Coronary MRA

                  935.       Coronary Magnetic Resonance Angiography Using Non Spatially Selective Navigator Excitations at 3T

Harsh K. Agarwal¹, Khaled Z. Abd-Elmoniem¹, Michael Schär¹, ², Sebastian Kelle¹, Matthias Stuber¹, Jerry L. Prince¹

¹Johns Hopkins University, Baltimore, Maryland, USA; ²Phillips Medical System, Best, Netherlands

B0 inhomogeneity, reduced T2* and off-resonant fat excitation may adversely affect the performance of 2D selective navigator technology, particularly at higher magnetic field strength. Furthermore, localization of a 2D selective navigator is user dependent and time consuming. A navigator technique which uses non-spatially selective excitation in conjunction with local surface coil navigator signal reception is proposed to circumvent the aforementioned issues.

Xin Liu¹, Xiaoming Bi, Nondas Leloudas, Renate Jerecic, James C. Carr, Debiao Li

¹Northwestern University, Chicago, Illinois, USA

To compare contrast-enhanced whole-heart coronary MRA at 3.0T and steady-state free precession imaging at 1.5 T, 11 healthy volunteers underwent both 3.0 T and 1.5 T coronary MRA using 3D FLASH with slow infusion of MultiHance and 3D TrueFISP, respectively. SNR, CNR, image quality, and coverage of coronary segments were analyzed and compared statistically. Contrast-enhanced whole-heart coronary MRA at 3.0 T demonstrated higher CNR, less acquisition time, and better depiction of coronary segments compared to non-contrast SSFP coronary MRA at 1.5 T.

Lan Ge¹, Xiaoming Bi¹, Peng Lai¹, Hua Peng¹, Andrew Larson¹, Debiao Li¹

¹Northwestern University, Chicago, Illinois, USA

During contrast-enhanced coronary MRA, the blood signal changes during the contrast injection time will result in image artifacts, blurring and relatively low SNR, when the k-space segments from different cardiac cycles are combined to reconstruct the final image as ¡ time-averaged¡±. Thus, it is important to acquire data during maximal blood signal enhancement from firstmpass, therefore requiring relatively high temporal resolution. This work demonstrated the feasibility of HYPR PR for time-resolved, contrast-enhanced coronary MRA with an increased temporal resolution. HYPR processed coronary artery images around the time of peak blood signal enhancement significantly improved CNR and suppressed artifacts compared to conventional composite images.

Lan Ge¹, Aya Kino¹, Xiaoming Bi², Xin Liu¹, Natasha Berg¹, Renate Jerecic², James Carr¹, Debiao Li¹

¹Northwestern University, Chicago, Illinois, USA; ²Siemens Medical Solutions, Chicago, Illinois, USA

Cardiac MRI is a promising tool to detect and evaluate myocardial ischemic disease. Various MRI techniques have been developed, including functional cine scan, first pass myocardial perfusion (FPMP), coronary MRA, and delayed enhancement imaging. Nevertheless, these techniques have not been applied in the same imaging session for a comprehensive examination. The study demonstrated the feasibility of a comprehensive protocol ¡ One Stop Shop¡± for cardiac MRI at 3T. The ¡ one stop shop¡± cardiac MRI was successfully acquired in 4 volunteers with an average time of 43 minutes.

Shigehide Kuhara¹, Tomohisa Okada², Shotaro Kanao², Ayako Ninomiya¹, Saori Satou¹, Toshikazu Kamae², Kimio Goto², Kaori Togashi²

¹Toshiba Medical Systems, Otawara-shi, Japan; ²Kyoto University Hospital, Kyoto, Japan

We have developed a visual feedback (VFB) system that displays the breathing level to the patient, thus permitting the patient to adjust his or her breathing level. The present study was undertaken to investigate the usefulness of the VFB system in WH MRCA studies, aiming to perform abdominal band-free examinations for clinical use.Using the VFB system, WH MRCA can be performed with less difficulty and without prolonging the scan time, and, in particular, multi-breath-hold WH MRCA with VFB provides the best image quality in the shortest practical time.

Himanshu V. Bhat¹, Sven Zuehlsdorff², Xiaoming Bi², Renate Jerecic², Debiao Li¹

¹Northwestern University, Chicago, Illinois, USA; ²Siemens Medical Solutions, Chicago, Illinois, USA

Contrast enhanced coronary artery magnetic resonance angiography (CMRA) at 3T has recently shown very promising results. Imaging time on the order of five minutes was reported for a whole heart acquisition. Further reduction of the imaging time is required to make CMRA more robust and clinically applicable. Segmented EPI is a method which can be exploited to provide a significant speed gain for CMRA and has previously been reported at 1.5T. In this work the feasibility of 3D segmented EPI for breath-hold high resolution CMRA at 3T has been shown. Segmented EPI is a promising technique for contrast enhanced coronary artery imaging at 3T.

Axel Bornstedt¹, Vinzenz Hombach¹, Marc Kouwenhoven, Volker Rasche¹

¹University of Ulm, Ulm, Germany

Possible improvement of image acquisition time for whole-heart coronary angiography at isotropic spatial resolution by application of a 32-element cardiac coil is presented. It is shown that parallel imaging factors of up to 7.5 can be applied without severe image degradation.

Zhaohui Han¹, Nicholas N. Kuzma¹

¹University of Rochester, Rochester, USA

We report the first confirmed NMR signals from hyperpolarized Xenon-129 (HP Xe-129) dissolved in atherosclerotic plaques of mouse aortas. HP Xe-129 has a broad range of properties that make it a biosensor of choice to characterize biological systems. These properties include high sensitivity to molecular environments, high solubility in biological tissues, and several orders of magnitude increase in NMR signal intensity by optical pumping. Exploring the use of xenon for atherosclerosis diagnostics, we have performed ex-vivo NMR on excised mouse aortas. We have detected characteristic signals from HP Xe-129 dissolved in the atherosclerotic plaques of the apolipoprotein E-deficient mouse.

Thanh D. Nguyen¹, Pascal Spincemaille¹, Matthew D. Cham¹, Jonathan W. Weinsaft¹, Martin R. Prince¹, Yi Wang¹

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

The objective of this study was to compare the performance of diaphragm navigator (DNAV) and cardiac fat navigator (FNAV) in free-breathing SSFP 3D CMRA. Imaging was performed in 16 volunteers at 1.5T using the PAWS real-time gating algorithm. Interpretable CMRA was obtained in all subjects with FNAV gating (0% failure rate) and only 14 subjects with DNAV gating (12% failure rate). Compared to DNAV gating, FNAV gating was found to provide more effective motion suppression, significantly better image quality (P<0.01), and a 30% improvement in average navigator efficiency (P=0.002).

Thanh D. Nguyen¹, Pascal Spincemaille¹, Matthew D. Cham¹, Jonathan W. Weinsaft¹, Martin R. Prince¹, Yi Wang¹

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

The objective of this work was to compare the performance of the restrospective (RETRO), accept/reject (A/R), diminishing variance algorithm (DVA), and phase ordering with automatic windows selection (PAWS) gating techniques in SSFP 3D coronary MRA. Experiments were performed in 10 volunteers without breath coaching and 15 vessels were imaged. PAWS provided significantly better image quality than A/R (P=0.02), DVA (P=0.01) and RETRO (P=0.002). PAWS and DVA were the most efficient algorithms, providing an approximately 20% and 41% higher navigator efficiency compared to A/R (P=0.01) and RETRO (P<0.001).

Sebastian Feuerlein¹, Martin Jeltsch², Oliver Klass², Hans-Juergen Brambs², Martin HK Hoffmann²

¹University of Ulm, Ulm, Germany; ²University of Ulm, Germany

The aim of our study was to investigate whether a modern respiratory biofeedback system using different diaphragm positions and supplemental oxygen could significantly increase navigator efficiency while maintaining image quality compared to conventional respiratory gated MRCA.According to our initial experiences such a Navigator Echo Biofeedback significantly increases navigator efficiency and thereby decreases total imaging time by about 40% compared to a conventional free breathing acquisition strategy.

Myocardial Perfusion

                   995.       Automated Breathing Motion Correction in First-Pass Myocardial Perfusion MRI

Julien Milles¹, Rob J. van der Geest¹, Michael Jerosch Herold², Johan HC Reiber¹, Boudewijn PF Lelieveldt¹

¹Leiden University Medical Center, Leiden, Netherlands; ²Brigham & Women's Hospital, Boston, USA

In this work, we present a fully automatic algorithm for registration of perfusion data that is robust against the large contrast variations during bolus passage, and does not require manual interaction or ROI definition. Validation experiments on 45 MR perfusion studies demonstrate: 1)a high robustness, 2)a substantial reduction in LV center motion after registration, with an average motion of 0.64 ± 0.46 pixel, 3)an increase in the percentage of studies with a motion below 1 pixel from 32% before to 88% after registration and 4)a substantial improvement due to registration of the correlation and NMSE of perfusion curves compared to manually derived perfusion curves. We conclude that the ICA-based registration shows an excellent accuracy, robustness and computation speed, adequate for use in a clinical environment.

Orhan Unal¹, ², Julia Velikina¹, Charles A. Mistretta¹

¹University of Wisconsin, Madison, Wisconsin, USA

In myocardial perfusion imaging, the goal is to differentiate contrast kinetics of normal and ischemic myocardium. The required high temporal resolution does not allow for acquisition of a fully sampled dataset in each time frame. As a result, reconstructed images suffer from loss of spatial resolution and/or low signal-to-noise ratio (SNR) and undersampling artifacts. Recently developed HighlY constrained backPRojection (HYPR)-based techniques in combination with a hybrid radial/Cartesian acquisition can provide relatively artifact free images with high SNR and high temporal resolution for large undersampling factors.

Zungho Zun¹, Eric C. Wong², Krishna S. Nayak¹

¹University of Southern California, Los Angeles, California , USA; ²University of California, San Diego, La Jolla, California , USA

We demonstrate myocardial perfusion imaging at 3 T using arterial spin labeling with background suppression. The measured perfusion rate follows a non-central chi distribution as verified by experiments. The number of averages that provides quantification of myocardial blood flow with a specific percentage confidence interval is roughly 200 times greater than that required for ASL cerebral blood flow measurement with the same confidence. ASL cardiac perfusion imaging is demonstrated in healthy volunteers with measured myocardial blood flow values matching literature values.

Taehoon Shin¹, Houchun Harry Hu¹, Samuel S. Valencerina², Gerald M. Pohost¹, Krishna S. Nayak¹

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

Inadequate spatial coverage of the left ventricle in current 2D multi-slice first-pass myocardial perfusion imaging (MPI) limits the complete assessment of under-perfused volume.  In this work, we investigate 3D MPI, an attractive alternative that provides contiguous volumetric coverage as well as greater signal-to-noise ratio efficiency.  In healthy volunteers, we demonstrate 3D MPI with 2D parallel imaging at 3 Tesla. Complete coverage of the left ventricle in diastole with 3x4.5x10 mm³ spatial resolution is achieved.  Excellent image quality was obtained, allowing semi-quantitative analysis and visualization of perfusion indices across the whole heart.

Uwe Heinrichs¹, Tobias Frauenrath¹, Jane Francis Utting¹, Gabriele A. Krombach¹, Rolf W. Günther¹, Thoralf Niendorf¹

¹University of Aachen, University Hospital, Aachen, Germany

This study examines the feasibility of free breathing, black blood prepared, cardiac gated spin echo based imaging (UFLARE) at 3.0 T to generate BOLD contrast and T₂*-maps of the myocardium, without geometric distortions. Phantom and volunteer studies demonstrate the geometrical integrity and high image quality obtained with UFLARE - even for strong T₂*-weighting. Results show that T₂*-weighted UFLARE may present a realistic alternative to contrast agent studies of myocardial perfusion, which avoids the drawbacks of EPI and gradient echo based imaging. Furthermore, the approach may be extended to map T₂*, quantify myocardial iron content, and assess endothelial function. 

Stefan Weber¹, Andrea Kronfeld¹, R. Peter Kunz¹, Kerstin Münnemann¹, Georg Horstick¹, Karl Friedrich Kreitner¹, Wolfgang G. Schreiber¹

¹Mainz University Medical School, Mainz, Germany

Quantitative first-pass myocardial perfusion imaging was performed using three TSENSE-accelerated (R=2) pulse se-quences (SR-TrueFISP, SR-TurboFLASH and SR-segEPI). Myocardial blood flow (MBF) was calculated using XSIM using the MMID4 model. All calculated MBF values were the in range expected for young healthy volunteers. However, SR-TurboFLASH and SR-segEPI yielded significant smaller MBF values than SR-TrueFISP. Furthermore, MMID4-fit quality was considerably lower using the SR-TurboFLASH or SR-segEPI pulse sequence compared to SR-TrueFISP. Therefore, under the shown experimental setting SR-TrueFISP seems to be the method of choice for quantitative myocar-dial perfusion imaging.

Jens Vogel-Claussen¹, Kakuya Kitagawa¹, Joao A.C. Lima¹, David A. Bluemke¹

¹Johns Hopkins Hospital, Baltimore, Maryland, USA

Traditionally gadolinium contrast is administered on a weight-based dosing scheme (mmol/kg). The purpose of this study was to evaluate the impact of this traditional weight-adjusted dosing scheme on the myocardial blood pool contrast concentration for quantitative analysis of first pass perfusion MRI. Traditional weight-adjusted dosing scheme for quantitative analysis of first pass SSFP perfusion MRI using 0.075mmol/kg does not result in uniform left ventricular blood pool contrast concentration. At doses >11ml gadopentetate dimeglumine the T2 effects appear to artificially decrease the measured LV contrast concentration, which cannot be fully corrected by the test bolus.

Frank Kober¹, Mark Cole², Martine Desrois¹, Carole Lan¹, Patrick J. Cozzone¹, Kieran Clarke², Monique Bernard¹

¹CNRS UMR N  6612, Faculté de Médecine, Université de la Méditerranée, Marseille, France; ²University Laboratory of Physiology, Oxford, UK

Isoproterenol is a beta-adrenoreceptor agonist used in animal models to study the mechanisms of cardiac hypertrophy and failure. In this longitudinal CMR study, time-dependent changes in myocardial perfusion were assessed by arterial spin labeling along with morphologic and functional parameters using multimodal cardiac MRI during continuous administration of IP over 7 days. Multimodal MRI has shown that hypertrophy during IP administration is preceded by strong MBF increase and that cardiac function is maintained on a high level even at day 7. This indicates a mismatch between cardiac function and perfusion that might play a major role in the process of ventricular adaptation.

Glenn Reynolds¹, Micheal Jerosch-Herold², Sandeep N. Gupta³

¹GE Healthcare, Boston, Massachusetts, USA; ²Brigham and Women's Hospital, Boston, Massachusetts, USA; ³GE Global Research Center, Niskayuna, USA

Quantitative techniques, such as Fermi analysis, are used to assess ischemic regions of the myocardium and require Arterial Input Functions (AIF) and Myocardial Response Functions (MRF). First-pass perfusion imaging, using a single bolus of Gadolinium (Gd) contrast agent, results in a linear Myocardial Response Function and a less than linear Arterial Input Function. The purpose of this study is to demonstrate a method to generate a linear AIF using a combination of imaging, calibration, and Bloch simulation.

Dirk Graafen¹, Kerstin Münnemann¹, Stefan Weber¹, Wolfgang G. Schreiber¹

¹Mainz University Medical School, Mainz, Germany

Using Computational Fluid Dynamics methodology dispersion of a contrast agent bolus was simulated in a straight vessel with different stenoses under steady state conditions. Two different perfusion conditions were examined: resting condition (inlet-velocity v = 0.1m/s) and stress condition (constant inlet-pressure p = 1010 Pa) both realizing a myocardial perfusion reserve of 5 in the absence of a stenosis. The dispersion in resting condition is greater than under stress conditions. Therefore, resting myocardial blood flow may be more underestimated in quantitative myocardial perfusion studies than stress perfusion.

Kyle Stephan McCommis¹, Thomas A. Goldstein¹, Robert J. Gropler¹, Jie Zheng¹

¹Washington University School of Medicine, St. Louis, Missouri, USA

A newly developed quantitative MR perfusion technique was evaluated in normal and stenotic dogs, at rest and during Dipyridamole or Dobutamine-induced hyperemia. Myocardial perfusion maps were constructed to assess changes in both myocardial blood flow (MBF) and volume (MBV). Stenosis caused gradual attenuations of both hyperemic MBF and MBV in the stenosis subtended region. Interestingly, these parameters were also attenuated in the normal remote myocardial regions. These effects may imply coronary steal and auto-regulation, but further study is necessary. Measurements of both MBF and MBV may allow for more comprehensive diagnoses of coronary artery stenosis and better treatment planning.

Sven Zuehlsdorff¹, Tongbai Meng², Ying Sun³, Peter Kellman⁴, Jens Guehring⁵, Sonia Nielles-Vallespin⁶, Christine H. Lorenz², Renate Jerecic¹

¹Siemens Medical Solutions USA, Inc., Chicago, USA; ²Siemens Corporate Research, Inc., Baltimore, USA; ³National University of Singapore, Singapore, Singapore; ⁴National Institutes of Health/NHLBI, Bethesda, USA; ⁵Siemens Corporate Research, Inc., Princeton, USA; ⁶Siemens AG Medical Solutions, Erlangen, Germany

The clinical use of first-pass MR myocardial perfusion imaging has shown a significant increase over the last years due to improved image quality and overall increased sequence performance in terms of spatial resolution and speed. Semi-quantitative or quantitative analysis of perfusion images usually requires significant user interaction and expertise that results in prohibitive long times to evaluate myocardial perfusion images. The goal of this work was to implement and test a framework for inline perfusion analysis which completely eliminates the need for user interaction and presents semi-quantitative parameter maps immediately after the scan together with the reconstructed images. The framework was tested on volunteers and patients. The feasibility of a fully automated perfusion analysis was demonstrated.

Nathan Allen Pack¹, ², Edward VR DiBella¹

¹University of Utah, Salt Lake City, USA

A model-independent deconvolution method, which uses iterative minimization and regularization to estimate the impulse response and myocardial blood flow from dynamic contrast-enhanced MRI perfusion images, was evaluated.  Variations in the delay time between blood and tissue enhancement changed flow estimates up to 10%. The inclusion of a vascular blood signal was found to linearly increase estimates of blood flow using this deconvolution method.  The use of model-independent analysis with noisy pixelwise dynamic MRI perfusion data resulted in flow estimates ~12% higher than flow estimates from large (200 pixel) regions.

Myocardial Viability Methods & Applications

Wolfgang G. Rehwald¹, Michael Salerno, Enn-Ling Chen, Robert M. Judd, Raymond J. Kim

¹Siemens Medical Solutions, Chapel Hill, North Carolina, USA

We developed a dark blood delayed enhancement technique for MR imaging of myocardial viability employing a triple IR preparation. We implemented the method on a 1.5T clinical MRI scanner and evaluated it in dogs. The contrast between infarct and blood pool improved dramatically compared to the clinical gold standard (GS) while still providing adequate infarct SNR. Compared to previous double-preparation dark blood techniques timing was more flexible and SNR was improved. The slice-selective inversions and readout could be played when the heart was in a similar position providing a homogeneous preparation and reliable blood signal suppression.

Kai Nassenstein¹, Frank Breuckmann², Christina Bucher¹, Gernot Kaiser², Thomas Konorza², Gerd Heusch³, Joerg Barkhausen¹

¹University Hospital Essen, Essen, Germany; ²Universitiy Hospital Essen, Germany; ³University Hospital Essen, Germany

Detection of structural myocardial abnormalities in non-ischemic diseases is difficult by cardiac MRI, because non-ischemic diseases typically cause multifocal myocardial damages and affect only a small amount of myocardium. To estimate how much myocardial damage is necessary to detect focal myocardial lesions by late enhancement, experimental coronary microembolization was performed in 18 pig as a model for a multifocal myocardial pathology. Our results show, that focal myocardial lesions exceeding 5% of myocardium per section could be detected in vivo by late gadolinium enhancement in 86%.

J.B.M. Warntjes¹, ², J. Kihlberg¹, J. Engvall¹

¹Center for Medical Imaging Science and Visualization (CMIV), Linköping, Sweden; ²Division of Clinical Physiology, Linköping, Sweden

To establish the optimal inversion delay time for inversion recovery images in a Late Gadolinium Enhancement examination is challenging. The optimal delay time depends on many factors such as patient weight, contrast dose and time after contrast injection. A method is presented that allows to visualize LGE images with an inversion delay time that is free to choose. All these images are based on a single breath-hold scan. The procedure is implemented into the PACS visualization software and helps to interactively find the optimal inversion delay for all subsequent MR scans.

Anja Hennemuth¹, Achim Seeger², Ola Friman¹, Stephan Miller², Heinz-Otto Peitgen¹

¹MeVis Research, Bremen, Germany; ²University Hospital Tübingen, Tübingen, Germany

Late enhancement imaging is proven to be suitable for the assessment of myocardial infarctions. The purpose of our work was the development and evaluation of an automatic and thus reproducible method to detect and quantify non-viable tissue in short axis late enhancement images. The developed methods are based on an intensity distribution model combined with a watershed segmentation. For evaluation 15 datasets were segmented by experts, with the common 3s-method and with the new automatic algorithms. Segment-based volume analysis correlated well for automatic and manual segmentations but not for 3s-results. Overlap comparisons also indicate the appropriateness of the developed algorithms. Further improvements are possible by consideration of long axis images

Ola Friman¹, Anja Hennemuth¹, Heinz-Otto Peitgen¹

¹MeVis Research, Bremen, Germany

A Rician-Gaussian mixture model for segmenting scarred tissue in delayed enhancement MRI images is presented. The parameters in the mixture model are fitted using the Expectation-Maximization algorithm, which also is detailed in the abstract. It is shown that the Rician-Gaussian model fits well, as is predicted by theory. However, partial volume effects due to the large voxel size in delayed enhancement images broaden the fitted distributions. Future work to combat the influence of partial volume effects involves extending the mixture model and introducing spatial context via a Markov Random Field.

John J. Sheehan¹, George Lin¹, Jim Conners¹, Mark J. Alberts¹, Karin Dill¹, Reed A. Omary¹, Richard A. Bernstein¹, James C. Carr¹

¹Northwestern University, Chicago, USA

CVMR is a non invasive method for the detection of intracardiac thrombi and is clinically advantageous in the detecting non thrombotic findings, including prothrombotic conditions. 106 patients with a suspected CES had CVMR for the detection of intracardiac thrombi. CVMR revealed 10 thrombi in n=9  patients. In 9 patients echocardiography was positive in n=2, indeter. in n=2 and negative in n=5. Additional findings associated with thrombus formation were n=19 (20%) for CVMR and n=7 (7%) for echocardiography.  CVMR should be considered as part of the routine evaluation with echocardiography in the assessment of patients with suspected CES.

Jason E. Taclas¹, John V. Wylie¹, Reza Nezafat¹, Thomas H. Hauser¹, Mark E. Josephson¹, Warren J. Manning¹, Dana C. Peters¹

¹Beth Israel Deaconess Medical Center, Boston, Massachusetts, USA

Radio frequency pulmonary vein ablation is a treatment for atrial fibrillation which provides an electrical block between the left atrium and the pulmonary veins.  Electroanatomic mapping system CARTOMERGE (Biosense, Webster) records sites of RF application, and registers these sites to angiographic data in real time to help guide the procedure.  After the procedure, late gadolinium enhancement (LGE) cardiovascular MR can be used to image scar generated by RF ablation.  We have developed a tool using Visual Toolkit (Kitware Inc.) to register electroanatomic mapping data to angiographic data, and render it with LGE scar data. 

Robert S. Oakes¹, ², Eugene Kholmovski³, ⁴, Edward V.R. DiBella³, ⁴, Nathan Segerson⁴, Eric Nathaniel Fish⁴, Christopher J. McGann⁴, Rob S. MacLeod², ⁴, Nassir F. Marrouche⁵

¹University of Utah School of Medicine, Salt Lake City, USA; ²University of Utah, Salt Lake City, USA; ³Utah Center for Advanced Imaging Research, Salt Lake City, USA; ⁴University of Utah Schol of Medicine, Salt Lake City, USA; ⁵University of Utah Schol of Medicine, Salt Lake, USA

We report an MRI method to define responders to pulmonary vein isolation (PVI) in patients with atrial fibrillation (AF).  Thirty one patients were imaged using a delayed enhancement MRI scan protocol.  Novel image processing methods were then applied to visualize and quantify the amount of enhancement in two and three dimensions.  Patients with extensive enhancement on MRI were much more likely to suffer recurrence than patients without enhancement.  MRI appears to offer a feasible way to define areas of fibrosis in patients with AF and shows great promise in predicting responders to PVI treatment.

Maythem Saeed¹, Alastair J. Martin¹, Phlilip Ursell¹, Loi Do¹, Matthew Bucknor¹, Charles B. Higgins¹, David Saloner¹

¹University of California San Francisco, San Francisco, California , USA

Cardiac dysfunction in ischemic cardiomyopathy is most likely due to decreased perfusion and increased collagen synthesis. Therefore, an approach that alters perfusion and fibrosis may open up a new therapeutic avenue. Accordingly, we investigated the angiogenic and antifibrotic effects of intramyocardially delivered hepatocyte growth factor (HGF) gene in acute infarction. Perfusion and viability MR imaging was used to non-invasively assess the therapy. Histopathological methods were used to confirm MR findings. Our novel study demonstrated the biological effects of HGF gene in infarcted myocardium. The beneficial effect of HGF gene includes angiogenesis and formation of peninsulas/islands of viable cells in peri-infarcted myocardium. The MR study provides comprehensive assessment of myocardial perfusion, viability and function after gene therapy.

RF Transmit Array Methods

Glen Morrell¹

¹University of Utah, Salt Lake City, USA

A framework is presented for evaluation of coil configurations for transmit SENSE.  For subsampled Cartesian k-space, a matrix A of coil sensitivity values is inverted at each spatial location to calculate the individual per-coil excitation profiles which add to give the desired excitation.  We show through mathematical analysis and Monte Carlo simulation that the condition number of A, evaluated point by point in the spatial domain, is a good measure of the fidelity of excitation that can be achieved by a transmit SENSE coil configuration in the presence of noise in the coil sensitivity profiles.

Geoff Jacobs¹, Abel Assefa¹, Jacob Willig-Onwuachi¹

¹Grinnell College, Grinnell, USA

While Q measurements are used widely to characterize RF coils and the theory involved is well understood, there is surprisingly little literature on how to most accurately perform these measurements.  This project is an initial attempt to compare several common methods of measuring Q with each other and with expected theoretical results.  A basic model for simulation is presented.

Jonathan C. Sharp¹, Scott B. King², Donghui Yin², Vyacheslav Volotovskyy², Boguslaw Tomanek¹

¹National Research Council Institute for Biodiagnostics, Calgary, Canada; ²National Research Council Institute for Biodiagnostics, Winnipeg, Canada

The abstract shows how slice selection can be achieved using only pair of RF coils and a single-channel transmitter system. This method uses two or more uniform amplitude RF volume coils designed with B1 phase gradients. A train of refocusing pulses, applied alternately between the two coils, provides a k-space trajectory. Interleaved small tip angle pulses with a sinc envelope results in slice selection. In addition, the method may be accelerated to reduce the echo train length by use of two or more coils for the small tip angle pulses following each refocusing pulse.

Scott B. King¹, Mike J. Smith¹, Ulrich Fontius², Peter Latta¹, Jarod Matwiy¹, Franz Schmitt², Boguslaw Tomanek¹

¹Institute for Biodiagnostics, Winnipeg, Canada; ²Siemens Medical Solutions, Erlangen, Germany

B1-shimming can be accomplished with hardware-only using a transmit-array with appropriately driven amplitude and phase of each array element, or using 2D/3D RF spatially selective pulses.  On a new Siemens 8-channel transmitter 3T Trio-Tim MRI system, we have shown that B1-shimming using the hardware approach using an azimuthal distribution of array elements is unable to adequately achieve uniform excitation and therefore future Tx-array coils will require elements to be placed along the axis of the coil as well.  2D/3D RF pulses can effectively produce uniform excitation but require Tx-arrays (Tx-SENSE) to minimize acceptable pulse lengths.

Hellmut Merkle¹, Shumin Wang¹, Peter van Gelderen¹, Tie Q. Li¹, Joseph Murphy-Boesch¹, Alan P. Koretsky¹, Jeff H. Duyn¹

¹National Institutes of Health, Bethesda, Maryland, USA

Multiple resonant loop elements placed within the interior of a circularly polarized volume transmit coil have been used to modify the field profile over a conductive phantom.  The phantom has been filled with low dielectric material and saline to approximate the dielectric and conductive properties of the head.  A B1 mapping sequence has been used to provide quantitative information about the modified transmit B1 field.

Mike J. Smith¹, Scott B. King¹, Peter Latta¹, Jarod Matwiy¹, Ulrich Fontius², Franz Schmitt², Boguslaw Tomanek¹

¹National Research Council Institute for Biodiagnostics, Winnipeg, Canada; ²Siemens Medical Solutions, Erlangen, Germany

Transmit array systems can generate better 2D excitation profiles in shorter time than single channel transmitters. We used our new 3T Siemens 8-channel transmit system in enhancing parallel imaging when the sample exceeds the ROI. When excitation artifact signals are below the image noise level, g-factor maps are significantly improved, and effective reduction factors larger than the receive array are possible.

Yeun Chul Ryu¹, Jung Ho Hyun¹, Jong Soek Oh¹, Yong Gwon Kim¹, Soo Yeol Lee², Heung Kyu Lee³, Chang Hyun Oh¹

¹Korea University, Seoul, Republic of Korea; ²Kyung-Hee University, Suwon, Republic of Korea; ³KAIST, Dae-jeon, Republic of Korea

A Subject-dependent rapid RF field mapping method for parallel RF transmission.

Tsung-Hui Chang¹, Zhi-Quan Luo¹, Xiaoping Wu², Can Akgun², Thomas Vaughan², Kamil Ugurbil², ³, Pierre-Francois Van de Moortele²

¹University of Minnesota, Minneapolis, Minnesota, USA; ²University of Minnesota Medical School, Minneapolis, Minnesota, USA; ³Max Planck Institute for Biological Cybernetics, Germany

There has been an increasing interest in constraining transmit B1 shimming with specific absorption rate (SAR) limits, especially at high magnetic field. Since most of the existing methods rely on solving a nonconvex optimization problem, they are typically faced with two difficulties: Only local optimum solutions are obtained, and they are susceptible to the chosen initial points for optimization. Here we introduce a two stage optimization method where a reliable initial point is acquired in the first stage by a convex semidefinite relaxation (SDR) approximation method. A high quality B1 shimmed map then can be obtained in the second stage optimization using the SDR initial points. The presented technique is verified with simulations for a 16-channel transmit coil array at 7T with a human head model.

RF Systems & Components

Bernd Stoeckel¹, Andreas Potthast¹, Niels Oesingmann¹, Daniel Sodickson², Ray Lee², Davide Santoro², Glyn Johnson², Thomas Heumann³, Thomas Arnold³, Michael Wullenweber³, Matthias Buettner⁴

¹Siemens Medical Solutions USA, Inc., New York, New York, USA; ²Center for Biomedical Imaging, New York, New York, USA; ³Siemens Medical Solutions, Inc., Erlangen, Germany; ⁴Astrum IT, Erlangen, Germany

We present the expansion of a commercially available Siemens 3T MR scanner to a system, which enables the online acquisition and reconstruction of images from up to 128 receive channels and which offers multi-nuclear capabilities. The combination of hyperpolarization and a very high number of receive elements is very interesting for two reasons: The initially available SNR is increased and the SNR loss traditionally associated with accelerated parallel imaging may be mitigated or even eliminated. As an example first volunteer lung images with hyperpolarized helium are shown which were acquired with a 128 element receive array at ³He frequency.

Stephan Biber¹, Peter Baureis², Jan Bollenbeck¹, Phillip Höcht¹, Hubertus Fischer¹

¹Siemens Medical Solutions, Erlangen, Germany; ²University of Applied Sciences Wuerzburg Schweinfurt, Germany

A new prototype system for analog optical signal transmission between local coil and the receiver is presented in this paper. The system allows to transmit 4 MRI receiver signals on the same optical fiber and implements an optical link with high dynamic range using low cost vertical cavity surface emitting laser diodes (VCSEL). The advantage of an optical transmission concept is the fact that large bundles of copper cables including cable traps for shield wave suppression can become redundant.

Gary X. Shen¹, Juan Wei, Yong Pang

¹The University of Hong Kong, Hong Kong, People's Republic of China

This work investigates MRI application using the recent developed technology of WLAN 802.11n (draft) for multiple RF channel EPI. The number of RF channel, signal dynamic range, bandwidth, ADC sampling rate, direct digital synthesizer and digital signal processing control are analyzed and discussed in detail.  FDM modulation is used to increase the number of RF channels before ADC. The theoretical analysis and bench test results show that 802.11n can be used for ultra-fast EPI (20 images/s) with 64 RF channels. The MR signal dynamic range could be up to 120 dB.

Klaus Markus Huber¹, Martin Hemmerlein¹, Stephan Biber¹, Ralph Oppelt¹, Karsten Wicklow¹

¹Siemens AG, Erlangen, Germany

The new 7T preamplifier is capable of handling input powers of up to -15.6dBm although consuming less than 250mW of dc power. Its noise figure is as low as 0.6dB. An integrated cable trap suppresses common mode signals and thus makes the design of high channel array coils much easier. The actual size of the amplifier PCB is only about 41mm x 18mm. With a first 3-channel array equipped with the new preamplifiers, outstanding signal to noise ratios have been achieved in a Siemens 7T MRI-system.

Cecilia Possanzini¹, Marco Boutelje¹

¹Philips Medical Systems, Best, Netherlands

In this paper, we show the influence of magnetic field on preamplifiers carrying a GaAs FET (field effect transistor). The S-parameters of the amplifier are measured with magnetic field parallel and perpendicular to the carriers in the FET and compared with data at zero magnetic field. The difference in behaviour of the preamplifier with and without magnetic field can be explained by the occurrence of Hall effect.

Victor Taracila¹, Keerthi Shet², Fraser Robb¹

¹GE Healthcare, Aurora, USA; ²Ohio State University, Columbus, USA

The most modern decoupling method is the preamp decoupling, which consists in bringing a low impedance preamp as close as possible to the coil element and inserting it into a parallel resonance trap which itself is in series with the coils element. This method allows decoupling of the coil elements placed far from each other, which exhibits week coupling and cannot be decoupled by other means. Although there are many interpretation of this techniques based on transmission line theory, we consider that an eigenvalue interpretation is closer to the multiple-degree electrical oscillator, which the MR RF coil represents par excellence.

Victor Taracila¹, Fraser Robb¹

¹GE Healthcare, Aurora, USA

At high frequency the current flows only on the surface of the conductor due to the well known skin effect, the excess of copper will not lower the noise coming from the coil itself. Moreover, very large copper strips will shield the subject (object) under study from transmit RF field introducing significant B1 field distortion. This problem can be solved for one dimension and even two dimensions for a rectangular cross-section using complicate mathematical methods. Presented method is simpler and can easily be applied to any cross section of the copper strip.

Patrick J. Smith¹, Dario Mager¹, Ute Loeffelmann¹, Jan G. Korvink¹

¹University of Freiburg, Freiburg, Germany

To the knowledge of the authors, inkjet printing has not yet been used to produce MRI RF transmit and receive coils, possibly because of the variability of the fabrication technology, and the lack of low resistivity metal-containing inks. In this paper, we discuss the questions and responses facing an inkjet-based processing route for MRI coils, and report on the experimental steps that have been undertaken to demonstrate the veracity of the answers. 

Ronald D. Watkins¹, William Doherty², Pavel Voskoboynik

¹Stanford Medical School, stanford, California , USA; ²MicroSemi Corp, Lowell, USA

Authors here have developed and demonstrated greatly improved RFPIN Diode components for reducingsusceptibility artifacts at high field

Victor Taracila¹, Fraser Robb¹

¹GE Healthcare, Aurora, USA

A receive RF coils contains material like Kapton™ tape, Delrin™ rods, Polycarbonate, Nylon etc., which do contain a certain amount of proton signal. The ideal situation would be to eliminate completely all Hydrogen from all building components. However, it could be costly and sometimes unnecessary if the amount and relaxation times T1 and T2 are correlated with existing gradient slew rates, RF pulse width and the very matter we intend to image – living organisms.

Refaat E. Gabr¹, Michael Schar¹, ², Arthur D. Edelstein³, Paul A. Bottomley¹, William A. Edelstein¹

¹Johns Hopkins University, Baltimore, Maryland, USA; ²Philips Medical Systems, Cleveland, Ohio, USA; ³University of California, Berkeley, Berkeley, California , USA

We have undertaken MRI dynamic range (DR) measurements in order to determine compatibility of MR multicoil images with fiber optic signal transmission. Our results indicate that such links should be viable at 3T and possibly higher.

MR-Guided Interventions (Non-Thermotherapy)

                  1197.     Cardiovascular MR Imaging is a Platform for Percutaneous Transendocardial Delivery and Assessment
                                of Gene Therapy

Maythem Saeed¹, Alastair J. Martin¹, Alexis Jacquier¹, Loi Do¹, Matthew Bucknor¹, Charles B. Higgins¹, David Saloner¹

¹University of California San Francisco, San Francisco, California , USA

MR fluoroscopy was used for guiding transendocardial delivery of plasmid VEGF and for evaluating myocardial perfusion, viability and function in occlusive infarction. A mixture of plasmid VEGF and LacZ or plasmid LacZ and Dy-DTPA-BMA were delivered into the border and core of MR hyperenhanced infarcted region. At 7-8 wks, plasmid VEGF increased ejection fraction, perfusion, vascularity and decreased infarcted region on MR imaging and histochemical staining. MR imaging was successfully used in guiding delivery of genes and assessing myocardial viability and function. The effectiveness of this approach most likely stems from VEGF effects on neovascularization in scar tissue.

David Arthur Woodrum¹, Thomas Link², Wesley D. Gilson², Christine H. Lorenz², Dara L. Kraitchman², Jeff WM Bulte², Aravind Arepally²

¹Mayo Clinic, Rochester, Minnesota, USA; ²Johns Hopkins University School of Medicine, Baltimore, Maryland, USA

Magnetic resonance (MR)-trackable magnetocapsules (MCs) were created to simultaneously immunoprotect pancreatic beta cells and non-invasively monitor, in real-time, portal vein delivery and engraftment using MR imaging (MRI). The purpose of this study was to demonstrate the feasibility of MR-guided targeted biopsy of MCs containing human islets to assess viability and function.  MR-compatible 14g biopsy needle was tracked (using a passive artifact from the needle) and steered into the four quadrants of the  liver using  a real-time sequence through a percutaneous access. By utilizing MRI, needle placement close to the MCs was ensured and core biopsies were obtained.

Sherif G. Nour¹, ², Jens O. Heidenreich¹, Jamal J. Derakhshan, Simi Paul¹, Fadi W. Abdul-Karim¹, Mark A. Griswold¹, Vikas Gulani¹, John Jesberger, Philip A. Linden¹, Jeffrey L. Duerk¹

¹University Hospitals Case Medical center / Case Western Reserve University, Cleveland, Ohio, USA

Esophageal cancer affects 5/100,000 people in the USA.  5-year survival rate is <10% despite therapeutic advances. Accurate staging and appropriate treatment rely on determining the depth of tumor invasion through the esophageal wall. Here, a catheter-mounted receiver coil was used to obtain in-vivo high-resolution images of esophageal wall. The performance of various MR-pulse sequences was tested, imaged segments were harvested, and images were correlated with histology. Results show that high-resolution esophageal wall MRI is feasible with commercially-available imaging catheters and there is high correlation with histology in delineating various layers of normal esophageal wall as required for future MR staging.

Jaane Rauschenberg¹, Wolfhard Semmler¹, Michael Bock¹

¹Deutsches Krebsforschungszentrum (dkfz), Heidelberg, Germany

In percutaneous MR-guided interventions imaging can be significantly accelerated by inner volume excitation or outer volume suppression, which restricts the phase encoding FOV. The combination of outer volume suppression with steady state pulse sequences such as PSIF is challenging, as conventional suppression pulses lead to a violation of the steady state condition. Here, a new integrated outer volume suppression technique, which maintains the steady state, is presented and evaluated in phantom and volunteer experiments.

Makiya Matsumoto¹, Yuu Shoji¹, Yuichiro Matsuoka², Etsuko Kumamoto¹, Kagayaki Kuroda², ³, Toshiya Kaihara¹

¹Kobe University, Kobe, Japan; ²Institute of Biomedical Research and Innovation, Kobe, Japan; ³Tokai University, Hiratsuka, Japan

We have devised image fusion techniques for the integrated MR-endoscope system, in which real time MR scanning is performed with an intraluminal surface coil during endoscopic observation and surgery for gastrointestinal applications. To navigate the scope and to match the scope coordinates with that of the scanner, wireless resonators with Gadolinium-absorbed gel were developed. For the image fusion, the MR images were deformed numerically with a fifth order, two dimensional approximation function obtained from the optical distortion of the scope view, after appropriate Affine-transformations for scaling and trapezoidal conversion were applied.

Ann-Kathrin Homagk¹, Sven Müller¹, Peter Hallscheidt², Marc-Andre Weber¹, Wolfhard Semmler¹, Michael Bock¹

¹Deutsches Krebsforschungszentrum (dkfz), Heidelberg, Germany; ²University Heidelberg, Heidelberg, Germany

The present study proposes a method of acquiring intra-arterial images using additional MR tracking information. An imaging and tracking pulse sequence was implemented and tested on a healthy anesthetized pig. The position of the tracking catheter tip was continuously extracted from projection data to use them for retrospective gating of the acquired images. For motion compensation, a range of positions was defined which were accepted as input for the subsequent image calculation. A comparison between motion-corrected images and uncorrected images showed that the use of projection data for motion correction leads to a significant improvement in image quality.

Orhan Unal¹, ², Julia Velikina¹, Charles M. Mistretta¹

¹University of Wisconsin, Madison, Wisconsin, USA

Catheter tracking and visualization require both good spatial and temporal resolution.  Recently developed HighlY constrained back PRojection (HYPR)-based techniques in combination with radial or hybrid radial/Cartesian acquisition techniques provides relatively artifact free images with large undersampling factors and is therefore well-suited for  MRI-guided passive catheter tracking and visualization applications.

Haydar Celik¹, Michael A. Guttman², Onur Kocaturk², Christina Saikus², Kanishka Ratnayak², Anthony Faranesh², Andrew Derbyshire², Robert Lederman², Ergin Atalar¹, ³

¹Bilkent University, Ankara, Turkey; ²National Institude of Health, Bethesda, Maryland, USA; ³Johns Hopkins University, Baltimore, Maryland, USA

In MRI-guided vascular interventions, visualization of interventional devices is rather difficult. For this purpose many tracking techniques have been developed. In one of these studies, reverse polarized signal was acquired using a receive-only birdcage coil in order to separate the anatomical information from a catheter which contains a receive coupled RF (RCRF) coil.In present work, the reverse polarization method has been implemented to phased array coils and real-time experiments have been conducted on phantom using TSENSE  algorithm. As a result, reverse polarization mode of image, which consists of the catheter, and color-coded image are obtained. In order to show safety profile, phantom heating experiments were conducted.

Ioannis Koktzoglou¹, ², Sotirios A. Tsaftaris³, Sven Zuehlsdorff⁴, Debiao Li², Aggelos K. Katsaggelos³, Rohan Dharmakumar²

¹Evanston Northwestern Healthcare, Evanston, Illinois, USA; ²Feinberg School of Medicine, Northwestern University, Chicago, Illinois, USA; ³Northwestern University, Evanston, Illinois, USA; ⁴Siemens Medical Solutions, Chicago, Illinois, USA

Automated tracking of a passive stiletto catheter for regenerative myocardial therapy under the MR environment may improve the accuracy ofthe procedure. We report successful implementation of automated computer-assisted tracking for this purpose in a controlled phantom study.

Donald W. McRobbie¹, Marc Rea¹

¹Imperial College Healthcare NHS Trust, Charing Cross Hospital, UK

Inductively coupled fiducial markers have an inconsistent appearance on MR images.The Non-Excitory Steady-State Interference Elimination (NESSIE) pulse sequence has been developed to remove this imaging sequence dependence. The contrast of the fiducials can be controlled independently from the background anatomy using NESSIE pre-pulses, potentially improving MR tracking during interventions.

Juergen Rahmer¹, Sascha Krueger¹

¹Philips Research Europe, Hamburg, Germany

Increasingly, interventional procedures rely on MR imaging for tracking of catheters in the body. Active catheter visualization allows highlighting the tip position in a standard MR image, but requires additional equipment and dedicated catheters. In contrast, passive visualization requires only minimal device modification. Typically, susceptibility markers are attached to the device to cause local contrast in the image. However, this contrast effect is permanent and can negatively affect the anatomical information in the MR image. We propose to use short-T2 material with T2 shorter than about 1 ms as a passive marker. This material does not interfere with standard MR imaging, but can be visualized using ultrashort echo-time (UTE) sequences. This work demonstrates the passive visualization of the full catheter length using 3D UTE imaging of short-T2 rubber bands inserted into the catheter.

Reiner Umathum¹, Michael Bock¹

¹Deutsches Krebsforschungszentrum (dkfz), Heidelberg, Germany

Active device tracking is challenging in anatomical areas such as the lung where no MR signal is present. To overcome this limitation, a signal reservoir needs to be integrated into the tracking coil; however, these reservoirs require frequent re-filling. In this work we propose semi-solid rubber materials for MR tracking which are long-term stable. Due to their short T2* relaxation times, pulse sequences with ultra-short TE below 1 ms need to be used for device tracking.

Stephan Fandrey¹, Steffen Weiss², Jörg Müller¹

¹Hamburg University of Technology, Hamburg, Germany; ²Philips Research Europe , Hamburg, Germany

A modified micro Helmholtz coil was designed for intravascular imaging and tracking with the objective to achieve maximum homogeneity of the coil sensitivity in the surrounding medium. The optimal design was determined using finite element simulations. Foil-based micro-Helmholtz-coils were fabricated with this design using planar micro systems technology and mounted on a 5F catheter tube. The sensitivity pattern of the coils was evaluated in phantom experiments and found to be conform with the simulation results. The use of micro systems technology allows coil fabrication in large batches with high reproducibility.

Lizabeth Y. Li¹, William R. Overall¹, Greig C. Scott¹, John M. Pauly¹

¹Stanford University, Stanford, California , USA

We apply the use of a grounding pad in combination with a wire receiver antenna to increase signal intensity in front of the wire tip for use in interventional MRI, offering a significant improvement in look-ahead visualization over a guidewire receiver antenna only.  The ground pad directs the current density off of the wire, increasing the receiver B1 field and the resulting signal. With different positioning of the return current path from the ground pad, we can also selectively increase the signal intensity at locations near the front of the guidewire.

Christian Seebauer¹, Florian Wichlas¹, Jens Rump¹, Jens Pinkernelle¹, Ioannis Papanikolaou¹, Tobias Jung¹, Sascha Chopra¹, Ulf Teichgräber¹, Hermann Josef Bail¹

¹Charité, Berlin, Germany

Osteochondrosis dissecans is a localized subchondral aseptic bone necrosis. However, its pathogenesis is still controversial. Various operative techniques for the treatment of osteochondral lesions of the talus with varying success have been reported. MRI yields additional information on the vitality and stability of the osteochondral fragment and allows differentiating the surrounding tissue, providing information on the fixation of the fragment. Here, we propose an innovative method for the retrograde drilling of necrotic areas in the talus using a MR-compatible drilling guide.

Matthew Smith¹, Xu Zhai¹, Ray Harter², Sean Fain¹

¹University of Wisconsin, Madison, Wisconsin, USA; ²Marvel Medtech LLC, Madison, Wisconsin, USA

The performance of image-guided devices for breast procedures is dependent on how well the tissue is stabilized during interventions. Novel MR-compatible devices for breast interventions have been previously developed by various groups using conventional 2D compression plates for breast tissue immobilization during a procedure. However, these devices typically distort the anatomy and cause discomfort for many women. We find that tissue is sufficiently immobilized during a trocar insertion with a 3D tissue immobilization concept using circumferential air bladders compared to an insertion without immobilization. Patient studies are planned to optimize the bladder shape and evaluate the required stabilization pressure.

Zion Tsz Ho Tse¹, Abbi Hamed¹, Michael Lamperth¹

¹Imperial College London, London, UK

In this research piezoelectric sensors are utilized for force sensing and developed into an array for the purpose of inner-body palpation using an endoscope system. The sensors are shown to be MR compatible and highly sensitive and will output a charge signal on the slightest deformation (micron scale), however this property and their highly capacitive nature also entail that they are very sensitive to interference from the noisy scanner environment and hence significant attention must be paid into attenuating unwanted signals before they can present a serviceable transduction solution.

Harald Busse¹, Robert Trampel¹, ², Wilfried Gründer³, Nikita Garnov³, Jochen Fuchs¹, Tim-Ole Petersen¹, Thomas Kahn¹, Michael Moche¹

¹Leipzig University Hospital, Leipzig, Germany; ²Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany; ³Leipzig University, Leipzig, Germany

The ongoing development of powerful MR imaging techniques also allows for advanced possibilities to guide and control minimally invasive interventions. Various navigation concepts have already been described for practically all regions of the body. Most diagnostic scanners, however, do not allow the physician to guide the instrument inside the magnet and, consequently, the patient needs to be moved out of the bore. The purpose of this work was to present a concept for real-time navigation with automatic patient registration and interventional control for a closed-bore scanner and to provide first estimates on the overall targeting accuracy in an experimental setup.

Marc A. Rea¹, ², H Elhawary², Zion T. Tse², Donald W. McRobbie, ¹², Michael Lamperth², Ian Young²

¹Charing Cross Hospital, London, UK; ²Imperial College London, London, UK

This abstract reports the implementation of an imaging methodology enabling real-time 3-dimensional tracking of devices using only passive micro-coil fiducials and image processing. A modified FLASH sequnce was executed with a minimum update rate of 1.8 seconds, sufficient for tracking movements up to 7mm/s. Using a MRI-compatible manipulator, prostate phantom lesions were successfully targeted with a maximum accuracy of 0.5mm.

Bernhard C. Meyer¹, Alexander B. Brost², Liron Yatziv³, Norbert Strobel⁴, Wesley Gilson⁵, Karl Juergen Wolf¹, Jonathan S. Lewin⁶, Frank K. Wacker, ¹⁶

¹Charite - Universitätsmedizin Berlin, Berlin, Germany; ²Universität Erlangen, Erlangen, Germany; ³Siemens Corporate Research, Princeton, New Jersey, USA; ⁴Siemens Medical Solutions, Forchheim, Germany; ⁵Siemens Corporate Research, Baltimore, Maryland, USA; ⁶Johns Hopkins University, Baltimore, Maryland, USA

Purpose: To test the use MR images augmented on C-arm CT images to guide needle punctures in phantoms. Method: 10 punctures with 6 mm radius ring targets and 2 punctures of 15-20 mm cysts were performed. MR based fluoroscopic guidance was used.Result: The needle tip was correctly placed in 8/10 rings with a maximum distance to the center of 5mm. The cyst punctures were successful with a needle deviation of 10 and 8 mm.Conclusion: The combination of preacquired MR images, C-arm CT and fluoroscopy allows puncture of lesions not visible with fluoroscopy and hardly accessible using MR.

Margarita Con Lima¹, James Bean, Thomas Collins

¹Lima Institute for Advanced Studies, Lima, Peru

An advanced, MR compatible communication device for interventional MR is presented.  It is made of space age polymers and fibers, and can transmit audio data up to three meters with only 98% losses.

B1 Mapping & Correction

Kelvin Chow¹, Ian Paterson¹, Richard Thompson¹

¹University of Alberta, Edmonton, Canada

The double angle method for mapping radiofrequency (B₁) fields was adapted for use in the heart and lungs using the HASTE pulse sequence and free breathing acquisition.  At 1.5T, significant heterogeneity was found within the chest, with flip angle variations >20  between the right and left heart, as well as ~10  between the base and apex.  The right lung experiences an average flip angle ~20  less than the left lung, with the largest spatial gradients around the heart.  The HASTE approach to mapping B₁ fields is applicable to other regions and higher field strengths (with minor modifications).

Kim Shultz¹, Greig Scott¹, John Pauly¹

¹Stanford University, Stanford, California , USA

RF current density imaging, useful for ablation treatment planning and RF safety evaluation, requires mapping a large range of B₁ magnitudes in the presence of strong off-resonance effects. Standard B₁ mapping methods like the double angle method require many repetitions to achieve the necessary dynamic range and lose accuracy in B₁ phase measurement as off-resonance effects increase. An adiabatic excitation that is not brought fully to on-resonance creates transverse magnetization approximately linearly dependent on B₁ magnitude over a large range of B₁ values. Off-resonance effects can be corrected using a second excitation with a negative frequency offset. 

Trevor Wade¹, ², Brian Rutt¹

¹Robarts Research Institute, London, Canada; ²University of Western Ontario, London, Canada

A new method for B1 mapping is introduced that takes advantage of fast 3D imaging at low flip angles.  By sampling the inversion recovery curve using two different, small tip angle RF pulses, an expression is obtained for both a corrected T1 and the achieved flip angle. This results in a B1 mapping technique that is very fast and efficient at low flip angles and does not require long TR values.

Hans-Peter Fautz¹, Mika Vogel¹, Patrick Gross¹, Adam Kerr², Yudong Zhu³

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

A composite technique for B1 mapping of transmit coil arrays is proposed that uses all coil elements for signal excitation whereas the flip angle produced by the individual coils is encoded using a magnetization preparation. Only one flip angle encoding scan is required per coil element plus one reference scan that is used for the calibration of all coils. The dynamic range of the excitation pulse of the imaging part is reduced over the FOV allowing a significant increase in precision with which low B1 field amplitudes from single coils can be determined.

Alexis Amadon¹, Nicolas Boulant²

¹CEA-Saclay, Gif-sur-Yvette, France; ²Siemens Medical Solutions, Saint-Denis, France

In high-field MRI, more and more sequences make use of the prior knowledge of B0- and B1-maps of the object under study. This abstract describes a fast method to simultaneously measure 3D maps of the B0 and B1 fields. Examples of brain maps acquired with this method are given at 3T.

Houchun Harry Hu¹, Kyunghyun Sung¹, Krishna S. Nayak¹

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

A retrospective approach to compensate variations in signal intensity due to RF transmit and receive inhomogeneity is described for proton-density-weighted abdominal imaging.  The model utilizes rapid, low-resolution acquisitions to measure the RF transmit and receive fields.  Corrections for RF-induced signal shading are not only helpful in improving image quality, but are essential in applications requiring signal-intensity-based quantification.  We hypothesize that this method can accurately remove signal intensity non-uniformity typically encountered in multi-coil abdominal imaging, and demonstrate its performance at 3 Tesla.  Signal non-uniformity was reduced from 46% to 4% and 25% to 15% in phantom and in vivo experiments, respectively.

RF Pulse Design

Sonal Josan¹, ², Elena Kaye², John Pauly¹, Bruce Daniel², Kim Butts Pauly²

¹Stanford University, Stanford, California , USA; ²Stanford University, USA

Half RF excitation pulses used in ultrashort T2 imaging are very sensitive to eddy currents which distort its slice profile. The purpose of this work is to improve slice selectivity of the half RF in the presence of eddy currents, by using quadratic phase RF saturation to suppress any out of slice magnetization, thus providing a simple robust method for accurate T2* quantitation.

Takayuki Abe¹, Takeuchi Hiroyuki¹, Takahashi Tetsuhiko¹

¹Hitachi Medical Corporation, Kashiwa, Japan

Robust fatsat is important in clinical applications. With 1.5-T or higher MRI scanners, an adiabatic inversion pulse is used for reducing the residual fat shown on an image caused by B1 inhomogeneity. Since this pulse is a 180‹-one, a long TI is required resulting in a dead time of the measurement. So, we developed a new fast fatsat pulse called H-Sinc, which excites at a flip angle near 90‹ and is insensitive to B1 inhomogeneity. Because H-Sinc can easily induce an arbitrary FA and does not require TI, it shows faster fat suppression than that of the adiabatic inversion pulse.

Randy Lee Tyson¹, Jonathan Sharp¹, Boguslaw Tomanek¹

¹National Research Council of Canada, Calgary, Canada

Low-ripple RF pulse design through numerical optimization using the coefficients of the discrete cosine or Fourier transform of the SINC waveform as initial parameters. Bloch simulation and optimization of waveform parameters fitted to a target slice profile results in decreased rippling in both the pass and stopbands.

Joëlle Karine Barral¹, John Mark Pauly¹, Dwight George Nishimura¹

¹Stanford University, Stanford, California , USA

The Shinnar-Le Roux algorithm has been widely used to reduce the pulse design problem to a well-known finite impulse response digital filter design problem when a rectangular slice profile is desired. However, there is no intrinsic limitation to such a profile. An adaptation of the SLR algorithm is proposed, which allows the design of large flip-angle pulses producing arbitrarily-shaped excited slice profiles accurate in both, magnitude and phase. This design should benefit alternative schemes like wavelet encoding which so far suffer from magnitude and phase distortions of the excited magnetization profiles, especially when SNR-efficient large flip-angles are desired.

Ziqi Sun¹, Jay L. Zweier¹

¹The Ohio State University, Columbus, USA

A selective adiabatic full passage (AFP) pulse train, consisted of two 90 degree hyperbolic secant (HS1_R20) pulses of alternate frequency sweep (AFS) at the two ends of the pulse train, and two low power HS1_R20 AFP pulses in the middle, was developed for spin refocusing in a 3D spin echo sequence. In comparison to an amplitude-modulated refocusing pulse, AFP-AFS pulse train substantially improved signal sensitivity and uniformity, which is ascribed to the effective compensation of the nonlinear phase dispersion and flip-angle errors associated with selective AFP pulses. This interpretation is supported by the theoretical analysis and Bloch equation simulation results.

Carsten Warmuth¹, Robert Krieg¹

¹Siemens Medical Solutions, Erlangen, Germany

A temporal mismatch between gradient and RF waveforms causes strong N/2 ghosting in two-dimensional excitations when using echo-planar k-space trajectories. We implemented a quick delay calibration approach preceding each sequence using 2D RF respiratory navigators.

Martin G. Busch¹, ², Jürgen Finsterbusch¹, ²

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

This work presents 2D-selective RF excitations based on the PROPELLER trajectory. The trajectory consists of segments of parallel lines which are rotated to one another. The center of k-space is covered by all segments which yields large flip angles for all segments. Profile blurring in the presence of off-resonance effects like chemical shift or magnetic field inhomogeneities is avoided by using non-selective refocusing RF pulses between the individual lines of each segment.

Nathaniel James Powell¹, Malgorzata Marjanska¹, Julien Valette², Pierre-Gilles Henry¹, Michael Garwood¹

¹University of Minnesota, Minneapolis, USA; ²CEA-neurospin, Gif-sur-Yvette, France

Outer Volume Suppression (OVS) is sometimes needed in MR imaging and, more often, in spectroscopy.  MR Spectroscopic Imaging (MRSI) studies of the brain and other anatomical regions generally require some form of OVS to ensure that the strong lipid signals from the layer of skin and subcutaneous fat do not interfere with the desired signals from deeper tissues of interest.  Typically when OVS is used in brain MRSI, the standard approach involves sequentially applying multiple slice-selective (1D) suppression pulses at different angles around the periphery of the object, an approach that has drawbacks in terms of efficiency and Specific Absorption Rate (SAR).  Our method employs a single two-dimensional pulse to suppress an elliptically shaped annulus in one shot, saving both time and SAR, and in some cases providing a suppression pattern that more closely matches the anatomy.

SSFP Methods

Brian A. Hargreaves¹

¹Stanford University, Stanford, California , USA

Balanced SSFP imaging sequences provide high signal-to-noise, rapid imaging, and diagnostically useful contrast, but their clinical utility is limited primarily by the dark-band artifacts resulting from off-resonance effects, especially at 3.0T.  This work demonstrates that using an unbalanced gradient immediately before and after the readout, combined with a modified reconstruction, a signal with balanced SSFP characteristics and no dark band artifacts can be obtained.  The technique is demonstrated in phantoms and human scans.

Hsu-Lei Lee¹, Krishna S. Nayak¹

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

An efficient initial preparation is critical for reducing the transient signal fluctuation in non-continuous steady-state imaging.  For conventional SSFP, LeRoux derived a Fourier relation between RF amplitude increments and the resulted oscillatory residues. In alternating TR SSFP the Fourier relation is also altered and the preparation scheme has to be modified. We utilized the SU2 formalism to build a simplified model for alternating TR SSFP sequences and present a design method that can be applied to arbitrary repetition times and RF phase cycling combinations. This approach is used to design stabilizer sequences for ATR-SSFP and wideband SSFP.

Jochen Leupold¹, Jürgen Hennig¹, Klaus Scheffler²

¹University Hospital Freiburg, Dept. of Diagnostic Radiology, Medical Physics, Freiburg, Germany; ²MR Physics, Department of Medical Radiology, University Hospital of Basel, Basel, Switzerland

RF-spoiled gradient echo sequences are widely used in clinical MRI. However, the theoretical description of the measured signal is not trivial, as a pseudo steady state of the voxel magnetisation is built up. We demonstrate that the well known occasional occurrence of ghost artifacts is a direct consequence of the existence of the PSS, and that there is no rigid rule for the needed moment of the spoiler gradient of an RF spoiled gradient echo sequence.

Tolga Çukur¹, Michael Lustig¹, Dwight Georger Nishimura¹

¹Stanford University, Stanford, California , USA

Balanced steady-state free precession (SSFP) imaging offers high SNR efficiency within short scan times, but suffers from banding artifacts in the presence of strong field inhomogeneities. A common approach is to combine multiple SSFP images with different phase-cycling schemes. There is an inherent trade-off between SNR and the level of banding artifact reduction for most combination methods. In this work, accurate estimates of the SSFP sensitivities are obtained from multiple phase-cycled acquisitions using the magnitude-weighted complex-sum combination as an initial reference. Banding-free images with true SSFP contrast are reconstructed without sacrificing SNR performance.

                  1361.     Synthesis of Multiple Phase Cycled SSFP Images to Remove Band Artifacts as Well as to Improve
                                 the SNR   by Use of a Spectral Decomposition

Kwan-Jin Jung¹, ²

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

The band artifacts in SSFP can be removed by MIP of the multiple acquisitions of the phase cycled SSFP images. MIP, however,  cannot take the effect of averaging of the multiple images to improve the SNR. There were two reported methods, i.e., the complex sum and the squared sum, that had claimed to achieve both objectives. These two methods, however, could not be confirmed to achieve the major objective of the band artifacts removal. A new method to achieve both effects has been developed by use of the spectral decomposition of each phase cycled image into the low and high frequency components. The low frequency components are processed by MIP, while the high frequency components are averaged. The MIP of the low frequency components and the average of the high frequency components are summed, and results in the synthesized image that is free from the band artifacts with the improved SNR.

New Methods for Generating Contrast

Tolga Çukur¹, William Overall¹, Dwight Georger Nishimura¹

¹Stanford University, Stanford, California , USA

Methods for generating positive-contrast images from susceptibility-induced magnetic field variations are useful for applications such as the guiding of interventional devices or imaging of super paramagnetic iron oxide (SPIO)-labeled cells. SSFP-based methods, such as fast low-angle positive-contrast SSFP (FLAPS), offer fast imaging and flow insensitivity. However, the level of background suppression can be limited. In this work, we employ the alternating repetition time (ATR) SSFP stop-band to suppress the signal from on-resonant spins. The ATR sequence provides positive contrast with robust suppression for a wide range of flip angles and tissue parameters.

Angus Zoen Lau¹, Charles Henry Cunningham¹

¹Department of Medical Biophysics, University of Toronto, Toronto, Canada

Conventional imaging of superparamagnetic iron oxide (SPIO) particles yields negative contrast in the region surrounding the particles. Recent positive contrast techniques refocus this dephased signal using an off-resonance spin echo sequence. We investigate the possibility of imaging off-resonance short T₂ spins using a short TE spiral pulse sequence containing a spectrally-selective self-refocusing RF pulse. A short T₂ agar gel model using gadolinium as the field perturber is used to evaluate the performance of this sequence. We demonstrate that the short TE sequence is able to refocus increased signal from off-resonance short T₂ spins compared to a spectrally-selective spin echo sequence.

Shannon Heather Kolind¹, Burkhard Mädler², Alex Lloyd MacKay¹

¹University of British Columbia, Vancouver, Canada; ²Philips Medical Systems,, Vancouver, Canada

We compared results from a recently introduced 3D multi-echo T₂-relaxation imaging technique to the standard single-slice multi-echo T₂-relaxation measurement most common in the literature. For 10 healthy controls, myelin water fraction correlated strongly between the techniques (slope=1.00,intercept=-0.91%,R²=0.89,p<0.0001) with significant differences only occurring in peripheral brain. Geometric mean T₂ was not significantly different in any brain structure examined except minor forceps. SNR was generally higher using the 3D technique. In conclusion, results from the 3D multi-echo T₂-relaxation technique were generally consistent with single-slice results, and achieved 7 times greater brain coverage in similar scan times with higher SNR.

Christine Lucas Tardif¹, D Louis Collins¹, G Bruce Pike¹

¹Montreal Neurological Institute, Montreal, Canada

The stability and accuracy of the voxel-based morphometry (VBM) process is crucial in large studies to improve the power of statistical results and minimize regional bias. In this study, we evaluated three protocols, FLASH, MP-RAGE and MDEFT, from the perspective of VBM. Nine volunteers were scanned twice for each protocol. We performed a VBM analysis of grey matter (GM) density between the sequences, as well as GM density scan-rescan variability. The results show large areas of significant difference in GM density and variability between the three protocols, and suggest that MDEFT is best suited to this type of analysis.

José P. Marques¹, ², Tobias Kober¹, ³, Wietske van der Zwaag¹, ², Gunnar Kruegger¹, ³, Rolf Gruetter¹, ⁴

¹Ecole Polytechnique Federale de Lausanne, Lausanne, Switzerland; ²University of Lausanne, Lausanne, Switzerland; ³Siemens Medical Solutions-CIBM, Lausanne, Switzerland; ⁴University of Lausanne and Geneva, Switzerland

To improve bias field correction as well as contrast between tissues with different T₁’s, the MPRAGE sequence was modified into a MP2RAGE (Magnetization Prepared with 2 RApid Gradient Echoes), which accommodates two gradient echo modules after each inversion, GRE_TI1 and GRE_TI2, which are characterized by their two different inversion times (TI1 and TI2) and flip angles α ₁ and  α ₂). The use of the combined images resulted in higher contrast between CSF, GM and WM.

Julien Sénégas¹, Wei Liu², ³, Hannes Dahnke¹

¹Philips Research Europe, Hamburg, Germany; ²National Institutes of Health, Bethesda, USA; ³Philips Research North America, Briarcliff Manor, USA

This work presents a new acceleration method that exploits the temporal correlation of the k-space signal at different echo times to reduce the number of phase-encoding steps in a multi-echo spin-echo sequence. The extension to the case of multiple coils is addressed. The approach is evaluated with respect to SNR and reconstruction artefacts, and compared to SENSE and GRAPPA. For single coil acquisitions, the proposed method represents a new way of reducing the scan time that was not accessible with sensitivity encoding approaches. For multiple coils acquisitions, it extends the existing methods and allows an appreciable improvement in reconstruction quality, limiting especially noise amplification.

Christian Hoehl¹, Nouri Elmiladi¹, Jessica Mende¹, Karl Maier¹

¹Rheinische Friedrich-Wilhelms-Universität, Bonn, Germany

A new method of contrast may be achieved by using different nanoparticles as high frequency transmitters. Activated by ultrasound radiation, the nanoparticles emit electromagnetic waves within their nearest neighborhood. This produces a change in the relaxation rate, depending on the ultrasound parameters and the physical properties of the tissue.

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

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

We present a closed-form prescription for NMR pulses and delays which maximizes heteronuclear hyperpolarization through transfer of spin order from a coupled nuclear singlet state.  This result is applicable to the preparation of MRI contrast agents through the method of Parahydrogen-Induced Polarization.  The sequence is derived using a geometric analogy to the evolution of orthogonal density matrix components.  We find that the sequence is optimal in that it minimizes sequence duration, and therefore T2-like relaxation, and is less sensitive to pulse imperfections than other prescriptions studied.

Andreas Stadlbauer¹, Wilma van der Riet², Markus Scheidegger³, Stephan Gruber⁴, Erich Salomonowitz¹

¹Landesklinikum St.Poelten, St. Poelten, Austria; ²European MRI Consultancy (EMRIC), Strasbourg, France; ³Philips AG Medical Systems, Zurich, Switzerland; ⁴Medical University Vienna, Vienna, Austria

We determined the optimal systolic and diastolic time delays (TD) for ECG-triggering of a non-contrast enhanced MRA (TRANCE) in patients with peripheral arterial disease. We specified the requirements for the systolic images with minimal and the diastolic images with maximum arterial SNR. An ECG-triggered T2w TSE (VISTA) sequence was performed in 12 patients using variable trigger delays for systolic and diastolic phase. The SNR in the popliteal and anterior tibial arteries of systolic and diastolic images were determined. The optimal systolic TD was 14–21ms smaller, the optimal diastolic TD 200ms apart from time of maximum peak velocity.

Twinkle Gupta¹, Saurabh Shah², Sumeet Virmani, Reed Omary¹, Andrew Larson¹

¹Northwestern University, Chicago, Illinois, USA; ²Siemens Medical Solutions, Chicago, Illinois, USA

We investigated the use of spectrally selective hard pulses (SSHP) for positive contrast MRI of paramagnetic markers. Our primary focus was to study the effects of specific SSHP-parameters (targeted off-resonance and no. of binomial sub-pulses=N) upon marker conspicuity, background and fat suppression. Studies were performed using simulations and phantom models. Background suppression increased with increasing N and targeted off-resonance. Marker conspicuity was higher at lower N and off-resonance values. Fat suppression maximized at 100Hz off-resonance. SSHP pulses offer a promising method for positive contrast MR. Further developments are needed to demonstrate the utility of this technique for in-vivo clinical applications.

Zhiqiang Li¹, Donglai Huo², Eric Aboussouan², Xiaoli Zhao³, John Karis², Leland Hu², Zhu Li³, James G. Pipe²

¹GE Healthcare, Phoenix, Arizona , USA; ²Barrow Neurological Institute, Phoenix, Arizona , USA; ³GE Healthcare, Waukesha, Wisconsin, USA

Spin Echo (SE) sequence is widely used for T1-weighted brain imaging. T1 FLAIR FSE also has drawn significant attention recently due to superior SNR and CNR, and improved lesion conspicuity. However, these techniques are significantly impaired in regions affected by flow artifact, which is even worse in post-contrast imaging. Turboprop is a technique insensitive to flow artifact. In this work, we combine FLAIR preparation with turboprop to produce superior T1 contrast with minimized flow artifact. This technique can be used for both pre- and post-contrast T1 imaging, and has several advantages over current clinical T1 SE protocols.

Rexford David Newbould¹, Marcus T. Alley¹, Stefan Ropele², Roland Bammer¹

¹Stanford University, Stanford, California , USA; ²Medical University Graz, Graz, Austria

Recent work has presented the concept of magnetization transfer effects in balanced SSFP (bSSFP) acquisitions for calculating MTR maps from variation in the steady-state magnetization. Another quantity of interest is the variation in T1 of the free pool under varying saturation of the bound pool. By prefacing the bSSFP train with an inversion pulse, the transient magnetization progresses towards steady-state with an apparent rate constant known as T1*, which can be related to underlying T1.  In this work IR-bSSFP relaxometry is performed with varying direct RF power deposition, which can measure both the variation in steady-state levels and underlying T1.

Kyunghyun Sung¹, Hsu-Lei Lee¹, Houchun Harry Hu¹, Krishna S. Nayak¹

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

Magnetization transfer (MT) in 3T cardiac balanced steady state free precession (bSSFP) imaging can possibly produce a new type of contrast to detect infarcts and inflammation of the myocardial tissue. We evaluate the range of MT ratio (MTR) with different RF pulse elongations by a constant TR (3.6ms – 5.6ms). We also perform multiple measurements to establish variation range in 3T cardiac imaging. The study shows a MTR of 12 – 19% in 3T bSSFP cardiac imaging while a MTR of 30 – 50% has been reported in brain applications.

Sergio Andres Uribe¹, Reza Razavi¹, Tobias Schaeffter¹

¹Kings College London, London, UK

In this work we present a new method that allows the integration of magnetization preparation sequences into b-SSFP sequence without disturbing the steady state. Gradients and rf pulses are integrating into steady state free precession sequence after certain time interval without the need to stop and re-establish the steady state. An example for fat suppression has been implemented and tested in phantom and volunteers. Other preparation techniques can also benefit from this method, such as regional saturation (REST) or spin labeling.

Catherine Judith Moran¹, Walter F. Block²

¹University of Wisconsin, Madsion, USA; ²University of Wisconsin, Madison, USA

Linear Combination SSFP is one of a number of methods that provide fat-water separation in balanced SSFP.  Fat/water separation in this dual pass method suffers when the magnitude of the signal varies between passes.  Instead, we recognize the phase difference between the two passes robustly falls into two categories, 0 for water and 180 degrees for fat.  Unlike previous phase-sensitive methods, common mode sources of phase error cancel out. Thus an extremely simple reconstruction method can classify fat voxels and remove them.  We demonstrate the method in dramatic improvements of contrast in breast and knee imaging.

Walter Robert Thurmond Witschey II¹, Mark A. Elliott, Jeremy Magland¹, Ravinder Reddy

¹University of Pennsylvania, Philadelphia, USA

Steady-state free precession gridding exploits the external field sensitivity of the balanced gradient echo sequence to generate a uniform image grid. A simple gradient pulse added to the balanced gradient echo sequence 'unbalances' the sequence and generates a grid with spacing proportional to the gradient pulse amplitude and gradient. Potential applications include steady-state diffusion weighting and persistent motion sensitivity.

Philipp Mörchel¹, Gerd Melkus¹, Markus Kotas¹, Michael Flentje¹, Peter Michael Jakob¹

¹University of Würzburg, Würzburg, Germany

A new method for the simultaneous measurement of T1 and T2* maps is presented. This method is robust against variations in flip angles due to B1 inhomogeneities of the resonator. Moreover, this sequence is time efficient because both parameters can be measured in the time the T1 measurement would take. Another advantage is that there no misregistration artifacts due to motion during a consecutive measurement of the single parameters T1 and T2*.

Roman Fleysher¹, Lazar Fleysher¹, Songtao Liu¹, Oded Gonen¹

¹NYU School of Medicine, New York, USA

We present a method of T1 estimation designed to be free of systematic errors caused by B1 inhomogeneity in which the value of T1 is extracted from three 3D spoiled-gradient-recalled-echo (SPGR) images acquired with EPI readout. The method provides protocols optimized for precision in T1 given available scan time. The precision is comparable to that of the adopted two-SPGR method in the same time and spatial resolution. If the two-SPGR methods are combined with B1 measurement for accuracy, then TRITONE supersedes them in precision in the same total time.

Bernard Siow¹, Li Sun¹, Andrew M. Blamire¹

¹Newcastle University, Newcastle upon Tyne, UK

Intermolecular multiple quantum coherence imaging sequences have recently been shown to provide a fundamentally different contrast mechanism to conventional MRI. A numerical study of intermolecular zero-quantum coherence imaging has shown sensitivity to susceptibility gradients at selected distance scales. In this study, an iZQC sequence was implemented and iZQC signal verified. The sequence was used to investigate sensitivity to susceptibility gradients at selected distance scales in structured samples. Images show contrast in areas where susceptibility gradients are present. Furthermore, contrast was modulated by the distance scale selected. Further results suggest that contrast is modulated by specific resonant frequency difference at distance scale selected.

Akira Sasao¹, Mika Kitajima¹, Toshinori Hirai¹, Hirofumi Fukuoka¹, Tomoko Okuda¹, Tomoyuki Okuaki², Shutaro Saiki², Yasuyuki Yamashita¹

¹Kumamoto University, Kumamoto, Japan; ²Philips Medical Systems, Japan

We measured T2^* values of the five white matter structures, the genu and splenium of the corpus callosum, posterior limb of the internal capsule (PLIC), occipital white matter adjacent to the lateral ventricle and frontal white matter, at 3T using a multi-echo fast field echo sequence with and without the main field inhomogeneity (B₀ inhomogeneity) correction in 20 neurologically normal cases. The mean T2^* values of all measured white matter regions using B0 inhomogeneity correction was statistically longer than those without B₀ inhomogeneity correction. The mean T2^* value of PLIC using B₀inhomogeneity correction was longest among the five regions.

Compressed Sensing

Kai Tobias Block¹, Martin Uecker¹, Jens Frahm¹

¹Biomedizinische NMR Forschungs GmbH, Goettingen, Germany

Finite sampling of k-space causes ringing artifacts from signal truncation at the border of the measured k-space. This work demonstrates that the simple assumption of a piecewise-constant object can be exploited to extrapolate the data in k-space beyond the measured part. The assumption translates into a total variation minimization problem which can be solved with a non-linear optimization algorithm and allows for a significant reduction of truncation artifacts without compromising resolution. In the presence of substantial noise, a modified approach offers edge-preserving denoising which, in addition to supplementing data, also tolerates slight deviations from the measured data.

Chen Zhao¹, Tao Lang¹, Jim Ji¹

¹Texas A&M University, College Station, USA

Both parallel MR Imaging (pMRI) and compressed sensing (CS) can significantly reduce image acquisition time in MRI, the former by utilizing multi-channel receivers and the latter by utilizing the sparsity of MR images in a transformed domain. An integrated approach was developed to use CS as a regularization tool to improve the pMRI reconstruction. Reconstruction results from in-vivo data show that CS can significantly suppress the aliasing artifacts and improve SNR, but with very minor resolution loss. Comparison with the truncated SVD and the Tikhonov  regularization were presented.

Martin Uecker¹, Kai Tobias Block¹, Jens Frahm¹

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

Recent algorithms for autocalibrated parallel imaging estimate the coil sensitivities and the image at the same time. Regularization is commonly employed to counteract the low SNR due to the reconstruction process (quantified by the g-factor map). In contrast to linear reconstruction methods which are restricted to a regularization related to the L2-norm of the image, nonlinear regularization methods like the L1-norm in combination with a sparsity transform or total variation are known to suppress noise much more efficiently. This work demonstrates how L1-wavelet regularization can be incorporated into an autocalibrating parallel imaging algorithm based on a regularized nonlinear inversion method.

Bing Wu¹, Rick P. Millane¹, Richard Watts¹, Philip Bones¹

¹University of Canterbury, Christchurch, New Zealand

Compressed sensing is used in recovering images from multiple receiver coil data sets, and the results are compared to those obtained from using a direct image recovery method (SENSE). It is shown that compressed sensing achieved better reconstruction results than the conventional SENSE approach at high acceleration factors, since the latter suffers badly from deteriorated SNR.

Tao Lang¹, Jim Ji¹

¹Texas A&M University, College Station, USA

Compressed Sensing (CS) is an effective approach to fast imaging by utilizing the signal sparsity in a transformed domain. We developed a CS imaging method for dynamic contrast enhanced MRI. Specifically, a difference operator is applied to the successive temporal data frames to enhance the spatial signal sparsity for CS reconstruction. The new algorithm method is assessed using simulated and in-vivo dynamic imaging data. The result shows that the new method provides higher resolution than zero-padded Fourier reconstruction and the Keyhole method, and it results in reduced artifacts and noise than conventional CS reconstruction where no temporal information is used.

Diego Hernando¹, Justin Haldar¹, Leslie Ying², Kevin King³, Dan Xu³, Zhi-Pei Liang¹

¹University of Illinois at Urbana-Champaign, Urbana, Illinois, USA; ²University of Wisconsin-Milwaukee, Milwaukee, Wisconsin, USA; ³G.E. Medical Systems, Milwaukee, Wisconsin, USA

Interventional MRI (I-MRI) is an important dynamic imaging application, allowing the guidance of therapeutic procedures, which requires high frame-rate and near-real-time reconstruction. Compressed sensing (CS) allows high-resolution reconstruction from a reduced number of samples by exploiting the sparsity of the signal. In this work, CS is tailored to ¡ maximize¡± the sparsity in each frame while satisfying the inherent causality constraints in I-MRI reconstruction, so that high-quality images can be obtained from a small number of samples.

Orhan Unal¹, ², Jie Tang¹, Guang-Hong Chen¹

¹University of Wisconsin, Madison, Wisconsin, USA

Recently, a novel reconstruction technique called PICCS (Prior Information Constraint Compressed Sensing) has been developed. This technique enables accurate reconstruction of high SNR images with large undersampling factors (>50 in 2D). In this work, PICCS and its utility in MR perfusion imaging is investigated. In PICCS, the sparsity of the difference image between the target image and the high SNR prior image is exploited.  When this PICCS sparsity is combined with other known sparsifying transforms such as total variation (TV) norm used in compressed sensing (CS), PICCS enables accurate reconstruction of images using fewer projections than required by CS and standard filtered backprojection (FBP) reconstruction

Luca Marinelli¹, Christopher J. Hardy¹, Daniel J. Blezek¹, ²

¹General Electric, Niskayuna, USA; ²Mayo Foundation and Clinic, Rochester, USA

Parallel imaging methods reduce acquisition time at the expense of aliasing artifacts and image SNR.  Compressed sensing has emerged as a framework for exact reconstruction for compressible signals and images. A novel approach is proposed to adapt distributed compressed sensing algorithms to accelerated multi-coil MR imaging. The algorithm employs coil sensitivity maps and can be combined with parallel imaging to achieve high acceleration factors. We demonstrate the convergence properties of the method on a numerical phantom and apply it to FGRE imaging.

Ricardo Otazo¹, Fa-Hsuan Lin², Graham Wiggins², Ramiro Jordan¹, Stefan Posse¹

¹University of New Mexico, Albuquerque, New Mexico, USA; ²Massachusetts General Hospital, Boston, Massachusetts, USA

Parallel MRI reconstruction is formulated as a superresolution problem where acceleration is performed by acquiring the low spatial resolution representation of the object being imaged and coil sensitivity maps are acquired with higher target spatial resolution. The increase in spatial resolution will be determined by the degree of sensitivity variation within the low resolution voxel. The method is applicable to receiver arrays with a large number of small elements which provide strong spatial variation of the coil sensitivity maps. Superresolution Sensitivity-Encoding (SURE-SENSE) represents a powerful alternative to standard SENSE for the same acquisition time and it is advantageous for low spatial resolution where the aliasing from intra-voxel coil sensitivity variation is removed. We show feasibility of the method for human brain imaging using receiver arrays with 32 and 96 elements.

Richard Z. Shilling¹, Marijn E. Brummer²

¹Georgia Insitute of Technology, Atlanta, Georgia, USA; ²Emory University, Atlanta, Georgia, USA

Most super-resolution reconstruction (SRR) techniques proposed for MRI to date have involved multiple multi-slice acquisitions with sub-pixel shifts in the slice-selection or phase-encoding direction.  Another SRR technique has been developed called “Multi-stack” (MS), which instead of using sets of parallel shifted scans (PS), uses multiple multi-slice stacks of the same object, scanned at different orientations, rotated about the frequency-encoding axis.  In this work we compare these methods with a real scanning experiment using 2-D Inversion Recovery acquisition.

André Fischer¹, ², Felix Breuer², Martin Blaimer², Nicole Seiberlich¹, Peter Michael Jakob¹, ²

¹University of Wuerzburg, Wuerzburg, Germany; ²Research Center for Magnetic Resonance Bavaria e.V., Wuerzburg, Germany

In recently published works, convex optimization procedures were chosen for recovering missing data with Compressed Sensing (CS). This method depends on a proper adjustment of parameters in the functional to be optimized; this is often computationally expensive and reduces the advantage of accelerated data acquisition. In this work, a CS reconstruction algorithm based on a nonconvex procedure is introduced for MR imaging. No parameter determination is necessary, thus reducing the computational load. A preliminary application in the field of MR is shown, as well as a demonstration which indicates that the nonconvex technique offers similar results as a convex optimization.

Kevin F. King¹

¹GE Healthcare, Waukesha, Wisconsin, USA

Compressed sensing uses L1-norm minimization in a sparse transform space to reconstruct randomly undersampled k-space data.  Parallel imaging uses L2-norm error minimization to incorporate receive B1 information into the reconstruction of undersampled multicoil k-space data.  An L1-norm penalty function has also been used for denoising and regularization of parallel imaging and non-Cartesian k-space reconstructions.  These methods are overlapping and complimentary and can be combined by reconstructing randomly undersampled multicoil k-space data with both L1- and L2-norm minimization.  This combines compressed sensing and L1-regularized parallel imaging and denoises the image.  The resulting acceleration can be higher than with either technique alone.

Yoshifumi Yamada¹, Satoshi Ito¹

¹Utsunomiya University, Utsunomiya, Japan

This report presents a novel data interpolation method in the phase-scrambling Fourier transform (PSFT) imaging technique. A coarsely sampled PSFT signal which an aliasing artifact is produced in the reconstructed image can be interpolated to obtain fully encoded signal by using an iteration method based on the Gerchberg's algorithm known as a super resolution technique. Modification of the Gerchberg's algorithm is made to apply the method to interpolation of PSFT signal. Numerical simulation using MR images shows that alias-free images are reconstructed from the interpolated PSFT signal by using this technique.

Tissue Segmentation & Localization

Ricardo J. Ferrari¹, Anne L. Martel¹, ²

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

The main objective of this work is to use bilateral asymmetry analysis to detect abnormal global changes in the breast parenchymal flow. The analysis, which is based on the differences in texture and directional edge information obtained from the left and right MR breast images, aims to help radiologists in the detection of breast cancer in women with a high risk of developing such disease. Preliminary results of the proposed method has shown an average accuracy of 75.0% on the classification when applied to 40 MRI cases (20 benign and 20 malignant).

Chunming Li¹, Zhaohua Ding¹, Mary M. Zutter¹, John C. Gore², Thomas Yankeelov³

¹Vanderbilt University, USA; ²Institute of Imaging Science, USA; ³Vanderbilt University, Nashville, USA

We apply our recently developed local binary fitting (LBF) method of image segmentation to the problem of separating angiogenic vessels amid a heterogeneous intensity background and show how it yields superior results than common methods of vessel segmentation which include simple intensity thresholding and the Chan-Vese model.  The LBF algorithm is less sensitive to noise than thresholding and Chan-Vese model, provides more accurate segmentation and can be applied to images with substantial intensity inhomogeneities.

Suyash Prakash Awate¹, James C. Gee¹

¹University of Pennsylvania, Philadelphia, USA

We present a novel, fast, and robust brain-tissue segmentation method.

Yutong Liu¹, ², Mariano Uberti¹, Michael Boska¹

¹University of Nebraska Medical Center, Omaha, Nebraska, USA

We developed a technique to extract mouse brain from in vivo head MRI based on an active contour model (“snake”). Most brain extraction techniques developed for human head MRI that employ the image intensity and gradient cannot be directly used for mouse. We include higher level image constraints (“pins”) in the energy function of the snake. The definition of the pins and the initial snakes are automated in this technique. Results have demonstrated that pinned snakes can accurately extract the brain from 3D head MRI with minimal user intervention.

Gernot Reishofer¹, Stefan Ropele¹, Franz Ebner, Rudolf Stollberger²

¹Medical University Graz, Graz, Austria; ²Technical University Graz, Graz, Austria

Previous studies have shown the potential of independent component analysis to separate the signal assigned to macro vessels from tissue signal. This enables an algorithm for minimizing macro-vessel signal in DCE-MRI perfusion imaging to avoid overestimation of relative blood flow and blood volume. The reconstruction step for restoring corrected DCE-MRI time series involves the solution of ill-posed linear systems. Direct back transformation leads to noise amplification in the reconstructed dynamic time series which results in noisy hemodynamic parameter maps. We found that Tikhonov regularization improves the visualization of hemodynamic parameter changes without affecting the quality of macro-vessel minimization through ICA

Voreak Suybeng¹, Borys Shuter¹, David Stringer², Ashraf Kassim¹, Shih-Chang Wang¹

¹National University of Singapore, Singapore, Singapore; ²National University Hospital, Singapore, Singapore

4D DCE-MRI studies contain temporal information which may be useful in segmenting the kidney, liver and spleen. Three cluster analysis techniques (Kohonen Neural Network, Fuzzy K-Means, Expectation-Maximization) and Factor Analysis of Dynamic Structures (FADS) are compared. All performed similarly in segmenting kidney parenchyma from other abdominal organs. FADS appeared to be most sensitive to the temporal information producing TICs for kidney compartments that were of more physiological interest.

Peter Jan Koopmans¹, Rashindra Manniesing², Wiro J. Niessen², Max A. Viergever³, Markus Barth⁴, ⁵

¹F.C. Donders Centre for Cognitive Neuroimaging, Radboud University, Nijmegen, Netherlands; ²Department of Medical Informatics and Radiology, Rotterdam, Netherlands; ³Image Sciences Institute, University Medical Center Utrecht, Netherlands; ⁴F.C. Donders Centre for Cognitive Neuroimaging, Radboud University, Nijmegen, Netherlands; ⁵Erwin L. Hahn Institute for Magnetic Resonance Imaging, University Duisburg-Essen, Germany

MR venography data of the brain were acquired at 3T and 7T using susceptibility weighted imaging (SWI). Two automated vein segmentation filters, the Utrecht vesselness filter and Vessel Enhancing Diffusion (VED), were tested. VED results were in excellent agreement with manual segmentation. Results at 3T were marginally better than those at 7T due to increased inhomogeneity at the latter field strength. However, the acquisition time at 3T was 50% longer. SWI at 7T shows good promise for future MR venography.

James P. Withers¹, Mark E. Bastin¹, Amos J. Storkey¹

¹University of Edinburgh, Edinburgh, UK

The segmentation of brain MRI volumes is complicated by noise as well as partial volume voxels that contain a mixture of tissue types. Automated algorithms that perform partial volume segmentation often rely on average tissue class maps that may not represent the target anatomy, especially blood vessels and cortical sulci. These small structures can also be smoothed out by aggressive homogeneity constraints or confined to tissue boundaries only. In this abstract, robust techniques for minimizing aliasing through super-resolution, quantifying mixtures over a small number of discrete mixture states, and detecting tube-like structures using differential geometry are examined.

Stephanie Bricq¹, ², Christophe Collet¹, Jean-Paul Armspach²

¹LSIIT - UMR CNRS 7005, Illkirch, France; ²LINC - UMR CNRS 7191, Strasbourg, France

We present a new automatic robust method to estimate parameters to segment brain MR images  (WM, GM, CSF) and MS lesions using the Hidden Markov Chain (HMC) model. For this aim, we use the Trimmed Likelihood Estimator (TLE) to extract outliers and propose to include a priori information brought by a probabilistic atlas. Tests on Brainweb images with MS lesions have been carried out.

David Scott Wack¹, ², Michael Dwyer¹, Sara Hussein¹, Christopher Caiola¹, Peter Hojczyk¹, Robert Zivadinov¹

¹State University of New York at Buffalo, Buffalo, USA

A method which uses Stochastic Discrimination for the detection of MS lesions was trained using 40 scans from MS subjects acquired from 4 scanners.  Thirty-nine scans  were used as a test set.  The pattern recognition results were subsequently used in a stage which created an outline of the lesions in a format accepted by JIM (Java Image Manipulation) software, allowing for further editing of the results.  A comparison of the un-edited ROIs and expert drawn ROIs had median Kappa value of .61.  ROIs from either method predicted over 93% of the lesions found in the other.

Holly C. Canuto¹, ², Marko Velic³, Charles I. McLachlan³, Mikko I. Kettunen¹, ², Anant S. Krishnan¹, Andre' A. Neves¹, ², Maaike M. de Backer¹, De-en Hu¹, Michael P. Hobson¹, ³, Kevin M. Brindle¹, ²

¹University of Cambridge, Cambridge, UK; ²Cancer Research UK, Cambridge, UK; ³Metropolis Data Consultants, Cambridge, UK

The capability to perform reliable and accurate automated segmentation of MR images does not currently exist within the MR imaging community. Bayesian image analysis methods have been used successfully in the astronomy community for more than thirty years to de-noise images and enhance feature recognition.  We have developed a Bayesian Multi-Region Segmenter (BMRS) to provide reliable assessment of contrast heterogeneity within murine lymphomas and potentially enhance the sensitivity of contrast agent detection.

Ana Rodriguez-Soto¹, ², Julio Carballido-Gamio², Thomas M. Link², Sharmila Majumdar²

¹Universidad Iberoamericana, Mexico City, Mexico; ²University of California, San Francisco, San Francisco, California , USA

In this work a semi-automatic cartilage segmentation technique is presented. The bone cartilage interface was extracted by Canny edge detection and was used as initialization for the segmentation of the articular surface. First and second order gradients of pixel intensities were used to move the control points of the splines representing the cartilage contours. A self correction technique based on local cartilage thickness values of adjacent slices was implemented to reduce user interaction. The algorithm was tested with 10 human knees with different levels of OA which were segmented twice. The global CV was 0.57% for 3D cartilage thickness and 1.69% for cartilage volume demonstrating the good performance of the technique.

Sylvia Drabycz¹, J Ross Mitchell¹

¹University of Calgary, Calgary, Canada

Previous studies have demonstrated that local spectral analysis using the polar S-transform (PST) can help discriminate between normal-appearing white matter (NAWM) and active as well as inactive multiple sclerosis (MS) lesions on MR images. However, the very large number of computations required for even moderately sized images relegates the PST to a region-of-interest approach, which reduces the resulting spectral resolution. We introduce a novel spectral analysis technique that uses circularly symmetric windows in the frequency domain and averages the contribution of each complex spectral contribution. Our method produces a complex, and fully invertible, 3D local frequency domain with complete spectral resolution.

Ankur Purwar¹, RKS Rathore¹, RK Gupta², MK Sarma¹, G Bayu¹, D Rathore¹, R Trivedi², JK Singh¹, Anup Singh¹, S Verma¹

¹IIT Kanpur, Kanpur, India; ²SGPGIMS Lucknow, Lucknow, India

De-scalping the brain is a very useful procedure with enormous applications in visualization, surface rendering, decreasing the complexity of subsequent processing algorithms, and the like. Many applications related to brain imaging either require, or benefits from the ability to accurately segment brain from the non-brain tissue. For example, (a) in the registration of b0-images to DW images in DT-MRI, both b0 and DW images often contain considerable portions of eyeballs, skin etc. that cause problems in the registration process, which gets improved once these non-brain parts of the images are removed, (b) a second application of de-scalping is in tissue-type segmentation, which helps in isolating brain tissue from other parts of the stack such as CSF, (c) in the removal of strong ghosting effects which can occur with EPI and (d) creating probabilistic atlases from large groups of subjects. This work describes an automatic procedure for de-scalping in the brain for MRI for scans with axial orientation. The procedure has effective in all types of brain imaging techniques including T1, T2W, PD, and T2W-EPI.

Zhiqiang Lao¹, Dinggang Shen¹, Elias R. Melhem¹, Nick R. Bryan¹, Christos Davatzikos¹

¹University of Pennsylvania, Philadelphia, Pennsylvania, USA

We present a computer-assisted white matter lesions (WMLs) segmentation method, based on local features extracted from multi-parametric MRI sequences. A Support Vector Machine (SVM) classifier is first trained on expert-defined WML’s, and is then used to classify new scans. Subsequent post-processing analysis further reduces false positives by utilizing anatomical knowledge and measures of distance from the training set. Cross-validation on 35 patients from 3 different imaging sites with WMLs of varying sizes, shapes and locations tests the robustness and accuracy of the proposed segmentation method, compared to the manual segmentation results from two experienced neuroradiologists.

Spectroscopy Methodology:  Miscellaneous

Meng Gu¹, Dirk Mayer¹, Edith Sullivan, Adolf Pefeeferbaum, Daniel Spielman¹

¹Stanford University, Stanford, USA

A reproducibility study of proton magnetic resonance spectroscopic imaging of the human brain was conducted to evaluate the reliability of an automated 3D in vivo spectroscopic imaging acquisition and associated quantification algorithm.  The complete study consisted of 7 healthy adult subjects scanned within a one month interval to assess inter-subject variability and one subject scanned six times to assess intra-subject variability. The observed inter-subject and intra-subject coefficients-of-variation from five regions-of-interest were comparable to those reported for single-voxel acquisitions from similar ROIs. These results demonstrate that reproducible whole-brain 1H-MRSI data can be robustly obtained with the proposed methods.

Eric Albert Mellon¹, Stephen J. Pickup¹, Gamliel Isaac¹, Sueng Cheol Lee¹, Edward J. Delikatny¹, Ravinder Reddy¹, Jerry D. Glickson¹

¹University of Pennsylvania, Philadelphia, Pennsylvania, USA

Selective Homonuclear Multiple-Quantum Coherence-transfer (Sel-MQC) lactate spectroscopy has shown excellent promise for the detection of lactate with full fat and water suppression. While 2D versions of this technique have been published on animal scanners and very recently a 3D technique has been proposed, to our knowledge these techniques have not been implemented on a clinical scanner. Here a 2D phase encoded, 1D Hadamard encoded Sel-MQC sequence is demonstrated in single and double quantum modes for the detection of physiologic levels of lactate in phantoms on a 3T clinical scanner with a standard coil for the goal of future clinical studies.

Toru Shirai¹, Yukari Yamamoto¹, Yoshitaka Bito¹, Satoshi Hirata²

¹Hitachi, Ltd., Kokubunji-shi, Japan; ²Hitachi Medical Corporation, Kashiwa-shi, Japan

We propose a new technique for simultaneous acquisition of metabolite and water signals in CSI. The pulse sequence of this technique includes a CHESS pulse whose amplitude is switched alternately in accordance with phase encoding steps in order to reverse the polarity of the water signal. Because the water signal is shifted to the top and bottom of the reconstructed image, the metabolite signal is separable from the water signal. The results of phantom experiments showed that this technique was effective in suppressing water signal, which suggests the usefulness of the proposed method.

Joonsung Lee¹, Borjan Gagoski¹, Elfar Adalsteinsson¹

¹Massachusetts Institute of Technology, Cambridge, Massachusetts, USA

Estimates of cortical brain metabolites using chemical shift imaging (CSI), especially those of NAA, are severely hampered by strong, subcutaneous lipid signals. by reducing the spatial side lobes and using the variable-density spiral trajectories, lipid signals can be bound spatially to reduce the amount of lipid signal leakage. A 3D filter was designed and demonstrated in variable-density spiral-based trajectory for volumetric brain CSI. Excellent lipid suppression was achieved via the filter function without minimal SNR tradeoffs.

Eugene Ozhinsky¹, ², Sarah J. Nelson¹, Daniel B. Vigneron¹

¹University of California San Francisco, San Francisco, California , USA

To improve coverage of the brain, while simultaneously reducing lipid contamination, we have developed a technique for automatic optimization of sat band position and orientation that simultaneously maximizes coverage of subcutaneous fat and minimizes the portion of brain tissue that is removed. This allows to prescribe a much larger PRESS box and provides MR spectra from a significantly larger volume of the brain.

Claudiu Schirda¹, Fernando Boada²

¹State University of New York at Buffalo , Buffalo, New York, USA; ²University of Pittsburgh, Pittsburgh, Pennsylvania, USA

Fast spectroscopic imaging enables resolving the spectral information at much higher spatial resolutions compared to classical Chemical Shift Imaging (CSI) experiments, in shorter scan times. Besides fast CSI techniques that sample k-space on a Cartesian grid (e.g. Dixon), a number of non-conventional approaches employing non-Cartesian trajectories have been developed. Among them are stochastic trajectories, PREP, spiral and rosette trajectories. We theoretically analyze and experimentally demonstrate the off-resonance effects associated with these non-conventional encoding techniques.

Yao Fu¹, Hacene Serrai¹

¹National Research Council Institute for Biodiagnostics, Winnipeg, Canada

This work describes a new MRSI technique resulting from combination of wavelet encoding and parallel imaging. The purpose of this work is to reduce acquisition time in MRSI and minimize the SNR loss inherent to parallel imaging. Similar to Fourier encoding with parallel imaging where a predetermined number of k-space lines are skipped depending on the speed factor R, fewer lines in the wavelet domain, which resemble to k-space for wavelet encoding, are omitted and reconstructed from the acquired signals. As a result, reduction of acquisition time achieved by a factor R with minimum loss of SNR as compared to Fourier encoding.

Claudiu Schirda¹, Fernando Boada²

¹State University of New York at Buffalo , Buffalo, New York, USA; ²University of Pittsburgh, Pittsburgh, Pennsylvania, USA

Fast spectroscopic imaging enables resolving the spectral information at much higher spatial resolutions compared to classical Chemical Shift Imaging (CSI) experiments, in shorter scan times. Besides fast CSI techniques that sample k-space on a Cartesian grid (e.g. Dixon), a number of non-conventional approaches employing non-Cartesian trajectories have been developed. Among them are stochastic trajectories, PREP, spiral and rosette trajectories. We theoretically analyze and experimentally demonstrate the off-resonance effects associated with these non-conventional encoding techniques.

Claudiu Schirda¹, ², Fernando Boada²

¹State University of New York at Buffalo , Buffalo, New York, USA; ²University of Pittsburgh, Pittsburgh, Pennsylvania, USA

Nonalcoholic fatty liver disease (NAFLD) affects an estimated 14% to 30% of the general population in the United States, with an important number of these patients progressing to nonalcoholic steatohepatitis (NASH), cirrhosis or die of liver failure. The earliest manifestation of NAFLD/NASH is hepatic steatosis (fatty infiltration of the liver) and regrettably, the utility of liver biopsy is very limited. MRI techniques based on a two-point or multi-point Dixon methods were demonstrated for fat-water separation and fat fraction quantification in liver. However, FT based techniques are much more susceptible to even minimal amounts of motion compared to methods using center-out k-space trajectories. We propose the use of Rosette Spectroscopic Imaging (RSI) for fat-water separation in liver.

Venkata Veerendranadh Chebrolu¹, Huanzhou Yu², Ethan K. Brodsky¹, Charles McKenzie, Scott B. Reeder¹

¹University of Wisconsin Madison, Madison, USA; ²GE Healthcare, Menlo Park, California , USA

This study investigates the impact of T2* decay on the ability of MRI using water-fat separation techniques to quantify fatty infiltration of the liver. A new generic analytical model, that includes the effects of line-width caused by T2* decay, is developed to the quantity the amount of different chemical species present in a specimen using MRI. The model is used to demonstrate the effect of line-width on the quantification of fat-fraction in non-alcoholic fatty liver disease. Our results demonstrate the need for methods that compensate for the effects of T2* decay when attempting to measure hepatic fat content.

Daniel Coman¹, Fahmeed Hyder¹

¹Yale University, New Haven, USA

In the last decade, a new non-invasive method for temperature and pH determination was developed based on temperature and pH dependencies of the ¹H chemical shifts emanating from TmDOTP^5-. More recently a similar temperature-sensitive probe was introduced, based on temperature dependence of the methyl ¹H chemical shift of TmDOTMA^-, which is pH independent. Here we report detection of TmDOTMA^- in the rat brain and compare in vivo temperature distributions obtained using TmDOTP^5- and TmDOTMA^- complexes. Additionally we discuss advantages and/or disadvantages of using these two complexes as exogenous probes for temperature and pH sensing in rat brain.

Jan Weis¹, Lars Johansson¹, Francisco Ortiz-Nieto¹, Håkan Ahlström¹

¹Uppsala University Hospital, Uppsala, Sweden

The unsuppressed water line is almost exclusively used as a concentration reference in quantitation of the muscle lipids. In spectroscopy of the skeletal muscle, the necessity for relaxation correction of water reference is a fundamental disadvantage. Alternatively, the fat signal can be used as the internal standard. In this study we compare both methods for the determination of lipid content in human skeletal muscle: relaxation effects sensitive water referenced single-voxel 1H MRS and relaxation effects robust high-spatial-resolution MRSI with fat (yellow bone marrow) as the internal or vegetable oil as the external concentration reference.

Patrick J. Bolan¹, Michael Garwood¹, Mark A. Rosen², Anthony Levering³, Jeffrey D. Blume⁴, James Gimpel³, Laura J. Esserman⁵, Nola Hylton⁵

¹University of Minnesota, Minneapolis, Minnesota, USA; ²University of Pennsylvania, Philadelphia, USA; ³ACRIN, Philadelphia, USA; ⁴Brown University, Providence, Rhode Island, USA; ⁵University of California, San Franciso, San Francisco, California , USA

The American College of Radiology Imaging Network (ACRIN) and the I-SPY network are supporting a multi-site clinical trial using quantitative single voxel spectroscopy (SVS) to quantify early response to neoadjuvant therapy in 140 patients. Ensuring that the spectroscopic [tCho] measurements can be performed with sufficient accuracy and precision to address this clinical aim is a difficult challenge, especially considering that field strength, scanner manufacturer, breast coil, and sequences will vary between sites. The goal of this abstract is to describe the design of the study and its quality control measures and present initial findings from the quality control phantom scans.

Hee-Won Kim¹, Gerald M. Pohost¹, Padmini Varadarajan², Milena Ocon¹, Rafit Drori³, Patrick Colletti¹, Adina E. Zeidler¹

¹University of Southern California, Los Angeles, California , USA; ²Loma Linda University Medical Center, Loma Linda, California , USA; ³Hillel Yaffe Medical Center, Hadera, Israel

The change of high energy phosphate (HEP) level in the heart of Latino Type-1 diabetes mellitus (DM) patients was evaluated during rest and stress at 3T in order to identify impaired left ventricular function that is associated with reversible, exercise-induced metabolite alteration in Type-1 DM patients. The significant drop of PCr/ATP was found during handgrip stress among 28% of the patient compared with normal control. It may suggest systemic microvascular disorder as bioenergetic changes present among Type-1 DM patients and the non-invasive 31P cardiac MRS may facilitate the prediction of cardiomyopathy in such patients.

Quantitative MRS of Glutathion 2 ¹³C Labeled Metabolites

Li An¹, Yan Zhang¹, David M. Thomasson¹, Lawrence L. Latour¹, Eva H. Baker¹, Jun Shen¹, Steven J. Warach¹

¹National Institutes of Health, Bethesda, Maryland, USA

In this work, we present a fully automated spectral registration method for frequency, phase, and linear baseline corrections. Instead of using a Lorenzian and/or Gaussian curve, target spectra for spectral registrations are generated from the data by a fitting and selection process. Spectral registrations are performed using broader ranges of spectral data instead of just one peak in the spectrum. A linear baseline correction is also done in the spectral registration process, which is important for GSH measurement where residual water baseline is a bigger problem than for other J-difference editing experiments.

Changho Choi¹, Nicholas J. Coupland², Peter Seres²

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

A new strategy for yield enhancement of glutathione (GSH) double-quantum (DQ) filtering for detection of the 2.95-ppm resonances is proposed. Following the double-quantum preparation and DQC encoding, a spectrally-selective 180° RF pulse is employed to interchange DQC and ZQC, followed by another DQC encoding step. With the large bandwidth of adiabatic 180° pulses for space localization, an edited signal amplitude ~100% with respect to the 90°-acquired multiplet can be achieved at an optimal sequence time. Preliminary results of numerical simulation and phantom tests are presented.

Shaolin Yang¹, Yihong Yang¹

¹National Institute on Drug Abuse, NIH, Baltimore, Maryland, USA

The reduced form of glutathione (GSH) is a major intracellular antioxidant. Due to its strong spectral overlap with other metabolites, spectral editing or LCModel analysis of short-echo PRESS or STEAM spectra has been suggested to measure GSH. We propose here a new strategy to resolve the spectral overlap between GSH and other metabolites by optimizing the timing parameters of a standard STEAM sequence through spectral simulation. The results show that the proton resonances of GSH at 2.54 ppm can be resolved from other metabolite resonances at 3T and 4T, which provides a potential to significantly improve the quantification of GSH. 

Li An¹, Yan Zhang¹, David M. Thomasson¹, Lawrence L. Latour¹, Eva H. Baker¹, Jun Shen¹, Steven J. Warach¹

¹National Institutes of Health, Bethesda, Maryland, USA

This work demonstrates the ability to consistently measure GSH in human brain on a commercial 3 T scanner with a scan time of nine minutes. A relatively long TE of 131 ms yields good GSH signal, smaller water baseline, as well as near in-phase GSH and co-edited NAA/NAAG resonances. A novel spectral registration method was also developed to perform frequency, phase, and linear baseline corrections without human intervention. From scanning normal volunteers and stroke patients, this work appears to be a step forward toward practical use of GSH measurement in clinical studies.

Melissa Terpstra¹, Dinesh Deelchand¹, Ivan Tkac¹, Pierre-Gilles Henry¹

¹University of Minnesota, Minneapolis, Minnesota, USA

Quantification of ascorbate (Asc) and glutathione (GSH) concentrations may be useful to study neurodegenerative disease. T₂ might differ between patients and controls. The goal of this study was to measure T₂ along with Asc and GSH concentrations in the human occipital cortex in vivo. Asc and GSH resonances were detected at multiple echo times using DEW MEGA-PRESS edited ¹H MR spectroscopy at 4 T. Reasonable T₂ were measured in one subject without compromising signal to noise for detection of Asc and GSH. Data can be averaged over several subjects to reduce uncertainty in measurement of T₂.

Ming Lu¹, ², Suhanti Banerjee¹, Gerald M. Saidel¹, Xin Yu¹

¹Case Western Reserve University, Cleveland, USA

To examine the effect of cytosolic redox state (NADH/NAD⁺) on malate-aspartate (M-A) shuttle activity in hearts, oxidative cardiac metabolism was studied using ¹³C NMR spectroscopy. Isolated rat hearts were perfused with either 2.5mM [2-¹³C]acetate or 2.5mM [2-¹³C]acetate plus glucose and lactate to increase cytosolic NADH/NAD⁺. Tricarboxylic acid (TCA) cycle flux (V_TCA) and M-A shuttle flux (V_M-A) were determined by least-square fitting of a novel multi-domain kinetic model of cardiac metabolism to dynamic ¹³C NMR spectra. Our results showed that V_TCA was unaltered by changes in cytosolic redox state, whereas V_M-A increased by 38% at high cytosolic NADH/NAD⁺.

Bernard Lanz¹, Kai Uffmann¹, Matthias T. Wyss², Bruno Weber², ³, Alfred Buck², Rolf Gruetter¹, ⁴

¹Ecole Polytechnique Fédérale de Lausanne, Lausanne, Switzerland; ²PET Center, University Hospital Zurich, Zurich, Switzerland; ³University Zurich, Zurich, Switzerland; ⁴Universities of Lausanne and Geneva, Lausanne and Geneva, Switzerland

Combining PET and NMR for cerebral labeling studies in vivo requires a common  model to compare and combine the results given by both techniques. The aim of the present study is to develop a metabolic model of neuro-glial metabolism suitable to assess metabolic rates resulting from an infusion of acetate labeled at the C1 position and to apply this model to experimental radiotracer time courses following  11C[1]-acetate infusion. We developed and applied successfully a model based on traditional NMR modeling to simulate and fit PET data.

Dinesh K. Deelchand¹, Chris Nelson¹, Alexander A. Shestov¹, Kamil Ugurbil¹, Pierre-Gilles Henry¹

¹University of Minnesota, Minneapolis, Minnesota, USA

In this work the feasibility of measuring neuronal-glial metabolism in rat brain in vivo using co-infusion of [1,6-¹³C₂]glucose and [1,2-¹³C₂]acetate was investigated. The distinct ¹³C spectral pattern observed in glutamate and glutamine directly reflected the fact that glucose was metabolized primarily in the neuronal compartment and acetate in the glial compartment. Time courses of concentration of singly and multiply-labeled isotopomers of glutamate and glutamine were obtained. We expect that dynamic metabolic modeling of these new ¹³C isotopomer data using two-compartment neuronal-glial models will lead to a more precise determination of metabolic rates, particularly the rate of glutamate-glutamine cycle.

Multi-Modality Contrast Agents

Hisanori Kosuge¹, Masahiro Terashima¹, Sarah Sherlock, Jin Hyung Lee, Hongjie Dai, Michael V. McConnell

¹Stanford University, Stanford, California , USA

The novel graphite/metal core-shell nanocrystals (CN) are effectively taken up by macrophages in vitro and in vivo. This particles show promising properties as MRI contrast agents for noninvasive evaluation of atherosclerosis.

Chuqiao (Tom) Tu¹, Ryan Ngao¹, Angelique Louie¹

¹UC Davis, Davis, California , USA

We have developed a small molecule gadolinumm contrast agent that changes conformation in response to redox activity, thereby modulating relaxivity.  The synthesized contrast agent responds to NADH at biologically relevant concentrations.  Significant changes in MR image contrast are produced in response to chemical reductants and oxidants.

Matti M. van Schooneveld¹, Esad Vucic², Rolf Koole¹, Yu Zhou², Cheuk Y. Tang², Klaas Nicolay³, Andries Meijerink¹, Zahi A. Fayad², Willem J.M. Mulder²

¹Utrecht University, Utrecht, Netherlands; ²Mount Sinai School of Medicine, New York, New York, USA; ³Eindhoven University of Technology, Eindhoven, Netherlands

We investigated the pharmacokinetics and bio-distribution of a novel silica-based multimodal nanoparticle. This probe consists of a quantum dot (for fluorescence imaging) incorporated in a 35 nm silica nanoparticle, which is coated by paramagnetic (for MRI) and pegylated lipids to overcome the limited bio-applicability of silica particles. MRI, optical techniques, relaxation measurements, and ICP-MS were used to determine the blood half-life value and the tissue distribution at anatomical and cellular level. A 10-fold increase in the half-life value was observed in case the particles were lipidic coated as compared to non-coated particles (t1/2 = 17 min vs. 180 min).

                       1698.     Kinetics of Avidin-Induced Clearance of Biotinylated Bimodal Liposomes for Improved MR Molecular
                                 Imaging

Geralda A.F. van Tilborg¹, Gustav J. Strijkers¹, Emilie M. Pouget¹, Chris P.M. Reutelingsperger², Nico A.J.M. Sommerdijk¹, Klaas Nicolay¹, Willem j.m. Mulder³

¹Eindhoven University of Technology, Eindhoven, Netherlands; ²University of Maastricht, Maastricht, Netherlands; ³Mount Sinai School of Medicine, New York, USA

Bimodal liposomes, carrying large amounts of gadolinium-lipids, were recently proposed as potent contrast agents for MR molecular imaging. These nanoparticles display a long circulation time, which enables massive accumulation at the targeted site but also keeps the target to background ratio low for a prolonged period. In this study an avidin chase was designed to rapidly clear paramagnetic biotinylated liposomes from the blood circulation. Avidin-induced alterations in the biodistribution and blood clearance kinetics were studied. The ability to rapidly clear circulating contrast agent opens up exciting possibilities to study targeting kinetics and to optimize nano-particulate contrast agent formulations.

Marlies Oostendorp¹, ², Kim Douma², Tilman M. Hackeng², Anouk Dirksen², Mark J. Post², Marc A.M.J. van Zandvoort², Walter H. Backes¹, ²

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

Paramagnetic quantum dots (pQD), Gadolinium-loaded fluorescent semiconductor nanoparticles, are becoming more popular as molecular MRI contrast agent. Besides decreasing the local T₁ relaxation time, it is expected that the semiconductive properties of pQD result in field inhomogeneities, which contribute to T₂ relaxation time shortening. It is therefore unknown whether T₁ or T₂ based methods show the strongest contrast enhancement upon pQD injection. Here, we applied a quantitative molecular MRI method to analyze tissue T₁ and T₂ and the changes thereof induced by pQD targeted to the angiogenic tumor vasculature. Significant differences were only found for T₁ based effects.

Donghoon Lee¹, Conroy Sun¹, Omid Veiseh¹, Chen Fang¹, Stacey Hansen², Miqin Zhang¹, James Olson², Richard Ellenbogen¹

¹University of Washington, Seattle, Washington, USA; ²Fred Hutchinson Cancer Research Center, Seattle, Washington, USA

Nanoprobes were developed for MR and optical target imaging of intracranial tumors. Currently, nanoconjugates have been limited in target imaging for intracranial tumors due to insufficient accumulation within tumor and inability to cross the blood brain barrier (BBB). Our developed nanoconjugates were iron oxide nanoparticles coated with PEG (polyethylene glycol)-chitosan conjugated with chlorotoxin. Images acquired by MR and optical imaging revealed specific particle uptakes by intracranial tumors for 2-5 days post injections and evidence of the nanoprobes passing through the BBB for mouse model with BBB intact. Histology data also supported the results.

Wilfried Reichardt¹, Robert Zeiser², Jürgen Hennig¹, Dominik von Elverfeldt¹

¹University of Freiburg, Freiburg, Germany; ²Stanford University School of Medicine, Stanford, USA

Synopsis:Insight into the trafficking pattern of dendritic cell (DC) populations into lymphoid and non-lymphoid tissues after allogeneic hematopoietic cell transplantation (aHCT) is critical to understand the complex processes that regulate graft-versus-host disease. We were able to generate a platform for in vivo bioluminescence (BLI) and magnetic resonance imaging (MRI) based cell trafficking studies by labelling luciferase transgenic DC with superparamagnetic iron oxide (SPIO) Nano particles bound to a murine IgG antibody.

                  1702.     Dual Modality Imaging of Phosphatidylcholine-Specific Phospholipase C in DU145 Prostate Cancer
         Cells and Solid Tumors

Theresa Meganne Mawn¹, Daniel-Joseph Leung¹, Nancy Beardsley¹, Anitoliy Popov¹, E. James Delikatny¹

¹University of Pennsylvania, Philadelphia, Pennsylvania, USA

We have employed a dual modality approach to detect changes in lipid metabolites and phosphatidylcholine-specific phospholipase C (PC-PLC) activity in prostate cancer cells and solid tumors. MRS studies have consistently revealed the presence of elevated levels of phosphocholine (PCho) in various types of cancer. However, it remains uncertain whether these elevations are due to choline transport and choline kinase, or occur catabolically through the action of PC-PLC. By combining MRS with a near-infrared (NIR) PC-PLC-activated optical imaging probe, we are able to directly image the relative contributions of the anabolic and catabolic pathways to PCho elevations in cancer.

Kim Anne Radermacher¹, Nelson Beghein¹, Sebastien Boutry², Sophie Laurent², Luce Vander Elst², Robert N. Muller², Benedicte F. Jordan¹, Bernard Gallez¹

¹Université Catholique de Louvain, Brussels, Belgium; ²University of Mons-Hainaut, Mons, Belgium

The aim of the study was to develop a molecular marker for non invasive diagnosis and monitoring in the early stages of inflammation. An E-selectin ligand was coupled to ultrasmall particles of iron oxide (USPIO). After intravenous injection of the grafted or ungrafted USPIO, their concentration was evaluated in inflamed muscle ex vivo by an EPR X-band and in vivo by an EPR L-band, as well as by MRI. All three methods demonstrate that the specific targeting of grafted USPIO was twice higher in inflamed tissues than using the ungrafted ones.

Non-Proton MRI

Robin A. de Graaf¹, Peter B. Brown¹, Kevin L. Behar¹

¹Yale University, New Haven, Connecticut, USA

Despite the rich abundance of oxygen in biologically relevant compounds, oxygen detection by NMR is limited largely by the low natural abundance of oxygen-17 (0.037%) and the broad resonances (spin 5/2). However, the favorable T1 relaxation constants, together with the increased spectral dispersion and sensitivity at high magnetic fields allowed the detection of natural abundance 17O NMR spectra from rat brain in vivo. A large number of resonances originating from phosphate, sulfate and carbonyl groups were readily detected.

Jesus Pacheco¹, ², Dawen Zhao¹, Vikram D. Kodibagkar¹, Angelina Contero¹, Debabrata Saha³, Sebastian Cerdan², Ralph P. Mason¹

¹University of Texas Southwestern Medical Center, Dallas, Texas, USA; ²Instituto de Investigaciones Biomédicas "Alberto Sols" - CSIC, Madrid, Spain; ³University of Texas Southwestern Medical Center, Dalas, Texas, USA

Hypoxia is recognized to influence tumor response to therapy. ¹⁹F MRI using the reporter molecule hexafluorobenzene has been demonstrated to measure pO₂ dynamics in tumors growing in animals- the  FREDOM  approach. Here, we explore the use of  FREDOM   to evaluate pO₂ distributions and response to hyperoxic gas breathing in H460 lung tumor xenografts growing in rats. Oxygenation is found to range from hypoxia to well oxygenated and the H460 tumors are found to be highly responsive to hyperoxic gas breathing. The regions, which are initially best oxygenated showed greatest response. These results show further application of ¹⁹F MRI oximetry.

Maxim Terekhov¹, Alexander Scholz², Ursula Wolf¹, Julien Rivoire¹, Wolfgang Schreiber¹

¹Mainz University Medical School, Mainz, Germany; ²Mainz University Medical School,, Mainz, Germany

MRS and MRI of drugs are emerging tools in the development and understanding of pharmacons. Some drugs, e.g. the inhaled anesthetics such as sevoflurane, embody fluorine. In these cases, 19F MRI allows for in vivo imaging.without background signal. We established a large animal model for NMR/MRI investigations of fluorinated anesthetics at 1.5T medical scanner that may be considered as a further step towards human studies. This imaging and spectroscopy techniques may allow for a regional analysis of pharmacokinetic and better understanding of molecular interactions of inhaled anesthetics.

Daniel Kalthoff¹, ², Ursula Wolf¹, Kerstin Münnemann¹, Wolfgang Günther Schreiber¹

¹Mainz University Medical School, Mainz , Germany; ²Max Planck Institute for Neurological Research, Cologne, Germany

A method of phase contrast MRI was implemented to study flow of ³He and especially ¹⁹F gases which are currently used in lung imaging. Experiments were performed on a custom-made gas flow phantom that produced a continuous gas flow with in-vivo-like characteristics in flow rates and geometry. Turbulent and laminar flow were observed as expected in C₄F₈ and ³He respectively. A comparison of MRI data to simultaneous flow sensor measurements showed excellent agreement and verified the accuracy of the presented method. Thus, phase contrast MRI of ¹⁹F and ³He appears to suitable for quantitative investigation of lung ventilation.

Oliver Lips¹, Jochen Keupp¹

¹Philips Research Europe, Hamburg, Germany

¹⁹F-MRI bears potential as a tracer technique for applications in diagnosis and therapy monitoring when combined with fluorine-labeled agents. However, ¹⁹F-MRI suffers from chemical shift (CS) artifacts due to multi-line spectra of typical ¹⁹F-agents or the presence of multiple agents exhibiting different CS. Hence, countermeasures are needed, which do not spoil the SNR or significantly increase scan time. A concept is presented, which utilizes different coil sensitivity profiles to correct for CS artifacts applying a SENSE-like algorithm in the frequency encoding direction. Simulations and initial experiments were performed. The results indicate, that the method has potential to improve multi-line ¹⁹F-imaging.

Xin Liu¹, Zhong-Xing Jiang¹, Y. Bruce Yu², Eun-Kee Jeong¹

¹University of Utah, Salt Lake City, USA; ²University of Maryland, College Park, USA

1H MR imaging of water protons in tissue is widely used for studying the pharmacokinetics of Gd-chelate based endogenous contrast agent. The concentration of the delivered compound is indirectly measured by analyzing the change in the proton MRI signal intensity. this method is not practical in most of dynamic MR imaging due to the long imaging time for T1 mapping. In this report, 19F MRI is presented as an alternative method to access the drug quantity. The concentration of fluorine compound is directly proportional to the 19F signal in density weighted 19F-MRI.

Joerg Magerkurth¹, Ulrich Pilatus¹

¹Johann Wolfgang Goethe Universität, Frankfurt a. M., Germany

Voxelwise coregistration for combining quantitative 31P and 1H MR spectroscopy ishampered by the rather large voxel size and the poor point-spread-function (PSF) of 31PMRSI data.  The problem can be addressed by minimizing the partial volume effects usingimage segmentation and linear regresson analysis as described previously for 31P  and 1H. The approach was evaluated to determine concentration differences between gray and whitematter for the three components of the trimethylamine (tCho) signal at 3.2 ppm in 1HMRSI. This study demonstrates that application of tissue segmentation in combination with quantitative 31P and 1H MRSI can resolve the distribution of these components in gray andwhite matter.

Aiming Lu¹, Ian C. Atkinson¹, Xiaoong Joe Zhou¹, Theodore Claiborne¹, Keith R. Thulborn¹

¹Univ. of Illinois at Chicago, Chicago, Illinois, USA

31P MRI can provide quantitative bioenergetic information of human brain non-invasively. However, quantitative determination of the spatial distribution of 31P is challenging due to its low MR sensitivity. With the increased sensitivity at 9.4T, initial 31P images with reasonable SNR have been obtained on a phantom with concentration comparable to that of PCr in human brain within 10 minutes using a RARE sequence, which demonstrates the potential of quantitative 31P metabolic imaging at 9.4T.

Tyler Meldrum¹, Monica A. Smith¹, Leif Schroeder¹, Thomas Lowery¹, David Wemmer¹, Alex Pines¹

¹University of California, Berkeley, Berkeley, California , USA

We have developed a modular approach to xenon-based biosensor synthesis that facilitates optimization of the sensors for increasing sensitivity.  By synthesizing a biosensor with a negatively-charged side chain, together with increased temperature (37 C) and the HYPER-CEST detection method, we can detect biosensor concentrations as low as 10 nM.

                  1743.     Measurement of Pressure from the Diffusion Coefficient of 3He Gas

Jim M. Wild¹, Steven R. Parnell¹

¹University of Sheffield, Sheffield, UK

Non-invasive measurement of pressure with NMR could have a variety of applications in-vivo (e.g. measurement of blood pressure) and in non-medical applications (e.g. pressure jets in fluid dynamics). In this work an inverse relation between the pressure of 3He and the measured (apparent) diffusion coefficient (D) was observed using intermediate diffusion time pulsed gradient spin echo (PGSE) methods. Extension of the technique for measuring pressure in-vivo in micro bubbles is proposed using different gas mixtures and short time scale diffusion sequences.

Matthew David Robson¹, Stefan Neubauer¹

¹Oxford University, Oxford, UK

We have implemented and evaluated a 3D stack of spokes acquisition at 3T using an 8-channel phased array sodium coil.  This incorporates an ultra-short TE readout, and can be run as an SSFP sequence.  This combination of technologies boosts the signal to noise, and provides higher image quality than have previously been demonstrated.

Diffusion Acquisition

Emine Ulku Saritas¹, Jin Hyung Lee¹, Dwight G. Nishimura¹

¹Stanford University, Stanford, California , USA

The optimization of b-value for two-point apparent diffusion coefficient (ADC) estimation schemes were previously investigated, assuming high signal-to-noise ratio (SNR) imaging and considering the effects of T₂-weighting. Here, we investigate both the dependence of the optimum b-value on the SNR of the imaging scheme and also on the T₂ of the tissue of interest. The results of this work are especially important for high-resolution diffusion-weighted imaging, which intrinsically suffers from low SNR.

Minzhi Gui¹, Zhe Hu¹, Konstantinos Arfanakis¹

¹Illinois Institute of Technology, Chicago, Illinois, USA

In the image reconstruction process for Turboprop diffusion tensor imaging (DTI), weights representing the quality of the data are estimated for each k-space blade (quality weighting, QW), and used during reconstruction to reduce the contribution of corrupted blades to the final image. This weighting procedure reduces artifacts due to: bulk motion, odd/even echo fluctuations, and other sources of k-space data variation between blades. In this work, the degree to which the QW also compensates for the increase in the total variance of the diffusion tensor due to cardiac-induced brain pulsation is investigated in both simulations and experiments on humans.

Huiling Peng¹, Gady Agam¹, Minzhi Gui¹, Konstantinos Arfanakis¹

¹Illinois Institute of Technology, Chicago, USA

Accurate spatial normalization in diffusion tensor imaging (DTI) requires data with minimal artifacts. Conventional spin-echo echo-planar DTI (SE-EPI-DTI) suffers from severe B₀-related image artifacts, which are prominent in certain brain regions, such as the brainstem, the frontal and temporal lobes. In contrast, Turboprop-DTI is relatively immune to image artifacts caused by magnetic field inhomogeneities. In this study, the performance of spatial normalization of DTI data obtained with SE-EPI-DTI and Turboprop-DTI was evaluated. The results demonstrated that, Turboprop-DTI may lead to more accurate spatial normalization than SE-EPI-DTI in regions where the latter suffers from magnetic field inhomogeneity artifacts.

Yeji Han¹, JinYoung Hwang², HyunWook Park²

¹University Hospital Freiburg, Freiburg, Germany; ²Korea Advanced Institute of Science and Technology, Republic of Korea

The proposed method aims to generate high-resolution diffusion-weighted (DW) images from a reduced number of projection data acquired with any kind of existing radial acquisition sequences. After acquiring less than a half of the required number of projection data, a reasonable estimation of the unacquired part of the projections is performed. After that, the proposed process refines the estimated projection data utilizing information of a full-resolution non-DW image. By using the proposed method, the acquisition time can be decreased while preserving the image quality. The simulation and experiment results show that the proposed method gives high-resolution DW images from the reduced-view projection dataset.

Hiroyuki Kabasawa¹, Shigeki Aoki², Yoshitaka Masutani², Osamu Abe², Kuni Ohtomo²

¹GE Yokogawa Medical Systems, Hino-shi, Japan; ²The university of Tokyo, Bunkyo-ku, Japan

We demonstrated that VERSE implementation to PROPELLER reduced SAR and that helped to increase volume coverage and reduce scan time to clinically acceptable range. Slice coverage was increased from less than slice to 28 slices with 8sec of repetition time. The proposed method has potential for acquiring diffusion tensor images without compromising both image geometry accuracy and slice coverage.

Xianfeng SHI¹, Eugene Kholmovski¹, Eun-Kee Jeong¹

¹University of Utah, Salt Lake City, USA

3D single-shot Diffusion-Weighted STimulated-EPI (3D ss-DWSTEPI) suffers from low SNR because it uses only half the diffusion-weighted magnetization stored into the longitudinal plane by a 90o RF and spoils another half. The degree of susceptibility-induced distortion in EPI-type acquisitions is inversely proportional to the speed of the k-space traveling in phase-encoding direction. To improve SNR and reduce the susceptibility-induced distortion, 3D ss-DWSTEPI in-plane readout is shorten by reducing FOV in phase-encoding direction. Parallel imaging technique has been implemented in 3D ss-DWSTEPI to reduce the geometric distortion and a method has been developed to improve SNR by utilizing whole diffusion-weighted magnetization.

Irvin Teh¹, ², David J. Larkman¹, Xavier Golay², ³

¹Imperial College London, London, UK; ²Singapore Bioimaging Consortium, Singapore, Singapore; ³National Neuroscience Institute, Singapore

Diffusion weighted imaging is widely used for studying tissue integrity and structure. Single shot DW-EPI is commonly used to improve motion robustness and efficiency, but suffers from lower resolution and magnetic field inhomogeneity effects. At high fields, image distortion and short T2* make multishot FSE methods more attractive. Periodically Rotated Overlapping ParallEL Lines with Enhanced Reconstruction (PROPELLER) has been demonstrated in humans as a self navigating and efficient sequence robust to motion artifacts and image distortion. Here, we present the novel application of DW-FSE-PROPELLER in in-vivo imaging of mice at 9.4T.

Sheng Fang¹, Kui Ying¹

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

Multi-shot diffusion-weighted imaging suffers from motion-induced phase error that varies from shot to shot. Conjugate-gradient (CG) based reconstruction may not adequately correct this error if phase estimation is not accurate enough. Besides, the CG algorithm only utilizes the phase information from the navigator image. In this work, we propose an effective k-space weighted multi-channel regularization algorithm based on Tikhonov regularization to fully utilize acquired navigator data and further reduce the residual motion-induced aliasing artifacts left by standard Tikhonov regularization. Both single-coil and multi-coil simulations have been performed to demonstrate the feasibility of the proposed method.

Tzu-Cheng Chao¹, Yi-Jui Liu², Teng-Yi Huang³, Hsiao-Wen Chung¹, Cheng-Yu Sandy Chen⁴

¹National Taiwan University, Taipei, Taiwan; ²Feng-Chia University, Taichung, Taiwan; ³National Taiwan University of Science and Technology, Taipei, Taiwan; ⁴Tri-Service General Hospital, Taipei, Taiwan

Susceptibility-induced geometric distortions in diffusion tensor EPI limit the image resolution in regions showing strongly inhomogeneous magnetic fields. In this work, we proposed an alternative inner volume excitation scheme for EPI with reduced FOV, which employed a pre-inversion approach suitable for multi-slice acquisition without severe compromise in SNR. Simulations and experimental results for diffusion tensor imaging in the cerebellum showed that this approach could provide simultaneous signal uniformity and good SNR throughout all slices. Susceptibility-related distortions were effectively minimized via a shortening of the data acquisition window to achieved DTI of the cerebellum at 0.86mm in-plane resolution.

Paul S. Morgan¹, Ron J. Coxon¹, Josef Habib¹, Penny A. Gowland¹, Richard Bowtell¹

¹University of Nottingham, Nottingham, UK

Application of diffusion-weighted (DW) MRI on a 7 T scanner holds the promise of significantly increased spatial resolution. However the challenges involved in implementing DW-EPI at 7 T require re-evaluation of the best sequence for obtaining DW images. Here we have therefore implemented four different DW-EPI sequences at 7 T and compared the results in terms of signal-to-noise and spatial distortion, as well as studying the effect of parallel imaging. Twice-refocused DW-EPI acquisitions were found to offer the best compromise between SNR and spatial distortion.

Samantha J. Holdsworth¹, Stefan Skare¹, Rexford D. Newbould¹, Anders Nordell², Roland Bammer¹

¹Stanford University, Palo Alto, California , USA; ²Karolinska University Hospital, Stockholm, Sweden

“Short-Axis readout Propeller EPI” (SAP-EPI), its dual-blade variant (dual-blade SAP-EPI), and Readout-Segmented EPI (RS-EPI) have been proposed as variants of EPI for high resolution diffusion-weighted (DW) imaging. While there are many similarities between these sequences, there are various differences between them which affect the scan efficiency and overall image quality. The purpose of this abstract is to make an initial assessment of these schemes with regard to diffusion preparation/acquisition ratio, normalized scan time, and image quality for a typical set of scan parameters we have been using for high resolution GRAPPA-accelerated DW imaging.

Jeroen Siero¹, Hans Hoogduin¹

¹University Medical Center Groningen / University of Groningen, Groningen, Netherlands

A non-uniform gradient direction scheme for DTI is presented based on the non-uniform fiber orientations in the human brain. A comparison with a uniform 60 direction scheme shows an increase in FA values and a decreased error for the dominant fiber directions.

Rexford David Newbould¹, Stefan Skare¹, Roland Bammer¹

¹Stanford University, Stanford, California , USA

Magnitude averaging is very common in diffusion imaging, as it avoids random phase offsets in the data from the diffusion-sensitizing gradients. Noise in magnitude MR images is Rician distributed. In low SNR magnitude images, such as diffusion images, the Rician distribution approaches a Rayleigh distribution.  This introduces a severe non-zero bias of the signal expectation value and underestimates the calculated ADC. The deflection of the diffusion attenuation can reduce the CNR and mimic biexponential decay. Combining magnitude diffusion images for signal averaging or to create isotropic DWI images does not lower this mean, whereas combining complex diffusion images does.

Ha-Kyu Jeong¹, ², Adam W. Anderson¹

¹Vanderbilt University, Nashville, Tennessee, USA

In this study, high angular resolution diffusion imaging (HARDI) measurements performed at 7T were corrected for distortions due to B0 field inhomogeneities using a static field map. After the correction, the anatomy of individual brain slices and corresponding fiber tracts are much closer to those of multishot gradient echo images, which have relatively small distortions. The improvements in spatial accuracy of the HARDI data are critical for quantitative applications of diffusion MRI at high field strengths.

Ralf Luetzkendorf¹, Tobias Moench¹, Maurice Hollmann¹, Sebastian Baecke¹, Claus Tempelmann¹, Joerg Stadler², Johannes Bernarding¹

¹University of Magdeburg, Magdeburg, Germany; ²Leibniz Institute for Neurobiology, Magdeburg, Germany

Higher B0 fields are expected to increase the signal-to-noise ratio which should lead to higher spatial resolution of DTI. At high fields B1 inhomogeneities are strong and result in sever distortions and signal voids in DTI. We used parallel imaging techniques combined with data post-processing to reduce distortions and to reach a high spatial resolution. The same starting maps and waypoint maps to track the apparent fibers of the same volunteer were used in both 3T and 7T to ensure the comparability of the data.

Allen W. Song¹, Zoe Englander¹, Bin Chen¹

¹Duke University, Durham, North Carolina, USA

DTI images suffer from fluid contaminations from CSF and vasculature, especially in the ventral brain regions, periventricular space and cortical surface, because of the hyper-intensity of fluid in the T2 weighted baseline image. Early methods for removing these confounds have relied on inversion recovery technique which could result in a large reduction in signal-to-noise ratio and temporal resolution. Here we propose a method that applies flow-sensitive diffusion weighting to the baseline scan, and evaluate its advantage on improving the delineation of white matter anisotropy. Much improved characterization of white matter anisotropy is found with further separation from the gray matter.

Ted P. Trouard¹, ², John Wade Totenhagen¹, Ali Bilgin¹

¹University of Arizona, Tucson, Arizona , USA

Compressed Sensing (CS) reconstruction enables generation of images from dramatically undersampled MRI data.  In this work, CS methodology is applied to undersampled diffusion-weighted radial-FSE datasets to evaluate the utility of CS to reduce scan times in multishot DW-radial-FSE acquisitions.  DW-radial-FSE has been shown to produce diffusion-weighted images with high spatial resolution and little artifact from motion and/or magnetic field inhomogeneities, but requires considerably more acquisition time than single-shot DW-MRI methods.  CS reconstruction of severely undersampled radial-FSE data is shown to produce diffusion-weighted images and ADC maps without the streaking artifacts generated by filtered backprojection or regridding reconstruction.

Seong-Eun Kim¹, Eun-Kee Jeong¹, Tae Ho Kim¹, J Rock Hadley¹, Emilee Minagla¹, Jeff Anderson¹, Dennis L. Parker¹

¹University of Utah, Salt Lake City, USA

In this work we present the in vivo DTI measurement of the normal human optic nerve using two different interleaved multiple inner volume diffusion weighted EPI sequences and a 20 channel dedicated optic nerve coil on 3T system. The results indicate that both 2D IMIV-DWEPI and 2D rFOV-DWEPI allow reliable DTI measurement of optic nerve and ocular muscle. Quantitative analysis of diffusivity and anisotropy in the human optic nerve area appears to be feasible and has potential to enable more sensitive and specific detection and monitoring of structural changes caused by pathology including multiple sclerosis, optic neuritis, and neurofibromatosis.

Hua Guo¹, ², Allen W. Song², Ed X. Wu¹

¹The University of Hong Kong, Pokfulam, Hong Kong; ²Duke University, Durham, North Carolina, USA

Due to computation and design complexity, there is no design so far for isotropic and background gradient independent diffusion gradients. In this abstract, we present a simple but robust method to design orientation invariant and background magnetic field insensitive diffusion gradients. Although the diffusion weighting efficiency is reduced, this design provides a set of diffusion gradients that may be used to map ADC efficiently and accurately or in presence of severe field inhomogeneity such as induced by metal implants.

Joelle E. Sarlls¹, Stefano Marenco¹, Carlo Pierpaoli¹

¹National Institutes of Health, Bethesda, Maryland, USA

Sub-millimeter isotropic voxel diffusion tensor images (DTI) of the optic chiasm are presented, in vivo, for the first time. Data are acquired with a modified radial-FSE sequence, which allows anatomical details of structures in the brain to be studied  with a voxel size about 10 times smaller than that currently achievable with EPI-based DTI. These images clearly depict the decussation of the medial fibers in the optic chiasm. Geometric distortion and susceptibility artifacts are virtually eliminated in these high-resolution DTI maps.

Diffusion & Perfusion Studies in Animal Models

                   1867.     Rapid High-Field Diffusion MR Histology: Image-Based Phase Correction for Diffusion-Weighted RARE

Julian Michael Tyszka¹, Lawrence R. Frank

¹California Institute of Technology, Pasadena, California , USA

Two phase correction schemes for eddy-current ghosting in high-field diffusion-weighted RARE imaging are presented. The schemes are tailored for efficient MR diffusion histological studies of samples such as fixed rodent brains, in which physiological motion is absent and per-acquisition navigation can be eliminated. The effectiveness of both approaches is demonstrated in rat brain samples.

                  1868.     Altered Fiber Connectivity in Adult Brain of PAX6 Knock-Out Mice Revealed by DTI in Vivo

Susann Boretius¹, Anastassia Stoykova¹, Roland Tammer¹, ², Jens Frahm¹, Thomas Michaelis¹

¹Max-Planck-Institut für biophysikalische Chemie, Göttingen, Germany; ²DFG Center of Molecular Physiology of the Brain (CMPB), Göttingen, Germany

Cortex specific conditional knock-out of transcription factor PAX6 at the beginning of neurogenesis in mice results in almost full abolishment of the upper cortical layers. To assess putative structural abnormalities in the cortex at maturity, pairs of mutant and control mice were studied by diffusion tensor imaging (DTI). Tracking of major neuronal fiber bundles revealed a lack of interhemispheric connectivity as well as pronounced fiber reorganization in the septal region and fimbria. These findings indicate an important role of PAX6 in establishment of the intracortical fiber connectivity. 

                  1869.     Quantifying Cerebral Blood Volume Over the Mouse Cerebral Cortex Using Micro-CT Co-Registered
                                to an MRI Anatomical Atlas

Brige P. Chugh¹, Jason P. Lerch¹, Lisa X. Yu¹, R Mark Henkelman¹, John G. Sled¹

¹Mouse Imaging Centre,Toronto Centre for Phenogenomics, Hospital for Sick Children, Toronto, Canada

Measurement of the cerebral blood volume (CBV) in local regions of the mouse brain will be useful to describe the phenotypes of models of neurodegenerative diseases that alter microvasculature including Alzheimer’s disease. A limitation of current methods to determine CBV in mice, including MRI methods sensitive to magnetic susceptibility changes, is that they do not provide absolute quantification of CBV over the entire mouse cerebral cortex. To overcome this difficulty, we developed a method to determine absolute CBV over the mouse cortex using micro-computed tomography (micro-CT) co-registered to an MRI anatomical brain atlas.

                  1870.     Diffusion Tensor Imaging Detects and Characterizes Proliferative Diabetic Retinopathy in the
                                 Murine Retina

Saurav Chandra¹, Angelos Barmpoutis¹, John Forder¹

¹University of Florida, Gainesville, USA

The specific aim in this study is to prove that the alterations in the eye following damage (by laser damage and diabetic retinopathy) can be detected by DTI. This is done by comparing the fractional anisotropies in the retinas of controls versus diabetic eyes. The mice had been subjected to Type I diabetes. The experiments were successful in determining significant differences between the fractional anisotropies of retinas in controls compared to the lasered as well as diabetic retinas. Future studies will involve in vivo imaging of mice with various stages of the disease.

                  1871.     In Vivo Quantification of Lamina T1, T2, and Apparent Diffusion Coefficient in the Mouse Retina at 11.74T

Junjie Chen¹, Qing Wang¹, Huiying Zhang¹, Xiaoxia Yang¹, Jian Wang², Bruce A. Berkowitz³, Samuel A. Wickline¹, Sheng-Kwei Song¹

¹Washington University, St. Louis, Missouri, USA; ²Chinese Pharmaceutical Association, Beijing, People's Republic of China; ³Wayne State University, Detroit, Michigan, USA

Quantitative MRI of the mouse retina (~ 250 µm thick) is challenging due to its limited thickness.  In the present study, T1 and T2 relaxation time constants, and the directional apparent diffusion coefficient (ADC) in three MR-detected retina layers of C57/BL6 mice were measured at 47 x 47 x 400 µm ³ resolution.  This study establishes, for the first time, normative metrics of T1, T2, and ADC of the mouse retina. These MR parameters are expected to be useful in future evaluation of developmental and pathological alterations of retinal cell layers in mice. 

Yingxia Li¹, Haiying Cheng¹, Qiang Shen¹, Moon K. Kim², Darin E. Olson¹, Peter M. Thule¹, Machelle T. Pardue², Timothy Quyen Duong¹

¹Emory University, Atlanta, Georgia, USA; ²Department of Veterans Affairs Medical Center, Atlanta, Georgia, USA

Vision loss due to retinal degeneration is a major problem in clinical ophthalmology. We have previously reported a thinning of the retina and perturbed BOLD fMRI responses to physiologic challenges in the retina of an animal model of progressive retinal degeneration (Royal-College-of-Surgeons rats). In this study, we extend previous findings by studying basal blood flow (BF) and physiologically induced BF changes in RCS rat retinas and age-matched controls at 90x90x1500-ƒÝm. Quantitative BF was measured using the continuous arterial-spin-labeling technique. MRI provides quantitative BF data without depth limitation and large field-of-view and could complement existing retinal imaging techniques.

John Carr¹, David Buckley¹, Jean Tessier², Geoff Parker¹

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.

Fernando F. Paiva¹, ², Erica C. Henning¹, Alberto Tannus², Afonso C. Silva¹

¹National Institutes of Health, Bethesda, Maryland, USA; ²Universidade de Sao Paulo, Sao Carlos, Brazil

Cerebral blood flow under basal and hypercapnic condition (5% and 10% CO2) were obtained in normotensive and hypertensive rats using a continuous ASL EPI sequence with a a three coil system comprised of a homogeneous volume excitation coil, a receive-only quadrature surface coil and a dedicated labeling RF coil. Higher CBF values were found in hypertensive rats both under normocapnia and hypercapnia. While the vascular reactivity under 5% carbon dioxide was not significantly different amongst both strains, the hypertensive rats displayed impaired reactivity when submitted to 10% CO2.

Fernando F. Paiva¹, Bojana Stefanovic¹, Yoshiyuki Hirano¹, Alberto Tannus², Afonso C. Silva¹

¹National Institutes of Health, Bethesda, Maryland, USA; ²Universidade de Sao Paulo, Sao Carlos, Brazil

Dynamic arterial spin labeling (DASL) is an efficient way to quantitatively measure cerebral blood flow (CBF), the transit-time tƒnof the endogenous labeled arterial water and T1 of the brain. In this work, we combine DASL with fMRI experiments during somatosensory stimulation in rats. We show that the combination of DASL with fMRI constitutes an interesting approach to obtain, in a single experiment, dynamic quantification of both resting and functional hemodynamics with improved efficiency and SNR. This new methodology is poised to become a versatile experimental plat-form for studying the spatial and temporal characteristics of functional cerebral hemodynamics.

Xiao-xia Du¹, Fu-chun Lin², Hao Lei¹

¹Wuhan Institute of Physics and Mathematics, Chinese Academy of Science, Wuhan, People's Republic of China; ² Wuhan Institute of Physics and Mathematics, Chinese Academy of Science, Wuhan, People's Republic of China

Agonists/antagnoists of alpha2-adrenoceptor and imidazoline receptor are known to have the ability to affect cerebral blood flow (CBF). Xylazine, clonidine and moxonidine have different receptor affinities for alpha2-adrenoceptors and imidazoline receptors. In this study, continuous arterial labeling (CASL) perfusion imaging was used to investigate the effects of xylazine, clonidne and moxonidine on CBF in rats. The results showed that, at the dosages used, the three drugs induced similar cardiovascular responses, but significantly different effects on CBF.

Samir Kamalesh Amin¹, Faridis Serrano¹, Tomoya Terashima¹, Lawrence Chan¹, Lingyun Hu¹, Robia G. Pautler¹

¹Baylor College of Medicine, Houston, Texas, USA

Diabetes is associated with microvascular disease and believed to increase cerebrovascular complications, particularly acute ischemic strokes. Studies have shown that both acute and chronic hyperglycemia cause cerebral blood flow (CBF) reduction, however, little is known about the mechanism of change. We used an in vivo mouse model of diabetes not only to evidence reduced CBF utilizing Flow-sensitive Alternating Inversion Recovery (FAIR) arterial spin labeling (ASL), but also to determine if this vascular effect is reversible by modifying glucose levels with insulin. In conclusion, we observed CBF reduction in a diabetic mouse model (STZ) and witnessed an “insulin rescue” in CBF.

Martin M. Pike¹, ², William D. Rooney², Xin Li², Christine Neumann Stoops¹, G Yancey Gillespie¹, Charles S. Springer²

¹University of Alabama at Birmingham, Birmingham, USA; ²Oregon Health & Science University, Portland, USA

Recent evidence suggests that contrast agent (CA) permeation into normal brain parenchyma is detectable with high-field MRI, which should increase detection sensitivity via tissue 1H2O T1 increases, concomitant with greater CA-induced T1 reductions.  During gradual, sustained GdDTPA2- infusions, we observed substantial image intensity increases in the normal mouse brain ventricular system, and in non-ventricular parenchyma, using ultra high-field MRI (8.5T).   These data provide clear and convincing evidence of standard monomeric Gd(III) chelate MRI CA permeation into normal brain parenchyma, consistent with improved  CA detection sensitivity at high-field,  potentially enabling implementation of powerful new CA approaches in normal and diseased brain.

Riikka Johanna Immonen¹, Juha Yrjänheikki², Taneli Heikkinen², Leena Tähtivaara², J Puoliväli², R I. Grundy³, T Tuinstra⁴, A Phinney⁴, B Van Vliet⁴, Olli Gröhn¹

¹A.I.Virtanen Institute for Molecular Sciences, University of Kuopio, Kuopio, Finland; ²Cerebricon Ltd., Kuopio, Finland; ³Cerebricon Ltd., London, UK; ⁴Solvay Pharmaceuticals BV, CP Weesp, Netherlands

Hypoperfusion may have a role in the secondary injury cascade following traumatic brain injury. We measured relative cerebral blood volume (CBV) changes and diffusion in rat brain following cortical impact injury aiming to find out more about the hemodynamics in the perifocal area surrounding the lesion. The Δ R2 and Δ;R2* maps using iron oxide contrast agent and average diffusion maps were acquired at 1h, 2h, 4h, 1d, 2d, 3d, 4d, 7d and 14d after TBI. Both Δ R2 and Δ;R2* showed acute drop in CBV in the perifocal area, which then started to recover. Yet even after 14d, ΔR2 remained decreased.

Kim C.C. van de Ven¹, Rick M. Dijkhuizen¹, Ivo A.C.W. Tiebosch¹, Annette Van der Toorn¹

¹University Medical Center Utrecht, Utrecht, Netherlands

Velocity Selective Arterial Spin Labeling (VSASL) is an ASL-based perfusion MRI technique that magnetically labels blood spins based ontheir velocity instead of their spatial position. Thus far VSASL perfusionimaging has not yet been assessed for application in an animal model. We measured VSASL subtraction signal in in vivo rat brain at 9.4T, which diminished to noise level after asphyxiation. Our study demonstrates that VSASL can be successfully applied in rat brain studies.

Rui Liu¹, Ping Hou², Kurt H. Bockhorst², Renjie He², Ponnada A. Narayana²

¹University of Houston, Houston, Texas, USA; ²University of Texas Medical School at Houston, Houston, Texas, USA

Synopsis: A method for automatic determination of the perfusion parameters based on dynamic susceptibility contrast MRI is proposed and applied to determine the hemodynamic parameters. Our algorithm is based on the singular valued decomposition (SVD) technique in which the thresholds are automatically determined. In addition, our algorithm automatically determines the arterial input function by searching the whole image volume. Application of this method to rodents yielded a value of the cerebral blood volume that is very close to the value determined using the microsphere technique, considered to be the “gold standard”.

Animal Models & Cerebral Ischemia

Susanne Wegener¹, Tommy Shute¹, Eric C. Wong¹

¹University of California San Diego, La Jolla, USA

One of the unresolved problems in stroke therapy is the heterogeneity of the disease and of the individual resistance to ischemic damage. We used longitudinal MRI to study ischemia tolerance, induced in rats by preconditioning with 3-nitroproprionic acid (3NPA). When preconditioned rats were subjected to transient ischemia, the immediate deficit on apparent diffusion coefficient maps was dramatically reduced within 30min of occlusion, together with better maintained residual blood flow (CBF) in the ischemic area compared to controls. Besides, CBF was lower on the unaffected side in NPA animals. MRI might have the potential to detect a “signature” of ischemia tolerance.

Qiang Shen¹, Meghana Bhatta¹, Ryan Jessee¹, Timothy Q. Duong¹

¹Emory Unviersity, Atlanta, Georgia, USA

Ischemic stroke occurs when cerebral blood flow (CBF) falls below a critical threshold, resulting in energy failure which subsequently manifests into a reduction in the water apparent diffusion coefficient (ADC). However, how the ADC is affected by CBF over time and under different ischemic conditions remains poorly understood. In this study, ADC and CBF in rat models following permanent and transient (30-min, 60-min and 90-min) MCAO were systematically studied at multiple time points after stroke. These results provide novel insight into the relationship between ADC and CBF under different ischemic durations and time after stroke.

Qiang Shen¹, Hanh Nguyen¹, Timothy Q. Duong¹

¹Emory Unviersity, Atlanta, Georgia, USA

We investigated postischemic hyperperfusion in rats subjected to three different occlusion durations. Diffusion, perfusion and T2 imaging during acute and sub-acute phase were acquired. Hyperperfusion was observed predominantly in the core after 24 hrs post-occlusion in the 30-min MCAO group, some in the 60-min group, and none in the 90-min group. No hyperperfusion was observed in normal tissues. Tracking tissue fates indicates that subsequently salvaged tissue has normal CBF (no hyperperfusion) whereas tissue destined to infarct showed hyperperfusion. These results suggest that chronic phase hyperperfusion is due to changes in blood-brain permeability which leads to increased CBF in infarct regions.

Yoji Tanaka¹, Tsukasa Nagaoka¹, Govind Nair¹, Timothy Q. Duong¹

¹Emory University, Atlanta, Georgia, USA

Stroke disrupts blood brain barrier, resulting in changing vascular permeability which is expected to affect cerebral blood flow (CBF) quantification. However, permeability changes are generally not taken into account in perfusion imaging of stroke. The aim of this study is to examine how changes in permeability affect CBF measured by arterial-spin-labeling and dynamic-susceptibility-contrast MRI on the same animals under different experimental conditions which included stroke, hypercapnia and permeability disruption by mannitol. We concluded that permeability changes markedly affects the CBF, it affects the two methods very differently, and CBF MRI of stroke needs to take permeability changes into account.

Chris McCabe¹, William Matthew Holmes¹, Lindsay Gallagher¹, Willy Gsell², Anna F. Dominiczak¹, Barrie Condon¹, I Mhairi Macrae¹

¹University of Glasgow, Glasgow, UK; ²Imperial College London, London, UK

In the present study we investigated the spatiotemporal evolution of stroke in the SHRSP rat and compared this to its normotensive control the WKY rat.  We have demonstrated that the SHRSP rat has significantly less penumbral tissue than the WKY within 1 hour of stroke onset and by 6 hours the SHRSP has no diffusion/perfusion mismatch.  These results could have important implications for the management of stroke patients with pre-existing hypertension and suggest  ischaemic damage could  progress at a faster rate in the presence of known risk factors such as hypertension.

Ona Wu¹, Jie Lu¹, Yasu Egi¹, Guangping Dai¹, Yoshi Murata¹, Joseph B. Mandeville¹, John J. Marota¹, A Gregory Sorensen¹, Rick M. Dijkhuizen², Kenneth K. Kwong¹, Eng H. Lo¹, Aneesh Singhal³

¹Massachusetts General Hospital, Charlestown, Massachusetts, USA; ²University Medical Center Utrecht, Utrecht, Netherlands; ³Massachusetts General Hospital, Boston, Massachusetts, USA

In rodent and pilot human acute stroke studies, normobaric oxygen therapy (NBO, or inhaled 100% oxygen) transiently improved neurological deficits and diffusion-weighted imaging abnormalities. We used serial MRI in rodent stroke models to investigate the hemodynamic effects of NBO in different brain regions. NBO induced an early, reproducible change in cerebral blood volume Δ CBV) in different brain regions, with ‘salvageable’ tissue showing increased CBV. This effect diminished over 3 hours. Whether this effect indicates a mechanism (e.g a hemodynamic “reverse steal phenomenon”) or a consequence of ‘salvaged’ tissue requires further investigation.

Shuning Huang¹, Dmitriy Atochin², Guangping Dai², Paul Huang², Bruce R. Rosen², Young Ro Kim²

¹MIT/MGH, Cambridge, Massachusetts, USA; ²MGH, Charlestown, Massachusetts, USA

Restoration of cerebral blood supply at early stage of stroke is critical for salvaging brain tissues at risk. We characterized vascular transformation during the reperfusion phase of transient ischemia using steady-state intravascular contrast agent techniques. Previous studies have shown that relative cerebral blood volume (rCBV) measured from Δ R2 and Δ R2*, the transverse relaxation rate change before and after the contrast agent administration, reflects total and micro-vascular CBV. The goal of this study was to understand cerebrovascular changes (rCBV and vessel size) in transient stroke mouse models using an intravascular contrast agent (superparamagnetic iron oxide nanoparticles) at 9.4T.

Ivan Vorisek¹, ², Norbert Zoremba³, Ales Homola¹, ², Karel Slais¹, ², Eva Sykova¹, ²

¹Institute of Experimental Medicine ASCR, Prague, Czech Republic; ²Charles University, 2nd Medical Faculty, Prague, Czech Republic; ³University Hospital RWTH Aachen, Aachen, Germany

The aim of the present study was to quantify the changes in ECS diffusion parameters during recovery from transient ischemia by TMA⁺-diffusion and MRI measurements and to describe their time course. The data were correlated with the results of DC-potential recordings and measurements of extracellular potassium levels. To the best of our knowledge, the diffusion parameters of the ECS during recovery from transient ischemia have not yet been studied in vivo.

Hongxia Lei¹, ², Carole Berthet³, Lorenz Hirt³, Rolf Gruetter¹, ⁴

¹Ecole Polytechnique Federale de Lausanne, Lausanne, Switzerland; ²University of Lausanne, Lausanne, Switzerland; ³Centre Hospitalier Universitaire Vaudois, Lausanne, Switzerland; ⁴University of Geneva, Lausanne, Switzerland

High magnetic fields increase sensitivities and thus using a recently installed 14.1 T/26cm MR system, we sought to determine the feasibility of studying lesion developing and neurochemical changes following 30 min of ischemia in mouse brain. At 8 and 24h after the insult, T2-weighted images presented lesion developing and localized spectra were obtained with 0.035ppm linewidths in the stroke region after B0 homogenneities adjustment with 2nd-order shimming. Prominent changes include a transient doubling of brain Gln, postulated to reflect Glu excitotoxicity, and decreases in several compounds such as NAA, Glu, Tau, as well as increases in acetate (attributed to NAA breakdown) and Lactate.

                  1942.     Radiation Pretreatment Result in a Dramatic Increase in ADC After Ischemic Brain Injury in Rats

Elena Titova¹, Arash Adami¹, Robert Ostrowski¹, Serafin Lalas¹, Roman Vlkolinsky¹, John H. Zhang¹, Gregory Nelson¹, Andre Obenaus¹

¹Loma Linda University School of Medicine, Loma Linda, California , USA

Radiation is routinely used in general medical practice significantly increases risk of stroke among patients but the response of the brain after such kind of injury is still insufficiently studied. We report that a single radiation (8Gy) exposure at 10 days prior to middle cerebral artery occlusion appears to reduce cytotoxic edema but dramatically increased ADC values at later time points. Thus, stroke outcomes in previously irradiated patients differ from the normal population and relatively low doses of radiation result in less brain edema in acute stages.

Ursula I. Tuor¹, ², Min Qiao¹, Sanju Lama², Salma Shivji², Kumud Deka², Tadek Foniok¹

¹National Research Council - Institute for Biodiagnostics (West), Calgary, Canada; ²University of Calgary, Calgary, Canada

We hypothesized that mild cerebral hypoxia-ischemia producing selective white matter injury would have accompanying markers of injury in white matter reflecting mechanisms of damage.  T2 MRI was used in our mild model of cerebral hypoxia-ischemia to screen animals with relatively selective white matter injury.  In such animals, substantial white matter injury appeared irreversible as demonstrated by an increased labelling of cells with TUNEL and a reduced myelination detected with O4 and MBP.  Inflammatory responses associated with microglia activation (ED1) were observed selectively in the whiter matter at 48hrs post a mild hypoxia ischemia.

Sanju Lama¹, ², Min Qiao², Tadek Foniok², Ursula I. Tuor, ¹²

¹University of Calgary, Calgary, Canada; ²National Research Council Institute for Biodiagnostics (West), Calgary, Canada

In adults, stroke injury is affected by gender.  We hypothesized that neonates  would also exhibit gender-differences in the extent, distribution or degree of vasogenic edema detected with MRI following cerebral hypoxia-ischemia.  Seven day old rats were subjected to cerebal hypoxia-ischemia and T2 imaging to determine infarct size and edema.  Three days post insult, the ischemic lesion volume was similar in male and female pups, however, T2 values in ischemic parietal cortex were less in female than male rat pups.  The greater edema in males supports a sexual dimorphism regarding the mechanisms of hypoxic-ischemic injury in immature brain.

Chrystelle Po¹, Sebastien Fau², Christiane Charriaut-Marlangue², Philippe Meric¹, Brigitte Gillet¹

¹ICSN/CNRS, Gif-sur-Yvette, France; ²UMR-CNRS 7102, Paris, France

The evolution of the injured tissues after a transient focal cerebral ischemia in neonate rats was investigated by serial measurements of T2 maps and DTI. Since day 7 after ischemia up to day 21, a significant low T2 value area on T2 maps and significant perturbations of tissue organization evidenced on maps of anisotropy fraction and of first eigenvector of DTI appeared in the previously ischemic area. At day 21, this area correlated with a significant increase of cell density on corresponding cresyl-violet stained brain slices suggesting that these data reflect tissue healing processes such as gliosis.

Kimmo Jokivarsi¹, Juha-Pekka Niskanen¹, ², Heidi Gröhn³, Shalom Michaeli⁴, Michael Garwood⁴, Risto A. Kauppinen⁵, Olli H. Gröhn¹

¹A.I.Virtanen Institute for Molecular Sciences, Kuopio, Finland; ²University of Kuopio, Kuopio, Finland; ³Kuopio University Hospital, Kuopio, Finland; ⁴University of Minnesota, Minneapolis, Minnesota, USA; ⁵Dartmouth Medical School, Hanover, New Hampshire, USA

Rotating frame relaxation obtained with adiabatic pulses allows for the contribution of dipolar interactions and exchange effects to be modulated by changing amplitude and phase modulation of the RF pulses. Using a known two-site exchange model the inherent correlation times, population sizes and exchange correlation times can be computed.  We investigated the water dynamics during evolving cerebral ischemia in a rat permanent stroke model. Our results show that both adiabatic T1 ρ and T2 ρ MRI reveal cerebral ischemia early on. Changes in fitted exchange parameters are indicative of increase in free water content and lytic damage to tissue.

Tzy-Haw Wu¹, ², Hsiao-Fang Liang², Chin-I Chen, ²³, Chung-Yi Hsu⁴, Sheng-Kwei Song²

¹National Taiwan University Hospital, Taipei, Taiwan; ²Washington University School of Medicine, St. Louis, Missouri, USA; ³Taipei Medical University - Wanfang Municipal Hospital, Taipei, Taiwan; ⁴Graduate Institute of Neuroscience, Taipei Medical University, Taipei, Taiwan

SynopsisWe generated a mouse stroke model to determine if DTI can differentiate the different severities of ischemic stoke in the early stage. In this study, the areas of infarcted cortex and external capsule (EC) were evaluated by DTI. The result showed that axial diffusivity and radial diffusivity in the area of infarcted EC are statistically different among the three ischemic stroke groups of different severities. The infarcted cortex volume increases with the increased severity of stroke. ADC values in the infarcted cortex are not as sensitive as axial diffusivity or radial diffusivity in the area of infarcted EC. The DTI predicted axonal injury in EC was also validated by immunohistochemistry. In conclusion, DTI provides a good quantitative measure to evaluate the severity of early stage ischemic stroke in mice, especially the measurement in the area of infarcted EC.

Alzheimer's Disease & Mild Cognitive Impairment

Qin Chen¹, Ling Zou, Qiang Yuan¹, Zheng-Yan Li, Luo Ou-Yang, Wei-Wei Zhang, Li-Jun Jiang, Dong Zhou¹, Qi-Yong Gong

¹West China Hospital of Sichuan University, Chengdu, People's Republic of China

Despite many efforts to draw a clear cut picture of the disease, a comprehensive characterization of grey and whiter matter changes in mild Alzheimer's disease (AD) is still not available. In this study, we aimed to characterize the brain changes in mild AD by the combined used of VBM and DTI. Our results suggest atrophy of cortical and subcortical structures and nerurodegeneration of specific fibre tracts may contribute to neurological deficits in AD. This approach guide future research investigating the relation between the brain areas involved and the clinical features in different phases of the disease.

Julio Acosta-Cabronero¹, Guy B. Williams¹, Peter J. Nestor¹

¹University of Cambridge, Cambridge, UK

The corpus callosum (CC), which contains millions of inter-hemispheric axons, is known to become atrophic early in AD, and its integrity is assumed to be associated with global cognitive performance. In this study, we extracted several DTI measures at the midline CC, and compared their relationships with CC area, as a marker of disease-related atrophy, and with global cognitive data. Radial diffusivity exhibited the strongest correlation to brain atrophy, whereas global cognition was better predicted by mean diffusivity measures. We found that high diffusivity integrals were systematically better predictors of both atrophy and global cognition than mean ROI values.

Darryl H. Hwang¹, Witaya Sungkarat¹, Manbir Singh¹

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

Study was conducted to detect common regions in the brain where FA was reduced both in MCI and AD when compared to normals, but reduced more in AD than MCI. Pathways in a normalized space were generated from these regions by sorting normalized DTI tractography. Resulting pathways are consistent with the biology of Alzheimer Disease.

Yu Zhang¹, ², Norbert Schuff¹, ², An-Tao Du¹, Howard J. Rosen², Joel H. Kramer², Maria Luisa Gorno-Tempini², Bruce L. Miller², Michael W. Weiner¹, ²

¹Center for Imaging of Neurodegenerative Diseases, VA Medical Center, San Francisco, California , USA; ²University of California, San Francisco, California , USA

Alzheimer’s disease (AD) and frontotemporal dementia (FTD) are two common dementias but difficult to be differentiated. In this study, diffusion tensor imaging (DTI) was used with tractography-based analysis and global analysis in 13 AD patients, 12 FTD, and 13 control (CN) subjects. DTI analyses showed that FTD had significant lower fractional anisotropy (FA) than CN predominantly in the anterior (frontal) and temporal brain; AD had significant lower (FA) than CN predominantly in the posterior (parietal) and temporal brain. These distinct FA patterns of FTD and AD may aid the differential diagnosis between the two types of dementias.

Donna J. Cross¹, Yoshimi Anzai¹, Jeffery Stevenson¹, Kenneth R. Maravilla¹, Elaine R. Peskind, Satoshi Minoshima¹

¹University of Washington, Seattle, Washington, USA

Wen-Tung Wang¹, Sang-Pil Lee¹, Mary L. Michaelis², In-Young Choi¹

¹University of Kansas Medical Center, Kansas City, Kansas, USA; ²University of Kansas, Lawrence, Kansas, USA

The levels of neurochemicals can be altered through various pathological mechanisms, thus providing insights into the disease progression. Neurochemical profiles were acquired from the hippocampus of triple-transgenic mice with mutations in APP, PS1, and tau using ultra-short echo time ¹H MRS at 9.4 T. Our preliminary study showed significant decreases of glutathione, a critical antioxidant, in the 19 months old transgenic mouse brain in vivo, indicating increased oxidative stress in AD.

Malgorzata Siger¹, ², Norbert Schuff¹, Xiaoping Zhu¹, Michael Weiner¹

¹Veterans Administration Medical Center, San Francisco, California , USA; ²Medical University of Lodz, Lodz, Poland

In this study, we used spectroscopic imaging (MRSI)  to determine whether both MCI and AD show systematic regional patterns of metabolite abnormalities in white matter and gray matter. Results of our study suggests vulnerability of white matter in the pathology of AD and MCI as indicated by widespread increased myo-inositol in white matter regions. Furthermore, increased myo-inositol may be an even more robust and sensitive marker for MCI and AD than NAA, which was not significantly reduced in white matter compared to aging

Jacob Penner¹, ², Raul Rupsingh¹, ², Matthew Smith³, Jennie Wells, ²³, Michael Borrie, ²³, Rob Bartha¹, ²

¹Robarts Research Institute, London, Canada; ²University of Western Ontario, London, Canada; ³Lawson Health Research Institute, London, Canada

Short echo time LASER localized proton magnetic resonance spectroscopy at 4.0 Tesla was used to quantify the levels of glutamate (Glu), N-acetyl aspartate (NAA), and creatine (Cr) in ten newly diagnosed Alzheimer disease patients at baseline and after 4 months of Galantamine treatment. A significant increase was found in the Glu/NAA ratio (p < 0.05) after 4 months, while there were no significant changes in the ratio of Glu/Cr, the absolute Glu, NAA, or Cr concentrations, or cognitive scores.

Andreana P. Haley¹, Kathleen L. Fuchs², Sarah Andrea Dunham³, Carol A. Manning², Jack Knight-Scott³

¹The University of Texas at Austin, Austin, USA; ²University of Virginia, Charlottesville, Virginia , USA; ³Children's Healthcare of Atlanta, Atlanta, Georgia, USA

In this work, we demonstrate that differences in regional metabolic patterns can improve the differentiation between amnestic and non-amnestic mild cognitive impairment. Proton MRS measurements in the anterior and posterior cingulate gyrus provide unique metabolic profiles for both diagnoses indicating these sub-types are also unique neuropathologies.

Brenda Lynn Bartnik Olson¹, Matthew Wagner¹, William Britt¹, Wolff Kirsch¹, Barbara A. Holshouser¹

¹Loma Linda University, Loma Linda, California , USA

This study compares global brain, white and gray matter volumes as well as cerebral metabolite levels and ratios in tracking cognitive decline in twenty MCI and 13 cognitively normal control subjects. Structural 3D T1-weighted MRI and single voxel 1H MRS of the posterior cingulated gyrus (PCG) studies were performed over a 2-year period with repeat studies approximately every 12 months and neurocognitive testing every 6 months. Results show that the conversion to AD is associated with an increased rate of GM loss and decreases in NAA and Glx levels, compared to cognitively normal aging adults. Baseline NAA and Glx levels had a higher predictive accuracy than tissue or CSF volume.

                  2018.     Correlation Between Global Severity Scales in Cognitive Impairment (GDS and CDR) and Magnetic
         Resonance H1 Spectroscopy, Perfusion Weighed Imaging and Diffusion Weighted Imaging

Jorge Humberto Davila Acosta¹, Nicolas Nicolas Fayed¹, Antonio Oliveros Cid¹

¹Hospital Quiron, Zaragoza, Spain

The clinical evaluation, the diagnosis and the following of patients with AD and MCI is very complex. The use of clinical tools as GDS and CDR help to classify these patients, to compare them and to evaluate the response to the treatment. Some studies have found a good correlation between these clinical scales and the pathologic stages in AD described by Braak. The use of MRS, PWI and DWI in patients with AD and MCI has showed structural and metabolic changes. These changes can help in the clinical evaluation of patients with AD and MCI if a good correlation with the clinical stages is demonstrated.

                  2019.     Association Between the Apolipoprotein E ε4 Gene Polymorphism and Cerebral Ventricular Dilatation
        Measured from Serial Magnetic Resonance Imaging in Subjects Enrolled in the Alzheimer’s Disease
        Neuroimaging Initiative  [Not Available]

Sean M. Nestor¹, Raul Rupsingh¹, Vittorio Accomazzi, Michael J. Borrie², Matthew Smith³, Jennie Wells², Jennifer Fogarty², Robert Bartha¹

¹Robarts Research Institute, London, Canada; ²University of Western Ontario, London, Canada; ³Lawson Health Research Institute, London, Canada

The current study compares ventricular dilatation between carriers for the apolipoprotien E ε4 allele (ε4+) and subjects with other polymorphisms (ε4-), in subjects with Alzheimer disease(AD) and mild cognitive impairment (MCI), over a short interval. Six-month longitudinal data were selected from the Alzheimer’s Disease Neuroimaging database, which included 502 subjects (elderly controls, MCI, and AD). Baseline and six-month unprocessed 3D T1-weighted MP-RAGE 1.5 Tesla MR images, neurocognitive, and genetic measures were acquired. Ventricular volume was computed using a semi-automated region-growing algorithm (Cedara Software). Ventricular dilatation was significantly greater in the AD ε4+ group compared to AD ε4- subjects (p< 0.05).

                  2020.     Shifting Regional Atrophy Rates in the Progression from Normal Aging to Alzheimer’s Disease Quantified
        by Fluid Registration

Jasper D. Sluimer¹, Wiesje M. van der Flier¹, Giorgos B. Karas¹, Ronald A. van Schijndel¹, Josephine Barnes², Richard G. Boyes², Keith S. Cover¹, Silvia D. Olabarriaga³, Nick C. Fox², Philip Scheltens¹, Hugo Vrenken¹, Frederik Barkhof¹

¹VU University Medical Center, Amsterdam, Netherlands; ²University College London, London, UK; ³Academic Medical Center, University of Amsterdam, Amsterdam, Netherlands

In 64 patients with Alzheimer’s disease (AD), 44 patients with mild cognitive impairment (MCI) and 34 controls, we determined regional atrophy rates in six different brain regions by applying Fluid nonlinear registration software to 3D coronal MPrage images from two timepoints. In MCI, the temporal lobe shows the greatest atrophy rate. In AD patients, the medial temporal lobe shows an atrophy rate comparable to MCI, while the remaining part of the temporal lobe demonstrates an even higher rate of atrophy. Moreover, atrophy also accelerates in parietal and occipital lobes.

David L. Henderson¹, John Frederic Schenck¹, Natalie A. Staples¹, Earl A. Zimmerman²

¹General Electric Global Research, Niskayuna, USA; ²Albany Medical College, Albany, USA

Quantitative T2-mapping holds great promise for increasing the utility of MRI in the management of neurodegenerative diseases (NDDs), such as Alzheimer's disease, which affect millions of patients. Although many studies have the potential of quantitative brain imaging, they usually require prohibitive levels of time-consuming postprocessing analysis to provide useful results. We show examples of semi-automated image analysis capable of quantifying regional brain atrophy in various NDDs that may greatly reduce this bottleneck and promote the widespread use of MRI in these conditions.

                  2022.     Sodium MRI Enhancement of the Medial Temporal Lobe in Mild Alzheimer’s Disease

Eric Albert Mellon¹, David T. Pilkinton¹, Reddy Shashank Beesam¹, Christopher M. Clark¹, Elias R. Melhem¹, Walter R. Witschey¹, Arijitt Borthakur¹, Ravinder Reddy¹

¹University of Pennsylvania, Philadelphia, Pennsylvania, USA

There is substantial interest in quantitative techniques for the study of Alzheimer’s Disease (AD) to track the progression of disease for both prognosis and to monitor new therapies for AD. Towards this, presented is a study of changes in the medial temporal lobes by sodium MRI within the brains of patients with AD versus non-demented elderly controls on a clinical 3T scanner. An optimized Gradient Recalled Echo readout shows a statistically significant 12% enhancement between 4 mild AD patients and 5 age-matched control subjects. Work is already underway to elucidate the mechanisms and specificity of this enhancement.

Robert John Dawe¹, David A. Bennett², Julie A. Schneider², Sunil Vasireddi¹, Konstantinos Arfanakis¹

¹Illinois Institute of Technology, Chicago, Illinois, USA; ²Rush University Medical Center, Chicago, Illinois, USA

One advantage of postmortem MRI over in vivo imaging is that the in vitro tissue can be held nearly motionless indefinitely, making it possible to perform very high resolution imaging. This opportunity is exploited in the current study. Thirty-eight cadaveric human brains were scanned, and the T2 values and volumes of the hippocampi were correlated with the subjects' MMSE scores and which hemisphere was imaged (right or left). It was found that in postmortem brain specimens, low MMSE scores are associated with lower hippocampal T2 values.

Sarah Pachtman¹, Himachandra Chebrolu¹, Charles Dennis Smith¹, Peter Andrew Hardy¹

¹University of Kentucky, Lexington, USA

                  2025.     Quantitative Comparison of Cerebral Blood Volume Between Patients with Alzheimer's Disease and
                                 Elderly Controls

Jinsoo Uh¹, Kelly Lewis-Amezcua¹, Kristin Martin-Cook¹, Myron Weiner¹, Ramon Diaz-Arrastia¹, Hanzhang Lu¹

¹University of Texas Southwestern Medical Center, Dallas, USA

Recent study suggests that Alzheimer’s Disease (AD) is featured by vascular factors, which may be useful for better understanding and early marker of AD. Utilizing a novel MR technique, Vascular-Space-Occupancy (VASO)-MRI, we quantitatively assessed Cerebral Blood Volume (CBV) of the entire brain on mild AD patients compared with age-matched controls. In particular, we identified brain regions that show significant differences between the two groups using ROI and voxel-by-voxel analyses based on elastic co-registration. Our study revealed that mild AD patients manifest significant reduction of CBV by 10-20% particularly at limbic system and basal ganglia regions.

delia Lenzi¹, ², Laura Serra³, patrizia Pantano¹, eraldu Paulesu, franco Giubilei¹, roberta Perri³, gian luigi Lenzi¹, carlo Caltagirone³, emiliano Macaluso³, marco Bozzali³

¹Universita"La Sapienza", Rome, Italy; ²Fondazione Santa Lucia, IRCCS , Rome, Italy; ³Fondazione Santa Lucia, IRCCS, Rome, Italy

Amnesic Mild Cognitive Impairment (a-MCI) is considered as a frequent prodromal state of Alzheimer’s disease. Aim of this study was to investigate, using fMRI, the patterns of activation in patients with a-MCI when performing multiple tasks that selectively engage specific cognitive domains. 15 patients with MCI and 10 sex- and age-matched group of healthy controls were studied  during tasks assessing  memory functions, spatial attention, and empathic ability. During all tasks there was a significantly increased activation in the same network observed in  healthy controls. This might reflect an initial compensation that explains the maintenance of performance in patients with MCI.

John A. Ronald¹, ², Yuanxin Chen¹, Lisa M. Bernas¹, ², Robert A. Hegele¹, ², Kem A. Rogers², Brian K. Rutt¹, ²

¹Robarts Research Institute, London, Canada; ²University of Western Ontario, London, Canada

We demonstrate a novel model of Alzheimer’s disease by feeding rabbits a low-level cholesterol diet for extended periods of time.  These animals develop extracellular beta-amyloid rich plaques similar to compact plaques found in humans, and these plaques also reliably accumulate iron.  High-resolution MRI of excised brains revealed signal voids throughout the brain parenchyma that directly correlated to iron-laden plaques in matched tissue sections.  Minimal voids were seen in control brains.  This is both the first successful attempt at direct imaging of plaques formed in a large animal model and first successful clinical field-strength imaging of plaques in any model. 

                  2028.     Improved Magnetic Resonance Microimaging of Individual Amyloid Plaques in Alzheimer's Transgenic
                                 Mice

Ryan Chamberlain¹, Denise Reyes², Geoffry L. Curran², Thomas M. Wengenack², Joseph F. Poduslo², Michael Garwood¹, Clifford R. Jack²

¹University of Minnesota, Minneapolis, Minnesota, USA; ²Mayo Clinic College of Medicine, Rochester, Minnesota, USA

One of the principal pathologies of Alzheimer’s disease is amyloid plaques, the reduction of which has been identified as a major therapeutic objective.  Plaques appear hypointense relative to background tissue on both T₂- and T₂^*-weighted images.  The major limitations of existing approaches are long scan time and low CNR.  This investigation analyzes the ability to increase the CNR over a previously verified spin echo sequence by summing multiple spin echoes and using multiple spin echo susceptibility weighted imaging.

Palamadai Nilakantan Venkatasubramanian¹, ², Gheorghe Iordanescu¹, Alice M. Wyrwicz¹

¹ENH Research Institute, Evanston, Illinois, USA; ²Northwestern University Feinberg School of Medicine, Chicago, Illinois, USA

Amyloid plaques in two APP transgenic mouse lines were imaged directly without the use of external contrast agent.  Plaque characteristics such as size, number and regional distribution were different between the two transgenic lines.

Yuanxin Chen¹, John A. Ronald¹, ², Lisa M. Bernas¹, ², Robert A. Hegele¹, ², Kem A. Rogers², Brian K. Rutt¹, ²

¹Robarts Research Institute, London, Canada; ²University of Western Ontario, London, Canada

Our lab has shown that direct visualization of Alzheimer’s disease (AD) plaques in a cholesterol-fed rabbit model of AD is possible using clinical-field strength ex vivo MRI.  These rabbits formed plaques with dramatic iron accumulation. Here, we performed MRI in an additional two AD models including rabbits fed a high-cholesterol diet for a short time and an AD transgenic mouse. We demonstrate that amyloid plaques formed in these models without iron deposits were not detectable on MR imaging. This suggests that the ability to detect AD plaques in animal models is due to their clear association with excess iron.

Samir Kamalesh Amin¹, Cynthia Massaad¹, Lingyun Hu¹, Eric Klann², Robia G. Pautler¹

¹Baylor College of Medicine, Houston, Texas, USA; ²New York University, New York City, New York, USA

A-beta peptide accumulation is pivotal in Alzheimer’s disease (AD) pathogenesis. Few studies have addressed A β involvement in oxidative impairment of cerebrovascular regulation in AD. We studied this aspect of A β pathogenesis using transgenic mice with Flow-sensitive Alternating Inversion-Recovery EPI arterial spin labeling. We wanted to determine if overexpression of the reactive oxygen species scavenger SOD-2 in APP mice leads to reversal of cerebral blood flow attenuation. We found that oxidative damage on cerebrovasculature occurs early in the pathogenesis and can be “rescued” with SOD-2 overexpression. Our present findings support the idea that ROS formation precedes plaques in the AD pathogenesis.

Nicolau Beckmann¹, Stefan Zurbruegg¹, Catherine Cannet¹, Christelle Gérard¹, Dorothee Abramowski¹, Karl-Heinz Wiederhold¹, Matthias Staufenbiel¹

¹Novartis Institutes for BioMedical Research, Basel, Switzerland

MRI was used to detect in APP23 mice the effects of cerebral amyloid angiopathy. Endorem® was administered i.v. 24h before imaging. Foci of signal attenuations became apparent in the brain cortex as well as in thalamic regions of 16-month-old APP23 mice. An age-dependent increase in the number of foci displaying attenuated signal was observed in older animals. With a few exceptions, these foci were absent in age-matched, wildtype littermate controls. Histology revealed that, at sites of signal loss detected by MRI, iron was localized in or around damaged vessels, entrapped in microglia/macrophages.

Rose-Ann Blenman¹, Marie Holahan¹, Denise Welsh¹, Jacquelynn Cook¹, Richard Hargreaves¹, Donald Williams¹

¹Merck & Co. Inc., West Point, Pennsylvania, USA

Proton MRS biomarkers were assessed in young adult and geriatric Rhesus monkeys.  Since geriatric Rhesus monkeys (≥ 24 years) have been characterized as models of neurodegenerative disease, our hypothesis was that geriatrics would have altered biochemical profiles suggesting neuronal degeneration and glial activation.  MRS in the posterior cingulate of geriatric Rhesus showed that NAA/tCr decreased by ≈15%, mI/tCr increased by ≈30%, and tCho/tCr decreased by ≈20% as compared to the young adult monkeys (≤15 years).  Results also showed that NAA/tCr, mI/tCr and tCho/tCr are highly correlated with age and that tCho/tCr is correlated with NAA/tCr and NAA/mI.

Laurent Uldry¹, Isabelle Bourquin¹, Andrea Brioschi¹, Joseph Ghika¹, Philippe Maeder¹, Reto Meuli¹, Eleonora Fornari¹

¹Centre Hospitalier Universitaire Vaudois and University of Lausanne, Lausanne, Switzerland

During an fMRI study, we compared 15 newly diagnosed Alzheimer (AD) patients with 15 age-matched subjects and investigated higher-order visual areas to detect AD-related malfunction before the appearance of clinical signs. Our paradigm addresses visual interhemispheric integration functions, with stimuli obeying (IG) or not (OG) the Gestalt principle. We found that the interhemispheric integration process (IG vs. OG contrast) induced BOLD increase within VP/V4 areas. AD patients showed the same activation pattern as control subjects, but characterized by a reduction of intensity and extent, possibly due to an already decreased myelination of cortico-cortical connections involved in spatial integration.

Takashi Yoshiura¹, Tomoyuki Noguchi¹, Akio Hiwatashi¹, Osamu Togao¹, Koji Yamashita¹, Eiki Nagao¹, Hidetaka Arimura¹, Tomoyuki Okuaki², Ivan Zimine², Marc van Cauteren², Hiroshi Honda¹

¹Kyushu University, Fukuoka, Japan; ²Philips Medical Systems, Japan, Tokyo, Japan

Nine patients with MCI, seven patients with AD and seven normal subjects were studied using a quantitative ASL pulse sequence (QUASAR) and 3T MRI. Absolute CBFs were measured in whole gray matter region, bilateral thalami and bilateral inferior parietal (IP) cortices. Relative CBF values of bilateral IP cortices normalized by whole gray matter CBF and by thalamic CBF were also obtained. Both absolute and relative CBFs in the right IP cortex were significantly decreased in AD group. Moreover, relative CBF in right IP normalized by whole gray matter CBF was significantly decreased in MCI group in comparison with normal group.

Developmental Brain Disorders

Malek Makki¹, Michael Behen¹, Elizabeth Primeau¹, Benjamin Wilson¹, Arpi Bhatt¹, Harry Chugani¹

¹Wayne State University, Detroit, Michigan, USA

The neurobiological abnormalities underlying Tourette syndrome (TS) remain unknown despite various biochemical, neuroimaging, neurophysiological, and genetic studies that suggest a role of the basal ganglia and related thalamic and cortical regions. MRI studies of TS have reported volume reductions in the basal ganglia, frontal cortex, hippocampal gyrus, and white matter. We investigated interhemispheric brain asymmetry in children with TS using voxel-based analysis of diffusion tensor indices. Results show abnormal white matter integrity in regions comprising the fronto-striato-thalamic circuit in children with TS as compared to age-matched healthy controls, extending previous work that has shown this circuit as involved in TS.

Malek Makki¹, ², Mike Behen¹, Arpi Bhatt¹, Harry Chugani¹

¹Wayne State University, Detroit, USA

Despite a plethora of evidence suggesting that the fronto-striato-thalamic (FST) circuitry is abnormal in patients with Tourette Syndrome (TS), the structural integrity of white matter pathways has not yet been directly investigated. We used DT-MRI to test the hypothesis that TS may be associated with specific white matter abnormalities in the lentiform nuclei, thalamus, and caudate nucleus, and that individuals with TS will show decreased FA and increased ADC in fibers of FST as compared to healthy controls. We also examined whether abnormalities on DTI are associated with tic severity and obsessive-compulsive disorder and attention deficit hyperactivity disorder comorbidity.

Otto Muzik¹, Malek Makki², Darshan Pai², Anita Dias³, Jing Hua², Harry Chugani¹

¹Wayne State University, Detroit, Michigan, USA; ²Wayne State University, USA; ³CHM PET Center, Detroit, Michigan, USA

This paper describes a method for the quantitative assessment of fiber tract connectivity strength based on Bayesian probabilistic tractography. For each fiber path a normalized probability value is calculated which characterizes the probability of connection between two predefined cortical areas. We applied this method to DTI data derived from children with Tourette Syndrome (TS) and pediatric controls to evaluate the connectivity strength between subcortical and cortical regions. Our findings indicates an abnormal connectivity pattern between the head of caudate and frontal lobe regions in children with TS as compared to control children.

Yann Gagnon¹, Tim Devito¹, Janet Hendry¹, N Gelman¹, N Rajakumar², P Williamson², R Nicolson², Dick Drost¹

¹University of Western Ontario, Lawson Health Research Institute, London, Canada; ²University of Western Ontario, London, Canada

The transverse relaxation time (T2) is a quantitative parameter of magnetic resonance imaging which is indicative of the molecular environment in brain tissue.  In this study, T2 times of lobar white matter were evaluated in a group of patients with autism.  Images obtained at 3 Tesla using the GESFIDE technique were spatially normalized for statistical analysis.  Patients in this study had an increase in left-sided white matter T2 as well as an increase in frontal and parietal white matter T2.

Tyler Bridgeland Jones¹, Lauren Kenworthy, ¹², Laura K. Case¹, Shawn C. Milleville¹, Peter Anthony Bandettini¹, Alex Martin¹, Rasmus Matthias Birn¹

¹National Institutes of Health, Bethesda, Maryland, USA; ²Childrens National Medical Center, Washington, District Of Columbia, USA

Adolescents with autism spectrum disorders (ASD) show decreased functional connectivity relative to typically developing individuals, supporting a model of underconnectivity in autism.  Employing an overt fluency task to probe language and executive deficits typical of ASD, we aim to determine what factors contribute to the decreased correlations of spatially remote BOLD fMRI time series observed in ASD.  By sequentially regressing out various factors, including differential task activation, trial-to-trial task response variability, and certain scanner artifacts, we conclude that the disruption in functional connectivity in ASD is at least partially due to differences in task-unrelated, or “spontaneous,” neuronal fluctuations.

Jee Eun Lee¹, David Hsu¹, Andrew L. Alexander¹, Mariana Lazar², Erin D. Bigler³, Janet E. Lainhart³

¹University of Wisconsin, Madison, Wisconsin, USA; ²New York University School of Medicine, New York, New York, USA; ³University of Utah, Salt Lake City, Utah, USA

A novel probabilistic method for diffusion tensor tractography (W-matrix tractography) was developed to study underconnectivity in autistic children. This algorithm incorporates measures of intravoxel and intervoxel structure and allows simultaneous multi-directional branching of white matter tracts. In preliminary testing comparing 43 autistic subjects with 34 matched controls, we find a strong tendency to segregate the autistic from the control populations, particularly in the younger subjects. With further tuning, W-matrix tractography may be useful in the early diagnosis of autism.

Manzar Ashtari¹, Joel Bregman², Shana Nichols², Carolyn McIlree³, Linda Spritzer², Andrew Adesman⁴, Melissa Narain⁵, Babak Ardekani⁶

¹Children's Hospital of Philadelphia, Philadelphia, Pennsylvania, USA; ²Fay J. Lindner Autism Center, Bethpage, New York, USA; ³University of Vermont, Burlington, USA; ⁴Schneider Children's Hospital, New Hyde Park, New York, USA; ⁵North Shore LIJ Health System, Glen Oaks, New York, USA; ⁶Nathan Kline Institute of Psychiatry, Orangeburg, New York, USA

We have applied diffusion tensor imaging to a whole-brain voxelwise analysis and tractography in a group of high-functioning youth with Autism Spectrum Disorder (ASD) and demographically matched controls.  Results from tractography and voxelwise analysis showed increased FA in the posterior portion of the cingulate bundle (limbic system).  Increased FA positively correlated with the social-emotional reciprocity and autistic mannerism of the ASD patients. Our data suggest that alterations in the limbic system secondary to suboptimal connectivity may lead to core impairments of social interaction and behavior associated with the autism phenotype.

Song Lai¹, Fumiko Hoeft², Jianrong Shi¹, John Lackey¹, Udomchai Techavipoo¹, Adam Flanders¹, David Roeltgen¹, Allan L. Reiss², Judith Ross¹

¹Thomas Jefferson University, Philadelphia, Pennsylvania, USA; ²Stanford University, Stanford, California , USA

A fMRI study was conducted to investigate the functional brain differences between prepubertal boys with KS (8-12 years) and age-matched control boys, and to examine function associated with androgen replacement in KS, so as to improve our understanding of cognitive deficits associated with KS and androgen deficiency. Relative to the age-matched control group, boys with KS showed aberrant activation patterns in a regional specific manner, depending on the cognitive operation (language and motor functions). Aberrant activation patterns showed ‘normalization (i.e., increase in activation)’ with androgen treatment.

Song Lai¹, John Lackey¹, Jianrong Shi¹, Udomchai Techavipoo¹, David Roeltgen¹, Adam Flanders¹, Fumiko Hoeft², Allan L. Reiss², Judith Ross¹

¹Thomas Jefferson University, Philadelphia, Pennsylvania, USA; ²Stanford University, Stanford, California , USA

A voxel-based morphometry study was conducted to investigate the structural brain differences between prepubertal boys with KS (8-12 years) and age-matched control boys in order to obtain insight of the underlying neuroanatomy of the KS cognitive phenotype. Boys with KS were found to have smaller brain volumes in regions related to cognitive functions of interest, including the left insula, ventrolateral prefrontal and temporal lobes, [language function], the bilateral anterior limb of the internal capsule (ALIC), globus pallidus, pre and post-central gyri and corpus callosum [for bimanual motor function], and the dorsolateral prefrontal and parietal lobes [working memory function.

Zhaoying Han¹, ², Lynn Fuchs¹, Nikki Davis¹, Christopher J. Cannistraci², Adam W. Anderson², John C. Gore², Benoit M. Dawant¹

¹Vanderbilt University, Nashville, Tennessee, USA; ²Vanderbilt University Institute of Imaging Science, Nashville, Tennessee, USA

Anatomical differences between normal children (NC) and children with math difficulties (MD) are investigated with deformation based morphometry in a population comprised of 20 pairs of age and gender matched NC and MD children. High resolution MR images have been acquired and one population average has been computed using a non-rigid registration technique that produces dense deformation fields. Statistically significant differences in the deformation fields between the two populations were found in several areas that have been previously reported in functional studies as related to computation. These findings may explain the differences that have been observed in functional studies.

Vincent Jerome Schmithorst¹, Scott Kerry Holland¹

¹Children's Hospital Medical Center, Cincinnati, Ohio, USA

Children with severe-to-profound unilateral sensorineural hearing loss (USNHL) show deficits in higher-order auditory processing tasks.  We investigate the task of receptive speech using functional MRI (fMRI) in children with right and left USNHL.  The paradigm consisted of a “modified token” task.  Subjects viewed an arrow moving to point to two shapes shown on a video screen.  Subjects would respond if the speaker correctly described the motion of the arrow.  Using group Independent Component Analysis (ICA), activation was shown in auditory processing and higher-order integrative regions.  Results also indicate the preferential formation of auditory processing pathways ipsilateral to the hearing ear.

Vincent Jerome Schmithorst¹, Scott Kerry Holland¹

¹Children's Hospital Medical Center, Cincinnati, Ohio, USA

Children with right unilateral sensorineural hearing loss (USNHL) show greater deficits in academic performance as compared to children with left USNHL.  We investigate possible differences in white matter microstructure using diffusion tensor MRI (DTI).  Children with right USNHL displayed greater fractional anisotropy (FA) in the genu and splenium of the corpus callosum, in frontal regions bilaterally, and in the left occipital lobe.  Children with left USNHL displayed greater FA in temporo-parietal white matter in the left hemisphere.  Results indicate preferential formation of inter-hemispheric pathways in children with right USNHL, and intra-hemispheric pathways in the left hemisphere in children with USNHL.

Vincent Jerome Schmithorst¹, Scott Kerry Holland¹

¹Children's Hospital Medical Center, Cincinnati, Ohio, USA

We used functional MRI (fMRI) in conjunction with an audiological test performed in-scanner to investigate the source of auditory processing deficits in children with unilateral sensorineural hearing loss (USNHL).  The task involves interpretation of speech-in-noise at varying levels of SNR.  Using group Independent Component Analysis, networks were found related to task difficulty.  Two attentionally-related components were found with activation in the frontal lobes.  Two auditory-related components were also found, with activation in Wernicke’s area and its RH homolog, and activation in the left IFG (BA 45/47).  Differences were also found between children with left and right USNHL.

Zhifeng Kou¹, Meng Li¹, Quan Jiang, Navid Seraji, Yang Xuan¹, E Mark Haacke¹, Harry T. Chugani, Csaba Juhasz, Jiani Hu¹

¹Wayne State University School of Medicine, Detroit, Michigan, USA

This study was to determine whether MRSI can improve detection of frontal lobe involvement in children with SWS. Among 16 children (age: 0.9 -10.4 years) with unilateral SWS, 8 children presented normal-appearing frontal lobes on conventional MRI, but 7 of them showed abnormal NAA and/or choline content in the frontal lobe of the affected hemisphere. Lower frontal lobe gray matter NAA was associated with earlier onset of seizures (r = 0.76; p = 0.04) and was an excellent predictor of motor function (r=-0.89, p<0.001). MRSI is more sensitive than conventional structural MRI for detection of frontal lobe involvement in SWS.

Animal Models: Demyelination & Regeneration in Brain & Spine

Amnon Bar-Shir¹, Ian D. Duncan², Yoram Cohen¹

¹Tel-Aviv University, Tel-Aviv, Israel; ²University of Wisconsin, Madison, USA

High b-value q-space diffusion MRI was reported to be sensitive to myelination disorders. In the present study we used high b-value q-space diffusion MRI and conventional DTI to describe, for the first time,  the diffusion characteristics of myelin-deficient (md) rat brains and their age-matched controls. This study clearly demonstrates that myelin affects, significantly, all three diffusion indices obtained from q-space DWI (i.e. displacement probability and kurtosis). This study also shows that FA, extracted from conventional DTI, blurs the differences between the two groups.

Cheryl R. McCreary¹, ², Thorarin A. Bjarnason², Viktor Skihar¹, ², J Ross Mitchell¹, ², V Wee Yong¹, ², Jeff F. Dunn¹, ²

¹Unversity of Calgary, Calgary, Canada; ²University of Calgary, Calgary, Canada

Identification of remyelination is important in the evaluation of potential treatments of demyelinating diseases like multiple sclerosis. Areas of demyelination with spontaneous remyelination have been observed in mice after a local injection of lysolecithin. The aim of this study was to determine if multi-exponential T2 analysis and magnetization transfer imaging, both indicative of myelin content, could detect changes in myelination, particularly remyelination, of the cervical spinal cord in mice treated with lysolecithin.  We found that the short and intermediate T2 components showed significant changes over time and had began to return to control levels, while the MTR remained lower over the time course studied.

Istvan Pirko¹, Yi Chen¹, Jeremiah McDole¹, Scott Dunn², Diana Lindquist², Aaron J. Johnson¹

¹University of Cincinnati, Cincinnati, Ohio, USA; ²Cincinnati Childrens Hospital Medical Center, Cincinnati, Ohio, USA

Brain atrophy is widely recognized in MS, but its pathomechanism remains unclear. Extensive central atrophy has not been reported in MS models. We demonstrate significant brain atrophy in TMEV infected SJL mice, an accepted MS model. Brain parenchymal fraction and ventricular volume analysis showed significant atrophy as early as 3 months, whereas motor deficits became significant at 4 months after induction. Atrophy progressed until 6 months, motor disability until 9 months after induction. NAA decrease accompanied the progressive atrophy. This model will enable us to investigate the pathomechanism of brain atrophy, and may lead to novel therapies addressing MS-related neurodegeneration.

Laura Adela Harsan¹, Dominik Paul¹, Dominik von Elverfeldt¹, Jerome Steibel², Jürgen Hennig¹, Said M. Ghandour²

¹University Hospital Freiburg, Freiburg, Germany; ²Université Louis Pasteur, Strasbourg, France

In vivo DT-MRI was used to quantify the demyelination extent and the effects of a novel thyroid hormone (T3) based therapy, applied to induce recovery in the chronic demyelinated mouse brains. Long-term cuprizone treatment in mice resulted in severe and irreversible brain demyelination. The pathology caused changes of DT-MRI derived parameter values, including loss of anisotropy and increase of radial water diffusion values, D(radial). T3 hormone injections restored progressively a normal level of myelin. The microstructural reorganization of the white matter fiber tracts during the remyelination involved gradual recovery of the white matter anisotropy and restoration of normal D(radial) values.

Mingqiang Xie¹, Regina C. Armstrong², Sheng-Kwei Song³

¹Washington University in St. Louis, St. Louis, Missouri, USA; ²Uniformed Services University of the Health Sciences, Bethesda, Maryland, USA; ³Washington University, St. Louis, Missouri, USA

SynopsisMultiple sclerosis (MS) is an inflammatorily repeated or chronic demyelinating disease.  Although populations of oligodendrocyte progenitors (OP) persist in the chronically demyelinating lesions, they typically fail to differentiate into myelinating oligodendrocytes.  Previous study showed that absence of fibroblast growth factor 2 (FGF2) promotes OP differentiation into remyelinating oligodendrocytes and enhances remyelination.  In this study, to monitor the spatial and temporal progression of spontaneous remyelination non-invasively, in vivo DTI biomarker of myelination, radial diffusivity was measured longitudinally in cuprizone treated live FGF2 -/- mice throughout the course of recovery period.  The result showed that radial diffusivity in FGF2 -/- mice increased significantly following 12 weeks of cuprizone ingestion and returned to the control level at the end of the 12-week recovery period.  Therefore, DTI detection was a sensitive measure for in vivo detection of the improved repair of chronic demyelination.

Raoul Oude Engberink¹, Elga de Vries², Annette van der Toorn¹, Susanne van der Pol², Christien Dijkstra², Erwin Blezer¹

¹University Medical Center Utrecht, Utrecht, Netherlands; ²VU Medical Center Amsterdam, Amsterdam, Netherlands

We studied the fate of USPIO longitudinally in a rat model of multiple sclerosis. Repetitive MRI was performed within a 6h period following a single intravenous USPIO injection both at onset as peak of the disease. Contrast effects in the brain are detected in 1h after injection and histological validation at 6h shows extra-cellular USPIO in the brain parenchyma. Signal changes in the brain are no longer detected 72h after injection and imaging of the cervical lymph nodes reveals USPIO accumulation over time. This study identifies USPIO as a marker for BBB damage in an early time frame.

Shu-Wei Sun¹, Hsiao-Fang Liang¹, Anne H. Cross¹, Sheng-Kwei Song¹

¹Washington University School of Medicine, St Louis, Missouri, USA

In this study, the feasibility of using axial and radial diffusivities in characterizing the primary and secondary axonal damage of EAE mice was evaluated. The decreased axial followed by increased radial in optic tract (OT) suggested that OT was likely damaged by the Wallerian degeneration.  Since Wallerian degeneration can be delayed in Wlds mice, the delayed damage of OT in Wlds EAE, i.e., one month later than the damage to optic nerve (ON), supported that OT damage is a secondary degeneration originating from the initial ON damage.

Shu-Wei Sun¹, Hsiao-Fang Liang¹, Anne H. Cross¹, Sheng-Kwei Song¹

¹Washington University School of Medicine, St Louis, Missouri, USA

Axonal and neuronal loss is the primary cause of the permanent disabilities in patients with MS.  However, the relation between axonal and neuronal damage is still not clearly. Derived from DTI, axial and radial diffusivities have been used to detect axonal and myelin in optic nerve from EAE mice, an animal model for human MS.  In this study, significant correlations between retinal ganglion cells (RGC, the cell body of optic nerve axons) loss and DTI abnormalities to optic nerve were demonstrated.  This study suggested a causal relationship between the observed axonal and neuronal injuries in EAE mice.

Matthew D. Budde¹, Sheng-Kwei Song¹

¹Washington University, Saint Louis, Missouri, USA

A decrease in axial diffusivity correlates with both axonal damage and hindlimb motor function in the spinal cord white matter of mice with experimental autoimmune encephalomyelitis (EAE), a model of Multiple Sclerosis.  However, mice with EAE have varying degrees of long-term impairment as well as heterogeneous distributions of lesions.  In the current study, both DTI parameter maps and histological sections stained for axonal damage were registered to a common space to address the relationship between localized axonal damage and neurological impairment in EAE.  Both modalities demonstrate axonal damage is present throughout the white matter, not solely within lesions.

                  2114.     High Resolution ¹H NMR Spectroscopic Based Metabolomic Urine Analysis of EAE, an Experimental Murine
                            Model of Multiple Sclerosis

Harold G. Parkes¹, Sarah Romero Shorter², Po-Wah So³, David Baker⁴, Gavin Giovannoni⁴, Gareth Pryce⁴, Klaus Schmierer¹

¹University College London, London, UK; ²Birkbeck College, London, UK; ³Hammersmith Hospital, Imperial College London, London, UK; ⁴Queen Mary University of London, London, UK

EAE is an experimentally induced autoimmune disease in mice, exhibiting a similar disease progression as multiple sclerosis (MS) pathology in man including different pathology categorised as acute, chronic, remitting and relapse.  ¹H NMR spectroscopic analysis of urine showed significantly decreased excretion of certain metabolites, and specific to each pathology type.  Thus, urine biomarkers may be generated by ¹H spectroscopy to aid diagnosis and monitoring as well as pathological classification of clinical MS.

sebastien Serres¹, Yanyan Jiang¹, Damian Tyler¹, Daniel Anthony¹, Nicola Sibson¹

¹University of Oxford, Oxford, UK

Multiple sclerosis (MS) is a neurodegenerative disease characterised by a chronic inflammatory disorder associated with demyelination. MS can display a relapsing-remitting form of disease in which bacterial infection may play a role. The aim was to discover whether a systemic infection can reactivate a quiescent MS-like lesion in the brain by using MRI and immunohistochemistry. Early increase of regional cerebral blood volume (rCBV) surrounding the lesion provides an insight into systemic reactivation of MS-like lesions and revealed the reactivation of a previously quiescent CNS lesion can be induced by a systemic inflammatory response in a clinically-relevent model of MS.

Madhavi Pai¹, ², Peter F. Bousquest², Annette J. Schwartz², Bradford L. Mcrae², Christine M. Nelson², Vincent P. Hradil¹, Bryan F. Cox¹, Gerard B. Fox¹, Chih-Liang Chin¹

¹Abbott Laboratories, Abbott Park, Illinois, USA; ²Abbott Bioresearch Center, Worcester, Illinois, USA

Experimental autoimmune encephalitis (EAE) is a widely-used animal model of multiple sclerosis.  Recently, it has been demonstrated USPIO-enhanced MRI allows monitoring macrophage infiltration in vivo during the disease course of EAE.  Here, we investigated the occurrence of lesions and its implications on observed neurological deficits in relapsing-remitting EAE.  Results indicate USPIO-labeled lesions occurred in distinct CNS regions at various phases: brainstem (acute), no lesion (remission), and cerebellum and spinal cord (relapse), which reveal the involvement of spino-olivocerebellar pathways (cerebellar control of the limb posture and movement), in EAE.  Our data provide important insights into further understanding of this disease model.

Mixed Brain Pathology

Gabriele Ende¹, Nuran Tunc-Skarka¹, Derik Hermann¹, Mareen Hoerst¹, Sigi Walter¹, Matthias Ruf¹, Katharina Kraus¹, Sabine Klein¹, Karl Mann¹

¹Central Institute of Mental Health, Mannheim, Germany

Previous MRS studies have left the question how and when the Cho increase with alcohol consumption reverts into decreased Cho in abstinent alcoholics. We studied alcohol dependent patients early during detoxification and 2 weeks into abstinence, heavy alcohol drinkers and light to medium social drinkers at 3T. We see first evidence for our hypothesis that alcohol triggers non-linear dynamic changes of the Cho concentration: Cho initially increases with alcohol consumption but starts to decrease when alcohol abuse starts and significantly decreases during alcohol early abstinence. In contrast NAA shows a trend towards recovery within the first 2 weeks of abstinence.

Nuran Tunc-Skarka¹, Mareen Hoerst¹, Tim Wokrina¹, ², Gabriele Ende¹

¹Central Institute of Mental Health, Mannheim, Germany; ²now at Bruker BioSpin MRI GmbH, Ettlingen, Germany

It has been previously shown at a 1.5T scanner that in alcohol dependent patients after detoxification the Cho and its ratio to Cr is below-normal and increases with duration of abstinence. Now we aimed to replicate the correlation at a 3T scanner and also wanted to examine whether the correlation is caused by Cho concentration change or if it might be an effect of altered relaxation. We could corroborate a significant positive correlation between the Cho concentration in the FWM and alcohol consumption of the last two weeks (R = 0.823, p = 0.000).

Anderson Mon¹, Timothy Craig Durazzo¹, Stefan Gazdzinski², Dieter Johannes Meyerhoff¹

¹UCSF, San Francisco, California , USA; ²CIND, San Francisco, California , USA

Synopsis: We used longitudinal perfusion MRI analysis to study cerebral gray matter perfusion changes in short-term abstinent alcoholics. The subjects were divided into cigarette-smoking alcoholics and non-smoking alcoholics. We observed that cerebral perfusion significantly increased in the non-smoking alcoholics over the first month of abstinence from alcohol but did not change in the smoking alcoholics. This suggests that cigarette smoking modulates cerebral perfusion recovery.

Christine Delmaire¹, Marie Vidailhet², Maxime Descoteaux³, Demian Wassermann³, Frederic Bourdain⁴, Christophe Lenglet³, Sophie Sangla², Axel Terrier⁵, Rachid Deriche³, Stéphane Lehéricy⁶

¹Roger Salengro Hospital, Lille, France; ²Pitie Salpetriere Hospital, Paris, France; ³INRIA, Sophia Antipolis, France; ⁴Foch Hospital, France; ⁵Saint Antoine Hospital, Paris, France; ⁶University Pierre and Marie Curie-Paris 6, Paris, France

In this study, we investigated white matter abnormalities using diffusion tensor imaging (DTI) in 26 patients with focal hand dystonia compared with 26 healthy volunteers. SPM analysis showed that patients had increased FA values bilaterally in the area of the posterior limb of the internal capsule. FA abnormalities were located along the fiber tracts connected to the primary motor and sensory areas. These FA abnormalities were likely to reflect a specific disturbance of the white matter pathways that carry afferents and efferents to the primary sensory motor cortex.

Kieren Grant Hollingsworth¹, Ahmed M. El-Sharkawy¹, Zia U. Khan¹, Andrew M. Blamire¹, Roy Taylor¹, David E. Jones¹, Julia L. Newton¹

¹Newcastle University, Newcastle Upon Tyne, UK

Autonomic dysfunction (AD) is a frequent finding in the autoimmune liver disease Primary Biliary Cirrhosis (PBC). This study examined structural lesions in 29 early stage PBC patients. Lesions were quantified on T2w images by two trained, independent observers according to the Scheltens scale. Autonomic function of the patients was assessed using a continuous beat-to beat Taskforce system. Total lesion load and lesion load in the globus pallidus was found to correlate with impaired baroreflex sensitivity. This is consistent with AD arising in PBC secondary to central effects. 

Daniel T. Franc¹, Bryon Mueller, Joline Dalton, Cameron Naughton, John W. Day, Kelvin O. Lim

¹University of Minnesota, Minneapolis, USA

Significant reductions in fractional anisotropy throughout cerebral white matter were observed in three Myotonic Dystrophy populations.

Joshua Francis Betz¹, ², Stephen L. Seliger, David Lefkowitz, Jiachen Zhuo, Shari R. Waldstein, Leslie I. Katzel, Rao P. Gullapalli

¹University of Maryland Baltimore County, Baltimore, USA; ²University of Maryland, Baltimore, USA

Chronic Kidney Disease is associated with a number of factors implicated in structural and functional changes in the brain. The goal of the study is to elucidate relationships between clinical markers of renal function and structural changes related to aging. After controlling for the effects of age, clinical markers of renal function and cardiovascular health were related to lower region-specific measures of white and gray matter, and higher region-specific measures of cerebrospinal fluid in patients with Chronic Kidney Disease. Our research indicates that Chronic Kidney disease is related to morphological changes which are detectable through quantitative cranial MRI.

Myriam Chaumeil¹, ², Hélène Bataille¹, Fawzi Boumezbeur¹, Julien Valette¹, Anne-Catherine Bachoud-Levi³, Philippe Hantraye², Vincent Lebon¹, ², Pierre Brugières³

¹NeuroSpin, Gif-sur-Yvette, France; ²MIRCen, Fontenay-aux-Roses, France; ³Hôpital Henry Mondor, Créteil, France

Metabolic abnormalities associated with Huntington’s Disease (HD) can be detected by NMR spectroscopy. Up to now ¹H spectroscopy in HD has revealed almost systematic decreases in N-acetyl-aspartate and occasional increases in lactate and glutamate+glutamine. Although the glial marker myo-inositol could be a potential marker of degeneration, it has not been quantified yet in the adult form of HD. In this study short echo time ¹H spectroscopy of the striatum was combined with LCmodel analysis in order to quantify myo-inositol in HD patients. We observed significantly elevated Ins in the striatum and strong correlation of the Ins/NAA ratio with clinical scores.

Radu Serban Jasinschi¹, Ahmet Ekin¹, Adriaan van Es², Mark Augustinus van Buchem², Rene Engbers³, C. Damkat¹, Caroline Jurgens², Jeroen van der Grond²

¹Philips Research, Eindhoven, Netherlands; ²Leiden University Medical Center, Leiden, Netherlands; ³Tass, Eindhoven, Netherlands

We investigate two MRI features computed in the Basal Ganglia (BG) of Huntington’s disease (HD) patients. One feature is the hypointensity load in T2-weighted (T2-w) images, and the other one corresponds to gray matter (GM) loss detected in T1-weighted (T1-w) images. We automatically select BG regions [2] given by a set of grid cells which define the region-of-interest (ROI). The T2-w hypointensity load within the ROI is computed by comparing image brightness with a threshold determined based on ROC curve analysis. The tissue segmentation was implemented based on a K-harmonic means clustering algorithm. We tested these two features on 28 subjects, 14 HD patients and 14 controls. We found out that, on average, the percentage of hypointense pixels in the BG is 3.95 times higher for HD patients compared to the controls. Also, the HD patients have 87 % less GM compared to controls, and about 8 % more white matter (WM), significant at p = 0.05 in t-test.

Stephen J. Sawiak¹, Guy Barnett Williams¹, Nigel I. Wood¹, A J. Morton¹, T A. Carpenter¹

¹University of Cambridge, Cambridge, UK

Mouse models of disease are of increasing importance, in part because of the comparative ease of developing transgenic variants. Growing amounts of data make traditional analysis techniques such as histology prohibitive. Whilst MRI offers the potential for in-vivo investigation, the analysis problem is not yet fully addressed. Here we present our work in extending automated voxel-based morphometry techniques to the mouse brain, using the popular statistical parametric mapping (SPM) package. Although our work is applicable to any mouse brain model, here we present our methods and results from the R6/2 model of Huntington's disease with 87 three-dimensional, high resolution datasets.

Ines Blockx¹, Nadja Van Camp, ¹², Marleen Verhoye¹, Johan Van Audekerke¹, Huu Phuc Nguyen³, ⁴, Stephan Von Horsten³, Olaf Riess⁵, Annemie Van der Linden¹

¹University of Antwerp, Antwerp, Belgium; ²Commissariat a L'Energie Atomique - CEA, Orsay, France; ³Medical school of Hannover, Hannover, Germany; ⁴University of Tubingen , Germany; ⁵University of Tubingen, Germany

Diffusion tensor imaging has been widely used for detailed analyses of tissue morphology and pathology. Previous DTI studies have successfully demonstrated neurodegenerative changes in a lesion model of Huntington Disease. Transgenic models provide the advantage of mimicking the human pathology more closely and in addition they display a more progressive course of the disease. We aimed at revealing the first signs of neurodegeneration by investigating changes of microstructure and neuroconnectivity in young presymptomatic (2 months) transgenic animals. The results of this study will be highly valuable for pre-clinical screening as well as for the understanding of the specific underlying pathogenesis.

Jiangyang Zhang¹, Kenichi Oishi¹, Qi Peng¹, Qing Li¹, Michael I. Miller², Susumu Mori¹, Wenzhen Duan¹

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

We studied the progression of brain atrophy in R6/2 mice, a model of Huntington’s disease, using in vivo longitudinal MRI. High resolution images captured macroscopic changes in brain morphology from three weeks to 12 weeks after birth in the same animals. Quantitative analysis revealed atrophy in the striatum, motor and pimiform cortex, as well as enlargement of the lateral ventricles in 5 week old mice. Longitudinally, the rates of brain atrophy were not uniform. Our results suggested that atrophy in the striatum was most active at around 5 weeks old.

Byeong-Yeul Lee¹, Jeffrey Vesek¹, Shoaib Alam¹, James R. Connor¹, Qing X. Yang¹

¹Penn State Univeristy College of Medicine, Hershey, Pennsylvania, USA

Restless leg syndrome (RLS) is a sensory-motor disorder coupled with iron deficiency in the RLS brain. It is not known whether or not RLS brain volume is different from normal control. We hypothesize that insufficient iron concentration in RLS may contribute to the global regional change. To test our hypothesis, we conducted a voxel-based morphometry study on RLS. Our data demonstrate that a significant volumetric increase was found in the gray and white matter in RLS. Thus, we may speculate that iron deficiency contribute to the increase in volume in RLS patients compared to controls.

Byeong-Yeul Lee¹, Jeffrey Vesek², Elana Farace², Shoaib Alam², James R. Connor², Qing X. Yang²

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

Restless legs syndrome (RLS) is a neurological disorder in which iron deficiency in central nervous system is a contributing factor. However, in vivo human studies which link iron deficiency to brain metabolism in RLS brain have not been previously reported. Using in vivo MRI and proton MRS, we investigated the effect of insufficient iron contents on the neurochemical metabolism in RLS.The result showed that decreased in R2 values and increase in various metabolite concentrations in the RLS brain. Thus, the findings of our study may provide insight into neurochemical derangements resulting from iron deficiency in RLS.

Giovanni Rizzo¹, ², Pasquale Montagna¹, Caterina Tonon¹, Roberto Vetrugno¹, Claudia Testa¹, Giuseppe Plazzi¹, Federica Provini¹, Bruno Barbiroli¹, Raffaele Lodi¹

¹University of Bologna, Bologna, Italy

Restless legs syndrome (RLS) is a common disorder in the general population. The pathophysiology of RLS is poorly understood. Recent studies have suggested a potential thalamic involvement. We performed a ¹H-MRS study in patients with RLS and normal controls selecting a volume of interest at the level of the medio-posterior region of the thalamus, in order to identify metabolic changes in this structure. In RLS patients we detected a significant reduction in NAA/Cr and in the absolute concentration of NAA. Our preliminary spectroscopic data confirm a thalamic involvement in RLS patients.

Jana Godau¹, Katharine Schweitzer¹, Adriana Di Santo¹, Daniela Berg¹, Uwe Klose¹

¹University Hospital Tuebingen, Tuebingen, Germany

Restless Legs Syndrome (RLS) is a disorder, for which brain iron deficiency is suspected. In this study with RLS patients, we examined alterations of the T2 relaxation, which is expected to occur with reduced iron content. T2-values were calculated from 2 echoes from a Turbo spin-echo sequence. Circular ROIs were bilaterally placed in ten selected regions. In every examined brain region the average T2 value histogram plots were remarkably shifted to higher T2 intensities in the RLS patient group. Increased T2 values in RLS patients compared to controls suggest decreased iron content in all examined brain regions.

Margaret R. Lentz¹, Vallent Lee¹, Susan V. Westmoreland², Eva-Maria Ratai¹, Elkan F. Halpern³, R. Gilberto Gonzalez¹

¹Massachusetts General Hospital, Charlestown, USA; ²New England Primate Research Center, Southborough, USA; ³Massachusetts General Hospital, Boston, USA

¹H MRS provides multivariate data on metabolites in the HIV-infected brain, and can benefit from the use of factor analysis (FA). We applied FA to identify patterns in the metabolic processes underlying the pathogenesis of neuroAIDS, using frontal cortex extracts from SIV-infected macaques moribund with AIDS. One factor could discern between SIV/AIDS animals and healthy controls. Another factor could distinguish animals with encephalitis, and was significantly different across classifications of encephalitis severity. We demonstrate the strengths of FA and discuss the implications that the data raise regarding the analysis of in vivo MRS.

Napapon Sailasuta¹, Kimberly Shriner², Brian Ross¹

¹Huntington Medical Research Institutes, Pasadena, California , USA; ²Huntington Hospital, Pasadena, California , USA

First report of direct measurement of brain glutamate concentrations in HIV patients