Perfusion & Permeability
 

Wednesday 3 June 2015

Jeiran Jahani¹, Timothy M Shepherd¹, Glyn Johnson¹, Valerij G Kiselev², and Dmitry S Novikov^1
1Department of Radiology, New York University School of Medicine, New York City, New York, United States, ²Department of Radiology, Medical Physics, University Medical Center Freiburg, Freiburg, Germany

Indicator dilution techniques have been at the heart of vascular disease diagnostics for decades. The current approaches consider the first pass of the contrast and cannot account for recirculation. This complicates data interpretation as the first pass and recirculation boluses overlap. For the first time, we develop a global approach to contrast circulation within the body with account for all organs and the topology of the circulation system. This framework allows us to obtain the explicit form of contrast kinetics in individual organs. It also provides us with AIF without directly measuring it, and a practical way to map the local absolute CBV without the need for the first pass extraction.

Fernando Calamante¹, André Ahlgren², Matthias J.P. van Osch³, and Linda Knutsson^2
1The Florey Institute of Neuroscience and Mental Health, University of Melbourne, Melbourne, Victoria, Australia, ²Department of Medical Radiation Physics, Lund University, Lund, Sweden, ³Department of Radiology, C.J. Gorter Center for high field MRI, Netherlands

The % blood-volume occupied by red-blood-cells is known as hematocrit (Hct). Whereas it is straightforward to measure Hct in large arteries (e.g. blood sample), it is very challenging to do it in microvasculature (cerebral Hct). Currently, this can only be done using invasive methods (e.g. PET, SPECT, autoradiography), but their use is very limited. Local variations in cerebral Hct have been reported in various brain abnormalities (e.g. stroke, tumours). We propose an MRI method to image cerebral Hct, which relies on combining data from two MRI measurements: one that provides Hct-weighted, and another one Hct-independent values of the same parameter.

Steven Sourbron^1
1University of Leeds, Leeds, UK, United Kingdom

A tracer-kinetic field theory has been proposed for DCE- and DSC-MRI that may reduce error and provide new information by spatial modelling of indicator transport. In this study a 2D simulation of the simplest one-compartment field model is used to investigate whether all model parameters are uniquely identifiable. The discrete system is written as a large sparse linear inverse problem and solved iteratively using constrained gradient descent. A well-defined optimum is found but the reconstructed fields are not sufficiently accurate for a simulation of ideal conditions. It is concluded that a better reconstruction algorithm is required, or additional constraints must be imposed.

Li Zhao¹, Weiying Dai¹, and David Alsop^1
1Radiology, Beth Israel Deaconess Medical Center & Harvard Medical School, Boston, MA, United States

Arterial spin labeling provides important perfusion information, but it suffers from intrinsically low SNR and spatial resolution. Since high quality T1 weighted images are collected in routine MRI procedures, we present a method to enhance ASL image by incorporating its local similarity with anatomic images. The proposed method was evaluated on five tumor patients and results preserved the low resolution contrast of ASL with improved CNR and spatial resolution.

Jack T Skinner^1,2, Paul L Moots³, and C Chad Quarles^1,2
1Radiology and Radiological Sciences, Vanderbilt University Medical Center, Nashville, TN, United States, ²Institute of Imaging Science, Vanderbilt University Medical Center, Nashville, TN, United States, ³Neurology, Vanderbilt University Medical Center, Nashville, TN, United States

Leakage of contrast agent can confound DSC-MRI measurements, necessitating correction techniques. Parameters relating to contrast extravasation (K₂, K_a) can be extracted from these techniques. Dual gradient-echo acquisitions also mitigate leakage effects and allow the extraction of T₁-weighted data and subsequently the computation of DCE-MRI parameters K^trans and v_e. Using dual-echo DSC-MRI, these parameters were simultaneously compared. A poor voxel-wise correlation was found between K₂ and K_a andK^trans, potentially attributed to competing T₁ and T₂* leakage effects and echo time dependence. A strong correlation was observed, however, between K₂ and K_a and v_e, suggesting a relationship with the extravasation space.

Ashley M Stokes¹ and C. Chad Quarles^1
1Institute of Imaging Science, Radiology and Radiological Sciences, Vanderbilt University, Nashville, TN, United States

Contrast agent leakage in tumors can severely reduce the reliability of DSC perfusion measures due to competing T₁ effects. While dual gradient-echo (GE) sequences can provide robust T₁-insensitive GE hemodynamic measures, no analogous method exists for T₁-insensitive spin-echo (SE) measures. Here, we propose a simplified spin- and gradient-echo (sSAGE) approach that utilizes a combined dual GE and SE sequence and an analytical solution for T₁-insensitive ∆R₂^* and ∆R₂ timecourses. As this approach only requires acquisition and storage of three echoes and does not rely upon computationally demanding non-linear fitting, it could facilitate more rapid clinical translation and adoption of SAGE-based DSC-MRI.

Natenael B Semmineh¹, Ashley M Stokes¹, John C Gore¹, and C Chad Quarles^1
1Institute of Imaging Science, Vanderbilt University, Nashville, TN, United States

Recently, we demonstrated that T2* and T2 weighted contrast enhanced MRI signals, when acquired at contrast agent equilibrium, are predominantly influenced by the cellular properties of tissue. We now propose a novel imaging approach that utilizes these T2* and T2 effects to derive an estimate of mean cell size in a voxel. Using simulations, in vitro and in vivo studies we demonstrate that the ratio of ΔR2* and ΔR2 is related to cellular size and may be used to estimate mean cell size in a voxel.

Shalini A. Amukotuwa^1,2, Fernando Calamante², and Roland Bammer^1
1Department of Radiology, Stanford University, Stanford, CA, United States, ²The Florey Institute of Neuroscience and Mental Health, University of Melbourne, Parkville, VIC, Australia

Combined spin- and gradient-echo (SAGE) dynamic susceptibility contrast (DSC) perfusion-weighted imaging (PWI) has considerable benefits over conventional single-echo spin-echo (SE) or gradient-echo (GRE) DSC PWI. However, mismatch criteria are based on GRE DSC PWI. Our aim is to investigate the comparability of the second gradient-echo of SAGE with GRE DSC PWI in patients with suspected stroke. Our findings indicate that, with matching scan parameters, the SAGE 2nd gradient-echo yields comparable results to conventional GRE DSC PWI; therefore SAGE data can be calibrated with GRE DSC data, and postprocessing algorithms optimized through previous clinical trials can be applied to SAGE.

Marco Castellaro¹, Alessandra Bertoldo¹, Denis Peruzzo², Filippo Arrigoni², and Matthias Van Osch^3
1Department of Information Engineering, University of Padova, Padova, Italy, ²Department of Neuroimaging, Research institute IRCCS "E. Medea", Bosisio Parini, Lecco, Italy,³C.J. Gorter Center for High Field MRI, Radiology, Leiden University Medical Center, Leiden, Netherlands

Arterial spin labelling (ASL) is a technique that permits to estimate perfusion non-invasively. The readout phase of ASL sequence can follow several schemes. In particular, to sample the entire kinetic curve of the inflowing blood the use of a Look&Locker (LL) read-out has been proposed. The LL approach suffers, however, from a progressively lower Signal-to-Noise Ratio (SNR) for later post-labeling delays (PLDs) due to T1-relaxation and the fact that the label experienced multiple excitation pulses. This work presents an approach to optimize the SNR over all PLDs of a LL read-out scheme by introducing a variable flip-angle (FA) sweep.

Magdalena Sokolska¹, Xavier Golay¹, and David Thomas^1
1UCL Institute of Neurology, London, United Kingdom

Pseudo-continuous Arterial Spin Labeling (ASL) (pCASL) is emerging as a method of choice for the non-invasive measurement of tissue perfusion in research and clinical practice. It is not clear, what the effect of imperfect angulation and mis-placement will have on the labeling efficiency and consequently, on CBF estimation. Therefore this work aims to address this question by simulation and in vivo experiment and quantitatively assess the effect of the angulation of the labeling plane on pCASL labeling efficiency.

Thomas Lindner¹, Ulf Jensen-Kondering¹, Olav Jansen¹, Matthias JP van Osch², and Michael Helle^3
1Department of Radiology and Neuroradiology, UKSH, Kiel, Germany, ²Department of Radiology, LUMC, C. J. Gorter Center for High Field MRI, Leiden, Netherlands, ³Philips Research, Hamburg, Germany

In this study, we present a method for time-resolved artery-selective non-contrast enhanced magnetic resonance angiography. The presented method is based on superselective arterial spin labeling (ASL), which allows tagging of a single artery. Time-resolved imaging is achieved by increasing the post labeling delay for each image, resulting in a temporal resolution of 100ms. To keep the acquisition time to an acceptable amount, keyhole reduced imaging is applied, where the keyhole factor as well as the flip angle were optimized using numerical simulations. This method was successfully applied in healthy volunteers, visualizing the individual flow territories of the intracranial arteries.

R. Marc Lebel^1,2, Ajit Shankaranarayanan³, Eric E. Smith⁴, Cheryl McCreary², Richard Frayne², Weiying Dai⁵, and David C Alsop^5
1GE Healthcare, Calgary, Alberta, Canada, ²Radiology, University of Calgary, Calgary, Alberta, Canada, ³GE Healthcare, Menlo Park, California, United States, ⁴Clinical Neurosciences, University of Calgary, Calgary, Alberta, Canada, ⁵Radiology, Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, Massachusetts, United States

Pseudo-continuous arterial spin labeling (PCASL) is a method for non-contrast perfusion imaging that tags blood flowing across a labeling plane. The labeling plane is active for only a finite period of time prior to imaging. A 1.5 s labeling time is typical. It is well known that the signal-to-noise ratio increases with longer labeling times but additional benefits exist. We describe improved temporal stability and demonstrate a reduced sensitivity to arterial transit delays with longer labeling. Overall we recommend using a label duration that is much longer than what is commonly done.

Tianrui Luo¹ and Luis Hernandez-Garcia^2
1University of Michigan, Ann Arbor, Michigan, United States, ²FMRI Laboratory, University of Michigan, Ann Arbor, Michigan, United States

A method to measure Inflow Velocity Density of brain tissue based on velocity selective ASL is proposed. The method uses modified BIR-8 pulses to generate both cosine and sine velocity dependent profiles in the observed magnetization. Fourier analysis of the resulting data yields the distribution of flow velocities in the arterial blood supply to a given voxel.

Liliana Lourenco Caldeira ¹, Seong Dae Yun¹, Nuno André da Silva¹, Christian Filss¹, and N. Jon Shah^1,2
1Institute of Neuroscience and Medicine (4), Forschungszentrum Juelich, Jülich, Germany, ²RWTH Aachen University, Faculty of Medicine, Department of Neurology, JARA, Aachen, Germany

The arterial input function (AIF) is essential for quantification in MRI and PET imaging. The ground truth for AIF estimation is arterial cannulation, which is a difficult procedure. Alternatively, an image-derived AIF can be estimated using MRI and/or PET images, but a reasonable temporal resolution of dynamic image series is necessary (<2s). In PET, high temporal resolution is limited (>5s) and images are rather noisy. Here, we propose a method to combine simultaneous MR-PET data to estimate the AIF. The MR AIF is based on an EPI with keyhole (EPIK) sequence and then can be converted to a PET AIF.

Lydiane Hirschler^1,2, Clément Stéphan Debacker^1,2, Jérôme Voiron², Jan Warnking^1,3, and Emmanuel Luc Barbier^1,3
1Université Grenoble Alpes, Grenoble Institut des Neurosciences, Grenoble, France, ²Bruker Biospin, Ettlingen, Germany, ³Inserm, U836, Grenoble, France

Performing pseudo-continuous arterial spin labeling (pCASL) at very high magnetic field is challenging as B0 inhomogeneities at the labeling plane, away from the isocenter, strongly affect the spins’ phase and thus inversion efficiency. This study shows that, for an unbalanced pCASL, performing a label and a control phase increment correction improves perfusion signal, corrects asymmetry between brain hemispheres arising when labeling far from isocenter at high field, and preserves high image quality at the same time.

Stefan Hindel¹, Nico Verbeek², Anika Sauerbrey¹, and Lutz Lüdemann^1
1Strahlenklinik und Poliklinik, Universitätsklinikum Essen, Essen, North Rhine-Westphalia, Germany, ²Heinrich-Heine-Universität Düsseldorf, Düsseldorf, North Rhine-Westphalia, Germany

We accessed a method for voxel-based partial volume correction for DCE-MRI. The spiral like trajectory of the complex blood signal was plotted in the complex plain and fitted by a logarithmic spiral to calculate the spiral center. The results demonstrate a strong correlation between echo time and accuracy of the analyzed data. Our results illustrate the limitations for using the method when applied to a 3D-GRE (DCE-MRI) sequence. The T2* 2D-GRE AIF analyzing method helps detecting and correcting partial volume artifacts which affect measurements of the arterial input function.

Wen-Chau Wu^1,2
1National Taiwan University, Taipei, Taiwan, ²Department of Medical Imaging, National Taiwan University Hospital, Taipei, Taiwan

Depicting the primary flow territories of the brain is of clinical importance because collateral circulation is a common finding in patients with steno-occlusive disease. In this study, we proposed to generate statistical mapping of the flow territories of the left/right internal carotid arteries and vertebral arteries by using multi-phase vessel-encoded pseudocontinuous arterial spin labeling and the general linear model. The described method allows comparison between flow territories and/or subject populations to be carried out in a statistical framework similar to that in functional MR imaging.

Tammo Rukat¹ and Stefan A Reinsberg^1
1Department of Physics and Astronomy, University of British Columbia, Vancouver, British Columbia, Canada

A fully Bayesian model mixing method for the estimation of haemodynamic parameters from DCE-MRI is being assessed. In particularly we examine the capability of weighing models of different complexity, such that the resulting parameter can be expected to be more accurate than the estimate from any single model. The Watanabe-Akaike information criterion (WAIC) is derived from the posterior likelihood distributions of the model parameters, which was sampled by adaptive MCMC. WAIC serves to calculate model mixing weights. This method is shown to be superior to the choice of any single model.

Thaís Roque¹, Amit Mehndiratta², Lawrence Kenning³, Martin Lowry³, and Michael Chappell^1
1Institute of Biomedical Engineering IBME, University of Oxford, Oxford, Oxfordshire, United Kingdom, ²Centre for Biomedical Engineering, Indian Institute of Technology Delhi, New Delhi, India, ³Centre for MR investigations CMRI, University of Hull, Hull, United Kingdom

In order to derive brain haemodynamic parameters such as CBF and the residue function, DSC-MRI analysis typically involves a deconvolution problem. If the blood-brain barrier is disrupted, these parameters are artificially modified due to contrast agent (CA) leakage. In this work, a model free, non-parametric deconvolution method was modified (mCPI) to account for CA leakage. In simulations, mCPI accurately estimated haemodynamic parameters in the presence of CA leakage. Glioma clinical data analysis yielded different shapes of residue function for different tumour areas, which permits assessment of flow heterogeneity within and around the tumour.

Adam Michael Bush¹, Gregory Lee², Matt Borzage¹, Vincent Schmithorst², Scott Holland², and John Wood^1
1Children's Hospital Los Angeles USC, Los Angeles, California, United States, ²Pediatric Neuroimaging Research Consortium, Cincinnati Children's Hospital Medical Center, Cincinnati, Ohio, United States

To quantitatively assess PCASL labeling efficiency (LE) we used a technique that labels blood in the feeding vessels and measures the perfusion signal a short distance further downstream in the blood pool. We measured LE in a flow phantom and in-vivo, then compared these results to Bloch simulations of PCASL efficiency. Phantom LE measurements agree with Bloch predictions of LE with respect velocity. Generally, in-vivo LE could be explained by arterial velocity, however, in several in-vivo measurements, LE variability was significantly lower than predicted by velocity alone. These data suggest measurement of LE may be necessary for accurate ASL quantification.

Erin K Englund¹, Zachary B Rodgers¹, Thomas F Floyd², and Felix W Wehrli^1
1Department of Radiology, University of Pennsylvania, Philadelphia, PA, United States, ²Department of Anesthesiology, Stony Brook University, Stony Brook, NY, United States

The blood flow waveform in the peripheral circulation differs substantially from that of the central vasculature. During baseline flow conditions, blood flows both antegrade and retrograde. During reactive hyperemia, average blood flow increases substantially, although the maximum velocity remains relatively unchanged. Here, we explored the implications of varying blood flow velocity at baseline and during reactive hyperemia on pCASL labeling efficiency by measuring signal in the femoral artery during an ischemia-reperfusion paradigm. Results suggest average labeling efficiency to be unaffected by changes in average blood flow velocity, likely because the maximum velocity is relatively unchanged.

Kathrin Lorenz^1,2, Toralf Mildner¹, Torsten Schlumm¹, and Harald E. Möller^1,2
1Max Planck Institute for Human Cognitive & Brain Sciences, Leipzig, Germany, ²Faculty of Physics and Earth Sciences, University of Leipzig, Saxony, Germany

Although pseudo-continuous arterial spin labeling (pCASL) has become the recommended choice for non-invasive perfusion measurements in the human brain, perfusion quantification is still challenging. For instance, the detected ASL signal change, and hence the estimated perfusion value, is directly proportional to the labeling efficiency . Purpose of the current study is to investigate requirements for a high reproducibility of  measurements. It is shown that this property can be improved significantly, if both the slice position and the sampling rate of the experiment are optimized. Thereby, procedural details towards measuring  robustly in one additional pre-scan were obtained.

Shin-Lei Peng^1,2, Pan Su^1,3, Fu-Nien Wang², Yan Cao⁴, Rong Zhang⁵, Hanzhang Lu^1,3, and Peiying Liu^1
1Advanced Imaging Research Center, University of Texas Southwestern Medical Center, Dallas, TX, United States, ²Department of Biomedical Engineering and Environmental Sciences, National Tsing Hua University, Hsinchu, Taiwan, ³Biomedical Engineering Graduate Program, UT Southwestern Medical Center, TX, United States, ⁴Department of Mathematical Sciences, University of Texas at Dallas, Richardson, TX, United States, ⁵Institute for Exercise and Environmental Medicine, Texas Health Presbyterian Hospital Dallas, Dallas, TX, United States

PC-MRI is a noninvasive technique for quantifying whole-brain CBF. However, PC-MRI measured velocity map is susceptible to partial voluming, leading to biases in CBF estimation. This work firstly aimed to optimize in-plane resolution of PC-MRI for CBF quantification. Furthermore, we assessed effects of non-perpendicular imaging slice orientation on CBF quantification. Results showed in-plane resolution of 0.5 mm could serve as an optimal protocol for quantifying whole-brain CBF. Moreover, non-perpendicular positioning of the imaging slice on the targeted artery could result in overestimated CBF. But if the slice orientation is within 10¢X of the ideal angulation, the bias is negligible.

Ina Nora Kompan^1,2 and Matthias Guenther^1,2
1Fraunhofer MEVIS, Bremen, Bremen, Germany, ²mediri GmbH, Heidelberg, Baden-Württemberg, Germany

Pharmacokinetic (PK) modeling is used in DCE MRI to quantify tissue physiology. Here, optimal sampling design based on the Fisher information approach is applied to the Tofts model to find important time points for model fitting. For an assumed underlying parameter distribution it is found that fast sampling during the first two minutes after contrast agent onset is important for fitting accuracy. It is also shown that optimal sampling schemes outperform equidistant sampling schemes for small number of sampling points with respect to fitting accuracy.

Michael Ingrisch¹, Michael Peller¹, Birgit Ertl-Wagner², Maximilian F Reiser², and Olaf Dietrich^1
1Josef-Lissner-Laboratory for Biomedical Imaging, Institute for Clinical Radiology, Ludwig-Maximilians-University Hospital, München, Germany, ²Institute for Clinical Radiology, Ludwig-Maximilians-University Hospital, München, Germany

Acceleration of the image acquisition is an ever important technique in dynamic contrast-enhanced (DCE) MRI. Conventional parallel imaging is limited by the SNR loss at higher acceleration factors and by potential reconstruction artefacts. A recently introduced method, CAIPIRINHA, promises a more robust image reconstruction by modifying the k-space sampling schema. In the present study, we perform phantom measurements to identify the optimal CAIPIRINHA sampling pattern for a 3D DCE MRI perfusion measurement with coverage of the entire neurocranium at high temporal resolution, and we demonstrate the application of CAIPIRINHA acceleration for cerebral DCE MRI in patient measurements

Moran Artzi^1,2, Gilad Liberman^1,3, Guy Nadav^1,4, Deborah T Blumenthal⁵, Orna Aizenstein¹, and Dafna Ben Bashat^1,6
1Functional Brain Center, Tel Aviv Sourasky Medical Center, Tel Aviv, Israel, ²Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel, ³Department of Chemical Physics, Weizmann Institute, Rehovot, Israel, ⁴Functional Brain Center, Tel Aviv University, Tel Aviv, Israel, ⁵Neuro-Oncology Service, Tel Aviv Sourasky Medical Center, Tel Aviv, Israel, ⁶Sackler Faculty of Medicine and Sagol School of Neuroscience, Tel Aviv University, Tel Aviv, Israel

The effect of dynamic contrast enhancement (DCE) acquisition duration on the pharmacokinetic-parameters estimation was investigated using simulated and real data obtained from seven patients with glioblastoma. DCE parameters were estimated using DUSTER, a method for DCE Up-Sampled-Temporal-Resolution, and with scan duration from 6 to 35 minutes. This study suggests an optimized protocol that can be easily implemented in routing clinical setup. Scan duration of 20 minutes was found to be sufficient to obtain reliable and accurate pharmacokinetic parameters from DCE, especially kep. Results emphasize that care should be taken when using model selection with data acquired with short scan duration.

Chong Duan¹, Jesper F Kallehauge², Carlos J Perez-Torres³, Kari Tanderup^4,5, Larry Bretthorst³, Joseph JH Ackerman^1,3, and Joel R Garbow^3
1Chemistry, Washington University, Saint Louis, Missouri, United States, ²Medical Physics, Aarhus University, Aarhus, Denmark, ³Radiology, Washington University, Saint Louis, Missouri, United States, ⁴Radiation Oncology, Washington University, Saint Louis, Missouri, United States, ⁵Oncology, Aarhus University, Aarhus, Denmark

Many DCE-MRI data models employ a directly determined upstream arterial input function (AIF). The actual AIF at the capillary inlet of any given region of interest is both delayed and dispersed compared with the upstream AIF measured in a major feeding vessel. We build a constrained local AIF (cLAIF) model to better model DCE-MRI data, accounting for both delay and dispersion. Voxel-wise model selection analyses show the cLAIF is preferred over the upstream measured AIF for ~80% of the voxels.

Yi Dang¹, Jia Guo², Jue Zhang^3,4, and Eric Che Wong^5
1Magnetic Resonance Imaging Research Center,Institution of Psychology, Chinese Academy of Sciences, Beijing, Beijing, China, ²Department of Bioengineering, University of California San Diego, CA, United States, ³Academy for Advanced Interdisciplinary Studies, Peking University, Beijing, China, ⁴College of Enigneering, Peking University, Beijing, China, ⁵Department of Radiology and Psychiatry, University of California San Diego, CA, United States

Conventional vessel-encoded arterial spin labeling (VEASL) is a territorial perfusion imaging technique to identify the perfusion territories of cerebral arteries simultaneously with prior knowledge of their positions. Recently, random vessel encoded arterial spin labeling (R-VEASL) was proposed to automate the scan prescription process. Our study compared the output of R-VEASL with the perfusion maps acquired with conventional VEASL. Quantitative comparison of perfusion territories between R-VEASL and conventional VEASL has shown that perfusion territories of both methods agree reasonably well. However, it seems that R-VEASL performs better than conventional VEASL when detecting perfusion region with mixed perfusion.

Edward Ashton¹ and Jill Fredrickson^2
1VirtualScopics, Inc., Rochester, NY, United States, ²Genentech, Inc., South San Francisco, CA, United States

Dual-baseline DCE-MRI data for 43 patients drawn from three Phase 1 clinical trials are analyzed both with and without conversion of data from signal intensity values to millimolar gadolinium concentration. Both model-based (KTrans) and model-free (AUCBN) parameters are calculated. Coefficients of variability are estimated for both parameters using both methods, and systematic bias between the two methods is calculated. Variability is approximately 25% lower for both parameters using signal intensity, primarily because conversion to gadolinium concentration is less robust to patient motion and other image quality issues. Absolute values of both parameters are systematically higher using signal difference methods.

Chang Kyung Lee¹, Bohyun Kim¹, Nieun Seo¹, Jeong Kon Kim¹, In Seong Kim², Berthold Kiefer³, and Kyung Won Kim^1
1Radiology, Seoul Asan Medical Center, Seoul, Seoul, Korea, ²Siemens Healthcare, Seoul, Seoul, Korea, ³Siemens Healthcare, Erlangen, Erlangen, Germany

We compared the breathing motion artifact and signal stability of conventional VIBE, Radial-VIBE with KWIC view-sharing, and CAIPIRINHA-VIBE with moving phantom to simulate breathing motion. Among the three sequences, the CAIPIRINHA-VIBE showed the best image quality, in that the phantom images maintained its shape and just moved its position anteriorly/posteriorly. Thus, the motion correction could align the CAIPIRINHA-VIBE images very well. In terms of signal stability, the CAIPIRINHA-VIBE with motion correction was the best. In conclusion, CAIPIRINHA-VIBE is quite feasible for free-breathing DCE-MRI with high image quality and high temporal resolution.

Dimo Ivanov¹, Anna Gardumi¹, Benedikt A Poser¹, Josef Pfeuffer², and Kâmil Uludağ^1
1Department of Cognitive Neuroscience, Maastricht University, Maastricht, Netherlands, ²Application Development, Siemens Healthcare, Erlangen, Germany

In this study, a pCASL and two PASL variants at 3 T are compared with a PASL approach at 7 T to determine, which of the techniques is most suited for simultaneous high temporal resolution functional CBF and BOLD measurements. The 3 T techniques show no significant differences in either grey matter perfusion or temporal SNR values. In contrast, the 7 T variant tested delivers significantly lower grey matter CBF and temporal SNR values, while offering improved BOLD sensitivity. This indicates that further technical improvements might be necessary for 7 T ASL to be able to reach its full potential.

Shuang Yang¹, Tianyi Qian², Jianwei Xiang³, Yingchao Liu⁴, Peng Zhao⁴, Josef Pfeuffer⁵, Guangbin Wang¹, and Bin Zhao^1
1Shandong Medical Imaging Research Institute, Shandong University, Jinan, Shandong, China, ²MR Collaborations NE Asia, Siemens Healthcare, Beijing, China, ³Shandong Medical Imaging Research Institute, Taishan Medical University, Jinan, Shandong, China, ⁴Neurosurgery, Shandong provincial Hospital Affiliated to Shandong University, Shandong, China, ⁵Application Development, Siemens Healthcare, Erlangen, Germany

A 3D multi-TI arterial spin-labeling (mTI-ASL) protocol had been applied to measure the blood perfusion of patients with brain tumor. This study present the feasibility of mTI-ASL for brain tumors in the clinical environment and compare it with dynamic-susceptibility contrast-enhanced perfusion imaging (DSC) to test the hypothesis whether mTI-ASL could substitute DSC in brain tumor cases. The results shows the performances of mTI-ASL in measuring CBF and timing information of blood flow were consistent with DSC. Thus, the perfusion information and the results obtained by dynamic contrast-enhanced technique could be obtained with only one single contrast agent injection.

Gena Matta^1,2, Andrew D Robertson¹, Sandra E Black^1,3, and Bradley J MacIntosh^1,3
1Canadian Partnership for Stroke Recovery, Sunnybrook Research Institute, Toronto, Ontario, Canada, ²University of Waterloo, Waterloo, Ontario, Canada, ³University of Toronto, Toronto, Ontario, Canada

Arterial-spin labelling (ASL) and phase-contrast angiography (PCA) provide two non-invasive methods to quantify the total cerebral blood flow. The primary aim of this study was to assess the agreement of CBF_ASL and CBF_PCA in two cerebrovascular cohorts: overt stroke and small vessel diseases (SVD), and to (2) assess the repeatability of each technique within a stroke cohort. We observed significant differences between the total CBF estimates produced by each modality, therefore we state that caution is warranted when comparing total CBF between modalities among patients with cerebrovascular disease. Intra-subject repeatability was high, however, for both approaches.

Sudipto Dolui^1,2, Marta Vidorreta¹, Ze Wang^3,4, David A. Wolk¹, and John A. Detre^1,2
1Department of Neurology, University of Pennsylvania, Philadelphia, Pennsylvania, United States, ²Department of Radiology, University of Pennsylvania, Philadelphia, Pennsylvania, United States, ³Hangzhou Normal University, Hangzhou, Zhejiang, China, ⁴Department of Psychiatry and Radiology, University of Pennsylvania, Pennsylvania, United States

We compared ASL MRI acquired from MCI patients and elderly control subjects using PASL and pCASL 2D EPI and pCASL background suppressed (BS) 3D spiral imaging. The methods demonstrate strong correlations between mean CBFs in different ROIs with strongest correlation between measures obtained using either same labeling (pCASL) or imaging strategy (2D). Temporal SNR of 3D BS is higher than the other approaches although its GM-WM contrast is lower than 2D pCASL. Observed control>patient differences in PCC CBF increased with the use of pCASL BS 3D vs pCASL non BS 2D and were not evident in PASL data.

Sudipto Dolui^1,2, Ze Wang^3,4, David A. Wolk¹, and John A. Detre^1,2
1Department of Neurology, University of Pennsylvania, Philadelphia, Pennsylvania, United States, ²Department of Radiology, University of Pennsylvania, Philadelphia, Pennsylvania, United States, ³Hangzhou Normal University, Hangzhou, Zhejiang, China, ⁴Department of Psychiatry and Radiology, University of Pennsylvania, Pennsylvania, United States

The averaging procedure in ASL MRI to overcome low SNR can be undermined by large artifacts present in only a small number of tag-control pairs. We proposed a novel method, named structural correlation based outlier rejection (SCOR), for removing outlier pairs based on i) structural similarity between mean CBF and individual CBF maps and ii) mean GM CBF of individual maps outside physiologically meaningful range. The performance of SCOR is assessed using repeated control scans obtained at 3 months interval from the ADNI database. Compared to alternative options, SCOR demonstrates superior performance by providing much better agreement between the two sessions.

Karsten Sommer^1,2, Dominik Bernat¹, Regine Schmidt¹, and Laura M. Schreiber^1
1Department of Radiology, Johannes Gutenberg University Medical Center, Mainz, Rhineland-Palatinate, Germany, ²Max Planck Graduate Center with the Johannes Gutenberg University Mainz, Mainz, Rhineland-Palatinate, Germany

While blood bolus dispersion has been shown to affect cerebral ASL, this effect has not yet been adressed for myocardial ASL. In this contribution, we employ computational fluid dynamics simulations in a realistic coronary artery model to evaluate the extent of dispersion in myocardial ASL and its impact on resulting myocardial blood flow values.

Jalal B. Andre¹, Greg Wilson¹, Yoshimi Anzai¹, Mahmud Mossa-Basha¹, Michael N. Hoff¹, and Satoshi Minoshima^1
1Radiology, University of Washington, Seattle, WA, United States

We investigate the diagnostic applicability and performance of three-dimensional stereotactic surface projections (3D-SSP), a commonly used tool for molecular brain PET imaging, as applied to arterial spin labeling in clinical MR examinations. Expert reader assessment found that the 3D-SSP method was generally preferred over the traditional method of ASL display. We conclude that 3D-SSP statistical mapping is feasible in a clinical population and enables quantitative data extraction and reliable localization of perfusion abnormalities by means of stereotactic coordinates in a condensed display. This technique may provide added value in the clinical assessment of a variety of neurological pathologies.

Mike-Ely Cohen^1,2, Isabelle Lajoie², Kenneth Dyson², Olivier Lucidarme^1,3, Richard D. Hoge^2,4, and Frédérique Frouin^1,5
1Laboratoire d'imagerie biomedicale, Sorbonne Université Univ Paris 06, Inserm, CNRS, Paris, France, ²Centre de recherche de l’institut universitaire de gériatrie de Montréal, Montréal, Quebec, Canada, ³Service de Radiologie Polyvalente Diagnostique et Oncologique, CHU Pitié-Salpêtrière, AP-HP, Paris, France, ⁴McConnell Brain Imaging Centre, Montreal Neurological Institute - McGill University, Quebec, Canada, ⁵CEA/I2BM/SHFJ, IMIV, Orsay, France

The purpose of this study was to optimize a pseudo-Continuous Arterial Spin Labeling (pCASL) approach to quantify hepatic perfusion among healthy volunteers. Six volunteers were examined using a 3T Siemens scanner, pCASL sequences with 4 to 20 measurements in a plane orthogonal to portal vein with a post label delay (PLD) of 600 ms were acquired. For two subjects, four additional PLD (from 1000 to 1600 ms) were tested. Results show a more robust estimation when using 20 measurements. Furthermore the five PLD values provided hepatic perfusion weighted differently by the hepatic artery and the portal vein blood flows.

Marijn van Stralen¹, Margreet F Sanders², Hanke J Schalkx³, Maurice A van den Bosch³, Clemens Bos⁴, Peter J Blankestijn², Tim Leiner³, and Esben Thade Petersen^3
1Image Sciences Institute, UMC Utrecht, Utrecht, Utrecht, Netherlands, ²Dept of Nephrology, UMC Utrecht, Utrecht, Netherlands, ³Dept of Radiology, UMC Utrecht, Utrecht, Netherlands, ⁴Image Sciences Institute, UMC Utrecht, Utrecht, Netherlands

Renal perfusion imaging using contrast media injection is potentially nephrotoxic. Arterial spin labeling (ASL), employing endogenous contrast, was shown using pulsed and pseudo-continuous labeling strategies. We investigated the feasibility of velocity selective ASL (VS-ASL), which eliminates label planning and possibly improves perfusion SNR by labeling closer to the target tissue. We show that renal VS-ASL is feasible, clearly showing the expected corticomedullary perfusion distribution in healthy volunteers. However, VS-ASL is sensitive to motion, especially during labeling. It could benefit from outlier rejection techniques or prospective and retrospective motion correction.

Yoojin Lee¹ and Tae Kim^1
1Department of Radiology, University of Pittsburgh, Pittsburgh, Pennsylvania, United States

The dynamics of the arterial spin labeling (ASL) provides information on the kinetics of water exchange between tissue and blood, transit time, and the tissue relaxation time. In this study, multi-band technique was applied to look-locker ASL to get the perfusion dynamics for whole brain with a sufficient temporal resolution. We found the hypertensive group has significantly longer capillary transit time and maximal arterial arrival time in most ROIs, while the arterial transit time was not significantly different, compared to the normotensive group. We demonstrated that our technique enables the detection of regional hypertension-induced cerebrovascular impairment in whole brain.

Sau May Wong¹, Eleana Zhang², Harm J. van de Haar¹, Julie E.A. Staals², Cécile R.L.P.N. Jeukens¹, Paul A.M. Hofman¹, Robert J. van Oostenbrugge², Jacobus F.A. Jansen¹, and Walter H. Backes^1
1Radiology, Maastricht University Medical Center, Maastricht, Limburg, Netherlands, ²Neurology, Maastricht University Medical Center, Maastricht, Limburg, Netherlands

Cerebral small vessel disease(cSVD) affects the small vessels in the brain, which can potentially lead to ischemic stroke and dementia. The blood-brain barrier (BBB) permeability is expected to play a pivotal role in the pathophysiology. We performed dual temporal resolution DCE-MRI to quantify BBB permeability. Significant higher permeability was found for cSVD patients compared with controls in the normal appearing brain tissue, which suggests that these regions might be at risk for further transition into white matter hyperintensities. Longitudinal studies can now be initiated to investigate the role of increased BBB permeability in the course of cSVD.

Chun-Xia Li¹, Doty Kempf¹, Leonard Howell¹, and Xiaodong Zhang^1
1Yerkes Imaging Center, Yerkes National Primate Research Center, Emory University, Atlanta, GA, United States

Ketamine and isoflurane are widely used for maintaining general anesthesia and induce a global CBF increase in humans and animals. However, it is not clear how differently the CBF is affected between the anesthetics. In the present study, adult rhesus monkeys were used to examine the effects of ketamine and isoflurane on CBF using the continuous arterial spin-labeling (CASL) perfusion MRI. The present study revealed the different effects of ketamine and isoflurane on CBF in monkeys under maintenance doses, suggesting the use of ketamine or isoflurane might interfere with the experimental outcome in neurovascular diseases and brain function studies.

Chun-Xia Li¹, Sudeep Patel¹, and Xiaodong Zhang^1
1Yerkes Imaging Center, Yerkes National Primate Research Center, Emory University, Atlanta, GA, United States

Isoflurane is popularly used for long-duration anesthesia in clinic practice or preclinical researches. It is known that long-duration anesthesia could cause neurocognitive dysfunction in animal and human. However, the potential mechanism remains unclear. In the present study, the pseudo continuous arterial-spin-labeling (pCASL) technique was used to evaluate the isoflurane anesthesia effect on cerebral blood flow (CBF) of adult rhesus monkeys under maintenance dosage for 4 hours. The results demonstrate that long-duration isoflurane exposure could cause general CBF decrease in most brain areas, which may contribute to the neurocognitive dysfunction observed in clinical or preclinical studies.

Andrew P Hale¹, Charlotte E Buchanan¹, Johannes van Lieshout², Penny A Gowland¹, Paul L Greenhaff³, and Sue T Francis^1
1Sir Peter Mansfield Imaging Centre, University of Nottingham, Nottingham, United Kingdom, ²School of Biomedical Sciences, University of Nottingham, Nottingham, United Kingdom, ³Faculty of Medicine & Health Sciences, University of Nottingham, Nottingham, United Kingdom

Low to moderate intensity dynamic exercise in healthy, young individuals has been directly linked to enhanced cerebral neuronal activity. Here we assess the feasibility of using MRI to measure global and regional changes in CBF to steady-state exercise. We demonstrate that it is possible to assess global and regional brain responses to exercise involving a large muscle mass at workloads of up to 50%VO2max. Moreover, at low intensity workloads (30%VO2max) it appears that there is an uncoupling of regional blood flow from brain activity, suggesting that cardiac output is the primary determinant of regional brain blood flow under these conditions.

Adam Bush¹, Matthew Borzage¹, Thomas Coates¹, and John Wood^1
1Children's Hospital Los Angeles USC, Los Angeles, California, United States

In sickle cell disease (SCD) stroke is common, however accurate quantification of basal cerebral oxygen consumption (CMRO2) is unknown. We measured brain volume, cerebral blood flow (CBF) and cerebral venous saturation (SvO2) in patients with SCD using phase contrast and TRUST MRI to address questions regarding cerebral metabolism. We found patients with SCD to have 167% the CBF of healthy controls, a higher oxygen delivery and a calculated CMRO2. Our study is the first to demonstrate elevated cerebral metabolism in SCD, mirroring increases in global resting energy expenditure and peripheral metabolic rate described by others.

Lirong Yan¹, Songlin Liu¹, Jia Guo², David S Liebeskind¹, Jeffrey L Saver¹, Noriko Salamon³, Neal Yao¹, Sunil Sheth¹, Conrad Liang¹, Eric C Wong², and Danny JJ Wang^1
1Neurology, UCLA, Los Angeles, CA, United States, ²UCSD, San Diego, CA, United States, ³Radiology, UCLA, Los Angeles, CA, United States

In the current study, the feasibility of random vessel-encoded ASL (rVE-ASL) was evaluated by comparison with the standard vessel-encoded ASL (VE-ASL) in healthy subjects as well as in ischemic stroke patients. Our results showed the perfusion territory maps of rVE-ASL were well matched with those of VE-ASL. The collaterals detected by rVE-ASL were consistent with DSA in stroke patients, suggesting the potential clinical utility of rVE-ASL to characterize collateral circulation in stroke.

David Shin¹, Burak Ozyurt¹, Jerod Rasmussen², Juan Bustillo³, Theodorus Van Erp², Jatin Vaidya⁴, Daniel Mathalon⁵, Bryon Mueller⁶, James Voyvodic⁷, Douglas Greve⁸, Judith Ford⁵, Gary Glover⁹, Gregory Brown¹, Steven Potkin², and Thomas Liu^1
1University of California, San Diego, La Jolla, CA, United States, ²University of California, Irvine, Irvine, CA, United States, ³University of New Mexico, Albuquerque, NM, United States, ⁴University of Iowa, Iowa City, IA, United States, ⁵University of California, San Francisco, San Francisco, CA, United States, ⁶University of Minnesota, Twin Cities, Minneapolis, MN, United States, ⁷Duke University, Durham, NC, United States, ⁸Harvard Medical School, Massachusetts General Hospital, Charlestown, MA, United States,⁹Stanford University, Stanford, CA, United States

Resting state CBF maps (n=234) collected from the multisite FBIRN Phase 3 Study (http://www.birncommunity.org) were used as learning features in a support vector machines (SVM) classification technique to evaluate its performance in differentiating Schizophrenic patients (SCZ) from healthy controls (CNT). For feature extraction, clusters of voxels with group differences (SCZ vs. CNT) were first identified (t-test, p<0.01) and the mean CBF values across these clusters were used as the training data. The sensitivity, specificity, and accuracy for the leave-one-out cross validation were found to be 75.9%, 75.5%, and 75.7%, respectively. The results suggest that the CBF map acquired from a 5-minute ASL scan combined with SVM may be a useful complimentary tool for diagnosis of SCZ.

Daniel Aguirre-Reyes^1,2, Juan P. Arab³, Marco Arrese³, Rodrigo Tejos³, Pablo Irarrazaval¹, Cristian Tejos¹, Sergio Uribe⁴, and Marcelo E. Andia^4
1Biomedical Imaging Center - Electrical Engineering Department, Pontificia Universidad Catolica de Chile, Santiago, Region Metropolitana, Chile, ²Computational Sciences and Electronic Department, Universidad Tecnica Particular de Loja, Loja, Loja, Ecuador, ³Gastroenterology Department, School of Medicine, Pontificia Universidad Catolica de Chile, Santiago, Chile, ⁴Radiology Department, School of Medicine, Pontificia Universidad Catolica de Chile, Santiago, Chile

Patients with chronic liver diseases generally have portal hypertension (PH) characterized by an increased liver resistance to blood flow that mainly affect the portal system. In order to determine the severity of PH some invasive methods has been proposed. In this work we present a method to indirect estimate the severity of PH by the quantification of the intrahepatic portal vein blood flow with a non-contrast, flow depend technique (TIR-ASL).