantify and model metabolite levels in gliomas. As well as variability in cell density and necrosis, there is metabolic variability that renders individual metabolite ratios and metabolite concentrations unsatisfactory biomarkers. For gauging glioma grade, pattern recognition analysis of all metabolites should generally be used. Peter Dechent1, Hans Christoph Bock1, Arne Wrede1, Jan-Hendrik Buhk1, Alf Giese1, Gunther Helms1, Jens Frahm2, Herwig Strik1, Michael Knauth1, Kai Kallenberg1 1University Göttingen, Göttingen, Germany; 2Biomedizinische NMR Forschungs GmbH am Max-Planck-Institut fuer biophysikalische Chemie, Göttingen, Germany Glioblastoma multiforme (GBM) is the most common primary brain tumor with invasive growth and mainly of astrocytic origin. Previous studies in GBM patients identified tumor cells in macroscopically normal brain parenchyma. We used localized proton MRS to assess alterations of cerebral metabolite within the contralateral hemisphere of GBM patients as potential markers of GBM cells. MRS revealed elevated concentrations of myo-inositol and glutamine in GBM patients relative to control subjects as well as low-grade glioma patients. This finding most likely represents a mild astrocytosis indicating early neoplastic changes in line with reports of glioma cells in otherwise inconspicuous brain parenchyma. Martin Krssak1, Georg Widhalm1, Stephan Gruber1, Wolfgang Dietrich1, Stephan Wolfsberger1, Daniela Prayer1 1Medical University of Vienna, Wien, Austria Clinical feasibility of 2D spectroscopic imaging and possibility of integration of metabolic maps with 3D anatomical data was evaluated in different regions of human brain in 25 patients with tumors. SI yielded robust data from intracerebral tumors positioned above and in the height of lateral ventricles. This data could be used for the integration into neurosurgery planning with the help of additional multi-slice T2weighted data set into 3D T1weighted anatomic data. Additional shimming procedure and/or special selection of spectroscopic VOI has to be considered for the SI measurements of the meningial tumors and/or lesions near the skull (base). 2820. Gadolinium Effect on Choline Signal in Brain Tumors as a Function of T1 Enhancement Maria Otaduy1, Rita Pincerato1, Eduardo Lima1, Katia Pincerato1, Luciana Borges1, Claudia Leite1 1Medical School of the University of São Paulo, São Paulo, Brazil Choline measured by MRS is a tumor marker and its quantitation is important for diagnostic and follow up purposes of brain tumors. It has been described that paramagnetic contrast administration has the potential to change its intensity. Some experimental studies suggest that Gd has a minimal, if any effect on Cho signal. In our study we observed Cho changes after Gd injection as a function of T1 enhancement, which indirectly reflects Gd concentration, and our results are in agreement with theoretical simulations that show Cho changes to be strongly dependent on the Gd concentration reaching the tissue. 2821. Using 7T Susceptibility-Weighted Imaging to Aid in the Characterization of High-Grade Gliomas Janine M. Lupo1, Susan M. Chang1, Sarah J. Nelson1 1University of California, San Francisco, USA Susceptibility-weighted imaging (SWI) is a powerful tool for high resolution imaging of the vasculature, allowing the detection of vessels or microhemorrhages as small as 100-200 μ at 7 Tesla. This study investigated the unique contrast present in SWI images of high-grade glial tumors by analyzing the overlap of 7T SWI hypointensity with the T2 hyperintensity region, contrast enhancing lesion, and perfusion abnormality at 3T. The addition of SWI highlighted heterogeneity within the contrast enhancing lesion and regions of elevated blood volume and/or microvascular leakage. 2822. Evaluation of Astrocytic Brain Tumors with 7T Versus 1.5T MRI Christoph Moenninghoff1, 2, Oliver Kraff1, 2, Jens M. Theysohn1, 2, Stefan Maderwald1, 2, Elke R. Gizewski1, 2, Jens P. Regel1, Mark E. Ladd1, 2, Michael Forsting1, 2, Isabel Wanke1, 2 1University Hospital Essen, Essen, Germany; 2University Duisburg-Essen, Essen, Germany Twelve patients with histologically confirmed astrocytic brain tumors (WHO grade II-IV) underwent pre-treatment MRI at 1.5T and 7T. T 2, PD and T2* weighted images were performed. The signal gain of 7T MRI was used to achieve a higher spatial resolution. Twelve gliomas were identified by both MR modalities. Due to the increased susceptibility assumed microvascularity and microhemorrhages were superiorly visualized in 80% of 10 high-grade gliomas by 7T MRI in comparison to 1.5T MRI. As angiogenesis is one marker for tumor aggressiveness 7T MRI might become a valuable tool for glioma grading and assessment of tumor microvascularity in vivo. 2823. BOLD Response in Pediatric Medulloblastoma Patients During Radiation Treatment Ping Zou1, Thomas Merchant1, Amar Gajjar1, Robert Ogg1 1St. Jude Children's Research Hospital, Memphis, Tennessee, USA Visual fMRI was used to evaluate the acute effect of craniospinal radiation on the BOLD response in children being treated for medulloblastoma. Activation in the primary visual cortex was evaluated at four time-points during the six-week course of radiation therapy (RT). The activated brain volume decreased significantly (p = 0.02) during the first week of RT, and remained low in the following weeks of RT. The change in BOLD response suggests an early acute effect of RT on hemodynamic responses to visual cortex activation. 2824. Evaluation of Reproducibility of FMRI Maps in Patients with Proven Low-Grade Brain NeoplasmsRachel DiAnne McKinsey1, Gabe Heredia2, Sean Fain2, Beth Meyerand2, Wolfgang Tome2 1University of Wisconsin Madison, Madison , USA; 2University of Wisconsin Madison, Madison, USA The purpose of this study is to demonstrate the test re-test reliability of fMRI within subjects with benign and biopsy proven low-grade neoplasms. FMRI activation maps with 3 different thresholds (selected standard and +/- 20%) were generated for 9 subjects with benign and biopsy proven low-grade brain neoplasms using a GE Signa 1.5T magnetic resonance imaging (MRI) scanner. The imaging protocol included a 3DSPGR T1 brain volume, and 8 BOLD-weighted functional scans. Expected fMRI responses were examined in the lateral inferior/middle frontal gyri, bilateral superior temporal gyri, bilateral primary sensorimotor cortices, and posterior occipital pole. |