Hui Zhang¹, Tim B Dyrby², and Daniel C Alexander^1
1Department of Computer Science & Centre for Medical Image Computing, University College London, London, United Kingdom, ²Danish Research Center for Magnetic Resonance, Copenhagen University Hospital, Hvidovre, Denmark

This work presents a technique for axon diameter mapping in fiber crossing regions. Earlier works assume a model of axons with a single dominant orientation. Although valid for many white matter areas, this approach prevents the mapping of regions with crossing fibers that are widespread in the brain and exhibit multiple dominant orientations. A solution to this problem is prerequisite for mapping axon diameter mapping over whole brain. We address this challenge with a crossing-fiber model that enables the simultaneous estimation of crossing configuration and microstructural features. We demonstrate the approach with synthetic and ex vivo brain imaging data.

Dmitry S. Novikov¹, and Els Fieremans^1
1Radiology, NYU School of Medicine, New York, NY, United States

Tissues are often viewed as tight packings of practically impermeable cells. While the intra- and extracellular water fractions can be determined using a straightforward biexponential fit, quantifying the cell size distribution and packing geometry with low-q diffusion metrics has been a notoriously hard problem. We develop a novel analytical approach for relating the diffusivity in the extra-axonal space to the volume fraction of axons and their size heterogeneity that agrees with Monte Carlo simulations. Our formalism quantitatively distinguishes between demyelination and axonal loss in the diffusion coefficient of the extra-axonal space transverse to the fibers.

Takayuki Obata¹, Jeff Kershaw², Daigo Kuroiwa¹, Sayaka Shibata², Abe Yoichiro³, Masato Yasui³, and Ichio Aoki^2
1Center for Charged Particle Therapy, National Institute of Radiological Sciences (NIRS), Chiba, Chiba, Japan, ²Molecular Imaging Center, NIRS, ³Department of Pharmacology, Keio University, Tokyo, Japan

We performed multi-b-value and multi-diffusion-time DWI on AQP4-nonexpressed (noAQ) and AQP4-expressed (AQ) cells to investigate the effect that CMP has on diffusion-weighted magnetic resonance signal. Significant signal differences between the noAQ and AQ cells were observed only at high b-values. Diffusion-time dependent ADC changes were also observed, but there were no significant differences between the noAQ and AQ cell samples.

Frank Q Ye¹, David A Leopold¹, Mustafa Irfanoglu², Carlo Pierpaoli², and Afonso C Silva^3
1NIMH, National Institutes of Health, Bethesda, Maryland, United States, ²NICHD, National Institutes of Health, ³NINDS, National Institutes of Health

It is known that axons in the cerebral cortex show distinct organization patterns, which in principle can be mapped by diffusion MRI. Up to now, diffusion imaging has not yet proved its value in study the microstructure of cortical gray matter. In this study, q-ball diffusion data of high spatial resolution and high SNR were acquired from a fixed marmoset brain. Interesting diffusion contrasts are observed in various cortical regions that could potentially be used to delineate cortical layers or to parcellate cortical structures.

Francesca Branzoli¹, Aranee Techawiboonwong², Andrew Webb¹, and Itamar Ronen^1
1C. J. Gorter Center for High Field MRI, Department of Radiology, Leiden University Medical Center, Leiden, Netherlands, ²Department of Electrical Engineering, Mahidol University, Bangkok, Thailand

The phenomenon of diffusion of molecules in biological tissues directly relates to the characteristic length scales of tissue compartments and it has been recognized as a powerful contrast mechanism for deriving information on white matter microstructure. In this study, for the first time the interaction between proton transverse relaxation times (T2) and apparent diffusion coefficients (ADC) was investigated, by means of diffusion-weighted spectroscopy, for both water and metabolites in the human brain at 7T.

Itamar Ronen¹, Robert Rengelink², Ece Ercan¹, and Andrew Webb^1
1C.J. Gorter Center for High Field MRI, Department of Radiology, Leiden University Medical Center, Leiden, Netherlands, ²Department of Physics, Leiden University, Leiden, Netherlands

In this work we investigated the diffusion properties of brain metabolites in a well-organized white matter fiber bundle, i.e. the corpus callosum, at 7T. The results indicate a clear difference in compartmentation among metabolites and raise interesting questions when results are fitted to conventional models of diffusion in given geometries.

Henry H. Ong¹, and Felix W. Wehrli^1
1Laboratory for Structural NMR Imaging, Departement of Radiology, University of Pennsylvania School of Medicine, Philadelphia, PA, United States

Q-space imaging offers potential for indirect assessment of white matter architecture but is complicated by signal from both extra- and intra-cellular spaces (ECS and ICS). Here, we use an empiric method to separate ECS and ICS signals based on different characteristics of hindered and restricted diffusion in white matter tracts of healthy mouse spinal cords. This method circumvents problems with modeling or multi-exponential fitting to distinguish the two signals. The measured mean axon diameter show excellent agreement with histology. The results demonstrate the feasibility of this method to accurately separate and characterize diffusion in the ECS and ICS.

Evren Ozarslan^1,2, Michal E Komlosh^1,2, and Peter J Basser^1
1STBB / PPITS / NICHD, National Institutes of Health, Bethesda, MD, United States, ²Center for Neuroscience and Regenerative Medicine, USUHS, Bethesda, MD, United States

Double pulsed field gradient (double-PFG) MR technique could be an important tool that provides microstructural information from tissue. In this study, orientation information obtained from an independent DTI acquisition was incorporated into a theoretical double-PFG framework enabling accurate measurements of cell size. DTI data, acquired in tandem, may reduce the sampling burden of double-PFG acquisitions when the underlying fiber orientation is known approximately, as is the case in spinal cord. The method is used on a celery data set to estimate the cell size in the vascular bundles of the celery stalk.

Blake Walters¹, and Jae Kim^1
1Thunder Bay Regional Research Institute, Thunder Bay, Ontario, Canada

We use a Monte Carlo simulation of diffusion in a tissue model to analyze healthy/malignant contrast in DWI with oscillating gradient spin echo (OGSE) sequences. Results show that using OGSE's contrast due to the increased nucleus-to-cell volume ratio (NCR) in malignant tissue amounts to up to ~30% of total contrast at high (>= 40 G/cm) gradient strengths. Total contrast is maximized at an optimum OGSE frequency. At high frequencies (>1kHz) contrast is due only to changes in tissue cellularity. The qualitative behaviour of contrast with OGSE frequency can be predicted using a single calculated effective diffusion coefficient.

Daniel Barazany¹, and Yaniv Assaf^1
1Neurobiology, Tel Aviv University, Tel Aviv, Israel

Inversion recovery diffusion tensor imaging (IR-DTI) framework presented in this work shows that integrating different tissue characteristics allow more accurate definition of tissue compartments. IR-DTI framework is applicable, and requires only two DTI scans with different TIs which will provide means to differentiate between different sub-voxel diffusion components. The analysis of multiple IR-DTI dataset is done by a bi-tensor model, where each component has its own T1 and diffusion characteristics. In this work we were able to differentiate between the optic and sciatic nerves, where each attribute distinct T1 and diffusivity characteristics.

Henry H. Ong¹, and Felix W. Wehrli^1
1Laboratory for Structural NMR Imaging, Departement of Radiology, University of Pennsylvania School of Medicine, Philadelphia, PA, United States

Q-space imaging (QSI) offers potential for indirect assessment of axonal architecture (i.e. mean axon diameter) in white matter through examination of the displacement probability density function (d-PDF). A major limitation is that the maximum gradient amplitude available commercially is in-sufficient for the displacement resolution required to accurately study axons. In this work, we developed a subvoxel processing-based interpolation algorithm to increase the apparent displacement resolution and validated it with experimental data from d-PDFs of healthy mouse spinal cords. The results show that this method may be used to interpolate d-PDFs up to a factor of four without loss in accuracy.

Chu-Yu Lee^1,2, Kevin M. Bennett³, and Josef P. Debbins^1,2
1Electrical Engineering, Arizona State University, Tempe, AZ, United States, ²Neuroimaging Research, Barrow Neurological Institute, Phoenix, AZ, United States, ³School of Biological and Health Systems Engineering, Arizona State University, Tempe, AZ, United States

The non-monoexponential DWI decay at high b-values has been attributed to the multiplicity of water diffusion rates, which may provide the information about the water compartmentation. The common way to compute the diffusion rates is through the multi-exponential analysis. The bi-exponential model assumes two diffusion rates. Considering the multiple length scales in tissues, the distributed exponential model makes no assumption about the number of diffusion rates, and can be empirically described by the stretched exponential model (α-DWI). Those fitting models: bi-exponential and α-DWI have a few parameters, and have been shown to correlate with the pathology. However, it remains difficult to pin down the underlying biophysical mechanisms. In addition, the multi-exponential relation is phenomenological, because each diffusion rate is no longer associated with a mono-exponential decay when diffusion time is long (>30 ms in a clinical DWI). In this study, the distribution of ‘effective’ diffusivity of water molecules diffusing in a simulated cell structure was created using Monte Carlo simulation. DW experiments were also simulated, and the DWI signals were fitted by the bi-exponential and α-DWI models. We studied how the fitted parameters: Dfast, Dslow, Vfast (fraction of Dfast) of bi-exponential fit, DDC, α of α-DWI, tracked the distribution of effective diffusivity when the microstructures were changed. This may give insights into the relationship between the phenomenological fitting models and the tissue structure.

Jing-Rebecca Li¹, Denis Grebenkov², Cyril Poupon³, and Denis Le Bihan^3
1Equipe DEFI, INRIA-Saclay, Palaiseau Cedex, France, ²Laboratoire de Physique de la Matiere Condensee, CNRS -- Ecole Polytechnique, PALAISEAU CEDEX, France, ³Neurospin, CEA, Saclay, France

We give an approximate analytical formula for the long time apparent diffusion coefficient (ADC) of the diffusion MRI signal attenuation in two and three dimensions. From the long time ADC measurements before and after cell swelling, we are able to use the analytical formula to accurately and robustly estimate the change in the cellular volume fraction

Gisela E Hagberg¹, Ilgar Mamedov¹, Anthony Powers¹, Michael Beyerlein¹, Hellmut Merkle², Goran Angelovski¹, and Nikos K Logothetis^1
1MPI Biological cybernetics, Tubingen, Germany, Germany, ²LFMI-NINDS, National Institutes of Health, Bethesda, MD, United States

Measurements of the diffusion properties of substances in the cerebral extra-cellular space (ECS) can be used to study drug delivery/clearance and brain tissue structure. Currently used methods have high sensitivity, but are limited to single spatial points or are performed post mortem, or has a limited depth penetration. Here we explore the use of MRI during direct infusion of T1 relaxating agents and mathematical modelling for investigating ECS diffusion.

Hernan Jara¹, Stephan William Anderson¹, Jorge A Soto¹, and Osamu Sakai^1
1Radiology, Boston University Medical Center, Boston, Massachusetts, United States

Purpose: To experimentally validate a correlation time diffusion coefficient (DCT) algorithm that incorporates the main phenomena affecting T1 relaxation: molecular kinetics, paramagnetic and magnetization transfer effects. Materials and Methods: The DCT technique was applied to the brains of eleven subjects of varying age and compared to standard pulsed field diffusion (DPFG) MRI. Results: DCT and DPFG were in good agreement for all patient ages studied. Conclusion: A correlation time diffusion qMRI technique that does not involve pulse-field-gradient diffusion sensitization, thus yielding high spatial resolution and SNR with limited vulnerability to motion artifact, has been developed and validated.

Dang Van Nguyen¹, Denis Grebenkov², Cyril Poupon³, Denis Le Bihan³, and Jing-Rebecca Li^4
1CMAP, Ecole Polytechnique, Palaiseau Cedex, France, ²PMC, Ecole Polytechnique, Palaiseau Cedex, France, ³Neurospin, CEA Saclay, Gif-sur-Yvette cedex, France, ⁴Equipe DEFI, INRIA Saclay, Palaiseau Cedex, France

Diffusion MRI can give useful information on cellular structure and structural change. We show that the effective diffusion tensor obtained by mathematical homogenization theory is a good approximation to the long time apparent diffusion tensor under realistic DMR scanning conditions for both isotropic and anisotropic diffusion and general geometries. The homogenized diffusion tensor is obtained by solving three steady-state Laplace equations, which is a more computationally efficient approach than long time simulation in the time domain, either via Monte-Carlo simulation or numerical solution of the time-dependent Bloch-Torrey PDE.

Evren Ozarslan^1,2, Timothy M Shepherd³, Cheng Guan Koay⁴, Stephen J Blackband⁵, and Peter J Basser^1
1STBB / PPITS / NICHD, National Institutes of Health, Bethesda, MD, United States, ²Center for Neuroscience and Regenerative Medicine, USUHS, Bethesda, MD, United States,³University of California, San Francisco, CA, ⁴University of Wisconsin, Madison, WI, ⁵University of Florida, Gainesville, FL

An anomalous diffusion model, inspired by the theory of diffusion in fractal and disordered media, is used to characterize the temporal scaling (TS) properties of diffusion MRI data acquired with multiple diffusion times. A robust computational procedure was devised to obtain accurate estimates even when the signal falls below the Rician noise floor. The interplay between diffusion anisotropy and TS was investigated. The findings suggest the adequacy of the model as well as the reproducibility of estimates. The TS parameters could be used as new and useful markers of tissue microstructure that could be altered by numerous processes and pathologies.

Darryl McClymont¹, Andrew Mehnert², Adnan Trakic¹, Dominic Kennedy³, and Stuart Crozier^1
1University of Queensland, Brisbane, QLD, Australia, ²Chalmers University of Technology, Sweden, ³Queensland X-Ray, Australia

Recent research suggests that diffusion-weighted MRI can be used to improve the sensitivity and specificity of dynamic contrast-enhanced MRI for the detection of breast cancer. However, to date the methods proposed for determining a representative ADC value for a suspicious lesion are highly varied. We propose an automated method based on the converging squares algorithm, which is a noise-robust minimum finding technique. We present an evaluation of the method for computing a representative ADC. The method is also compared to ensemble averaging of ADC values over the entire lesion and the selection of the global minimum ADC value.

April M Chow¹, Victor Ai², Polly SY Cheung³, Siu Ki Yu¹, and Gladys G Lo^2
1Medical Physics & Research Department, Hong Kong Sanatorium & Hospital, Happy Valley, Hong Kong SAR, China, ²Department of Diagnostic and Interventional Radiology, Hong Kong Sanatorium & Hospital, Happy Valley, Hong Kong SAR, China, ³Breast Care Center, Hong Kong Sanatorium & Hospital, Happy Valley, Hong Kong SAR, China

Diffusion-weighted imaging (DWI) characterizes the random microscopic motion of molecules and enables assessment of tissue microstructure. This technique has been widely used to characterize malignant and benign breast lesions; however, most studies only involve implementation of monoexponential model which assumes free and unrestricted water diffusion. In this study, using intravoxel incoherent motion (IVIM) biexponential analysis, we showed that both molecular water diffusion and blood microcirculation contribute to the alteration in apparent diffusion changes in malignant breast lesions at 3 T. With the capability to quantify the diffusion and perfusion effects separately, IVIM analysis may be valuable for characterizing in malignant breast lesions in vivo non-invasively without the use of contrast agents.

Johannes M Froehlich¹, Maria Triantafyllou¹, Giuseppe Petralia¹, Daniel GQ Chong¹, Peter Vermathen¹, Frederic D Birkhaeuser², Achim Fleischmann³, Urs E Studer², and Harriet C Thoeny^1
1Department of Radiology, University Hospital, Bern, Switzerland, ²Department of Urology, University Hospital, Bern, Switzerland, ³Institute of Pathology, University, Bern, Switzerland

Diffusion-weighted MRI(DW-MRI) may help to increase the diagnostic confidence of lymph node staging in patients with bladder or prostate cancer. In an ongoing prospective clinical trial 87 patients, all with normal sized lymph nodes prior to inclusion, underwent DW-MRI followed by surgical treatment with radical lymphadenectomy. A first prospective reading yielded a sensitivity of 56%, specificity of 64.5%, PPV of 39%, NPV of 78.4% and diagnostic accuracy of 62.1%, respectively. Three independent readers blinded for all clinical information performed a second reading with even slightly improved results. In comparison to morphological imaging DW-MRI improves the diagnostic confidence.

Jan Hansmann¹, Andreas Lemke², Jens Wambsganss¹, Mathias Meyer¹, Miriam Reichert¹, Stefan O. Schoenberg¹, and Ulrike I. Attenberger^1
1Institute of Clinical Radiology and Nuclear Medicine, University Medical Center Mannheim, Mannheim, BW, Germany, ²Computer Assisted Clinical Medicine, University Medical Center Mannheim, Mannheim, BW, Germany

Diffusion kurtosis parameter maps of the abdomen were obtained in 118 consecutive patients at 1.5 and 3 Tesla with 6 b-values (b=0-100-500-1000-1500-2000 s/mm²). Kurtosis values in normal parenchyma and a variety of kidney and liver lesions (cysts, benign and malignant tumors) were assessed and compared using the Wilcoxon-Rang-Sum test. Results show statistically signifcant differences in kurtosis values between normal parenchyma and the assessed lesions. Thus, kurtosis imaging of liver and kidney lesions is feasible at 1.5 and 3 Tesla and allows for a differentiation of various lesions compared to normal parenchmya based on kurtosis values.

Neil Peter Jerome¹, Jessica K R Boult¹, Matthew Orton¹, David J Collins^1,2, Simon P Robinson¹, and Dow-Mu Koh^2
1CRUK-EPSRC Cancer Imaging Centre, The Institute of Cancer Research, Sutton, Surrey, United Kingdom, ²Department of Radiology, Royal Marsden Hospital, Sutton, Surrey, United Kingdom

The superior hardware and development of clinical MR imaging systems make them attractive for preclinical studies; limited signal, and thus spatial resolution, for small animal work can in part be overcome by use of parallel imaging with combinations of vendor coils. We demonstrate the use of a clinical 1.5 T system in acquiring diffusion images of normal rat kidney with isotropic resolution 1.5 mm³, and show sensitivity to changes in diffusion and flow parameters obtained from fitting ADC and IVIM diffusion models following administration of hydralazine (systemic vasodilator), furosemide (loop diuretic), and angiotensin II (systemic vasoconstrictor).

Jinxia Zhu¹, Claudia Lenz¹, Markus Klarhöfer¹, Oliver Bieri¹, Klaus Scheffler^2,3, and Gregor Sommer^4
1Radiological Physics, University of Basel Hospital, Basel, Switzerland, ²MRC Department, MPI for Biological Cybernetics, Tübingen, Germany, ³Department of Neuroimaging and MR-Physics, University of Tübingen, Tübingen, Germany, ⁴Department of Radiology, University of Basel Hospital, Basel, Switzerland

Compared to the traditional mono-exponential model, the application of more complex models (stretched exponential model, diffusional kurtosis model, and bi-exponential model) to analyze the renal diffusion weighted MRI data can result in valuable additional insight into pathological processes, for example, benign cysts and renal cell carcinoma. In this research, we compared these four different models to demonstrate that renal pathologies can be differentiated from healthy tissue by applying higher order diffusion models to DWI data consisting of multiple diffusion weightings.

Gerrit H van de Maat¹, Maarten AD Vente², Johannes FW Nijsen², and Chris JG Bakker^2
1Image Sciences Institute, University Medical Center, Utrecht, Utrecht, Netherlands, ²Department of Radiology and Nuclear Medicine, University Medical Center Utrecht, Netherlands

The influence of holmium-166 loaded microspheres (Ho-PLLA-MS) on diffusion weighted imaging (DWI) was investigated using an ex-vivo pig liver into which a non-radioactive therapeutic amount of Ho-PLLA-MS was administered. Apparent diffusion coefficient (ADC) maps were constructed from DWI data and R2* maps were constructed from multi gradient echo data, prior to and after administration of Ho-PLLA-MS. No significant changes were observed in ADC values at locations where microspheres did lodge in the liver tissue according to the R2* maps. This indicates that treatment response measurements after radioembolization with Ho-PLLA-MS using DWI are not hampered by the presence of the microspheres

Daniel Peter Burke¹, Kaye Williams², John Charles Waterton¹, Muhammed Babur², and James O'Connor^1
1Imaging Sciences, University of Manchester, Manchester, United Kingdom, ²School of Pharmacy, University of Manchester, Manchester, United Kingdom

Cancer drug development requires novel imaging biomarkers to detect response to an intervention and enable treatment efficacy to be gauged at an early time point. The Apparent Diffusion Coefficient (ADC) has been postulated as a biomarker of cell death. The relationship between ADC and necrosis or apoptosis requires investigation to further validate ADC. In preclinical models of LoVo colorectal cancer ADC was found to increase in response to radiotherapy treatment and to be correlated with necrosis but not apoptosis (p=0.001 vs. p=0.09). Chemotherapy was observed to decrease ADC with no correlation between ADC and necrosis or apoptosis (p=0.30 vs. p=0.28).

Martijn Intven¹, Onne Reerink¹, and Marielle E.P. Philippens^1
1Radiation Oncology, UMC Utrecht, Utrecht, Netherlands

In rectal cancer, DW-MRI is increasingly used to select candidates for organ-sparing treatment after neo-adjuvant therapy. Quantitative response evaluation or prediction is based on serial DW-MRI. Assessment of the precision of ADC values is mandatory to distinguish therapy related response from measurement variations. In this study, in 14 patients, the repeatability of DW-MRI in rectal cancer was assessed at 1.5T. A repeatability coefficient for the rectal tumor of 10.0% was found for measurements in one MRI protocol, compared to 6.6% in the reference structure, the prostate central gland. Differences in ADC values of >10% can be interpreted as therapy related response in rectal cancer.

Valentin Hamy¹, Simon Walker-Samuel², David Atkinson¹, and Shonit Punwani^1
1Centre for Medical Imaging, UCL, London, United Kingdom, ²Centre for Advanced Biomedical Imaging, UCL, London, United Kingdom

The problem of apparent diffusion coefficient (ADC) estimation from Rician distributed diffusion-weighted magnetic resonance (DW-MR) data is addressed. The least squares (LS) algorithm, widely used in clinical practice, is known to produce biased estimates as it considers the noise as normally distributed. Maximum likelihood (ML) can provide a more robust alternative. In this study based on prostate cancer DW-MR, we compared LS and ML efficiency, for signal to noise ratios typical of the different types of tissue. The ML approach provided significantly less biased estimates than the LS, potentially allowing better accuracy in prostate cancer grading from MR images.

Juan Parra-Robles¹, and Jim M Wild^1
1Academic Unit of Radiology, University of Sheffield, Sheffield, United Kingdom

In this work, we use histological sections to generate computer models of acinar airways. Using these models, finite element computer simulations of 3He and 129Xe gas diffusion in the lungs are implemented. The results of these simulations are presented here through maps of microscopic magnetization and diffusivity distributions. This approach to the simulation and visualization of the results helps provide a better understanding of the different length scales and diffusion regimes present in lung diffusion MR experiments.

Juan Parra-Robles¹, Xiaojun Xu¹, Martin H Deppe¹, Steven R Parnell¹, Helen Marshall¹, and Jim M Wild^1
1Academic Unit of Radiology, University of Sheffield, Sheffield, United Kingdom

Short-range diffusion of 3He in the lungs is non-Gaussian and the b-dependence of the ADC has been used to extract morphometric information about the geometry of acinar airways. Due to the lower diffusivity of 129Xe, it may be more sensitive than 3He to airway changes caused by emphysema. However, the measurement of ADC at multiple b values is challenging for 129Xe due to the lower SNR and the longer diffusion times required to obtain large b values. In this work we present preliminary results of the measurement of the b-dependent ADC of 129Xe in healthy human volunteers.

Archontis Giannakidis¹, Alexander I Veress², Mustafa Janabi¹, James P O'Neil¹, Osama M Abdullah³, Edward W Hsu³, and Grant T Gullberg^1,4
1Radiotracer Development and Imaging Technology, Lawrence Berkeley National Laboratory, Berkeley, California, United States, ²Department of Mechanical Engineering, University of Washington, Seattle, Washington, United States, ³Bioengineering, University of Utah, Salt Lake City, Utah, United States, ⁴Radiology, University of California San Francisoc, San Francisco, California, United States

DT-MRI delineation of microstructural alterations and temporal evolution of ventricular wall mechanics in hypertensive left ventricular hypertrophy

Yin Wu^1,2, Chao Zou^1,2, Wei Liu^1,2, Wei-Qi Liao^1,2, and Ed X. Wu^3,4
1Paul C. Lauterbur Research Center for Biomedical Imaging, Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Shenzhen, Guangdong, China, ²Key Lab of Health Informatics, Chinese Academy of Sciences, Shenzhen, Guangdong, China, ³Laboratory of Biomedical Imaging and Signal Processing, The University of Hong Kong, Pukfulam, Hong Kong, ⁴Department of Electrical and Electronic Engineering, The University of Hong Kong, Pukfulam, Hong Kong

The free Gaussian process of water diffusion assumed in conventional DTI was found to not apply with increase of diffusion strength. In this study, effect of b-value on detecting myocardial structural alteration was assessed in infarcted rabbit models. Non-monoexponential diffusion manner was confirmed in both infarct and control groups. FA and radial diffusivity were found to alter significantly in infarct and/or adjacent regions, and the sensitivity to detect their statistical differences among all the groups were b-value dependent with the greatest total number of statistical significances occurred at b-values of 750 to 2000 s/mm². The study demonstrated the important effect of diffusion strength on DTI index characterization and emphasized the necessity of optimizing b-value for better monitoring and detecting myocardial structural degradation.

Caleb Robert Dulaney¹, Juebin Huang², Manohar Roda³, Alexander P Auchus², and Judy Rose James^3
1School of Medicine, University of Mississippi Medical Center, Jackson, MS, United States, ²Department of Neurology, University of Mississippi Medical Center, Jackson, MS, United States, ³Department of Radiology, University of Mississippi Medical Center, Jackson, MS, United States

This study explores the ability of DTI to analyze the functional architecture of a normal thigh muscle and explore the changes in DTI parameters from within various muscle compartments. DTI parameters in normal thighs were quantified and significant measureable differences between muscles in different parts of the thigh have been reported. DTI could potentially be used to detect subtle changes in skeletal muscles in diseases such as inflammatory myopathies or muscular dystrophies. These novel imaging markers may shed light on early detection, differential diagnosis, and monitoring of skeletal muscle diseases.

Patrick Hiepe¹, Daniel Güllmar¹, Christian Ros¹, Tobias Siebert², Alexander Gussew¹, Reinhard Rzanny¹, and Jürgen R. Reichenbach^1
1Department of Diagnostic and Interventional Radiology 1, Medical Physics Group, Jena University Hospital, Jena, Thuringia, Germany, ²Institute of Sportscience - Science of Motion, Friedrich-Schiller University, Jena, Thuringia, Germany

Knowledge about the 3D architecture of muscles is important, first, for a deeper understanding of contraction dynamics and muscle deformation, and second, as a prerequisite for the development of realistic finite-element muscle models. DTI provides valuable information about muscle architecture in normal and diseased states. In this work we introduce a framework including an MR compatible ergometer with monitoring and visual feedback options to perform DTI measurements during voluntarily muscle contractions at different joint angles. Data acquisition was performed in the right lower leg of one healthy volunteer where we found in activated muscles significant changes of DTI parameters.

Paul Martin Murphy¹, Nikolaus Szeverenyi¹, Claude Sirlin¹, and Mark Bydder^1
1Radiology, University of California - San Diego, San Diego, CA, United States

In liver, the signal decay of diffusion-weighted MR images has been modeled both as a mono-exponential function, representing apparent diffusion, and as a bi-exponential function, representing the separate attenuation of perfusion. Parameters of these models vary in several hepatic disease processes. Early diffusion theory predicted that signal from small vessel perfusion may decay as a sinc function, rather than as an exponential. We demonstrate this form of damped oscillation in a flow phantom and in liver in vivo. This effect may allow measurement of small vessel flow velocity in hepatic disease if temporal confounders can be excluded as its cause.

S. Andrea Wijtenburg¹, Stephen McGuire², David Sherman³, Laura M Rowland¹, and Peter Kochunov^1
1Maryland Psychiatric Research Center, Department of Psychiatry, University of Maryland School of Medicine, Baltimore, MD, United States, ²Aerospace Medicine Consultation Division, Dayton, OH, United States, ³Neuroradiology, 59th Medical Wing, Lackland AFB, TX, United States

We used a multimodal MR protocol, which combined DTI with ¹H MRS and 3D white matter lesion mapping to ascertain changes in biologically important spectroscopy markers and white matter lesion burden associated with changes in FA values. The data were collected in two samples: middle-aged and elderly healthy adults. Our findings demonstrate that age-related decline in FA values is associated with a decline in NAA concentrations in both middle-aged and elderly individuals.

Sonya Bells¹, Luke Dustan¹, David J McGonigle^1,2, C John Evans¹, and Derek K Jones^1
1School of Psychology, CUBRIC, Cardiff, Wales, United Kingdom, ²School of Biosciences, Cardiff University, Cardiff, Wales, United Kingdom

Tract based spatial statistics provides a unique tool to assess brain connectivity. However, the stability of performance-microstructure measurements and hemispheric asymmetries are unknown and is explored here using bootstrapping (100 permutations, 24 participants (omit 4)). The cross-correlation among all 100 permutations were calculated and have a heterogeneous cross-correlation matrix. Furthermore, significant white matter tracts were very different among permutations demonstrating low stability. On the other hand, hemispheric symmetric measurements were far more stable. Importantly, we showed using a resampling method that for three different behavioural measures the stability for a literature average number of participants (N=20) was low.

Shir Hofstetter¹, Ido Tavor¹, Shimrit Tzur-Moryosef¹, Tamar Blumenfeld Katzir¹, and Yaniv Assaf^1
1Tel Aviv University, Tel Aviv, Israel, Israel

Previous studies in our lab found changes in DTI parameters after short term learning task in humans and rats, mainly in the hippocampus. We investigated whether changes in the fornix accompany these modifications, using TBSS. Changes in FA and MD were found in humans after 2 hours of learning task, and reduction in MD in rats undergoing short training in Morris water maze. Additionally, correlation of MD change in the hippocampus with changes in the fornix was detected in all groups. This work provides first indication of short term white matter plasticity that can be detected with DTI.

Keigo Shimoji¹, Shigeki Aoki¹, Osamu Abe², Takanori Uka³, Yoshifumi Tamura⁴, Koji Kamagata¹, Koichi Asahi¹, Masaaki Hori¹, Atsushi Nakanishi¹, and Yasmin Hasina^1
1Department of Radiology, Juntendo University Graduate School of Medicine, Tokyo, Japan, ²Department of Radiology, Nihon University Graduate School of Medicine, Tokyo, Japan, ³Department of Neurophysiology, Juntendo University Graduate School of Medicine, Tokyo, Japan, ⁴Department of Metabolism and Endocrinology, Juntendo University Graduate School of Medicine, Tokyo, Japan

We explore the regional pattern of white matter alteration in metabolic subjects. In addition, we investigate whether white matter alteration was related to BMI. TBSS analysis revealed significantly lower FA values in metabolic subjects compared to normal control subjects in the part of the right external capsule, the entire corpus callosum and part of deep white matter of the right frontal lobe. By using tract-specific analysis, the mean FA value of right inferior fronto-occipital fasciculus was significantly lower in metabolic subjects compared with normal control subjects. A significantly statistical negative correlation was observed between BMI and FA values of right inferior fronto-occipital fasciculus.

Charvi Shetty¹, Yi-Ou Li¹, Julia Owen¹, Matthew Malter Cohen², BJ Casey², and Pratik Mukherjee^1
1Radiology, UCSF, San Francisco, CA, United States, ²Sackler Institue for Developmental Psychobiology, Weill Medical Colege of Cornell University, New York, NY, United States

In this work, we use seed-voxel correlation (SVC) as a means to assess the microstructural correlations between white matter tracts in the normal, adult brain. We compare our results with SVC to those obtained with independent component analysis (ICA), as previously published. We find 36 similar pairs of maps and we are able to uncover frontal fiber tracts not found previously. The results reveal that ICA and SVC yield complimentary, as well as, overlapping maps of spatially correlated white matter tissue.

Kousaku Saotome^1,2, Akira Matsushita¹, Tomonori Isobe³, Eisuke Satou⁴, Satoru Osuka³, Yoshiyuki Ishimori⁵, Akira Matsumura¹, and Yoshiyuki Sankai^1
1Center for Cybernics Research, University of Tsukuba, Tsukuba, Ibaraki, Japan, ²Tsukuba Medical Center Hospital, Tsukuba, Ibaraki, Japan, ³University of Tsukuba, Tsukuba, Ibaraki, Japan, ⁴Kitasato University, Sagamihara, Kanagawa, Japan, ⁵Ibaraki Prefectural University of Health Sciences, Ami, Ibaraki, Japan

 

Lingchih Lin¹, Jianhui Zhong^1,2, and Walter G. O'Dell^3
1Department of Physics and Astronomy, University of Rochester, Rochester, NY, United States, ²Department of Imaging Sciences, University of Rochester, Rochester, NY, United States, ³Department of Radiation Oncology, University of Florida, Gainesville, FL, United States

The distributions of inter- and intra- distances within and between groups of subjects are compared by Euclidean, Squared Euclidean, and Log-Euclidean based distance functions for DT-MRI to calculate the exact p-values in the permutation testing. A novel approach by approximating the tail distributions of distance functions as the test statistics from a generalized Pareto model (GPD) is proposed to increase the sensitivity of detection and computation efficiency at the same time. Higher ratios of deterioration induced by multiple sclerosis are detected from Log-Euclidean based distance functions compared to mean and median based distance functions in a larger portion of regions of corpus callosum and corona radiata. Densities of distribution function are quantified by parameters estimated from maximizing likelihood, and compared with methods of moments, and probability weighted moments.

Mustafa Okan Irfanoglu^1,2, Frank Q Ye³, Evren Özarslan¹, David Leopold³, Afonso C Silva⁴, and Carlo Pierpaoli^1
1NIH, NICHD, Bethesda, MD, United States, ²Center for Neuroscience and Regenerative Medicine, Uniformed Services University of the Health Sciences, Bethesda, MD, United States, ³NIH, NIMH, Bethesda, MD, United States, ⁴NIH, NINDS, Bethesda, MD, United States

Diffusion in the brain cortex has a complicated structuture. Traditional image axes based coordinate frameworks might not be suitable to analyze finer level details of this process and a new representation may be needed. In this work, we propose to analyze the diffusion properties relative to a "cortex based coordinate framework". We use an ex-vivo marmoset data to analyze the diffusion both wit the tensor and spherical harmonics model.

Jiri Keller^1,2, Aaron Michael Rulseh¹, Arnost Komárek³, Iva Latnerová¹, Robert Rusina⁴, Jiri Klempír⁵, Katerina Zárubová⁶, and Josef Vymazal^1
1RDG, Na Homolce Hospital, Prague 5, Prague, Czech Republic, ²Neurology, 3rd Medical Faculty, Charles University, Prague, Prague 10, Czech Republic, ³Faculty of Mathematics and Physics, Charles University, Prague, Czech Republic, ⁴Department of Neurology, Faculty Thomayer Hospital, Czech Republic, ⁵Department of Neurology and Center of Clinical Neuroscience, 1st Faculty of Medicine, Charles University, ⁶Department of Neurology, 2nd Faculty of Medicine, Charles University, Czech Republic

To facilitate the assessment of fractional anisotropy (FA) in daily clinical practice, we propose a new non-linear colour look-up table (LUT) based on healthy subject data (in the present study 76 volunteers). Our approach can be used to build a site-specific LUT, especially relevant considering the variability of signal-to-noise ratio. The LUT has been tested on a cohort 17 multiple system atrophy (MSA) subjects, 13 Parkinson disease subjects and 17 healthy volunteers. Three blinded radiologists achieved an average sensitivity of 88% (65-100%) and specificity of 93% (80-100%) in differentiating MSA from other groups solely using this method.

Joong Hee Kim¹, and Sheng-Kwei Song^1
1Radiology, Washington University School of Medicine, St. Louis, MO, United States

The sensitivity of DTI derived anisotropy, diffusivity, and tractography to primary and secondary white matter injury was examined. Rat spinal cords undergoing right lateral transection or sham operation were examined using DTI at acute and 7 days after injury. At acute, the axial diffusivity localized the epicenter with severe diffusivity reduction where diffusion anisotropy failed to detect the injury. At 7 days after injury, both axial diffusivity and anisotropy visualized the diffuse axonal injury from epicenter in close agreement with histological validation. However, the main orientation of axonal fiber was largely preserved failing to detect the axonal injury.

Amritha Nayak^1,2, Lindsay Walker^1,2, Carlo Pierpaoli³, and The Brain Development Cooperative Group^4
1STBB-PPITS, National Instritutes of Health, NICHD, Bethesda, MD, United States, ²Center for Neuroscience and Regenerative Medicine (CNRM), Bethesda, MD, United States,³National Instritutes of Health, NICHD, Bethesda, MD, United States, ⁴www.NIH-PediatricMRI.org

We have evaluated the Atlas based image analysis using Large Deformation Diffeomorphic Metric Mapping (LDDMM) on DTI age specific group average brains that have been created from a large cohort (0-22 yrs) of subjects that were recruited as part of the NIH MRI Study of Normal Brain Development project. Our goal has been to understand the utility of the atlas based method in accurately handling and extracting diffusion values that are representative both of the underlying structure and normal development.

Robert J. Harris^1,2, Robert M. Prins³, Whitney B. Pope¹, Timothy M. Cloughesy⁴, Linda M. Liau³, and Benjamin M. Ellingson^1,2
1Radiological Sciences, David Geffen School of Medicine, University of California - Los Angeles, Los Angeles, CA, United States, ²Biomedical Physics, David Geffen School of Medicine, University of California - Los Angeles, Los Angeles, CA, United States, ³Neurosurgery, David Geffen School of Medicine, University of California - Los Angeles, Los Angeles, CA, United States, ⁴Neurology, David Geffen School of Medicine, University of California - Los Angeles, Los Angeles, CA, United States

Immunogenic dendritic cells have recently been studied as a novel therapy for human brain tumors. In this study, we examine MRI anatomical and diffusion data for ten patients before and after dendritic cell treatment. Functional diffusion maps were calculated using serial diffusion images. One patient with a clear partial response showed a decrease in the volume of decreasing ADC voxels coupled with an increase in increasing ADC voxels during the time following treatment, while two patients with progressive disease showed opposite trends. This suggests that a steadily decreasing volume of decreased ADC following therapy is associated with a favorable response.

Yogesh Rathi¹, Oleg Michailovich², Sylvain Bouix³, Martha Shenton³, and Carl-Fredrik Westin^3
1Harvard Medical School, Boston, MA, United States, ²ECE, University of Waterloo, Waterloo, Canada, ³Harvard Medical School

In this work, we propose a novel method for obtaining a T1-weighted MR image from a diffusion MRI scan. Existing measures of diffusion, such as, FA, entropy, etc. do not provide enough contrast between gray matter and CSF regions. The proposed algorithm produces images that can better delineate the different tissue types just like a T1 image. The predicted T1 images can be used in registration, segmentation and visualization of fiber tracts.

Gerard Thompson¹, Sha Zhao¹, and Alan Jackson^1
1University of Manchester, Manchester, United Kingdom

Diffusion tensor imaging (DTI) has been widely used to study white matter in health and disease. Quantitative magnetization transfer (qMT) imaging has also been used to a lesser extent. We provide evidence that fractional anisotropy (FA) form DTI and bound fraction (f) from qMT provide complementary information in white matter. Importantly, f remains high even in areas of reduced FA due to fiber crossing. A combination of f and FA may therefore provide complementary information about white matter in health and disease, and allow interrogation of regions of fiber crossing and sub-cortical regions not normally assessable using FA alone.

Van J Wedeen¹, Lawrence L Wald¹, Julien Cohen-Adad¹, M Dylan Tisdall¹, Jennifer A McNab¹, Ruopeng Wang¹, Thomas Witzel¹, Ralph Kimmlingen², Eva Eberlein², Philipp Hoecht³, Boris Keil¹, Juergen Nistler², Dietmar Lehne², Keith Heberlein³, Herbert Thein², Franz Schmitt², Jack Van Horn⁴, Arthur Toga⁴, and Bruce R Rosen^1
1A. A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital, Harvard Medical School, Charlestown, MA, United States, ²Siemens Healthcare, Erlangen, Germany, ³Siemens Healthcare, Boston, MA, United States, ⁴Laboratory of Neuro Imaging, Dept. of Neurology, UCLA, Los Angeles, CA, United States

 

Robin Martin Heidemann¹, Alfred Anwander¹, Thorsten Feiweier², Cornelius Eichner¹, Ralf Lützkendorf³, Johannes Bernarding³, Thomas R Knösche¹, and Robert Turner^1
1Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany, ²Siemens Healthcare Sector, Erlangen, Germany, ³Otto von Guericke University, Magdeburg, Germany

An isotropic sub-millimeter spatial resolution can be achieved in diffusion MRI using advanced acquisition strategies and hardware at 7T. It is shown that an isotropic resolution of 1 mm results in micro-structural maps with more details, especially in the cortex, compared to 1.5 mm isotropic resolution. We show results based on diffusion acquisitions with 800 µm isotropic resolution, showing fine structures even in basal regions of the brain, such as the Thalamus.

Wenchuan Wu¹, Sheng Fang², Chun Yuan^1,3, and Hua Guo^1
1Center for Biomedical Imaging Research, Department of Biomedical Engineering, School of Medicine, Tsinghua University, Beijing, China, ²Institute of nuclear and new energy technology, Tsinghua Univerisity, Beijing, China, ³Department of Radiology, University of Washington, Seattle, WA, United States

Combined variable density spiral (VDS) and improved CORNOL was proposed to acquire high spatio-temporal resolution diffusion tensor images. The oversampled data from VDS central k space provides self-navigation capability, thus they can be used to correct motion induced phase error induced by motion sensitive diffusion gradients. Additionally, VDS is a suitable candidate for nonlinear reconstruction as the undersampling artifacts are incoherent. CORNOL is a kind of nonlinear reconstruction methods that can effectively suppress incoherent aliasing artifacts while maintaining image structure details for highly undersampled data. Preliminary results show that with a reduction factor of 4, VDS and CORNOL combination for high resolution diffusion tensor imaging has less artifacts and better structure details compared to CG-SENSE.

Denis Kokorin^1,2, Martin Haas¹, Stefanie Buchenau¹, Jürgen Hennig¹, and Maxim Zaitsev^1
1Department of Radiology, Medical Physics, University Medical Center Freiburg, Freiburg, Germany, ²International Tomography Center, Novosibirsk, Russian Federation

Parallel transmission in combination with spatially selective excitation allows a reduction of field of view in the phase encoding direction. In this study this principle was examined for multi slice inner volume imaging and the advantages for artifact suppression in EPI are presented. The method was tested in vivo for DWI applications on Siemens MAGNETOM TRIO human system with 8 channel TxArray extension.

Murat Aksoy¹, Samantha J Holdsworth¹, Rafael O'Halloran¹, and Roland Bammer^1
1Center for Quantitative Neuroimaging, Department of Radiology, Stanford University, Stanford, CA, United States

We propose a novel high resolution diffusion-weighted imaging sequence that combines readout-segmented (RS)-EPI and short-axis propeller (SAP)-EPI readouts. The RS-EPI image provides a high resolution diffusion-weighted image, whereas the SAP-EPI readout acts as a navigator for both the RS-EPI and SAP-EPI readout and is also used to form a second diffusion weighted image with a higher T2-weighting and thus additional diffusion-weighted contrast.

J. L. Holtrop¹, A. T. Van², and B. P. Sutton^1,3
1Bioengineering, University of Illinois at Urbana-Champaign, Urbana, IL, United States, ²Electrical and Computer Engineering, University of Illinois at Urbana-Champaign, Urbana, IL, United States, ³Beckman Institute, University of Illinois at Urbana-Champaign, Urbana, IL, United States

High resolution diffusion weighted imaging can provide more accurate information on structural connectivity of small fiber structures. Achieving high resolution with diffusion weighting provides many challenges. Among these challenges is the need to achieve sufficient SNR. We propose an approach to diffusion imaging that increases SNR efficiency by employing a multi-slab strategy which is used to optimize the TR and combine this with a multi-shot 3D encoding and motion induced phase error correction. The strategy, employed along with advances in image reconstruction, enable the acquisition of 1mm isotropic diffusion weighted images. Examples are shown for a whole brain study.

Valentina Taviani¹, Jean H. Brittain², Bruce D. Collick³, Diego Hernando¹, Nathan S. Artz¹, and Scott B. Reeder^1,4
1Department of Radiology, University of Wisconsin, Madison, WI, United States, ²Applied Science Laboratory, GE Healthcare, Madison, WI, United States, ³GE Healthcare, Waukesha, WI, United States, ⁴Department of Medical Physics, University of Wisconsin, Madison, WI, United States

Diffusion-prepared (DP) fast spin-echo (FSE) techniques could potentially produce distortion-free, diffusion-weighted images with high SNR and high resolution. However, the large diffusion-sensitizing gradients applied during the preparation can produce eddy-current-induced phase shifts at the time of magnetization tip-up, resulting in an unwanted modulation of the transverse magnetization. In this work we demonstrate the improved eddy current properties of a twice-refocused diffusion preparation scheme when compared to previously reported once-refocused preparations and we show preliminary phantom data obtained with a DP single-shot FSE technique and the twice-refocused diffusion-encoding scheme.

Zoltan Nagy¹, Nikolaus Weiskopf¹, and David L Thomas^2
1Wellcome Trust Centre for Neuroimaging, University College London, London, United Kingdom, ²Institute of Neurology, Department of Brain Repair and Rehabilitation, University College London, London, United Kingdom

In the twice-refocused spin echo sequence for diffusion encoding several pathways may lead to a measurable signal, of which, only the twice-refocused spin echo should be acquired. Here we show that there is adequate signal from FID of the 1st RF pulse and care must be taken to make sure the diffusion-weighting and crusher gradients do not cancel each other for any of the diffusion-encoding directions. A general solution is hard to achieve because the gradient timings depend on the TE, gradient amplitude, etc but one solution may be to always use crushers orthogonal to the diffusion encoding direction.

Daniel C Alexander¹, and Tim B Dyrby^2
1Centre for Medical Image Computing, Dept. Computer Science, UCL, London, United Kingdom, ²Danish Research Centre for Magnetic Resonance, Copenhagen University Hospital Hvidovre, Hvidovre, Denmark

We present a simple method to avoid bias introduced by crusher and slice select gradients in stimulated echo diffusion MRI. We demonstrate the necessity of using such a compensation in classical diffusion tensor imaging and ActiveAx axon diameter index mapping.

Stephen R Yutzy¹, Sudhir K Pathak², Kevin Jarbo², Walter Schneider², and Fernando E Boada^1
1Radiology, University of Pittsburgh, Pittsburgh, PA, United States, ²Psychology, University of Pittsburgh, Pittsburgh, PA, United States

As white matter tractography grows in popularity, its frequently lengthy acquisition times become more problematic. We propose the q-space trajectory, where multiple points in q-space are sampled within each TR, as a reconstruction-independent method for accelerating these acquisitions. Comparable orientation distribution functions are obtained from a conventional DSI sequence and an accelerated q-space trajectory version of the same diffusion directions. The q-space trajectory technique is compatible with other acceleration methods such as compressed sensing and multi-band imaging.

Bruno Madore¹, Jr-Yuan Chiou¹, Renxin Chu¹, Tzu-Cheng Chao², and Stephan E. Maier^1
1Department of Radiology, Harvard Medical School, Brigham and Women's Hospital, Boston, MA, United States, ²Department of Computer Science and Information Engineering, National Cheng-Kung University, Tainan, Taiwan

A fast-imaging strategy is presented, to accelerate data acquisition in segmented MR diffusion imaging. Navigator echoes are typically employed to correct for motion effects in multi-shot diffusion imaging. A spatially-smooth representation of the object is obtained from a small region around k-space center, and only phase information is used for motion correction. On the other hand, in the field of accelerated MRI, spatially-smooth signal is often called ‘prior knowledge’, and only magnitude information in an x-y-frequency space is used for regularization purposes. The proposed approach fully exploits the available navigator signal, phase and magnitude, toward accelerated motion-corrected diffusion imaging.

Yin Wu^1,2, Yan-Jie Zhu^1,2, Wei Liu^1,2, Qiu-Yang Tang^1,2, Yi-Shuo An^1,2, Ed X. Wu^3,4, Leslie Ying⁵, and Dong Liang^1,2
1Paul C. Lauterbur Research Center for Biomedical Imaging, Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Shenzhen, Guangdong, China, ²Key Lab of Health Informatics, Chinese Academy of Sciences, Shenzhen, Guangdong, China, ³Laboratory of Biomedical Imaging and Signal Processing, The University of Hong Kong, Pukfulam, Hong Kong, ⁴Department of Electrical and Electronic Engineering, The University of Hong Kong, Pukfulam, Hong Kong, ⁵Department of Electrical Engineering and Computer Science, University of Wisconsin, Milwaukee, WI, United States

DTI is a powerful tool to probe microstructure of biological tissues, but usually suffered from lengthy data acquisition. In the current study, theory of distributed compressed sensing (DCS) was applied to test its feasibility of accelerating DTI data sampling. Reconstruction performance was found to be related with SNR and reduction factor based on simulation study. Good reconstruction accuracy was numerically and visually achieved even at high accelerating rate of 4 for the experimental data. All the results indicate the feasibility of DCS to speed DTI data acquisition, which would greatly help to broaden its potential practical applications in the future.

Jonathan I. Sperl¹, Marion I. Menzel¹, Ek T. Tan², Kedar Khare², Kevin F. King³, Christopher J. Hardy², and Luca Marinelli^2
1GE Global Research, Garching n. Munich, BY, Germany, ²GE Global Research, Niskayuna, NY, United States, ³GE Healthcare, Waukesha, WI, United States

Diffusion spectrum imaging (DSI) not only provides angular information about diffusivity in the brain but also radial information such as diffusional kurtosis. Due to the non-Gaussian noise distribution in DSI, a standard least-squares fitting of diffusion and kurtosis tensor induces bias on the fitted tensor elements and the subsequently derived scalar measures such as mean kurtosis. This work is intended to show that compressed sensing reconstruction in q-space, which is used to accelerate DSI by enabling undersampled acquisitions, also helps to reduce the bias on the data and by this means improve the estimation of kurtosis.

Rüdiger Stirnberg¹, Tony Stöcker¹, and N. Jon Shah^1,2
1Institute of Neuroscience and Medicine - 4, Forschungszentrum Jülich, Jülich, Germany, ²Department of Neurology, Faculty of Medicine, JARA, RWTH Aachen University, Aachen, Germany

Several publications have addressed the analysis of anisotropic error propagation in DTI, i.e. as a function of fibre orientation. Recently, model-free diffusion imaging methods requiring high angular resolved diffusion imaging acquisitions have gained in popularity, e.g. q-ball imaging (QBI). In this abstract, QBI error anisotropy without and with noise (Monte Carlo simulation) is compared for different encoding scheme types with 30 to 240 directions and constant number of excitations. It is shown that QBI error anisotropy converges to a minimum with increasing numbers of diffusion weighting directions provided an encoding scheme type offering a high degree of uniformity is used.

Cheng Guan Koay¹, Samuel A Hurley¹, and M. Elizabeth Meyerand^1
1Department of Medical Physics, University of Wisconsin-Madison, Madison, WI, United States

Diffusion MRI measurements are acquired sequentially with unit gradient directions that are distributed uniformly on the unit sphere. The ordering of the gradient directions has significant effect on the quality of dMRI-derived quantities. Even though several methods have been proposed to generate optimal orderings of gradient directions, these methods are not widely used in clinical studies because of the two major problems of computational inefficiency. In this work, the authors propose two extremely efficient and deterministic methods to solve these problems.

Julien Sénégas¹, Thomas G Perkins^2,3, Jochen Keupp¹, Christian Stehning¹, Gwenael Herigault⁴, Mariah Smith-Miloff³, and Shahid M Hussain^3
1Philips Research Laboratories, Hamburg, Germany, ²Philips Healthcare, Cleveland, Ohio, United States, ³University of Nebraska Medical Center, Omaha, Nebraska, United States, ⁴Philips Healthcare, Best, Netherlands

Careful choice of the b-values for in vivo measurements of ADC with IVIM is required to obtain non-biased ADC values with maximal precision for a given acquisition time. The proposed Monte Carlo methodology allows targeting the b-values to the organ-specific perfusion regime. Our results showed that b-value sampling schemes designed to minimize noise propagation can significantly outperform common sampling schemes such as regular distributions of b-values: In the presented study, a 77% increase in ADC SNR was observed just by modifying two b-values from the original diffusion protocol, confirming the prediction of the Monte Carlo simulations.

Marco Lawrenz¹, and Jürgen Finsterbusch^1
1Institute of Systems Neuroscience, University Medical Center Hamburg-Eppendorf, Hamburg, Germany, Germany

Double-wave-vector diffusion-weighting (DWV) experiments are a promising tool to investigate the tissue microstructure by considering the signal amplitude vs. the relative angle between the two wave vectors. While for short mixing times between the two diffusion-weighting periods involved, the signal modulation is proportional to the cell size, it increases with the cell eccentricity at long mixing times. Here, it is shown that with an appropriate direction combination scheme all elements of the recently presented tensor model for DWV experiments at long mixing times can be determined which allows the estimation of compartment sizes and cell eccentricities within a single experiment.

Ivana Drobnjak¹, and Daniel C Alexander^1
1Center for Medical Image Computing, Department of Computer Science, University College London, London, London, United Kingdom

We explore whether variable, rather than fixed, orientation of diffusion-gradients improves sensitivity to coherently oriented pore size. We optimise the shape and orientation of the gradient waveform to minimise the expected variance of parameter estimates of a simple white-matter model. Results suggest that: varying orientation in a plane perpendicular to the cylinders does not increase sensitivity to model parameters; variation in a parallel plane increases sensitivity significantly because it improves the diffusion constant estimates; similar improvement in estimates can be achieved without optimising orientation, but by having one measurement in the parallel and the others in the perpendicular direction.

Karen Mooney¹, John P. Mugler III², Gordon D. Cates Jr.^1,2, W. A. Tobias¹, and G. Wilson Miller^2
1Physics, University of Virginia, Charlottesville, Virginia, United States, ²Radiology, University of Virginia, Charlottesville, Virginia

The sensitivity of hyperpolarized helium to diffusion provides an opportunity to generate diffusion weighting based on asymmetric signal attenuation in a balanced SSFP sequence, by placing diffusion-sensitizing gradients in selected TR windows. We present a pulse sequence designed to take advantage of this effect, and use it make accurate measurements of the helium diffusion coefficient at very short timescales.

Silvia De Santis¹, and Derek K Jones^1
1CUBRIC, Cardiff University, Cardiff, United Kingdom

We propose a simple and easily-implemented routine to perform QA, and will make the analysis software freely available. The routine comprises diffusion-weighted (DW) acquisitions on a phantom along different gradient directions at different b-values; each repetition lasts less than 5 minutes. Linearity of the logarithm of the diffusion signal vs b-value, uniformity of Gmax across the field-of-view, mutual agreement of gradient power across the three logical axes and temporal stability are tested to spot potential issues that can lead to inaccuracy in DTI and tractography studies.

Rafael O'Halloran¹, and Roland Bammer^1
1Radiology, Stanford University, Stanford, CA, United States

In diffusion-weighted steady state free precession imaging combining data from multiple shots to form an image leads to artifactual signal dropout. The cause of this is random, motion-induced phase from shot to shot. A simple and intuitive way to understand this phenomenon is by analogy to RF spoiling, in which “random” phase is imposed on the transverse magnetization by changing the phase of the RF excitation pulse. This analogy is presented pictorially and discussed in the context of a comparison of a single-shot EPI sequence to a multi-shot SSFP sequence in a human volunteer.

Christopher L Welsh^1,2, Edward W Hsu¹, Ganesh Adluru², Jeffrey S Anderson², and Edward VR DiBella^2
1Department of Bioengineering, University of Utah, Salt Lake City, Utah, United States, ²UCAIR, Department of Radiology, University of Utah, Salt Lake City, Utah, United States

Diffusion Tensor Imaging (DTI) is useful for characterizing tissue microstructure, but suffers from low temporal resolution and the associated low spatial resolution and SNR. A model-based strategy is presented to reconstruct undersampled multi-coil DTI data, accelerating acquisition. The reconstruction is performed by fitting the diffusion tensor directly to the acquired data via minimizing a spatially regularized cost function. The proposed approach, along with traditional compressed sensing, GRAPPA and SENSE, are compared against the fully sampled case. SENSE performs the best in terms of fiber orientation estimation, while the model-based approach performs the best in terms of FA estimation.

Siawoosh Mohammadi¹, Zoltan Nagy¹, Chloe Hutton¹, Oliver Josephs¹, and Nikolaus Weiskopf^1
1Wellcome Trust Centre for Neuroimaging, UCL Institute of Neurology, University College London, London, United Kingdom

Vibrations induced by strong diffusion gradients in DTI, cause an echo shift in k-space and consequential signal-loss. Here we present a simple method to correct these vibration artefacts using phase-encoding reversal (COVIPER) by combining two images with reversed PE direction, each weighted by a function of its local tensor fit error. COVIPER was validated against low vibration reference data, resulting in an error reduction of about 72% in FA maps. When COVIPER is combined with other corrections based on phase encoding reversal, they provide a comprehensive correction for eddy currents, susceptibility-related distortions and vibration artefacts in DTI.

Christopher J. Wargo¹, and John C. Gore^1
1Department of Radiology and Radiological Sciences, Vanderbilt University Institute of Imaging Science, Nashville, TN, United States

In this abstract, outer volume suppression (OVS) is combined with SENSE parallel imaging to accelerate diffusion tensor imaging (DTI) in the human brain at 7T. OVS provides additional reduction in the DTI data set beyond that achieved using SENSE by constraining the FOV to localized regions in the brain, targeting specifically the midbrain, lentiform, and basal ganglia. In-plane resolutions of 1mm x 1mm x 2 mm were acquired using six diffusion directions and a b value of 1000 s/mm2. Resulting DTI-OVS images are of high quality with minimal or no artifacts using a single shot EPI sequence, enabling localized FA and ADC measurement.

Guangqiang Geng¹, Roland Henry², and Caroline Rae^1,3
1Neuroscience Research Australia, Sydney, NSW, Australia, ²Departments of Neurology, Radiology and Biomedical Imaging, Bioengineering, UCSF, San Francisco, California, United States, ³UNSW, Sydney, NSW, Australia

Magnetic resonance imaging (MRI) with phase array coils and an increased number of receiver channels improves signal-to-noise ratio (SNR). This would benefit imaging methodologies sensitive to noise such as diffusion tensor imaging (DTI). In order to verify the improved imaging performance with increased acquisition channels, we quantify the uncertainty of DTI measurements for anisotropy through statistical analysis of human brain data. We have proved that the uncertainties of DTI results are consistently reduced with 32-channel coil compared to 8 channels, and this will potentially reduce the necessary scan time and SAR while maintaining the reliability of DTI measurements.

S Ennar¹, H Tuo¹, and I Sickle^1
1CIRBUC, Ffidrac, Ffidrac, Iceland

In this work, we demonstrate a remarkable effect in diffusion MRI. When pure dihydrogen monoxide (maintained at room temperature) is sampled with a carefully-calibrated sampling scheme comprising six sampling vectors (arranged according to a model of electrostatic repulsion), peaks in the orientational density function are distributed with incredible regularity. Despite the absence of apparent structure outside the scanner, our analysis shows that a combination of high B0-field (3000 mT), and rapidly changing magnetic field gradients, leads to alignment of the eigenvectors of the tensor produce a regular crystalline dodecahedral arrangement. (Or perhaps not)

Dan Xu¹, Brice Fernandez², Kenichi Kanda³, Robert D. Peters¹, Joe K. Maier³, and R. Scott Hinks^1
1Applied Science Lab, GE Healthcare, Waukesha, WI, United States, ²Applied Science Lab, GE Healthcare, Munich, Germany, ³MR Engineering, GE Healthcare, Waukesha, WI, United States

Homodyne processing with adaptively selected lowpass filter for phase map estimation was previously shown effective in reducing worm-like and shading artifacts sometimes seen in partial Fourier based diffusion weighted echo planar imaging (DW-EPI). Occasionally, image shading artifacts can also appear in DW-EPI images due to gradient induced table vibrations. In this paper, we show that a k-space centroid based adaptive homodyne method can be used to effectively mitigate the vibration related image shading artifacts.

Bertram L Koelsch¹, Kayvan R Keshari², Renuka Sriram², Daniel B Vigneron^1,2, and John Kurhanewicz^1,2
1Graduate Program in Bioengineering, UC Berkeley - UCSF, San Francisco, CA, United States, ²Radiology, UCSF, San Francisco, CA, United States

Most diffusion weighted imaging and spectroscopy studies observe proton since carbon-13 is limited by low SNR, resulting from the nuclei’s low gamma and its low natural abundance. We take advantage of the signal enhancement achieved with dissolution DNP and molecular motion to acquire diffusion weighted spectra of hyperpolarized carbon-13. The apparent diffusion coefficient (ADC) measured for thermally polarized 13C urea was similar to that measured by hyperpolarized 13C urea. In comparing thermally polarized and hyperpolarized 13C urea diffusion we show the feasibility of using hyperpolarized diffusion weighted spectroscopy in future bioreactor and animal studies to quantify transport and enzyme kinetics.

Christian Baumgartner¹, Ofer Pasternak², Sylvain Bouix², Carl-Fredrik Westin³, and Yogesh Rathi^2
1Information Technology and Electrical Engineering, ETH Zürich, Zürich, Switzerland, ²Psychiatry Neuroimaging Laboratory, Harvard Medical School, Boston, MA, United States, ³Laboratory of Mathematics in Imaging, Harvard Medical School, Boston, MA, United States

In this work, we describe a tractography method that simultaneously estimates multiple diffusion tensors and an isotropic component, we term as free water. The model consists of three Gaussian tensors, one of which represents isotropic diffusion of free water, that are fitted to the DWI-signal using an unscented Kalman filtering framework. By this means, each estimation is guided by those previous, resulting in an inherent regularization of the tracts. The proposed method can be useful in tracing fibers through edema or lesions, where traditional tractography algorithms fail.

Paul Armitage¹, Susana Muñoz Maniega², James Bridson², Michael Poon², and Mark Bastin^2
1University of Sheffield, Sheffield, South Yorkshire, United Kingdom, ²University of Edinburgh, Edinburgh, United Kingdom

Neighborhood tractography was implemented using a fast marching algorithm to provide automated extraction of white matter fibers of interest. Inter- and intra-subject reproducibility of resulting tract volume, and FA measurements from segmented white matter fibres were obtained on a series of normal healthy subjects. Variability of the automated method was found to be comparable to previous studies, including those using manual methods, and 83% of fibres were considered to provide acceptable representations of the tract of interest.

Jadrian Miles¹, and David H. Laidlaw^1
1Computer Science Department, Brown University, Providence, Rhode Island, United States

We present an easy-to-implement method for computing uncertainty in deterministic DTI tractography. The model is derived from Monte-Carlo simulation studies of the effect of diffusion-weighted image noise and Q-space sampling on streamline orientation variability. The result is a straightforward equation for the growth of uncertainty that is linear in the arc-length distance from a streamline seed point.

Robert Elton Smith^1,2, Jacques-Donald Tournier^1,2, Fernando Calamante^1,2, and Alan Connelly^1,2
1Brain Research Institute, Florey Neuroscience Institutes, Heidelberg, Victoria, Australia, ²Department of Medicine, University of Melbourne, Melbourne, Victoria, Australia

Diffusion MRI streamlines tractography methods suffer from a range of methodological inadequacies, which can result in differences between fibre-tracking results and the underlying biology. We propose a modular addition to streamlines tractography, which makes effective use of tissue segmentation from an anatomical contrast image, to improve the biological accuracy of the reconstructed connectome. Interestingly, this mechanism improves fibre-tracking results when state-of-the-art diffusion models and tracking algorithms are used, whilst highlighting the inadequacies of inferior models and methods.

Yuichi Suzuki¹, Yoshitaka Masutani^1,2, Kenji Ito^1,2, Kenji Ino¹, Katsuya Maruyama³, Thorsten Feiweier⁴, Yasushi Watanabe¹, Yoshirou Satake¹, Masami Goto¹, Akira Kunimatu^1,2, Masaaki Akahane¹, Keiichi Yano¹, and Kuni Ohtomo^1,2
1Department of Radiology, The University of Tokyo Hospital, Bunkyo-ku, Tokyo, Japan, ²Radiology and Biomedical Engineering, The University of Tokyo Graduate School of Medicine and Faculty of Medicine, Japan, ³Siemens Japan K.K., Japan, ⁴Siemens AG, Germany

Our goal is depicting pyramidal tract with our anisotropic MPG and to compare the ability of depiction anisotropic MPG with Jones (isotropic) MPG. We performed visual assessment of the tractography result by each MPG, and compared the DICE factors. In visual assessment, anisotropic MPG depicted pyramidal tract in the all cases while the isotropic 60 MPG depicted 100% of CST, but 75% of CBT. Therefore, our anisotropic MPG has better depiction ability. In the comparison of DICE factors, our anisotropic MPG was slightly superior to that of the isotropic MPG though there was no significant difference between each DICE factors.

Lawrence R. Frank¹, and David Meyer^2
1Radiology, UCSD, San Diego, CA, United States, ²Mathematics, UCSD, La Jolla, CA, United States

We have developed a fiber tractography method that computes the maximum entropy trajectories between locations and depends upon the global structure of the diffusion tensor field. Computation of the pathways requires only solving a simple eigenvector problem for which efficient numerical routines exist, and a simple iterative computation. This method has potential significance for a wide range of applications, including studies of brain connectivity.

Myron Zhang^1,2, Ken E. Sakaie¹, and Jones Stephen^1
1Imaging Institute, The Cleveland Clinic, Cleveland, OH, United States, ²Physics, Cornell University, Ithaca, NY, United States

Maps of whole-brain anatomical connections generated by tractography prove valuable for identifying targets for resection in the treatment of phamacoresistant epilepsy. Unfortunately, current implementations have difficulty identifying a number of important connections while relying on intuitively appealing but ad-hoc logic. In this contribution, we present a logical formulation of probabilistic tractography that lends itself to fast implementation. The method identifies connections throughout the entire brain and may prove important for presurgical planning and other medical applications.

Parnesh Raniga¹, Kai Wong¹, Donald Tournier², Alan Connelly^2,3, and Olivier Salvado^1
1CSIRO Preventative Health National Research Flagship ICTC, The Australian e-Health Research Centre, Brisbane, Queensland, Australia, ²Brain Research Institute, Florey Neuroscience Institutes (Austin), Melbourne, Victoria, Australia, ³Department of Medicine, University of Melbourne, Melbourne, Victoria, Australia

We present a GPU accelerated version of an established CSD-based multi-fibre tractography algorithm that is 25 times faster at computing full brain tractography maps.

Kerstin Pannek¹, David Raffelt², Olivier Salvado³, and Stephen Rose^1
1The University of Queensland, Brisbane, Queensland, Australia, ²Brain Research Institute, Australia, ³The Australian E-Health Research Centre, Australia

A number of diffusion tractography derived scalar maps have been introduced, including track density imaging and average pathlength maps. We extend this technique by incorporating the directional information contained within the streamlines to build an angular track image (ATI). The ATI reveals information about angular streamline distribution. Information other than streamline number, such as streamline length, FA within streamlines, etc can also be incorporated in the calculation of the ATI to generate quantitative maps that can easily be compared across participants. ATIs provide more detailed information than scalar tractography maps by incorporating a directional component.

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

Abstract: A method to obtain the principal diffusion directions from the orientation density function (ODF) determined using the fourth order tensor, where the local maximum method fails, is developed. This method is applied to analyze the human brain diffusion data using high angular resolution diffusion imaging (HARDI) data. Application of the proposed method to HARDI datasets, a clear crossing pattern of superior longitudinal fasciculus (SLF) and cerebral spinal tract (CST) that was not seen with the local maximum method was demonstrated.

Sjoerd B Vos¹, Max A Viergever¹, and Alexander Leemans^1
1Image Sciences Institute, University Medical Center Utrecht, Utrecht, Netherlands

We present a new tractography visualization for HARDI data that captures the local configuration of “crossing fibers” and displays it along the fiber trajectory. This tract visualization object can be considered as the continuous envelope of voxel-based three-dimensional HARDI glyphs along the fiber pathway (also known as a hyperstreamline), combining the local diffusion information derived from these glyphs, e.g., the fiber orientation distribution, with the global anatomical information obtained from tractography. These HARDI hyperstreamlines create a more complete visualization, especially for tracts that traverse through areas of complex fiber configurations, aiding in our understanding of the white matter architectural configuration.

Ofer Pasternak¹, Yogesh Rathi¹, Martha Shenton¹, and Carl-Fredrik Westin^1
1Harvard Medical School, Boston, MA, United States

We present a method to extract crossing-angle maps from HARDI analysis. Applying the method on first-episode schizophrenia patients demonstrates a significant group-difference comparing with matched controls. In addition, we demonstrate that FA is not sensitive to the crossing-angle. Based on these findings we suggest the crossing-angle as a novel structural marker sensitive to unique HARDI information. The crossing-angle is expected to be influenced by large brain deformations such as tumors and sever trauma. However, in the absence of such major confounds we expect crossing-angle abnormalities to reflect developmental differences or local cytoskeletal dysfunction that changed the configuration of the crossings.

Ken E. Sakaie¹, Jian Lin¹, Lael Stone², and Mark J. Lowe^1
1Imaging Institute, The Cleveland Clinic, Cleveland, OH, United States, ²Mellen Center, The Cleveland Clinic, Cleveland, OH, United States

Tissue microstructure measurements specific to a tractography-defined white matter fascicle may serve as a clinically relevant measure in multiple sclerosis (MS). One problem is substantial variability along a fascicle, particularly if it includes regions near cortex and deeper in the brain. We present a methodology for segmenting a fascicle identified by probabilistic tractography and demonstrate its use in the transcallosal motor pathway of healthy subjects and MS patients. Using this method, tissue microstructure measurements can be made in a piecewise manner in order to reduce the overall variance and improve sensitivity to local differences.

Peter Savadjiev¹, Yogesh Rathi¹, Martha E. Shenton¹, Sylvain Bouix¹, and Carl-Fredrik Westin^2
1Psychiatry Neuroimaging Laboratory, Brigham and Women's Hospital, Harvard Medical School, Boston, MA, United States, ²Radiology, Brigham and Women's Hospital, Harvard Medical School, Boston, MA, United States

We introduce a novel method for computing multi-scale fibre tract shape and geometry based on the differential geometry of curve sets. By measuring the variation of a curve's tangent vector at a given point in all directions orthogonal to the curve, we obtain a ``2D dispersion orientation distribution function (ODF)'' at that point. That is, we compute a function on the unit circle which describes the extent to which the fibres disperse, or fan, along each direction on the circle. Our formulation is then easily incorporated into a continuous scale-space framework.

Lisa Willats¹, David Raffelt¹, Robert Elton Smith^1,2, Jacques-Donald Tournier^1,2, Alan Connelly^1,2, and Fernando Calamante^1,2
1Brain Research Institute, Florey Neuroscience Institutes, Melbourne, Victoria, Australia, ²Department of Medicine, University of Melbourne, Victoria, Australia

Recently several novel image contrasts derived from whole-brain fibre tracking-data (tractograms) have been introduced. The novel contrasts of these track-weighted imaging (TWI) methods may provide important information for clinical studies. However, before they can be used reliably to generate quantitative measures, it is important to characterise the within-subject intra-session and inter-session reproducibility, and between-subject variability. In this work we investigate the within-subject reproducibility (both intra-session and inter-session), and between-subject variability of TWI for a number of different contrasts across multiple subjects.

Fernando Calamante^1,2, Young-Don Son³, Jacques-Donald Tournier^1,2, Taek-Hyun Ryu³, Se-Hong Oh³, Alan Connelly^1,2, and Zang-Hee Cho^3
1Brain Research Institute, Florey Neuroscience Institutes, Heidelberg, Victoria, Australia, ²Department of Medicine, University of Melbourne, Melbourne, Victoria, Australia,³Neuroscience Research Institute, Gachon University of Medicine and Science, Incheon, Korea

We describe the first combination of PET images and whole-brain diffusion MRI fibre-tracking data, to generate super-resolution track-weighted PET maps (with 250µm isotropic resolution). This method is based on the recently proposed technique of super-resolution track-weighted imaging (TWI), which combines the information from whole-brain fibre-tracking with a reference image, to generate a super-resolution track-weighted version of that image. The method is illustrated on 11C-DASB, a radioligand with high affinity for serotonin transporters. The proposed technique of TW-PET opens up new possibilities in molecular imaging, by combining the high sensitivity to brain biomarkers from PET and the super-resolution capabilities of TWI.

Thijs Dhollander^1,2, Louise Emsell^1,3, Wim Van Hecke^1,3, Frederik Maes^1,2, Stefan Sunaert^1,3, and Paul Suetens^1,2
1Medical Imaging Research Center (MIRC), K.U.Leuven, Leuven, Belgium, ²Department of Electrical Engineering (ESAT), K.U.Leuven, Leuven, Belgium, ³Department of Radiology, University Hospitals of the K.U.Leuven, Leuven, Belgium

Track-density imaging (TDI) is a recently proposed technique to achieve super-resolution from diffusion weighted imaging datasets, by performing massive fiber tracking and counting the tracks in each voxel of a high resolution grid. In this work, the effect of noise in the data on the final TDI is investigated. The findings indicate that caution is needed: many discovered patterns of structures might mostly be caused by noise rather than true anatomy, although they look plausible and do not resemble the traditional impression of noise. Tools such as bootstrapping, combined with certain maps can highlight regions where caution is due.

Chieh-En Tseng¹, Yung-Chin Hsu², Yu-Chun Lo³, and Wen-Yih Isaac Tseng^3,4
1Biomedical Imaging and Radiological Sciences, National Yang-Ming University, Taipei, Taiwan, ²Institute of Biomedical Engineering, National Taiwan University, Taipei, Taiwan,³Center for Optoelectronic Biomedicine, National Taiwan University College of Medicine, Taipei, Taiwan, ⁴Department of Medical Imaging, National Taiwan University Hospital, Taipei, Taiwan

We proposed an automatic tract-specific analysis based on a diffusion spectrum imaging (DSI) template. This template-based approach entails transformation of native DSI data to the template, and performs tract-specific analysis on the transformed DSI data. The approach inevitably blurs out the data and may lead to errors in quantitative analysis. To investigate this question, the template-based approach was compared to a native-space approach. Fifteen subjects were analyzed. The results showed that mean generalized fractional anisotropy (GFA) was lower in the template-based approach, but the two approaches showed strong correlation. The results support the use of template-based approach in tract-specific analysis.

Sylvain Louis Merlet¹, Rachid Deriche¹, Kevin Whittingstall², and Maxime Descoteaux^3
1Athena Project-Team, INRIA, Sophia Antipolis, Méditerranée, France, ²Radiology department, Université de Sherbrooke, Québec, Canada, ³Sherbrooke Connectivity Imaging Laboratory, Computer Science Departement, Université de Sherbrooke, Québec, Canada

Diffusion MRI enables the quantification of water diffusion, influenced by the structure of biological tissues. While recent advances enable to recover complex fiber geometries using diffusion measurements along various sampling schemes some older MR systems work with limited gradient tables ,designed for Diffusion Tensor Imaging (DTI). DTI is an over simplification of water molecules diffusion that cannot resolve crossing fibers. We show that new diffusion signal modeling and processing techniques enable to capture complex angular structure of the diffusion process even from a reduced gradient direction set arising from an older MR system.

Namgyun Lee¹, Bryce Wilkins¹, and Manbir Singh^1
1Radiology and Biomedical Engineering, University of Southern California, Los Angeles, CA, United States

A novel compressed sensing approach for Diffusion Spectrum Imaging incorporating a unique sampling pattern is presented. This approach relies on minimizing the l1 norm of spatial finite-differences of an ensemble average propagator. Results of simulation studies show significant improvements with our approach compared to partial Fourier approach for accurately estimating underlying crossing fibers and reducing false positives. Human tractography studies show recovery of the cingulum, fornix, and other tracts in crossing-fiber areas with quality comparable to a fully-sampled reference (DSI203).

Zhenyu Zhou¹, Yijun Liu², Guang Cao¹, Karen M. von Deneen², and Dongrong Xu^3
1Global Applied Science Laboratory, GE Healthcare, Beijing, Beijing, China, ²McKnight Brain Institute, University of Florida, Gainesville, FL, United States, ³MRI Unit, Columbia University, New York, NY, United States

Recently, the tract-based analysis of white matter fibers has raised interests from the neurology and clinical neuroscience community since this methodology provides quantitative analysis of the properties of the specific fiber bundles, which provide a useful abstraction of the white matter structures and a clear identification of neural fibers. In order to benefit from the tract-based analysis, many clustering algorithms of the fiber tracts have been proposed. However, most approaches require a user initialization. In this paper, we propose a novel cluster method to automatically group brain white matter fibers into biologically meaningful neural tracts.

Young Beom Kim^1,2, Chrystelle Po², Daniel Kalthoff², and Mathias Hoehn^2
1Advanced Media Lab, SAIT, Yongin, Gyeonggi, Korea, ²In-vivo-NMR Lab, Max-Planck-Institute for Neurological Research, Cologne, Germany

Fiber tracking in small laboratory animals is of great interest as it is expected to open doors to understanding connectivity in controlled paradigms which can also be assessed with independent means. These fiber tracking investigations have been performed using DTI because of the demand to combine high resolution, low signal-to-noise of small voxels, and speed of scan time for minimization of anesthesia period. Diffusion spectrum imaging provides a more complex analysis of fiber circuits than the so far used DTI approach, discriminating also crossing fibers in the brain.

Longchuan Li¹, Matthew F Glasser², Todd M Preuss³, James K Rilling³, Frederick W Damen⁴, and Xiaoping Hu^1
1School of Medicine, Emory University/Gerogia Institute of Technology, Atlanta, GA, United States, ²Department of Anatomy and Neurobiology, Washington University, St Louis, MO, United States, ³Center for Translational and Social Neuroscience, Emory University, Atlanta, GA, United States, ⁴Department of Biomedical Engineering, Georgia Institute of Technology, Atlanta, GA, United States

Synopsis: Tracing studies in primates show that the primary motor cortex exhibits homotopic callosal projections, with (1) the strength of the callosal projections showing a medial to lateral gradient and (2) sectors of M1 having the densest connection with the homotopic contralateral sectors. Here we used two advanced tractography algorithms to trace these fibers in humans to see whether they replicate the observations in primates. While the medial sector showed significant stronger connections, the densest homotopic projection to the contralateral sectors were not observed in our results, indicating the need for more robust tractography methods for quantitative brain mapping studies.

Burkhard Mädler¹, Thomas E. Schläpfer^2,3, Jaak Panksepp⁴, and Volker A. Coenen^1
1Neurosurgery, University Bonn, Bonn, NRW, Germany, ²Psychiatry and Psychotherapy, University Bonn, Bonn, NRW, Germany, ³Psychiatry and Mental Health, Johns Hopkins University, Baltimore, MD, United States, ⁴Comparative Anatomy, Physiology and Pharmacology, Washington State University - College of Veterinary Medicine, Pullmann, WA, United States

The neuroanatomy of the MFB shows remarkable differences between rodents and humans. In rodents the MFB is large but a compact heterogeneous pathway that is a massive connection highway for integrating lower and higher brain functions whereas in humans the MFB depicts as a truly bipartite structure. It is evident that the MFB contains the major brain system that is responsible for a variety of drug addictions. Thus, one could imagine that inadvertently caused stimulation of the MFB during STN DBS-stimulation could be addictive but also the possibility of localized DBS of the slMFB may skirt some of the other brain processes that produce addiction.

Nico D. Papinutto^1,2, Sebastiano Galantucci^2,3, Roland G. Henry², Jorge Jovicich¹, and Maria Luisa Gorno-Tempini^2
1CIMeC, University of Trento, Mattarello, TN, Italy, ²University of California San Francisco, San Francisco, CA, United States, ³Scientific Institute and University Hospital San Raffaele, Milano, MI, Italy

The anterior temporal lobe (ATL) is crucial for higher order language functions, such as semantic memory, and it is involved in behavioral regulation. Evidences of ATL damage are found in many neurological diseases, but little is known about the structural connections of this area with the rest of the brain. In this study, by using DTI on 21 healthy subjects, we explore the architecture of the left ATL connectivity with many ipsilateral regions of the brain (in particular the areas known to have a key role in language) and we segment the ATL based on these connectivity patterns.

Colin R. Buchanan¹, Krzysztof Gorgolewski¹, Cyril R. Pernet², Amos J. Storkey¹, and Mark E. Bastin^2
1School of Informatics, University of Edinburgh, Edinburgh, United Kingdom, ²School of Molecular & Clinical Medicine, University of Edinburgh, Edinburgh, United Kingdom

We assessed the intra and intersubject variability of graph-theoretic measures of brain connectivity obtained from diffusion MRI (dMRI). Our method involves: 1) segmentation of cortical regions from high-resolution 3D T1-weighted volume scans; 2) construction of structural networks from dMRI data using established tractography algorithms with seed points placed in cortical regions; 3) graph-theoretic analysis of cortico-cortical connections; 4) quantification of between-subject and within-subject differences of network metrics using a percentile bootstrap technique. Results show that four commonly used network metrics can be produced consistently between imaging sessions with interscan difference of less than 6%.

Linda Marrakchi-Kacem^1,2, Christine Delmaire³, Pamela Guevara^2,4, Fabrice Poupon^2,5, Jérôme Yelnik^1,6, Sophie Lecomte^2,5, Pauline Roca^2,5, Alan Tucholka⁷, Alexandra Durr^1,8, Jean-François Mangin^2,5, Marie Chupin¹, Stéphane Lehéricy^1,9, and Cyril Poupon^2,5
1UPMC, CRICM, UMR-S975, Inserm, U975, CNRS, UMR 7225, Institut du Cerveau et de la Moëlle épinière, Paris, France, ²NeuroSpin, CEA, Gif-Sur-Yvette, France, ³Department of Neuroradiology, CHU Lille, Lille, France, ⁴University of Concepción, Concepción, Chile, ⁵IFR 49, Gif-Sur-Yvette, France, ⁶Département de Neurologie, Centre d’Investigation Clinique, Groupe Hospitalier Pitié-Salpêtrière, APHP, Paris, France, ⁷Department of Radiology, CHUM Notre-Dame Hospital, Montreal, Quebec, Canada, ⁸Department of Genetics, Groupe Hospitalier Pitié-Salpêtrière, APHP, Paris, France, ⁹Center for Neuroimaging CENIR, Groupe Hospitalier Pitié-Salpêtrière, APHP, Paris, France

In this study, we focused on the connectivity of the basal ganglia inferred from tractography. We proposed a novel method for the selection of the fiber tracts linking two brain structures from a whole brain tractogram obtained using streamline tractography. This method takes into account the basal ganglia circuits and provides only the direct connections between a pair of brain regions of inerest. It allows removing the effect of indirect connections from connectivity based studies. We used surface connectivity atlases as an application to illustrate the impact of this new tract selection approach on the analysis of the connections between the basal ganglia and the cortex.

Sarah Mang¹, Benjamin Bender², and Uwe Klose^2
1E071, German Cancer Research Center DKFZ, Heidelberg, DE, Germany, ²MR Research Group, Department of Diagnostic and Interventional Neuroradiology, University Hospital Tuebingen, Tuebingen, Germany

Comparison of anatomical labels in the Thalamus with DTI based segmentation.

Frederick William Damen^1,2, Longchuan Li¹, and Xiaoping Hu^1,2
1Biomedical Imaging Technology Center, Emory University, Atlanta, GA, United States, ²Department of Biomedical Engineering, Georgia Institute of Technology, Atlanta, GA, United States

The prospect to unveil the network architecture of the human brain has recently achieved large progress through the advancements in diffusion MRI and tractography. However, such effort in chimpanzees is lacking. Chimpanzees are our closest living relatives, and comparing their brain networks with those of humans may shed light on the unique cognitive and language abilities owned by humans. Here, we investigated the effects of parcellation schemes on the chimpanzee cerebral cortex as the first step to establish a framework to derive and compare brain connectivity networks across three species.

Pamela Guevara^1,2, Delphine Duclap^1,3, Cyril Poupon^1,3, Linda Marrakchi-Kacem^1,4, Josselin Houenou^1,5, Marion Leboyer⁵, Denis Le Bihan¹, and Jean-Fraçois Mangin^1,3
1Neurospin, CEA, Gif-sur-Yvette, France, ²University of Concepción, Concepción, Chile, ³IFR 49, Gif-sur-Yvette, France, ⁴UPMC, CRICM, UMR-S975, Inserm, U975, CNRS, UMR 7225, ICM, Paris, France, ⁵AP-HP, Univ. Paris-East, Dept. of Psychiatry, INSERM, U955 Unit, France

We present a HARDI human brain multi-subject bundle atlas derived from a two-level intra-subject and inter-subject clustering strategy. Each atlas bundle corresponds to several inter-subject clusters labeled by an expert in neuroanatomy to account for subdivisions of the underlying pathway often presenting large variability across subjects. An atlas bundle is represented by the multi-subject list of the centroids of all intra-subject clusters in order to get a good sampling of the shape and localization variability. The atlas contains 36 bundles deep white matter bundles belonging to brain hemispheres and the corpus callosum, and 47 superficial white matter bundles in each hemisphere.

Bryce Wilkins¹, Namgyun Lee¹, Kyungmin Nam¹, Darryl Hwang¹, and Manbir Singh^1
1Radiology and Biomedical Engineering, University of Southern California, Los Angeles, California, United States

A novel Independent Component Analysis (ICA) based approach to resolving multiple fiber directions per voxel in diffusion-weighted MRI analysis is quantitatively compared against four alternatives: Generalized q-Sampling Imaging, Constrained Spherical Deconvolution, analytical Q-Ball Imaging, and Higher-Order Tensors. We investigate the performance of the various methods when processing limited sample data, as is likely to be acquired in clinical studies due to constrained scan time. Results of two phantom datasets and a human study are presented, revealing consistently higher metric scores for the ICA-based approach, especially in the case of limited gradient sampling.

Monica Bucci¹, Sri Nagarajan², Eduardo Caverzasi², Kelly Westlake², Bagrat Amirbekian², and Roland Henry^3
1neurology, ucsf, san francisco, ca, United States, ²ucsf, ³neurology, ucsf

This study was designed to investigate and map the structural inter-regional connectivity of the hand motor network, by combining functional Magnetoencephalography Imaging (MEGI) with High angular resolution diffusion imaging MRI (HARDI) probabilistic tractography methods. We were able to map consistently the normal structural hand motor network in 20 controls. Transcallosal fiber tracts connecting Hand Primary Motor Cortex (Hand M1) to the contralateral Primary Motor Cortex and Supplementary Motor Area were not found in all the control subjects and interestingly the MEGI of the subjects without these connections showed a pattern of bilateral activation during both hands motor tasks.

Luke Bloy¹, Madhura Ingalhalikar², Robert T Schultz³, Timothy P.L. Roberts⁴, and Ragini Verma^2
1Section of Biomedical Imaging, University of Pennsylvania, Philadelphia, PA, United States, ²Section of Biomedical Imaging, Univeristy of Pennsylvania, Philadelphia, PA, United States, ³Center for Autism Research, Children's Hospital of Philadelphia, Philadelphia, PA, United States, ⁴Lurie Family Foundation's MEG Imaging Center, Children's Hospital of Philadelphia, Philadelphia, PA, United States

We create and compare an automated atlas creation method applied to DTI and HARDI data. The method works by generating a population average DTI dataset which is then parcellated using a normalized cuts segmentation algorithm. The framework is demonstration through its application to a population of adolescent subjects. The resultant atlas is compared to one created using HARDI datasets acquired on the same subjects. The HARDI atlas is better able to capture regions of cortical WM and those of complex WM such as fiber crossings. Additionally, these two atlases can be used in a population for comparative analysis between modalities.

Yung-Chin Hsu¹, Ching-Han Hsu², and Wen-Yih Isaac Tseng^3
1National Tsing Hua University, Hsinchu, Taiwan, Taiwan, ²National Tsing Hua University, ³National Taiwan University College of Medicine and Hospital

To our best knowledge, no method has been proposed to construct the template which is intrinsically from a group of DSI datasets. In the current study, we proposed an algorithm to iteratively estimate the group-specific DSI template, which incorporated a LDDMM-based method capable of spatially transforming between two DSI datasets. Seventy DSI datasets were recruited into the template estimation procedure. The results show the proposed method could construct the DSI template where many important fibers could be tracked out on it, indicating the effectiveness of the method.

Bryce Wilkins¹, Namgyun Lee¹, and Manbir Singh^1
1Radiology and Biomedical Engineering, University of Southern California, Los Angeles, California, United States

The FiberCup diffusion phantom, designed to mimic a coronal cross-section of the human brain, was developed for quantitative evaluation of multi-fiber diffusion models and tractography algorithms. The data available, however, is insufficient to allow comparison against all diffusion-weighted MRI methods, such as Diffusion Spectrum Imaging. We outline development of an accurate and flexible q-Space simulation of the FiberCup phantom, and illustrate its versatility by synthesizing data and processing it according to Diffusion Tensor Imaging, Diffusion Spectrum Imaging, Constrained Spherical Deconvolution and analytical Q-Ball Imaging methods. Results of tractography are presented.

Penny L Hubbard^1,2, Feng-lei Zhou¹, Stephen J Eichhorn³, Tim B Dyrby⁴, and Geoff J M Parker^1,2
1Imaging Sciences, Manchester Academic Health Science Centre, The University of Manchester, Manchester, United Kingdom, ²The Biomedical Imaging Institute, The University of Manchester, Manchester, United Kingdom, ³Physics and Astronomy, The University of Exeter, United Kingdom, ⁴Danish Research Centre for Magnetic Resonance, Copenhagen University Hospital Hvidovre, Hvidovre, Denmark

There is a need for a physical biomimetic phantom to act as a gold standard to allow a full characterisation and validation of the different diffusion acquisition methods, models, tracking algorithms, and microstructure measures. Here we introduce the use of co-electrospinning, a technique that allows the deposition of hollow, aligned, micron-sized fibres which mimic the microstructural and bulk characteristics of white matter tracts. We have demonstrated that the measured diffusivity is within the approximate range of biological tissues and that different diameter electrospun hollow fibres lead to different diffusivities and anisotropies, as well as evidence of fibre alignment.

Pim Pullens^1,2, Alard Roebroeck¹, and Rainer Goebel^1,2
1Maastricht Brain Imaging Center, Maastricht, Netherlands, ²Brain Innovation BV, Maastricht, Netherlands

Determining data quality and MR stability over time is an important issue in DW-MRI. Until now, this was done with isotropic phantoms or human subjects. We investigated the temporal DW-MRI characteristics of an anisotropic phantom, as well as the agreement in DW-MRI measures in 5 different anisotropic phantoms.

Michael Bach^1,2, Klaus Fritzsche³, Sena Minjoli², Bram Stieltjes¹, and Frederik Bernd Laun^1,2
1Quantitative Imaging-based Disease Characterization, German Cancer Research Center, Heidelberg, Germany, ²Dept. of Medical Physics in Radiology, German Cancer Research Center, Heidelberg, Germany, ³Medical Imaging and Biological Informatics, German Cancer Research Center, Heidelberg, Germany

The phantom presented here consists of six circular fiber strands of different cross sections (5x5 mm^2, 3x3 mm^2, 2.5x2.5 mm^2, 2x2 mm^2, 1.5x1.5 mm^2, 1x1 mm^2). These fiber strands are characterized by a homogeneous and high FA (up to 0.9). The precision, with which the fiber strands can be produced, opens up new possibilities with regard to the validation of the whole diffusion tensor imaging process. Presented data shows resolution and partial volume effects in DTI and tracked fibers. For example, a rotation of the field of view leads to more pronounced partial volume effects and therefore thicker strands.

Lindsay Walker^1,2, Michael Curry¹, Amritha Nayak^1,2, Nicholas Lange³, Carlo Pierpaoli¹, and The Brain Development Cooperative Group^4
1PPITS/STBB/NICHD, NIH, Bethesda, MD, United States, ²CNRM, USUHS, Bethesda, MD, United States, ³Harvard Schools of Medicine and Public Health, Boston, MA, United States, ⁴www.NIH-PediatricMRI.org

The NIH MRI Study of Normal Brain Development (PedsMRI) aims to study human brain development (birth – 18 years) using MRI modalities including diffusion tensor imaging (DTI). In preparation for the upcoming PedsMRI data release, including DTI data for the first time, we performed an analysis of variability on the PedsMRI phantom DTI data with the aim of providing high quality DTI imaging data for the public data repository.

Daniel M Krainak¹, Jennifer L Hufton², Nadia M Biassou³, David M Thomasson², and Sunder S Rajan^1
1CDRH/OSEL/DP, U.S. Food & Drug Administration, Silver Spring, MD, United States, ²NIAID, NIH, Bethesda, MD, United States, ³CC, NIH, Bethesda, MD, United States

We assess the reproducibility and reliability of diffusion tensor imaging (DTI)-derived metrics, such as fractional anisotropy, using a previously reported polyester fiber based anisotropic phantom. Our results indicate that DTI-metrics are stable across time (days) and slice location when acquired perpendicular to the fiber orientation. The measurements (fractional anisotropy, FA; axial diffusivity, AD; and radial diffusivity, RD) obtained from various fiber orientations were not consistent, and possibly influenced by the orientation of the fibers relative to the angulation of the acquisition slice.

Els Fieremans¹, Antonio Pires¹, and Jens H. Jensen^1
1CBI, Radiology, New York University School of Medicine, New York, NY, United States

Heat-treated heavy dairy cream is shown to have a diffusion coefficient and a diffusional kurtosis similar to values for in vivo brain and may therefore serve as a practical and inexpensive phantom for testing diffusional kurtosis imaging (DKI) protocols intended for neuroimaging applications. A feature of the phantom is that the fat/water shift may be exploited to provide a nontrivial consistency check. This phantom may be particularly useful as a calibration reference for multicenter and longitudinal investigations.

Birgitte Fuglsang Kjølby¹, Søren Christensen², Irene Klærke Mikkelsen¹, Kim Mouriden¹, Peter Gall³, Valerij G Kiselev³, and Leif Østergaard^1
1CFIN, Department of Neuroradiology, Aarhus University Hospital, Aarhus, Denmark, ²Department of Neurology and Radiology, University of Melbourne, Melbourne, Australia,³Department of Diagnostic Radiology, Medical Physics, University Hospital Freiburg, Freiburg, Germany

Perfusion DSC-MRI is becoming increasingly important in the study of neurological diseases, in particular acute stroke. Perfusion estimates are compared and studied to establish a perfusion threshold that may guide the selection of patients for thrombolysis. Therefore perfusion measurements must be optimized to detect and distinguish subtle levels of hypoperfusion. The precision and accuracy of derived perfusion values rely critically on the noise regularization used in the deconvolution process. We compare existing methods with a new method optimized for precision and show that it improves CBF differentiation in hypoperfused tissues both in individuals and in group studies.

David Bonekamp¹, Xu Li^1,2, Richard Leigh³, Peter C. van Zijl^1,2, and Peter B. Barker^1,2
1Radiology and Radiological Science, Johns Hopkins University School of Medicine, Baltimore, MD, United States, ²FM Kirby Research Center for Functional Brain Imaging, Kennedy Krieger Institute, Baltimore, MD, United States, ³Neurology, Johns Hopkins University School of Medicine, Baltimore, MD, United States

Perfusion-weighted imaging (PWI) performed using bolus injection of Gd-DTPA is traditionally analyzed based on changes in brain signal relaxation times (T2*,T2 and T1) associated with passage of contrast through the brain. Here, the feasibility of dynamic Quantitative Susceptibility Mapping (QSM) of cerebral blood flow (CBF) is shown. The expected linear susceptibility-contrast concentration relationship and the independence of contrast location (intra- or extravascular) on susceptibility are advantages over existing methods. The presented pilot study results show qualitative agreement between R₂^* and  CBF perfusion images for white matter. Current limitations exist for CBF determination in the brain periphery (gray matter).

Matthew R Orton¹, James A d'Arcy¹, Keiko Miyazaki¹, Dow-Mu Koh², David J Collins³, and Martin O Leach^1
1CR-UK and EPSRC Cancer Imaging Centre, Institute of Cancer Research, Sutton, Surrey, United Kingdom, ²Department of Radiology, Royal Marsden Hospital, Sutton, Surrey, United Kingdom, ³Clinical MRI Unit, Royal Marsden Hospital, Sutton, Surrey, United Kingdom

DSC-MRI methods are designed to measure vascular properties by analyzing the first-pass curves obtained from dynamic T2*-weighted imaging. The presence of recirculation and leakage will bias these measures, which can be reduced by cropping the data to exclude these features. However, useful information may be contained in the cropped data and manually selecting the cropping point is time consuming. In this abstract we present an empirically motivated model that can be fitted to DSC-MRI data that avoids the need to define cut-off times and accurately fits all the data from all pixels, including first-pass, recirculation and leakage components.

Amit Mehndiratta¹, Bradley J MacIntosh², David E Crane², Stephen J Payne¹, and Michael A Chappell^1
1Institute of Biomedical Engineering, University of Oxford, Oxford, Oxfordshire, United Kingdom, ²Medical Biophysics, University of Toronto, Toronto, ON, Canada

A vascular model-based Bayesian solution for perfusion quantification of DSC-MRI has previously been proposed. However, estimates from the method are potentially biased due to the use of priors estimated from the data within the analysis. Here a modification is proposed that introduces a further parameter into the model resulting in higher accuracy and providing an independent estimate of MTT.

Matthew R Orton¹, James A d'Arcy¹, Keiko Miyazaki¹, Nina Tunariu^1,2, David J Collins^1,2, and Martin O Leach^1
1CR-UK and EPSRC Cancer Imaging Centre, Institute of Cancer Research, Sutton, Surrey, United Kingdom, ²Clinical MRI Unit, Royal Marsden Hospital, Sutton, Surrey, United Kingdom

First-pass curves from DSC-MRI data can be characterized by fitting them to a model, from which summary parameters can be derived – in this case the procedure is effectively a de-noising operation. A less widely used application is to fit a model that includes specific parameters describing the tissue properties, in a similar manner to that routinely used with DCE-MRI data. Either way, the success of the technique depends on how well the model describes the data. We propose a novel model to describe DSC-MRI first-pass data, and demonstrate that it gives improved fits compared with two established models.

Michael A Chappell^1,2, Amit Mehndiratta¹, Stephen J Payne¹, and Fernando Calamante^3
1Institute of Biomedical Engineering, University of Oxford, Oxford, United Kingdom, ²FMRIB Centre, University of Oxford, Oxford, United Kingdom, ³Brain Research Institute, Florey Neuroscience Institutes, Melbourne, Victoria, Australia

Macro vascular (MV) contamination from contrast agent in major arteries is a significant source of bias in Dynamic Susceptibility Contrast (DSC) perfusion measurements. In this work we propose a model-based approach for MV correction that includes a MV component within a previously proposed vascular model for DSC data.

Haoyu Wang¹, Yanming Yu², Jiani Hu³, E. Mark Haccke³, Wei Xing⁴, and Shanglian Bao^1
1Beijing Key Lab of Medical Physics and Engineering, Peking University, Beijing, China, ²1Logging technique research institute, great wall drilling company, China National Petroleum Corporat, Beijing, China, ³Department of Radiology, Wayne State Univerisity, Detroit, MI, United States, ⁴Department of Radiology, Third Affiliated Hospital, Suzhou University, Changzhou, Jiangsu, China

Dual-echo pulse sequence has been proposed to simultaneously acquired T₁ - DCE and T₂^* - DSC MRI. However, there is no comparative investigation of T₂^* effects on permeability and T₁ effect on perfusion between normal and tumor tissues of brain. The goal of this work is to fill the gap. Preliminary results show that the difference of relative rCBV for healthy tissue with and without T₁ correction is about 20%, but for tumor it is about 153%. The results indicate that T₂^* effects on permeability and T₁effects on perfusion are much greater in tumor than in normal brain tissue.

Ke Zhang¹, Seong Dae Yun¹, Irene Neuner^1,2, Christian Filss¹, Karl-Josef Langen¹, and Nadim Joh Shah^1,3
1Institute for Neuroscience and Medicine, Medical Imaging Physics (INM-4), Forschungszentrum Jülich, Jülich, Germany, ²Faculty of Medicine, Department of Psychiatry and Psychotherapy, JARA, RWTH Aachen University, Aachen, Germany, ³Faculty of Medicine, Department of Neurology, JARA, RWTH Aachen University, Aachen, Germany

For dynamic susceptibility contrast (DSC) based perfusion, parallel EPI (pEPI) has been the traditional method for readout. However, image artefacts such as distortions in the peripheral region are still problematic. EPIK which combines the keyhole imaging scheme with multishot EPI approaches has been presented to reduce the geometric distortion. In this work, we compare EPIK and pEPI with T₂^*-weighted images in a dual-shot contrast agent DSC study. Three representative human brain tumour cases demonstrate that the feasibility of DSC using EPIK. EPIK can offer less distortion in peripheral region, higher perfusion contrast in DSC, and same temporal resolution as pEPI.

Thomas Christen¹, Deqiang Qiu¹, Wendy Wei Ni¹, Georges Hankov¹, Zungho Zun¹, Michael Moseley¹, and Greg Zaharchuk^1
1Department of Radiology, Stanford University, Stanford, California, United States

Cerebral blood volume (CBV) maps are usually acquired using Dynamic Susceptibility Contrast imaging which inherently limits the spatial resolution and signal to noise ratio of the images. In the present study, we used ferumoxytol (AMAG Pharmaceuticals, Inc., Cambridge, MA) an FDA-approved compound to obtain steady-state high-resolution quantitative CBV maps in 5 healthy volunteers. The results show excellent contrast between white and gray matter as well as fine highly-detailed vascular structures. An average blood volume of 4% was found in the brain of all volunteers, consistent with prior literature values.

C. Chad Quarles¹, Feng Weng¹, Mohammed Tantawy¹, Rosie Jiang², Keiko Takahashi², Chuan-Ming Hao², Todd Peterson¹, Raymond Harris², and Takamune Takahashi^2
1Institute of Imaging Science, Vanderbilt University, Nashville, TN, United States, ²O'Brien Mouse Kidney Physiology and Disease Center, Vanderbilt University, Nashville, TN, United States

Renal relative blood volume (RBV) measurements could provide a valuable tool to characterize abnormal renal perfusion in mouse models of kidney disease. We optimized MRI methods for spin-echo based RBV mapping in mouse kidneys and assessed the reproducibility of RBV measurements acquired on consecutive days. The mean kidney RBV measured on consecutive days was 19.97 ± 1.50 and 19.86 ± 1.62, yielding a concordance correlation coefficient of 0.94, indicating that this approach is highly reproducible. Interestingly, the use of a spin-echo sequence preferentially weighted towards microvessels, revealed that microvascular RBV within regions of the medulla were higher than that found in cortex.

Marcel Oei¹, Bozena Goraj¹, Anton Meijer¹, Jan-Jurre Mordang¹, Albert Idema², Sandra Boots-Sprenger³, Hendrik Laue^1,4, and Mathias Prokop^1
1Radiologie, Radboud University Nijmegen Medical Centre, Nijmegen, Gelderland, Netherlands, ²Neurosurgery, Radboud University Nijmegen Medical Centre, Nijmegen, Gelderland, Netherlands, ³Pathology, Radboud University Nijmegen Medical Centre, Nijmegen, Gelderland, Netherlands, ⁴Fraunhover Mevis, Bremen, Germany

rCBV values are normalized to an internal reference standard. The normalization step of rCBV calculation is subjective and introduces variability. 2 observers measured rCBVtumor derived from DSC-MR images in 17 glioma patients on 2 occasions. rCBV was normalised using 6 internal reference standards of normal apearing white and gray matter. Inter- and intraobserver reproducibility were quantified with intraclass correlation coefficient (R) and the coefficient of variation (CV). Inter- and intraobserver CV ranged from 7% - 17%. ICC showed good intra- (R=.69-.85) and interobserver variability (R=.76) for contralateral putamen. The least added variability was gained by normalization in the contralateral putamen.

Natenael B Semmineh¹, Junzhong Xu², Jerry Boxerman³, and C Chad Quarles^2
1institute of imaging science, vanderbilt university, nashville, TN, United States, ²institute of imaging science, Vanderbilt university, ³Alpert Medical School of Brown University

Synopsis: 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 ∆R2* is non-linearly related to cell density, with the greatest influence occurring at cell volume fractions between 40 – 60%. We also show that the spacing of cells within a voxel can alter ∆R2* values by as much as 50% for spacing variations on the order of the cell size.

Nicolas Pannetier^1,2, Clement Debacker^1,3, Franck Mauconduit^1,3, Christen Thomas⁴, and Emmanuel Barbier^1,3
1Grenoble Institut des Neurosciences, Universite Joseph Fourier, Grenoble, France, ²Department of Radiology & Biomedical Imaging, University of California, San Francisco, CA, United States, ³U836, INSERM, Grenoble, France, ⁴Department of Radiology, Stanford University, Stanford, CA, United States

We investigated how the spatial distribution of the vessels impacts the relaxation rate T2* of the MR signal. We used simulation in 2D with different constraints on the vessels distribution. We found that at 7T, the distribution of the vessels may impacts for about 8% the SO2 estimates. Moreover we found an interesting linear relationship between T2* and a spatial frequency that characterizes the distribution of the vessels.

Wendy W Ni^1,2, Deqiang Qiu¹, Thomas Christen¹, Heiko Schmiedeskamp¹, Roland Bammer¹, Michael E Moseley¹, and Greg Zaharchuk^1
1Department of Radiology, Stanford University, Stanford, California, United States, ²Department of Electrical Engineering, Stanford University, Stanford, California, United States

In this study, we describe our initial human experience with brain perfusion imaging using the ultrasmall paramagnetic iron oxide (USPIO) ferumoxytol as a T2* agent in steady state. We found that both ΔR2* and ΔR2 have strong linear dependence on dosage, and that the effect of ferumoxytol on T2 decay is relatively small. We also demonstrated the use of saline as a dilution bolus following ferumoxytol injection to measure relative cerebral blood volume in humans, finding low SNR and physiological effects to be the main issues in this technique.

Swati Rane¹, Samet Kose², Stephan Heckers², Malcolm Avison¹, and John C Gore^1,3
1Radiology and Radiological Sciences, Vanderbilt University Institute of Imaging Science, Vanderbilt University, Nashville, TN, United States, ²Psychiatry, Vanderbilt University, Nashville, TN, United States, ³Biomedical Engineering, Vanderbilt University, Nashville, TN, United States

Cihat Eldeniz¹, Weili Lin², and Hongyu An^2
1unc at chapel hill, chapel hill, nc, United States, ²UNC at Chapel Hill

Paramagnetic deoxyhemoglobin within capillaries and veins can induce magnetic field variations around these vessels. MR signal decay is governed by both static and diffusion induced dephasing effects. It has been suggested that diffusion effects highly depends on MR pulse sequence, echo time, and vessel size. Previous studies have proposed models with simplistic assumptions such as uniform vessel radii within a voxel, or explore the morphology of cerebral cortex microvasculature that might help modeling; however, a realistic mathematical model that can embrace the vast variety of microvasculature in various regions of the human brain has not been established to date. This also complicates the quantification of diffusion effects on the measurement of hemodynamic parameters. In this study, we aimed to shed light on how diffusion affects the MR signal under different vascular configurations pertaining to distinct regions.

Tobias Heye¹, Daniel T. Boll¹, Mustafa Bashir¹, and Elmar M. Merkle^1
1Department of Radiology, Duke University Medical Center, Durham, NC, United States

DCE-MRI pharmacokinetic parameter (Ktrans, Kep, Ve, iAUGC) calculation requires the T1 relaxation time (RT) of the tissue to convert signal intensity into a gadolinium concentration. Commercially available DCE-MRI post-processing software allow T1-RT measurement by variable flip angle sequences or input of a reference value. This study assesses the differences between calculations by both methods in 15 DCE-MRI cases. Additionally the behavior of pharmacokinetic parameters with changing T1-RT is measured. There is a considerable difference (6.6-54.9%) between calculations using T1-RT measurement vs. a reference value. Increasing T1-RT yield lower pharmacokinetic parameter values within the same DCE-MRI data set.

Sandeep N Gupta¹, Ehud J Schmidt², Robert Mulkern³, Andriy Fedorov², Ileana Hancu¹, Yingxuan Zhu¹, Clare Tempany-Afdhal², and Fiona Fennessy^2
1GE Global Research Center, Niskayuna, NY, United States, ²Radiology, Brigham and Women's Hospital, Boston, MA, United States, ³Radiology, Boston Children's Hospital, Boston, MA, United States

Dynamic Contrast Enhanced MRI has shown promise in non-invasive assessment of tumor vascular properties with applications in prostate cancer staging and treatment monitoring. Analysis of DCE requires knowledge of pre-contrast T1 maps. T1 mapping using Variable Flip Angle imaging is inaccurate due to B1 inhomogeneity. We present a simple approach for correcting VFA T1 maps by incorporating knowledge of approximate T1 of known reference tissue. This method can be used to produce clinically usable T1 maps in prostate. Our method was tested in 8 prostate patients and verified using a multiple-TR approach and the Bloch-Siegert B1 mapping method.

Irene Klærke Mikkelsen¹, David Alberg Peters², and Anna Tietze^1,3
1Centre for Functionally Integrative Neuroscience, Aarhus University, Aarhus, Denmark, ²Dept. of Clinical Engineering, Central Denmark Region, ³Dept. of Neuroradiology, Aarhus University Hospital

Dynamic Contrast Enhanced Perfusion Weighted Imaging (DCE-PWI) is usually measured with a 3D FLASH sequence. The required baseline T1 measurement is also measured with the 3D sequence, and for the sake of speed, the measurement is often performed by measuring the signal for a range of flip angles instead of as a function the inversion (or saturation) time. This work investigates how off-set in flip angles in the presence of B1 inhomogeneities propagates into large errors in the T1 estimates. The errors are markedly reduced when the measurement is performed as a function of time, even though a preparation pulse is then needed.

Anita Banerji^1,2, Alexandra Morgan^1,2, Yvonne Watson^1,2, Giovanni A Buonaccorsi^1,2, and Geoff J M Parker^1,2
1Imaging Sciences, University of Manchester, Manchester Academic Health Science Centre, Manchester, United Kingdom, ²Biomedical Imaging Institute, Manchester, United Kingdom

Voxel wise fitting of the extended Kety model to DCE-MRI data allows estimation of a median K^trans value for a tumour. However, the estimation of K^trans for liver tumours is likely to be effected by respiratory motion. Using synthetic data with motion based on biomechanical modelling we have assessed the robustness of model fitting to motion and the benefit of using a model-driven registration. We have found that median K^trans values are largely insensitive to motion and that heterogeneity measures are likely to benefit from translation only registration. However, more complex registrations should be used with caution.

Eric Bultman¹, Debra Horng², Ethan K Brodsky², Kevin M Johnson², Walter F Block^1,2, and Scott B Reeder^1,3
1Biomedical Engineering, University of Wisconsin-Madison, Madison, WI, United States, ²Medical Physics, University of Wisconsin-Madison, Madison, WI, United States,³Radiology, University of Wisconsin-Madison, Madison, WI, United States

Quantitative perfusion imaging of hypervascular liver lesions performed over multiple breath-holds (BHs) requires compensation for motion between BH periods. In this work, we utilize a contrast-enhanced 3D radial acquisition with 4s true temporal footprint to image the entire abdomen of four healthy volunteers and four patients with hepatocellular carcinoma. We demonstrate that suspension of respiratory motion during sequential BHs produces images free of motion artifact, and that rigid registration of BH-averaged image volumes provides excellent correction for subject motion and variations in residual lung volume between BHs.

Yiqun Xue¹, Jiangsheng Yu¹, Mark A Rosen¹, and Hee Kwon Song^1
1Radiology, University of Pennsylvania, Philadelphia, Pennsylvania, United States

Temporally constrained reconstruction (TCR) has recently been developed for dynamic MR imaging to achieve high image quality at high undersampling factors. However, whether this technique can achieve accurate quantitative MRI measurements has not been investigated. In this work, an enhanced simulation experiment was conducted to evaluate of the accuracy of TCR for the assessment of DCE-MRI perfusion parameters. Compared to simple undersampled and KWIC methods, the TCR method yields more accurate and precise perfusion measures, and is much less sensitive to different levels of image noise, Ktrans values, and tumor sizes.

Magne Mørk Kleppestø¹, Christopher Larsson¹, Raimo Aleksi Salo¹, Jonas Vardal¹, Kine Mari Bakke², Knut Lote³, Petter Brandal³, I. Andre Rasmussen¹, and Atle Bjørnerud^1,2
1The Intervention Centre, Oslo University Hospital, Rikshospitalet, Oslo, Norway, ²Department of Physics, University of Oslo, Oslo, Norway, ³Departement of Neuro-oncology, Oslo University Hospital, Rikshospitalet, Oslo, Norway

Dynamic contrast-enhanced MRI is an established method of assessing brain hemodynamics and blood-brain-barrier integrity in brain tumors. The kinetic properties of the tumor is estimated from the measured signal intensity change following the administration of a contrast agent. However, there is no consensus on the temporal parameters required to obtain accurate and reproducable estimates from such measurements. Temporal resolution and total sampling duration affects the quality of the resulting parameter maps and is of practical importance in the clinic. The present work investigates how these parameters are best chosen to achieve reliable results in high-grade brain tumors.

Xia Zhao¹, Yiqun Xue¹, Ramesh Paudyal¹, Jiangsheng Yu¹, Hyun Seon Kang¹, Sarah Englander¹, Thomas Ferrara¹, Harish Poptani¹, Amit Maity², Ramesh Rengan², Rosen Mark¹, Ping Wang¹, and Hee Kwon Song^1
1Department of Radiology, Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania, United States, ²Department of Radiation Oncology, Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania, United States

In this work, the standard concentration-based method of computing tissue perfusion was compared with a simpler signal difference method in patients with lung tumors using a 3D radial DCE-MRI sequence. Repeatability of K^trans is slightly higher in the standard method, and is similar to those observed in previous studies using Cartesian based sequences. The results indicate that even in the absence of intrinsic T1 or accurate flip angle information perfusion parameters can be determined using the signal difference methodology with only minor loss of measurement precision.

Ping Wang¹, Jiangsheng Yu¹, Yiqun Xue¹, Xia Zhao¹, Mark Rosen¹, and Hee Kwon Song^1
1Radiology, University of Pennsylvania, Philadelphia, PA, United States

In conventional DCE-MRI, flip angle accuracy is important in accurately determining perfusion parameters. It has been shown previously that large flip angle deviations can occur throughout the body, resulting in erroneous perfusion measures. Recently, a simpler method for computing perfusion was proposed, in which the absolute signal difference is used without the need for baseline T1, contrast agent concentration measurements, or flip angle assumptions. In this study numerical simulations were used to systematically compare the performance of the signal difference methodology with the conventional technique in the presence of flip angle deviations.

Rejean Lebel¹, Eric Poulin¹, and Martin Lepage^1
1Centre d'imagerie moléculaire de Sherbrooke, Université de Sherbrooke, Sherbrooke, Quebec, Canada

We developed a new method that exploits the existence of blood/tissue equilibrium (BTE) points in DCE-MRI to automatically extract the shape of the AIF from the tissue concentration (CT) maps. This method is (1) non invasive, (2) does not require the assumption of Ktrans or ve in a specific tissue, (3) is automated (no user input/ROI tracing) and (4) corresponds to the average AIF as “perceived” by the voxels inside the volume of interest. In animals, the method is shown to yield accurate AIFs compared to manual blood samples analysed by inductively coupled plasma mass spectrometry.

Jin Zhang¹, Alana Amarosa¹, Andrew B. Rosenkrantz¹, and Sungheon Kim^1
1Radiology, New York University School of Medicine, New York, New York, United States

Arterial Input Function (AIF) plays an important role in pharmacokinetic model analysis of dynamic contrast enhanced (DCE)-MRI data. Reference tissue approaches have been proposed as a means to estimate AIF using a two compartment model and literature values of transfer constant and interstitial space volume. In this study, neural network was introduced to delineate reference tissue concentration curve and consequently to reconstruct AIF. Clinical and preclinical DCE-MRI studies were performed to compare different reference tissue approaches. The results demonstrated the proposed method has improved reproducibility as well as flexibility in estimating different shapes of AIF.

Julio Cárdenas-Rodríguez¹, Christine M Howison², and Mark D. Pagel^3
1Chemistry and Biochemistry, The Arizona Cancer Center, University of Arizona, Tucson, AZ, United States, ²Arizona Research Laboratories, University of Arizona, Tucson, AZ, United States, ³Department of Biomedical Engineering, Chemistry and Biochemistry, and The Arizona Cancer Center, University of Arizona, Tucson, Arizona, United States

We have developed a new Linear Reference Region (LRRM) model for DCE-MRI, and compared its performance with the standard Non-Linear Model (NLRRM) using simulations and pre-clinical DCE-MRI data. The LRRM estimate the relative Ktrans of tow tissues more accurately than the standard NLRM at coarser temporal resolution (128 sec vs. 32 sec), and lower SNR (15 vs 30). These results show that our LRRM algorithm can be used to translate the Reference Region Model to clinical setting.

Matthias Christian Schabel^1,2
1Advanced Imaging Research Center, Oregon Health & Science University, Portland, OR, United States, ²Radiology, Utah Center for Advanced Imaging Research, Salt Lake City, UT, United States

We have developed a four site linear exchange model (4SLX) for pharmacokinetic analysis of DCE-MRI data. This model incorporates separate compartments for intracellular, extracellular extravascular, blood plasma, and erythrocyte water. Comparison of parameter estimates obtained using a simpler two-compartment exchange (2CX) model that assumes the fast water exchange limit with those obtained from the 4SLX in a human glioma show modest but statistically significant changes in parameter estimates.

Manav Bhushan¹, Julia Schnabel¹, Lydia Tanner², Fergus Gleeson³, Sir Michael Brady², and Mark Jenkinson^4
1Institute of Biomedical Engineering, Oxford University, Oxford, United Kingdom, ²Department of Radiation Oncology and Biology, Oxford University, ³Department of Radiology, Churchill Hospital, ⁴Centre for Functional MRI of the Brain, Oxford University

We present a novel Bayesian framework for non-rigid motion correction and Pharmacokinetic (PK) parameter estimation in dynamic contrast-enhanced MRI. We use our algorithm to co-register image volumes from dceMRI scans acquired for Colorectal cancer patients before, and after 5 weeks of chemoradiotherapy and estimate the PK-parameter maps for the same. We then classify each patient as a responder or non-responder to therapy on the basis of the difference between the pre- and post-therapy distributions of PK-parameters. We show that there is a significant benefit in using motion-correction within this framework, and also compare the results obtained using two AIFs.

Gregory Z. Ferl¹, Shiv J. Acharya¹, James P.B. O'Connor², Geoffrey J.M. Parker², and Ruediger E. Port^1
1Development Sciences, Genentech, Inc., South San Francisco, CA, United States, ²Imaging Sciences Research Group, School of Medicine, University of Manchester, Manchester, United Kingdom

Here, we present a population level mathematical model that describes the time course of K^trans response to a single dose of the therapeutic monoclonal antibody bevacizumab based on Dynamic Contrast Enhanced MRI data from a previously published post-licensing study, where K^trans is a composite measure of vascular permeability to contrast agent, surface area and rate of tissue perfusion. We show that the model is capable of describing a K^trans response characterized by rapid decrease during the first 4 hours post-dose, followed by a slower return to baseline over 12 days.

Ram KS Rathore¹, Rakesh Kumar Gupta², Prativa Sahoo¹, Rishi Awasthi², Divya Rathore¹, and Bhaswati Roy^2
1Mathematics & Statistics, Indian Institute of Technology, Kanpur, Kanpur, Uttar Pradesh, India, ²Radiodiagnosis, Sanjay Gandhi Post Graduate Institute of Medical Sciences, Lucknow, India, Lucknow, Uttar Pradesh, India

A three compartment tissue model - Leaky Tracer Kinetic Model (LTKM) is introduced to take care of the varied tissue natures in the whole body applications of DCE-MRI. In particular it takes care of vasculature in accessible leakage and provides a consistent determination of the perfusion parameters. Applications of it elsewhere have led to a better grading of tumors.

Mikael Fredrik Forsgren^1,2, Olof Dahlqvist Leinhard^1,2, Gunnar Cedersund^3,4, and Peter Lundberg^1,2
1Depts of Radiation Physics, Linköping University and Radiation Physics, UHL County Council of Ostergotland, Linköping, Sweden, ²Center for Medical Image Science and Visualization (CMIV), Linköping University, Linköping, Sweden, ³Depts of Clinical and Experimental Medicine (IKE), Diabetes and Integrated Systems Biology, Linköping University, Linköping, Sweden, ⁴School of Life Sciences, Freiburg Institute of Advanced Sciences, Freiburg, Germany

Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI)using hepatocyte-specific contrast agents (CA) has been proposed as promising methods for liver function characterization. A human whole body pharmacokinetic model for DCE-MRI analysis has previously been proposed, there is however issues with parameter uncertainties in models aiming towards mimicking complex biological systems. The aim of this study was to derive ad hoc constraint that can reduce the model uncertainty. Our results suggest that a lowest amount of CA residing within the blood pool after 3 h is a likely candidate that might be used on a variety of liver DCE-MRI models.

Matthias Christian Schabel^1,2
1Advanced Imaging Research Center, Oregon Health & Science University, Portland, OR, United States, ²Radiology, University of Utah, Salt Lake City, UT, United States

We describe a new impulse response model for analysis of DCE-MRI that mathematically unifies the Tofts-Kety (TK), Extended Tofts-Kety (ETK), Adiabatic Tissue Homogeneity (ATH), and Two Compartment Exchange (2CX) models. This model introduces a distribution of capillary transit times and includes a parameter describing the shape of this distribution that continuously varies between the ATH and 2CX limits. In vivo results in three human brain tumors suggest that this shape parameter takes on values spanning the range of possible values and may provide an imaging biomarker characterizing disorder of the capillary vascular bed.

Dennis Lai Hong Cheong^1,2, Thian Chor Ng^3,4, Bo Zhang^1,5, Bingwen Zheng¹, Eugene Mun Wai Ong⁴, and Soo Chin Lee^6,7
1Clinical Imaging Research Center, A*STAR & National University of Singapore, Singapore, Singapore, ²Neuroradiology Department, National Neuroscience Institute, 308433, Singapore, ³Clinical Imaging Research Center, A*STAR & National University of Singapore, Singapore, Singapore, Singapore, ⁴Department of Radiology, National University of Singapore, 119074, Singapore, ⁵Quantitative Image Processing Group, SBIC/A*STAR, 138671, Singapore, ⁶Department of Haematology-Oncology, National University Health System, 119074, Singapore, ⁷Cancer Science Institute, 117456, Singapore

Pixel-by-pixel concentration time curves from a breast tumor were extracted and analyzed by two compartmental models (Tofts model and extended Tofts model), and two distributed parameter models (adiabatic approximation to tissue homogeneity model and two-compartment axially distributed parameter model). All models are able to fit the data although distributed parameter models have slightly better fittings. Parameters Ktrans and ve and impulse residue functions of the models were being compared. In pixel-by-pixel analysis of DCE MRI data in tumors, our preliminary data suggests that distributed parameter models are better than compartmental models by having better fittings and more stable parameter values.

Jin Zhang¹, and Sungheon Kim^1
1Radiology, New York University School of Medicine, New York, New York, United States

Due to lack of gold standard to measure pharmacokinetic model parameters, it remains uncertain how accurately these model parameters represent the tissue microenvironment. Recently pharmacokinetic models with inter-compartmental water exchange effects have been introduced. However, it has not been shown how reliably the parameter estimation can be done with such models with additional parameters for water exchange. Hence, we investigated the influence of initial values and noise on the model parameter estimation with simulated MRI data. Our results demonstrate that numerical simulation studies can be used to assess the accuracy and precision in the model parameter estimation.

Benjamin Leporq¹, Sorina Camarasu-Pop¹, Frank Pilleul¹, and Olivier Beuf^1
1CREATIS; CNRS UMR 5220; INSERM U1044; INSA-Lyon; UCBL Lyon 1, Villeurbanne, Rhône-Alpes, France

This study presents a MR acquisition protocol and a processing method using distributed computing on the European Grid Infrastructure (EGI) to allow 3D liver perfusion parametric mapping after DCE-MR imaging with the MS-325 blood pool agent. In this preliminary study, processing speed, reproducibility and accuracy were assessed. The presented method allows 3D liver perfusion quantification in a reasonable processing time, suitable for clinical study in a research context. While the distributed processing method was validated compared to ROI-based quantification, such fully automatic processing requires higher image quality which is now achievable on the latest 3T MRI systems available.

Yuxiang Zhou¹, Manickam Kumaravel¹, Vipulkumar S Patel¹, Kazim Sheikh², and Ponnada A Narayana^1
1Diagnostic & Interventional Imaging, University of Texas Health Science Center at Houston, Houston, Texas, United States, ²Department of Neurology, University of Texas Health Science Center at Houston, Houston, Texas, United States

The effect of spatial resolution, number of diffusion gradient encoding directions (DGED), and number of repetitions on the visualization and quantification of peripheral nerves on the diffusion tensor imaging (DTI) in human forearm was investigated. The fractional anisotropy (FA) maps at higher spatial resolution allowed the visualization of superficial radial, median, and ulnar nerves consistently on all the scans, whereas only ulnar and median nerves were clearly visualized at lower spatial resolution. Our results indicate the visualization and the quantification of the diffusion anisotropy of forearm nerves is strongly influenced by the SNR, spatial resolution, and number of DGED. These studies help in the identification of optimum scan parameters for the DTI of peripheral nerves.

Teodora-Adriana Perles-Barbacaru¹, Michel Sarraf¹, Regine Farion¹, Marie France Nissou¹, Boudewijn van der Sanden¹, Francois Berger¹, and Hana Lahrech^1
1Grenoble Institute of Neurosciences, INSERM U836, Grenoble, France

A blood volume fraction (BVf) in a C6 rat brain tumor model is quantified with the Rapid Steady State T₁ (RSST₁) technique and Gd-DOTA using a novel pharmacokinetic model to account for transendothelial contrast agent leakage. The BVf is validated with a ΔR₂* steady state technique employing a blood pool contrast agent and histologically. The advantage of the RSST₁ technique over conventional dynamic contrast enhanced techniques is that it does not require measurement of the arterial input function, making BVf quantification straightforward and easy to apply in the clinical routine.

Yoni Cohen¹, Hagit Dafni¹, Inbal Biton¹, Tal Raz¹, and Michal Neeman^1
1Weizmann Institute of Science, Rehovot, Israel

 

Simon Walker-Samuel¹, Rajiv Ramasawmy¹, Peter Johnson², Jack Wells¹, Bernard Siow¹, Barbara Pedley², and Mark F. Lythgoe^1
1Centre for Advanced Biomedical Imaging, University College London, London, United Kingdom, ²Cancer Institute, University College London, United Kingdom

Tumours exhibit raised interstitial fluid pressure (IFP), which acts as a barrier to drug delivery and results in convection currents that radiate from the centre of the tumour to the periphery. We report a novel method named extra-vascular convectography (EVAC) for measuring interstitial fluid velocity (IFV), that does not require a contrast agent, and have evaluated it in two tumour models. The method consists of vascular nulling using a double inversion preparation, a recovery delay, and a velocity encoding readout. Using EVAC we have identified both radial and laminar interstitial convection patterns with a median velocity of 0.28 mm/s.

Maximilian N. Voelker¹, Jan M. Burg², Peggy Schlupp³, Ulf Maeder², Alexander M. Koenig¹, and Johannes T. Heverhagen^1
1Diagnostic Radiology, Philipps University, Marburg, Hessen, Germany, ²Institute of Medical Physics and Radiation Protection, University of Applied Sciences Giessen-Friedberg, Giessen, Germany, ³Institute of Biopharmaceutical Technology, University of Applied Sciences Giessen-Friedberg, Giessen, Germany

A topically applied submicron emulsion on porcine skin samples was used as a carrier system for the contrast agent. Contrast enhanced MRI enables the measurement of its penetration into the skin. Measuring the relaxation time change due to the penetration of the contrast agent into the skin allowed the quantification of the drug carrier system. Compared with common optical and analytical assessment of skin penetration, contrast enhanced MRI has an advantage in unlimited imaging depth. Moreover, it provides an easy and precise quantification despite the worse resolution.

Jill Fredrickson¹, Natalie Serkova², Richard Carano¹, Shelby Wyatt¹, Andrea Pirzkall¹, Colin Weekes², Jeffrey Silverman³, Lee Rosen⁴, and Alex de Crespigny^1
1Genentech, South San Francisco, CA, United States, ²University of Colorado Hospital, Denver, CO, ³Landmark Imaging, LLC, Santa Monica, CA, ⁴Premiere Oncology, Santa Monica, CA

Vessel size imaging (VSI) metrics acquired using ferumoyxtol and stock pulse sequences in a Phase 1 patient population with advanced solid tumors are compared to pre-clinical results. Despite using half the ferumoxytol dose, % changes in both R2 and R2* are much greater in human liver metastases than in xenografts. Average vessel density is lower in humans due to the greater increase in ∆R2*. Clinical VSI is feasible using stock pulse sequences in a Phase 1 population; in particular, the high level of ferumoxytol in the normal liver does not prevent the measurement of VSI metrics in liver metastases.

Steven K.K. Chow¹, Ann D King¹, David K.W. Yeung¹, Jing Yuan¹, Kunwar Bhatia¹, Anil T Ahuja¹, Alexander C Vlantis², and Brian K.H. Yu^3
1Department of Imaging and Interventional Radiology, The Chinese University of Hong Kong, Sha Tin, N.T., Hong Kong, ²Department of Otorhinolaryngology Head and Neck Surgery, The Chinese University of Hong Kong, Hong Kong, ³Department of Clinical Oncology, The Chinese University of Hong Kong

We used DCE-MRI to examine whether this technique can be used to predict treatment response in head and neck cancer. We examined 19 primary and 21 metastatic nodal lesions. Despite no significant difference was found between Ktrans parameters of tumors with disease control and disease failure, there was a trend towards disease failure in tumors with a lower Ktrans mean, Ktrans skewness and Ktrans kurtosis. This study is ongoing and presently we could only report that there appears to be a trend towards treatment failure in primary and nodal metastases with lower Ktrans means.

Andrew B Gill¹, David J Bowden¹, Richard T Black¹, Andrew N Priest¹, Ilse Joubert¹, Martin J Graves¹, and David J Lomas^1
1Radiology, Addenbrooke's Hospital & University of Cambridge, Cambridge, United Kingdom

This pilot study investigated the ability of DCE-MRI indices to differentiate between tumour and ‘unaffected’ regions in the livers of patients with hepatocellular carcinoma. Data was collected from 6 patients and 9 healthy volunteers using a saturation-prepared dual-acquisition sequence, and analysed with a dual-input single-compartment pharmacokinetic model. The arterial fraction was found to be significantly higher in tumour tissue than in ‘unaffected’ tissue and significantly higher than that in volunteers. Mean total perfusion was found to be lower in tumours than in the livers of volunteers, probably due to the degree of necrosis in the tumour central regions.

Alana R. Amarosa¹, Hersh Chandarana¹, Stella K. Kang¹, Mary T. Bruno¹, Konstantinos Arhakis¹, William C. Huang², Edgar F. Suan¹, and Sungheon Kim^1
1Radiology, NYU School of Medicine, New York, New York, United States, ²Urology, NYU School of Medicine, New York, New York, United States

The purpose of this study was to investigate the feasibility of dynamic contrast enhanced perfusion MRI with pharmacokinetic modeling to generate and analyze parameters, ktrans, ve, vp, for discrimination of renal cell cancer (RCC) subtypes. In this prospective study 22 patients with known or suspected renal tumors underwent preoperative DCE-MRI at 1.5T. The generalized kinetic model was used for pharmacokinetic analysis to obtain ktrans, ve, vp for total of 23 lesion. Significant differences were found in ktrans between clear cell RCC and oncocytic tumors (p = 0.038) and between clear cell and papillary RCCs (p = 0.074).

Wenchao Cai¹, Feiyu Li¹, Yi Dang², Jue Zhang^2,3, Xiaoying Wang^1,2, and Xuexiang Jiang^1
1Radiology, Peking University First Hospital, Beijing, Beijing, China, ²Academy for Advanced Interdisciplinary Studies, Peking University, Beijing, Beijing, China, ³College of Engineering, Peking University, Beijing, Beijing, China

Purpose:To identify age-related changes and differences in the kinetic parameters obtained from DCE-MRI within the prostate gland in heathy adult men. Methods:A total of 68 heathy male volunteers were recruited into the study. The patients were divided into three age groups(10~30years,31~50years and >50years). The DCE-MRI examinations were performed on a clinical 3.0 T MR scanner. Post-procession of DCE-MRI data was conducted in the Matlab to obtain the Ktrans,Vep and Ve values. Results:The Variance(ANOVA) analysis identified significant differences of Ktrans , Kep among the three age groups within the PZ (all P<0.05),While Ve showed no statistically significant difference. Conclusion:Our results shows that there is a decrease trend of Ktrans and Kep along with age increase within the prostate PZ of healthy adults.

Rishi Awasthi¹, Rakesh Kumar Gupta¹, Deepak Tripathi², Vikas Agarwal², Vinita Agrawal³, Prativa Sahoo⁴, and Ram KS Rathore^4
1Radiodiagnosis, Sanjay Gandhi Post Graduate Institute of Medical Sciences, Lucknow, India, Lucknow, Uttar Pradesh, India, ²Immunology, Sanjay Gandhi Post Graduate Institute of Medical Sciences, Lucknow, India, Lucknow, Uttar Pradesh, India, ³Pathology, Sanjay Gandhi Post Graduate Institute of Medical Sciences, Lucknow, India, Lucknow, Uttar Pradesh, India, ⁴Mathematics & Statistics, Indian Institute of Technology, Kanpur, Kanpur, Uttar Pradesh, India

This study was performed on 73 patients who had inflammation in knee joint with age ranging from 18-65 years. DTI and DCE-MRI were performed on a 3T MR scanner. On discriminant analysis, DCE derived BV and DTI derived FA values were found to be significant discriminators between tubercular and non-tubercular inflammation with canonical correlation of 0.851. We were able to correctly classify a total of 98.6% of original grouped cases into tubercular and non-tubercular knee arthritis with very narrow range of overlapping values.

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

Territorial arterial spin labeling (TASL) has been known for its capability of noninvasively measuring regional blood flow supplied by a single or a subset of feeding arteries, which in the human brain is currently only achievable by using interventional digital subtraction angiography. The prior determination of target vessels for TASL demands familiarity of vascular anatomy and can be difficult when variations are caused by diseases. Here we propose to expand the use of pseudocontinuous ASL and harmonic encoding to detect flow territory, arterial source location, and off resonance with minimal prior planning.

Sung-Hong Park¹, Bumwoo Park¹, Jung-Hwan Kim¹, Tiejun Zhao², and Kyongtae Ty Bae^1
1Radiology, University of Pittsburgh, Pittsburgh, PA, United States, ²MR Research Support, Siemens Healthcare, Pittsburgh, PA, United States

Blood flow signals are commonly measured with arterial spin labeling (ASL) technique and represented in a scalar quantity. We tested the feasibility of mapping blood flow directionality in a vector and a tensor forms using a new ASL technique, alternate ascending/descending directional navigation. Blood flow directions mapped in the vector form was from brain center to lateral, anterior, and posterior and from feet to head. Although the directions of the primary eigenvector calculated from the tensor form were heterogeneous between regions, regional clusters of the same directionality were consistent across subjects. Further studies are necessary to validate the new measurement.

Aaron Oliver-Taylor¹, Chris Randell², Roger J Ordidge³, and David L Thomas^4
1Department of Medical Physics and Bioengineering, University College London, London, England, United Kingdom, ²PulseTeq Products Division, Renishaw PLC, United Kingdom,³Centre for Neuroscience, University of Melbourne, Melbourne, Victoria, Australia, ⁴Institute of Neurology, University College London, London, England, United Kingdom

Vessel selective continuous arterial spin labelling can be achieved using a single surface coil placed over one of the carotid arteries. However the labelling B₁ field extends across the neck causing contralateral labelling. Presented is a method using parallel transmission to nullify the B₁ field at the contralateral artery. Both phantom and in-vivo results show a reduction in contralateral labelling is possible, improving specificity without losing the advantageous high labelling efficiency and no magnetisation transfer effects of separate-coil CASL.

Jon-Fredrik Nielsen¹, and Luis Hernandez-Garcia^1
1Biomedical Engineering, University of Michigan, Ann Arbor, Michigan, United States

Arterial Spin Labeling (ASL) provides quantitative and reproducible measurements of cerebral blood flow, and is an attractive method for functional MRI. Most existing ASL fMRI protocols are based on 2D multislice or 3D spin-echo, and suffer from low image signal-to-noise ratio (SNR) or through-plane blurring. 3D ASL with multi-shot (segmented) readouts can improve the SNR efficiency relative to 2D multislice, and does not suffer from T2-blurring. We characterize the temporal SNR of a segmented 3D spiral ASL sequence, and investigate the effects of RF phase-cycling scheme and flip angle schedule on the ASL time-course signal. We conclude that RF-spoiled segmented 3D spiral ASL with a cubic flip angle schedule starting at 15o can produce high-quality functional activation maps with only modest through-plane blurring.

Li Zhao¹, Xiao Chen¹, Samuel W. Fielden¹, Frederick H. Epstein¹, John P. Mugler III², Josef Pfeuffer³, Manal Nicolas-Jilwan², Max Wintermark², and Craig H. Meyer^1,2
1Biomedical Engineering, University of Virginia, Charlottesville, Virginia, United States, ²Radiology, University of Virginia, Charlottesville, Virginia, United States, ³Siemens, Erlangen, Germany

To achieve high SNR, we develop 3DTSE with stack spiral acquisition. To shorten scan time, compress sensing is modified to perform on spiral trajectory. CBF and MTT are estimated from dynamic data properly and off-line reconstruction with acceleration rate of 3 shows little difference.

Xiaoyun Liang¹, Alan Connelly^1,2, Jacques-Donald Tournier^1,2, and Fernando Calamante^1,2
1Brain Research Institute, Florey Neuroscience Institutes, Melbourne, VIC, Australia, ²Department of Medicine, University of Melbourne, Melbourne, VIC, Australia

 

Marta Vidorreta¹, Ze Wang², Ignacio Rodríguez³, John A. Detre^4,5, and María A. Fernández-Seara^1
1Functional Neuroimaging Laboratory, CIMA, University of Navarra, Pamplona, Navarra, Spain, ²Department of Psychiatry, University of Pennsylvania, Philadelphia, Pennsylvania, United States, ³Instituto de Estudios Biofuncionales, UCM, Madrid, Spain, ⁴Department of Neurology, University of Pennsylvania, Philadelphia, Pennsylvania, United States, ⁵Deparment of Radiology, University of Pennsylvania, Philadelphia, Pennsylvania, United States

ASL can be implemented combining different labeling schemes and readout sequences. A consensus is emerging in the ASL community concerning the use of pseudo-continuous ASL as the optimum labeling strategy. However, there is no agreement to date on the choice of readout sequence. In this work, we have compared the performance of 3 different sequences: two background suppressed 3D fast spin echo based sequences (GRASE and FSE spiral) and the standard 2D EPI. The SNR of the 3D sequences is greatly improved with respect to 2D EPI. The spiral based readout offers a further increased in temporal SNR that translates in higher statistical power for task-activation studies.

David L Thomas¹, Enrico De Vita^1,2, Xavier Golay¹, and Maria A Fernandez-Seara^3
1Department of Brain Repair and Rehabilitation, UCL Institute of Neurology, London, London, United Kingdom, ²Lysholm Department of Neuroradiology, National Hospital for Neurology and Neurosurgery, London, London, United Kingdom, ³Center for Applied Medical Research, University of Navarra, Pamplona, Navarra, Spain

3D GRASE pCASL is a method for obtaining non-invasive perfusion maps of the brain with good SNR. A CPMG RF refocusing pulse phasing scheme is typically chosen for the 3D GRASE readout, to maximise robustness to B₁ inhomogeneity. However, this method fails when vascular crusher gradients are implemented, due to violation of the CPMG condition. In this work, we implement an alternative refocusing pulse phasing scheme which is insensitive to the initial phase of the transverse magnetisation and still robust to B₁ inhomogeneity, enabling use of vascular crusher gradients and improving the accuracy of perfusion quantification in 3D GRASE ASL.

Jan Petr¹, Georg Schramm¹, Frank Hofheinz¹, Jens Langner¹, Jörg Steinbach¹, and Jörg van den Hoff^1
1PET center, Department of Radiopharmacy, Helmholtz-Zentrum Dresden-Rossendorf, Dresden, Germany

Partial volume correction becomes an important issue in arterial spin labeling. It is usually done using the partial volume maps obtain by segmenting the high resolution T₁-weighted images. This brings additional issues with imperfections in segmentation and corrections of image distortion. We propose an alternative by using T₁ relaxation times obtained from the Look-Locker sequence to directly obtain the partial volume ratios. These ratios show higher agreement with the perfusion data than the standard T₁-weighted images segmentation.

Mustafa Cavusoglu^1,2, Rolf Pohmann², and Kamil Uludag^3
1Biomedical Engineering, ETH Zurich, Zurich, Switzerland, ²High Field MR Center, Max Planck Institute, Tuebingen, Germany, ³Department of Cognitive Neuroscience, Maastricht Brain Imaging Centre (MBIC), Maastricht, Netherlands

The arterial transit time is a key local variable which has to be included in the kinetic models to estimate the CBF (f) in arterial spin labeling. Due to cardiac pulsation, non-uniform cross-sectional flow profile and complex vessel networks, the distribution of transit delay time has a statistical nature instead of being uniform. In this study, we have investigated the regional effects of magnetization dispersion for varying distances between tagging and imaging regions by implementing pulsed (PASL), pseudo-continuous (PCASL) and dual-coil continuous (DC-CASL) ASL encoding schemes.

André Ahlgren¹, Ronnie Wirestam¹, Freddy Ståhlberg^1,2, and Linda Knutsson^1
1Dept. of Medical Radiation Physics, Lund University, Lund, Skåne, Sweden, ²Dept. of Diagnostic Radiology, Lund University, Lund, Skåne, Sweden

To improve estimated residue function characteristics and cerebral blood flow (CBF) quantification in arterial spin labeling (ASL), we recently proposed the implementation of nonlinear stochastic regularization (NSR) in model-free ASL. Here we present quantitative results from an improved implementation and compare it to the more common singular value decomposition (SVD) methods. SVD based methods yielded unrealistic residue functions and reasonable (cSVD) or low (oSVD) CBF estimates. NSR yielded realistic residue functions and robust perfusion maps but somewhat low CBF values.

Jia Guo¹, and Eric C. Wong^2
1Bioengineering, University of California San Diego, La Jolla, California, United States, ²Radiology and Psychiatry, University of California San Diego, La Jolla, California, United States

We explore here the dependence of the velocity selective ASL signal on the tagging time and the velocity cutoff. We note that the peak of the ASL signal occurs at lower TI than expected, suggesting significant tissue water exchange or a shorter than expected bolus length. We also note oscillations in the inflow curve in some subjects that may provide information on the rate of deceleration in arteries.

Zungho Zun¹, Brian A. Hargreaves¹, and Greg Zaharchuk^1
1Radiology, Stanford University, Stanford, CA, United States

We demonstrate that velocity-selective arterial spin labeling (VSASL) can be improved particularly for 2D multislice cerebral blood flow (CBF) imaging. (1) Stimulated echo removal using variable gradient spoiler reduced the temporal noise in CBF time series by 27%. (2) Improved slice profile using matched-phase RF pulse reduced the slice spacing to one third of the imaging slice thickness for near-contiguous multislice. (3) Timing optimization reduced the temporal noise by 34% due to increased SNR efficiency. (4) Off-resonance correction minimized blurring from spiral trajectory. These changes are expected to drastically improve the utility of VSASL for clinical studies.

Minghui Tang¹, and Toru Yamamoto^2
1Graduate School of Health Sciences, Hokkaido University, Sapporo, Hokkaido, Japan, ²Faculty of Health Sciences, Hokkaido University, Sapporo, Hokkaido, Japan

 

Alexander Graeme Gardener¹, and Peter Jezzard^1
1FMRIB Centre, University of Oxford, Oxford, United Kingdom

Dynamic Subtraction VASO has been proposed as a method of measuring cerebral blood volume. Paired images are acquired - in the first global inversion then slice-selective inversion are applied; the second employs two slice-selective inversions back-to-back. Images are acquired at blood null time (TI1). Subtracting the first set from second allows for quantification of CBV. This study extends this to acquire additional images at longer delay (TI2) with CBF weighting, improving time efficiency of the sequence and allowing for simultaneous CBV/CBF estimation, to give additional concurrent insight into cerebral haemodynamics. Values found are comparable to existing separate measurement methods.

Sophie Schmid¹, Esben T. Petersen², Jeroen Hendrikse², Andrew Webb¹, and Matthias J.P. van Osch^1,3
1C.J. Gorter Center for High Field MRI, Dept. of Radiology, Leiden University Medical Center, Leiden, Netherlands, ²Dept. Radiology and Radiotherapy, University Medical Center Utrecht, Utrecht, Netherlands, ³Leiden Institute of Brain and Cognition, Leiden, Netherlands

A new method called Inflow QUIXOTIC (IQ) is introduced, which employs a pulsed ALS module at T1blood*ln(2) seconds before VS-labeling to null the arterial pool. By nulling the arterial signal, the venous pool will be labeled exclusively by a single VS-module. IQ with and without QUIPSS was compared to QUIXOTIC. IQ enables selection of the local venous blood pool, showed a higher SNR, and provides also the opportunity to not only estimate the venous, but also the arterial T2. Whereas T2 measurements in large vessels (both arteries and veins) were feasible with our current sequence, tissue measurements were hampered by artifacts.

David D Shin¹, Thomas T Liu¹, Richard B Buxton¹, Guo Jia², and David J Dubowitz^1
1Center for Functional MRI, University of California, San Diego, La Jolla, California, United States, ²Department of Bioengineering, University of California, San Diego, La Jolla, California, United States

A novel approach to TRUST (T2 relaxation under spin tagging) for measuring venous O2 saturation (Yv) is presented. Critical to the measurement of Yv is a calibration between R2v and Yv. Although several authors have published calibration data, this are critically technique dependent which limits its universal applicability. We present a theoretical framework and propose a method to measure a R2 difference between two vessels (deltaR2) which removes some of the potential systematic errors from a calibration that is not locally applicable. This reduces the number of fit parameters from 6 to only 3. Initial tests in vivo reveal a better estimate of Yv and OEF than those obtained by assuming universal applicability of all 6 fit parameters. An novel imaging technique (“SAVANT”) is presented that directly measured the required deltaR2 by Simultaneous Arterial and Venous AcquisitioN of T2

Erin K Englund¹, Michael C Langham², Cheng Li¹, Emile R Mohler³, Thomas F Floyd⁴, and Felix W Wehrli^2
1Department of Bioengineering, University of Pennsylvania, Philadelphia, PA, United States, ²Department of Radiology, University of Pennsylvania, Philadelphia, PA, United States, ³Department of Cardiology, University of Pennsylvania, Philadelphia, PA, United States, ⁴Department of Anesthesiology, Stony Brook University School of Medicine, Stony Brook, NY, United States

To make use of the dead time during the post-labeling delay inherent in ASL sequences, a hybrid PASL/MR susceptometry sequence was developed capable of measuring perfusion, a microvascular parameter, and venous oxygen saturation (S_vO₂), a macrovascular parameter, with 2-second temporal resolution. Repeated measurements of perfusion alone, S_vO₂ alone, and PASL/S_vO₂ combined revealed no significant difference between the perfusion measurements (PASL vs. PASL/S_vO₂), or S_vO₂ (susceptometry vs. PASL/S_vO₂). These data suggest that the combined method is capable of quantifying perfusion and oxygen saturation simultaneously, thus providing measures of vascular physiology on both the macrovascular and microvascular level.

Patrick W Hales¹, and Chris A Clark^1
1Imaging & Biophysics Unit, University College London, London, United Kingdom

Multi-TI ASL and multi-b-value DWI measurements were made in ten healthy adult brains, and repeated after 20 min in five subjects. Estimates of capillary volume fraction (vbw) were derived from the DWI data using the IVIM model. vbw data were used in a two-compartment ASL model to estimate cerebral blood flow (CBF), bolus arrival time (BAT) and capillary permeability surface product (PS). CBF was most heterogeneous in the frontal lobe, vbw was most heterogeneous in the occipital lobe. By incorporating measured vbw values in a two-compartment ASL models, PS measurements showed improved heterogeneity and repeatability compared to previously published studies.

Andreas Wetscherek¹, Bram Stieltjes², Wolfhard Semmler¹, and Frederik Bernd Laun^1
1Dept. of Medical Physics in Radiology, German Cancer Research Center, Heidelberg, Germany, ²Quantitative imaging based disease characterization, German Cancer Research Center, Heidelberg, Germany

To investigate the origin of the IVIM model’s diagnostic capabilities we measured the signal attenuation in the pancreas for bipolar and flow compensated diffusion gradients. The diffusion weighting time T was varied while TE was kept fixed. Using a simple model relying on microscopic parameters of blood flow in microvasculature, we were able to reproduce the observed T-dependence in the flow compensated DWI data and the strong signal decay at small b-values for the bipolar DWI data. We found the characteristic time  > 100 ms implying that only a fraction of the blood exhibits directional changes during T.

Oscar San Emeterio Nateras¹, Yi Zhang^1,2, Qi Peng¹, Eric Muir², and Timothy Duong^1,2
1Radiology, University of Texas Health Science Center at San Antonio, San Antonio, TX, United States, ²Research Imaging Institute, University of Texas Health Science Center at San Antonio, San Antonio, TX, United States

This study reports human retina-choroidal blood flow values across different age groups. The BF was obtained using a quantitative MRI technique previously developed, performed on a 3.0T Phillips Achieva System. The images were acquired with high resolution BF MRI (0.5x0.8x6 mm3) on a single central-axial slice bisecting the optic nerve head and fovea. The study aims to evaluate BF MRI on human retina to learn the effects of aging on retina-choroidal vasculature. BF MRI of the retina provides quantitative data and has the potential to complement and cross-validate existing retinal imaging techniques.

Laura M. Parkes¹, Helen Beaumont¹, and Laurence Abernethy^2
1Biomedical Imaging Institute, University of Manchester, Manchester, Manchester, United Kingdom, ²Department of Radiology, Alder Hey Children's Foundation Trust, Liverpool, United Kingdom

For accurate quantification of cerebral blood flow (CBF) using Arterial Spin Labelling, multiple time-point measurements are required, allowing estimation of arrival time and CBF. A standard sequence is time-inefficient as no signal is collected during the delay period. The Look-Locker readout offers a more efficient approach, but complicates quantification. The aim of this work is to develop a simplified quantification model for the Look-Locker signal using a single blood compartment model. The model is validated using data from a healthy volunteer at a range of flip angles and is applied to 9 infants during routine clinical practice.

Maria A. Fernández-Seara¹, Marta Vidorreta¹, Maite Aznarez-Sanado^1,2, Federico Villagra¹, and Maria A. Pastor^1
1Neuroscience, Center for Applied Medical Research - University of Navarra, Pamplona, Navarra, Spain, ²Radiology, University of Pennsylvania, Philadelphia, PA, United States

Most previous reproducibility studies of Arterial Spin Labeling (ASL) have assessed within-session or short-term reproducibility (at 1 week intervals) in young volunteers. The purpose of this study was to assess long-term reproducibility at monthly intervals, in a group of older adults. Whole brain CBF measurements showed good reproducibility for the duration of the study period, with a wsCV of 10%. This information can be useful for planning longitudinal studies that are usually carried out during periods of several months.

Songlin Yu¹, Rui Wang², Dong Zhang³, Rong Wang³, Shuo Wang³, Zhentao Zuo², Lin Ai⁴, Bo Wang², Danny J. J. Wang⁵, and Jizong Zhao^3
1Beijing Tiantan Hospital, Capital Medical University, beijing, beijing, China, ²The State Key Laboratory of Brain and Cognitive Science, Institute of Biophysics, Chinese Academy of, ³Beijing Tiantan Hospital, Capital Medical University, ⁴Beijing Neurosurgical institute, ⁵Department of Neurology University of California Los Angeles Ahmanson-Lovelace Brain Mapping Center

The major quantitative calculation error of regional cerebral blood flow (rCBF) using arterial spin labeling (ASL) may exist in the differences in arterial transit arrival time. In this study we applied pseudocontinuous ASL (pCASL) with four postlabel delay (PLD) times and correlated the CBF results with that of CT perfusion to determine the optimal PLD that can best reflect the hemodynamics in moyamoya patients. The results show that pCASL images obtained at 2 to 2.5s best correlate with CTP. CBF values obtained at 1.5s may underestimate CBF values while the values obtained at 3000ms begin to decrease, likely due to the relaxation of the labeled signal.

Dennis F Heijtel¹, ElsMarieke van de Giessen², Matthan W Caan¹, and Aart J Nederveen^1
1Radiology, Academic Medical Center, Amsterdam, Netherlands, ²Nuclear Medicine, Academic Medical Center, Amsterdam, Netherlands

 

Weiying Dai¹, Gopal Varma¹, Rachel Scheidegger¹, Ajit Shankaranarayanan², Gottfried Schlaug³, and David Alsop^1
1Radiology, Beth Israel Deaconess Medical Center, Boston, MA, United States, ²Global Applied Science Laboratory, GE Healthcare, Menlo Park, CA, United States, ³Neurology, Beth Israel Deaconess Medical Center, Boston, MA, United States

We report the measurement of resting physiologic fluctuations in volumetric arterial spin labeling (ASL) blood flow images of normal volunteers. When sufficient background suppression is employed, coherent spatial patterns of fluctuations consistent with resting state networks are the dominant contributor to structured noise. All 7 networks identified by independent component analysis (ICA) are consistent with the networks reported using blood oxygenation level-dependent (BOLD) fMRI. ASL studies of resting state networks may be useful for characterization of temporal frequencies and regions of the brain poorly studied with BOLD. Better understanding of resting fluctuations may be necessary for accurate analysis of multi-scan ASL studies.

Rishma Vidyasagar¹, Arno Greyling², Yvonne Zebregs², Richard Draijer², and Laura M Parkes^1
1School of Cancer and Enabling Sciences, University of Manchester, Manchester, United Kingdom, ²Unilever R&D Vlaardingen, Vlaardingen, Netherlands

Black tea is widely consumed around the world and consists of dietary compounds such as flavonoids and caffeine. Studies have shown flavonoids impact endothelial nitric oxide synthase which lead to vascular relaxation and increased blood flow in the brachial medial artery, whilst clinical studies have observed quicker recovery from vascular diseases that correlate with black tea consumption. Caffeine has been shown to reduce cerebral blood flow and is known to be an adenosine antagonist. The primary aim of this study was to use arterial spin labelling methods to observe effects that these compounds in tea have on cerebral perfusion.

Joost PA Kuijer¹, Petra JW Pouwels¹, Ajit Shankaranarayanan², David Alsop³, Frederik Barkhof⁴, and Rudolf Verdaasdonk^1
1Physics and Medical Technology, VU University Medical Center, Amsterdam, Netherlands, ²Global Applied Science Laboratory, GE Healthcare, Menlo Park, CA, United States,³Radiology, Beth Harvard Medical School, Israel Deaconess Medical Center, Boston, MA, United States, ⁴Radiology, VU University Medical Center, Amsterdam, Netherlands

Caffeine intake is a known confounder in ASL studies. A test-retest study is presented with intervention using a single serving of coffee and a control group, after a 4-hour abstinence. Ten minutes after coffee intake a significant effect of caffeine on CBF was found. No significant change in CBF was found for the control group. Additionally, a lower group variance was found during retest, suggesting a scanner habituation time may lower group variance.

Tamara Fong¹, Weiying Dai², Li-Wen Huang³, Leo Waterson³, Sharon Inouye⁴, and David Alsop^2
1Neurology, Beth Israel Deaconess Medical Center, Boston, MA, United States, ²Radiology, Beth Israel Deaconess Medical Center, Boston, MA, United States, ³Psychiatry, Beth Israel Deaconess Medical Center and Harvard Medical School, ⁴Gerontology, Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, MA, United States

The cholinergic system plays a key role in human cognition. In vivo imaging of cholinergic function has the potential of advancing our understanding of its effects on regional and network activity. Here we examine the cholinergic system in vivo using pharmacologic arterial spin-labeling (ASL) perfusion MRI. Fifteen subjects were administrated anticholinergic drugs (mecamylamine or scopolamine or combined) to induce cholinergic blockade. ASL was used to measure cerebral blood flow (CBF) after administration. In vivo results showed cholinergic blockade caused significant CBF changes and worse cognitive performance.

Marjorie Villien^1,2, Laurent Lamalle³, Irène Troprès³, Thomas Rupp⁴, François Estève^1,5, Alexandre Krainik^1,6, Patrick Levy⁴, Jan Warnking^1,2, and Samuel Vergès^4
1U836, INSERM, Grenoble, France, ²Grenoble Institut of Neurosciences, Université Joseph Fourier, Grenoble, France, ³SFR1, Université Joseph Fourier, Grenoble, France, ⁴HP2 Laboratory (INSERM U 1042), Université Joseph Fourier, Grenoble, France, ⁵European Synchrotron Radiation Facility, Grenoble, France, ⁶Department of NeuroRadiology and MRI, CHU Grenoble, Grenoble, France

The aim of this study was to assess the effects of altitude acclimatization on cerebral blood flow (CBF) and cerebro-vascular reactivity to CO2 (CVR). Eleven healthy male subjects were investigated at sea level prior to and after 7 days at 4400m using pseudo-continuous ASL during a block-design inhalation paradigm to measure basal CBF and CVR. EtCO2 was significantly decreased post-altitude, while basal CBF increased in most subjects. CVR did not change significantly. Isocapnic CBF, obtained correcting for changes in basal capnia using measured CVR maps, was significantly increased in all subjects, by up to 40%, after 7 days at 4400m.

Ashley D Harris¹, Kevin Murphy², Claris Diaz¹, Neeraj Saxena³, Judith E Hall³, Thomas T Liu⁴, and Richard G Wise^1
1CUBRIC, School of Psychology, Cardiff University, Cardiff, United Kingdom, ²CUBRIC, School of Psycholog, Cardiff University, Cardiff, United Kingdom, ³Department of Anaesthetics, Intensive Care and Pain Medicine, Cardiff University, Cardiff, United Kingdom, ⁴Centre for Functional Magnetic Resonance Imaging & Department of Radiology, University of California San Diego, La Jolla, California, United States

Cerebral blood flow data (ASL acquisition) was collected before, during and after a hypoxic challenge. Based on a model developed for blood flow in the middle cerebral artery, the dynamics of the CBF response to hypoxia is then used to quantify CBF dynamics in response to a hypoxic challenge and during the return to normoxia. The CBF response to hypoxia differs regionally and in comparison to bulk arterial flow. Interpretations and implications of these finding are then discussed.

Christian Federau¹, Philippe Maeder¹, Kieran O'Brien², Patrick Browaeys¹, Markus Klarhoefer³, Reto Meuli¹, and Patric Hagmann^1
1Dept. of Radiology, CHUV, Lausanne, Switzerland, ²University of Geneva, CIBM, Switzerland, ³Siemens, Switzerland

Quantitative brain perfusion measurement remains a major challenge to currently available MR perfusion methods. The intravoxel incoherent motion (IVIM) method offers a non-invasive, alternative method to measure brain perfusion quantitatively, based on a bi-exponential fit of the relative signal obtained with an standard diffusion sequence for multiple b-values, usually between 0 and about 1000 s/mm2. We challenged the IVIM method with graded hypercapnia and hyperoxgenation in 7 healthy volunteers and report a statistically significant, graded increase of the IVIM perfusion parameters under hypercapnia, namely of the perfusion fraction f and of the pseudo-diffusion coefficient D*.

Yoshikazu Okamoto¹, Tomonori Isobe², and Yuji Hirano^3
1University of Tsukuba Hospital, Tsukuba, Ibaraki, Japan, ²University of Tsukuba, ³University of Tsukuba hospital, Japan

The soleus muscle (SOL) and the flexor digitorum profundus muscle (FDP) was scanned to cover b-factors from 0 to 3,500 s/mm2 of diffusion weighted image. A statistical comparison between monoexponential and biexponential fits was performed, and a statistically improved fit (P < 0.05; F test) was observed in all cases. The fast and slow diffusion fractions and fast and slow apparent diffusion coefficient (ADCfast and ADCslow) were 0.92 �} 0.05, 0.08 �} 0.05, 1.68 �} 0.14, and 0.26 �} 0.17, for SOL, and 0.95 �} 0.01, 0.05 �} 0.01, 1.75 �} 0.22, and 0.01 �} 0.02, respectively, for FDP.

Hou-Jen Chen^1,2, and Graham A Wright^1,2
1Sunnybrook Research Institute, Toronto, Ontario, Canada, ²Medical Biophysics, University of Toronto, Toronto, Ontario, Canada

This work presented an optimization of Arterial Spin Labeling for peripheral limbs. The arterial and perfusion responses after exercise during temporary ischemia were characterized. Perfusion was also measured with different post-labeling delays in a paradigm where the hyperemia was prolonged. The blood velocity was used to estimate the ideal sequence repetition time and spatial extent of labeling. The image samples at different delays were used to estimate transit delay. Quantification of perfusion and the sensitivity to peripheral hyperemia can be improved by physiological optimization based on the data.

Michael C LANGHAM¹, Erin K Englund¹, Cheng Li², Thomas F Floyd³, Emile R Mohler III⁴, and Felix W Wehrli^1
1Radiology, University of Pennsylvania, Philadelphia, PA, United States, ²Radiology, University of Pennsylvania, Philadelphia, PA - Pennsylvania, United States, ³Anesthesiology, SUNY Stondy Brook, Stony Brook, New York, United States, ⁴Vascular Medicine Section, University of Pennsylvania, Philadelphia, PA, United States

Microvascular dysfunction, associated with peripheral arterial disease, has previously been assessed by quantifying post-ischemic perfusion in calf muscle with continuous arterial spin labeling (CASL). We demonstrate that the saturation inversion-recovery (SATIR) PASL outweighs the SNR gain of CASL significantly by reducing the confounding BOLD effect in temporally resolved quantification of post-ischemic muscle perfusion. The two main features of SATIR PASL that reduce the imbalance in the tissue magnetization caused by the BOLD effect between successive images are higher temporal resolution and the slice-selective saturation pulse played out at the end of each pulse sequence cycle.

Alexander David Cohen¹, Peter S LaViolette², Kimberly Pechman¹, and Kathleen M Schmainda^1,2
1Biophysics, Medical College of Wisconsin, Wauwatosa, WI, United States, ²Radiology, Medical College of Wisconsin, Wauwatosa, WI, United States

The Intravoxel Incoherent Motion (IVIM) theory assumes diffusion within live tissue is more complex than simple random Brownian motion. The theory includes a faster water diffusion component representing the microcirculation of blood through capillaries, and there remains some debate as to what this perfusion fraction actually represents. Rats injected with a C6 glioma were imaged with multiple-bvalue DWI and DSC. IVIM and rCBV were compared in tumor versus normal brain. rCBV and IVIM metrics were found to be heightened in tumors compared to normal brain indicating IVIM may be biologically relevant in brain tumors.

Moritz Schneider¹, Michael Ingrisch¹, Matthias Moser², Heidrun Hirner², Clemens Cyran², Konstantin Nikolaou², Maximilian Reiser², and Olaf Dietrich^1
1Josef Lissner Laboratory for Biomedical Imaging, Department of Clinical Radiology, LMU Ludwig Maximilian University of Munich, Munich, Germany, ²Department of Clinical Radiology, LMU Ludwig Maximilian University of Munich, Munich, Germany

Intravoxel incoherent motion (IVIM) MRI is performed by fitting a biexponential model function to the measured diffusion-weighted signal attenuation. While the conventional biexponential model has 4 free parameters including the initial signal at b-value 0, several studies use only the signal attenuation relative to this initial value, discriminating, thus, artificially the measurement at b=0 and at higher b-values. In this study, the Akaike Information Criterion is used to determine whether the 3- or 4-parameter model is more appropriate for typical IVIM data. The comparison results in a very strong preference of the 3-parameter approach (relative attenuation).

Sabrina Doblas¹, Philippe Garteiser¹, Mathilde Wagner^1,2, Valérie Vilgrain^1,2, Bernard E. Van Beers^1,2, and Ralph Sinkus^1
1U773-CRB3, INSERM, Paris, France, ²Radiology Department, Beaujon University Hospital, Clichy, France

The oscillating-gradient spin-echo (OGSE) technique is based on the replacement of classical motion-encoding gradients by oscillating gradients, to obtain DWI experiments with short evolution times and corresponding spatial scales in the order of the micron. This lead to higher ADC compared with ADC determined using a classical EPI-DWI acquisition with relatively long evolution time. In vivo experiments were conducted in rats presenting different hepatic lesions, and showed the interest of using OGSE to gather a new dimension of information on tumor microenvironment and vasculature.

Patrick Winter¹, Fabian Gutjahr², Thomas Kampf², Xavier Helluy², Cord Meyer², Volker Herold², and Peter Michael Jakob^2
1Universität Würzburg, Kürnach, Bavaria, Germany, ²Universität Würzburg

In this work, we introduce a retrospectively gated inversion recovery snapshot FLASH sequence with radial k-space sampling for robust myocardial T1 mapping in small rodents. The sequence was tested at 7T on rats and mice yielding T1 maps with good agreement with literature. In rats, retrospective gating using internal navigator was as successful as using an In this work, we introduce a retrospectively gated inversion recovery snapshot FLASH sequence with radial k-space sampling for robust myocardial T1 mapping in small rodents. The sequence was tested at 7T on rats and mice yielding T1 maps with good agreement with literature. In mice, the internal navigator was often not sensitive enough to allow retrospective triggering. In rats, retrospective gating using internal navigator was as successful as using an external ECG signal.

Eric R Muir¹, Damon Cardenas¹, Guang Li¹, John Roby¹, and Timothy Duong^1,2
1Research Imaging Institute, UT Health Science Center, San Antonio, TX, United States, ²Ophthalmology, UT Health Science Center

Understanding brain physiology and function under hyperbaric-oxygen could be useful since hyperbaric-oxygen has been used as treatment for various neurological diseases. Under hyperbaric conditions, there will be more oxygen, which is vasoconstrictive on cerebral vasculature, so cerebral blood flow (CBF) may be reduced. The aim of this study was to use arterial spin labeling MRI to study the effect of hyperbaric-oxygen on CBF in rats. A hyperbaric chamber was built that could be put in the MRI scanner, allowing MR acquisition during hyperbaric conditions.

Fabian Tobias Gutjahr¹, Thomas Kampf², Patrick Winter², Cord Meyer², Christian Herbert Ziener³, Xavier Helluy², Peter Michael Jakob^2,4, and Wolfgang Rudolf Bauer^5
1Experimental Physics 5, University of Wuerzburg, Wuerzburg, Bavaria, Germany, ²Experimental Physics 5, University of Wuerzburg, ³DKFZ, Heidelberg, ⁴Magnetic Resonance Bavaria, Wuerzburg, Bavaria, Germany, ⁵Medizinische Klinik und Poliklinik I, Universitätsklinikum Wuerzburg, Wuerzburg, Bavaria, Germany

A retrospectively triggered Inversion Recovery Snapshot Flash sequence is proposed for robust quantification of perfusion in murine myocardium. An external ECG signal is recorded for retrospective triggering. The sequence which maintains strict Look-Locker condition can be implemented on virtually any MRI system as it does not need advanced triggering capabilities. A reconstruction based on the framework of partially separable functions using Principle Component Analysis is used to interpolate missing k-space data. Quantitative T1 and perfusion maps agree well with literature.

Guillaume Duhamel¹, Virginie Callot¹, and Patrick J. Cozzone^1
1CRMBM, UMR 6612, CNRS, Aix-Marseille University, Marseille, France

Mouse models of human brain diseases are extensively studied with long scan time multimodal MR protocols that include functional, metabolic and structural approaches. To further describe the pathologies, there is a real need for a high resolution, sensitive method which allows assessing quantitative whole brain perfusion within a reasonable scan time. This work presents the use of the ubpCASL technique in combination with fast imaging sequences for the acquisition of high resolution quantitative CBF maps within a short scan time protocol. The pCASL inversion efficiency was measured and multi-orientation acquisitions were investigated

Basil Künnecke¹, Thomas Bielser¹, and Markus von Kienlin^1
1CNS Research, F. Hoffmann-La Roche Ltd, Basel, Switzerland

Continuous arterial spin labelling (CASL) has proven a valuable tool for assessing cerebral perfusion as a proxy for neural activity, but it inherently lacks multi-slice capabilities. Newer variants such as pseudo CASL and sine wave-modulated CASL have been proposed to remedy this deficiency. Here we have implemented and for the first time applied these modalities for 3D volumetric perfusion mapping in rats at 4.7T and 9.4T. Off-centre-labelling and optimisation of labelling conditions espe¬cially for the control scan, together with RARE-based data readout, provided high-quality and high–sensitivity perfusion maps in only 4 minutes measurement time with nearly distortion-free image geometry.

Megan Johnston¹, Zhenlin Zheng², Joseph Maldjian³, Christopher T. Whitlow³, Michael J. Morykwas², and Youngkyoo Jung^1,3
1Biomedical Engineering, Wake Forest School of Medicine, Winston-Salem, North Carolina, United States, ²Plastic Surgery, Wake Forest School of Medicine, Winston-Salem, North Carolina, United States, ³Radiology, Wake Forest School of Medicine, Winston-Salem, North Carolina, United States

Swine brain perfusion is a good model for that of humans due to the similar proportions of gray and white matter. In order to measure swine perfusion, PCASL was implemented using this model. Our investigation identified parameters of the swine PCASL procedure that required optimization, as follows: The arterial blood velocity was measured to optimize tagging efficiency by adjusting acquisition parameters. Multiple post-labeling delays were collected so that the CBF quantification would be less sensitive to varying transit delays across the brain. Model-specific blood T₁ and the M₀ ratio of blood-to-white matter were measured for the quantification of CBF.

Olivier Reynaud¹, and Luisa Ciobanu^1
1Neurospin, CEA, Saclay, France

This study uses the FENSI technique as an alternative to ASL to characterize and quantify non-invasively tumor microvascular flux in a rat glioma model (9L). We highlight a significant increase (12–48%) in tumor blood flux at early developmental stages (tumor size < 2mm). At late stages (tumor size > 3mm) we observe tumor compartmentalization with the presence of a hypovascularized core (55±11µL/min/cm²) and a peripheral region presenting similar blood flux (83±23µL/min/cm²) as the healthy contralateral side. The different regions found in our FENSI flux maps correlate well with histological results which reflect endothelial cell concentrations (CD31 staining).

Xiao Wang¹, Ming Lu¹, Xiao-Hong Zhu¹, Yi Zhang¹, and Wei Chen^1
1Center for Magnetic Resonance Research, Department of Radiology, University of Minnesota Medical School, Minneapolis, MN, United States

In the present study, the absolute CBF increase induced by transient mild hypercapnia in the rat brain was directly compared using the SR-T₁^app method at 9.4T and 16.4T. The results show a consistent R₁^app, and thus CBF induced by hypercapnia although the absolute T₁^app was significantly longer at 16.4T than that at 9.4T under both normocapnia and hypercapnia conditions. The magnetic field strength independence of R₁^app, along with a good agreement of CBF change calculated with the SR-T₁^appmethod and with the CASL technique further demonstrates that the SR-T₁^app method provides a noninvasive, simple and efficient way to image and quantify CBF change induced by physiological and pathological perturbations in the rat brain.

Yen-Yu Ian Shih¹, Guang Li¹, Bryan H De La Garza¹, Lin Wang², and Timothy Q Duong^1
1Research Imaging Institute, University of Texas Health Science Center at San Antonio, San Antonio, Texas, United States, ²Devers Eye Institute, Legacy Clinical Research and Technology Center, Portland, OR, United States

: This study reports the cross validation of retinal and choroidal blood flow in rats using microsphere and ASL MRI technique. The retinal and choroidal blood flows were 1.2 and 8.1 ml/g/min, respectively as measured by MRI and 9.1 and 73 £gl/min as measured by microsphere. Taking the wet weight of the retina (5-35 mg), the retinal blood flow measured by microsphere in the present study ranged 0.23 ¡V 1.64 ml/g/min, in reasonable agreement with the MRI results. ASL MRI can be used to longitudinally to image retinal and choroid blood flow in vivo.

Jack A Wells¹, Bernard Siow¹, Mark F Lythgoe¹, and David L Thomas^2
1UCL Centre for Advanced Biomedical Imaging, UCL Division of Medicine and Institute of Child Health, London, United Kingdom, ²Department of Brain Repair and Rehabilitation, UCL Institute of Neurology, London. ML and DT are joint senior authors.

The transverse decay of the FAIR ASL signal at 5 inflow times (TIs) was measured in the rat brain at 9.4T. Bi-exponential behavior was observed that appears to derive from different T2 values associated with the intra-vasculature (IV) and extra-vascular (EV) compartments. A two compartment bi-exponential model was used to assess the relative contribution of the IV and EV compartments to the ASL signal at the 5 TIs, without needing to assume a constant T2 for labeled blood water in the vessels. This novel method was applied to estimate tissue transit time and the oxygen saturation on the arterial side of the vasculature