Qinxiang Li¹, He Wang², and Yunfei Zha^1
1REN MIN HOSPITAL OF WUHAN UNIVERSITY, WUHAN, HUBEI, China, ²MR Research China, GE Healthcare, Shanghai, China

This study investigated the differences in quantitative perfusion parameters of thoracolumbar vertebral bone marrow(VBM) in adults in relation to vertebral level and age using dynamic contrast enhancement MRI(DCE-MRI). We found that the quantitative parameters (Ktrans,Kep and Ve )were significantly associated with vertebral level and significantly influenced by age, which could lay a foundation for the further research of vertebral marrow perfusion.

Milica Medved¹, Aytekin Oto¹, Xiaobing Fan¹, Federico D. Pineda¹, Russell Z. Szmulewitz², and Gregory S. Karczmar^1
1Radiology, University of Chicago, Chicago, Illinois, United States, ²Medicine, University of Chicago, Chicago, Illinois, United States

In applications of the two compartment model to tissue contrast agent uptake and washout, it is critical to determine the arterial input function (AIF). The AIF cannot always be measured directly, and approximate methods are sometimes used to derive it. For example, it can be derived by using muscle – with known K^trans and v_e values – as a reference tissue. Published values for skeletal muscle are used, typically measured in an accessible muscle, such as the calf. We show that K^trans and v_e vary between gluteal and deep pelvic muscles, and that thus skeletal muscles should be characterized individually.

Katherine C. Gao¹, Govind Nair¹, Colin D. Shea¹, Pietro Maggi^1,2, Souheil J. Inati³, Luca Massacesi², and Daniel S. Reich^1
1National Institute of Neurological Disorders and Stroke, Bethesda, Maryland, United States, ²Department of Neurological and Psychiatric Sciences, University of Florence, Florence, Italy, ³Natiional Institute of Mental Health, Bethesda, Maryland, United States

Patients with MS typically develop lesions around inflamed blood vessels in the brain. Since permeability and perfusion of MS lesions have been shown to change with time, ascertaining these parameters within individual MS lesions at a particular time point may provide a snapshot of the underlying inflammatory and pathphysiological processes. By acquiring a dual-echo T1-weighted dynamic scan during a single bolus of MRI contrast, we simultaneously estimated permeability and perfusion of chronic and active lesions in MS patients. As expected, enhancing lesions had the highest permeability. Additionally, some chronic lesions had non-zero permeability, which may reflect inflammatory activity that is no longer visibly detectable as enhancement in post-contrast T1-weighted images. Longitudinal studies are underway to track changes in lesion permeability and perfusion over time and will determine whether this methodology can be used to assess the effects of disease-modifying therapies that reduce inflammation and maintain the blood-brain-barrier.

Parinaz Massoumzadeh¹, Yi Su¹, Joshua S. Shimony¹, Andrei Vlassenko¹, Jonathan McConathy¹, and Tammie Benzinger^1
1Mallinckrodt Institute of Radiology, Washington University in St. Louis, St. Louis, MO, United States

Cerebral blood flow (CBF) has been compared using PET and MR, methods but not using a PET/MR dual-modality imaging system. Here, we compare CBF measurement obtained using three methods: dynamic susceptibility contrast (DSC) MR, Dynamic F18-FDG, and O-15 water PET scans for patients with brain tumor. We use a PET/MR scanner (Siemens Biograph mMR). MR based CBF were found to be correlated with PET CBF at both voxel and regional level, although the correlation was lower than the correlations between [015]PET and FDG imaging. Further investigation is ongoing to investigate the difference between PET and MR based CBF measurements.

Fabian Kording¹, Wedegaertner Ulrike¹, Nils Daniel Forkert², Jan Sedlacik¹, Gerhard Adam¹, and Chressen K. Much^1
1Department of Diagnostic and Interventional Radiology, University Medical Center Hamburg-Eppendorf, Hamburg, Germany, Germany, ²Department of Computational Neuroscience, University Medical Center Hamburg-Eppendorf, Hamburg, Germany, Germany

The assessment placental perfusion using dynamic contrast enhanced MRI is of increasing importance for early prediction of fetal growth restrictions and vascular complications during pregnancies. To assess differences in the placenta perfusion in experimental studies analysis is performed as an average over the whole placenta. However, the mouse placenta consists of two different functional zones and differentiating the two compartments may lead to additional important information. In this work placenta functional compartments were successfully differentiated using time-to-peak values of the contrast agent time curve. Perfusion analysis of both compartments revealed significant differences which could be beneficial for further studies.

Wilson B. Chwang¹, Rajan Jain^1,2, Siamak P. Nejad-Davarani^3,4, A.S.M. Iskander¹, Ashley VanSlooten⁵, Lonni Schultz⁵, James R. Ewing^3,6, Ali S. Arbab^1,7, and Hassan Bagher-Ebadian^3,6
1Department of Radiology, Henry Ford Hospital, Detroit, MI, United States, ²Department of Neurosurgery, Henry Ford Hospital, Detroit, MI, United States,³Department of Neurology, Henry Ford Hospital, Detroit, MI, United States, ⁴Department of Biomedical Engineering, University of Michigan, Ann Artbor, MI, United States, ⁵Department of Public Health Sciences, Henry Ford Hospital, Detroit, MI, United States, ⁶Department of Physics, Oakland University, Rochester Hills, MI, United States, ⁷Department of Radiology, Wayne State University, Detroit, MI, United States

The purpose of this study was to investigate parameters of vascular physiology such as vp, Ktrans, and ve in a rat glioma model using two different contrast agents, an intravascular or blood pool agent (gadofosveset) and an extravascular agent (gadopentetate dimeglumine, Gd-DTPA) using DCE T1 nested model selection. We found that Ktrans was significantly lower using gadofosveset compared to Gd-DTPA, while vp and ve measures were not statistically different. But more importantly these parameters were measured with high agreement using DCE T1 NMS, demonstrating the stability of nested model selection in DCE-MRI and its importance for assessing the tumor microenvironment.

Huijun Chen¹, Matthew L. Olson², Savannah C. Partridge³, and William Kerwin^3
1Tsinghua University, Beijing, Beijing, China, ²University Of Washington, Seattle, WA, United States, ³University of Washington, Seattle, WA, United States

The Extended Graphical Model (EGM) has the ability to acquire accurate pharmacokinetic parameters (contrast transfer constant: Ktrans) using a shorter scan time than the widely used modified Kety/Tofts model in dynamic contrast enhanced (DCE) MRI of carotid atherosclerotic plaque. However, in the area of oncology, where DCE MRI and pharmacokinetic modeling are widely used in clinical research, the capability of the EGM has not been investigated. In this study, we sought to evaluate the EGM in breast imaging for use with a novel hybrid DCE MRI that provides separate high spatial and high temporal resolution datasets with full bilateral coverage.

Yi-Ying Wu¹, Chen-Hao Wu^1,2, Chih-Ming Chiang¹, Clayton Chi-Chang Chen^1,3, and Jyh-Wen Chai^1,4
1Department of Radiology, Taichung Veterans General Hospital, Taichung City, Taiwan (R.O.C), Taiwan, ²Institute of Biomedical Engineering, National Taiwan University, Taipei City, Taiwan(R.O.C.), Taiwan, ³Department of Biomedical Engineering, Hung Kuang University, Taichung City, Taiwan (R.O.C.), Taiwan, ⁴College of Medicine, China Medical University, Taichung City, Taiwan (R.O.C), Taiwan

Venous output function (VOF) from the larger and straighter superior sagittal sinus, with similar shape of the concentration time curves as AIF, has shown the effectiveness in permeability measurements of brain neoplasms. However, there are still some errors from VOF that may influence the quantification of tissue permeability. The aim of this study is to investigate the clinical applicability of permeability measurement with DCE-MRI, by using a combination of the scaled AIF with VOF and the normalization to the plasma volume of the adjacent white matter, to estimate Ktrans of brain metastatic lesions in the longitudinal follow-up.

Kourosh Jafari-Khouzani¹, Kyrre E. Emblem^1,2, Jayashree Kalpathy-Cramer¹, Atle Bjørnerud^2,3, Mark Vangel¹, Elizabeth R. Gerstner⁴, Kathleen M. Schmainda⁵, Tracy T. Batchelor⁴, Bruce Rosen¹, and Steven M. Stufflebeam^1
1Athinoula A. Martinos Center for Biomedical Imaging, Department of Radiology, Massachusetts General Hospital, Harvard Medical School, Boston, MA, United States, ²The Intervention Centre, Rikshospitalet, Oslo University Hospital, Oslo, Norway, ³Dept of Physics, University of Oslo, Oslo, Norway,⁴Massachusetts General Hospital Cancer Center, Harvard Medical School, Boston, MA, United States, ⁵Department of Biophysics, Medical College of Wisconsin, Milwaukee, WI, United States

This study evaluates the repeatability of dynamic susceptibility contrast (DSC) based perfusion imaging using a double baseline MRI acquisition setup in 31 adult patients (18 M, 13 F, age 23-72, mean 56) with newly diagnosed glioblastoma. The images were acquired using a dual-echo, combined gradient-echo (GE) and spin-echo (SE) echo planar imaging sequence. Cerebral blood volume (CBV) and cerebral blood flow (CBF) maps were generated and evaluated within tumor regions (enhancing tumor from T1-weighted images and whole tumor including edema from FLAIR image). Repeatability was evaluated using intraclass correlation coefficients. High repeatability was obtained for mean values of perfusion maps within tumor regions.

I Chi Liu^1,2, Tsong-Hai Lee³, Ching-Chung Liang⁴, and Ho-Ling Liu^1,5
1Department of Medical Imaging and Radiological Sciences, Chang Gung University, Taoyuan, Taiwan, ²Department of Medical Physics, Koo Foundation Sun Yat-Sen Cancer Center, Taipei City, Taiwan, ³Department of Neurology and Stroke Center, Chang Gung Memorial Hospital and Chang Gung University College of Medicine, Taoyuan, Taiwan, ⁴Department of Obstetrics and Gynecology, Chang Gung Memorial Hospital, and Chang Gung University College of Medicine, Taoyuan, Taiwan, ⁵Department of Medical Imaging and Intervention, Chang Gung Memorial Hospital, Taoyuan, Taiwan

Accurate estimation of arterial input function (AIF) is crucial for the CBF quantification using DSC-MRI. In small animals, AIF is difficult to identify due to the inevitable partial-volume effects. This study aimed to investigate the feasibility of using a venous output function (VOF) for calibrating the AIF in rats, and evaluate the deviation caused by the selection of VOF within sagittal sinus. The results showed that using the ratio of areas under the curve from VOF versus AIF after a gamma-variate fitting yielded the most stable CBF calibration. The uncertainties of resulted CBF values were in the range of 8-28%.

Dennis Lai Hong Cheong¹ and Thian C. Ng^1,2
1Clinical Imaging Research Center, SBIC/A*STAR & National University of Singapore, Singapore, Singapore, ²Department of Diagnostic Radiology, National University of Singapore, Singapore, Singapore

In the analysis of DCE MRI, a functional curve representing the arterial or vascular input function is beneficial. Empirical formulations depend on the shape of the input function, and its parameters provide little physiologic meaning. Horsfield physiologically based model is good but its application is limited to human DCE MRI study using Gd-DTPA. We have developed a physiologically based model without this limitation. We demonstrate its ability to fit data measured by both DCE MRI and DCE CT, and at different blood vessel locations. It has potential application to animal studies, and to other measurements (e.g. PET). Unphysiological parameter values may assist in identifying errors in measured input functions, especially in DCE MRI where quantification is problematic.

Allen T. Newton^1,2, Jack T. Skinner¹, and Christopher C. Quarles^1
1Radiology and Radiological Sciences, Vanderbilt University Institute of Imaging Science, Nashville, TN, United States, ²Radiology and Radiological Sciences, Monroe Carell Jr. Children's Hopsital at Vanderbilt, Nashville, TN, United States

We present a simple method by which investigators performing multi-echo DSC-MRI perfusion studies can identify voxels saturating into the noise floor at later echo times, and show the effect of removing these voxels from analysis on AIF estimation. Furthermore, our study evaluates these methods in a population of pediatric patients, demonstrating their applicability in an actual clinical context.

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

We recently presented a method for a quantitative determination of the arterial input function in the carotid arteries. The AIF-measurement plug-in in the standard DSC perfusion sequence used the apparent shift in the arterial position in one-dimensional projection images and was optimised for high doses of contrast. Here we demonstrate an alternative AIF measurement plug-in for low-dose clinical measurements based on fast two-dimensional readout and the phase AIF. Efficient background suppression using inversion recovery remains a crucial component of the method. We present a comparative discussion of both methods in relation to the dose, tracer type, and the field strength.

Gregory J. Wilson¹, Charles S. Springer, Jr.², Mark Woods^2,3, Sarah Bastawrous^1,4, Puneet Bhargava^1,4, and Jeffrey H. Maki^1
1Radiology, University of Washington, Seattle, WA, United States, ²Advanced Imaging Research Center, Oregon Health and Science University, Portland, OR, United States, ³Chemistry, Portland State University, Portland, OR, United States, ⁴Radiology, Puget Sound VAHCS, Seattle, WA, United States

Measurements of longitudinal relaxation rates (R₁) of ¹H₂O in whole blood with gadolinium-based contrast reagent concentrations ([CR]) in the range of 2 to 18 mM provide evidence of water exchange across the erythrocyte cell membrane. The dependence of R₁ on [CR] was accurately predicted by the two-site-exchange (2SX) model in the fast exchange regime (FXR). Contrast reagents with high protein binding affinity (gadobenate dimeglumine and gadofosveset trisodium) were more sensitive to the water exchange kinetics because of their higher relaxivities. This sensitivity to water exchange across the cell membrane may enable future studies of erythrocyte function.

Martin M. Pike¹, Mohan L. Jayatilake¹, Xiaoyan Wang², Merryl R. Lobo¹, Xin Li¹, Matthias C. Schabel¹, William D. Rooney¹, Dale J. Christensen³, Jerry D. Glickson⁴, Rosalie C. Sears², Wei Huang¹, and Charles S. Springer, Jr.^1
1Advanced Imaging Research Center, Oregon Health & Science University, Portland, Oregon, United States, ²Molecular and Medical Genetics, Oregon Health & Science University, Portland, Oregon, United States, ³Oncotide Pharmaceuticals, Research Triangle Park, North Carolina, United States,⁴Radiology, University of Pennsylvania, Philadelphia, Pennsylvania, United States

The metabolic dimension of DCE-MRI is explored by using targeted therapy on a genetically modified mouse spontaneous breast cancer. DCE-MRI suggests a metabolic flux decrease upon therapy.

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 Physics, University of Toronto, Toronto, ON, Canada

Perfusion analysis could provide tissue residue function by deconvolution of the observed MR signal with AIF; however with SVD it is challenging due to non-physiological oscillations. The Control Point Interpolation method has been demonstrated to provide smooth residue functions in perfusion analysis and is sensitive to pathological variations. Our analysis of eight atherosclerotic disease patients showed variation in residue function among healthy and ischemic tissue which was quantified by calculating the time taken for the residue function to drop to 50% and 10% of its maximum value. Capillary transit time heterogeneity information was also extracted from this residue function by calculating transit time distribution for tissue.

Amit Mehndiratta¹, Fernando Calamante², Bradley J. MacIntosh³, David E. Crane³, Stephen J. Payne¹, and Michael A. Chappell^1
1Institute of Biomedical Engineering, University of Oxford, Oxford, Oxfordshire, United Kingdom, ²Melbourne Brain Centre - Austin campus, Brain Research Institute, Heidelberg, Victoria, Australia, ³Medical Physics, University of Toronto, Toronto, ON, Canada

Exponential residue function is commonly used in DSC MRI simulations for cerebral haemodynamic approximation to validate novel deconvolution methodologies. However, the haemodynamics is significantly altered in pathology where exponential approximation might be no more valid. In the study in-vivo observed residue function with non-parametric CPI method were approximated with four commonly used analytical expressions where bi-exponential expression was found to be more realistic approximation of normal and pathological variation in haemodynamics.

Amit Mehndiratta¹, Fernando Calamante², Bradley J. MacIntosh³, David E. Crane³, Stephen J. Payne¹, and Michael A. Chappell^1
1Institute of Biomedical Engineering, University of Oxford, Oxford, Oxfordshire, United Kingdom, ²Melbourne Brain Centre - Austin campus, Brain Research Institute, Heidelberg, Victoria, Australia, ³Medical Physics, University of Toronto, Toronto, ON, Canada

Bolus dispersion is a significant problem in perfusion imaging because of remote AIF measurement. It is challenging to accurately estimate CBF in presence of dispersion; in this study our motivation was to investigate the accuracy in CBF estimate with a number of deconvolution methods, and also investigate the possibility to separate the effects of dispersion from cerebral haemodynamics using these techniques. The ambiguity between true residue function and dispersion could not be resolved with any of the deconvolution methods, but accurate estimates of cerebral perfusion can be achieved with an effectively non-parametric CPI approach even in presence of dispersion.

Ka-Loh Li¹, Xiaoping Zhu¹, and Alan Jackson^1
1The Wolfson Molecular Imaging Centre, The University of Manchester, Manchester, United Kingdom

Uncertainty analysis for error propagation in five semi-quantitative DCE-MRI parameters was performed. Five patients with NF2 were used for evaluation. Percent deviation (PD) distributions from Monte Carlo simulation were compared with in vivo data. Results: 1) The sum of SE and SErel had greater precision than non-summed metrics for persistent type SI(t); (2) The “normalized” parameters had poorer precision than non-normalized metrics;(3) Rse1/se2 was more accurate for persistent type than washout type; (4) Rse1/se2, showed wide PD distribution. In summary, comparison of in-vivo analysis with Monte Carlo simulation supports the findings of previous studies on synthetic data.

Xin Li¹, Seymur Gahramanov², Charles S. Springer, Jr.¹, William D. Rooney¹, and Edward A. Neuwelt^2
1Advanced Imaging Research Center, Oregon Health & Science University, Portland, Oregon, United States, ²Department of Neurology, Oregon Health & Science University, Portland, Oregon, United States

Conventional gradient echo Dynamic-Contrast-Enhanced (DCE) MRI protocols acquire data at short but finite echo time (TE). During contrast reagent (CR) bolus passage, T2* related MRI signal reduction (mainly due to susceptibility gradient introduced by CR) may not be negligible even with short TE, especially at ultra-high field. Using a sequential multi-session DCE-MRI data collection at 11.75 T, we demonstrate that the T2* signal loss in DCE-MRI is tissue dependent and most pronounced for tissues with low CR extravasation which transiently experience large intra-voxel susceptibility gradients.

Jeiran Jahani¹, Glyn Johnson², Valerij G. Kiselev³, and Dmitry S. Novikov^4
1New York University School of Medicine, New York City, New York, United States, ²University of East Anglia, East Anglia, United Kingdom, ³University Medical Center Freiburg, Freiburg, Germany, ⁴Radiology, New York Univeristy School of Medicine, New York City, New York, United States

We present a nonlinear algorithm for a simultaneous reduction of noise in a series of dynamic maps, such as those acquired during bolus passage in DCE MRI. Our algorithm utilizes a universal distribution of eigenvalues of a random covariance matrix. Thus we are able to objectively remove the “noisy” eigenvalues and preserve the significant principal components. We test our method in a numerical dynamic phantom to show that the noise-reduction preserves all structural information, and further apply it to DCE MRI. Our method selectively reduces noisy components of the images across time while preserving the informative structures of contrast dynamics.

Jennifer Moroz¹, Piotr Kozlowski^2,3, and Stefan A. Reinsberg^1
1Physics and Astronomy, UBC, Vancouver, BC, Canada, ²Radiology, UBC, Vancouver, BC, Canada, ³MRI Research Centre, UBC, Vancouver, BC, Canada

A high-temporal resolution arterial input function (AIF) is desired for pharamacokinetic modelling, which may be achieved with a projection-based approach. However, local tissue enhancement will affect the accuracy of the AIF. Tissue enhancement may be visualized from the projections used to estimate the AIF, if acquired radially. This work investigates three methods for radial reconstruction and evaluates their potential for measuring tissue enhancement. An error profile, a plot comparing the profile of the measured enhancement (radial image) with the expected result, was calculated to evaluate the techniques. The NFFT technique had the lowest errors, making it attractive for our application.

Dattesh D. Shanbhag¹, Suresh E. Joel¹, Ming-Ching Chang², Kumar T. Rajamani¹, Sandeep Narendra Gupta², and Rakesh Mullick^1
1GE Global Research, Bangalore, Karanataka, India, ²GE Global Research, Niskayuna, NY, United States

In this work, we investigated a block-wise PCA based approach to reconstruct DCE-MRI data. It is based on the hypothesis that there exists overlapping temporal information within a spatial neighborhood which can be exploited to separate noise from true contrast enhancement while maintaining tissue heterogeneity. We demonstrate that PCA based reconstruction of dynamic DCE data produced smooth parametric maps while preserving the lesion conspicuity, compared to pixelated maps produced using voxel-by voxel analysis. This will improve the accuracy of DCE-MRI quantification and enhance the sensitivity of the method in clinical scenario.

Wei Cao¹, Wei Li¹, Shonagh K. O’Leary-Moore², Kathleen K. Sulik², G. Allan Johnson³, and Chunlei Liu^1,4
1Brain Imaging and Analysis Center, Duke University, Durham, NC, United States, ²Bowles Center for Alcohol Studies,University of North Carolina, Chapel Hill, NC, United States, ³Center for In Vivo Microscopy, Duke University, Durham, NC, United States, ⁴Department of Radiology, Duke University, Durham, NC, United States

The primary effects of prenatal alcohol exposure are on the brain development and the cognitive and behavioral deficits that ensue. Magnetic susceptibility imaging was used to assess its impact to mouse brains. Two groups of mice (n=3, postnatal day 80) were characterized: an ethanol group and a control group. Two main white matter fiber bundles, anterior commissure and corpus callosum, were analyzed. The orientation dependence of magnetic susceptibility was found significantly decreased in ethanol exposed brains compared to their age-matched controls in both regions. This may be interpreted as a loss of myelination due to prenatal alcohol exposure.

Ya-Jian Cheng¹, Junjie Chen², Shelton D. Caruthers², and Samuel A. Wickline^2
1Biomedical Eng., Washington University In Saint Louis, Saint Louis, Missouri, United States, ²Washington University In Saint Louis, St Louis, MO, United States

Heart muscle, the most important muscle in our body, demonstrates mysterious mechanical function during each heart beat. The heart wall generates up to 50% wall thickening with lower than 15% of fiber shortening. A laminar-like muscle structure, a.k.a. sheets, was found to contribute to such function. Here, we accessed the sheet reorientation function in the most used dystrophy model, mdx mice, with the Langendorff perfused viable hearts and mighty 11.7T diffusion tensor MRI.

Yurui Gao^1,2, Ann S. Choe^1,2, Xia Li¹, Iwona Stepniewska³, and Adam W. Anderson^1,2
1Institute of Imaging Science, Vanderbilt University, Nashville, Tennessee, United States, ²Department of Biomedical Engineering, Vanderbilt University, Nashville, Tennessee, United States, ³Department of Psychological Sciences, Vanderbilt University, Nashville, Tennessee, United States

In a previous study, we validated DTI-tractography-derived measures of primary motor area (M1) corticocortical (CC) connectivity by comparing the ex vivo DTI connectivity with the histological ground truth. Ex vivo acquisitions typically provide higher image quality than in vivo experiments, since scan times can be much longer and motion is usually not an issue. The goal of the present study was to get a more realistic understanding of the limitations of in vivo DTI measures of CC connectivity by comparing the reliability of in vivo and ex vivo DTI data acquired from the same squirrel monkey.

Jutta Janke¹, Andreas Wetscherek¹, Bram Stieltjes², and Frederik B. Laun^1,2
1Dpt. Medical Physics in Radiology, German Cancer Research Center (DKFZ), Heidelberg, Baden-Württemberg, Germany, ²Quantitative Imaging-Based Disease Characterization, German Cancer Research Center (DKFZ), Heidelberg, Baden-Württemberg, Germany

The Apparent Diffusion Coefficient (ADC) of lipids may be a valuable biomarker for various diseases, e.g. diabetes. The low ADC of lipids requires the use of long diffusion weightings, which results in low signal-to-noise ratios. Compensating the low SNR by thicker slices lead to unexpected artifacts, which usually do not appear in water diffusion experiments. In particular, we observed pronounced signal drops that we attribute to tissue pulsation and imperfect slice selections due to residual eddy currents. We compensated for these signal drops by using only the maximal signal intensity of all acquired images for the calculation of the ADC.

Hans-Peter Mueller¹, Ina Vernikouskaya², Albert Ludolph¹, Jan Kassubek¹, and Volker Rasche^2,3
1Department of Neurology, Ulm University, Ulm, Baden Wuerttemberg, Germany, ²Core Facility Small Animal MRI, Ulm University, Ulm, Baden Wuerttemberg, Germany, ³Experimental Cardiovascular Imaging, Ulm University, Ulm, NA, Germany

The SNR gain of cryogenically cooled resonators (CCR)is supposed to generate high fidelity DTI data with high in-plane resolution and thin slices in order to get close to isotropic voxel resolution. That way, accurate fractional anisotropy (FA) mapping and improved isotropic 3D DTI fiber reconstruction could be obtained. The aim of this study is to keep the overall scanning time at approximately 30 minutes in order to enable the application of the scan protocol to in-vivo cohort studies aiming at DTI comparisons of murine brain at the group level.

Roger Bourne¹, Dominic Pang¹, Andre Bongers², Carl Power², Paul Sved¹, and Geoffrey Watson^3
1University of Sydney, Sydney, NSW, Australia, ²University of New South Wales, Sydney, NSW, Australia, ³Royal Prince Alfred Hospital, Sydney, NSW, Australia

This work compares a high spatial resolution diffusion kurtosis analysis of two whole fixed prostates with single b-value DTI of the same voxels. Kurtosis was found to be high in low diffusivity glandular tissue and low in stromal tissue. Variance of the kurtosis correlated strongly with DTI-derived fractional anisotropy.

Jennifer Mei¹, Hsiao-Fang Liang², and Shu-Wei Sun^2
1University of California, Loma Linda, CA, United States, ²Loma Linda University, Loma Linda, CA, United States

The calcium-dependent aquaporin water channel (AQP) 0 in ocular lens fiber cells facilitates fluid circulation for nourishing the vascular lens to maintain transparency. The goal of this study is to evaluate whether AQP0 malfunction can be detected non-invasively by Diffusion Tensor Imaging (DTI). We used Manganese ions (Mn2+) to alter the AQP0 function in mouse lens. A reduction of radial diffusivity was found in Mn2+-affected lens, compared to the controls.

Farida Grinberg¹, Ezequiel Farrher¹, Luisa Ciobanu², and Nadim Jon Shah^1,3
1Institute of Neuroscience and Medicine 4 - Medical Imaging Physics, Forschungszentrum Juelich, Juelich, Germany, ²Neurospin, CEA, Gif-sur-Yvette, France,³Department of Neurology, Faculty of Medicine, JARA, RWTH Aachen University, Aachen, Germany

We analyse the sensitivity of the stretched- exponential model in providing contrast for ischemic lesions in animal stroke models. Good fits in the extended range of the diffusion weightings were obtained. In frame of the proposed analytical models, the stretched-exponential function is indicative of anomalous diffusion mechanisms and provide complementary means to infer valuable microstructural information. We obtained promising results in terms of image contrast and better understanding of pathological changes in the ischemic lesions.

Xingju Nie¹, Edward S. Hui¹, Jens H. Jensen¹, Joseph A. Helpern^1,2, Ann-Charlotte E. Granholm-Bentley², Heather A. Boger², and Maria F. Falangola^1,2
1Radiology and Radiological Science, Center for Biomedical Imaging, Medical University of South Carolina, Charleston, SC, United States, ²Neurosciences, Medical University of South Carolina, Charleston, SC, United States

The goal of this study was to utilize a recently developed cerebral microenvironment modeling (CMM) method to investigate the morphological abnormalities in the prefrontal cortex of young Ts65Dn mice. CMM relates diffusional kurtosis imaging (DKI) data to specific biophysical tissue characteristics. We demonstrated that CMM parameters are sensitive indicators of changes in the complexity of the neurite architecture associated with abnormal brain development and maturation in this model.

Mohammed Salman Shazeeb^1,2, Stuart Howes², Sivakumar Kandasamy², Christopher H. Sotak^1,2, and George Pins^2
1Radiology, University of Massachusetts Medical School, Worcester, MA, United States, ²Biomedical Engineering, Worcester Polytechnic Institute, Worcester, MA, United States

The in vivo evaluation of the remodeling of soft biomaterial implants often involves surgical removal of the implant for subsequent histological assessment. This approach is very resource intensive, it is often destructive, and imposes practical limitations on how effectively these materials can be evaluated. MRI has the potential to non-invasively monitor the remodeling of collagen scaffolds. This study investigated the development of a model system to evaluate the remodeling of implanted collagen scaffolds using MRI and conventional histological techniques. Significant correlations were found between the MRI and histological parameters demonstrating that MRI is sensitive to specific remodeling parameters.

Yoshikazu Okamoto¹, Tomonori Isobe¹, and Yuji Hirano^1
1University of Tsukuba Hospital, Tsukuba, Ibaraki, Japan

How slow and fast diffusion components work in the skeletal muscle?

Faezeh Sanaei Nezhad¹, Anahita Fathi Kazerooni^2,3, Anwar R. Padhani⁴, Hamid Soltanian Zadeh¹, and Hamidreza Saligheh Rad^2,3
1School of Electrical and Computer Engineering (ECE), College of Engineering, University of Tehran, Tehran, Iran, ²Medical Physics and Biomedical Engineering, Tehran University of Medical Sciences, Tehran, Iran, ³Quantitative MR Imaging and Spectroscopy Group, Research Center for Cellular and Molecular Imaging, Tehran University of Medical Sciences, Tehran, Iran, ⁴Paul Strickland Scanner Centre, Mount Vernon Cancer Centre, London, United Kingdom

Bone marrow metastases are common manifestation of many malignant cancers. In order to gain deeper insight into patient’s state of disease and the treatment response, the bone marrow cellularity can be assessed by the histogram analysis and monitoring of the ADC-maps acquired by DW-MRI. However, correct decision making about the effectiveness of therapy is dependent on the ROIs selected from the ADC-maps, raising concerns about the susceptibility of manual ROI placement to human errors. In this work, an automatic bone marrow registration/segmentation histogram analysis approach was applied to the ADC-maps, to show its superiority over the manual ROI analysis method.

Nathan White¹ and Anders M. Dale^2
1University of California, San Diego, La Jolla, CA, United States, ²Radiology and Neurosciences, University of California, San Diego, La Jolla, CA, United States

The apparent restricted water signal, derived from restriction spectrum imaging (RSI) data, provides improved conspicuity and delineation of high-grade brain tumors and reduced sensitivity to edema compared with traditional ADC. However, the biological origin of the restricted water signal remains poorly understood. In this study, we use Monte Carlo simulations to evaluate the specific role of the nuclear volume fraction and cell diameter on RSI signal contrast. The results suggest that the nuclear volume fraction is an important cellular characteristic for both RSI and ADC contrast, that may explain in part the lack of apparent restriction in low-grade tumors.

Novena Rangwala¹, David Hackney¹, and David C. Alsop^1
1Department of Radiology, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA, United States

A model of diffusion signal decay in the transverse plane is proposed and applied to high b-value diffusion-weighted (DW) images of the cervical spinal cord (SC) in vivo. This model assumes that axons within SC are not perfectly aligned and the transverse signal decay is a function of the angle made by the axons with the cord axis. The results show that the average axonal dispersion in the cord is ~10°, and decreases to ~6.5° within the posterior funiculus of white matter, indicating that the model is sensitive to regional differences in axon structure.

Michael Zhao¹, Esther Myint², Geoffrey Watson³, and Roger Bourne^1
1University of Sydney, Sydney, NSW, Australia, ²Laverty Pathology, Sydney, NSW, Australia, ³Royal Prince Alfred Hospital, Sydney, NSW, Australia

Diffusion weighted microimaging of fixed prostate tissue reveals distinct diffusivity differences between epithelium, stroma, and acinal spaces1. In normal glandular tissue changes in the relative volumes of epithelium and stroma explain ~60% of the variation in signal fraction of a biexponential model of diffusion signal decay. We hypothesise that microscopic tissue volume changes may also explain the observed strong correlation between cancer Gleason grade and apparent diffusion coefficient measured in vivo. In this pilot study we used semi-automated morphometry methods to quantify partial volumes of epithelium, stroma and acinal space in H&E-stained histological images of benign and cancerous prostate tissue and compared these results with segmentation and statistical analysis of diffusion weighted microimages of normal tissue.

Yu-Chun Lin¹, Jiun-jie Wang², Che-Ming Wu³, and Shu-Hang Ng^1
1Department of Medical Imaging and Intervention, Chang Gung Memorial Hospital, Linkou, 333, Taiwan, ²Department of Medical Imaging and Radiological Sciences, Chang Gung University, Taoyuan, 333, Taiwan, ³Department of Otolaryngology, Chang Gung Memorial Hospital, Linkou, Taiwan, Taiwan

The study evaluated the diagnostic value of multishot fast spin-echo periodically rotated overlapping parallel lines with enhanced reconstruction (PROPELLER) diffusion weighted imaging (DWI) in detecting cholesteatoma. The surgical pathologic findings were used as the reference standard in 23 patients. The images using PROPELLER DWI had significantly lower susceptibility and geometric artifacts, which resulted in higher sensitivity and accuracy compared with those using the single-shot EPI DWI. Both two techniques had 100% specificity in diagnosis. The multishot fast spin-echo PROPELLER DWI provided better diagnostic utility in detecting middle ear cholesteatoma compared with conventional EPI-DWI.

Habib Rahbar^1,2, Matthew L. Olson^1,2, Constance D. Lehman^1,2, and Savannah C. Partridge^1,2
1Radiology, University of Washington, Seattle, WA, United States, ²Radiology, Seattle Cancer Care Alliance, Seattle, WA, United States

Diffusion weighted (DW) breast MR holds promise as an adjunct technique to standard dynamic contrast-enhanced breast MR, a biomarker of disease subtypes, and a non-contrast screening tool. In this educational exhibit, we review the technical considerations and current evidence for clinical applications of this promising MR technique. We also highlight areas requiring further research to allow widespread implementation of this technique into clinical breast MR protocols.

Jun-Cheng Weng^1,2, Sih-Yu Lin¹, and Wen-Yih Isaac Tseng^3,4
1School of Medical Imaging and Radiological Sciences, Chung Shan Medical University, Taichung, Taiwan, ²Department of Medical Imaging, Chung Shan Medical University Hospital, Taichung, Taiwan, ³Center for Optoelectronic Biomedicine, National Taiwan University College of Medicine, Taipei, Taiwan, ⁴Department of Medical Imaging, National Taiwan University Hospital, Taipei, Taiwan

It is known that water signal decay in a MR diffusion experiment in neuronal tissues, at sufficiently high diffusion weighting, appears to be non-mono-exponential, thus complicating even further the interpretation and assignment of the different components to actual physiological compartments. Therefore, we used 2D bi-Gaussian model (i.e. slow and fast components) to fit our QPI data. After 2D Fourier transformation of the slow and fast Gaussian curved surfaces of signal decay, respectively, two Gaussian curved surfaces of displacement distribution (i.e. narrow and broad components) were obtained. Intracellular and extracellular information could then be extracted from the narrow and broad Gaussian displacement distributions, respectively. Our results demonstrated that bi-Gaussian fitting QPI produced reasonable distribution of relative axonal diameters of CC in normal human brain.

Zaiyi Liu¹, Changhong Liang¹, and Xin Chen^1
1Dept. of radiology, Guangdong General Hospital, Guangzhou, Guangdong, China

ADC is a promising imaging biomarker for characterizing liver lesions and predicting and monitoring the response of hepatic diseases to treatment. However, there's a wide variety in liver ADC values which have been reported, even in normal liver. Knowledge of the characteristic of ADC measurement in normal liver is of great importance for accurate interpretation of ADC changes. Our results showed that the ADC values are influeced by both DWI techniques and anatomical location of the liver, thus we should use the same technique and the same location when evaluate the ADC values changes in diffuse liver disease after treatment.

Timo De Bondt¹, Wim Van Hecke², Jelle Veraart³, Alexander Leemans⁴, Jan Sijbers³, Stefan Sunaert⁵, Yves Jacquemyn⁶, and Paul M. Parizel^1,7
1Radiology, Antwerp University Hospital, Antwerp, Antwerp, Belgium, ²icoMetrix, Leuven, Belgium, ³Physics, University of Antwerp, Antwerp, Antwerp, Belgium,⁴Image Sciences Institute, University Medical Center Utrecht, Utrecht, Netherlands, ⁵Radiology, University Hospitals of the Catholic University Leuven, Leuven, Belgium, ⁶Gynaecology and Obstaetrics, Antwerp University Hospital, Antwerp, Antwerp, Belgium, ⁷Radiology, University of Antwerp, Antwerp, Antwerp, Belgium

Khader M. Hasan¹ and Ponnada A. Narayana^1
1Diagnostic and Interventional Imaging, University of Texas Health Science Center at Houston, Houston, Texas, United States

Due to its close interplay with the vasculature, cerebrospinal fluid (CSF) serves as a chemical conduit in intracranial pressure regulation and chemical function. Whole brain, sulcal and ventricular CSF volumes have been used as a marker of normal brain development and aging and atrophy in a host of pathologies, including closed traumatic brain injuries. In addition CSF chemical and metabolic content serves as a robust biomarker of neural degeneration. A comprehensive analysis of normative CSF compartmental volumetry, relaxometry and diffusimetry has not been described previously using quantitative magnetic resonance imaging (qMRI). In this report, we applied novel and validated atlas-based segmentation methods for a detailed investigation of age dependent qMRI changes in the CSF compartments in both men and women

Tetsu Niwa¹, Koki Kusagiri², Noriko Aida², Taro Takahara³, Tomoaki Nagaoka⁴, and Yutaka Imai^1
1Radiology, Tokai University School of Medicine, Isehara, Kanagawa, Japan, ²Radiology, Kanagawa Children's Medical Center, Yokohama, Kanagawa, Japan,³Biomedical Engineering, Tokai University School of Engineering, Isehara, Kanagawa, Japan, ⁴Electromagnetic Compatibility Laboratory, Applied Electromagnetic Research Institute, National Institute of Information and Communications Technology, Koganei, Tokyo, Japan

We assessed whether diffusion-weighted reversed fast imaging with steady-state precession (DW-PSIF) can demonstrate cerebrospinal fluid flow. DW-PSIF showed signal reduction at the cerebrospinal fluid (CSF), which was corresponded to CSF flow on phase-contrast imaging. DW-PSIF with higher moment revealed less imaging quality and less signal-to-noise ratio. DW-PSIF can be a method to assess CSF flow in a relatively good imaging quality and in a short time.

Jack Harmer¹, Nicolas Toussaint¹, Kuberan Pushparajah¹, Christian T. Stoeck², Rachel W. Chan³, Reza Razavi¹, David Atkinson³, and Sebastian Kozerke^1,2
1Division of Imaging Sciences and Biomedical Engineering, Kings College London, London, United Kingdom, ²Institute for Biomedical Engineering, University and ETH Zurich, Zurich, Switzerland, ³Centre for Medical Imaging, University College London, London, United Kingdom

Advances in diffusion acquisition schemes employing stimulated echoes now allow for robust DTI of the beating heart. When used in conjunction with respiratory navigation, data can be acquired during free breathing making it potentially feasible in patients. We demonstrate how in-vivo DTI data can be acquired in a patient with a systemic right ventricle, allowing novel insights into the adaptation of myofibre architecture in a morphological right ventricle supporting systemic circulation. This is to our knowledge the first record of in-vivo DTI of a systemic right ventricle. The technique has considerable potential for furthering our understanding of congenital heart disease.

Valentina Di Marco¹, Marco Palombo^1,2, and Silvia Capuani^1,2
1Physics Department, Sapienza University, Rome, Rome, Italy, ²CNR IPCF UOS Roma, Physics Department, Sapienza University, Rome, Rome, Italy

We investigated the potential of Gaussian and non-Gaussian diffusion methods to obtain information about the microstructural complexity and the water compartmentalization in free and trabecular bone-marrow by investigating apparent diffusion coefficient (ADC) of the water and the fat component as a function of diffusion time (). ADCs of the fat were found to be constant as a function of , while the intracellular and the extracellular water exhibits a different ADC behavior as a function of  in free and trabecular bone marrow. Internal gradient at the interface between water and bone affects extracellular ADC behaviour which is non-Gaussian.

Eiko Yamabe¹, Ryo Miyagi², Toshinori Sakai³, Toshiyasu Nakamura⁴, and Hiroshi Yoshioka^1
1Department of Radiological Sciences, University of California Irvine, Orange, CA, United States, ²Department of Orthopaedic Surgery, Miyoshi City National Insurance Nishi-Iya Clinic, Tokushima, Japan, ³Department of Orthopedics, Institute of Health Biosciences, The University of Tokushima Graduate School, Tokushima, Japan, ⁴Department of Orthopaedic Surgery, Keio University School of Medicine, Tokyo, Japan

We qualitatively visualized the median nerve with diffusion tensor imaging in various acquisition positions/parameters and quantitatively investigated changes in diffusion values (FA, ADC) of the median nerve at different locations of the carpal tunnel in healthy volunteers. The median nerve at the carpal tunnel including a small nerve fiber such as a motor branch was well appreciated on fiber tracking images. Comparable fiber tracking images and FA/ADC values were obtained in supine and prone positions without any statistical differences. These results show a potential for their clinical use.

Qunfeng Tang¹, Hongwei Chen¹, and Hui Liu^2
1Radiology, Wuxi People Hospital, Nanjing Medical University, Wuxi, Jiangsu, China, ²MR Collaboration NE Asia, Siemens Healthcare, Shanghai, Shanghai, China

In the brain application, IVIM is a method capabling of simultaneously measuring CBF, CFV and ADC. In this study, we evaluated this IVIM based perfusion by comparing with the traditional DSC based perfusion measurement in 3 types of tumors. Initial results shows the consistent in CBF measurement between IVIM and DSC based methods except 18% cases which are considered to be contributed by flow effect for IVIM measurement.

Tao Jiang¹, Peng Wang¹, Shi-Yuan Liu¹, Weibo Chen², and Queenie Chan^3
1Department of Radiology, Shanghai Changzheng Hospital, Shanghai, China, ²Philips Healthcare, Shanghai, China, ³Philips Healthcare, Hong Kong, China

This study mainly focus on whether the the methyl-guanine methyl transferase(MGMT) promoter methylation status is associated with a specific imaging feature and several imaging parameters (ADC, FA and perfusion parameters) of HGGs.

Mami Iima^1,2, Olivier Reynaud¹, Tomokazu Tsurugizawa¹, Luisa Ciobanu¹, Jing-Rebecca Li¹, Françoise Geffroy¹, Boucif Djemai¹, Masaki Umehana³, and Denis Le Bihan^1
1Neurospin, CEA Saclay, Gif-sur-Yvette, Ile-de-France, France, ²Diagnostic Imaging and Nuclear Medicine, Kyoto University Graduate School of Medicine, Kyoto, Kyoto, Japan, ³Kyoto University Faculty of Medicine, Kyoto, Kyoto, Japan

IVIM MRI images were acquired in 14 rats implanted with a 9L glioma model using a 17.2T MRI scanner. IVIM maps clearly highlighted areas with high and low fraction perfusion within tumors which were generally heterogeneous, as confirmed by histology. Perfusion parameters could easily be obtained using a 2-steps processing approach. Diffusion parameters obtained using a biexponential 2-compartments diffusion model, such as ADCo and fslow, additionally provided information in tissue structure. However, the diffusion parameters estimated from the biexponential model were found to be very sensitive to noise and to the initial values used for data fitting.

Yonggang Lu¹, Jacobus F.A. Jansen², Yousef Mazaheri¹, Hilda E. Stambuk³, and Amita Shukla-Dave^1
1Medical Physics Department, Memorial Sloan-kettering Cancer Center, New York, NY, United States, ²Radiology Department, Maastricht University, Maastricht, Maastricht, Netherlands, ³Radiology Department, Memorial Sloan-kettering Cancer Center, New York, NY, United States

This study proposes optimal model mapping to characterize tumor microcirculation with diffusion weighted imaging in head and neck cancer. At each voxel within tumor tissue, an optimal model was determined to quantify tumor microcirculation from several common diffusion models based on the Bayesian Information Criterion. For individual tumors, voxel percentage for each model and the optimal model map were generated. The results were promising in characterizing the heterogeneity of tumor microcirculation. Future studies with large cohort of patients need to assess the use of this method in clinical applications such as tumor diagnosis, staging and treatment monitoring.

Moritz Jörg Schneider¹, Michael Ingrisch¹, Birgit Ertl-Wagner², Maximilian F. Reiser², and Olaf Dietrich^1
1Josef Lissner Laboratory for Biomedical Imaging, Institute for Clinical Radiology, LMU Ludwig Maximilian University of Munich, Munich, Bavaria, Germany,²Institute for Clinical Radiology, LMU Ludwig Maximilian University of Munich, Munich, Bavaria, Germany

Intravoxel incoherent motion (IVIM) MRI can be used to obtain information about tissue microcapillary perfusion properties based on diffusion-weighted acquisitions. Therefore, we compare perfusion-related parameters obtained by IVIM measurements with cerebral blood volume and flow (CBV,CBF) as well as semi-quantitative parameters retrieved from dynamic-contrast-enhanced (DCE) MRI in tumorous and normal-appearing white matter. The IVIM perfusion fraction f correlates well with CBV and CBF, indicating that IVIM yields perfusion-related information. All of the DCE parameters correlate strongly with the product D*×f, which reflects a “makeshift flow”and is therefore likely to be the best indicator for tissue-perfusion changes.

Stephanie L. Barnes^1,2, Christopher C. Quarles^1,2, and Thomas E. Yankeelov^1,2
1Radiology and Radiological Sciences, Vanderbilt University, Nashville, TN, United States, ²Vanderbilt University Institute of Imaging Sciences, Vanderbilt University, Nashville, TN, United States

Quantitative DCE-MRI analysis does not account for diffusion of the contrast agent (CA) within a voxel. This work utilizes a finite element model of CA diffusion to investigate the effect of diffusion on intra-voxel CA distribution, and the resulting effect on voxel SI and parameterization error. The results show that decreasing homogeneity of CA distribution increase the difference in effective total voxel signal intensity (SI). Additionally, CA diffusion (over a range of relevant values) affects SI on a time scale on the order of common temporal resolutions. These results demonstrate the potential importance of CA diffusion in DCE-MRI analysis.

Christian Federau¹, Kieran O'Brien², Markus Müller³, Matthias Stuber^1,4, Reto Meuli¹, Philippe Maeder¹, and Patric Hagmann^1
1Radiology, CHUV, Lausanne, VD, Switzerland, ²CIBM, Université de Genève, Lausanne, VD, Switzerland, ³The Abdus Salam International Center for Theoretical Physics, Trieste, Trieste, Italy, ⁴CIBM, Lausanne, VD, Switzerland

We present evidence that the IVIM perfusion parameters (f, the perfusion fraction, D*, the pseudo-diffusion coefficient, and fD*, which is related to blood flow) depend on the cardiac cycle in the human brain, while the diffusion coefficient was found to be stable. This represents direct experimental evidence of pulsatile flow in the human brain microvasculature and correlates well with recently reported measurement of pulsatile blood flow in the microvasculature of the cortex of the mice brain by Santisakultarm et al (Am J Physiol Heart Circ Physiol, 2012) using two-photon microscopy.

Yoshikazu Okamoto¹, Graham J. Kemp², and Tomonori Isobe^1
1University of Tsukuba Hospital, Tsukuba, Ibaraki, Japan, ²University of Liverpool, Liverpool, Liverpool, United Kingdom

New measurement method of skeletal mucle perfusion by IVIM

Sangwoo Lee¹, Jeong Min Lee², Jeong Hee Yoon², and Hiroyuki Kabasawa^3
1Global Applied Science Lab, GE Healhcare, Gangnam-gu, Seoul, Korea, ²Radiology, Seoul National University Hospital, Jongno-gu, Seoul, Korea, ³Global Applied Science Laboratory, GE Healhcare, Hino-shi, Tokyo, Japan

IVIM (IntraVoxel Incoherent Motion) has been widely used in various clinical applications including liver fibrosis staging and hepatic tumor characterization. However, in those studies, there has been no consideration of T1 and T2 relaxation differences between blood and tissue within the IVIM signal model. It was reported that perfusion fraction can be significantly affected by imaging TE due to such relaxation differences. In this work, we demonstrated the feasibility of simultaneous estimation of tissue T2 map within single IVIM data acquisition and compensation of perfusion fraction using the acquired T2 map.

Kimihiro Ogisu¹, Akiko Fujita², and Toru Yamamoto^3
1Department of Radiology, National Hospital Organization Hokkaido Medical Center, Sapporo, Hokkaido, Japan, ²Graduate School of Health Sciences, Hokkaido University, Sapporo, Hokkaido, Japan, ³Faculty of Health Sciences, Hokkaido University, Sapporo, Hokkaido, Japan

Blood flow in capillaries has been believed to be intravoxel incoherent motion whose MR signal appears in diffusion weighted imaging at low b values as well as MR signal from free water. However, we found an increase in T2 with decreasing b values, and this increase cannot be interpreted by blood in capillaries. MR signal from interstitial fluid (ISF), the T2 of which is around 300 ms, agrees the T2 increase. Blood flow in capillaries is rather understood as intravoxel coherent motion by using anisotropic directional distribution model of vessel segments.

Andreas Wetscherek^1,2, Frederik B. Laun^1,3, Claudia Prieto^2,4, and Cristián Tejos^2,5
1Medical Physics in Radiology, German Cancer Research Center (DKFZ), Heidelberg, Germany, ²Biomedical Imaging Center, Pontificia Universidad Catolica de Chile, Santiago, Chile, ³Quantitative Imaging-Based Disease Characterization, German Cancer Research Center (DKFZ), Heidelberg, Germany, ⁴Division of Imaging Sciences and Biomedical Engineering, King's College London, London, United Kingdom, ⁵Department of Electrical Engineering, Pontificia Universidad Catolica de Chile, Santiago, Chile

The flow-compensated (FC) intra-voxel incoherent motion (IVIM) model allows probing the characteristic timescale and velocity of the incoherent motion using FC diffusion weighted MRI. To reduce the long acquisition times required, we applied a compressed sensing reconstruction exploiting sparsity of the Karhunen-Loève transform domain to a 3x undersampled data set. We were able to show that the CS reconstruction improved the accuracy of the parameter maps obtained from the undersampled data making their quality comparable to those obtained from non-undersampled data.

Xiaojie Wang¹, Matthew Cusick², Robert S. Fujinami², and Sheng-Kwei Song^3,4
1Chemistry, Washington University, St. Louis, MO, United States, ²Pathology, University of Utah, Salt Lake City, UT, United States, ³Radiology, Washington University, St. Louis, MO, United States, ⁴Hope Center for Neurological Disorders, Washington University, St. Louis, MO, United States

SJL-EAE mice were treated with LDK or vehicle from the onset of first relapse to the study end-point (37 day post immunization). 99-direction DBSI was performed to examine the spinal cord white matter pathology. Axon and myelin integrity of the spinal cord white matter was assessed using axial and radial diffusivity, respectively, while inflammation extent was evaluated using cell ratio and edema water ratio derived by DBSI. LDK diminished the disease activity through the anti-inflammatory, axon preservation and probably remyelination effect.

Yan Zhou¹, Jianyu Liu², He Wang³, and Nan Sun^4
1Peking University Third Hospital, Beijing, Beijing, China, ²Radiology Department, Peking University Third Hospital, Beijing, Beijing, China, ³MR Research China, GE Healthcare, Shanghai, Shanghai, China, ⁴MR Research China, GE Healthcare, Beijing, Beijing, China

In this study, multi-b value Diffusion Weighed imaging models of mono-exponential, bi-exponential (intravoxel incoherent motion, IVIM) and stretched exponential models were discussed in diagnosing endometrial cancer and cervix cancer, in order to find more reliable or more precise method to deal with DWI. The results showed good diagnostic value of Slow ADC in IVIM model, and showed potentially ability of using perfusion fraction as a non invasive method assessing perfusion. While stretched exponential model were questioned by some results from this study.

Naomi H. M. Douglas¹, Jessica M. Winfield¹, Nandita M. deSouza¹, David John Collins¹, and Matthew R. Orton^1
1Radiotherapy and Imaging, Institute of Cancer Research, Sutton, Surrey, United Kingdom

A phantom suitable for quality assurance in multi-centre clinical trials has been designed and built. Phantom development involved testing of: temperature stability; relaxivity dependence on concentrations of sucrose and manganese chloride; ADC dependence on concentrations of sucrose and manganese chloride. Sodium azide was added to each of the solutions as a bactericidal agent. The final phantom is of a modular design containing samples with physiologically relevant values of ADC, T1 and T2. This test object is being used to aid the validation of ADC as a biomarker for treatment response of tumours in multi-centre clinical trials.

Frank G. Zöllner¹, Sven Kaiser¹, Gerald Weisser², and Lothar R. Schad^1
1Computer Assisted Clinical Medicine, Medical Faculty Mannheim, Heidelberg University, Mannheim, Baden-Württemberg, Germany, ²Institute for Clinical Radiology and Nuclear Medicine, Medical University Center, Heidelberg University, Mannheim, Baden-Württemberg, Germany

Intra-voxel incoherent motion (IVIM) diffusion weighted imaging (DWI) is an emerging method to derive microscopic structural information non-invasively [1]. To allow for a broader usage, in this work, an OsiriX plug-in was developed to submit the DWI analysis and specific the IVIM model for integration in the clinical routine diagnostics.Based on 30 data sets of prostate, a t-test revealed, that there was no significant difference (p < 0.05) between the plugin and a reference implmentation.

Koji Sakai¹, Ryusuke Nakai¹, Kei Yamada², Jun Tazoe², Kentaro Akazawa², and Naozo Sugimoto^1
1Kyoto University, Kyoto, Kyoto, Japan, ²Kyoto Prefectural University of Medicine, Kyoto, Kyoto, Japan

Among these MR methods, the most clinically applicable may be the post processing of diffusion-weighted images (DWI). However, this DWI-based method might be influenced by the composition of the CSF, which can strongly affect its viscosity and diffusivity. The purpose of this study was to investigate the influence of protein content on DWI thermometry by using artificial CSF (ACSF) with variable protein concentrations. The protein content of ACSF increased viscosity and decreased the diffusion coefficient. Consequently, the results of DWI thermometry were influenced by the protein content in the range of 0.10 to 1.11 mg/ml (0.10 ºC – 0.42ºC).

Manil Chouhan^1,2, Alan Bainbridge³, Nathan Davies⁴, Rajiv Jalan⁴, Rajeshwar Mookerjee⁴, Simon Walker-Samuel², Mark F. Lythgoe², Shonit Punwani¹, and Stuart Taylor^1
1UCL Centre for Medical Imaging, University College London, London, London, United Kingdom, ²UCL Centre for Advanced Biomedical Imaging, University College London, London, London, United Kingdom, ³Department of Medical Physics, University College Hospitals NHS Trust, London, London, United Kingdom, ⁴UCL Institute for Liver and Digestive Health, University College London, London, London, United Kingdom

The lack of robust non-invasive techniques for repeated measurement of liver blood flow has restricted understanding of the vascular changes underpinning chronic liver disease and the development of therapies to address these changes. Phase-contrast MRI has potential for portal venous flow quantification and in this study we demonstrate repeatability of phase-contrast MRI and present encouraging initial data demonstrating reproducibility of measurements and validation of flow quantification against invasive transit-time ultrasound measurements.

Teodora-Adriana Perles-Barbacaru¹, Irène Troprès¹, David Chechin², Sylvie Grand¹, Jean-François Le Bas¹, Francois Berger¹, and Hana Lahrech^1
1Grenoble Institute of Neurosciences, University of Joseph Fourier, INSERM U 836, La Tronche, France, ²Philips Healthcare, Suresnes, France

The Rapid Steady State T₁ technique enables direct quantitative mapping of the cerebral blood volume fraction with a clinically approved Gd-DOTA dose during the first pass and without requirement of the arterial input function. It facilitates treatment monitoring and clinical studies in particular for the evaluation of antiangiogenic therapies.

Åsmund Kjørstad¹, Dominique M.R. Corteville¹, André Fischer², Frank G. Zöllner¹, and Lothar R. Schad^1
1CKM, Heidelberg University, Mannheim, Baden-Württemberg, Germany, ²Institute of Radiology, University of Wuerzburg, Würzburg, Bavaria, Germany

The Fourier Decomposition method is established as a non-invasive method for perfusion related information in the lung, and has shown promise in clinical studies. One drawback is the lack of meaningful quantification, making follow-up and comparison of patients difficult. We show that quantification of Fourier Decomposition maps is both possible and easy, requiring only a small amount of post-processing.

Chin-Tien Lu¹, Tzu-Chen Yeh², and Fu-Nien Wang^1
1Department of Biomedical Engineering and Environmental Sciences, National Tsing Hua University, Hsinchu, Taiwan, ²Department of Radiology, Taipei Veterans General Hospital, Taipei, Taiwan

D₂O can be used as a free diffusible tracer for measurement of tissue perfusion. Recently, an indirect strategy for D₂O perfusion MRI was proposed by monitoring the ¹H signal attenuation. In this study, we aimed to further utilize the characteristic of long time course of D₂O. Using syringe pump for stable infusion of 2ml/100g D₂O and proton-density weighted RARE for dynamic imaging, intact brain morphology and good perfusion contrast were achieved in location with severe susceptibility effect.

Nicolas Pannetier^1,2, Maja Sohlin³, Thomas Christen⁴, Lothar R. Schad⁵, and Norbert Schuff^1,2
1Department of Radiology, University of California San Francisco, San Francisco, CA, United States, ²Centre for Imaging of Neurodegenerative Diseases, Veteran Affairs Medical Centre, San Francisco, CA, United States, ³MR centre, Sahlgrenska University Hospital, Goteborg, Goteborg, Sweden, ⁴Department of Radiology, Stanford University, Stanford, CA, United States, ⁵Computer Assisted Clinical Medicine, Heidelberg University, Mannheim, Heidelberg, Germany

MRI is a powerful tool to investigate the microvasculature. Various techniques benefit from the entanglement of the contrasts that arise within a voxel and provide relevant biomarkers. However, the analysis relies on analytical models that present a discrepancy with Monte Carlo simulations. In this work we present a versatile approach directly based on a numerical model. The approach is evaluated on phantom data where vessels are mimicked by polyamide strings. We compared our model with analytical ones and we demonstrated that this approach could provide a mean to estimate the radius of the vessel without the use of contrast agent (CA).

Nicolas Pannetier^1,2, Benjamin Lemasson^3,4, Thomas Christen⁵, Greg Zaharchuck⁵, Emmanuel Luc Barbier^3,4, and Norbert Schuff^1,2
1Department of Radiology, University of California San Francisco, San Francisco, CA, United States, ²Centre for Imaging of Neurodegenerative Diseases, Veteran Affairs Medical Centre, San Francisco, CA, United States, ³Grenoble Institute of Neurosciences, Grenoble University, La Tronche, Rhone-Alpes, France, ⁴U836, INSERM, La Tronche, Rhone-Alpes, France, ⁵Department of Radiology, Stanford University, Stanford, CA, United States

The analytical description of the susceptibility-related MR dephasing is used to assess the microvasculature. However, analytical solutions are discordant with Monte Carlo simulations and this discrepancy lead to bias vascular estimates. Moreover recent results demonstrate that BVf and SO2 cannot be assessed at once in qBOLD. In this work, we proposed a new integrative approach to simultaneously assess blood volume, oxygenation and the radius of the vessel. We built up a densely sampled lookup table by simulating the MR signal provided by a Gradient Echo Sampling of the FID and SE sequence prior and post injection of a contrast agent. We acquired the corresponding sequence on rats with focal brain ischemia and we estimated the vascular parameters from the LT.

Wendy W. Ni^1,2, Thomas Christen¹, Zungho Zun¹, and Greg Zaharchuk^1
1Department of Radiology, Stanford University, Stanford, CA, United States, ²Department of Electrical Engineering, Stanford University, Stanford, CA, United States

In this study, we compare the measurement of reversible contribution R2'=1/T2' of the transverse relaxation rate using nine methods characterized by three approaches: asymmetric spin echo (ASE), Gradient-Echo Sampling of FID and Echo (GESFIDE), and combined (COMBO). We found that the mean value, spatial SNR and inter-subject variation of resulting R2' maps differed significantly between methods, highlighting the importance of making a careful choice in R2' measurement method for quantitative BOLD (qBOLD) oxygenation measurements.

Farida Grinberg¹, Ezequiel Farrher¹, Ivan I. Maximov¹, and Nadim Jon Shah^1,2
1Institute of Neuroscience and Medicine, Forschungszentrum Juelich, Juelich, Germany, ²Faculty of Medicine, JARA, RWTH Aachen University, Aachen, Germany

We report on the results of the application of several non-Gaussian (diffusion kurtosis, log-normal distribution function, stretched exponential function) and anomalous diffusion models to describe experimental data in the anisotropic fibre phantoms. We focus on the influence of fibre packing density on the quantitative metrics of these models and compare their sensitivity to this parameter. We examine the power-law dependences of the apparent diffusivity on time and fibre density in the frame of anomalous diffusion model.

Shlomi Lifshits¹, Assaf Horowits², Daniel Barazany², Saharon Rosset¹, and Yaniv Assaf^2
1Department of Statistics and Operations Research, Tel-Aviv University, Tel-Aviv, Israel, ²Department of Neurobiology, Tel-Aviv University, Tel-Aviv, Israel

We suggest formulating the AxCaliber framework as a statistical learning classification problem and show that biologically consistent results are achieved by the resulting axon diameter distribution probability maps.

Gemma Nedjati-Gilani¹, Torben Schneider², Bernard M. Siow^1,3, Mohamed Tachrount⁴, Andrew Davies², Kenneth J. Smith², Ying Li⁴, Olga Ciccarelli⁴, David L. Thomas⁴, Daniel C. Alexander¹, and Claudia Angela M. Wheeler-Kingshott^2
1Centre for Medical Image Computing, Department of Computer Science, University College London, London, United Kingdom, ²NMR Research Unit, Queen Square MS Centre, Department of Neuroinflammation, UCL Institute of Neurology, London, United Kingdom, ³Centre for Advanced Biomedical Imaging, University College London, London, United Kingdom, ⁴Department of Brain Repair and Rehabilitation, UCL Institute of Neurology, London, United Kingdom

This works tests the feasibility of parameters from diffusion MR such as exchange time _i and axonal density f as biomarkers for axonal loss and demyelination, using an animal model of axonal loss in the spinal cord. Preliminary results show a significant decrease in f in regions of axon loss. Exchange time also decreases, but is not necessarily significant. This suggests that f is a potentially useful biomarker of axonal degeneration. Future work will extend the methodology to investigate demyelination.

Alexander Ruh¹, Philipp Emerich¹, Dmitry S. Novikov², and Valerij G. Kiselev^1
1Department of Radiology, Medical Physics, University Medical Center Freiburg, Freiburg, Germany, ²Bernard and Irene Schwartz Center for Biomedical Imaging, Department of Radiology, New York University School of Medicine, New York, NY, United States

The correlation time of diffusing spins in magnetically heterogeneous tissues is often commensurate with the timing of MRI experiments. This results in a time dependent transverse relaxation rate that for long times approaches the monoexponential limit. Here we show that the power law of this approach represents the statistics of large-scale organization of magnetic microstructure. The origin of this effect is the self-averaging, which is inherent to diffusion narrowing, when spins are exploring microstructure via diffusion. The present results enable an analysis scheme alternative to a previously proposed spectral domain approach, which enlarges options for future quantifications.

Uran Ferizi^1,2, Torben Schneider², Eleftheria Panagiotaki¹, Gemma Nedjati-Gilani¹, Hui Zhang¹, Claudia Angela M. Wheeler-Kingshott², and Daniel C. Alexander^1
1Centre for Medical Image Computing and Department of Computer Science, University College London, London, England, United Kingdom, ²NMR Research Unit, Queen Square MS Centre, Department of Neuroinflammation, UCL Institute of Neurology, London, England, United Kingdom

Finding models that are more descriptive than the standard Diffusion Tensor is necessary in producing better disease biomarkers that provide more specificity to the physiological changes and sensitivity to the pathological impact of the disease in the body. This work investigates which compartment models of diffusion MRI are best at describing the signal from in-vivo human brain white matter, and how reproducible these results are across acquisition sessions.This study helps clinicians and medical physicists in choosing models for future in-vivo brain microstructure imaging.

Alessandro Proverbio¹, Bernard M. Siow², Daniel Alexander³, and Adam Gibson^1
1Department of Medical Physics and Bioengineering, University College London, London, United Kingdom, ²Centre for Advanced Biological Imaging, University College London, London, United Kingdom, ³Centre for Medical Image Computing and Department of Computer Science, University College London, London, United Kingdom

Non-invasive histology is a key research area. Diffusion NMR (dNMR) and Diffuse Optical Spectroscopy (DOS) can provide complementary information about the microstructure of a tissue using different contrast mechanisms. Here, we developed a biophysical model of the microstructure directly informed by dNMR and DOS signals, and we applied it to a sample to estimate three microstructural parameteres. This is a proof of concept that a low level combination of information obtained from dNMR and DOS can improve the accuracy of parameters estimatation. The model was developed in order to allow an extention to real tissues microstructure.

Sato Eida¹, Marc Van Cauteren², Yuka Hotokezaka¹, Makoto Obara², Tomoaki Okuaki², Ikuo Katayama¹, Miho Sasaki¹, Misa Sumi¹, and Takashi Nakamura^1
1Radiology and Cancer Biology, Nagasaki university, nagasaki, nagasaki, Japan, ²Philips Electronics, tokyo, tokyo, Japan

The area and integrity of the plasma membranes, as quantified by the total CPL, are critical determinants of the diffusion characteristics on the cellular level. The evolution over time of the ADC values is characteristic of whether the cells die via apoptotic or non-apoptotic (necrotic) pathway. Therefore, this can be used as a cancer biomarker for differentiating between different pathways of cell death, e.g. in testing the effects of cancer drugs in vitro.

Siawoosh Mohammadi¹, Karsten Tabelow², Thorsten Feiweier³, Joerg Polzehl², and Nikolaus Weiskopf^1
1Wellcome Trust Centre for Neuroimaging at UCL, ION UCL, UCL, London, London, United Kingdom, ²Weierstrass Institute for Applied Analysis and Stochastics, Berlin, Berlin, Germany, ³Healthcare Sector, Siemens AG, Erlangen, Bayern, Germany

Recent studies suggest that Diffusion Kurtosis Imaging (DKI) is more sensitive to gray microstructure than the well-known diffusion tensor imaging (DTI). However, DKI suffers from a low signal-to-noise ratio (SNR), since it is based on multiple and high b-value data. Thus, in-vivo high-resolution DKI with small voxel sizes has not been available on clinical scanners. We aimed to overcome the low SNR issue by using a novel version of the position orientation adaptive smoothing (POAS), which is separately applied on grey and white matter masks.

Alexander L. Sukstanskii^1
1Radiology, Washington University, St. Louis, Missouri, United States

Apparent diffusion coefficient (ADC) is analyzed for the case of oscillating diffusion sensitizing gradients. The exact analytical expressions are obtained for the high-frequency expansion of ADC for an arbitrary number of oscillations. The validity conditions of the high-frequency expansion of ADC are analyzed. It is shown that these conditions are substantially different for cos- and sin-type gradients. For the cos-type gradients, the high-frequency regime is reached whenT<<t_D (T is the period of oscillations, t_D is the characteristic diffusion time), whereas for the sin-type gradients, this regime is achieved when the total diffusion time t is smaller t_D.

Letizia Squarcina¹, Denis Peruzzo¹, Filippo Arrigoni², Fabio Triulzi³, and Alessandra Bertoldo^1
1Department of Information Engineering, University of Padova, Padova, Italy, Italy, ²Neuroimaging Unit, Scientific Institute, IRCCS Eugenio Medea, Bosisio Parini, Lecco, Italy, ³Neuroimaging Unit, Fondazione IRCCS Ca' Granda, Ospedale Maggiore Policlinico, Milano, Italy, Italy

The aim of this work is to understand if diffusion brain data are better described by a bi-exponential model rather than with the conventional mono-exponential. We investigate if all brain areas or only a critical portion of them benefit from the more complicated bi-exponential model and if the derived diffusion indexes obtained with the two models are comparable. The bi-exponential model resulted to fit the data significantly better in all brain areas, with a higher impact in term of diffusion indexes on those where the estimated fraction volume suggests two components.

Yuriko Suzuki¹ and Marc Van Cauteren^2
1Philips Electronics Japan, Minato-ku, Tokyo, Japan, ²Philips Healthcare Asia Pacific, Minato-ku, Tokyo, Japan

Diffusion Kurtosis Imaging (DKI) is a method to study the deviation of water diffusion from Gaussian distribution, which has been successfully applied to several brain diseases recently. However, in the organ with plenty of flow, DKI curve fitting is ruined. The aim of study is to define the DKI model which consider perfusion/flow component, and using this model, investigate how the conventional DKI fitting is affected by perfusion/flow. Our study has shown that large perfusion/flow fraction induced lower R2 of fitting, overestimated D and large error of estimated K.

Benoit Scherrer¹, Maxime Taquet^1,2, Onur Afacan¹, and Simon K. Warfield^1
1Computational Radiology Laboratory, Boston Children's Hospital, Harvard Medical School, Boston, MA, United States, ²ICTEAM Institute, Université catholique de Louvain, Louvain-la-Neuve, Belgium

Multiple works have shown that the diffusion attenuation in a voxel significantly deviates from a monoexponential decay. However, when imaging with a clinical scanner with long δ, it is not clear that the signal arising from a single fascicle is non-monoexponential. We hypothesize that the non-monoexponential behavior observed in voxels may reflect predominantly the presence of multiple fascicles with heterogeneous orientation and the presence of an additional compartment that is macroscopically isotropic. We imaged a region containing a single fascicle orientation, the body of the corpus and investigated the residual diffusion decay after subtracting the contribution of unrestricted diffusion.

Darya Morozov¹, Leah Bar², Nir Sochen², and Yoram Cohen^1
1School of Chemistry, Tel Aviv University, Tel Aviv, Israel, ²School of Mathematical Sciences, Tel Aviv University, Tel Aviv, Israel

Diffusion NMR is a powerful tool for gleaning microstructural information on opaque systems. We demonstrate that our modeling of d-PFG MR experiments, performed on a series of phantoms of increasing complexity where the ground truth is known a priori, can, without assuming the number of compartments, identify the number of restricted compartments, detect their sizes and determine their relative populations. In addition the model can identify free diffusion when presents in addition to the restricted compartments. This model was then used to study diffusion in excised nerves, showing that it provides a mean to obtain detailed microstructural information in nerves.

Darya Morozov¹, Leah Bar², Nir Sochen², and Yoram Cohen^1
1School of Chemistry, Tel Aviv University, Tel Aviv, Israel, ²School of Mathematical Sciences, Tel Aviv University, Tel Aviv, Israel

Diffusion NMR is a powerful tool for gleaning microstructural information on opaque systems. We demonstrate that our modeling of s-PFG MR experiments, performed on a series of phantoms of increasing complexity where the ground truth is known a priori, can, without assuming the number of compartments, identify the number of restricted compartments, detect their sizes and determine their relative populations. In addition the model can identify free diffusion when presents in addition to the restricted compartments. This model was then used to study diffusion in excised nerves, showing that it provides a mean to obtain detailed microstructural information in nerves.

Pim Pullens^1,2, Alard Roebroeck², Rainer Goebel², and Kamil Uludag^2
1Biomedical MR Imaging and Spectroscopy Group, Image Sciences Institute, University Medical Center Utrecht, Utrecht, Netherlands, ²Maastricht Brain Imaging Center, Maastricht, Netherlands

The susceptibility in white matter fiber is orientation-dependent and, thus, affects the gradient-echo signal. Rotating a fibrous phantom in the magnetic field, we also observed signal changes in diffusion-weighted spin-echo experiments. Consequently, diffusion parameters, such as FA and ADC, can be distorted by these orientation-dependent effects. In the study, Monte-Carlo simulations were performed to investigate the effect of susceptibility changes in a PGSE experiment. We found in our simulations that diffusion-weighted signal, FA and ADC change as a result of susceptibility differences. This could have an impact on human DW-MRI measurements especially at high and ultra-high fields.

Jens H. Jensen^1,2, Ali Tabesh^1,2, and Joseph A. Helpern^1,2
1Radiology and Radiological Science, Medical University of South Carolina, Charleston, SC, United States, ²Center for Biomedical Imaging, Medical University of South Carolina, Charleston, SC, United States

The diffusion orientation distribution function (dODF) can be used to estimate the directions of axonal fiber bundles and may be combined with white matter fiber tractography algorithms. Here a new analytical representation for the dODF is presented that includes the leading non-Gaussian corrections through the diffusional kurtosis tensor. By using numerical simulations, it is shown that including the non-Gaussian corrections both improves the accuracy of the predicted fiber bundle directions and allows for the direct quantification of intra-voxel fiber crossings. This explicit formula for the dODF may be of utility when diffusional kurtosis imaging data is available.

Tingting Wang¹, Hui Zhang¹, Matt G. Hall¹, and Daniel C. Alexander^1
1University College London, London, United Kingdom

Two classes of fibre orientation dispersion are observed in biological tissue: 1) Macroscopic dispersion, a population of straight fibres with different orientation, e.g. crossing and fanning fibres. 2) Microscopic dispersion, individual fibres with varying orientation, e.g. undulating fibres. Current reconstruction techniques implicitly assume macroscopic dispersion. Here we construct virtual tissue environments and conduct Monte-Carlo simulations to study differences that arise in the water dispersion, FA, and parameter estimates from biophysical models. Our results suggest that, the ODF can be recovered accurately in macroscopic dispersion, but not with small scale microscopic dispersion. We suggest considering such effects in future modelling works.

Gregory T. Baxter¹ and Lawrence R. Frank^1
1Radiology, UCSD, La Jolla, CA, United States

Sen and Basser developed a theoretical model for diffusion in white matter in the brain in the long-time limit. Their model represents white matter fascicles as coated cylinders in periodic arrays, with distinct axonal core, myelin sheath, and surrounding medium. We implement this white matter model in a Monte Carlo diffusion simulator to find results in the context of DTI. Anisotropy calculated from numerical simulations matches the theoretical model in the long-time limit, but anisotropy is significantly understated when diffusion time is short. Anisotropy is also sensitive to the concentration of spins in the myelin sheath.

Jing-Rebecca Li¹ and Denis Le Bihan^2
1Equipe DEFI, INRIA Saclay, Palaiseau Cedex, Palaiseau Cedex, France, ²NeuroSpin, CEA, Gif-sur-Yvette Cedex, Gif-sur-Yvette Cedex, France

We numerically simulated the magnetization of diffusing water molecules in a two compartment tissue model consisting of permeable cells, with cylindrical and spherical shapes, and extra-cellular space subject to the pulsed gradient spin echo (PGSE) sequence at various diffusion times by numerically solving the Bloch-Torrey partial differential equation using a finite volume spatial discretization coupled with a Runge-Kutta Chebyshev time-stepping method. The simulated results are consistent with experimental findings in rat cortex in vivo.

Junzhong Xu¹ and John C. Gore^1
1Institute of Imaging Science, Vanderbilt University, Nashville, TN, United States

Many practical anti-cancer treatments induce tumor microstructural variations at both supra- and subcellular levels, and detecting changes across different length scales is likely to provide more comprehensive information about tumor status. In this study, we used histology-based simulations and imaging in vivo of mouse xenografts to demonstrate the importance of probing various length scales of tumors. The results suggest that, in order to obtain comprehensive information about tumor status or to maximum the sensitivity for monitoring tumor response to treatment, diffusion measurements with a broad range of diffusion times may be necessary

Marco Palombo^1,2, Andrea Gabrielli³, Giancarlo Ruocco^1,2, and Silvia Capuani^1,2
1Physics Department, Sapienza University, Rome, Rome, Italy, ²CNR IPCF UOS Roma, Sapienza University, Rome, Rome, Italy, ³ISC-CNR, Rome, Rome, Italy

The anomalous diffusion parameter , which quantifies sub-diffusion processes was investigated as a function of the sphere-density . We demonstrated, by using numerical simulation and experimental results that, unlike conventional tortuosity investigations based on long diffusion time behavior of D(t)/D0, quantifies the global structural complexity (disorder) of heterogeneous systems. Moreover, we showed that  can be measured by using -varying PFG experiment which is faster than the diffusion time behavior of D(t)/D0 study. As a consequence we speculate that our approach could be used as a new tool to probe changes in microstructural properties in healthy and pathological tissues.

Kwan-Jin Jung^1
1Psychology, Carnegie Mellon University, Pittsburgh, PA, United States

The brain tissue is known to be bi-exponential in diffusion weighting at a high b value. The diffusion spectrum imaging (DSI) employs a wide range of b values with a peak value greater than 5000 and hence the diffusion in DSI should be considered to be bi-exponential. However, the conventional analysis of DSI has not considered the bi-exponential diffusion weighting. The bi-exponential diffusion weighting in DSI has been accounted for by replacing the acquired b with an adjusted b* which can be obtained by a simple nonlinear curve fitting. The bi-exponentially adjusted b* was experimentally confirmed to be effective in improving the fiber tracts particularly through the multiple fiber crossing regions.

Hazem Eltawell¹, Stephan Anderson¹, Osamu Sakai¹, and Hernan Jara^1
1Boston University Medical Center, Boston, MA, United States

The aim of our work was to discern the effects of the brain growth and development on global diffusion coefficients measured by correlation time technique DCT, in comparison to the standard DPFG technique. 27 subjects (0.5 to 24 years old) brains were scanned using 1.5 T scanner, DW-SE-sshEPI, and mixed-TSE pulse sequence, images were processed by qMRI algorithm to produce parametric maps used to calculate global DPFG and DCT histograms. Statistical analysis of the Peak DPFG and DCT histograms gives a power fit with different degrees of correlation, R2=0.92; R2=0.47 respectively, and maximum 17% age (myelination) related difference.

Sheryl L. Herrera¹, Trevor J. Vincent², Jonathan D. Thiessen^1,3, Richard Buist³, and Melanie Martin^2,4
1Physics and Astronomy, University of Manitoba, Winnipeg, Manitoba, Canada, ²Physics, University of Winnipeg, Winnipeg, Manitoba, Canada, ³Radiology, University of Manitoba, Winnipeg, Manitoba, Canada, ⁴Radiology, Physics and Astronomy, University of Manitoba, Winnipeg, Manitoba, Canada

Oscillating gradient spin echo (OGSE) sequences probe the shortest possible diffusion time scales so that the transition from restricted to hindered diffusion within the smallest structures can be detected. Here we simulate different geometries using OGSE sequences to determine the ability of the OGSE sequences to distinguish small pore sizes. As expected, OGSE frequencies corresponding to scales in the range between completely restricted and free diffusion best distinguish restriction sizes. Combining these measurements with other techniques such as AxCaliber and ActiveAx can allow for the inference of restriction sizes in the samples.

Greg D. Parker^1,2, David Marshall², Paul L. Rosin², Nicholas Drage³, Stephen Richmond³, and Derek K. Jones^1
1CUBRIC, School of Psychology, Cardiff University, Cardiff, South Glamorgan, United Kingdom, ²School of Computer Science, Cardiff University, Cardiff, South Glamorgan, United Kingdom, ³School of Dentistry, Cardiff University, Cardiff, South Glamorgan, United Kingdom

We present a novel method for robust estimation of fibre orientation distribution functions through diffusion weighted signal outlier rejection. Our algorithm combines aspects of the Richardson-Lucy spherical deconvolution with a non negative sparse coder to produce an adaptive signal dictionary that 'learns' compensations for common non-axially symmetric noise/corruption while preserving signals arising from complex fibre architecture. This has the effect of improving both fODF estimation and (by studying the dictionary adaptations) the robustness of outlier identification.

Chantal M.W. Tax¹, Ben Jeurissen², Max A. Viergever¹, and Alexander Leemans^1
1Image Sciences Institute, University Medical Center Utrecht, Utrecht, Netherlands, ²Department of Physics, University of Antwerp, iMinds-Vision Lab, Antwerp, Belgium

Spherical deconvolution techniques characterize complex fiber configurations within a voxel. Currently, the response function (RF) needed for deconvolution is derived from voxels with the highest fractional anisotropy (FA). Poor accuracy of FA in the high b-value and low SNR regime and the ad-hoc nature of selecting these voxels complicate RF estimation, which may lead to the detection of false positive peaks and altered peak magnitudes. In this work, the computation of the RF is optimized by excluding "crossing fibers" voxels in a recursive framework, which does not rely on FA. Feasibility is demonstrated on simulated and real diffusion MRI data.

Muwei Li¹, J. Tilak Ratnanather², Yajing Zhang¹, Hangyi Jiang¹, Kenichi Oishi¹, Michael I. Miller², and Susumu Mori^1
1The Russell H. Morgan Department of Radiology and Radiological Science, Johns Hopkins University School of Medicine, Baltimore, MD, United States, ²Center for Imaging Science, Johns Hopkins University, Baltimore, MD, United States

A knowledge-guided path-finding approach is proposed in this study. Combined with the automated brain segmentation, this approach can automatically generate 106 tracts based on our anatomical knowledge.This tool can be used to investigate tract-specific white matter anatomical states in various brain disorders.

Peter F. Neher¹, Bram Stieltjes², Ivo Wolf³, Hans-Peter Meinzer¹, and Klaus Hermann Fritzsche^1,2
1Medical and Biological Informatics, German Cancer Research Center (DKFZ), Heidelberg, Germany, ²Quantitative Image-based Disease Characterization, German Cancer Research Center (DKFZ), Heidelberg, Germany, ³Medical Informatics, Mannheim University, Mannheim, Germany

Evaluation of diffusion MRI-based tractography methods is challenging and existing approaches for quantitative assessment mainly focus on measures of general tract overlap or the presence of anatomically plausible connections between brain regions. We present an evaluation of the local fiber direction plausibility on a voxel-wise basis. We further propose a reference-based validation procedure that aims at evaluating tractograms obtained from data of clinical quality using high quality HARDI data as a reference.

Etay Ziv¹, Julia P. Owen¹, Yi-Ou Li¹, Eric J. Friedman², and Pratik Mukherjee^1
1University of California, San Francisco, San Francisco, California, United States, ²International Computer Science Institute, Berkeley, California, United States

We apply a Variational Bayesian approach to modularity analysis of the structural connectome in normal adult subjects and compare this VBMOD algorithm to the two most widely used module detection algorithms within the brain network community. We demonstrate the superiority of VBMOD to these existing methods in precision, accuracy and robustness to noise, both in module identification and cardinality inference. These findings are consistent over a broad range of thresholds used to binarize the networks. Our results are of interest to researchers in the connectomics and diffusion tractography literature.

Delphine Duclap¹, Benoit Schmitt¹, Alice Lebois¹, Pamela Guevara², Hui Zhang³, Clarisse Longo Dos Santos¹, Denis Le Bihan¹, Jean-François Mangin¹, and Cyril Poupon^1
1I2BM, CEA NeuroSpin, Gif-sur-Yvette, France, ²University of Concepción, Concepción, Chile, ³Department of Computer Science and Centre for Medical Image Computing, University College London, London, United Kingdom

Super-resolution techniques now provide an efficient strategy to improve the spatial resolution of individual diffusion maps. We extended this approach to a population of subjects and proposed a novel half-millimeter isotropic probabilistic atlas of the human brain connectivity, using the ARCHI database acquired in the frame of the European CONNECT connectome project. We demonstrate the efficacy of this approach combining a super-resolution technique based on high density streamline tractography and a robust segmentation technique of the white matter bundles.

Alice Lebois¹, Delphine Duclap¹, Benoit Schmitt¹, Clarisse Longo Dos Santos¹, Pamela Guevara², Hui Zhang³, Denis Le Bihan¹, Jean-François Mangin¹, and Cyril Poupon^1
1I2BM, CEA-NeuroSpin, Gif-Sur-Yvette, France, France, ²University of Concepcion, Concepcion, Chile, Chile, ³Department of Computer Science and Centre for Medical Image Computing, University College of London, London, United Kingdom

This work is focused on the construction of two probabilistic atlases of the quantitative T1 and T2 parameters of the human brain from the relaxometry data of the CONNECT/ARCHI database, an open MRI database of 79 healthy subjects built under the framework of the European CONNECT project which aims at inferring the connectome atlas of the human brain from functional and anatomical MRI data at 3T. We combined these atlases to density maps of the major white matter bundles we extracted from the diffusion data of the 79 subjects to compute the average T1/T2 relaxation times for each bundle.

Benoit Schmitt¹, Delphine Duclap¹, Alice LEBOIS*¹, Pamela Guevara², Hui Zhang³, Yaniv Assaf⁴, Fabrice POUPON¹, Denis Riviere¹, Yann Cointepas¹, Denis Le Bihan¹, Jean-François Mangin¹, and Cyril POUPON^1
1I2BM, CEA - NeuroSpin, GIF-SUR-YVETTE, FRANCE, France, Metropolitan, ²University of Concepcion, CONCEPCION, CHILE, Chile, ³Department of Computer Science & Centre for Medical Image Computing, University College London, LONDON, UNITED KINGDOM, United Kingdom, ⁴Department of Neurobiology, Tel-Aviv University, TEL-AVIV, ISRAEL, Israel

Providing a map of the human connectome is today a major challenge of diffusion. To reach this goal, two projects were launched: Human Brain Connectome in the USA and CONNECT project in Europe. This study was done in the frame of CONNECT project and included 79 subjects. In this study, we built the first probabilistic atlas of the human brain connectivity and diffusion quantitative statistics along bundles. We believe this novel atlas will help to infere the human brain connectome and further to provide some microstructural information for each bundle.

Maxime Taquet^1,2, Benoit Scherrer¹, Benoit Macq², and Simon K. Warfield^1
1Computational Radiology Laboratory, Harvard Medical School, Boston, Massachusetts, United States, ²ICTEAM Institute, Université catholique de Louvain, Louvain-la-Neuve, Belgium

DTI are unable to distinguish fascicles crossing in one voxel, making them impractical to study the white matter microstructure and the brain connectivity. Multi-tensor models are a valuable option in these contexts. However, their reconstruction requires acquisitions of DWI at multiple b-values, unlike all datasets acquired with a single-shell HARDI sequence. This results in a waste of money, effort and imaging time. In this study, we propose a method to circumvent the multiple b-values requirement, allowing the community to use their existing datasets to reconstruct multi-fascicle models and conduct further analyses of the white matter at no additional cost.

Stelios Angeli¹, Nicholas Befera², Gary Cofer², G. Allan Johnson², and Christakis Constantinides^1
1University of Cyprus, Nicosia, Nicosia, Cyprus, ²Radiology, Duke University Medical Center, Durham, NC, United States

The unique myocardial architecture, comprised of oriented laminar sheets and helical spiral tracts of fibers, accounts for its efficient contractile and torsional mechanical function. The inherent structural-functional associations of the heart also underline the potential significance of fiber tractography in remodeling or cellular disarray, following early, or late pathological states. Diffusion MRI tractography can provide 3-dimensional (3D) mapping of the myofiber structure . This study employs 3D, microscopic, spin-echo, diffusion-weighted MRI to construct a fiber atlas of the ex-vivo, fixed, C57BL/6 murine heart.

Su-Chun Huang^1,2, Gregory J. Pauley¹, Todd L. Richards¹, Neva M. Corrigan³, Dennis W. Shaw³, Alan A. Artru⁴, Annette Estes⁵, Stephen R. Dager³, and Natalia M. Kleinhans^1,3
1Integrated Brain Imaging Center, Department of Radiology, University of Washington, Seattle, WA, United States, ²Institute of Biomedical Engineering, National Taiwan University, Taipei, Taiwan, ³Department of Radiology, University of Washington, Seattle, WA, United States, ⁴Department of Anesthesiology, University of Washington, Seattle, WA, United States, ⁵Speech and Hearing Sciences, University of Washington, Seattle, WA, United States

Here we introduce a new analysis pipeline named tractography-based voxel-wise analysis (TBVA), which combines the advantages of TBSS and probabilistic tractography. This approach shows individual tract bundles in MNI standard space and performs voxel-wise comparisons within the tract bundle. We employed TBVA on the cingulum bundle (CB) in a group of participants with autism spectrum disorders (ASD) and age and IQ matched controls, and compared the results from conventional TBSS to the TSVA tractography results.

Farida Grinberg¹, Ivan I. Maximov¹, Ezequiel Farrher¹, and Nadim Jon Shah^1,2
1Institute of Neuroscience and Medicine 4 - Medical Imaging Physics, Forschungszentrum Juelich, Juelich, Germany, ²Department of Neurology, Faculty of Medicine, JARA, RWTH Aachen University, Aachen, Germany

Diffusion-based tractography is restricted by an uncertainty of determination of the diffusion directionality in the regions where fibre alignment anisotropy is low or fibre configuration is complex. In this work, we exploit the higher fractional anisotropy of the slow diffusion component which refers, in frame of the established models, to the restricted axonal water fraction. We show that advanced, model-based fibre tracking allows one to enhance reliability and visualisation of various fibre tracks, especially at the borders between white and grey matter. Potential benefits are discussed also in the context of combined functional and structural connectivity studies.

Franco Pestilli¹, Jason Yeatman¹, Ariel Rokem¹, Kendrick Kay¹, and Brian Wandell^2
1Psychology, Stanford University, Stanford, CA, United States, ²Psychlogy, Stanford University, Stanford, CA, United States

We propose a method to evaluate the accuracy of white-matter connectomes generated using diffusion weighted MRI and fiber tractography. The method predicts the diffusion data in each voxel as a weighted sum of the contributions from all the fascicles in that voxel. We derive several connectomes using different tractography algorithms (TEND, FACT and CSD). We evaluate each connectome using cross-validation with respect to a second set of diffusion measurements in the same subject. Connectomes generated using the CSD algorithm with an intermediate order of spherical harmonic basis set (4>lmax<16) predicts diffusion measurements better than data reliability.

Yung-Chin Hsu¹ and Wen-Yih Isaac Tseng^1,2
1Center for Optoelectronic Biomedicine, College of Medicine, National Taiwan University, Taipei, Taiwan, ²Graduate Institute of Brain and Mind Sciences, College of Medicine, National Taiwan University, Taipei, Taiwan

A registration method for unlabeled curves under the LDDMM framework is proposed, and this method is applied to register two right cingulum bundles for demonstration. The results show that the proposed method could effectively transform the tract bundles so that the global shapes are matched well. By registering the tract bundles to a common space, this method would be a valuable tool to investigate the brain morphology, which is a hot topic in neuroimaging studies in development, ageing or mental disorders.

Peter F. Neher¹, Bram Stieltjes², Frederik B. Laun³, Hans-Peter Meinzer¹, and Klaus Hermann Fritzsche^1,2
1Medical and Biological Informatics, German Cancer Research Center (DKFZ), Heidelberg, Germany, ²Quantitative Image-based Disease Characterization, German Cancer Research Center (DKFZ), Heidelberg, Germany, ³Medical Physics in Radiology, German Cancer Research Center (DKFZ), Heidelberg, Germany

We present an interactive graphical tool, Fiberfox, to generate arbitrary fiber configurations and corresponding diffusion weighted images (DWI). The intuitive user interface allows for the generation of any number of straight, bended, crossing, kissing, twisting and fanning fiber bundles, simply by placing 3D-ROIs used as waypoints for the synthetic fibers. From these fibers, a DWI is calculated applying a user defined number of gradient directions, b-value, SNR, image size and spacing. Fiberfox is implemented in the open source medical imaging interaction toolkit (MITK) and will also be available as free download included in the next release of MITK Diffusion.

Fulvia Palesi^1,2, Donald Tournier^3,4, Fernando Calamante^4,5, Niels Muhlert⁶, Gloria Castellazzi^2,7, Declan T. Chard⁶, Egidio Ugo D'Angelo^2,8, and Claudia Angela M. Wheeler-Kingshott^6
1Department of Physics, University of Pavia, Pavia, Italy, ²Brain Connectivity Center, IRCCS C. Mondino, Pavia, Italy, ³Melbourne Brain Centre - Austin Campus, The Florey Institute of Neuroscience and Mental Health, Heidelberg, Victoria, Australia, ⁴Department of Medicine, Austin Health and Northern Health, University of Melbourne, Melbourne, Australia, ⁵Melbourne Brain Centre - Austin Campus, Florey Institute of Neuroscience and Mental Health, Heidelberg, Victoria, Australia, ⁶NMR Research Unit, Queen Square MS Centre, Department of Neuroinflammation, UCL Institute of Neurology, London, United Kingdom,⁷Department of Industrial and Information Engineering, University of Pavia, Pavia, Italy, ⁸Department of Public Health, Neuroscience, Experimental Medicine, University of Pavia, Pavia, Italy

The study of cerebellum’s connectivity networks is challenging. This study assessed the feasibility of using advanced diffusion imaging methods to reconstruct whole contralateral cerebellar-cortical tracts in vivo. We combined two techniques: track reconstruction with constrained spherical deconvolution (CSD) algorithm and seed/target placement through the creation of super-resolution maps using track density imaging (TDI). We reconstructed tracts by placing ROIs both on b0 and TDI images: tracts appeared more anatomically accurate when using TDI maps rather than b0 images and inter-subject variability of cerebellar FA was found to be more reproducible when using ROIs placed on TDI.

Ezequiel Farrher¹, Farida Grinberg¹, A. Avdo Celik², Kaveh Vahedipour², and Nadim Jon Shah^2,3
1Institute of Neuroscience and Medicine 4 - Medical Imaging Physics, Forschungszentrum Juelich, Jülich, Germany, ²Institute of Neuroscience and Medicine 4 - Medical Imaging Physics, Forschungszentrum Jülich, Jülich, Germany, ³Department of Neurology, Faculty of Medicine, JARA, RWTH Aachen University, Aachen, Germany

We present a new multiple-fibre design of an anisotropic fibre phantom for diffusion MRI applications. The valuable feature of the new design relates to the integration of several geometrical configurations and fibre populations in a single device. We have demonstrated the application of the design in the analysis of high angular resolution diffusion imaging (HARDI) data analysis methods. The new phantom is shown to be an excellent tool for the investigation and comparison of several of HARDI methods as well as fibre tractography algorithms.

Qing Ji¹, John O. Glass², and Wilburn E. Reddick^1
1Radiological Science, St.Jude Children's Research Hospital, Memphis, TN, United States, ²Radiological Science, St. Jude Children's Research Hospital, Memphis, TN, United States

In this study, deterministic and probabilistic tractography algorithms for measuring brain structural connectivity were quantitatively evaluated on data from 8 healthy subjects. Eighty-two brain anatomic structures were obtained on each subject using FreeSurfer and connections between these structures were calculated by both tractography techniques. Neuroanatomical connectivity metrics between the two tractography methods were significantly correlated and fiber tracks from the deterministic algorithm were spatially similar to fiber probability distributions created using the probabilistic algorithm. While there was a high degree of correspondence between the approaches, the probabilistic algorithm provides more connections among brain structures and the additional information of uncertainty.

Peter F. Neher¹, Bram Stieltjes², Ivo Wolf³, Hans-Peter Meinzer¹, and Klaus Hermann Fritzsche^1,2
1Medical and Biological Informatics, German Cancer Research Center (DKFZ), Heidelberg, Germany, ²Quantitative Image-based Disease Characterization, German Cancer Research Center (DKFZ), Heidelberg, Germany, ³Medical Informatics, Mannheim University, Mannheim, Germany

In research settings it is common practice to acquire diffusion weighted images with an isotropic resolution. In contrast, clinical datasets are often acquired with a relatively high in-plane resolution of about 1-2 mm but much thicker slices (2-5 mm). The impact of such anisotropic image resolutions on the outcome of fiber tractography has not been systematically analyzed yet. This work uses software phantom data to quantify the impact of anisotropic voxels on nine different publicly available and commonly used algorithms.

Michael Zeineh¹, Samantha J. Holdsworth¹, Mansi B. Parekh¹, Stefan Skare², and Roland Bammer^3
1Stanford, Stanford, CA, United States, ²Neuroradiology, Karolinska University Hospital, Stockholm, Sweden, ³Stanford University, Stanford, CA, United States

Abnormal neuronal connectivity may be important in disease states involving the medial temporal lobes (MTL), including Alzheimer’s disease and temporal lobe epilepsy. Our prior work pushed the resolving power of DTI to near-mm isotropic voxel size and utilized a detailed segmentation to characterize the major subregions and pathways of the MTL. However, DTI data is subject to echo-planar (EPI) distortion that complicates coregistration with structural images, making segmentation difficult. Additionally, this distortion can create inaccuracies in tractography. In this study, we applied distortion-correction to our source EPI data and found improved structural accuracy.

Jennifer Andreotti¹, Kay Jann¹, Lester Melie-Garcia², Thomas Dierks¹, and Andrea Federspiel^1
1Dept. of Psychiatric Neurophysiology, University Hospital of Psychiatry / University of Bern, Bern, Bern, Switzerland, ²Neuroinformatics Department, Cuban Neuroscience Center, Havana, Havana, Cuba

Objective: to compare properties and variability of whole-brain structural networks weighted by connectivity density or by communicability. Methods: Graphs nodes are defined by automatic parcellation of the brain while edges are defined by connectivity density or communicability based on the tractography maps. Global and local network properties as well as their variability were computed. Results: results show larger variability between than within subjects for all properties and weightings. Between subjects' variability is slightly reduced in weighted communicability networks. Discussion: This study shows that network measures may be used to compare subjects. Measures extracted with different weightings correlate well.

Julia P. Owen¹, Etay Ziv¹, Polina Bukshpun², Nicholas Pojman², Mari Wakahiro², Jeffrey I. Berman³, Timothy Roberts⁴, Elliott Sherr², and Pratik Mukherjee^1
1Radiology, UCSF, San Francisco, CA, United States, ²Neurology, UCSF, San Francisco, CA, United States, ³Radiology, Children's Hospital of Philadelphia, Philadelphia, PA, United States, ⁴Radiology, CHOP, Philadelphia, PA, United States

In this study, we evaluate the test-retest reliability of graph theory techniques as applied to the structural connectome of the human brain. We explore unweighted and weighted graph metrics, as well as other measures of consistency, such as edge weights and module assignments. Two cohorts are used, one group (n=10) was scanned twice on the same scanner and the other group (n=5) was scanned once at two different sites, on the same model of scanner with identical acquisition parameters.

Yu Zhang¹, Xue Wang², Yufen Chen², Michael Rozenfeld², Puneet Opal³, and Todd Parrish^1,2
1Biomedical Engineering, Northwestern University, Evanston, IL, United States, ²Radiology, Northwestern University, Chicago, IL, United States, ³Neurology, Northwestern University, Chicago, IL, United States

This study how model selection between bedpostx and qboot and direction issue will affect the probabilistic diffusion tractography.

Mingtao Wang¹, Caterina Caffes², Stephen T.C. Wong^1,3, and Kelvin K. Wong^1,3
1Department of Systems Medicine & Bioengineering, The Methodist Hospital Research Institute, Houston, TX, United States, ²Texas Children's Hospital, Houston, TX, United States, ³Department of Radiology, Weill Cornell Medical College, New York, NY, United States

In vivo diffusion tensor imaging (DTI) is a powerful tool to understand neuroanatomical connectivity and to evaluate neuronal integrity in mouse models. To resolve the complex connectome in a small brain in vivo, high spatial resolution imaging are crucial. We present a high resolution in vivo DTI and tractography method for mouse brain at 150�~150�~150 ƒÊm3 resolution at 9.4T. The protocol includes a new low-cost elastomeric padding head restrain to significantly reduce head motion. High resolution isotropic diffusion tensor tractography offers high quality fractional anisotropy map and consistent fiber tracking for connectivity studies between brain regions.

Kazuyuki Endo¹, Tomohiko Horie¹, Isao Muro¹, Yutaka Imai², Kagayaki Kuroda³, Tetsuo Ogio⁴, Makoto Obara⁴, Marc Van Cauteren⁴, Thomas C. Kwee⁵, and Taro Takahara^6
1Department of Radiological Technology, Tokai University Hospital, Isehara, Kanagawa, Japan, ²Radiology, Tokai University Hospital, Isehara, Kanagawa, Japan, ³Information Science and Technology, Tokai University, Hiratuka, Kanagawa, Japan, ⁴Philips Healthcare, Minato-ku, Tokyo, Japan, ⁵Radiology, University Medical Center Utrecht, Utrecht, Utrecht, Netherlands, ⁶Biomedical Engineering, Tokai University, Isehara, Kanagawa, Japan

Improved Motion Sensitized Driven Equilibrium�iiMSDE�jis a kind of T2-prep pulse with motion probing gradients. it can be widely used for clinical applications including black-blood vessel wall imaging , lower extremity MR venography , or suppression of flow artifacts in contrast-enhanced studies for brain metastasis detection . The purpose of this study was to investigate basic appropriate parameters and evaluate the feasibility of this technique as �gFlow sensitive cine MR imaging�h. Flow sensitive cine MR imaging with iMSDE pre-pulse allows visualization of flow phenomena very clearly, which was not able with conventional cine MR imaging.

Javier Urriola Yaksic¹, Nyoman Dana Kurniawan¹, Aiman Al Najjar¹, and David C. Reutens^1
1Centre for Advanced Imaging, The University of Queensland, Brisbane, Queensland, Australia

Increasing numbers of channels developed for parallel imaging have enabled higher image signal-to-noise ratio and faster acquisition time, while increasing image quality with a reduction of geometrical distortions in EPI sequences. However, little is known about the effects of coil type and acceleration factor to the accuracy of fibertracking of the whole brain. Our results showed that the choice of coil types and acceleration factors are important in quantitative diffusion fibertracking. The increasing number of false aberrant tracks offsets the advantage of shorter acquisition time and inhomogeneity artefacts when the HARDI data was acquired using larger number of acceleration factors.

Kun Zhou¹ and Wei Liu^1
1Siemens Shenzhen Magnetic Resonance Ltd., Shenzhen, Guangdong, China

In diffusion-weighted imaging (DWI), B1 inhomogenetiy induced signal nonuniformity artifacts may degrade image quality and affect diagnosis. A modified DW-EPI sequence with build-in double-angle B1+ mapping mechanism is introduced to address this problem. With the modified sequence, the B1 inhomogeneity information can be calculated and used to reduce related artifacts in DWI images.

Tristan Anselm Kuder¹ and Frederik B. Laun^1,2
1Medical Physics in Radiology, German Cancer Research Center (DKFZ), Heidelberg, Germany, ²Quantitative Imaging-Based Disease Characterization, German Cancer Research Center (DKFZ), Heidelberg, Germany

NMR diffusion measurements allow collecting the signal from the whole sample. This mainly eliminates the problem of vanishing signal at increasing resolution. It has been a longstanding question if the exact shape of closed pores can be determined by diffusion measurements. Here, we present a method using short diffusion gradient pulses only, which is able to reveal the shape of arbitrary closed pores. Compared to former approaches, the method has reduced demands on relaxation times due to faster convergence to the diffusion long time limit and allows for a more flexible NMR sequence design since stimulated echoes can be used.

Robert Frost¹, Peter Jezzard¹, David A. Porter², and Karla L. Miller^1
1FMRIB Centre, Nuffield Department of Clinical Neurosciences, University of Oxford, Oxford, United Kingdom, ²Healthcare Sector, Siemens AG, Erlangen, Germany

Readout-segmented echo-planar imaging (rs-EPI) acquires robust, high-quality diffusion-weighted images using a navigated, multi-shot acquisition. However, the scan times are extended roughly by the number of readout segments compared to single-shot echo-planar imaging. Here we present simultaneous multi-slab acceleration in a 3D multi-slab rs-EPI sequence that enables full brain coverage in SNR-optimal TR of 1-2s, which is difficult to achieve with conventional 2D or 3D methods. The modified sequence is used to acquire 1.5mm isotropic trace-weighted data with whole-brain coverage.

Fernando Emilio Boada¹ and Steven Yutzy^2
1Radiology, New York University, New York, New York, United States, ²Radiology, University of Pittsburgh, Pittsburgh, Pennsylvania, United States

A sampling and reconstruction scheme for increasing angular resolution in Diffusion Spectrum Imaging is presented and demonstrated at 3T

Tzu-Cheng Chao^1,2, Jr-Yuan Chiou³, Cheng-Wen Ko⁴, Ming-Ting Wu⁵, Stephan E. Maier³, and Bruno Madore^3
1Department of Computer Science and Information Engineering, National Cheng-Kung University, Tainan, Taiwan, ²Institute of Medical Informatics, National Cheng-Kung University, Tainan, Taiwan, ³Department of Radiology, Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts, United States, ⁴Department of Computer Science and Engineering, National Sun Yat-Sen University, Kaohsiung, Taiwan, ⁵Department of Radiology, Kaohsiung Veteran General Hospital, Kaohsiung, Taiwan

An accelerated sampling strategy is proposed that can effectively reduce both geometric distortion and scan time in the context of the high Angularly Resolved Diffusion Imaging(HARDI) diffusion-weighted imaging method. The approach combines sampling strategies inspired from accelerated dynamic imaging and from the rossing Fiber Angular Resolution of Intra-Voxel structure (CFRI) algorithm to resolve crossing fibers in a manner that is fast and fairly insensitive to susceptibility-induced field variations and geometric distortion.

Hans Knutsson¹ and Carl-Fredrik Westin^1,2
1Department of Biomedical Engineering, Medical Informatics, Linköping University, Linköping, Sweden, ²Department of Radiology, BWH, Harvard Medical School, Boston, MA, United States

We have presented a novel method for generating evenly distributed samples in a part of q-space that can be pre-specified in a general way. We have demonstrated the feasibility for two shapes, a sphere and a cube. The results are interesting from several points of view. There is a market tendency for the samples to group in shells indicating that the present work may provide a preferable alternative to recently proposed shell-interaction schemes . The distributions attained for the cube case are far from Cartesian, this may be an advantage in a sparse reconstruction, e.g. compressed sensing, setting.

Mitsuharu Miyoshi¹, Yuji Iwadate¹, and Hiroyuki Kabasawa^1
1Global Applied Science Laboratory, GE Healthcare Japan, Hino, Tokyo, Japan

B0 inhomogeneity and eddy current produce image distortion in DW-EPI. Diffusion Prepared 3D FSE (DP-Cube) was developed in this study. Diffusion Prepared pulse sequence with 3 phase cycle were used to eliminate image distortion and banding artifact. Flow Compensated MPG gradient and peripheral trigger were used to make the sequence insensitive to motion. Measured ADC value was higher than literature and DW-EPI because of motion artifacts. Distortion free Diffusion weighted images were acquired with DP-Cube.

Ileana Ozana Jelescu¹, Denis Le Bihan¹, and Luisa Ciobanu^1
1NeuroSpin, Gif-sur-Yvette, Essonne, France

A 3D diffusion-prepared fast imaging with steady-state free precession (DP-FISP) sequence was implemented at 17.2T to perform apparent diffusion coefficient (ADC) measurements in small biological samples (i.e. single cells). The sequence was tested on phantoms with different T₁s, including values similar to those of biological tissues, and was compared with a standard diffusion-weighted spin-echo (DW-SE) sequence. The FISP flip angle was optimized for minimal T₁ bias in ADC estimation (< 10%) for b-values up to 600 s/mm². Our sequence is less prone to artifacts than EPI and eight times faster than DW-SE. Future improvements include extension to higher b-values.

Rafael O'Halloran¹, Florian Knoll², Kristian Bredies³, Rudolf Stollberger², and Roland Bammer^1
1Radiology, Stanford University, Stanford, CA, United States, ²Institute of Medical Engineering, Graz University of Technology, Graz, Austria, ³Department of Mathematics and Scientific Computing, University of Graz, Graz, Austria

3D DTI acquisitions can be lengthy and as the demand for higher-resolution and more encoding directions increases, strategies for reducing scan times will be critical for adoption of these techniques into the clinic. Here undersampled diffusion-weighted SSFP is combined with a TGV² constrained reconstruction to reduce the scan time by a factor of 4-8. The effect of undersampling on the FA and the direction of the principle eigenvector is investigated and compared to a conventional gridding reconstruction.

Jian Xu¹, George Russell Glenn², Himanshu Bhat³, Stephen F. Cauley⁴, Kawin Setsompop⁵, Ray Lee⁶, Ali Tabesh², Jens H. Jensen², Joseph A. Helpern², and Keith A. Heberlein^7
1Siemens Healthcare USA, New York, NY, United States, ²Department of Radiology and Radiological Science, Medical University of South Carolina, Charleston, SC, United States, ³Siemens Medical Solutions, Charlestown, MA, United States, ⁴A.A. Martinos Center for Biomedical Imaging, Dept. of Radiology, Massachusetts General Hospital, Boston, MA, United States, ⁵A.A. Martinos Center for Biomedical Imaging, Dept. of Radiology, MGH, Boston, MA, United States, ⁶Princeton University, Princeton, NJ, United States, ⁷Siemens Healthcare USA, Charlestown, MA, United States

Diffusional kurtosis imaging extends diffusion tensor imaging by quantifying the non-Gaussian behavior of water diffusion. Established DKI acquisition protocols typically use multi-slice 2D EPI with multiple b-values (¡Ý) and typically 30 gradient directions, which requires 7-15 minutes for full brain coverage. Recently a slice-acceleration method acquiring multiple slices simultaneously with the blipped CAIPIRINHA technique and individually reconstructed using the slice GRAPPA reconstruction method was introduced. In this study, the combination of DKI and the aforementioned multi-slice acceleration technique was performed and initial comparisons of the resultant parametric maps are reported.

Carl-Fredrik Westin^1,2, Markus Nilsson³, Ofer Pasternak¹, and Hans Knutsson^2
1Department of Radiology, BWH, Harvard Medical School, Boston, MA, United States, ²Department of Biomedical Engineering, Medical Informatics, Linköping University, Linköping, Sweden, ³Department of Medical Radiation Physics, Lund University, Lund, Sweden

In this work we estimate diffusion tensors from double-PFG diffusion MRI using two types of measurements 1) paris of gradients that are collinear, and 2) pairs of gradients that are non-collinear. It is known from angular double-PFG that compartment shape anisotropy can be measured with varying the angle between the gradient pairs. The ratio between the measurements with 90 degrees (non-collinear) and the one with 0 degrees (collinear), depends on the compartments shape anisotropy. This work provides and alternative to the model of angular double-PFG introducing two different estimation of the diffusion tensor, where the ratio between them is related to the compartment shape anisotropy.

Gwendolyn Van Steenkiste¹, Ben Jeurissen¹, Jan Sijbers¹, and Dirk H.J. Poot^2
1iMinds-VisionLab, University of Antwerp, Antwerp, Belgium, ²Biomedical Imaging Group Rotterdam, Erasmus MC - University Medical Center Rotterdam, Rotterdam, Netherlands

Diffusion MRI typically employs large voxels to obtain sufficiently high SNR. Given the large voxel sizes, many voxels consist of a mixture of signals from different anatomical structures. To reduce the partial volume effect and retain high SNR, we propose a super resolution acquisition and reconstruction technique that directly computes high SNR and high resolution DTI parameters from a set of low resolution diffusion MRI data sets. Using simulations we show our technique outperforms direct high resolution acquisition and current super resolution reconstruction techniques which don't take into account the underlying diffusion model.

Maira Tariq¹, Torben Schneider², Daniel C. Alexander¹, Claudia Angela M. Wheeler-Kingshott², and Hui Zhang^1
1Department of Computer Science & Centre for Medical Image Computing, University College London, London, United Kingdom, ²NMR Research Unit, Queen Square MS Centre, Department of Neuroinflammation, UCL Institute of Neurology, London, United Kingdom

NODDI (Neurite Orientation Dispersion and Density Imaging) is a recent diffusion MRI technique for directly quantifying microstructual indices of neurites using clinically feasible acquisition. It has been shown to be able to disentangle the key microstructural contributors to the standard measures from diffusion tensor imaging (DTI). This study assesses the scan-rescan reproducibility of NODDI in healthy subjects in comparison to DTI. The results show that NODDI has high reproducibility comparable to that of DTI.

Evren Ozarslan¹ and Thomas H. Mareci^2
1Department of Radiology, Brigham and Women's Hospital, Harvard Medical School, Boston, MA, United States, ²Department of Biochemistry and Molecular Biology, University of Florida, Gainesville, FL, United States

Estimations of apparent diffusion coefficient (ADC) and tensor (ADT) from diffusion-weighted acquisitions necessitate computations of, respectively, the b-value and b-matrix associated with the employed pulse sequence. We show that these important quantities can be conveniently computed for any gradient waveform using a simple algorithm that requires a few lines of code. This is achieved by representing the waveform by a piecewise constant function. With this representation, our technique complements the existing methods commonly used to compute the effects of restricted diffusion, and provides a consistent and convenient framework for studies that aim to infer the microstructural features of the tissue.

Carson Ingo¹, Richard L. Magin¹, Luis Colon-Perez², William Triplett³, and Thomas H. Mareci^3
1Department of Bioengineering, University of Illinois at Chicago, Chicago, IL, United States, ²Department of Physics, University of Florida, Gainesville, FL, United States, ³Department of Biochemistry and Molecular Biology, University of Florida, Gainesville, FL, United States

We modeled diffusion in neural tissue from the perspective of the continuous time random walk. The characteristic diffusion decay is represented by the Mittag-Leffler function, which does not make a priori assumptions about the governing statistics. We then used entropy as a measure of the anomalous features for the characteristic function. DWI experiments were performed on fixed rat brains using a spectrometer at 17.6 Tesla at b-values arrayed up to 25,000 s/mm^2. In white and gray matter regions, the Mittag-Leffler and entropy parameters demonstrated new information regarding sub-diffusion and produced different image contrast from that of the classical diffusion coefficient.

Virendra Mishra¹ and Hao Huang^1
1Advanced Imaging Research Center, University of Texas Southwestern Medical Center, Dallas, TX, United States

Single tensor fractional anisotropy (FA) values are significantly underestimated at the crossing-fiber regions. Single tensor FA, therefore, is biased in characterizing white matter integrity in clinical research. Tract analysis has obtained recent attention as individual white matter tracts have more clinical significance. In this abstract, we further explored an MTTA technique capable of restoring the FA values of the targeted tract along its path while correcting the bias in single tensor FA. Specifically, we directly tested MTTA with in vivo diffusion MRI acquired in routine clinical research and proposed weighted FA at the crossing-fiber voxels as an unbiased anisotropy metric.

Vincent J. Schmithorst¹, Jessica L. Wisnowski¹, and Ashok Panigrahy^1
1Radiology, Children's Hospital of Pittsburgh of UPMC, Pittsburgh, PA, United States

In crossing fiber regions, the Tract-Based Spatial Statistics (TBSS) algorithm for spatial normalization may be sub-optimal, as these voxels might not be projected onto the skeleton and thus excluded from analysis. We investigated the relation of functional network integration to DTI parameters and compared TBSS to an alternative technique involving normalization to a white matter template. While no region with a significant correlation was found with TBSS, a crossing-fiber region in the right hemisphere was detected via the alternative technique. Results suggest that normalization to a white matter template should be considered as a complementary analysis strategy to TBSS.

Kwan-Jin Jung¹, Nidhi Kohli², Fang-Cheng Yeh³, Timothy Keller⁴, and Tiejun Zhao^5
1Psychology, Carnegie Mellon University, Pittsburgh, PA, United States, ²Electrical and Computer Engineering, Carnegie Mellon University, Pittsburgh, Pennsylvania, United States, ³Biomedical Engineering, Carnegie Mellon University, Pittsburgh, Pennsylvania, United States, ⁴Psychology, Carnegie Mellon University, Pittsburgh, Pennsylvania, United States, ⁵MR R&D Collaborations, Siemens Medical Solutions USA, Pittsburgh, Pennsylvania, United States

The motion in diffusion imaging was more accurately estimated using simulated diffusion images rather than the b=0 images because the simulated diffusion images had the same contrast as the diffusion images. For diffusion imaging with a single shell b this method has been confirmed to be effective without extensive iteration even for a large amount of deliberate motion. Furthermore, this method has been extended to account for multiple exponential diffusion weighting in diffusion spectrum imaging where the b value ranges widely.

Burkhard Mädler¹ and Volker A. Coenen^2
1Neurosurgery, University Bonn, Medical Centre, Bonn, Germany, ²Divison Stereotaxy and MR-based OR Techniques, University Bonn - Medical Centre, Bonn, Germany

Diffusion-weighted MRI (DWI) is a non-invasive imaging technique that detects thermally driven, random motion of water molecules in living tissue and is able to characterize its interaction with cell membranes, macromolecules and potential diffusion barriers. Le-Bihan proposed a bi-exponential model to separate intra- and extracellular diffusion from incoherent motion of water molecules within randomly oriented capillaries – IVIM. The presence of noise may substantially affect IVIM parameter estimation. Mathematical inversion of multi-exponential processes like water diffusion in living tissue is a strongly ill-posed problem and the necessity for a large number of signal points and SNR is imminent for a reliable separation of more than one exponential time constant. The evaluation of current fitting modalities for IVIM data and investigating their limitations in relation to SNR and number of b-values is therefore prudent.

Che-Wei Chang^1,2, Chien-Chang Ho¹, and Jyh-Horng Chen^1,2
1Electrical Engineering, National Taiwan University, Taipei, Taiwan, ²Interdisciplinary MRI/MRS Lab, National Taiwan University, Taipei, Taiwan

This study presents a texture based analysis, Local Binary Pattern on Three Orthogonal Planes (LBP-TOP), to extract effective features from raw DTI data. Examples of sex classification and age estimation were used to demonstrate the performance of this method. A total 204 subject downloaded from NKI/Rockland Samples were used to evaluate those approaches. Our results show that the best sex classification accuracy is 0.81, and the best age estimation mean average error is 6.32 years. We demonstrated that LBP-TOP is capable of extracting effective information from DTI and could be a good candidate for classifying or evaluating neurological diseases based on raw DTI data.

Sjoerd B. Vos^1,2, Murat Aksoy², Julian R. Maclaren², Zhaoying Han², Samantha J. Holdsworth², Alexander Brost², Christoph Seeger^2,3, Alexander Leemans¹, and Roland Bammer^2
1Image Sciences Institute, University Medical Center Utrecht, Utrecht, Netherlands, ²Center for Quantitative Neuroimaging, Department of Radiology, Stanford University, Stanford, California, United States, ³Pattern Recognition Lab, Department of Computer Science, Friedrich-Alexander-University Erlangen-Nuremberg, Erlangen, Germany

Recent 3D DWI pulse sequences can achieve resolutions up to 1mm isotropic, at the expense of longer acquisition times per volume. To acquire HARDI data with these high resolutions, the large number of diffusion directions required for HARDI can be spread out over different imaging sessions. We present a validated approach to acquire high-angular and high-spatial resolution diffusion imaging over different sessions, suggesting a ‘head holder’ that fits tightly around the subject’s head and in the coil. Accurate subject repositioning is ensured in each session, resulting in seamless fusion of data from different sessions.

Siawoosh Mohammadi¹, Chloe Hutton¹, Zoltan Nagy¹, Oliver Josephs¹, and Nikolaus Weiskopf^1
1Wellcome Trust Centre for Neuroimaging at UCL, ION UCL, UCL, London, London, United Kingdom

One important and still unresolved artefact in diffusion MRI is physiological movement, which can lead to severe bias in the signal. We developed an easy-to-implement extension of the original diffusion tensor model to account for physiological noise using measures of peripheral physiology (pulse and respiration), the so-called extended tensor model. We tested the performance of four different extended tensor models with different physiological noise regressors on non-gated and gated diffusion tensor imaging data, and compared it to an established data-driven robust fitting method. We showed that the framework of the extended tensor model facilitates investigation into physiological noise in DTI.

Michael A. Levine¹, Pavlina Polaskova¹, Sara Maria Sprinkhuizen¹, Steven M. Stufflebeam¹, Bruce R. Rosen¹, Jayashree Kalpathy-Cramer¹, and Elizabeth R. Gerstner^1
1Athinoula A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital, Charlestown, MA, United States

Fifteen glioblastoma patients treated with combination cediranib and cilengitide were enrolled in a correlative magnetic resonance imaging substudy. Ventricular apparent diffusion coefficient values were found to be statistically different across vendors. Multi-center clinical trials should be particularly careful to ensure that results acquired from different vendors are comparable.

Jeff Kershaw^1,2, Christoph Leuze³, Ichio Aoki¹, Takayuki Obata⁴, Iwao Kanno¹, Hiroshi Ito¹, Yuki Yamaguchi², and Hiroshi Handa^2
1Molecular Imaging Centre, National Institute of Radiological Sciences, Chiba, Japan, ²School of Bioscience and Biotechnology, Tokyo Institute of Technology, Yokohama, Japan, ³Department of Neurophysics, Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany, ⁴Research Centre for Charged Particle Therapy, National Institute of Radiological Sciences, Chiba, Japan

Recent work applying the oscillating-gradient spin-echo sequence to diffusion-tensor imaging of in vivo rat brain has found substantial, approximately linear changes with frequency to the eigenvalues (EVs) of the apparent diffusion tensor and to the fractional isotropy (FA). While the FA is a well known function of the EVs, it is not immediately clear how the behaviour of the EVs relates to the trends in the FA. This work investigates how the FA is affected by changes to the EVs.

Elodie André¹, Christophe Phillips^1,2, Ezequiel Farrher³, Ivan I. Maximov⁴, Farida Grinberg³, Nadim Jon Shah^4,5, and Evelyne Balteau^1
1Cyclotron research center, University of Liège, Liège, Belgium, ²Department of Electrical Engineering and Computer Science, University of Liège, Liège, Belgium, ³Institute of Neuroscience and Medicine - 4, Forschungszentrum Juelich, Juelich, Germany, ⁴Institute of Neuroscience and Medicine - 4, Forschungszentrum Juelich GmbH, Juelich, Germany, ⁵Department of Neurology, Faculty of Medicine, RWTH Aachen University, JARA, Aachen, Germany

The use of high b-values in diffusion kurtosis imaging makes the derived parameters very sensitive to low signal to noise. Here we show the dependence of mean kurtosis on SNR and demonstrate that noise correction is a necessary step, leading to more reproducible metrics.

Ken E. Sakaie¹, Kecheng Liu², and Mark J. Lowe^1
1Imaging Institute, The Cleveland Clinic, Cleveland, OH, United States, ²Siemens Medical Solutions USA, Inc, Malvern, PA, United States

Diffusion tensor imaging provides quantitative measures of tissue integrity that can be biased by the properties of the noise distribution. Sum of squares is commonly used for reconstructing multi-coil imaging data and exhibits Rician bias for signals near noise floor, leading to substantial bias in diffusivity values. Here, we quantify the bias and demonstrate that a simple but effective correction that can be applied retrospectively.

Damien McHugh^1,2, Penny L. Hubbard^1,2, Geoffrey Parker^1,2, and Josephine H. Naish^1,2
1Centre for Imaging Sciences, University of Manchester, Manchester, United Kingdom, ²The University of Manchester Biomedical Imaging Institute, Manchester, United Kingdom

The effect that imaging gradients have on b-values in oscillating and pulsed gradient diffusion sequences are compared. The impact this has on parameters obtained from model fitting is then evaluated using synthetic data from three models: monoexponential decay, biexponential decay, and the kurtosis model. It is shown that the diffusion-weighting introduced by imaging gradients in oscillating gradient sequences has little impact on parameter estimates for each model. Conversely, ignoring the influence of imaging gradients in pulsed gradient sequences is shown to lead to significant errors in fitted parameters, especially for the biexponential model.

David C. Newitt¹, Ek T. Tan², Thomas L. Chenevert³, Lisa J. Wilmes¹, Suchandrima Banerjee⁴, Luca Marinelli², and Nola M. Hylton^5
1Radiology and Biomedical Imaging, University of California, San Francisco, San Francisco, CA, United States, ²Diagnostics and Biomedical Technologies, GE Global Research, Niskayuna, New York, United States, ³Radiology – MRI, University of Michigan, Ann Arbor, Michigan, United States, ⁴Applied Science Lab, GE Healthcare, Menlo Park, California, United States, ⁵Radiology and Biomedical Imaging, University of California San Francisco, San Francisco, CA, United States

Gradient nonlinearity (GN) is a significant source of error for quantitative diffusion MRI, and the extent of GN varies with the scanner used. This confounds results from multi-center and longitudinal studies required in clinical trials. GN effects are of particular concern in breast imaging, where the anatomy has large offsets from magnet isocenter. Retrospective GN correction (GNC) was evaluated in a multi-center setting with phantoms and in normal and breast cancer subjects. GNC significantly reduced the spatial-dependence of ADC values and improved quantitative accuracy of ADC, which could in turn improve the sensitivity of cancer-detection and cancer-monitoring in diffusion MRI.

Christopher Leslie Adamson¹, Richard Beare¹, Deanne Thompson¹, and Marc Seal^1
1Developmental Imaging, Murdoch Childrens Research Institute, Parkville, VIC, Australia

We present an improved method for correction of affine distortions in high b-value diffusion weighted imaging due to eddy currents and subject movement. We show that our method improves the alignment and scaling of the registration results. The improved registration has implications for DTI and tractography applications.

Pan Su¹, Xiaodong Ma², Wenchuan Wu², and Hua Guo^2
1School of Electronic and Information Engineering, Beihang University, Beijing, China, ²Center for Biomedical Imaging Research, Department of Biomedical Engineering, School of Medicine, Tsinghua University, Beijing, China

Phase inconsistencies among different shots in multi-shot diffusion imaging causes artifacts in final images. Oversampling in k-space center is thus required to remove these errors, however, it inevitably lengthens the readout window and introduces further blurring artifacts. In this study, a new method is proposed to estimate object function and motion-induced phase map in muti-shot spiral DWI jointly, and is validated by the simulation results. Our proof-of-concept work demonstrates that this method estimates the desired artifact-free image more accurately than self-navigator based technique when navigator information is limited, which has potential utilization in high resolution diffusion imaging.

Daniel Güllmar¹, Stefan Kämmer², and Jürgen R. Reichenbach^1
1Medical Physics Group, Institute of Diagnostic and Interventional Radiology I, Jena University Hospital - Friedrich Schiller University Jena, Jena, Germany,²Faculty of Computer Science and Automation, Technical University Ilmenau, Ilmenau, Germany

A simple and fast algorithm for assessment of DTI spike noise based on the background signal was implemented and applied to more than 1500 data sets. The determined quality measure was used as indicator for hardware issues and correlated to recorded incidents in the past. We found that all relevant hardware problems which forward spike noise had have been detected using this approach.

Jack Harmer¹, Christian T. Stoeck², Rachel W. Chan³, Nicolas Toussaint¹, Constantin Von Deuster², David Atkinson³, and Sebastian Kozerke^1,2
1Division of Imaging Sciences and Biomedical Engineering, Kings College London, London, United Kingdom, ²Institute for Biomedical Engineering, University and ETH Zurich, Zurich, Switzerland, ³Centre for Medical Imaging, University College London, London, United Kingdom

In recent work diffusion acquisition schemes that employ stimulated echoes (STE) have been used to allow diffusion tensor imaging of the beating heart. Higher order localized image based shimming can partially compensate for the increased field inhomogeneity at high field strength but, at the same time, compromises the effectiveness of fat saturation techniques such as SPIR across the entire field-of-view. We present a dual navigated STE sequence with Slice Selective Gradient Reversal (SSGR) applied to the second and third RF pulses in order to suppress the fat signal in in-vivo data of the human heart at 3T.

Karl G. Helmer¹, Ming-Chung Chou², Ronny I. Preciado³, Allen Song⁴, Jessica Turner⁵, Barjor Gimi⁶, and Susumu Mori^7,8
1Radiology, Massachusetts General Hospital, Charlestown, MA, United States, ²Department of Medical Imaging and Radiological Sciences, Kaohsiung Medical University, Kaohsiung, Taiwan, ³Massachusetts General Hospital, Charlestown, MA, United States, ⁴Radiology, Duke University, Durham, NC, United States,⁵Translational Neuroscience, The Mind Research Network, Albuquerque, NM, United States, ⁶Radiology and Medicine, Dartmouth-Hitchcock Medical Center, Lebanon, NH, United States, ⁷Radiology, Johns Hopkins School of Medicine, Baltimore, MD, United States, ⁸Kennedy Kreiger Institute, Baltimore, MD, United States

We report on a method using histogram-similarity measures to establish normative data which can be used in the “calibration” of diffusion-weighted imaging data collected at a single or multiple sites. This method can be used to quantitatively determine the quality of new data acquired over time by comparing it to the normative data. We have calculated histogram similarity using different metric types, both within and between sites and have shown that these metrics are sensitive measures for both fractional anisotropy (FA) and mean diffusivity (MD). Statistical significance of the results was determined using simulations of two different histogram distributions.

Hu Cheng^1
1Indiana University, Bloomington, IN, United States

Information of head rotation during DTI data acquisition is critical for reliable fiber tracking. However, the information can be erroneously extracted by decomposing the affine matrix from image registration because of image distortions induced by Eddy current. In this work, we examined the feasibility of retrospective motion correction by using an ultra-short EPI navigator via GRAPPA operators. Based on the simulated data from real EPI scans, it is possible to use only three K-space lines to obtain accurate between-volume head motion. Our results are encouraging for implementing this technique in DTI pulse sequence.

Daniel Guo Quae Chong^1
1Dept. of Diagnostic, Interventional and Pediatric Radiology (DIPR), Inselspital, Bern, Bern, Switzerland

DWI is inherently T₂ weighted and can caused problem such as T₂ shine through. Relax DWI is a novel sampling scheme that can estimate T₂ map with only one additional b value measurement compare to normal DWI. Relax DWI uses minimum TE for each b values. Additional data point at low b value and long TE assist in fit stability. T₂ shine through can be corrected by remove T₂ component from DWI at different b values. Examples like abdominal organs and peritoneal metastasis are shown.

Rafael O'Halloran¹, Anh Tu Van¹, Eric Aboussouan¹, Murat Aksoy¹, and Roland Bammer^1
1Radiology, Stanford University, Stanford, CA, United States

Diffusion-weighted SSFP Imaging is an alternative to single-shot spin echo EPI that can provide high resolution (<1.5mm isotropic) DWI and DTI images. Because Diffusion-weighted SSFP is a multi-shot technique, motion-induced phase errors can cause reconstruction issues and most critically, can cause loss of steady-state phase coherence. Phase navigation has been shown to help retrospectively to mitigate loss of image magnitude but errors remain because phase errors accrued during a particular TR propagate into all coherence pathways derived from the transverse phase states of that TR. This ultimately results in disruption of the steady state. In this work phase navigation is used to prospectively correct the rigid-body component of motion-induced phase and is shown to improve signal magnitude in a phantom-based feasibility study.

Venkata Veerendranadh Chebrolu¹, Dattesh D. Shanbhag¹, Ek Tsoon Tan², Patrice Hervo³, Marc-Antoine Labeyrie⁴, Catherine Oppenheim^4,5, and Rakesh Mullick^1
1Medical Image Analysis Lab, GE Global Research, Bangalore, Karnataka, India, ²GE Global Research, Niskayuna, NY, India, ³GE Healthcare, Buc, France,⁴Departments of Radiology and Neurology, Centre Hospitalier, Sainte-Anne, Paris, France, ⁵Université Paris Descartes, Paris, France

The b-factor used for DWI imaging could vary due to factors such as gradient non-linearity and concomitant field effects. These system imperfections may differ between different vendors and MRI systems. This variation would impact multi-center studies evaluating the efficacy of the PDM criteria for thrombolysis treatment decision. In this work we demonstrate that in presence of b-factor variability a joint DWI and ADC based method may offer more robust perfusion-diffusion mismatch assessment compared to ADC alone based DWI lesion segmentation method.

Oleg Posnansky¹, Myung-Ho In¹, and Oliver Speck^1
1BMMR, IEP, Otto-von-Guericke University, Magdeburg, SA, Germany

Single-shot spin-echo echo-planar imaging allows for very fast signal acquisition and is frequently used in diffusion-weighting (DW) MRI. However, it is prone to severe distortions due to local variations in the static magnetic field, eddy currents induced by the DW gradients, susceptibility, and chemical shift, especially at ultra-high magnetic field. Geometric incongruence in DW images may impose significant errors in calculations of diffusion tensors and their derived metrics. We demonstrate that these imperfections can be corrected effectively by a combination of eddy-current correction and improved point-spread function mapping.

Bilan Liu¹, Tong Zhu², and Jianhui Zhong^2
1Electrical and Computer Engineering, University of Rochester, Rochester, NY, United States, ²Imaging Sciences, University of Rochester, Rochester, NY, United States

Diffusion tensor imaging (DTI) is susceptible to numerous artifacts. Therefore the Quality control (QC) of the Diffusion tensor images is critical for image interpretation and diagnostic accuracy. The goal of this study is to develop an experimental protocol to help choose between existing QC tools by analyzing the precision and accuracy of fractional anisotropy (FA), mean diffusivity (MD) calculated by those tools in the presence of major DTI-specific artifacts. Both simulation data obtained using Monte Carlo simulation and human in vivo data were used to help assess the effectiveness of those QC tools.

Ivan I. Maximov¹, Farida Grinberg², Irene Neuner^1,3, and Nadim Jon Shah^1,4
1Institute of Neuroscience and Medicine 4, Forschungszentrum Juelich GmbH, Juelich, Germany, ²Institute of Neuroscience and Medicine 4, Forschungszentrum Juelich, Juelich, Germany, ³Department of Psychiatry, Psychotherapy and Psychosomatics, RWTH Aachen University, Aachen, Germany,⁴Department of Neurology, RWTH Aachen University, Aachen, Germany

Clinical diffusion imaging frequently suffers from multiple artefacts. As a consequence, many corrupted datasets cannot be used in further analysis and/or medical treatments. We have developed a robust post-processing framework that allows one to recover the degraded datasets and to return them to use. In order to demonstrate the advantages of the developed framework, the results obtained by the robust approach with other post-processing algorithms are compared. We show the benefits of the robust post-processing framework using voxelwise analysis by TBSS software from FSL package.

Ali Ersoz¹, Volkan Emre Arpinar², and L. Tugan Muftuler^2,3
1Department of Biophysics, Medical College of Wisconsin, Milwaukee, WI, United States, ²Department of Neurosurgery, Medical College of Wisconsin, Milwaukee, WI, United States, ³Center for Imaging Research, Medical College of Wisconsin, Milwaukee, WI, United States

We conducted a quantitative analysis of directional bias imposed on primary eigenvector (PE) estimations in DTI when gradient table is not reoriented after motion correction. We compared this bias with the inherent uncertainty in the estimation of orientation distribution functions. We found that 90% percent of the voxels in cingulum and CST tracts had substantially smaller directional adjustments on the PE than the uncertainty in their estimations. We also demonstrated that this directional bias might accumulate along the fiber tract and lead to inaccurate fiber tracking. The findings were validated using simulations.

Ek T. Tan¹, Seung-Kyun Lee¹, Jean-Baptiste Mathieu¹, Venkata Veerendranadh Chebrolu², Dattesh D. Shanbhag², Matthew A. Bernstein³, John Huston III³, Luca Marinelli¹, and John F. Schenck^1
1Diagnostics and Biomedical Technologies, GE Global Research, Niskayuna, NY, United States, ²Diagnostics and Biomedical Technologies, GE Global Research, Bangalore, India, ³Radiology, Mayo Clinic, Rochester, MN, United States

A head-only MRI system can have cost and performance advantages, especially in diffusion imaging. However, a dedicated head-only gradient coil tends to have a greater gradient field nonlinearity (GN) compared to a conventional whole-body gradient coil optimized for body imaging. This nonlinearity results in a larger spatial variation of the diffusion-encoding b-value. Diffusion imaging using a dedicated, head-only gradient coil without nonlinearity correction was found to produce significant, spatially-varying b-values that would result in highly inaccurate diffusion-weighted images, ADC and FA. However, the inaccuracies were correctable with GN correction employing 5th-order spherical harmonics. These results will be of interest in quantitative stroke and oncology imaging.

Dariya I. Malyarenko¹, Brian D. Ross¹, and Thomas L. Chenevert^1
1Radiology - MRI, University of Michigan, Ann Arbor, Michigan, United States

Relatively large ADC bias error, well exceeding measurement noise, can be observed for anatomical regions imaged distant from magnet isocenter. Gradient nonlinearity is a main source of bias. Our previously-described approach allowed substantial reduction of spatial ADC bias for media of arbitrary anisotropy using DWI encoding along any three orthogonal directions. This work focuses on practical aspects of ADC correction implementation via system nonlinearity tensor for derivation of correctors applicable to DWI or b-map generation. Correction performance was evaluated for isotropic and anisotropic media.

Jing Yuan¹, Qinwei Zhang¹, David Ka Wai Yeung¹, Anil T. Ahuja¹, and Ann D. King^1
1Department of Imaging and Interventional Radiology, Chinese University of Hong Kong, Shatin, NT, Hong Kong

Maximum likelihood estimation (MLE) is preferable as an unbiased estimator compared to least-squares estimation (LSE) due to the Rician noise distribution of magnitude MR image. Accurate noise estimation is essential for MLE that affects accuracy and precision of MLE, but is often hampered by many factors in practice. We investigate the effects of erroneous noise estimation on ADC estimation by MLE through simulation and ADC mapping of clinical DWI images. Results show that MLE accuracy and precision are significantly reduced by noise estimation error at low signal-to-noise ratios (SNRs), but exhibit fairly good robustness to such errors at high SNRs>10.

Ana-Maria Oros-Peusquens¹ and Nadim Jon Shah^1,2
1INM-4, Research Centre Jülich, Jülich, Germany, ²Faculty of Medicine, JARA, RWTH Aachen University, Aachen, Germany

We present a method combining acceleration of multi-contrast diffusion data using UNFOLD with denoising based on singular value decomposition (SVD). UNFOLD is ideally suited for acceleration of q-space acquisition because q-space information is extracted from the Fourier transform of the q-dependent diffusion signal. The information is largely contained in the q-space centre, a region unaffected by aliasing at moderate UNFOLD acceleration (2-4). Multi-direction HARDI signal also has Fourier components in a restricted interval and is thus very well suited for UNFOLD acceleration. SVD denoising works extremely well for q-space data and also, but less spectacularly, for quantitative information obtained with HARDI. In conclusion, acceleration and denoising of q-space and HARDI data is demonstrated, using simple methods with numerically undemanding reconstruction. The same method should work well for combinations of q-space and HARDI, such as diffusion spectrum imaging (DSI).