Stephen F. Cauley¹, Obaidah A. Abuhashem², Berkin Bilgic³, Itthi Chatnuntawech³, Julien Cohen-Adad⁴, Kawin Setsompop⁵, Elfar Adalsteinsson^2,5, and Lawrence L. Wald^5,6
1A.A. Martinos Center for Biomedical Imaging, Dept. of Radiology, Massachusetts General Hospital, Charlestown, MA, United States, ²Department of Electrical Engineering and Computer Science, MIT, Cambridge, MA, United States, ³Department of Electrical Engineering and Computer Science, Massachusetts Institute of Technology, Cambridge, MA, United States, ⁴Department of Electrical Engineering, Institute of Biomedical Engineering, Ecole Polytechnique de Montreal, Montreal, QC, Canada, ⁵A.A. Martinos Center for Biomedical Imaging, Dept. of Radiology, MGH, Charlestown, MA, United States, ⁶Harvard-MIT Division of Health Sciences and Technology, MIT, Cambridge, MA, United States

The substantial signal attenuation in DI images for large b-values can affect accurate calculation of orientation distribution functions (odf) and fiber tracks. In addition, the low signal-to-noise (SNR) observed at large b-values hinders the performance of popular denoising methods like the LMMSE estimator and the NLM filter. In this work we demonstrate the benefits of basis smoothing within a low-rank DWI estimation framework. Our method significantly reduces the dependencies on noisy basis vectors while preserving root-mean-square error (RMSE) relative to low-noise data (computed by averaging multiple acquisitions).

Florian Knoll¹, Tuomo Valkonen², Kristian Bredies³, and Rudolf Stollberger^1
1Institute of Medical Engineering, Graz University of Technology, Graz, Austria, ²Department of Applied Mathematics and Theoretical Physics, University of Cambridge, Cambridge, United Kingdom, ³Department of Mathematics and Scientific Computing, University of Graz, Graz, Austria

High resolution diffusion weighted imaging and diffusion tensor imaging with isotropic voxels are desirable for a large number of applications. Unfortunately, acquisitions of such data sets are challenging, due to the notoriously low SNR of DWI. In addition, even with fast sequences like single shot EPI, measurement times often become prohibitively long because of the large number of diffusion encoding directions that have to be acquired. In this work we introduce two higher-order variational denoising approaches to reconstruct DTI data with isotropic voxels from a single average which are based on Total Generalised Variation.

Russell Glenn¹, Ali Tabesh¹, and Jens H. Jensen^1
1Radiology and Radiological Sciences, Medical University of South Carolina, Charleston, SC, United States

Diffusion MRI (dMRI) measurements are positively biased by noise from use of magnitude reconstructed images. The effects of the noise bias increase with decreasing signal-to-noise ratio (SNR), which can be problematic in high resolution dMRI acquisitions. A simple, retrospective noise correction technique is described and a weighted linear least squares fitting algorithm is presented for diffusional kurtosis imaging (DKI). Noisy phantom data is analyzed in DKI datasets with variable SNR, and the results of the noise correction are compared to uncorrected and reference data. Noise correction substantially reduces the bias in kurtosis estimates.

Nikolaos Dikaios¹, Valentin Hamy¹, Shonit Punwani¹, and David Atkinson^1
1Centre for Medical Imaging, University College London, London, Greater London, United Kingdom

The scope of this project was to estimate the noise using an adaptation of the median-absolute-deviation (MAD) in the wavelet domain for the expected noise distribution, and calculate the diffusion coefficient (ADC) with a non linear regression (NR) algorithm that accounts for underlying noise.

Jelle Veraart¹, Jeny Rajan¹, Ronald R. Peeters², Alexander Leemans³, Stefan Sunaert², and Jan Sijbers^1
1Vision Lab, University of Antwerp, Antwerp, Belgium, ²Dept. Radiology, University Hospitals of Leuven, Leuven, Belgium, ³Image Sciences Institute, University Medical Center Utrecht, Utrecht, Netherlands

Diffusion magnetic resonance imaging (dMRI) is widely used to quantify water diffusion in biological tissue. The accuracy of diffusion model-specific measures will be limited if not accounting for the statistical distribution of the magnitude dMRI data. The prior knowledge of the underlying noise is thus of importance when aiming for more accurate dMRI analyses. Many existing techniques to estimate the noise parameters rely, however, on assumptions that are generally not fulfilled in actual dMRI. Therefore, a new approach is presented.

Olaf Dietrich¹, Anno Graser², Martina Karpitschka², Melvin D'Anastasi², and Maximilian F. Reiser^1,2
1Josef Lissner Laboratory for Biomedical Imaging, Institute for Clinical Radiology, Ludwig-Maximilians-University Hospital Munich, Munich, Germany, ²Institute for Clinical Radiology, Ludwig-Maximilians-University Hospital Munich, Munich, Germany

The purpose of this study was to analyze the influence of image noise and intravoxel-incoherent-motion (IVIM) effects on diffusional kurtosis imaging (DKI) outside the brain. DKI was performed in the prostate (b_max=2000s/mm²) and data was analyzed with and without fit of the noise level and with and without suppression of IVIM effects. Kurtosis values were substantially increased due to noise and also due to IVIM effects at low b-values. Including the image noise level into the kurtosis analysis and choosing a minimum b-value of 200s/mm² is recommend for DKI outside the brain.

Roland Bammer¹ and Rafael O'Halloran^1
1Center for Quantitative Neuroimaging, Department of Radiology, Stanford University, Stanford, CA, United States

Diffusion-weighted SSFP (dwSSFP) offers an efficient diffusion preparation and is a promising approach to high-resolution 3D DW MRI. As a multi-shot technique dwSSFP is affected by shot-to-shot phase inconsistencies as all multi-shot diffusion-weighted sequences are. However, because the steady state signal is composed of the constructive addition of many phase coherence pathways, motion-induced phase accrued during the application of a single diffusion-encoding gradient can propagate to subsequent phase pathways. This represents an additional mechanism of signal loss in dwSSFP. In this work, these errors are classified into two types and strategies for correcting each type are discussed.

Eric Aboussouan¹, Anh Tu Van¹, Rafael O’Halloran¹, Samantha J. Holdsworth¹, Marcus T. Alley¹, Murat Aksoy¹, and Roland Bammer^1
1Radiology, Stanford University, Stanford, California, United States

Synopsis: High-resolution diffusion-weighted imaging is limited to multi-shot acquisitions, which are subject to signal decay due motion-induced phase variations. These variations are caused by rigid-body (non-repeatable) and pulsatile (repeatable over the cardiac cycle) motion during the diffusion-encoding period. It is possible to prospectively correct this phase before the creation of the spurious echo pathways in RF-refocused sequences (e.g. FSE, SSFP). While linear and constant phase errors can be corrected with gradient blips and the RF phase, the non-linear phase component should be compensated by the application of an RF pulse designed to remove that particular phase profile. In this work we will investigate methods to estimate the non-linear component of the 3D motion-induced phase and compensate it using a time-efficient 3D RF pulse design.

Yen-Wei Cheng¹, Ming-Chung Chou², Wen-Yih Isaac Tseng³, and Hsiao-Wen Chung^4
1Department of Electrical Engineering, National Ilan University, IIran, Taiwan, ²Department of Medical Imaging and Radiological Sciences, Kaohsiung Medical University, Kaohsiung, Taiwan, ³Center for Optoelectronic Biomedicine, National Taiwan University, Taipei, Taiwan, ⁴Department of Electrical Engineering, National Taiwan University, Taipei, Taiwan

The purpose of this study was to simulate the randomly corrupted signals and evaluate the resultant ODF errors as well as error reduction rates by using a neighboring interpolation correcting method. The simulation was performed by generating corrupted signals in the gold standard QBI with random loss of signal intensity as well as random occurrence of pixel location. We found that the reduction rates of ODF errors by neighboring interpolation method were 82.62%, 79.32%, 77.67% for b = 2000, 3000, 4000 s/mm2 QBI datasets, so it suggests the proposed NI method is helpful to correct signal loss artifacts in QBI.

Ming-Chung Chou¹, Wan-Hsin Wen¹, Sheng-Fang Huang¹, Cheng-Wen Ko², and Ping-Hong Lai^3
1Department of Medical Imaging and Radiological Sciences, Kaohsiung Medical University, Kaohsiung, Taiwan, ²Department of Computer Science and Engineering, National Sun Yat-sen University, Kaohsiung, Taiwan,³Department of Radiology, Kaohsiung Veterans General Hospital, Kaohsiung, Taiwan

Diffusion kurtosis imaging (DKI) was demonstrated to successfully estimate the non-Gaussian distribution of water diffusion in vivo and has been applied in many clinical applications. Similar to diffusion tensor imaging, b-value is an important factor in the measurement of DKI indices and may have influential effects on them. Rapid DKI with three b-values was proposed to obtain DKI indices in clinically acceptable time, but the results may be affected by the choice of b-value. This study performed repeated measurement to investigate the effects of b-value on the reproducibility and accuracy of DKI indices and found that both reproducibility and accuracy were impacted by b-values.

Javier Urriola Yaksic¹, Nyoman Dana Kurniawan¹, Zhengyi Yang², and David C. Reutens^1
1Centre for Advanced Imaging, The University of Queensland, Brisbane, Queensland, Australia, ²Information Technology and Electrical Engineering, The University of Queensland, Brisbane, Queensland, Australia

Track density imaging (TDI) mapping method has been developed to dramatically increase the spatial resolution of diffusion-weighted imaging (DWI) data beyond the acquired resolution. However, it is not clear if TDI maps can be used to reliably to quantify differences in brain populations. This project aims to characterise short track TDI (stTDI) in terms of stability and reproducibility as a quantitative tool for group comparison of brain structures. Our results showed that stTDI produced more consistent profile in the white matter compared to the gray matter. Multiple stTDI reconstructions of a single dataset need to be generated then averaged to minimise the noise resulting from probabilistic fibertracking. Limited comparison between intra- and inter-subject variability suggests that stTDI could be developed as a method to measure differences in brain structures between two populations.

Alex Smith¹, Madhura Ingalhalikar¹, and Ragini Verma^1
1Section for Biomedical Image Analysis, University of Pennsylvania, Philadelphia, PA, United States

This study assesses the reliability of graph metrics extracted from structural networks constructed via probabilistic tractography, and the sensitivity of such networks to individual subjects. Both weighted and binarized networks were evaluated. It was found that subjects could be differentiated based on their weighted structural connectivity patterns, suggesting that such networks can be used in population studies. The intraclass correlation coefficients (ICCs) of the assessed metrics tended to be moderately high across network types, indicating longitudinal robustness across scan sessions. The ICCs tended to be lower in the weighted networks, possibly due to their decreased uniformity versus the binarized versions.

Jérémy Lecoeur¹, Madhura Ingalhalikar¹, and Ragini Verma^1
1Section of Biomedical Image Analysis, Department of Radiology, University of Pennsylvania, Philadelphia, PA, United States

This study investigates the stability of cortical connectivity based parcellation over time. Such parcellations provide a unique characterization of the brain and it is believed that any change can be interpreted as being introduced by pathology or development. Hence a parcellation profile for a person could be defined using this paradigm and that constitutes a first step in quantifying differences between subjects, facilitating a meaningful clinical study. We demonstrate that such parcellation for the primary visual cortex is replicable over time with high similarity scores between the obtained clusters with an unsupervised algorithm.

David Alexander Slater¹, Po-Wah So², Karthik Munikoti Prakash², Istvan Bodi³, Michel Modo^4,5, and Flavio Dell'Acqua^1,6
1NATBRAINLAB, Department of Neuroimaging, King's College London, Institute of Psychiatry, London, United Kingdom, ²Department of Neuroimaging, King's College London, Institute of Psychiatry, London, United Kingdom, ³Department of Clinical Neuropathology, King's College Hospital, London, United Kingdom, ⁴McGowan Institute for Regenerative Medicine, University of Pittsburgh, Pittsburgh, Pennsylvania, United States,⁵Department of Radiology, University of Pittsburgh, Pittsburgh, Pennsylvania, United States, ⁶NIHR Biomedical Research Centre for Mental Health at South London and Maudsley NHS Foundation Trust, King's College London, Institute of Psychiatry, London, United Kingdom

Diffusion imaging of fixed human brain tissue has the potential to reveal neuroanatomical details at a scale that remains largely unexplored. Here we investigated how the optimisation of tissue sample preparation and imaging parameters can increase data quality. Our findings suggest that phosphate-buffered saline (PBS) and specific concentrations of the contrast agent GD-DTPA have the potential to dramatically increase SNR-efficiency at short TE and TR. However, longer diffusion times also provide improved diffusion contrast and thus a balance is required between potential increases in SNR-efficiency and diffusion contrast. These findings may provide improved SNR, enhanced resolution and/or faster acquisitions.

Shogo Fujii¹, Etsuji Yamamoto¹, Yo Taniguchi², and Yoshitaka Bito^1,2
1Graduate School of Engineering, Chiba University, Chiba-shi, Japan, ²Central Research Laboratory, Hitachi, Ltd., Kokubunji-shi, Japan

The purpose of this study is to develop a DWI simulator, which can generate the images for the adult human brain size models in a reasonable time. In order to shorten the computation time, we proposed three approaches based on the principle of MRI to increase simulation efficiency and succeded in improving the efficiency approximately 140,000 times. As a result, simulation time for an adult human brain size model reduced to less than one hour.

Tim Sprenger^1,2, Brice Fernandez³, Jonathan I. Sperl¹, Vladimir Golkov^1,2, Michael Bach⁴, Ek T. Tan⁵, Kevin F. King⁶, Christopher J. Hardy⁵, Luca Marinelli⁵, Michael Czisch⁷, Philipp Sämann⁷, Axel Haase⁸, and Marion I. Menzel^1
1GE Global Research, Garching, Germany, ²Technical University Munich, Garching, Germany, ³GE Healthcare, Munich, Germany, ⁴German Cancer Research Center, Heidelberg, Germany, ⁵GE Global Research, Niskayuna, NY, United States, ⁶GE Healthcare, Waukesha, WI, United States, ⁷Max Planck Institute for Psychiatry, Munich, Germany, ⁸Technische Universität München, Garching, Germany

Conventional diffusion spectrum imaging (DSI) requires Nyquist or full-sampling of q-space, and hence suffers from long acquisition times. To overcome this limitation, compressed-sensing-accelerated DSI has been proposed recently. For this technique q-space is randomly undersampled and subsequently reconstructed exploiting the signal sparsity in an appropriate transform domain. This work studies the SNR-dependent performance of compressed sensing-accelerated DSI using a fiber crossing phantom and by that validates the superior signal recovery and denoising properties.

Stephan E. Maier¹ and Robert V. Mulkern^2
1Department of Radiology, Harvard Medical School, Brigham and Women's Hospital, Boston, MA, United States, ²Department of Radiology, Harvard Medical School, Children's Hospital, Boston, MA, United States

The egg presents a simple but relevant object for exploring biophysical aspects of common tissue contrast parameters like diffusion and transverse relaxation. The biexponential characterization of the white yolk containing latebra diffusion signal decay results in diffusion coefficients and signal fractions that are very similar to those found in brain. White yolk contains spheres with membranes. The size of the spheres observed in white yolk falls in the range of cell sizes typically encountered in tissues. The avian egg latebra, is thus an easily accessible model for tissue water diffusion

Jonathan Phillips^1,2 and Geoffrey David Charles-Edwards^3
1Medical Engineering and Physics, King's College London, 124-126 Denmark Hill, London, United Kingdom, ²Institute of Life Science 2, Swansea University, Swansea, Wales, United Kingdom, ³Medical Physics, St. Thomas' Hospital, Westminster Bridge Road, London, United Kingdom

Diffusion kurtosis imaging (DKI) is an extension of conventional diffusion-weighted MRI which fits a model to diffusion b-value images from which a value of diffusion kurtosis K, the degree to which diffusion is non-Gaussian, is obtained. We present a simple and robust DKI test object based upon colloidal spheres dispersed in water. A linear relationship between concentration of spheres and kurtosis is found. DKI measurements over a period of time showed excellent reproducibility indicating the suitability of these colloidal spheres as a long term test object for DKI.

Francesco Grussu¹, Torben Schneider¹, Hugh Kearney¹, Hui Zhang², David H. Miller¹, Olga Ciccarelli³, Daniel C. Alexander², and Claudia Angela M. Wheeler-Kingshott^1
1NMR Research Unit, Queen Square MS Centre, Department of Neuroinflammation, UCL Institute of Neurology, London, England, United Kingdom, ²Department of Computer Science and Centre for Medical Image Computing, University College London, London, England, United Kingdom, ³NMR Research Unit, Queen Square MS Centre, Department of Brain Repair and Rehabilitation, UCL Institute of Neurology, London, England, United Kingdom

We investigate the feasibility of Neurite Orientation Dispersion Imaging (NODDI) in the spinal cord. NODDI is a new technique that promises novel markers of neuronal density and dispersion, which may be informative in diseases like Multiple Sclerosis (MS) beyond routine Diffusion Tensor Imaging (DTI). Here we compare NODDI with standard DTI metrics in healthy volunteers and MS patients. NODDI disentangles the microscopic sources of DTI measures and provides better discrimination between very coherent and less coherent neural tissue structures. Further work will improve the acquisition protocol and study NODDI in a larger patient cohort.

Maria F. Falangola^1,2, David Guilfoyle³, Edward S. Hui¹, Xingju Nie¹, Ali Tabesh¹, Jens H. Jensen¹, Scott Gerum³, Caixia Hu³, John LaFrancois³, Heather Collins¹, and Joseph A. Helpern^1,2
1Radiology and Radiological Science, Center for Biomedical Imaging, Medical University of South Carolina, Charleston, SC, United States, ²Neurosciences, Medical University of South Carolina, Charleston, SC, United States, ³Nathan Kline Institute, Orangeburg, NY, United States

The goal of this study was to utilize a recently developed white matter modeling method compatible with diffusional kurtosis imaging to characterize the demyelination process in the corpus callosum of the cuprizone mouse model. We studied 10 cuprizone and 10 control mice for a period of 10 weeks. All diffusion metrics were derived from DKI datasets. Quantitative analysis of myelin density was performed using Solochrome staining. The histological correspondence shown in our results demonstrate the correspondence between the WMM metrics and a well established histological marker of myelin.

Lindsey A. Leigland¹, Marcus Cappiello², and Christopher D. Kroenke^1,3
1Advanced Imaging Research Center and Department of Behavioral Neuroscience, Oregon Health & Science University, Portland, OR, United States, ²Department of Physics, Oregon State University, Corvallis, OR, United States, ³Division of Neuroscience, Oregon National Primate Research Center, Oregon Health & Science University, Beaverton, OR, United States

The goal of this research was to improve understanding of the determinants of physical properties (spin relaxation, diffusion, etc.) in postmortem tissue, as well as further characterize the dependence of these molecular properties on common laboratory tissue processing procedures. Here we show that exposure of tissue to alkaline conditions can induce dramatic (>2-fold) and irreversible changes in the water apparent diffusion coefficient (ADC) without affecting the water 1H transverse relaxation, an effect that is potentially realized through modifications to biological membranes. This research provides an avenue toward developing further insight into the determinants of physical parameters observable by MRI.

Seongjin Choi¹, Ria Mazumder², Petra Schmalbrock¹, Michael V. Knopp¹, Richard D. White^1,3, and Arunark Kolipaka^1,3
1Radiology, The Ohio State University, Columbus, Ohio, United States, ²Electrical and Computer Engineering, The Ohio State University, Columbus, Ohio, United States, ³Internal Medicine, The Ohio State University, Columbus, Ohio, United States

Diffusion-based tractography has the potential to delineate complex cardiac muscle fiber architecture. We explored fractional anisotropy distribution, angle and length threshold dependency of tractography in a fixed heart specimen. Tracking method used in the brain research was applied straightforwardly. Among three tracking constraints, seeding FA value and minimum fiber length threshold noticeably affected the tractography results. z-component of principal eigenvectors was useful index for subepi/subendocardium segmentation. Three long fiber bundles of interest and subepi/subendocardium were segmented as a demonstration of this approach.

Constantin von Deuster^1,2, Christian T. Stoeck², Daniel Giese¹, Jack Harmer¹, Rachel W. Chan³, David Atkinson³, and Sebastian Kozerke^1,2
1Division of Imaging Sciences, King's College London, London, United Kingdom, ²Institute for Biomedical Engineering, ETH Zurich, Zurich, Switzerland, ³Centre for Medical Imaging, University College London, London, United Kingdom

Diffusion tensor imaging (DTI) of the beating human heart became feasible in the last years, employing Stimulated Echo (STEAM) sequences for diffusion acquisition. Several signal averages and diffusion encoding directions make the image formation process however fundamentally slow. Hence acquisitions for 3D whole heart fibre configurations are not clinically applicable as several slices are necessary. We could show that concurrent dual-slice single-shot EPI cardiac diffusion weighted imaging using multi-band STEAM and controlled aliasing in parallel imaging can be used to speed up scan speed by a factor of two.

Ria Mazumder¹, Seongjin Choi², Bradley Dean Clymer¹, Richard White^2,3, and Arunark Kolipaka^2,3
1Department of Electrical and Computer Engineering, The Ohio State University, Columbus, OH, United States, ²Deparment of Radiology, The Ohio State University, Columbus, OH, United States, ³Deptarment of Internal Medicine, Division of Cardiology, The Ohio State University, Columbus, OH, United States

Diffusion tensor imaging (DTI) is a technique used for fiber tracking. It has been rarely explored to track cardiac fibers .Previous studies have used formalin fixed heart for tracking fibers. However, formalin fixation tends to change the diffusion properties of tissues. This study explores the effects formalin fixation has on fiber tracking when compared to fresh hearts. DTI analysis parameter such as fractional anisotropy was further exploited to track the fibers in fresh and formalin fixed hearts. We have observed that formalin fixation makes diffusion pattern of the cardiac muscle fibers more isotropic and thereby lowers the FA value.

Eddy Solomon¹, Noam Nissan^1,2, Hadassa Degani³, and Lucio Frydman^1
1Chemical Physics, Weizmann Institute of Science, Rehovot, Israel, ²Biological Regulation, Weizmann Institute of Science, Rehovot, Israel, ³Biological Regulation, Weizmann Institue of Science, Rehovot, Israel

Breast imaging is particularly challenging owing to a number of factors, including the relatively high fat/water heterogeneities that characterize breast, the off-center character of the targeted organs, and the unavoidability of breathing/cardiac motions. Here we exploit SPEN’s abilities to deliver quality single-scan images even from challenging regions subject to all these problems, to explore its use in DW MRI. Diffusion SPEN measurements were carried out at 3T in female breast analyses, showing significant advantages in anatomical and diffusional information vis-à-vis EPI.

Koji Fujimoto¹, Aki Kido², Tomohisa Okada², Masato Uchikoshi³, and Kaori Togashi^2
1Department of Diagnostic Imaging and Nuclear Medicine, Kyoto University, Kyoto, Japan, ²Department of Diagnostic Imaging and Nuclear Medicine, Kyoto University Graduate School of Medicine, Kyoto, Japan,³Siemens Japan K.K., Tokyo, Japan

Uterine DTI was performed for nine healthy young women on a 3T scanner. VOIs were drawn for the outer myometrium (OM), junctional zone (JZ) and endometrium (EM). ADC, FA and maximum fiber length were compared. Fibers were classified into 12 groups and visually evaluated. ADC was highest for OM (1.12), followed by EM (0.97) and JZ (0.83). FA was highest for JZ (0.297), followed by OM (0.257) and EM (0.186). Fibers were longest in OM (42.0mm), followed by JZ (34.2mm) and EM (20.0mm). Circular orientation was observed in 50% (36/72) of fibers in OM and in 44% (32/72) in JZ.

Franz Schilling^1,2, Jan Henrik Ardenkjær-Larsen³, Martin A. Janich², Marion I. Menzel², and Lise Vejby Søgaard^3
1Institute of Medical Engineering, Technische Universität München, Garching, Germany, ²GE Global Research, Garching, Germany, ³Danish Research Centre for Magnetic Resonance, Copenhagen University Hospital Hvidovre, Hvidovre, Denmark

We suggest a new approach to characterize the metabolism by performing apparent diffusion coefficient (ADC) measurements of hyperpolarized metabolites in-vivo. A diffusion-weighted pulse sequence was developed based on the pulsed gradient spin echo with low flip angle (α = 20˚) slice selective excitation in combination with a band-selective universal-rotation refocusing pulse optimized by optimal control theory. In two rats, we acquired ¹³C-metabolite ADCs in 3 spatial directions from lactate, alanine and pyruvate in muscle tissue at 4 different positions. The analysis of diffusion properties of intracellularly produced hyperpolarized compounds might potentially be useful for characterizing pathological changes in tumor cells.

David Garbera¹, Jyoti Parikh², and Geoffrey David Charles-Edwards^3
1Imaging Sciences, King's College London, London, United Kingdom, ²Radiology, Guy's and St Thomas' NHS Foundation Trust, London, United Kingdom, ³Medical Physics, Guy's and St Thomas' NHS Foundation Trust, London, United Kingdom

We analyse use of diffusion-weighted MRI (DWI) scanning, and generation of apparent diffusion coefficient (ADC) maps, in predicting early response to neoadjuvant chemotherapy (NACT) in patients with breast cancer. It is thought that perfusion effects may confound accuracy of ADC values calculated from DWI. We gathered mean ADC values from pre and mid-treatment scans by generating a region of interest around a whole lesion. Early changes in mean ADC value were correlated with eventual radiological response to NACT. Early change in mean ADC value performed better than long-axis measurements in predicting response. Only ADC maps excluding perfusion were significantly better.

Eleftheria Panagiotaki¹, Simon Walker-Samuel², Bernard M. Siow^1,2, Peter Johnson³, Rosamund Barbara Pedley³, Mark F. Lythgoe², and Daniel C. Alexander^4
1Centre for Medical Image Computing, University College London, London, United Kingdom, ²Centre for Advanced Biomedical Imaging, University College London, London, United Kingdom, ³Cancer Institute, University College London, London, United Kingdom, ⁴Centre for Medical Image Computing, Dept Computer Science, University College London, London, United Kingdom

This work explores the diffusion weighted (DW) MR signal of cancer cell lines with a selection of multi-compartment diffusion models, adapting a method for diffusion model identification in brain tissue . Characterisation of colorectal cancer cells (CRC) can improve the response of the cells to treatment. Most previous cancer cell classification studies involve invasive histological methods that alter the state of the sample. This study uses DW-MRI to quantitatively characterise CRC cell lines in vivo based on their microstructure.

Karen Bae¹, Daniel Ying¹, David Yerushalmi¹, Alice Kim¹, and Jinha Park^1
1Diagnostic Radiology, City of Hope, Duarte, CA, United States

Diffusion-weighted magnetic resonance imaging (DW-MRI) has been suggested as a promising, non-invasive, yet quantitative tool to detect lesions. However, global implementation of DW-MRI as a means of assessing cancer faces challenges due to lack of quality assessments and no accepted standards for measurements and analysis. The purpose of this study was to address this issue by assessing a recently implemented liver DWI protocol at our institution - City of Hope National Medical Center. In this retrospective study, we investigated the apparent diffusion coefficient values of liver lesions and found a statistical difference between benign and malignant lesions (p<0.0001). These results were consistent with previously published reports and demonstrate the clinical benefit of adopting DWI as a standard sequence in a liver MRI protocol.

Xiawei Ou¹, Charles M. Glasier¹, Raghu H. Ramakrishnaiah¹, Sarah B. Mulkey², Vivien L. Yap², and Jeffrey R. Kaiser^2
1Radiology, Arkansas Children's Hospital; University of Arkansas for Medical Sciences, Little Rock, AR, United States, ²Pediatrics, Arkansas Children's Hospital; University of Arkansas for Medical Sciences, Little Rock, AR, United States

Cerebral white matter injury is very common in extremely low birth-weight (ELBW) infant survivors of intensive care and is associated with poor long-term neurological outcomes. In this study, we used DTI in addition to conventional MRI to evaluate two groups of ELBW infants: one was randomized to permissive hypercapnia ventilation during the first week of life, and the other was randomized to normocapnic ventilation. We found that while the average white matter score from conventional MRI for the hypercapnic group was not significantly different with that in the normocapnic group, DTI revealed lower FA values and higher MD values in the genu and splenium of hypercapnic subjects, suggesting that hypercapnia may have unfavorable effects on the white matter development of ELBW infants, and DTI may be more sensitive to detect these effects than conventional MRI.

Georgia Lockwood Estrin^1,2, Zhiqing Wu², Serena J. Counsell³, Mary A. Rutherford¹, and Joseph V. Hajnal^2
1Perinatal Imaging, Imperial College London, London, United Kingdom, ²Centre for the Developing Brain, Kings College London, London, United Kingdom, ³Centre for the Developing Brain, King's College London, London, United Kingdom

Fetal DTI offers the potential to provide detailed information on normal white matter development. A major limitation is movement, which is often not addressed in fetal MR studies. This study employs a fully motion corrected reconstruction method including distortion correction using a measured B0 field map to measure fractional anisotropy (FA) in 19 fetuses. Results were compared with ex utero preterm FA data of similar post-menstrual ages. The fetal FA values were consistent with the preterm neonates, including trends with age. This provides confidence that the method is reliable and offers the opportunity to study white matter development in utero.

Arzu Ceylan Has¹, Aslihan Taskiran Sag², Cagri Mesut Temucin³, Nese Oztekin², Fikri Ak², and Kader Karli Oguz^4
1National Magnetic Resonance Research Centre (UMRAM), Bilkent University, ANKARA, Turkey, ²Department of neurology, Ankara Numune Education and Research Hospital, ANKARA, Turkey, ³Department of Neurology, Faculty of Medicine, and Institute of Neurological Sciences and Psychiatry, Hacettepe University, ANKARA, Turkey, ⁴Department of Radiology and National Magnetic Resonance Research Centre (UMRAM), Hacettepe University and Bilkent University, ANKARA, Turkey

Neuropathic pain is caused by a lesion or a disease of the somatosensory system itself. It’s usually more prominent in acquired pathologies of nerve fibers like diabetic neuropathy (DNP) despite less severe degeneration than hereditary neuropathies (HNP). On the basis of differences in etiopathogenetic mechanisms, we hypothesized that grey and white matter alterations would reveal distinct patterns in patients with DNP and HNP.

Ido Tavor¹, Rotem Botvinik¹, and Yaniv Assaf^1
1Department of Neurobiology, Tel Aviv University, Tel Aviv, Israel

Diffusion MRI was recently shown to be sensitive to short term neuroplasticity in spatial learning. While spatial learning is expected to cause structural changes in limbic structures, the neuroplasticity aspects in higher cognitive domains are not straightforward. Thus in the current study we used DTI in order to investigate the neuroplasticity that accompanies motor learning. Non-musician subjects were scanned before and after a motor sequence learning task based on an electric piano keyboard. We found learning related changes in diffusion properties in motor system regions, suggesting that DTI can follow on short-term learning-related brain plasticity of an entire cognitive domain.

Khin Khin Tha¹, Satoshi Terae², Hiroyuki Hamaguchi², Kinya Ishizaka², Makoto Hirotani², Kazuyuki Minowa³, Yuriko Suzuki⁴, and Hiroki Shirato^1
1Department of Radiobiology and Medical Engineering, Hokkaido University Graduate School of Medicine, Sapporo, Hokkaido, Japan, ²Hokkaido University Hospital, Sapporo, Hokkaido, Japan, ³Hokkaido University Graduate School of Dental Medicine, Sapporo, Hokkaido, Japan, ⁴Philips electronics Japan, Ltd., Tokyo, Japan

Laterality in the DTI indices of the corticospinal tract (CST) of the brain and the spinal cord and their variation with handedness were evaluated in 8 left handers and 8 age- and gender-matched right handers. The DTI indices were compared across the two sides and between the two handers, at various CST levels. The results revealed a similar pattern of asymmetry in the indices between the brain and the spinal cord for both handers. The values differed between the two handers. The knowledge of laterality of the DTI indices and variation with handedness is important in interpretation of the indices.

Tetsuo Sato¹, Nobuyuki Maruyama², Toru Hoshida², and Kotaro Minato^1
1Nara Institute of Science and Technology, Ikoma, Nara, Japan, ²National Hospital Organization Nara Medical Center, Nara, Nara, Japan

In this research, we have calculated the correlation coefficient between the DTI parameter of uncinate fasciculus, fornix and WMS-R score. Correlations between DTI measures and memory performance suggest the relationships between the UF, fornix and function in memory tasks lateralization. Our finding matches previous reports on the correlation between FA in the left or L1 in the right UF and performance on visual memory.

Anja Maria Marschar¹, Tristan Anselm Kuder¹, Bram Stieltjes², and Frederik B. Laun^1,2
1Department of Medical Physics in Radiology, German Cancer Research Center (DKFZ), Heidelberg, Germany, ²Quantitative Imaging Based Disease Characterization, German Cancer Research Center (DKFZ), Heidelberg, Germany

The diffusional kurtosis (K_app) and the apparent diffusion coefficient (D_app) were measured in the human calf muscle. Maps of D_app and K_app are shown and the time dependence of both values is reported. A dependence on the diffusion direction is visible for both values, but a higher anisotropy can be observed for D_app than for K_app. The kurtosis decreases at longer diffusion time. Absolute values of K_app in the muscle tissue are much smaller than in white matter tissue in the brain.

Devon Middleton¹, Feroze Mohamed¹, Nadia Barakat², Scott Faro¹, Pallav Shah¹, MJ Mulcahey², Amer Samdani², and Jürgen Finsterbusch^3
1Radiology, Temple University, Philadelphia, PA, United States, ²Shriners Hospital for Children, Philadelphia, PA, United States, ³University Medical Center Hamburg-Eppendorf, Hamburg, Germany

DTI of the pediatric spinal cord (SC) has the potential to provide useful information on white matter integrity and to serve as an important biomarker for spinal cord injury (SCI). This study examines the evaluate the efficacy of an inner-field-of-view DTI sequence for imaging of the thoracic spinal cord in pediatric subjects, both normal and with SCI. Good quality DTI images were successfully collected for six healthy subjects and two subjects with SCI and DTI indices were examined for both groups.

Christopher D.J. Sinclair^1,2, Laura Mancini^2,3, and John S. Thornton^1,2
1MRC Centre for Neuromuscular Diseases, UCL Institute of Neurology, London, London, United Kingdom, ²Department of Brain Repair and Rehabilitation, UCL Institute of Neurology, London, London, United Kingdom,³Lysholm Department of Neuroradiology, National Hospital for Neurology and Neurosurgery, London, London, United Kingdom

The sciatic nerve is conveys motor and sensory signals in the lower limbs and abnormalities can include focal tumours or demyelination in peripheral neuropathies. We performed high-resolution anatomical imaging and diffusion tensor imaging (DTI) of the sciatic nerve in healthy subjects at 3T in order to visualize the course of the nerve and derive metrics such as the fractional anisotropy (FA). The mean FA of the nerve was 0.62±0.04, substantially higher than the surrounding muscle (0.22±0.02). FA maps provided good depiction of the course of the nerve and any observed variations may be valuable in monitoring conditions such as Charcot-Marie-Tooth disease.

Marco Reisert¹, Matthias Weigel¹, Els Fieremans², Valerij G. Kiselev¹, and Dmitry S. Novikov^3
1Dpt. of Radiology, Medical Physics, University Medical Center Freiburg, Freiburg, BW, Germany, ²Center for Biomedical Imaging, Dpt. of Radiology, New York University, New York, NY, United States, ³Center for Biomedical Imaging, Dpt. of Radiology, New York University School of Medicine, New York, NY, United States

As the inverse problem of quantifying tissue structure in every voxel is highly ill-posed, we regularize it by unifying sub-voxel modeling of dMRI signal at the mesoscopic scale with multi-voxel connectivity achieved with global fiber tracking. When MesoFT converges, we obtain the physically motivated fiber directions, connections, voxel-wise neurite densities, and, in principle, can incorporate other mesoscopic structural parameters. In particular, MesoFT yields strong anisotropy of extra-axonal diffusion and a notable tortuosity in corpus callosum.

Jian Cheng¹, Dinggang Shen¹, and Pew-Thian Yap^1
1Department of Radiology and Biomedical Research Imaging Center (BRIC), The University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, United States

In diffusion MRI, Spherical Deconvolution (SD) was proposed to estimate the fiber Orientation Distribution Function (fODF) based on spherical deconvolution using a single-fiber response function. The peaks or the shape of fODFs can be used to infer local fiber directions. Constrained Spherical Deconvolution (CSD), which takes into consideration the non-negative of the fODF, is the most widely used method among SD variants. Although CSD is capable of accurately determining fiber directions, it is susceptible to false positive peaks especially in the regions with low anisotropy. This is a common drawback of all existing SD-based methods. Moreover, in practice the fODF estimated using CSD still has significant negative values. We propose a method called Non-Negative Spherical Deconvolution (NNSD) to solve the above two problems. Based on a Riemannian framework of ODFs and Square Root Parameterized Estimation for non-negative definite Ensemble Average Propagator, NNSD is formulated such that the non-negativity of the fODF is guaranteed with largely reduced false positive peaks.The synthetic data and real data experiments demonstrated the improvement of NNSD over CSD.

Getaneh Bayu Tefera¹, Yuxiang Zhou¹, and Ponnada A. Narayana^1
1Diagnostic & Interventional Imaging, UT Houston, Houston, Texas, United States

High angular resolution diffusion imaging (HARDI) data was analyzed using 4th and 6th order tensors. The orientation diffusion function (ODF) was deconvolved using spherical basis set with order 4 and 6. The performance of this method was demonstrated by performing tractography around the centrum semiovale regions and corpus callosum. Our results indicate that sixth order tensor, convolved with six basis functions is superior to the fourth order tensor in resolving multiple crossing fibers

Kai-kai Shen¹, Stephen Rose¹, Jurgen Fripp¹, Katie McMachon², Greig de Zubicaray³, Nicholas Martin⁴, Paul Thompson⁵, Margret Wright⁴, and Olivier Salvado^1
1Australian eHealth Research Centre, CSIRO, Herston, Queensland, Australia, ²Centre for Advanced Imaging, University of Queensland, Brisbane, Queensland, Australia, ³School of Psychology, University of Queensland, Brisbane, Queensland, Australia, ⁴Queensland Institute of Medical Research, Brisbane, Queensland, Australia, ⁵Imaging Genetics Center, Laboratory of Neuro Imaging, UCLA, Los Angeles, CA, United States

We evaluated the test-retest reliability for measures based on the fibre orientation distribution (FOD) using constrained spherical deconvolution (CSD). We derived measurements in the peak direction identified from the FOD, and evaluated the test-retest reliability for these FOD peak amplitudes using the intra-class correlation (ICC) coefficient. We visualized the ICC coefficients on the FOD orientation plot. We compared the reliability of measuring the FOD peak amplitudes and the fractional anisotropy (FA) derived from the diffusion tensor imaging (DTI) model.

Qiuyun Fan^1,2, Erika Spangler³, Xin Hong⁴, Ha-Kyu Jeong⁵, Nicole Davis^2,6, Laurie E. Cutting^2,3, and Adam W. Anderson^1,2
1Department of Biomedical Engineering, Vanderbilt University, Nashville, TN, United States, ²Vanderbilt University Institute of Imaging Science, Nashville, TN, United States, ³Department of Special Education, Peabody College of Education and Human Development, Nashville, TN, United States, ⁴Singapore Bioimaging Consortium, Singapore, Singapore, ⁵Philips Healthcare, Korea Basic Science Institute, Cheongwon, Korea, ⁶Department of Radiology and Radiological Sciences, Vanderbilt University, Nashville, TN, United States

Diffusion tensor imaging (DTI) provides valuable information about neuronal tissues, such as fractional anisotropy (FA), but is limited to single-fiber populations. High angular resolution diffusion imaging (HARDI) methods have been developed to reveal complex white matter structures. The conventional spherical deconvolution approaches do not allow for estimates of fiber-specific response kernels in the case of crossing fibers. The multiple kernel spherical deconvolution (MKSD) method can resolve the orientations of multiple fiber populations and provide estimates of the diffusion properties intrinsic to each fiber. In this work, we developed fiber tracking algorithms based on MKSD fiber orientation distribution functions and studied the stability of the fiber-specific FA estimates obtained.

Michael Paquette¹ and Maxime Descoteaux^2
1SCIL, Universite de Sherbrooke, Sherbrooke, Quebec, Canada, ²SCIL, Université de Sherbrooke, Sherbrooke, Quebec, Canada

The q-space truncation induce in the classical acquisition scheme of Diffusion Spectrum Imaging leads to a modified version of the diffusion propagator. That modification can be inverted by deconvolution methods. This work present a separable deconvolution approximation to get better propagator. Those better propagators leads to more physically meaningful diffusion space metric and sharper orientation distribution functions.

Stamatios N. Sotiropoulos¹, Saad Jbabdi¹, Jesper L. Andersson¹, Mark W. Woolrich^1,2, Kamil Ugurbil³, and Timothy E.J. Behrens^1
1FMRIB Centre, University of Oxford, Oxford, United Kingdom, ²Oxford Centre for Human Brain Activity (OHBA), University of Oxford, Oxford, United Kingdom, ³Center for Magnetic Resonance Research, University of Minnesota, Minneapolis, MN, United States

The trade-off between signal to noise ratio and spatial specificity governs the choice of spatial resolution in diffusion-weighted magnetic resonance imaging. We present an approach for tackling this trade-off by combining data acquired both at high and low spatial resolution. We combine all data into a single Bayesian model to estimate the underlying fibre patterns, therefore, combining the benefits of each acquisition. We show that fibre crossings at the highest spatial resolution can be inferred more robustly using this model compared to a simpler model that operates only on high-resolution data, when both approaches are matched for acquisition time.

Ian Nimmo-Smith¹, Frank Fang-Cheng Yeh², and Eleftherios Garyfallidis^3
1MRC Cognition and Brain Sciences Unit, Cambridge, Cambridgeshire, United Kingdom, ²Biomedical Engineering, Carnegie Mellon University, Pittsburgh, PA, United States, ³University of Cambridge, Cambridge, Cambridgeshire, United Kingdom

GQI2 is a new non-parametric method for reconstruction of diffusion imaging data acquired with Q-space cartesian lattice design. Using simulations and human data we show it has superior ability to resolve angular directions in voxels with two or three crossing fiber directions. GQI2 also has efficiency advantages over diffusion spectrum imaging (DSI) in terms of computer memory and computation time.

Matteo Bastiani^1,2, Ana-Maria Oros-Peusquens², Daniel Brenner², Klaus Moellenhoff², Arne Seehaus^1,3, Avdo Celik², Jörg Felder², Andreas Matusch², Ralf Galuske³, Hansjürgen Bratzke⁴, Nadim Jon Shah^2,5, Rainer Goebel¹, and Alard Roebroeck^1
1Department of Cognitive Neuroscience, Maastricht University, Maastricht, Limburg, Netherlands, ²Institute of Neuroscience and Medicine (INM-4), Research Centre Jülich, Jülich, Nordrhein-Westfalen, Germany,³Department of Biology, TU Darmstadt, Darmstadt, Hesse, Germany, ⁴Department of Forensic Medicine, Faculty of Medicine, JWG-University, Frankfurt/Main, Hesse, Germany, ⁵Faculty of Medicine, JARA, RWTH Aachen University, Aachen, Germany

Ultra-high resolution diffusion weighted imaging (DWI) on ex vivo tissue represents a unique tool to investigate human brain anatomy at a microscopic scale. This study focuses on two separate aspects which can be derived from high isotropic resolution HARDI data analysis on post mortem human tissue: i) we reconstruct and analyze the short association fibers connecting human motor and premotor area and fanning cortical insertions in the gyral crown and ii) we distinguish the majority of cortical layers by automatic clustering of their diffusion characteristics. Both aspects are validated using myelin stains of the sectioned tissue.

Van J. Wedeen¹, Ruopeng Wang¹, Timothy G. Reese¹, Thomas Witzel¹, Julian Cohen-Adad², Bruce R. Rosen¹, and Lawrence L. Wald^1
1Martinos Center for Biomedical Imaging, Mass General Hospital, Harvard Medical School, Charlestown, MA, United States, ²Department of Electrical Engineering, Institute of Biomedical Engineering, Ecole Polytechnique de Montreal, Montreal, Quebec, Canada

It has been recently shown that the fiber pathways of the brain follow a natural curvilinear coordinate system, and being vanishingly improbable, this finding is conditionally self-validating. To test these results in human subjects, we obtained Q-ball and diffusion spectrum MRI and with a simple automated procedure mapped their grid structure. Grid structure was observed widely in the brain regions with all diffusion methods, though different methods emphasized different structures. This study indicates that grid structure is readily obtained in humans, and offers a promising strategy for tractography validation and the mapping of brain structure.

Mark Drakesmith¹, John Evans¹, Anthony David², and Derek Jones^1
1CUBRIC, Cardiff University, Cardiff, Wales, United Kingdom, ²Institute of Psychiatry, Kings College London, London, United Kingdom

The effect of false positives on the performance of blind tract clustering is unclear. From an idealised dataset of 6 pre-defined bundles, various distance metrics and clustering methods were tested across varying FP-rates by substituting a random proportion of ‘true’ tracts with FPs. Most methods deteriorate gradually with noise. Affinities computed from maximum (Hausdorff) and endpoint distances were most robust to noise. These methods showed FP misclassification were concentrated on the smallest fibre bundle while other methods showed more diffuse misclassification across adjacent bundles. These two distance metrics are therefore best for clustering noisy tractography datasets.

Christian Ros^1,2, Daniel Guellmar¹, Martin Stenzel², Hans-Joachim Mentzel², and Jürgen R. Reichenbach^1
1Medical Physics Group, Institute of Diagnostic and Interventional Radiology I, Jena University Hospital - Friedrich Schiller University Jena, Jena, TH, Germany, ²Pediatric Radiology, Institute of Diagnostic and Interventional Radiology I, Jena University Hospital - Friedrich Schiller University Jena, Jena, TH, Germany

With this contribution, we presented a new method for the analysis of DTI data sets that uses fiber bundles to enhance the quantitative analysis. By utilizing fiber bundles the occurrence of adverse interpolation effects at the boundaries of white matter structures as a result of the non-linear spatial normalization is prevented. The method was used to assess hemispheric differences in selected fiber bundles using FA maps of 46 healthy volunteers. A statistical analysis was performed and correction methods were employed to deal with statistical errors. In various bundles, statistically significant differences were observed that are in line with the literature.

Anna Varentsova¹, Shengwei Zhang², and Konstantinos Arfanakis^2
1Physics, Illinois Institute of Technology, Chicago, IL, United States, ²Biomedical Engineering, Illinois Institute of Technology, Chicago, IL, United States

Digital human brain white matter (WM) atlases play an important role in brain imaging research. Existing WM atlases have been generated either based on anatomical landmarks, thus mixing tracts with substantially different roles, or using DTI tractography, which fails in regions with crossing fibers. The purpose of this study was to develop a probabilistic WM atlas of the adult human brain by performing probabilistic tractography on an artifact-free high angular resolution diffusion imaging (HARDI) brain template constructed in ICBM-152 space. Presented preliminary results show that the information contained in the new atlas is in agreement with known anatomy.

Shengwei Zhang¹ and Konstantinos Arfanakis^1
1Department of Biomedical Engineering, Illinois Institute of Technology, Chicago, IL, United States

A comprehensive digital human brain atlas was generated in this work. The atlas contains high quality, artifact-free anatomical and diffusion MRI data, and detailed WM and GM labels, in the same space (ICBM-152). The atlas also contains a number of supporting maps of quantities describing the quality of the information provided in different brain structures. The new resource provides a flexible reference frame for integration of macro-structural, micro-structural and functional information about the human brain.

Yu-Chun Lo¹, Yu-Jen Chen¹, Yung-Chin Hsu¹, and Wen-Yih Isaac Tseng^1,2
1Center for Optoelectronic Biomedicine, National Taiwan University College of Medicine, Taipei, Taipei, Taiwan, ²Department of Medical Imaging, National Taiwan University Hospital, Taipei, Taipei, Taiwan

In this study, a standard procedure of tractography was used to develop an atlas of fiber tracts of the whole brain. This procedure entailed a diffusion spectrum imaging (DSI) template on which interested fiber pathways and their ROIs were determined and validated by experts. A total of 107 major white matter tracts were reconstructed. Based on this atlas, a template-based approach was proposed to enable an automated analysis of the microstructural integrity of the tracts identified in the atlas. The atlas and the automated tract analysis method may facilitate the connectome study in a large cohort.

Kerstin Pannek¹, Roslyn Boyd¹, and Stephen Rose^2
1The University of Queensland, Brisbane, Queensland, Australia, ²The Australian E-Health Research Centre, CSIRO, Brisbane, Australia

Connectomes are typically represented in matrix form. To improve visual assessment of the connectome, we suggest the use a streamtube representation of connections. Representation of connections contained in the connectome using streamtubes allows visualization of confidence in the existence of a connection (streamline number) and local attributes of the connections (e.g. FA). Tube thickness, colour and opacity can be fine-tuned for optimized visualization of cortical or deep structures. The opacity and colour of individual streamtubes can be adjusted to highlight subsets of connections, such as connections of altered connectivity.

Kiran K. Seunarine¹, Jonathan D. Clayden², and Christopher A. Clark^2
1Imaging and Biophysics Unit, UCL Institute of Child Health, London, United Kingdom, ²Imaging and Biophysics Unit, University College London, London, United Kingdom

Tractography allows the probing of brain connectivity. Limitations of the standard approaches are that they can be time consuming, require good anatomical knowledge and can be prone to false-positives. Neighbourhood tractography (NT) overcomes these limitations by using a priori information about the expected path of the tract through the brain. This work introduces a whole-brain extension to the NT algorithm to overcome some of the limitations of the method. We demonstrate the approach by segmenting four white-matter tracts. The segmentations show clear differences over single-seed NT, including fewer false-positives and a greater extent to the segmentation.

Mark J. Lowe¹, Ken E. Sakaie¹, Katherine Koenig¹, Lael Stone², Robert A. Bermel³, and Micheal D. Phillips^1
1Imaging Institute, Cleveland Clinic, Cleveland, OH, United States, ²Neurologic Institute, Cleveland Clinic, Cleveland, OH, United States, ³Neurologic Institute, The Cleveland Clinic, Cleveland, OH, United States

There are several diffusion and fiber tracking-related measures that are used as metrics of functional connectivity. Many of these depend on the tracking algorithm. This makes it difficult to compare different studies. We propose a very simple tensor-based measure as a robust metric of anatomic connectivity that can easily be combined with different tracking algorithms and results in a metric that can be compared across studies.

Alessandra Griffa^1,2, Richard Betzel³, Kim Q. Do⁴, Philippe Conus⁵, Patric Hagmann^1,6, and Jean-Philippe Thiran^1,6
1Signal Processing Laboratory (LTS5), Ecole Polytechnique Fédérale de Lausanne (EPFL), Lausanne, Vaud, Switzerland, ²Department of Radiology, University Hospital Center (CHUV) and University of Lausanne (UNIL), Lausanne, Vaud, Switzerland, ³Department of Psychological and Brain Sciences, Indiana University, Bloomington, Indiana, United States, ⁴Center for Psychiatric Neurosciences, Lausanne University Hospital (CHUV) and University of Lausanne, Lausanne, Vaud, Switzerland, ⁵Department of Psychiatry, Lausanne University Hospital (CHUV), Lausanne, Vaud, Switzerland, ⁶Department of Radiology, Univerity Hospital Center (CHUV) and University of Lausanne (UNIL), Lausanne, Vaud, Switzerland

Diffusion MRI, tractography and graph analysis allowed to characterize the brain structural architecture as a small-world, hierarchically modular network. The study of the brain modular topology is raising new interest, and could be a key approach for the understanding of neurodevelopmental disorder. In this framework, it is important to individuate representative partitions for whole groups of subjects. In this work we use information theory-derived measures to quantify the inter-subjects variability of structural network modular decomposition, and we propose different approaches (and particularly the consensus clustering algorithm) to individuate a group-representative partition.

Robert E. Smith¹, Jacques-Donald Tournier¹, Fernando Calamante¹, and Alan Connelly^1
1Brain Research Institute, Florey Institute of Neuroscience and Mental Health, Heidelberg, Victoria, Australia

SIFT (Spherical-deconvolution Informed Filtering of Tractograms) has recently been proposed as a method for improving the correspondence between a whole-brain fibre-tracking reconstruction and the underlying diffusion data. Here we evaluate its utility for quantitative tractography analysis (such as connectomics) by comparing properties of tractography reconstructions before and after application of SIFT to those estimated from post-mortem brain dissection.

Jennifer S.W. Campbell¹ and Bruce G. Pike^2
1McConnell Brain Imaging Centre, McGill University, Montreal, Quebec, Canada, ²McConnell Brain Imaging Centre, Montreal Neurological Institute, McGill University, Montreal, Quebec, Canada

The purpose of this study was to evaluate how well the residual bootstrap statistical technique predicts the variability in scan-rescan estimates of fibre orientation, using spherical deconvolution diffusion MRI. Here, we evaluated how well the fibre orientations obtained from different, coregistered datasets in the same subject fit the fibre probability distribution function obtained from the residual bootstrap technique. For major fibre tracts (i.e., FA>0.3), the correspondence between the observed variability in the fibre orientations and the variability predicted by the bootstrap was very good. For low FA, the residual bootstrap underestimates the scan-rescan repeatability.

Maxime Chamberland¹ and Maxime Descoteaux^1
1Computer Science, Université de Sherbrooke, Sherbrooke, Québec, Canada

Real-time fiber tractography consists in achieving the computation and simultaneous display of fiber tracts. By doing so, this permits the user to interactively tweak any parameters involved in the tracking process and therefore see the effect of each adjustement made on the resulting fibers. For neurosurgical planning, this feature is crucial since giving a pre-computed set of tracts in the hands of neurosurgeon can hide important information, which could have been explored and tuned in real-time.

Ryan P. Cabeen¹, Ksenia Andreyeva², Mark E. Bastin², and David H. Laidlaw^1
1Brown University, Providence, RI, United States, ²University of Edinburgh, Edinburgh, United Kingdom

The primary purpose of this work is to provide a resource of high angular resolution diffusion MRI datasets of 80 normal volunteers aged 25-64 together with neuropsychological testing data spanning general cognitive ability, memory, and information processing speed. To support the value of these data, we demonstrate statistical correlations between age and tract-based metrics calculated automatically from the data. The metrics are a function of fiber count, length, diffusion rate, and diffusion anisotropy; the statistical relationship to age was modeled with linear regression.

Jonathan D. Clayden¹, Susana Munoz Maniega², Amos J. Storkey³, Martin D. King¹, Mark E. Bastin⁴, and Christopher A. Clark^1
1Institute of Child Health, University College London, London, United Kingdom, ²Division of Clinical Neurosciences, University of Edinburgh, Edinburgh, United Kingdom, ³School of Informatics, University of Edinburgh, Edinburgh, United Kingdom, ⁴Centre for Clinical Brain Sciences, University of Edinburgh, Edinburgh, United Kingdom

This abstract introduces and describes TractoR, a freely available software package for MR image analysis implemented in R. The package has a particular emphasis on tractography and connectivity analysis.

Jonathan D. Clayden¹, Susana Munoz Maniega², Mark E. Bastin³, and Christopher A. Clark^1
1Institute of Child Health, University College London, London, United Kingdom, ²Division of Clinical Neurosciences, University of Edinburgh, Edinburgh, United Kingdom, ³Centre for Clinical Brain Sciences, University of Edinburgh, Edinburgh, United Kingdom

In this work, we compare two freely available, probabilistic and automated methods for segmenting white matter tracts from diffusion MRI data.

Elizabeth Zakszewski¹, Nagesh Adluru¹, Ned Kalin¹, and Andrew L. Alexander^1
1University of Wisconsin-Madison, Madison, WI, United States

Connectivity-based segmentation of the precuneus, an important connectome hub, is performed on DTI data of young rhesus monkeys. We observe how the proportion of connections from the precuneus to functionally different brain regions (limbic, somatosensory, and visual) shifts with increasing age. This gives us valuable information for creating models of brain development.

Stephen Jones¹, Erik B. Beall², Jorge A. Gonzalez-Martinez¹, Blessy Mathew¹, Dileep Nair¹, Imad Najm¹, Michael Phillips¹, Ken E. Sakaie¹, and Myron Zhang^1
1Cleveland Clinic Foundation, Cleveland, OH, United States, ²Cleveland Clinic, Cleveland, OH, United States

We study the correlation between diffusion weighted imaging (DWI) measures of structural brain connectivity and electrophysiological connectivity obtained from intracranial electrodes in epilepsy patients. We have collected data from four patients, each with around 150 electrode contacts implanted, and our analysis suggests that the correlation between DWI-based connectivity and electrophysiological connectivity is only modest and still uncertain, which may shape how we interpret DWI data and results.

Ken E. Sakaie¹, Mark J. Lowe¹, Katherine Koenig¹, Robert A. Bermel², Lael Stone², and Michael Phillips^1
1Imaging Institute, The Cleveland Clinic, Cleveland, OH, United States, ²Mellen Center for Multiple Sclerosis Treatment and Research, The Cleveland Clinic, Cleveland, OH, United States

As the primary efferent of hippocampus, diffusion tensor imaging (DTI) of fornix has potential as a biomarker for memory-related cognitive decline. As commonly-used spatial resolution for DTI risks partial volume averaging with surrounding CSF spaces, increases in diffusivity associated with disease that simply result from atrophy may be misinterpreted as change in tissue integrity. Here, we examine the use of high spatial resolution DTI of fornix to avoid partial volume averaging. The results suggest that 1mm isotropic voxels are sufficient for characterizing fornix in healthy controls but may be inadequate in multiple sclerosis patients experiencing atrophy.

Zungho Zun¹, Deqiang Qiu¹, Jarrett Rosenberg¹, and Greg Zaharchuk^1
1Radiology, Stanford University, Stanford, CA, United States

Arterial spin labeling (ASL) allows for quantitative measurements of cerebral blood flow (CBF), enabling longitudinal observation of CBF in the same subjects. In this work, monthly CBF measurements were performed in five normal volunteers for six months using pseudocontinuous ASL (PCASL) and velocity-selective arterial spin labeling (VSASL). Global CBF measurement of gray matter for six months was, on average, 61 ± 5 ml/100 g /min with PCASL and 44 ± 4 ml/100 g/min with VSASL. While CBF measurement with PCASL showed good agreement with literature values, both methods showed similar coefficient of variation.

Zungho Zun¹, Ajit Shankaranarayanan², and Greg Zaharchuk^1
1Radiology, Stanford University, Stanford, CA, United States, ²GE Healthcare, Menlo Park, CA, United States

Velocity-selective arterial spin labeling (VSASL) is a promising method for measuring cerebral blood flow (CBF) in patients with slow or delayed flow. Most previous works on VSASL were demonstrated using 2D imaging, and no previous work reported comparison of VSASL and PCASL, both with 3D imaging and with the same quantification method. In this study, VSASL was combined with 3D image acquisition and was performed in normal volunteers along with 3D PCASL. CBF maps acquired using two methods showed consistent results. Measured signal-to-noise ratio (SNR) from VSASL was about 30% lower than that of PCASL, but was adequate to provide good image quality.

Esben Thade Petersen^1,2, Jill Britt De Vis¹, Cornelis A.T. van den Berg², and Jeroen Hendrikse^1
1Department of Radiology, UMC Utrecht, Utrecht, Netherlands, ²Department of Radiotherapy, UMC Utrecht, Utrecht, Netherlands

In this work an optimal labeling and acquisition scheme for Arterial Spin Labeling is proposed. Repeated pulsed labeling during a multi time-point Look-Locker readout, allow full-brain acquisition while keeping optimum perfusion signal. The method gained 2-5 times the temporal SNR as compared to existing pCASL sequences while at the same time providing information about auxiliary parameters needed for quantification such as bolus arrival time, tissue T1, and arterial input function and blood volume. The method which is an extension of the model-free QUASAR ASL sequence was dubbed Turbo-QUASAR and it will boost any pulsed labeling approach (Pulsed, Velocity- or Acceleration-selective).

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

Thin slab Flow-sensitive alternating inversion recovery (FAIR) arterial spin labeling (ASL) can be advantageous when the geometry or transit delay of the arterial supply is not known, but FAIR has reduced signal compared to continuous labeling methods. Here, a modification of pseudo-continuous labeling that achieves labeling very close to both sides of a slice is presented. Labeling RF is applied centered on the slab and labeling planes on both sides of the slice are created. The performance of this approach is evaluated in single slice perfusion imaging of the brain.

James A. Meakin^1,2, Natalie L. Voets¹, Thomas W. Okell¹, and Peter Jezzard^1
1FMRIB Centre, University of Oxford, Oxford, United Kingdom, ²Department of Oncology, University of Oxford, Oxford, United Kingdom

Velocity selective arterial spin labeling (VSASL) is not diffusion balanced between the tag and control condition, which causes artefacts from free fluids to appear as positive signal in the perfusion weighted subtractions. This affects both the healthy brain (CSF) and in tumors (inflammation response, edema). Here we present a T2-FLAIR method to null long T1/T2 spins at the time of the velocity selective tag using a segmented BIR-4 inversion. We find that this method increases the accuracy of perfusion measurements with VSASL in healthy grey matter by reducing CSF partial volume effects.

Jia Guo¹ and Eric C. Wong^1
1University of California San Diego, La Jolla, California, United States

In this study, we demonstrated that it is feasible to use Velocity-Selective ASL (VSASL) to measure pulmonary blood flow (PBF) at 3.0 T with bSSFP and cardiac gating. Comparing with FAIR, the intravascular signal of VSASL is minimized by choosing a proper cutoff velocity (2cm/s in this study) without affecting the quantification, therefore providing a less biased thus more accurate estimate of PBF. We expect VSASL to benefit from the longer T2/T2* at a lower field (1.5 T) on measurement of PBF.

Yuriko Suzuki¹, Noriyuki Fujima², Hiroyuki Sugimori², Tetsuo Ogino¹, and Marc Van Cauteren^3
1Philips Electronics Japan, Minato-ku, Tokyo, Japan, ²Hokkaido University Hospital, Sapporo, Hokkaido, Japan, ³Philips Healthcare Asia Pacific, Minato-ku, Tokyo, Japan

Hemodynamic information is useful for the accurate diagnosis, effective treatment and follow-up for numerous neurovascular diseases. Recently developed non-contrast-enhanced time-resolved 4D MRA using arterial spin labeling (ASL) methods achieve both high spatial and temporal resolution with 3D coverage. However, these methods require labeled and control images to make the subtracted image, doubling the scan time. We present a new non-contrast-enhanced time-resolved 4D MRA, in which contrast and labeled images are acquired in a single Look-Locker like acquisition, making scan time half compared to the conventional ASL method.

Magdalena Sokolska¹, Aaron Oliver-Taylor¹, Xavier Golay², and David Thomas^1
1Department of Brain Repair & Rehabilitation, UCL Institute of Neurology, London, UK, United Kingdom, ²Department of Brain Repair & Rehabilitation, University College London, London, UK, United Kingdom

Arterial Spin Labeling (ASL) is a non-invasive imaging technique for quantifying blood perfusion in tissue. Pseudo-continuous ASL has recently emerged as the method of choice. However, concern has been raised about acoustic noise levels experienced by patients undergoing pCASL scans, leading to patient discomfort. This is especially a concern in imaging unsedated infants or elderly patients, where exposure to long lasting acoustic noise can increase patient unease in the scanner. This study investigates the Sound Pressure Level (SPL) produced by the pCASL sequence and how it can be reduced without significantly changing the inversion efficiency.

Michael Helle¹ and Tim Nielsen^1
1Philips Research Laboratories, Hamburg, Germany

Conventional pseudo-continuous arterial spin labeling (pCASL) employs trapezoidal gradient waveforms and requires high gradient strengths in combination with high slew rates for sufficient labeling of flowing blood spins, which results in mechanical vibrations and a high acoustic noise level. In this study, possibilities are presented on how to reduce the acoustic noise level and the mechanical strain of a pCASL sequence, either by optimizing the labeling interval duration or by applying a sinusoidal gradient waveform for the labeling sequence or both.

Randall B. Stafford^1,2, Se-Hong Oh^1,2, Myung-Kyun Woo³, Tiejun Zhao⁴, Kyoung-Nam Kim³, Young-Bo Kim³, Zang-Hee Cho³, John A. Detre^1,2, and Jongho Lee^1,2
1Radiology, University of Pennsylvania, Philadelphia, PA, United States, ²Center for Functional Neuroimaging, University of Pennsylvania, Philadelphia, PA, United States, ³Neuroscience Research Institute, Gachon University, Incheon, Incheon, Korea, ⁴Siemens Healthcare USA, University of Pittsburgh, Pittsburgh, PA, United States

Arterial spin labeling has the dual benefit of increased spin polarity and elongation of the T1 relaxation time of arterial blood at 7T. In order to realize these benefits at 7T, specific absorption rate limits must be mitigated. One way to reduce the SAR deposition is to eliminate the need for RF power during ASL control acquisitions. Here, we use a novel ASL head/labeling coil system with active detuning to eliminate magnetization transfer effects during labeling, thereby allowing us to remove the control RF. This reduces the SAR deposition by almost 50%, and allows for shorter TRs. We demonstrate the utility of this system for CASL in healthy volunteers with a SAR-minimum TR under 7 seconds.

Youngkyoo Jung^1,2, Megan E. Johnston², Christopher T. Whitlow^1,3, and Joseph A. Maldjian^1
1Radiology, Wake Forest School of Medicine, Winston-Salem, North Carolina, United States, ²Biomedical Engineering, Wake Forest School of Medicine, Winston-Salem, North Carolina, United States, ³Translational Science Institute, Wake Forest School of Medicine, Winston-Salem, North Carolina, United States

Pseudo-continuous ASL (PCASL) based vascular territory mapping enables the detection of source arteries. An automated vascular territory segmentation using connectivity information from both image space and label space is proposed. The algorithm was tested with data obtained at two locations with middle cerebral artery branching. The algorithm demonstrates the ability to resolve neighboring sources into separate territories with minimal computational burden.

Yi Dang¹, Jia Guo², Jue Zhang^1,3, and Eric C. Wong^4
1Academy for Advanced Interdisciplinary Studies, Peking University, Beijing, Beijing, China, ²Department of Bioengineering, University of California San Diego, San Diego, California, United States, ³College of Engineering, Peking University, Beijing, Beijing, China, ⁴Department of Radiology and Psychiatry, University of California San Diego, San Diego, California, United States

Random Vessel-encoded arterial spin labeling (R-VEASL) can be used to uniquely identify the locations of the source arteries in the tagging plane. However, localization of the decoded vascular sources is accurate in most but not all cases. In this work, the causes of this have been investigated. In our data, motion during the scan did not dominate the errors in the detecting process. The estimation of vessel locations was only slightly improved by selective filtering of data. In some cases, the detected vessel locations correlated closely with the vascular anatomy several millimeters inferior to the nominal tagging plane.

Yi Dang¹, Jia Guo², Jue Zhang^1,3, and Eric C. Wong^4
1Academy for Advanced Interdisciplinary Studies, Peking University, Beijing, Beijing, China, ²Department of Bioengineering, University of California San Diego, San Diego, California, United States, ³College of Engineering, Peking University, Beijing, Beijing, China, ⁴Department of Radiology and Psychiatry, University of California San Diego, San Diego, California, United States

In Random vessel encoded arterial spin labeling (R-VEASL), the correlation coefficient between acquired perfusion signal data and the theoretical signal model was used to detect the vessel locations. In this work, the accuracy of the theoretical signal model derived from Bloch simulation that is currently used to estimate vessel location in R-VEASL has been demonstrated. The theoretical model appears to generally fit the data well, and is insensitive to vessel velocity selection. This suggests that the method is robust to assumed velocity for the purpose of vessel detection, but that the data likely cannot be used to estimate flow velocities.

Li Zhao¹, Samuel W. Fielden¹, Xiao Chen¹, John P. Mugler, III², Josef Pfeuffer³, Manal Nicolas-Jilwan², Max Wintermark², and Craig H. Meyer^1
1Biomedical Engineering, University of Virginia, Charlottesville, Virginia, United States, ²Radiology, University of Virginia, Charlottesville, Virginia, United States, ³Siemens, Erlangen, Germany

Low SNR in ASL limits the achievable spatial resolution and the accuracy of perfusion maps. Dynamic ASL is time-consuming and also suffers from low SNR. Compressed sensing can improve image quality by enforcing spatial-domain sparsity. Compressed sensing can also enforce time-domain sparsity in dynamic ASL. Volunteer data are shown to demonstrate CS performance on single PLD PCASL images and multiple dynamic frames. The results show image SNR and image quality improvement. More importantly, the estimated CBF becomes more accurate and stable with compressed sensing image reconstruction.

Yihang Zhou^1,2, Jie Zheng³, Dong Liang⁴, and Leslie Ying^1,2
1Electrical Engineering, University at Buffalo, Buffalo, NY, United States, ²Department of Biomedical Engineering, University at Buffalo, Buffalo, NY, United States, ³Department of Radiology, Washington University, St. Louis, MO, United States, ⁴Paul C. Lauterbur Research Centre for Biomedical Imaging, Shenzhen, China

Arterial spin labeled (ASL) MRI method has been used as an alternative to first-pass perfusion imaging in assessing end organ perfusion. It has the advantage of avoiding administration of any contrast agent. However, this method is sensitive to motion artifacts due to prolonged data acquisition time for each T1 weighted image. The motion artifacts will greatly degrade the image quality and reduce the accuracy of perfusion measurements. In this study, the feasibility of accelerating ASL acquisition using compressed sensing is investigated. Results on calf muscle and cardiac experiments demonstrate compressed sensing is able to preserve the perfusion information with accelerated acquisition.

Tae Kim¹, Wanyong Shin², Erik B. Beall², Mark J. Lowe², Tiejun Zhao³, and Kyongtae Ty Bae^1
1Radiology, University of Pittsburgh, Pittsburgh, PA, United States, ²Imaging Institute, Cleveland Clinic, Cleveland, OH, United States, ³Siemens Medical Solution USA, INC., Siemens MediCare Healthcare USA, Pittsburgh, PA, United States

Multi-band (MB) excitation for data acquisition was successfully implemented into pulsed arterial spin labeling and evaluated on healthy volunteers at 3T. Perfusion quantification for MB excitation was highly comparable with that of a conventional perfusion acquisition. Our study demonstrates that the MB technique facilitates an accelerated acquisition of high resolution, whole-brain perfusion maps.

Hiroshi Toyoda^1,2, Guoxiang Liu², Yusuke Morito², and Yasuyoshi Watanabe^1
1RIKEN Center for Molecular Imaging Science, Kobe, Japan, ²National Institute of Information and Communications Technology, Kobe, Japan

A new 3D arterial spin labeling (ASL) acquisition using multi-slab 3D gradient and spin echo (GRASE) was proposed, which shortened scanning time without sacrificing measurement accuracy. We have shown that the ASL imaging based on a multi-slab 3D GRASE sequence provided equivalent results in the absolute quantification of the CBF to that of the single slab sequence. The feasibility of this new multi-slab 3D acquisition approach was demonstrated by its better time efficiency and equivalent quantification accuracy.

Erin K. Englund¹, Michael C. Langham¹, Cheng Li¹, and Felix W. Wehrli^1
1Department of Radiology, University of Pennsylvania, Philadelphia, PA, United States

In this work, a traditional pulsed arterial spin labeling (PASL) technique is combined with simultaneous image refocusing (SIR) EPI readout to measure dual-slice perfusion in the leg. PASL-SIR is evaluated by comparing results acquired in the two slices to an otherwise identical single-slice PASL sequence during a series of ischemia reperfusion paradigms in the leg. Time course parameters for the dual-slice PASL-SIR technique are in relative agreement with single-slice-measured perfusion and with literature reported values. Implementation of PASL-SIR in longitudinal studies may improve reproducibility by reducing errors due to subject repositioning.

Karthik Prabhakaran¹, Ryan D. Hopson¹, Mark A. Elliott¹, Kosha Ruparel¹, Raquel E. Gur¹, Ruben C. Gur¹, and John A. Detre^1
1University of Pennsylvania, Philadelphia, PA, United States

The use of frequency-selective fat saturation in EPI based perfusion imaging results in spurious measurements of cerebral blood flow (CBF) if the lipid suppression is poor and inconsistent. Frequency-selective water excitation provided superior lipid suppression and yielded higher CBF measurements when compared to frequency-selective fat-saturation in a spin-echo EPI pseudo-continuous arterial spin labeling (pCASL) application at 3T.

Megan E. Johnston¹, Joseph A. Maldjian², and Youngkyoo Jung^1,2
1Biomedical Engineering, Wake Forest University School of Medicine, Winston-Salem, North Carolina, United States, ²Radiology, Wake Forest University School of Medicine, Winston-Salem, North Carolina, United States

The arterial transit time is a valuable metric for both optimizing arterial spin labeling parameters and as a complementary diagnostic tool in cerebrovascular disease. Simulations were performed for three arterial spin labeling based methods in order to compare their efficiencies in estimating the arterial transit time and cerebral blood flow. The three methods included a method with variable TR, Hadamard-encoded ASL, and a method using Look-Locker acquisition. The variable TR method produced the lowest error in arterial transit time estimation and the lowest percentage error in cerebral blood flow measurement over the range of arterial transit times simulated.

Li Zhao¹ and Craig H. Meyer^1,2
1Biomedical Engineering, University of Virginia, Charlottesville, Virginia, United States, ²Radiology, University of Virginia, Charlottesville, Virginia, United States

Noise in low SNR ASL images is more accurately modeled as Rician rather than Gaussian. Least squares estimation is typically used in ASL, but this results in a biased estimate with Rician noise. This work describes a new maximum likelihood (ML) estimator and an optimal post-label delay (PLD) design for dynamic ASL assuming Rician noise. To verify the performance of CBF estimation, a simulation is performed based on low SNR dynamic ASL signal. The results show that the new ML estimator provides unbiased estimation and that optimal PLD design can reduce the variance of CBF estimation significantly.

David D. Shin¹, Ho-Ling Liu², and Thomas T. Liu^1
1Center for Functional MRI, University of California, San Diego, La Jolla, CA, United States, ²Chang Gung University, Taoyuan, Taiwan

We present a novel PCASL technique for correcting compromised inversion efficiency resulting from off-resonance effects and gradient imperfections at the tagging plane. The implemented sequence acquires the ASL signal using 8 RF phase offsets during the first 32 reps then switches to the regular tag/control mode to acquire the conventional PCASL signal. The inversion response curves from the first mode is used to estimate the tagging efficiencies for the main feeding arteries, which are in turn used during the quantification step to correct for underestimated CBFs associated with regular PCASL. The gray matter CBFs calculated from alpha-corrected PCASL are compared to those from OptPCASL which restores inversion efficiency but requires additional calibration scans.

Johanna Kramme¹ and Matthias Günther^1,2
1Fraunhofer MEVIS, Bremen, Germany, Germany, ²Faculty of Physics and Electronics, University of Bremen, Bremen, Germany, Germany

Precise arterial spin labeling T2 measurements are challenging because they are influenced by many factors like the choice of crusher gradients, the refocusing flip angle, and the number of fitted echoes. When extended to multi-TI T2 measurements, SNR at longer inflow times can be critical, especially if scan time limits the number of image averages. Here, a robust and reliable multi-TI T2 acquisition and fitting routine is presented that is fast enough to be implemented in clinical routine to determine T2 values for every individual patient. The T2 values could then be incorporate in two compartment models for permeability quantification.

Roman Fleysher¹, Mark E. Wagshul¹, Michael L. Lipton¹, and Craig A. Branch^1
1Gruss Magnetic Resonance Research Center, Department of Radiology, Albert Einstein College of Medicine, Bronx, NY, United States

Both Gaussian dissipation and single compartment Kety models are found to be inadequate for description of ASL experiments for two major reasons: 1) bolus disperses continuously as it travels down the vascular tree and must be modeled as such all the way to the capillary level and 2) exchange through the capillary wall is not instantaneous. We use more realistic simulations with continuous bolus dispersion and restricted water permeability of capillary walls to ascertain sensitivity of CASL, ITS-FAIR and Quasar ASL techniques to bolus dispersion. Quasar is found to handle bolus dispersion most effectively at 20 degree flip angle.

Kathrin Lorenz¹, Toralf Mildner¹, André Pampel¹, and Harald E. Möller^1
1Nuclear Magnetic Resonance Unit, Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Saxony, Germany

Inflow vascular-space-occupancy (iVASO) is a non-invasive method for the quantification of absolute arterial cerebral blood volume (aCBV). The present work is purposed to investigate determining factors of aCBV measurement like slow inflow of inverted blood from small vessels and inflow of fresh, i.e. non-inverted, blood in case of short arterial transit times. Transient effects related to the blood delivery by vessels of variable size were addressed by dedicated experiments. Such effects potentially lead to biased values in the aCBV measurement and were shown to vary between different regions in the human brain.

Elias Kellner¹, Roman Fleysher², Matthias Günther³, Marco Reisert¹, Peter Gall¹, and Valerij G. Kiselev^1
1Department of Radiology, University Hospital Freiburg, Freiburg, Germany, ²Gruss Magnetic Research Center, Department of Radiology, Albert Einstein College of Medicine, Bronx, NY, United States, ³Institute for Medical Image Computing, Fraunhofer MEVIS, 28359, Germany

Proper account for blood transport is imperative for quantitative analysis of pharmacokinetic and perfusion data in DSC, DCE, ASL MRI and CT/PET. The empirical nature of currently available models hinders prediction of delay and dispersion in higher generations of the vascular tree. We present an analytical framework for solving this problem based on the morphological parameters of vascular networks. Explicit expression with one fitting parameter is obtained for a simple scaling vascular tree. An ASL measurement in large cerebral arteries supports the model, but demonstrates weak disambiguation between several models using this vessel group.

Patrick W. Hales¹, Kim P. Phipps², Ramneek Kaur¹, Tina Banks³, and Christopher A. Clark^1
1Imaging and Biophysics, University College London, London, London, United Kingdom, ²Neuro-oncology Department, Great Ormond Street Hospital, London, London, United Kingdom, ³Radiology Department, Great Ormond Street Hospital, London, London, United Kingdom

A ‘two-stage’ model for quantification of CBF and arterial volume fraction in dynamic ASL data is presented. The two-stage model provides CBF values in healthy grey matter that are more in line with previously published data from PET studies (compared to standard ‘single stage’ ASL models), due to a reduction in ‘large vessel’ artefacts. Furthermore, the two-stage model is shown to provide a better fit to data collected in paediatric brain tumour patients, and provide novel contrast, in terms of arterial volume maps in and around the tumour region.

Feng-Xian Yan¹, Alex M. Wong², and Ho-Ling Liu^1,2
1Department of Medical Imaging and Radiological Sciences, Chang Gung University, Taoyuan, Taiwan, ²Department of Medical Imaging and Intervention, Chang Gung Memorial Hospital, Taoyuan, Taiwan

This study aimed to evaluate whether pseudo-continuous ASL (PCASL) perfusion method is sensitive to leaky vessels in brain tumors by comparing with DSC-MRI with and without leakage correction. Eight patients with contrast-enhancing tumors participated in this study. The results demonstrated that tumor/white matter (WM) CBF ratios obtained from PCASL had significant positive correlation with the leakage-corrected tumor/WM CBV ratios obtained from DSC-MRI(p=0.01), but not with the CBV ratios before correction(p=0.13). This study provides a direct evidence for supporting that using PCASL for evaluation of brain tumor perfusion is advantageous because it is insensitive to contract agent extravasation from tumor vessles.

Marco Castellaro¹, Amit Mehndiratta², Denis Peruzzo¹, Gianluigi Pillonetto¹, Esben Thade Petersen³, Xavier Golay⁴, Michael A. Chappell², and Alessandra Bertoldo^1
1Department of Information Engineering, University of Padova, Padova, PD, Italy, ²Institute of Biomedical Engineering, University of Oxford, Oxford, United Kingdom, ³Departments of Radiology and Radiotherapy, University Medical Center Utrecht, Utrecht, Netherlands, ⁴University College London, London, United Kingdom

QUASAR ASL allows the simultaneous estimation of Cerebral Blood Flow (CBF) and the residue function using deconvolution techniques. One vital aspect of quantification process is the estimation of delay between arterial input function and tissue. The error in delay might be propagated to CBF estimation. In this study the performance of a novel deconvolution method, Stable Spline (SS) was compared with the most commonly used method (oSVD) both on simulated and clinical data. SS showed a reliable estimation of delay, CBF along with physiologically realistic residue function compared to oSVD.

Xiaoyun Liang¹, Alan Connelly^1,2, and Fernando Calamante^1,2
1Brain Research Institute, Florey Institute of Neuroscience and Mental Health, Heidelberg, VIC, Australia, ²Department of Medicine, Austin Health and Northern Health, University of Melbourne, Melbourne, VIC, Australia

The SNR of fMRI images is critical for functional connectivity studies. Although ASL has some advantages over BOLD fMRI, the reliability for detecting networks may be compromised due to its intrinsic low SNR. In this study, we proposed a denoising method combining block-wise non-local means and dual-tree complex wavelet transform to enhance the SNR of ASL images. Simulations show that the proposed method was superior to discrete wavelet transform. The validity of the proposed method has been further confirmed by the more robust detection of functional connectivity from in vivo data. Overall, the proposed method can enhance the SNR of ASL data significantly and thus enable more reliable network detection.

Yongsheng Zhang¹ and Ze Wang^1
1Department of Psychiatry, University of Pennsylvania, Philadelphia, PA, United States

ASL MRI has an intrinsic low signal-to-noise-ratio, requiring an efficient denoising method. Spatial smoothing is efficient for suppressing random noise but at the expense of a resolution loss. We proposed a resolution preserving ASL denoising method based on the Markov random field theory. Our evaluation results showed that the proposed method can gain up to 4 fold SNR increase while still keep the structural details intact. By contrast, spatial smoothing induced severe blurring effects when a similar level of SNR increase was achieved.

Marjorie Villien^1,2, Alexandre Krainik³, Julien Bouvier¹, Matthias J.P. van Osch⁴, Laurent Lamalle⁵, Irène Troprès⁵, and Jan M. Warnking^1
1Grenoble Institut of Neurosciences, INSERM, Grenoble, France, ²Athinoula A. Martinos center for biomedical imaging, Massachusetts General Hospital, Charlestown, Massachusetts, United States, ³Clinique universitaire de neuroradiologie et d'IRM, CHU Grenoble, Grenoble, France, ⁴Leiden University Medical Center, Leiden, Netherlands, ⁵SFR1, Université Joseph Fourier, Grenoble, France

Our aim is to compare the performance of a variety of data analysis methods in order to maximize the robustness of ASL CVR mapping in the context of clinical exams and basic and clinical research. Here, we analyzed 56 sessions of ASL vasoreactivity data obtained in patients and healthy subjects using the classical block regressor and regressors based on the physiological state of the individual subjects. We also analyzed the effect of excluding data obtained during the transition periods between capnia levels, and of regressors modeling physiological noise. Regressors based on individual capnia timecourses consistently outperformed standard block regressors.

Clément S. Debacker^1,2, Jan M. Warnking^1,3, Jérôme Voiron², and Emmanuel Luc Barbier^3,4
1Team 5: Functional NeuroImaging and Brain Perfusion, Grenoble-Institute of Neuroscience, La Tronche, France, ²MRI, Bruker BioSpin, Ettlingen, Germany, ³U836, INSERM, La Tronche, France, ⁴Team 5: Functional NeuroImaging and Brain Perfusion, Grenoble Institute of Neurosciences, La Tronche, France

Arterial Spin Labeling (ASL) methods can quantify cerebral blood flow (CBF), but this requires an estimate of the inversion efficiency (IE). The purpose of this study is to estimate IE for three ASL methods (pulsed ASL (PASL), continuous ASL (CASL), and pseudo-continuous ASL (pCASL)) and two magnetic fields in the context of pre-clinical imaging. We also show that shim correction seems to increase IE.

Magdalena Sokolska¹, Xavier Golay², and David Thomas^1
1Department of Brain Repair & Rehabilitation, UCL Institute of Neurology, London, UK, United Kingdom, ²Department of Brain Repair & Rehabilitation, University College London, London, UK, United Kingdom

Arterial Spin Labelling (ASL) is a non-invasive method allowing quantitative measurement of CBF. Pseudo-continuous ASL (pCASL) is considered one of the best ASL techniques due to its high SNR. However, the accurate and reproducible CBF quantification using this technique can be challenging , because the labelling efficiency of the protons (α), which measures the effectiveness of the magnetic labelling of spins, depends on a number of subject-specific factors. The aim of this study was to compare different methods of labelling efficiency estimation of pCASL and evaluate error distribution on perfusion quantification resulting from α.

Harshan Ravi^1,2, Jinsoo Uh³, Peiying Liu¹, Lisa C. Krishnamurthy^1,2, and Hanzhang Lu^1
1Advanced Imaging Research Center, University of Texas Southwestern Medical Center, Dallas, Texas, United States, ²Department of Bioengineering, University of Texas at Arlington, Arlington, Texas, United States,³Department of Radiological Sciences, St. Jude Children’s Research Hospital, Memphis, Tennessee, United States

Venous blood oxygenation is an important parameter for the quantification of cerebral metabolic rate of oxygen (CMRO2), a key biomarker for brain tissue viability and functionality. T2 based methods has been reported to measure Yv in macroscopic veins. However, the oxygenation of microvasculature (capillary/venule), which renders spatially more specific information, is not yet investigated. In this study we propose a novel technique to isolate microvascular venous blood signal and measure its oxygenation, named micro-vasculature T2-relaxation-Under-Spin-Tagging (mTRUST). This technique complements and extends the applicability of existing methods targeting macroscopic veins, and may lay foundation to region specific CMRO2 mapping in future.

Joost P. A. Kuijer¹, Larissa W. van Golen², Marc C. Huisman³, Richard G. IJzerman², Frederik Barkhof³, Michaela Diamant², and Adriaan A. Lammertsma^3
1Physics and Medical Technology, NCA, VU University Medical Center, Amsterdam, NL, Netherlands, ²Diabetes Center, Internal Medicine, VU University Medical Center, Amsterdam, NL, Netherlands, ³Radiology and Nuclear Medicine, NCA, VU University Medical Center, Amsterdam, NL, Netherlands

Arterial Spin Labeling (ASL) was compared with H₂¹⁵O Positron Emission Tomography (PET), the current gold standard for non-invasive quantification of CBF, in patients with type 1 diabetes and healthy subjects. The ASL variant used was pseudo-continuous, with background suppression, and a 3D FSE spiral readout. Both methods were compared in terms of absolute quantification and regional differences.

David D. Shin¹, Burak I. Ozyurt², and Thomas T. Liu^1
1Center for Functional MRI, University of California, San Diego, La Jolla, CA, United States, ²Dept. of Psychiatry, University of California, San Diego, La Jolla, CA, United States

Since its introduction in 2011 ISMRM, the CBFBIRN Database and Analysis Pipeline (CBFDAP) has processed/stored more than 1300 ASL datasets contributed by 24 research studies. The availability of CBF data from many diverse conditions (e.g. Alzheimer’s disease, schizophrenia, bipolar disorder, etc.) and associated subject demographics/assessments provides a unique opportunity to explore and mine clinically useful information. In order to facilitate this process, we developed a host of automated group analysis tools and integrated them to the CBFDAP. Three types of group analysis are supported, Path 1: mean gray matter analysis; Path 2: Regional analysis with user-provided ROIs; Path 3: voxel-level standard space analysis. Examples from all three paths are presented.

Yen-Yu Ian Shih^1,2, Bryan H. De La Garza², Shiliang Huang², Guang Li², Lin Wang³, and Timothy O. Duong^4
1Experimental Neuroimaging Laboratory, Department of Neurology and Biomedical Research Imaging Center, University of North Carolina, Chapel Hill, NC, United States, ²Research Imaging Center, University of Texas Health Science Center at San Antonio, San Antonio, TX, United States, ³Devers Eye Institute, Legacy Research Institute, Portland, OR, United States, ⁴Research Imaging Center, UT Health Science Center at San Antonio, San Antonio, TX, United States

Continuous arterial spin labeling (CASL) MRI, widely used to non-invasively image blood flow (BF) of the brain, has been applied to image BF of the rat brain and retina but remains to be validated. The goal of this study was to use an established microsphere technique to cross-validate ASL MRI BF measurement in the rat brain and retina. By using a mixture of the two different sized fluorescent microspheres with two different colors, the cerebral, retinal and choroidal BF can be measured simultaneously in the same subject. Our data showed that BF values by MRI are in good accordance with the microsphere technique in the rat retina and cerebral cortex.

Esther A H Warnert¹, Ashley D. Harris¹, Kevin Murphy¹, Michael A. Chappell², Judith E. Hall³, and Richard G. Wise^1
1Cardiff University Brain Research Imaging Centre, Cardiff, South Glamorgan, United Kingdom, ²University of Oxford, Headington, Oxford, United Kingdom, ³Department of Anaesthetics, Intensive Care and Pain Medicine, School of Medicine, Cardiff, South Glamorgan, United Kingdom

We have performed a multi-inversion time pulsed ASL study with healthy subjects to obtain tissue perfusion kinetic curves of the brainstem. A two-compartment model was used that contains a tissue and intra-arterial component in order to model out macrovascular signal. Our study showed that the brainstem perfusion signal indeed has a relatively high macrovascular component, which can be model out to obtain tissue perfusion curves.

Chun-Xia Li¹, Sudeep Patel¹, Danny J.J. Wang², and Xiaodong Zhang^1
1Yerkes Imaging Center, Yerkes National Primate Research Center, Emory University, Atlanta, GA, United States, ²Ahmanson-Lovelace Brain Mapping Center, Department of Neurology, UCLA, Los Angeles, CA, United States

Non-human primates were widely used as various disease models in neuroscience studies and examined under isoflurane anesthesia. It is known that the cerebral blood flow (CBF) autoregulation can be disrupted under high isoflurane dosage and is more vulnerable in subcortical regions, but the regional specification of the effect remains poorly understood. In the present study, the pseudo continuous arterial-spin-labeling (pCASL) technique was used to evaluate the dose-dependent effect of deep isoflurane anesthesia on CBF of different brain structures. The result indicates that the CBF autoregulation in most cortical and subcortical regions of monkeys is impaired under 2% isoflurane.

Wen-Chau Wu^1,2, Shu-Chi Lin², and Kuan-Lin Chen^2
1Graduate Institute of Oncology, National Taiwan University, Taipei, Taiwan, ²Medical Imaging, National Taiwan University Hospital, Taipei, Taiwan

The feasibility of ASL MRI in cerebral perfusion measurement has been recognized in gray matter, but remains pending in white matter, mainly due to the tissue's low flow rate and long transit time. We tried to optimize the labeling duration () and post-labeling delay (PLD) of pseudocontinuous ASL experimentally and numerically, based on a spatial resolution (1.56x1.56x5 mm³) finer than commonly used (3.4x3.4x5 mm³) to avoid partial volume effect. Results showed that cerebral perfusion can be measured in about 60% of white matter and that the optimal value is 2000 ms for  and 1500-1800 ms for PLD.

Fabian Tobias Gutjahr¹, Thomas Kampf¹, Xavier Helluy¹, Christian Herbert Ziener², Peter M. Jakob^1,3, and Wolfgang Rudolf Bauer^4
1Experimental Physics 5, University of Wuerzburg, Wuerzburg, Bavaria, Germany, ²German Cancer Research Center, Heidelberg, Baden Wuerttemberg, Germany, ³Magnetic Resonance Bavaria, Wuerzburg, Bavaria, Germany, ⁴Medizinische Klinik und Polyklinik I, Universitaetsklinikum Wuerzburg, Wuerzburg, Bavaria, Germany

In this work a retrospectively triggered method for the quantification of perfusion and regional blood volume (RBV) in murine myocardium is demonstrated. Perfusion and RBV maps can be reconstructed on any desired position in the heart cycle. Furthermore the calculation of RBV maps requires a voxel by voxel comparison of a pre and post contrast agent T₁ map. A good match for both map can be guaranteed due to the ability to reconstruct T₁ maps of any desired position in the heart cycle. The obtained RBV and perfusion values are shown to be in agreement with previous studies.

Erin K. Englund¹, Michael C. Langham¹, Cheng Li¹, Emile R. Mohler², Thomas F. Floyd³, and Felix W. Wehrli^1
1Department of Radiology, University of Pennsylvania, Philadelphia, PA, United States, ²Department of Cardiology, University of Pennsylvania, Philadelphia, PA, United States, ³Department of Anesthesiology, Stony Brook University, Stony Brook, NY, United States

A novel method to simultaneously measure perfusion, venous oxygen saturation, and skeletal muscle T₂* using an interleaved pulsed arterial spin labeling and multi-echo GRE sequence, termed PASL/Ox-BOLD, is presented. The technique is assessed in healthy subjects during a series of ischemia reperfusion paradigms. Time course data is analyzed to investigate the kinetics of recovery following cuff-induced ischemia. Results indicate that PASL/Ox-BOLD is capable of faithfully measuring all three parameters at 2 second temporal resolution. The method is also used to investigate a small cohort of peripheral artery disease patients. In these patients, a blunted and delayed hyperemic response is detected.

Kiril Schewzow^1,2, Georg Bernd Fiedler^1,2, Martin Meyerspeer^1,2, Ewald Moser^1,2, and Albrecht Ingo Schmid^1,2
1Medical Physics and Biomedical Engeneering, Medical University of Vienna, Vienna, Wien, Austria, ²MR Centre of Excellence, Medical University of Vienna, Vienna, Wien, Austria

Adequate perfusion is essential for muscle function and tissue health. Vascular complications are a common problem in several diseases like diabetes mellitus. Arterial spin labeling represents a non-invasive assessment of tissue perfusion using magnetically labeled blood water as an endogeneous tracer. ASL has been used to measure dynamic changes in skeletal muscle perfusion during and after exercise (plantar flexion). In this study we demonstrate the feasibility of measuring ASL in aerobic exercise at 7 Tesla with a high temporal resolution and good SNR and high intra-individual reproducibility.

Rajiv Ramasawmy^1,2, Adrienne E. Campbell-Washburn¹, Sean Peter Johnson², Jack Wells¹, Rosamund Barbara Pedley², Simon Walker-Samuel¹, and Mark F. Lythgoe †^1
1Centre for Advanced Biomedical Imaging, University College London, London, Greater London, United Kingdom, ²Cancer Institute, University College London, London, Greater London, United Kingdom

We present a novel use of a multi-slice respiratory-triggered flow-sensitive alternating inversion recovery (FAIR) Look-Locker arterial spin labelling (ASL) technique to measure perfusion within a mouse liver. Measurements from the multi-slice technique were in good agreement with those from single-slice triggered FAIR Look-Locker ASL. As the acquisition time is the same for the multi-slice and single-slice sequences, the multi-slice technique demonstrates a three-fold increase in efficiency of liver coverage. Pre-clinical liver perfusion measurements will find utility in a number of disease models such as liver metastases and liver cirrhosis. As ASL does not require injected contrast agents, this technique is suitable for follow-up and repeated measurements to track subject response to therapies.

Huan Tan¹, Ioannis Koktzoglou^1,2, and Pottumarthi Vara Prasad^1,2
1NorthShore University HealthSystem, Evanston, IL, United States, ²The University of Chicago Pritzker School of Medicine, Chicago, IL, United States

A single breath-hold scan does not provide enough time for signal averaging needed by arterial spin labeling technique in renal perfusion measurement. In our study, we have implemented a novel two-dimensional (2D) navigator technique in concert with FAIR True-FISP to minimize respiratory motion via retrospective reconstruction and allow for multiple averages during free-breathing acquisition. The 2D navigator-gated approach achieved good image quality in patients while the breath-hold results were unusable. Given the limitation of using contrast agents in patients with compromised renal function, free breathing ASL offers a non-invasive method to evaluate renal perfusion.

Manil Chouhan^1,2, Rajiv Ramasawmy^2,3, Adrienne E. Campbell-Washburn², Alan Bainbridge⁴, Jack Wells², Nathan Davies⁵, Rosamund Barbara Pedley³, Raj Mookerjee⁵, Shonit Punwani¹, Stuart Taylor¹, Simon Walker-Samuel², and Mark F. Lythgoe^2
1Centre for Medical Imaging, University College London, London, Greater London, United Kingdom, ²Centre for Advanced Biomedical Imaging, University College London, London, Greater London, United Kingdom,³Cancer Institute, University College London, London, Greater London, United Kingdom, ⁴Department of Medical Physics and Bioengineering, University College London, London, Greater London, United Kingdom,⁵Institute for Liver and Digestive Health, University College London, London, United Kingdom

Non-invasive liver perfusion measurements could be used to monitor hepatic disease development and provide a functional biomarker for novel therapies. Arterial spin labelling (ASL) has not found extensive utility in the liver, mainly due to its dual vascular supply and susceptibility to respiratory motion. ASL can provide regional perfusion maps whereas with phase-contrast measurements taken in the portal vein can provide a bulk portal perfusion. Previous work has reported mouse liver regional perfusion maps but here we present rat liver ASL validated against phase-contrast measurements.

Hongxia Lei^1,2 and Rolf Gruetter^3,4
1University of Geneva, Geneva, Geneva, Switzerland, ²EPFL, Lausanne, Vaud, Switzerland, ³École Polytechnique Fédérale de Lausanne, Geneva, Geneva, Switzerland, ⁴University of Lausanne, Lausanne, Vaud, Switzerland

The aim of our study was to apply the continuous arterial spin labeling (CASL) technique on mouse before and during hypoglycemia at 9.4T. The elevated blood flow upon hypoglycemia was observed even when animals were under isoflurane anesthesia. This study opens possibility of investigating genetically modified models using multiple MR modalities at high magnetic fields.

Weiying Dai¹, Ajit Shankaranarayanan², and David C. Alsop^1
1Radiology, Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, MA, United States, ²Global Applied Science Laboratory, GE Healthcare, Menlo Park, CA, United States

Arterial transit delay (ATD) measurement has been proposed to eliminate the ATD errors in perfusion measurements. In elderly cohorts, potentially longer and more variable ATD may introduce systematic errors in perfusion measurement if not taken into account. A quick ATD measurement was performed in 59 elderly patients over 70 years old scheduled for elective surgical procedures. ATD was heterogeneous across different brain regions and significantly longer than the young in all regions except posterior regions. ATD measurement can improve the reliability and accuracy of perfusion measurement in the elderly population and thus should improve the sensitivity of clinical research studies involving elderly subjects.

Maria A. Fernandez-Seara¹, Elisa Mengual², Marta Vidorreta¹, Francis Loayza¹, Jaione Irigoyen¹, and Maria Pastor^1
1Neuroscience, CIMA - University of Navarra, Pamplona, Navarra, Spain, ²Anatomy, Medical School - University of Navarra, Pamplona, Navarra, Spain

Arterial spin labeled (ASL) perfusion MRI offers the possibility of measuring cerebral blood flow (CBF) and assessing resting state functional connectivity (FC) by means of evaluating the spatial patterns of synchronous spontaneous fluctuations in the CBF time series. Recent work has shown that ASL FC has statistical power comparable to that of BOLD FC and could provide a better characterization of low frequency fluctuations than BOLD. In this work, ASL FC has been used to study FC connectivity of the STN in healthy controls and to evaluate STN FC alterations in Parkinson’s disease.

Bin Chen¹, Kai Zhao², Bo Li³, Wenchao Cai², Xiaoying Wang^1,2, Jue Zhang^1,3, and Jing Fang^1,3
1Academy for Advanced Interdisciplinary Studies, Peking University, Beijing, China, ²Dept. of Radiology, Peking University First Hospital, Beijing, China, ³College of Engineering, Peking University, Beijing, China

Compressed sensing (CS) technique was introduced in this study in order to improve the temporal resolution in Dynamic Contrast-enhanced MR imaging. A real pulse sequence was modified with CS MRI scheme with undersampling in phase encoding, resulting in 2x, 3x, 4x, and 8x accelerations. Actual high temporal resolution was achieved, which could contribute more creditable quantitative renal perfusion measurements. In addition, the contrast-to-noise ratio of reconstructed images was effectively improved.

Ashley M. Stokes^1,2 and Christopher C. Quarles^1,2
1Institute of Imaging Science, Vanderbilt University, Nashville, TN, United States, ²Radiology and Radiological Sciences, Vanderbilt University, Nashville, TN, United States

A combined spin- and gradient-echo (SAGE) EPI method was applied in a rat glioma model to obtain absolute ΔR₂, ΔR₂* and ΔR₁ curves, thereby permitting derivation of dynamic susceptibility contrast (DSC) parameters. T₁-insensitive estimates of R₂ and R₂^* were acquired with SAGE and compared to conventional GE and SE DSC parameters, as well as independently measured R₂ and R₂^*. The normal brain GE and SAGE ΔR₂^* curves produced similar relative cerebral blood volume (rCBV) values, while tumor leakage effects led to substantially underestimated rCBV values from the GE ΔR₂^* curve. Dynamic SAGE acquisition can provide improved perfusion and permeability information.

Jack T. Skinner^1,2, Ryan K. Robison^2,3, and Christopher C. Quarles^1,2
1Radiology and Radiological Sciences, Vanderbilt University, Nashville, TN, United States, ²Institute of Imaging Science, Vanderbilt University, Nashville, TN, United States, ³Philips Healthcare, Highland Heights, OH, United States

Previous single-echo techniques for perfusion imaging have been extended to multiple-echo sequences to acquire more accurate estimates of R₂* with contrast injection. Recently, a spin- and gradient-echo (SAGE) acquisition, for simultaneous R₂ and R₂* mapping, was introduced. This current study implements a single-shot SAGE sequence that utilizes SENSE parallel imaging and partial k-space sampling, allowing short echo times (< 10ms) suitable for AIF characterization. Validation of the current protocol with conventional multiple-echo acquisitions was carried out in phantoms and in the brain. The suggested single-shot SAGE acquisition may provide a robust method for quantitative perfusion imaging in brain tumors.

Alexander Brost¹, Heiko Schmiedeskamp¹, Matus Straka¹, Jalal Andre², and Roland Bammer^1
1Center for Quantitative Neuroimaging, Stanford University, Stanford, CA, United States, ²Department of Radiology, University of Washington, Seattle, WA, United States

The spin- and gradient-echo (SAGE) EPI sequence was developed to estimate cerebral blood flow (CBF) and cerebral blood volume (CBV). The main purpose of the sequence is to provide T1-independent perfusion-weighted imaging maps. Gadolinium-based contrast agents cause T1-shortening, in particular, when contrast agent leaks into the extravascular-extracellular space. Leakage correction using SAGE EPI requires a separately acquired pre-bolus T1 map. To estimate the effects of an incorrect T1 map, we simulated the effects using a pharmacokinetic model. Leakage correction could be applied to multi-echo data that lack a properly determined prebolus T1 map without significantly compromising CBV and MTT estimates.

Emelie Lindgren¹, Ronnie Wirestam¹, Karin Markenroth Bloch^1,2, André Ahlgren¹, Matthias J.P. van Osch³, Danielle van Westen⁴, Yulia Surova^5,6, Freddy Ståhlberg^1,7, and Linda Knutsson^1
1Department of Medical Radiation Physics, Lund University, Lund, Sweden, ²Clinical science, Philips, Lund, Sweden, ³C.J.Gorter Center for high field MRI, Department of Radiology, LUMC, Leiden, Netherlands, ⁴Center for Medical Imaging and Physiology, Skane University Hospital Lund, Lund, Sweden, ⁵Department of Clinical Sciences, Lund University, Lund, Sweden, ⁶Department of Neurology, Skåne University Hospital, Lund, Sweden,⁷Lund University Bioimaging Centrum, Lund University, Lund, Sweden

Quantification of CBF in absolute terms is relevant in some clinical applications. However, absolute CBF obtained by dynamic susceptibility contrast MRI (DSC-MRI) is usually overestimated. One solution is to calibrate DSC-MRI CBF using a complementary CBV method. In this study, T1-based CBV (Bookend) and VASO were used to calibrate CBF from DSC-MRI, and results from calibrations were compared with CBF from pseudo-continuous ASL (pCASL). Both calibration methods showed absolute grey matter CBF values which were comparable with corresponding CBF values from pCASL.

Ming-Ching Chang¹, Sandeep Narendra Gupta¹, Laura I. Sacolick², Clare Tempany-Afdhal³, Fiona Fennessy³, and Ehud J. Schmidt^3
1GE Global Research Center, Niskayuna, NY, United States, ²GE Global Research Center, Munich, Germany, ³Radiology, Brigham and Women's Hospital, Boston, MA, United States

Dynamic Contrast Enhanced MRI (DCE-MRI) is used for the assessment of tumor vascular properties with application to prostate cancer staging and treatment monitoring. Analysis of DCE-MRI requires knowledge of pre-contrast T1 maps. T1 mapping using Variable Flip Angle imaging is inaccurate due to B1 inhomogeneity. We present improved T1 mapping and PK quantification in prostate DCE-MRI at 3T by incorporating B1 inhomogeneity correction using the Bloch-Siegert B1 mapping method. We validated the method on 26 subjects and demonstrated good T1 quantification and better PK maps by incorporating B1 correction into DCE-MRI quantification.

Brandon Zanette¹, Greg O. Cron^2,3, Thanh B. Nguyen^2,3, Mark E. Schweitzer^2,3, and Ian G. Cameron^2,3
1Carleton University, Ottawa, Ontario, Canada, ²Ottawa Hospital Research Institute, Ottawa, Ontario, Canada, ³Radiology, University of Ottawa, Ottawa, Ontario, Canada

For quantitative DCE-MRI, Gd concentration is conventionally estimated via a pre-DCE T1 map using variable flip angles (VFA), which allows conversion of DCE signal to concentration. This approach, however, is sensitive to B1 variations. Two techniques may potentially reduce errors: B1 mapping and the “Bookend Method” (involving a post-DCE T1 map). The results of this study show that, when estimating concentration in tissue for quantitative DCE-MRI, the best results will be achieved using the Bookend Method and accurate T1 maps. B1-corrected VFA does not consistently deliver enough T1 accuracy, indicating that further improvements in T1 mapping techniques are required.

Neil P. Jerome¹, James A. d'Arcy¹, Matthew R. Orton¹, Thorsten Feiweier², Dow-Mu Koh³, Martin O. Leach¹, and David John Collins^1
1Radiotherapy & Imaging, The Institute of Cancer Research, Sutton, Surrey, United Kingdom, ²Imaging & Therapy Division, Siemens AG, Healthcare Sector, Erlangen, Germany, ³Department of Radiology, Royal Marsden Hospital, Sutton, Surrey, United Kingdom

Diffusion-weighted MRI in the body must account for a microcirculation fraction, separate to self-diffusion, within imaging voxels. Explicit control of diffusion pulse length and delay allows reproducible application of diffusion gradients with varying echo times; calculation of mono-exponential T2 estimates with applied gradients of b=0 and b=200 s/mm² shows significant changes observed in liver, kidney and spleen. This suggests that the microcirculation component, with its own distinct T2, is being removed, allowing the generation of flow-free T2 maps more robustly estimating tissue T2s. This approach enables appropriate b-value choices when considering diffusion models that include or exclude microcirculation contribution.

Natenael B. Semmineh¹, Junzhong Xu¹, and Christopher C. Quarles^1
1Institute of Imaging Science, Vanderbilt University, Nashville, TN, United States

The use of DSC-MRI in tumors can be confounded by the assumption that a linear relationship, with a spatially uniform rate constant termed the vascular susceptibility calibration factor, exists between the contrast agent concentration and the measured transverse relaxation rate change. Using simulations we demonstrate that varying vascular morphology parameters can increase or decrease the vascular susceptibility calibration factors found in tumor-like vessel trees, leading to the overestimation or underestimation of the tumor blood volume.

Peter S. LaViolette¹, Mitchell Daun², Alexander D. Cohen³, Jennifer Connelley², and Kathleen M. Schmainda^1
1Radiology, Medical College of Wisconsin, Milwaukee, WI, United States, ²Neurology, Medical College of Wisconsin, Milwaukee, WI, United States, ³Biophysics, Medical College of Wisconsin, Milwaukee, WI, United States

Independent component analysis applied to DSC perfusion data allows the separation of arterial and venous perfusion. This study varies the number of ICA components modeled to determine which number generates the most repeatable arterial and venous maps. 68 patients with 5 repeated scans on the same scanner were compared. We conclude the modeling 3 components reliably generates arterial and venous maps.

Madhura Ingalhalikar¹, Bilwaj Gaonkar¹, Alex R. Smith¹, Robert T. Schultz², and Ragini Verma^1
1Section of Biomedical Image Analysis, Department of Radiology, University of Pennsylvania, Philadelphia, PA, United States, ²Center for Autism Research, Childrens Hospital of Philadelphia, Philadelphia, PA, United States

The study employs a novel method using an analytic approach to permutation tests on support vector machines for computing statistical significance maps on connectivity matrices. Permutation tests are critical for interpreting SVM output for high dimensional data. However performing these tests is time consuming and computationally expensive. However, the analytical approximation to these tests can yield the results quickly. We apply this method to investigate the differences between patients with autism and typically developing controls based on their structural connectivity networks. We find that patients with autism show lower connectivity mainly in the connections initiating from temporal regions like fusiform gyrus and insula as well between fronto-parietal tracts.

Matt Nathan Gwilliam¹, David John Collins¹, Martin O. Leach¹, Helen Young², and Matthew R. Orton^1
1CRUK and EPSRC Imaging Centre, Institute of Cancer Research, Sutton, London, United Kingdom, ²Astrazeneca, Macclesfield, Cheshire, United Kingdom

DCE-MRI is widely used in clinical trials of antiangiogenic and vascular disrupting agents in the assessment of treatment response. Fitting pharmacokinetic models gives well defined measures of vascular function, but difficulties in obtaining a patient-specific AIF can cause additional errors on PK metrics. This abstract demonstrates that a model-free PCA approach is as sensitive to DCE-MRI treatment changes as model-based PK measures in a group of patients receiving a VEGF inhibitor.

Marijn van Stralen¹, Hanke J. Schalkx², Harriët W. Mulder¹, Kenneth Coenegrachts³, Maarten S. van Leeuwen², Wouter B. Veldhuis², and Josien P.W. Pluim^1
1Image Sciences Institute, UMC Utrecht, Utrecht, Utrecht, Netherlands, ²Radiology, UMC Utrecht, Utrecht, Utrecht, Netherlands, ³Radiology, AZ St.‐Jan Brugge‐Oostende AV, Brugge, West-Vlaanderen, Belgium

Recent developments have enabled high-resolution dynamic contrast-enhanced MRI (DCE-MRI) of the entire liver. However, breathing and other organ motion hampers the current clinical use of DCE-MRI for perfusion analysis. Moreover, intensity change as a result of contrast perfusion challenges existing motion correction techniques. We propose a temporally consistent motion compensation method based on the integrated 3D + time alignment of the gradient magnitude of the image intensity. It is shown that time-consistent non-rigid registration outperforms temporal independent registration by measuring surface errors on manual liver segmentations. These results promise improved perfusion analysis (eg. pharmacokinetic modeling) of abdominal organs using DCE-MRI.

Jonathan P. Dyke¹, Pascal Spincemaille¹, Martin R. Prince¹, and Krishna Juluru^1
1Radiology, Weill Cornell Medical College, NY, NY, United States

Monitoring total liver volume is of importance to assess regeneration post-resection. Normal liver volume is also of prognostic importance in subject with tumor burden and inflammatory disease. A technique is presented to segment normal liver by calculating a voxel wise Pearson’s R2 correlation coefficient with slice specific Gd-EOB-DTPA (Eovist) dynamic contrast enhanced DCE-MRI uptake curves. Significant correlation was found with DCE segmented liver volume and manual tracing methods. Future application of this technique may be found in other organ systems and patient specific monitoring of organ volume over time.

Jihun Oh¹, Diego Martin², and Xiaoping P. Hu^3
1School of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, Georgia, United States, ²Department of Medical Imaging, University of Arizona, Tucson, Arizona, United States, ³Department of Biomedical Engineering, Georgia Institute of Technology and Emory University, Atlanta, Georgia, United States

This paper describes our work of using features derived from contrast-enhanced liver MR images for providing a quantitative assessment of chronic liver disease severity. We first examined the mean slope of contrast uptake in hepatobiliary phase and demonstrated that it is significantly correlated with fibrosis score. We also examined several texture measures in equilibrium phase using Gabor filtering and grey level co-occurrence matrix and built a supervised maximum a posteriori classifier using these features to predict the disease severity. The classifier was evaluated by cross-validation and shown to be highly robust in predicting fibrosis score.

Madhava P. Aryal^1,2, Tavarekere N. Nagaraja², Kelly A. Keenan², Hassan Bagher-Ebadian², Swayamprava Panda², Stephen Brown², and James R. Ewing^1,2
1Physics, Oakland University, Rochester, MI, United States, ²Henry Ford Hospital, Detroit, MI, United States

In this study, Logan plot graphical approach was applied to dynamic contrast enhanced MRI(DCE-MRI) data to estimate the distribution volume(VD) in a rat model of cerebral glioma. Test-retest VD values for 18 animal studies were quite stable. The test group mean of VD (7.94%) moved downward to (7.21%) in the retest group, but not significantly (p=0.21). The combined sample mean(±standard deviation) of VD was; (7.58±2.33)%. Also, the estimated VD showed a strong correlation to corresponding Patlak plot estimate (r=0.93, p<0.001). The mean tumor cellularity count was 2441 ± 850. Tumor cellularity and VD were highly correlated (r =0.76, p < 0.01). Thus, the Logan plot graphical approach can be a useful tool on DCE-MRI also for the diagnosis and evaluation of therapeutic response in tumors.

Dennis Lai Hong Cheong¹, Bo Zhang¹, Soo Chin Lee², and Thian C. Ng^1,3
1Clinical Imaging Research Center, SBIC/A*STAR & National University of Singapore, Singapore, Singapore, ²Department of Haematology-Oncology, National University Health System, Singapore, Singapore, ³Department of Diagnostic Radiology, National University of Singapore, Singapore, Singapore

Model based tracer kinetic analysis of DCE MRI relies on some curve fitting process. The reliability of voxel level measured parameter depends on two main criteria: sufficient contrast to noise ratio (CNR) and successful curve fitting. We present a method to filter out unreliable voxels with low CNR and poor quality of fit based on fraction of modeling information (FMI). We found that CNR and FMI are useful measures for rejecting voxels with unreliable parameter values that were either due to a low CNR or a poor curve fitting. This is a part of our DCE-MRI project on breast tumors.

Georges Hankov¹, Jill Fredrickson¹, Gregory Ferl¹, David Clayton¹, Alexandre Coimbra², and Alexandre de Crespigny^2
1Genentech Inc., South San Francisco, California, United States, ²Genentech Inc, South San Francisco, California, United States

The use of DCE-MRI in multicenter therapeutic trials requires standardization, streamlined processing, and a focus on reproducibility. The purpose of this work was to assess the necessity of modeling the AIF and estimating the time lag between tissue and AIF uptake, and also evaluated the reproducibility of the Signal Difference [SD] method, using simulations and patient data. We found that modeling the AIF results in underestimation of the parameters, especially plasma volume, vp. Estimating the time lag increases vp estimates, without altering the other kinetic parameters. The simpler [SD] method was slightly more reproducible than the regular apparent concentration method.

Guido H. Jajamovich¹, Claudia Calcagno¹, Hadrien A. Dyvorne¹, Henry Rusinek², Shimon Aronhime¹, and Bachir Taouli^1
1Mount Sinai School of Medicine, New York, NY, United States, ²NYU School of Medicine, New York, NY, United States

A reconstructed arterial input function (AIF) using prebolus injection of a low dose gadolinium contrast (1.3 mL) was quantified and compared with the measured AIF after the injection of the main bolus in DCE-MRI of the liver. The curve shape reconstructed using pre-bolus injection was observed to be of better quality than the measured main bolus AIF in the majority of cases. The quantified AIF using pre-bolus showed better parameter reproducibility and significantly higher AUC60 (area under the curve at 60 sec), peak and upslope and significantly lower FWHM (full width at half maximum) compared to main bolus AIF.

Elias Kellner¹, Irina Mader², Marco Reisert¹, and Valerij G. Kiselev^1
1Department of Radiology, University Hospital Freiburg, Freiburg, Germany, ²Section of Neuroradiology, University Hospital Freiburg, Freiburg, Germany

We recently presented a method for a quantitative determination of the arterial input function for DSC MRI. The method was verified in the pig model. In this work, we present measurements in patients with carotid stenosis and discuss the quantitative perfusion parameters obtained with our extended technique. Obtained cerebral blood flow is compared with the whole-brain flow evaluated using phase contrast MRI measurements in large arteries. Discussion focuses on issues, which are crucial for correct perfusion evaluation, such as vascular transport effects and water relaxation properties in tissue.

Tammo Rukat^1,2, Simon Walker-Samuel³, and Stefan A. Reinsberg^1
1Department of Physics and Astronomy, University of British Columbia, Vancouver, BC, Canada, ²Institut für Physik, Humboldt Universität zu Berlin, Berlin, Germany, ³Centre for Advanced Biomedical Imaging, University College London, London, United Kingdom

The reliability of small animal DCE-MRI data analysis with quantitative pharmakokinetic models depends strongly on the time course of contrast agent (CA) administration, known as arterial input function (AIF). In this study the dependence of the intrinsic limits of a parameter predictability on the width of the AIF-peak (i.e. the speed of CA administration) is quantified to provide investigators with a tool to determine reasonable limits for CA administration protocols.

Denis Peruzzo¹, Danilo Benozzo¹, Gianluigi Pillonetto¹, and Alessandra Bertoldo^1
1Department of Information Engineering, University of Padova, Padova, PD, Italy

We present fast NSR, an optimization of the original Non-linear Stochastic Regularization algorithm to quantify the residue function and the CBF in DSC-MRI. Fast NSR introduces a preliminary step that allows to overcome the limits in the original NSR implementation, such as the sensitivity to the starting points and the required computational time. In the preliminary step the optimal starting points and the stochastic component of the residue function are computed for each voxel. Fast NSR elaborates a whole subject in a couple of hours and is now suitable for the analysis of large data sets and in clinical context.

Natenael B. Semmineh¹, Junzhong Xu¹, Jerry Boxerman², Gary W. Delaney³, and Christopher C. Quarles^1
1Institute of Imaging Science, Vanderbilt University, Nashville, TN, United States, ²Alpert Medical School of Brown University, Providence, Rhode Island, United States, ³CSIRO Mathematical and Information Sciences, Clayton South, Vic, Australia

Brain tumor DSC-MRI studies can be confounded by extravascular T2* effects that occur when the contrast agent extravasates. The resulting signals are consequently influenced by the extravascular compartmentalization of the contrast agent and therefore may depend on the spatial distribution of tumor cells within tissue. Using simulations we demonstrate that the dependence of ∆R2* on tissue CA concentration is significantly influenced by to cell density. We also show that the spacing of cells and cellular cluster distribution within a voxel can alter ∆R2* values significantly for cell spacing variations on the order of the cell size.

Keiko Miyazaki¹, David John Collins¹, James A. d'Arcy¹, Dow-Mu Koh², Anwar R. Padhani³, Martin O. Leach¹, and Matthew R. Orton^1
1CR-UK and EPSRC Cancer Imaging Centre, The Institute of Cancer Research, Sutton, Surrey, United Kingdom, ²Department of Diagnostic Radiology, The Royal Marsden NHS Foundation Trust, Sutton, Surrey, United Kingdom, ³Paul Strickland Scanner Centre, Mount Vernon Hospital, Northwood, Middlesex, United Kingdom

A recently published DSC-MRI model that includes first-pass, re-circulation and leakage components was evaluated on serially acquired brain tumour data after pre-loading with Gd-contrast. Relationships between DSC-MRI and DCE-MRI kinetic parameters were explored. The tertiary amplitude parameter a3 of the novel model, which characterizes the equilibrium phase (where BBB is intact) or leakage (where BBB is disrupted), was found to correlate strongly with the interstitial space. Results suggest it may be possible to obtain information on tumour interstitial space, in addition to blood volume, blood flow and mean transit time, from a single DSC-MRI measurement by employing this novel model.

Cecile RLPN Jeukens¹, Harm J. van de Haar¹, Jacobus F.A. Jansen¹, Paul A.M. Hofman¹, C. Eleana Zhang², Julie Staals², Saartje Burgmans³, Frans RJ Verhey³, Robert J. van Oostenbrugge², and Walter H. Backes^1
1Radiology, Maastricht University Medical Centre, Maastricht, Limburg, Netherlands, ²Neurology, Maastricht University Medical Centre, Maastricht, Limburg, Netherlands, ³Alzheimer Centre Maastricht, Maastricht University Medical Centre, Maastricht, Limburg, Netherlands

Using Monte Carlo simulations, we aim to determine the optimal sampling settings for reliable permeability (K_i) and plasma volume (v_p) determination: 1) a single temporal resolution, i.e a single sample frequency (t), versus dual temporal resolution, i.e., a high initial sample frequency (t1=2.5s for 1:30min) followed by a lower sample frequency (t2=10s—3min), and 2) the scan duration. We find that using a dual temporal resolution protocol and a scan duration of 15-20 minutes allows a reliable K_i and v_p determination (5%/95% CI = -/+ 10%) for K_i and v_p values down to K_i=0.002 min^-1 and v_p = 0.01.

Ting-Ting Chang¹, Alex M. Wong², Feng-Xian Yan¹, Yu-Shi Lin¹, and Ho-Ling Liu^1,2
1Department of Medical Imaging and Radiological Sciences, Chang Gung University, Kwei-Shan, Tao-Yuan, Taiwan, ²Department of Medical Imaging and Intervention, Chang Gung Memorial Hospital, Kwei-Shan, Tao-Yuan, Taiwan

This study proposed to combine dynamic contrast enhanced (DCE) and arterial spin labeling (ASL) MRI for assessing tumor vessel permeability in a group of 11 pediatric patients. After co-registration of measured parametric maps, permeability-surface-product (PS) was calculated by Ktrans from DCE-MRI and CBF from ASL on voxel-by-voxel basis. The results showed that for small Ktrans values they were approximately equal to the resulted PS values. When Ktrans values were greater, they became increasingly underestimated than the PS values. The largest discrepancy between Ktrans and PS in this study was 13% in a patient with mean tumor Ktrans of 0.10 min-1.

Septian Hartono^1,2, Tong San Koh¹, Lei Zhou³, Quan Sing Ng¹, Puor Sherng Lee¹, Kai-Hsiang Chuang⁴, Yock Young Dan³, Laurent Martarello⁵, and Choon Hua Thng^1
1National Cancer Centre Singapore, Singapore, NA, Singapore, ²Nanyang Technological University, Singapore, NA, Singapore, ³National University Health System, Singapore, NA, Singapore, ⁴Singapore Bioimaging Consortium, Singapore, NA, Singapore, ⁵Roche Translational Medicine Hub, Singapore, NA, Singapore

We carried out a pilot study using DCE-MRI to assess liver cirrhosis in mice. Elevated values of extravasation parameters such as interstitial volume and vascular permeability were found in the cirrhotic mice, which was consistent to the capillarisation of liver sinusoidal endothelium found in liver cirrhosis.

Chih-Feng Chen¹, Ho-Ling Liu^2,3, Yu-Jie Huang⁴, Yuan-Hsiung Tsai¹, and Hsu-Huei Weng^1
1Department of Radiology, Chang Gung Memorial Hospital, Chiayi branch, and Chang Gung University of Science and Technology, Chiayi, Taiwan, ²Department of Medical Imaging and Radiological Sciences, Chang Gung University, Taoyuan, Taiwan, ³Healthy Aging Research Center, Chang Gung University, Taoyuan, Taiwan, ⁴Kaohsiung Chang Gung Memorial Hospital and Chang Gung University College of Medicine, Kaohsiung, Taiwan

This study assessed heterogeneity changes in peak enhancement ratio (PER) from DCE-MRI by calculating entropy in brain tumors before and after radiotherapy. There were 11 patients in the non-responder group and 12 patients in the responder group. The post- to pre-radiotherapy ratio of entropy was defined as therapeutic entropy ratio (TER). TER was significantly higher in the non-responder (1.14¡Ó0.14) than that in the responder (0.86¡Ó0.1). The corresponding area under the receiver operating characteristic curve was 0.98. In conclusion, calculating the heterogeneity changes in tumors via entropy could help diagnose response to therapy in clinical.

Thomas Christen¹, Deqiang Qiu¹, Wendy W. Ni¹, Michael E. Moseley¹, and Greg Zaharchuk^1
1Radiology, Stanford University, Stanford, California, United States

In this work, we acquired resting state BOLD image data in normal subjects and patients with cerebrovascular disease to study brain perfusion. The signal was analysed in two ways: (1) we looked at the amplitude of the fluctuations after high pass filtering of the signal, and (2) we looked at the correlation between the MR signal in major blood vessels (such as the sagittal sinus) and the rest of the brain. These maps demonstrated similar characteristics to CBF and CBV but did not require contrast.

Feng-Xian Yan¹, Tsong-Hai Lee², Pin-Hsun Huang¹, Kuo-Lun Huang², Ho-Fai Wong³, and Ho-Ling Liu^1,3
1Department of Medical Imaging and Radiological Sciences, Chang Gung University, Taoyuan, Taiwan, ²Department of Neurology and Stroke Center, Chang Gung Memorial Hospital, and Chang Gung University College of Medicine, Taoyuan, Taiwan, ³Department of Medical Imaging and Intervention, Chang Gung Memorial Hospital, Taoyuan, Taiwan

This study aimed to investigate whether the functional connectivity (FC) derived from BOLD signal fluctuation is correlated with deficient blood supply in patients with unilateral internal carotid artery (ICA) stenosis. Thirty-eight patients underwent resting-state (RS) fMRI and DSC perfusion MRI scans. This study found that the prolonged perfusion delay time (Tmax) was significantly correlated (p < 0.05) with reduced default-mode network connectivity in the inferior parietal cortex ipsilateral to the ICA stenosis without apparent CBF impairments. Further studies are required to understand whether such correlation is originated from neuronal or purely hemodynamic variations.

Jing Wang¹, Yudong Zhang², Jue Zhang¹, Xiaoying Wang³, and Jing Fang^4
1Academy for Advanced Interdisciplinary Studies, Peking Unversity, Beijing, China, ²Department of Radiology, The First Affiliated Hospital With Nanjing Medical University, Nanjing, China, Beijing, China, ³Department of Radiology, Peking University First Hospital, Beijing, China, ⁴Peking University, Beijing, China

This study developed a new approach for noninvasive renal perfusion and glomerular filtration rate mapping based on PASL using variable TE acquisitions on clinical 3T MR. There were six rabbits included in the initial experiments and each of them accepted 6 different TEs ASL scans. Based on the proposed method, the rabbit RBF, blood R2* and GFR maps were obtained simultaneously by a two-compartment model fitting. Furthermore, the non-contrast GFR values were compared with DCE-MRI results to explore the feasibility of VTE-ASL as a noninvasive assessment of renal status in a single examination.

Marjorie Villien^1,2, Do Kien N Guyen¹, Julien Bouvier¹, Cedric Mendoza³, Sylvie Grand³, Louise Fanchon¹, Emmanuel Luc Barbier⁴, Irène Troprès⁵, Jean-François Le Bas³, Alexandre Krainik³, and Jan M. Warnking^1
1Grenoble Institut of Neurosciences, INSERM, Grenoble, France, ²Massachusetts General Hospital, Harvard, Charlestown, Massachusetts, United States, ³Clinique universitaire de neuroradiologie et d'IRM, CHU Grenoble, Grenoble, France, ⁴Grenoble Institut of Neurosciences, INSERM U836, Grenoble, France, ⁵SFR1, Université Joseph Fourier, Grenoble, France

The goals of this study were to assess the reproducibility of Dynamic Susceptibility Contrast (DSC) MRI and Perfusion-CT (PCT) and to compare the CBF estimates obtained with DSC, PCT, and Arterial Spin Labeling (ASL). We included patients with cerebral tumor, routinely monitored with these imaging methods. We observed good agreement between DSC and PCT, and poor correlation between DSC and ASL. A more detailed analysis of the present data could yield insights into the respective strengths and weaknesses of each method depending on the vascular properties.

Elias Kellner¹, Irina Mader², Michal Mix³, Marco Reisert¹, Katharina Förster⁴, Thao Nguyen-Thanh², Daniel Nico Splitthoff¹, Peter Gall¹, and Valerij G. Kiselev^1
1Department of Radiology, University Hospital Freiburg, Freiburg, Germany, ²Section of Neuroradiology, University Hospital Freiburg, Freiburg, Germany, ³3Department of Nuclear Medicine, University Hospital Freiburg, Freiburg, Germany, ⁴4Department of Cardiovascular Surgery, University Hospital Freiburg, Freiburg, Germany

We recently presented a method for the quantitative determination of the arterial input function for bolus tracking perfusion imaging. Here we present evaluation of cerebral blood flow obtained with the new method in comparison with 15H2O Positron-Emission-Tomography (PET) in the porcine model (N=13). In contrast to previous studies, we do not employ any adjustable parameters. The CBF values obtained with both methods correlate significantly, but MRI CBF is systematically lower. Present results create a quantitative basis for discussing the role of delay and dispersion in as well as the principal difference between PET and MRI perfusion measurements.