Electronic Posters : Diffusion & Perfusion - Neuro
Click on to view the abstract pdf and click on to view the video presentation.
ADC & DTI Methods

Wednesday May 11th
Exhibition Hall  14:00 - 16:00 Computer 74

14:00 3891.   Diffusion model complexity reduces repeatability in multiple b-value DWI fitting : Impact of tumour volume and fitting methodology in a phase I clinical trial setting 
Matthew R Orton1, David J Collins1, Christina Messiou1, Jean Tessier2, and Martin O Leach1
1CR-UK and EPSRC Cancer Imaging Centre, Institute of Cancer Research, Sutton, Surrey, United Kingdom, 2Formerly with Early Clinical Development, AstraZeneca, Alderley Park, Macclesfield, United Kingdom

Diffusion Weighted Imaging with multiple b-values is being used more widely in many application areas, including clinical trials for assessment of novel cancer therapeutics. Complex models are being increasingly used to extract more information from the data, and the suitability and interpretation of these is a matter of current debate. In this abstract we present results obtained from data acquired in a phase I clinical trial setting which demonstrate the effect of the diffusion model complexity, fitting methodology and tumour volume on the repeatability of functional parameter estimates.

14:30 3892.   Evaluation of a Novel Continuously Distributed Diffusion Model in Normal Human Brain 
He Wang1, Yong Zhang1, and Guang Cao1
1Global Applied Science Laboratory, GE Healthcare, Shanghai, Shanghai, China, People's Republic of

A novel continuously distributed exponential model was explored to quantify diffusion in normal human brain. Diffusion-weighted images were acquired with twenty-one b-values ranging from 0 to 4000 s/mm2 to calculate the continuous distribution of diffusion coefficients in white matter, gray matter and cerebrospinal fluid. White matter showed different diffusion rates among three orthogonal orientations as contrast to CSF and gray matter with similar diffusion rates in all directions. More widely dispersed diffusion distribution in white matter benefits from the application of the continuously distributed exponential model. This technique may be used to characterize diffusion heterogeneity in different types of tumors.

15:00 3893.   New strategy for registering DW and non-DW images via tensor estimation metric 
Cheng Guan Koay1,2, Andrew L. Alexander1, and M. Elizabeth Meyerand1
1Department of Medical Physics, University of Wisconsin-Madison, Madison, WI, United States, 2STBB, National Institutes of Health, Bethesda, MD, United States

Registration of DW and non-DW images is a critical data-analytics step. In this study, we proposed a novel DW and non-DW image registration strategy that provides the much needed information on the goodness-of-registration of DWI images to non-DWI (T2-weighted image or template) in terms of tensor estimation metric within the registration process. This strategy requires very minimal modification to the existing acquisition of DTI except two distinct b-values and is built upon weighted linear least squares estimation for computational efficiency consideration. Reduction in registration error as computed from the metric was about 48.6% for the experimental data set we tested.

15:30 3894.   Statistical Comparison of DT-MRI Interpolation Methods Using Cardiac DT-MRI Data 
Jin Kyu Gahm1,2, Nicholas Wisniewski3, William S Klug4, Alan Garfinkel3,5, and Daniel B Ennis1,6
1Department of Radiological Sciences, University of California, Los Angeles, CA, United States, 2Department of Computer Science, University of California, Los Angeles, CA, United States, 3Department of Medicine, University of California, Los Angeles, CA, United States, 4Department of Mechanical and Aerospace Engineering, University of California, Los Angeles, CA, 5Department of Physiological Science, University of California, Los Angeles, CA, United States, 6Biomedical Engineering Interdepartmental Program, University of California, Los Angeles, CA, United States

DT-MRI interpolation is the process of estimating diffusion tensors at arbitrary points in space from regularly sampled tensor data. Tensor interpolation is important for tensor-based fiber tractography, registration, volume rendering, and computational model building. In this work, we use bootstrap statistical methods to compare four different DT-MRI interpolation methods’ accuracies for recovering the tensor shape (invariants) and orientation of unknown tensors from known tensor data. By using a cardiac DT-MRI dataset, we show the statistical bootstrapping results for the paired comparisons, and present recommendations for the selection of the DT-MRI interpolation method.

Thursday May 12th
  13:30 - 15:30 Computer 74

13:30 3895.   Six is Enough? Examining the Controversy of 6 versus 30 Diffusion Encoding Directions for Deterministic Tractography of Human Brain 
Catherine Lebel1, Thomas Benner2, and Christian Beaulieu3
1Biomedical Engineering, University of Alberta, Edmonton, AB, Canada, 2Athinoula Martinos Center for Functional and Structural Biomedical Imaging, Harvard University, Boston, MA, United States, 3Biomedical Engineering, University of Alberta, Edmonton, Alberta, Canada

Simulations suggest that using more than the minimum 6 diffusion-encoding gradient directions provides more robust DTI parameter estimates. Deterministic tractography is often criticized for using only 6 directions despite the lack of experimental evaluation. Diffusion measurements were compared between 6 versus 30 directions in healthy volunteers for 12 major white matter tracts. FA, MD and volume differences were very minimal (1/12 tracts had FA or MD differences, 2/12 had different volumes). These results suggest that 6-direction DTI data with adequate SNR is sufficient for FA, MD, and volume measurements and thus should not be considered inferior for population DTI studies.

14:00 3896.   Effect of SNR of DTI on the structural network 
Hu Cheng1, Dae-Jin Kim1, Olaf Sporns1, Yang Wang2, Jinhua Sheng2, and Andrew Saykin2
1Indiana University, Bloomington, IN, United States, 2Indiana University, Indianapolis, IN, United States

DTI has been used for mapping the structural network of human brain. The network is constructed by choosing various brain regions as nodes and fiber tracts connecting the regions as links. To investigate the effect of DTI SNR on the constructed network, a set of DTI data with different SNR was created from the 48 direction dataset. The result shows that SNR affects the accuracy of network while has less effect on the variation.

14:30 3897.   The Reproducibility and Correlation of Phase Errors in Diffusion Weighted Imaging with the Cardiac Cycle 
Rafael Luis O'Halloran1, Samantha Holdsworth1, and Roland Bammer1
1Radiology, Stanford University, Palo Alto, CA, United States

One of the fundamental problems with multi-shot diffusion weighted MRI is phase mismatch between shots. The ultimate source of these errors is motion, however, one component of it is reproducible and correlated with the cardiac cycle. In this work the reproducibility of cardiac induced phase is investigated with a rapid time-resolved single-shot gradient echo DW-EPI. Phase maps showed that, although there is a large component of the phase that is highly correlated with the cardiac cycle, there are still residual variations in the phase that do not correlate.

15:00 3898.   Informed RESTORE for Removal of Physiological Noise Artifacts in Low Redundancy DTI Data 
Lin-Ching Chang1, Lindsay Walker2, Babak Behseta3, and Carlo Pierpaoli2
1Department of Electrical Engineering and Computer Science, The Catholic University of America, Washington, DC, United States, 2STBB, NICHD, National Institutes of Health, Bethesda, Maryland, United States, 3Pediatric & Developmental Neuroscience Branch, NIMH, National Institutes of Health, Bethesda, Maryland, United States

Artifacts are common in diffusion weighted images (DWIs) especially those originating from cardiac pulsation in ungated acquisitions and from subject motion. Neglecting to account for them can result in erroneous diffusion tensor values. The Robust Estimation of Tensors by Outlier Rejection (RESTORE) is an effective method for improving tensor estimation presence of artifactual data points. However, RESTORE relies heavily on data redundancy (large DWIs datasets), which may not be acquired in some clinical settings. This paper proposes a method called informed RESTORE (iRESTORE) that incorporates the notion that physiological noise artifacts are more likely to results in signal drops rather than signal increases to achieve an accurate rejection of artifactual data points in low redundancy DWIs datasets.

Electronic Posters : Diffusion & Perfusion - Neuro
Click on to view the abstract pdf and click on to view the video presentation.
Dynamic Contrast Enhancement Methods (DCE-MRI)

Monday May 9th
Exhibition Hall  14:00 - 16:00 Computer 75

14:00 3899.   Arterial Input Functions in Dynamic Contrast-Enhanced MRI: Magnitude Versus Phase 
Paul Wessel de Bruin1, Maarten J. Versluis1, Erlangga Yusuf2, Monique Reijnierse1, and Matthias J.P. van Osch1
1Radiology, LUMC, Leiden, ZH, Netherlands, 2Rheumatology, LUMC, Leiden, ZH, Netherlands

In DCE-MRI contrast agent concentration curves can be phase- or magnitude-based. We show that neither is perfect and that a hybrid approach is needed to combine the advantages of both methods.

14:30 3900.   MR Estimation of Arterial Input Function (AIF) in Dual Gradient Echo Sequences Using an Adaptive Model Trained by Standard Radiological AIF 
Hassan Bagher-Ebadian1,2, Tavarekere N Nagaraja3, Robert Knight1,2, Ramesh Paudyal1, Siamak P Nejad-Davarani1, Stephen Brown4, Sawyam Panda1, Polly Whitton1, Joseph D Fenstermacher3, and James R Ewing1,2
1Neurology, Henry Ford Hospital, Detroit, MI, United States, 2Physics, Oakland University, Rochester, MI, United States, 3Anesthesiology, Henry Ford Hospital, Detroit, MI, United States, 4Radiation Oncology, Henry Ford Hospital, Detroit, MI, United States

Estimating the arterial input function of a contrast agent, the time-concentration curve in plasma, has long presented a problem in MR dynamic contrast Enhanced and dynamic susceptibility studies. A set of time courses of radiolabeled-Gd-DTPA CA-concentrations in arterial plasma (time–activity curve) of 13 animals has been recently measured and named as Standard-Radiological-AIF (SRAIF). Herein, an adaptive model was used for predicting the time trace of AIF from Dual-Gradient-Echo (DGE) signal measured in the normal area of rat brains. A set of physically-meaningful extracted features from DGE sequences, an artificial neural network was trained to estimate the time-trace of SRAIF.

15:00 3901.   Effects of Artery Input Function on Dynamic Contrast Enhanced MRI for Determining Grades of Gliomas 
Na Zhang1, Lijuan Zhang1, Xin Liu1, Hairong Zheng1, Jeffrey Carpenter2, and Bob L Hou2
1Paul C. Lauterbur Research Center for Biomedical Imaging, Shenzhen Institute of Advanced Technology, Chinese Academy of Science, Shenzhen, Guangdong, China, People's Republic of, 2Radiology, West Virginia University, Morgantown, WV, United States

To select an optimum brain artery for generating an artery input function (AIF) is the most important step for differentiating gliomas. We applied a modified two-compartment Tofts model to calcalate Ktrans and Ve, and compared the values obtained by using different AIFs in four grades of gliomas (determined by biopsy) for evaluating impacts of the AIFs on grading the tumors. We found that selecting an AIF from MCA is the best selection for grading gliomas.

15:30 3902.   Construction of a Model-Based High Resolution Arterial Input Function (AIF) Using a Standard Radiological AIF and the Levenberg-Marquardt Algorithm 
Hassan Bagher-Ebadian1,2, Azimeh Noorizadeh3, Siamak P Nejad-Davarani1,4, Ramesh Paudyal1, Tavarekere N Nagaraja5, Robert Knight1,2, Stephen Brown6, Joseph D Fenstermacher5, and James R Ewing1,2
1Neurology, Henry Ford Hospital, Detroit, MI, United States, 2Physics, Oakland University, Rochester, MI, United States, 3Mechanical Engineering, Nuclear Engineering, University of Shiraz, Shiraz, Fars, Iran, 4Biomedical Engineering, University of Michigan, Ann Arbor, MI, United States, 5Anesthesiology, Henry Ford Hospital, Detroit, MI, United States, 6Radiation Oncology, Henry Ford Hospital, Detroit, MI, United States

Estimating the arterial input function of a contrast agent (CA), the time-concentration curve in plasma, has long presented a problem in MR dynamic contrast Enhanced and dynamic susceptibility studies. A successful model based on the blood-circulatory system for the construction of an algorithm for estimating the CA time-concentration curve in arterial plasma has been previously suggested. A set of time courses of radiolabeled-Gd-DTPA CA-concentrations in arterial plasma (time–activity curve) of 13 animals has been recently measured. Herein, the radiolabeled-time-concentration curve was used to estimate a set of parameters for constructing a blood-circulatory model AIF in a high temporal resolution.

Tuesday May 10th
  13:30 - 15:30 Computer 75

13:30 3903.   Intraarterial MR Perfusion Imaging of Meningiomas: Comparison to Digital Subtraction Angiography 
Steven W Hetts1, Alastair J Martin1, Christopher F Dowd1, Van V Halbach1, Randall T Higashida1, Michael McDermott2, Soonmee Cha1, and David Saloner1
1Radiology, UCSF, San Francisco, CA, United States, 2Neurosurgery, UCSF, San Francisco, CA, United States

IA MR perfusion techniques appear to be a useful adjunct to DSA in determining tumor vascularity and the source of that blood supply during DSA guided preoperative embolization procedures. IA MR perfusion techniques appear to be more sensitive than DSA in detecting residual vascularized tumor. Further correlations with intraoperative observations and pathologic specimens are warranted to better assess the overall sensitivity and specificity of IA MR perfusion and to determine its overall utility in comparison to DSA and IV perfusion methods.

14:00 3904.   Blood volume fraction mapping for angiogenesis assessment in a novel human glioblastoma stem cell model 
Teodora-Adriana Perles-Barbacaru1, Feriel Tiar2, Laurent Pelletier2, Didier Wion2, Francois Berger2, and Hana Lahrech1
1INSERM U836, functional and metabolic neuroimaging, Grenoble Institute of Neurosciences, University Joseph Fourier, Grenoble, France, 2INSERM U836, brain nanomedicine group, Grenoble Institute of Neurosciences, University Joseph Fourier, Grenoble, France

To assess angiogenesis in a novel orthotopic mouse model derived from human glioblastoma stem cells, quantitative blood volume fraction (BVf) mapping was performed using dynamic Rapid Steady State T1 MRI with the clinically approved low molecular weight contrast agent Gd-DOTA. While the tumor occurrence was hardly detectable with T2-weighted imaging during the 2nd month of tumor development, the tumor BVf was increased (0.034 ± 0.010) compared to the contralateral cerebral BVf (0.023 ± 0.004) without increased vascular permeability to Gd-DOTA. This slowly growing tumor model can be useful for the preclinical evaluation of antiangiogenic therapies.

14:30 3905.   Comparison of the uptake of Gadolinium contrast agents between pre-clinical colorectal and other tumour models by dynamic contrast enhanced magnetic resonance imaging. 
Ian Wilson1, G S Almeida1, Huw D Thomas2, David R Newell2, and Ross J Maxwell1
1Newcastle MR Centre, Newcastle University, Newcastle Upopn Tyne, Tyne and Wear, United Kingdom, 2Northern Institute of cancer Research, Newcastle University, Newcastle Upon Tyne, Tyne and wear, United Kingdom

Our focus is on the development of MR and PET imaging agents for the evaluation of novel cancer treatments. The aim of this study was to evaluate the uptake of two similar gadolinium contrast agents Gadoteridol and Gadobutrol in various pre-clinical human tumour xenograph models using dynamic contrast enhanced magnetic resonance imaging. Colorectal pre-clinical tumour models show significantly lower Gadolinium contrast agent uptake than other tumour models. These provisional DCE MRI experiments helped us to identify model-specific tumour uptake patterns that we used for further development of MR contrast agents and PET tracers.

15:00 3906.   Quantitative assessment of perfusion and permeability in osteochrondritis dissecans lesions: feasibility and initial results 
Andreas P. Arnoldi1, Michael Ingrisch2, Sandra Utzschneider3, Maximilian F. Reiser1, and Sabine Weckbach1
1Department of Clinical Radiology, Ludwig-Maximilians-University Munich, Munich, Bavaria, Germany, 2Josef Lissner Laboratory, Department of Clinical Radiology, Ludwig-Maximilians-University Munich, Munich, Germany, 3Department of Orthopedics, Campus Grosshadern, Ludwig-Maximilians-University Munich, Munich, Germany

Our study on perfusion quantification in 6 patients with OCD lesions of the knee and talus using a two compartment model shows the feasibility of DCE-MRI to detect lesions in patients with OCD and to quantify perfusion and permeability. This quantitative approach showing elevated perfusion and permeability parameters provides additional information on the pathophysiology of OCD, might influence disease management and provide a non-invasive tool for therapy monitoring.

Wednesday May 11th
  13:30 - 15:30 Computer 75

13:30 3907.   Adaptive Neural Network for Direct Quantification of Longitudinal Relaxation Rate Change (Capital Greek DeltaR1) in T One by Multiple Read Out (TOMROP) Sequence 
Hassan Bagher-Ebadian1,2, Meser M Ali3, Ali Seyd Arbab3, Malek Makki4, Siamak P Nejad-Davarani1,5, Sawyam Panda1, Quan Jiang1,2, and James R Ewing1,2
1Neurology, Henry Ford Hospital, Detroit, MI, United States, 2Physics, Oakland University, Rochester, MI, United States, 3Radiology, Henry Ford Hospital, Detroit, MI, United States, 4Diagnostic Imaging, University of Children Hospital of Zurich, Zurich, Switzerland, 5Biomedical Engineering, University of Michigan, Ann Arbor, MI, United States

Direct and accurate measurement of the temporal change in the longitudinal relaxation rate after injection of a paramagnetic contrast agent has become increasingly important in MR perfusion studies. The Look-Locker (LL) sequence provides accurate T1 estimates which generally accomplished by nonlinear multi-dimensional curve fitting. However, these fitting methods are sensitive to initial guesses, and the errors of the parametric estimates. Herein, two Adaptive Neural Networks (ANNs) were trained using an analytical model of the LL signal. The model-trained ANNs were applied to the MR data acquired from animal model and human. Results were also compared with those of conventional methods.

14:00 3908.   A numerical advection-diffusion model to fit dynamic contrast-enhanced MRI (DCE-MRI) data 
Nicolas Michoux1, Denis Rommel1, and Emmanuel Lefrançois2
1IMAG - Radiology Department, Université Catholique de Louvain, Brussels, Brussels, Belgium, 2UMR 6253 UTC-CNRS, Université de Technologie de Compiègne, Compiègne, France

Fitting kinetic models to DCE-MRI data is a useful method to study mechanisms such as perfusion and diffusion in tissues. Compartmental models with well-mixed spaces and diffusion as sole mechanism of mass transport are commonly used for this purpose. However, advection mechanism may play an important role in mass transport. Therefore, we assess a model of microvascular structure built upon a network of elementary segments. The advection-diffusion equation is then solved with a finite element approach to fit perfusion data. Although the advection-diffusion model requires a precise dimensioning, preliminary results suggest that DCE-MRI data can be fitted with such model.

14:30 3909.   Wide variations in cellular-interstitial water exchange rates are within the experimental uncertainty of AIF variations in their effect on uptake curve shapes for DCE-MRI Modelling 
Matthew R Orton1, David J Collins1, and Martin O Leach1
1CR-UK and EPSRC Cancer Imaging Centre, Institute of Cancer Research, Sutton, Surrey, United Kingdom

Accurate quantification of DCE-MRI data involves modelling or measurement of three processes: an AIF; transfer of contrast between various compartments; generation of MR signals. Ambiguities between the AIF and the MR signal generation process on tissue curves are investigated. Simulations show that only small AIF changes are needed to account for changes in the signal curves generated by a very wide range of cellular-interstitial water exchange rates. We conclude that the AIF and the MR signal model are strongly coupled – very accurate characteristaion of the AIF is needed to have confidence in in-vivo estimates of cellular-interstitial water exchange rates.

15:00 3910.   A Pharmacokinetic Model enabling Modelling of DCE-MRI data of normal and cancerous Liver 
Matthew R Orton1, David J Collins1, and Martin O Leach1
1CR-UK and EPSRC Cancer Imaging Centre, Institute of Cancer Research, Sutton, Surrey, United Kingdom

We present a combined model that describes the contrast enhancement patterns seen in liver tissue and in liver tumours, resulting in a compact but comprehensive methodology enabling voxel-wise fitting over entire liver volumes. By modelling the whole organ a more detailed understanding of liver tumours may be obtained, in particular the complexities surrounding growth and infiltration at tumour boundaries that are particular to metastatic liver disease. The model is demonstrated on DCE-MRI data of a liver containing neuro-endocrine tumours

Thursday May 12th
  13:30 - 15:30 Computer 75

13:30 3911.   Bayesian Estimation Improves Plasma Volume Repeatability with Compartmental Modelling of DCE-MRI Data 
Matthew R Orton1, David J Collins1, Christina Messiou1, Jean Tessier2, and M O Leach1
1CR-UK and EPSRC Cancer Imaging Centre, Institute of Cancer Research, Sutton, Surrey, United Kingdom, 2Formerly with Early Clinical Development, AstraZeneca, Alderley Park, Macclesfield, United Kingdom

The Extended Kety model is widely used for modelling DCE-MRI data. It is well known that estimates of the plasma volume fraction are subject to large errors compared with other parameters obtained with this model, which limits the utility of such estimates in trials and clinical practice. In this abstract we present a Bayesian estimation methodology that reduces test-retest repeatability of estimates by around 50% in comparison to least-squares estimates. This results in a more reliable measure that has similar repeatability to DC-CT based measures, and therefore has the potential to detect smaller changes as a result of therapeutic interventions.

14:00 3912.   Comparison of the kinetic parameters estimated with different numerical methods in DCE-MRI 
Cing-Ciao Ke1, Shin-Lei Peng1, Chih-Feng Chen2, Ho-Lin Liu3, and Fu-Nien Wang1
1Biomedical Engineering and Environmental Sciences, National Tsing Hua University, Hsinchu, Taiwan, 2Radiology, Chang Gung Memorial Hospital, Chiayi, Taiwan, 3Medical Imaging and Radiological Sciences, Chang Gung University, Taoyuan, Taiwan

The previous study only demonstrated that liner least-squares (LLSQ) method may be more useful method than nonlinear least ¡Vsquares NLSQ method in solving modification of Tofts and Kermode. We performed computer simulations to assess the accuracy of estimating the kinetic parameters for LLSQ and NLSQ methods in solving Tofts and Kermode model. At lower signal-to-noise (SNR), the accuracy of parameters estimated with LLSQ method is better than NLSQ method. Besides, the calculation velocity of LLSQ method was seventeen times faster than NLSQ method.

14:30 3913.   Combined analysis of perfusion and capillary permeability by parametric analysis of the tissue residue function from DCE-MRI 
Atle Bjornerud1,2, Tuva Hope1, Christopher Larsson1, Frederic Courivaud1, Raimo Aleksi Salo1, Knut Lote3, and Inge Andre Rasmussen1
1Interventional Centre, Oslo University Hospital, Oslo, Norway, 2Dept. of Physics, Univ. of Oslo, Norway, 3Dept. of Oncology, Oslo University Hospital, Oslo, Norway

We present a method for combined analysis of perfusion and capillary permeability by parametric analysis of the tissue residue function obtained by vascular deconvolution. Transfer constant (Ktrans) and intra/extravascular volume fractions (vi/ve) were derived and compared to the same values obtained with a conventional two-compartment kinetic model (std. analysis) in three glioma patients undergoing a standard treatment regime, including a total of 13 longitudinal scans. The proposed method gave Ktrans and ve values in very good agreement with those obtained with std. analysis whereas vi values were systematically higher with the proposed method.

15:00 3914.   Utility of Non-model based ‘Semi-quantitative’ Indices derived from Dynamic Contrast Enhanced T1-weighted MR Perfusion in Differentiating Treatment Induced Necrosis from Recurrent Progressive Brain Tumor. 
Jayant Narang1, Rajan Jain1,2, Syed Ali Arbab3, and Abbas Babajani-Feremi3
1Neuroradiology, Henry Ford Health System, Detroit, MI, United States, 2Neurosurgery, Henry Ford Health System, Detroit, MI, United States, 3Radiology, Henry Ford Health System, Detroit, MI, United States

The purpose of this study was to assess the utility of non-model based ‘semi-quantitative’ indices derived from dynamic contrast-enhanced T1-weighted magnetic resonance perfusion (DCET1MRP) in differentiating treatment induced necrosis (TIN) from recurrent/progressive tumor (RPT). Practical impact of DCET1MRP on routine neuro-oncologic imaging practice is limited by the need of complicated multi-compartment physiological models and intensive computational requirements to derive pharmacokinetic metrics and hence, the lack of an easy to use commercially available software. Our results show that these non-model based indices which are relatively easy to derive, robust and reproducible can be effectively used to differentiate TIN from RPT.

Electronic Posters : Diffusion & Perfusion - Neuro
Click on to view the abstract pdf and click on to view the video presentation.
Perfusion & Permeability: DSC - Methods

Monday May 9th
Exhibition Hall  14:00 - 16:00 Computer 76

14:00 3915.   Reliable estimation of capillary transit time distributions at voxel-level using DSC-MRI 
Kim Mouridsen1, Leif Østergaard1, Søren Christensen2, and Sune Nørhøj Jespersen1
1Center for Functionally Integrative Neuroscience, Aarhus University, Aarhus University Hospital, Aarhus, Denmark, 2Department of Neurology, Royal Melbourne Hospital, Melbourne, Australia

DSC-MRI has proven very useful for characterizing tissue perfusion indices such as CBF, CBV and MTT. However, whereas CBF and MTT only represent average capillary flow and transit time, important physiological parameters such as flow heterogeneity and oxygen extraction capacity depend on the shape characteristics of the complete distribution of transit times. We demonstrate in a simulation study, that by using a Bayesian estimation approach and a parameterization of the residue function based on a vascular model, shape characteristics of the residue function can be reliably estimated for a range of microvascular flow patterns.

14:30 3916.   Does R2* increase or decrease when contrast agent extravasates? A simulation study. 
Nicolas Pannetier1,2, Clément Debacker1,2, Franck Mauconduit1,2, Thomas Christen1,3, and Emmanuel Luc Barbier1,2
1U836, INSERM, Grenoble, France, 2Grenoble Institut des Neurosciences, Université Joseph Fourier, Grenoble, France, 3Department of Radiology, Stanford University, Stanford, California, United States

Quantification of Dynamic Susceptibility Contrast MRI remains a challenge when contrast agent (CA) extravasates into the interstitium. Indeed, CA extravasation increases tissue R1 relaxation(signal enhancement) and competes with R2* increase (signal destruction) used to compute perfusion estimates. However, the consequences of CA extravasation on R2* are often overlooked. CA extravasation reduces the magnetic susceptibility difference at the intra-/extravascular interface and CA extravasation yields to the emergence of magnetic susceptibility differences at the interfaces between cells and interstitium.In this study, we evaluated these two competing R2* effects in case of blood-brain barrier permeability using numerical simulations for various porosities and cell sizes.

15:00 3917.   Variability of model-based blood volume correction and vessel permeability estimation in dynamic susceptibility contrast MRI: A computer simulation study 
Lin-Wei Hsu1, Yeng-Peng Liao1, and Ho-Ling Liu1,2
1Institute of Medical Physics and Imaging Science, Chang Gung University, Taoyuan, Taiwan, 2Department of Medical Imaging and Intervention, Chang Gung Memorial Hospital, Taoyuan, Taiwan

Mathematical models have been previously developed to correct for T1 and T2 effects originated from contrast agent extravasations in dynamic susceptibility contrast MRI, and showed additional potentials for assessing the vessel permeability. This study aimed to assess variability of the blood volume (BV) correction and permeability estimation using computer simulations based on varied SNR, permeability and whether baseline T1 is available. The results showed that quality of BV correction and permeability estimation depended on both SNR and severity of contrast agent leakage. Accurate measurement of baseline T1 is especially helpful for the situation of high permeability and low SNR acquisition.

15:30 3918.   An Efficient Computational Approach to Characterize DSC-MRI Signals Arising from Heterogeneous Vascular Networks 
Natenael B Semmineh1, Junzhong Xu1, and Christopher Chad Quarles1
1Radiology and Radiological Sciences, Vanderbilt University, Nashville, TN, United States

Given the heterogeneous nature of blood vessels within tissue and the dependence of susceptibility field gradients on vascular geometry the assumption that a linear relationship, with a spatially uniform rate constant termed the vascular susceptibility calibration factor (kp), exists between the CA concentration and the measured transverse relaxation rate change, could significantly impact the reliability of DSC-MRI hemodynamic measurements. We propose the use of an efficient computational method to estimate the extent of kp heterogeneity across normal and tumor tissue and to assess the reliability of DSC-MRI measures of blood volume and blood flow.

Tuesday May 10th
  13:30 - 15:30 Computer 76

13:30 3919.   Effect of Cerebral Hemodynamic Changes on DTI Quantitation: A Hypercapnia Study 
Abby Ying Ding1,2, and Ed X Wu1,2
1Laboratory of Biomedical Imaging and Signal Processing, The University of Hong Kong, Hong Kong, Hong Kong SAR, China, People's Republic of, 2Department of Electrical and Electronic Engineering, The University of Hong Kong, Hong Kong, Hong Kong SAR, China, People's Republic of

The accuracy of DTI quantitation can directly affect the interpretation of underlying biological microstructures. DTI measurements can be confounded by the presence of vasculature and blood perfusion. So far, little is known about the extent of such effects in quantitative DTI. In this study, we quantitatively examined the effect of hypercapnia on quantitative DTI indices in vivo in rodent brains. DTI indices were found to alter during hypercapnia, including various diffusivities and FA. These findings indicated that alterations in physiologic condition, vascular characteristics and hemodynamic regulations can affect the in vivo quantitation of various DTI indices. Therefore, caution must be taken in designing experiments and interpreting DTI indices.

14:00 3920.   The effects of myelin in FA and QSI indices: control vs. Long Evans shaker rat brains 
Debbie Anaby1, Ian D Duncan2, and Yoram Cohen1
1School of Chemistry, Tel Aviv University, Tel Aviv, Israel, 2School of Veterinary Medicine, University of Wisconsin-Madison, Madison, Wisconsin, United States

The contribution of myelin on the observed diffusion anisotropy of water in neuronal tissue is not clear. Here we used high-b-value q-space DWI to probe the effects of myelin on diffusion indices in Long Evans shaker (les) and control rat brains. We compared the QSI indices with the FA which was extracted from the low-b-value regime of the QSI data. We found that the QSI indices are significantly different between the groups while the FA blurs the differences between them. Therefore, high b-value QSI should be considered as a means for achieving better distinction between les and control brain tissues.

14:30 3921.   On the Time to Peak Factor of Dynamic Susceptibility Contrast of Microbubbles 
Shin-Lei Peng1, Chih-Kuang Yeh1, Chung-Hsin Wang1, Hsu-Hsia Peng1, and Fu-Nien Wang1
1Department of Biomedical Engineering and Environme, National Tsing Hua University, Hsin-Chu, Taiwan

Recently, microbubbles were used as a MR susceptibility contrast agent because of the susceptibility differences by the gas-liquid interface. However, the relationship between rCBV deriving from Gd-DTPA and that from microbubbles has not been investigated. In this study, we aim to investigate correlation between rCBV from both contrast agents and find the key factor influencing the correlation. The results showed that the time-to-peak for microbubbles increases, the correlation between microbubbles and Gd-DTPA increases. When time-to-peak of microbubbles increases, it may mean that more microbubbles aggregate and trap in local tissue vasculature. As a result, the stronger susceptibility effect will be produced and lead to high correlate with Gd-DTPA.

15:00 3922.   DSC MRI on Rat Model: Choosing the Integration Interval for Measuring CBV 
Yi-Ling Wu1, Chien-Chung Chen1, Yi-Chun Wu1, Chia-Hao Chang1, and Fu-Nien Wang1
1Biomedical Engineering and Environmental Sciences, National Tsing Hua University, Hsinchu, Taiwan

In this work, we take a closer look at the different integrated parts of CT, such as the anterior and posterior part of CT peak, direct summation of first pass part, the fitted gamma-variate curve (conventional CBV map), recirculation part, and finally the whole curve. Since the extravasation supposed to be minor in the short duration of DSC imaging, we confirmed the feasibility of integrating the total area under the CT curve including both the first pass and recirculation parts could further enrich the data points and to enhance the reliability of measured CBV.

Wednesday May 11th
  13:30 - 15:30 Computer 76

13:30 3923.   Altered Hemodynamics of Cortical Lesions in Multiple Sclerosis: a Dynamic Susceptibility Contrast MRI study using a Kernel-Based Deconvolution Algorithm 
Marco Castellaro1, Denis Peruzzo1, Massimiliano Calabrese2, Francesca Rinaldi2, Valentina Bernardi2, Alice Favaretto2, Irene Mattisi2, Paolo Gallo2, and Alessandra Bertoldo1
1Department of Information Engineering, University of Padova, Padova, Italy, 2Multiple Sclerosis Centre, Department of Neuroscience, University of Padova, Padova, Italy

In this Dynamic Susceptibility Contrast (DSC-MRI) study we investigated the hemodynamic changes that may occur in Cortical Lesions (CLs) in 20 patients with Relapsing Remitting multiple sclerosis. 91% of the CLs show a statistical significant decreased CBF (-53±21%, p<0.001) when compared to the Normal Appearing Grey Matter CBF values. The remaining 9% of the CLs show a statistical significant higher CBF (+39±30% p < 0.05). This confirms neuropathological observations showing a proportion of active CLs characterized by a high degree of inflammated cells. On the other hand, the ipoperfused CLs may suggests a neuronal damage and loss.

14:00 3924.   Tissue Similarity Map of Perfusion Weighted MR Imaging in the Study of Multiple Sclerosis 
E. M. Haacke1, Meng Li1, and Flavia Juvvigunta1
1Department of Radiology, Wayne State University, Detroit, Michigan, United States

Multiple Sclerosis is a chronic inflammatory demyelinating disease of the central nervous system. The cerebral perfusion in MS is reduced with the function of severity of the disease. In this study we use a new approach to processing PWI data that we call Tissue Similarity Maps. It can identify those lesions linked by the same vascular response to the contract agent. Our results showed that the TSM of nulling the MS lesions is successful in enhancing specifically MS lesions. It is a very useful means to reveal information about tissues otherwise difficult to see with conventional PWI processing approaches.

14:30 3925.   Evaluation of signal formation in local arterial input function measurements of DSC-MRI 
Egbert J.W. Bleeker1, Andrew G. Webb1, Marianne A.A. van Walderveen2, Mark A. van Buchem1,2, and Matthias J.P. van Osch1
1Radiology, C.J. Gorter Center for high field MRI, Leiden University Medical Center, Leiden, Netherlands, 2Radiology, Leiden University Medical Center, Leiden, Netherlands

Local or regional arterial input function (AIF) measurements aim for voxel specific AIFs from small arteries. These local AIFs are assumed to reflect the true input of the microvasculature much better than global AIFs. However, do the measured local AIFs reflect the true concentration-time curve of small arteries? The current study investigated this question with numerical modeling that simulated partial volume effects in local AIF measurements. In addition, local AIF candidates selected using angiograms were evaluated in vivo. The findings suggest that local AIF measurements do not reflect the true concentration-time curve in small arteries.

15:00 3926.   Comparison of Automatic Localized and Manual Global AIF Perfusion Imaging from DSC MRI by Vascular Territories 
Adam Martin Winchell1,2, Ralf B Loeffler1, Ruitian Song1, Himanshu Bhat3, Michael Hamm3, Alberto Broniscer4, and Claudia M Hillenbrand1
1Radiological Sciences, St. Jude Children's Research Hospital, Memphis, TN, United States, 2Biomedical Engineering, University of Memphis, Memphis, TN, United States,3Siemens Healthcare, Charlestown, MA, United States, 4Oncology, St. Jude Children's Research Hospital, Memphis, TN, United States

The aim of our study was to compare two DSC methods, a manual global and an automatic local AIF approach, and examine their calculated regional perfusion differences as a function of vascular territory.

Electronic Posters : Diffusion & Perfusion - Neuro
Click on to view the abstract pdf and click on to view the video presentation.
Diffusion Acquisition & Pulse Sequences

Monday May 9th
Exhibition Hall  14:00 - 16:00 Computer 77

14:00 3927.   Effect of truncated sampling on estimated fiber directions in q-space Imaging 
Bryce Wilkins1, Namgyun Lee1, and Manbir Singh1
1Radiology and Biomedical Engineering, University of Southern California, Los Angeles, California, United States

The effect of truncated-sphere sampling commonly used in DSI to estimate single fiber orientation errors is simulated using fully-sampled (7x7x7) q-space data (343 samples), truncated-sphere (203 samples), zero-padding to (21x21x21), noise-free and noisy (SNR = 30) scenarios. We show that the error of resolving a single fiber direction depends on the fiber orientation, and the “corners of q-space”, which are not sampled when truncated sampling schemes are used, help reduce error (mean error 3.5deg versus 3.9deg for the noise case and zero-padding used). As truncated-sphere sampling requires 40% less time, its use could justify the small increase in error.

Silvia De Santis1,2, Yaniv Assaf3, Christopher John Evans1, and Derek K Jones1
1CUBRIC, School of psychology, CARDIFF University, United Kingdom, 2Physics department, Sapienza University, Rome, Italy, 3Tel Aviv University, Israel

We have developed a comprehensive optimised CHARMED pipeline, comprising an optimised data acquisition scheme, based on the electrostatic repulsion algorithm, combined with optimised ordering and extended to different b-value shells, and an analysis approach that incorporates a model parsimony. Marked improvements in data quality are demonstrated using a bootstrap approach. Our proposed pipeline clearly improves the data quality and results in better confidence in the estimated parameters such as fibre orientation.

15:00 3929.   Harmonic analysis of spherical sampling in diffusion MRI 
Alessandro Daducci1, Jason McEwen2, Dimitri Van De Ville3,4, Jean-Philippe Thiran1, and Yves Wiaux2,4
1Signal Processing Laboratory (LTS5), École Polytechnique Fédérale de Lausanne (EPFL), Lausanne, Switzerland, 2Institute of Electrical Engineering, École Polytechnique Fédérale de Lausanne (EPFL), Lausanne, Switzerland, 3Institute of Bioengineering, École Polytechnique Fédérale de Lausanne (EPFL), Lausanne, Switzerland, 4Department of Radiology and Medical Informatics, University of Geneva (UniGE), Geneva, Switzerland

Diffusion MRI has become a powerful tool to non-invasively study white-matter integrity in the brain. Recently multi-shell spherical acquisitions have been advocated for mapping the diffusion signal, notably through its ODF, with a lower number of q-space samples, hence providing acceleration. In this context the spherical harmonic (SH) transform has gained a great deal of popularity. This study presents a theoretical framework and numerical simulations aiming to provide initial guidance in designing efficient multi-shell spherical sampling strategies in a model independent approach. It is based on the use of equiangular grids on the sphere, for which exact sampling theorems exist.

15:30 3930.   Effect of using super-resolution technique in slice direction on DTI fiber tractography 
Daniel Güllmar1, Christian Ros1, and Jürgen R Reichenbach1
1Medical Physics Group, Jena University Hospital, Jena, Thuringia, Germany

2-D multislice Diffusion-MRI typically suffers from lower resolution in the slice direction compared to in-plane resolution. Thus, we applied super-resolution technique in slice direction by acquiring overlapping slices (negative slice gap). Based on this dataset we evaluated the effect on DTI fiber tractography. We observe an overall improvement in diffusion property maps (FA, ADC, etc.) as well as in fiber tractography.

Tuesday May 10th
  13:30 - 15:30 Computer 77

13:30 3931.   High-Resolution Diffusion Imaging of the In Vivo Human Hippocampus 
Michael Zeineh1, Samantha Holdsworth1, Stefan Skare1, Scott Atlas1, and Roland Bammer1
1Stanford Univeristy, Stanford, CA, United States

For clinical entities such as Alzheimer’s disease and epilepsy, it would be beneficial to perform diffusion imaging of the medial temporal lobes with higher resolution than typically acquired for a more detailed analysis of hippocampal microstructure. In this study, we have compared 8-channel and 32-channel coil acquisitions at 3T while altering the direction of phase encoding and use of GRAPPA-accelerated echo-planar imaging. We demonstrate that an optimal balance of SNR, distortion, and symmetry is achieved using the 32-channel coil utilizing parallel imaging with phase-encoding in the supero-inferior direction.

14:00 3932.   Comparison of two alternative approaches for diffusion-weighted Readout-Segmented (RS)-EPI 
Samantha J Holdsworth1, Stefan Skare2, Murat Aksoy1, Rafael O'Halloran1, and Roland Bammer1
1Department of Radiology, Stanford University, Palo Alto, CA, United States, 2Clinical Neuroscience, Karolinska Institute, Stockholm, Sweden

Two approaches for Readout-Segmented (RS)-EPI have purported to increase the SNR for DWI. One fills k-space with full readout segments (RS-EPI-X), and the other which minimizes the TE with partial readout segments (RS-EPI-Y). Here we explore the scan efficiency of RS-EPI-Y and RS-EPI-X. Human brain data show that SNR efficiency is about equivalent, except for the case of a higher matrix size combined with twice-refocused DW preparation – whereby RS-EPI-Y has 1.6 times better scan efficiency than RS-EPI-X.

14:30 3933.   Multi Slice Localized Parallel Excitation for Abdominal and Pelvic EPI Applications in Humans 
Denis Kokorin1,2, Martin Haas1, Frederik Testud1, Jürgen Hennig1, and Maxim Zaitsev1
1Medical Physics, University Medical Center Freiburg, Freiburg, Germany, 2International Tomography Center, Novosibirsk, Russian Federation

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

15:00 3934.   High spatial-resolution DTI using 32 channel head coil at human 7 T 
1Vanderbilt University Institute of Imaging Science, Nashville, TN, United States, 2Radiology and Radiological Sciences, Vanderbilt University, Nashville, TN, United States

High spatial-resolution DTI at 7 T has not been commonly implemented using single-shot methods due to technical difficulties in reducing the effects of geometric distortions and/or blurring artifacts and B1-inhomogeneities that cause SNR decreases in some regions. We have previously suggested methods for diffusion-weighted multi-shot SENSE acquisitions with 2-D navigators and image-domain column-by-column reconstructions using a point spread function. With a 32 channel head coil, the B1-inhomogeneity induced regional SNR decrease can be ameliorated, and diffusion-weighted single-shot and multi-shot images show higher SNR compared with 16 channel coil data. Moreover, the suggested approach can delineate tiny white matter fiber structures at sub-millimeter resolution even in regions with relatively high B1-inhomogeneity, where conventional single-shot methods do not perform adequately.

Wednesday May 11th
  13:30 - 15:30 Computer 77

13:30 3935.   MR measurements of anomalous diffusion indices lower case Greek alpha and lower case Greek gamma by means of PGSTE techniques at varying of time and of gradient strength in phantoms 
Marco Palombo1, Andrea Gabrielli2, Silvia De Santis1, and Silvia Capuani1,3
1Physics Department, Sapienza University of Rome, Rome, Italy, 2ISC, CNR, Rome, Italy, 3IPCF UOS Roma, CNR, Rome, Italy

This work introduces a novel method to give more insides on anomalous diffusion processes of water in heterogeneous systems, by means of PGSTE techniques. When a competition between subdiffusive and superdiffusive processes occurs, it is possible to introduce a pseudo-RMSapproximately equaltlower case Greek alpha/lower case Greek gamma for which lower case Greek alpha/lower case Greek gamma grater or smaller than 1 defines superdiffusion and subdiffusion processes respectively. Experimental data obtained in micro-beads water suspension samples, demonstrate that lower case Greek alpha and lower case Greek gamma reflect some additional microstructural information which cannot be obtained using conventional diffusion procedures based on Gaussian diffusion. Specifically, lower case Greek alpha can provide information on the degree of order and/or disorder within the investigated systems.

14:00 3936.   Concatenated Double Wave Vector Diffusion Weighting Experiments 
Martin A. Koch1, and Jürgen Finsterbusch1
1Systems Neuroscience, University Medical Center Hamburg-Eppendorf, Hamburg, Germany

Double wave vector diffusion weighting may help to assess pore size and shape in tissue. It employs two successive diffusion weighting periods with independent gradient direction. With a short delay between the periods, the signal difference between parallel and antiparallel gradient orientation for restricted diffusion depends on the mean pore size. It is often very small when whole body gradients are employed, hampering routine application. Recently, a "concatenated" experiment was proposed where the double diffusion weighting is applied multiple times. It should provide a larger antiparallel-parallel difference. The predicted increase is investigated experimentally in excised pig spinal cord.

14:30 3937.   Human Brain Mapping of Orientationally Invariant Axonal Diameter Using Q-space Diffusion Tensor MRI 
Jun-Cheng Weng1,2
1School of Medical Imaging and Radiological Sciences, Chung Shan Medical University, Taichung, Taiwan, 2Department of Medical Imaging, Chung Shan Medical University Hospital, Taichung, Taiwan

Fundamental relationships between diffusion tensor imaging (DTI) and q-space imaging can be derived which establish conditions when these two complementary MR methods are equivalent. When the 3D displacement distribution is measured by q-space imaging with large displacement and small q vector, the result is similar to 3D Gaussian assumed in DTI. Combing displacement information from q-space imaging and fiber direction from DTI, distribution of axonal diameters and directions could be derived at the same time. Based on the assumption, the study proposed a novel technique, q-space diffusion tensor imaging (qDTI), to map orientationally invariant axonal diameter distribution of human brain. The goal could be achieved with any of two image reconstruction methods described below. One was tensor-based method. The 3D Gaussian displacement distribution could be obtained directly from the displacement tensor. The other was displacement projection method. The fiber directions were first calculated from conventional DTI, and the mean displacement as well as maximum diffusivity of water molecules along specific direction were then obtained with q-space imaging. The effective axonal diameter was defined as the average of several displacements projected to the direction of the fiber cross section. Our results demonstrated that two qDTI methods both produced reasonable distribution of orientationally invariant axonal diameters in human brain.

15:00 3938.   Measurement of axon radii distribution in orientationally unknown tissue using angular double-pulsed gradient spin echo (double-PGSE) NMR 
Wenjin Zhou1, and David Laidlaw1
1Brown University, Providence, RI, United States

We present an analytical diffusion model for measuring axon properties in biological tissue of unknown orientation using low-q angular double-pulsed gradient spin echo (d-PGSE) NMR. The axon properties we aim to recover are axon radii distribution, axon orientation, and axon volume fraction. Our simulation results demonstrate that clinically feasible acquisition protocols are sufficient to accurately recover these axon properties in typical human brain tissue range (1-5µm).

Thursday May 12th
  13:30 - 15:30 Computer 77

13:30 3939.   Diffusion Tensor Imaging with View Angle Tilting Technique for Distortion Correction 
Sinyeob Ahn1, Ki Sueng Choi1, and Xiaoping Hu1
1Biomedical Engineering, Georgia Institute of Technology and Emory University, Atlanta, GA, United States

Distortion in diffusion tensor imaging (DTI) causes inter-subject variance and intensity modulation, which hinders an accurate analysis of diffusion characteristics. View angle tilting (VAT) technique has been used to correct distortion in the readout direction in spin echo imaging. This work describes a method for correcting image distortion along the phase-encode (PE) direction using the VAT technique in twice-refocused DTI. Phase generated by the VAT gradient concurrently applied with the PE gradient offsets an unwanted phase by field inhomogeneity, leading to the correction of image distortion. B0 images and tractography were used to evaluate the performance of the proposed method.

14:00 3940.   Geometric Distortion Correction of DTI using Accelerated PSF Mapping based Reconstruction at 7 Tesla 
Myung-Ho In1, and Oliver Speck1
1Biomedical Magnetic Resonance, Otto-von-Guericke-University, Magdeburg, Germany

Although diffusion tensor imaging (DTI) allows the observation of molecular diffusion in tissues, it severely suffers from distortions due to B0 field inhomogeneity, susceptibility, chemical shift, as well as eddy current effects, especially at ultra high field such as 7T. These artifacts cause geometric and intensity distortions in MR image formation. Moreover, the distortions vary with the diffusion gradient direction. In this study, an acceleration point spread function (PSF) mapping based reconstruction is proposed in order to correct for direction-dependent distortions in diffusion tensor imaging. The results demonstrate that the proposed method can fully correct the direction-dependent distortions with high fidelity.

14:30 3941.   Robustness of echo planar imaging (EPI) distortion correction in diffusion tensor imaging using forward/reverse phase encode directional b=0 scans 
Wanyong Shin1, Erik B Beall1, Ken Sakaie1, Mingyi Li1, Dominic Holland2, Anders M Dale3, and Mark Lowe1
1Radiology, Imaging Institute, Cleveland Clinic, Cleveland, OH, United States, 2Neuroscience, University of California, San Diego, CA, United States, 3Radiology, University of California, San Diego, CA, United States

Recently, Holland et al. (Neuroimage, 2010) presented a fast, highly computationally efficient unwarping method using forward and reverse phasing encode (PE) directional EPI scans. In this study, we propose a single diffusion tensor imaging (DTI) scan in which a forward / reverse PE direction scan with b = 0 is implemented as a first acquisition. We test the reproducibility and robustness of the proposed unwarping technique from voxel-wise fractional anisotropy comparison using the proposed DTI sequence.

15:00 3942.   Implementation of real time motion correction in Diffusion Tensor Imaging 
ALKATHAFI ALI Alhamud1, Aaron Hess1, Matthew Dylan Tisdall2, Ernesta M. Meintjes1, and Andre J. van der Kouwe2
1University of Cape Town, Cape Town, South Africa, 2Department of Radiology, Harvard Medical School, MA, United States

This paper presents a novel technique to correct, in real time, spatial misalignment of diffusion volumes. A 3D low resolution navigator based on T1 weighted imaging is inserted between the diffusion volumes to compensate for rotation and translation in all 6 degrees of freedom using PACE. The DTI sequence was modified and ran three (Moco, NoMoco and Baseline) on a subject. The results show a substantial improvement in the diffusion tensor parameters with Moco, also the FA histogram shows a shift and a broadening in the FA distribution with NoMoco data, while a significant recovering with Moco was noted.

Electronic Posters : Diffusion & Perfusion - Neuro
Click on to view the abstract pdf and click on to view the video presentation.
Diffusion Applications, Non-Gaussian Diffusion & Diffusion Related Contrasts

Monday May 9th
Exhibition Hall  14:00 - 16:00 Computer 78

14:00 3943.   The drum is visible in nuclear magnetic resonance diffusion experiments 
Frederik Bernd Laun1, Wolfhard Semmler1, and Bram Stieltjes2
1Medical Physics in Radiology, German Cancer Research Center, Heidelberg, Baden-Württemberg, Germany, 2Quantitative Imaging-based Disease Characterization, German Cancer Research Center, Heidelberg, Baden-Württemberg, Germany

While nuclear magnetic resonance diffusion experiments are widely used to resolve the structure of cells and porous media, it has been elusive whether they can reveal the exact shape of boundaries confining the diffusion process. This question is mathematically closely related to Kacs “hear the drum” problem. Although the shape of the drum is not “hearable”, we show that the confining boundary in closed pores can indeed be detected using modified Stejskal-Tanner magnetic field gradients. These modified gradients can act like imaging gradients, and potentially allow novel imaging contrasts.

14:30 3944.   Diffusion Relaxation Correlation Spectroscopy at Ultra Short Echo Times Reveals Two Major Compartments in Human Cadaver Brain White Matter 
Bibek Dhital1, Marcel Gratz2, and Robert Turner1
1Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany, 2Faculty of Physics and Geoscienes, Department of Interface Sciences, University of Leipzig, Leipzig, Germany

Measuring multi-compartmental diffusion in brain tissue and estimating relative fractions of these compartments is complicated due to the influence of relaxation, restriction and exchange. Identifying different compartments with their relative fraction requires a wide range of echo times. Effects of restriction and exchange can be minimized by shortening the diffusion time. We performed Diffusion Relaxation Correlation Spectroscopy (DRCOSY) measurements with very short diffusion time (Δ = 1.2 ms) on excised human corpus callosum. Data were analyzed using a two dimensional inverse Laplace transform. Our results show that diffusion is multi-compartmental and that the fast diffusing compartment has a faster relaxation.

15:00 3945.   Renormalization Group Method: Effects of Diffusion Retarding on Intracellular Membranes 
Oleg Posnansky*1, Yuliya Kupriyanova1, and N. Jon Shah1,2
1Medical Imaging Physics, Institute of Neurosciences and Medicine - 4, Forschungszentrum Jülich GmbH, 52425 Jülich, Germany, 2Department of Neurology, Faculty of Medicine, JARA, RWTH Aachen University, Aachen, Germany

  Proton self-diffusion demonstrates superb sensitivity to the biological properties of living tissue. A complex interplay between different compartments and structures poses different amounts of hindrance to diffusing spins. The presence of restricting cellular structures, and variation in the integrity of the underlying neural tissue lead to non-Gaussian dynamics of the signal. Retarding effects on axon membranes were included in renormalization-group method. Microparameters of diffusivity in neurons were investigated. It is shown that the presence of axon membranes leads to non-Gaussian diffusion. A physical model determining intracellular properties from diffusion-weighted imaging is described.

15:30 3946.   Efficient numerical solution of the Bloch-Torrey equation for modeling multiple compartment diffusion 
Jing Rebecca Li1, Donna Calhoun2, Chun-Hung Yeh3, Cyril Poupon4, and Denis Le Bihan4
1INRIA-Saclay, Palaiseau Cedex, France, 2CEA, Saclay, France, 3National Yang-Ming University, Taiwan, 4CEA Neurospin, Saclay, France

We propose a numerical method for solving the Bloch-Torrey partial differential equation to compute the bulk magnetization of a sample under the influence of a diffusion gradient. We couple a mass-conserving finite element discretization in space with a stable time discretization using an explicit Runge-Kutta-Chebyshev method [1] . We are able to solve the Bloch-Torrey PDE in multiple compartments rapidly and accurately, making it a reasonable candidate as the forward solver in the inner iterative loop of an inverse problem solver going from signals to biological parameters.

Tuesday May 10th
  13:30 - 15:30 Computer 78

13:30 3947.   Constrained maximum likelihood estimator for more accurate diffusion kurtosis tensor estimates 
Jelle Veraart1, Wim Van Hecke2,3, Dirk H. J. Poot4, and Jan Sijbers1
1Vision lab, University of Antwerp, Antwerp, Antwerp, Belgium, 2Dept. of Radiology, University Hospitals of the Catholic University of Leuven, Leuven, Belgium, 3Dept. of Radiology, University Hospital Antwerp, Antwerp, Antwerp, Belgium, 4Biomedical Imaging Group Rotterdam, Erasmus MC, Rotterdam, Netherlands

A computational framework to obtain an accurate quantification of the Gaussian and non-Gaussian component of water molecules’ diffusion through brain tissues with diffusion kurtosis imaging (DKI) is presented. The DKI model quantifies the kurtosis on a direction-dependent basis, constituting a higher order diffusion kurtosis tensor, which is estimated in addition to the diffusion tensor. To reconcile with the physical phenomenon of molecular diffusion, both tensor estimates should lie within a physically acceptable range. We therefore suggest to estimate both diffusional tensors by maximizing the joint likelihood function of all Rician distributed diffusion weighted images while imposing a set of constraints.

14:00 3948.   Characterization of neural tissues in humans using diffusion kurtosis imaging 
Wenshu Qian1, Zhongping Zhang1, Ed Xuekui Wu2, Matthew M. Cheung2, Queenie Chan1,3, Pek-Lan Khong1, and Mina Kim1
1Diagnostic Radiology, The University of Hong Kong, Hong Kong, Hong Kong, China, People's Republic of, 2Electrical and Electronic Engineering, The University of Hong Kong, Hong Kong, Hong Kong, China, People's Republic of, 3Philips Healthcare, Hong Kong, Hong Kong, China, People's Republic of

Diffusion kurtosis imaging (DKI) was recently proposed to probe non-Gaussian diffusion property. The recent study demonstrated that various kurtosis estimates can reveal information different from diffusivity estimates by diffusion tensor imaging (DTI). However, it has never been reported whether directional kurtosis analysis in addition to DTI estimates improves tissue characterization in various human brain tissues. We aimed to characterize human neural tissues using kurtosis estimates and to compare them with conventional diffusivity estimates. Our results demonstrate that kurtosis estimates may help to better characterize neural tissues and detect their microstructural alterations complementary to the conventional diffusivity estimates.

14:30 3949.   Apparent Kurtosis in the Motional Narrowing Regime: Analytic Results for Closed Domains 
Frederik Bernd Laun1, Wolfhard Semmler1, and Bram Stieltjes2
1Medical Physics in Radiology, German Cancer Research Center, Heidelberg, Baden-Württemberg, Germany, 2Quantitative Imaging-based Disease Characterization, German Cancer Research Center, Heidelberg, Baden-Württemberg, Germany

We study the link between apparent kurtosis and eigensystem of the confining domain. It is shown that the fourth cumulant of the phase distribution for closed domains in the motional narrowing regime is properly described by a series expansion in inverse powers of the diffusion time. The coefficients are proportional to the product of two constants. One of them is solely determined by the eigensystem, while the other is solely determined by the temporal profile of the diffusion gradients. In conclusion, a clear relation between the apparent kurtosis and the eigensystem is established.

15:00 3950.   Estimation of the Axonal Density Using DKI: a Validation Study 
Els Fieremans1, Jens H Jensen1, Ali Tabesh1, and Joseph A Helpern1,2
1Center of Biomedical Imaging, Department of Radiology, New York University School of Medicine, New York, NY, United States, 2Center for Advanced Brain Imaging, Nathan S. Kline Institute, Orangeburg, NY, United States

Diffusional kurtosis imaging (DKI) has been proposed as a clinically feasible extension of DTI to probe restricted water diffusion in biological tissues. We exploit here the DKI information to estimate the axonal water fraction (AWF), a measure of axonal density. We validate its measurement against the fraction as obtained from bi-exponential fitting in healthy controls, showing a very good agreement. Whereas other advanced diffusion models typically require high b-values, we propose a DKI WM model based on b-values up to 2000 s/mm2. Assessing the axonal density using DKI is clinically feasible and could potentially provide important information on neurodegenerative disorders.

Wednesday May 11th
  13:30 - 15:30 Computer 78

13:30 3951.   Electrically Active In-Vitro Spinal Cords for the Study of Functional Diffusion Weighted Imaging 
Nitzan Tirosh1, and Uri Nevo1
1Biomedical Engineering, Tel Aviv University, Tel Aviv, Israel

DWI is explored in the last years as a tool for imaging brain activity. However the actual interaction between brain function and water is not yet understood and intensive investigation is therefore needed. We suggest the use of isolated and perfused spinal cords of newborn rats as a novel model for the study of the response of vital, mammalian neuronal tissues in the MRI simultaneously to electrical activity recording. This model enables to acquire high resolution and high SNR images and good ADC maps while extracting out multiple physiological factors that are not related to the neuronal mechanisms under question.

14:00 3952.   Brain tissue water comes in 2 pools: Evidence from diffusion and R2 measurements with USPIOs in non human primates 
Denis Le Bihan1,2, Olivier Joly3, Toshihiko Aso2, Lynn Uhrig3, Cyril Poupon1, Naoki Tani3, H Iwamuro3, Shin-ichi Urayama2, and Bechir Jarraya3
1I²BM, NeuroSpin, Gif-sur-Yvette, France, 2HBRC, Kyoto University, Kyoto, Japan, 3NeuroSpin, INSERM-AVENIR unit, Gif-sur-Yvette, France

Water diffusion in biological tissues is not free, as the signal attenuation is not monoexponential with diffusion-weigthing (b value). Some groups have successfully characterize this attenuation with a biexponential model, which suggests the presence of 2 water pools (fast and slow) in slow or intermediate exchange. However, this model is still controversial and the nature of the 2 remains elusive. Here we show that this 2 pools model also explains the T2 behavior of USPIO contrast agents in the non-human primate brain. This mechanism which also applies to blood deoxyhemoglobin may have implications for the interpretation of BOLD experiments. Finally, the possibility to target USPIO effects to a specific water pool using diffusion-weighting might offer new contrast avenues, especially for molecular imaging.

14:30 3953.   Magnetic susceptibility local variations affect γ-weighted maps contrast in brain 
Silvia De Santis1,2, Andrea Gabrielli3, Emiliano Macaluso4, Marco Bozzali4, and Silvia Capuani2,5
1CUBRIC, School of Psychology, CARDIFF, South Glamorgan, United Kingdom, 2Physics department, Sapienza University, Rome, Italy, 3via dei Taurini 19, ISC-CNR, Rome, Italy, 4Neuroimaging Laboratory Santa Lucia Foundation, Rome, Italy, 5IPCF UOS Roma, Sapienza University, Rome, Italy

The goal of the work was to investigate the correlation between the stretching exponential parameter, introduce to account for anomalous diffusion in brain tissue, and the magnetic susceptibility difference at tissue interface. A significant correlation was found between the two measurements, suggesting that local variations in magnetic susceptibility affect the diffusive motion. These results offer an innovative interpretation of water diffusion decay, and suggest that the peculiar contrast obtained using the stretched exponential model can be further explored and correlated to specific local microstructures.

15:00 3954.   Susceptibility-induced increase in apparent diffusion coefficient 
Dmitry S Novikov1, and Valerij G Kiselev2
1Radiology, NYU School of Medicine, New York, NY, United States, 2Diagnostic Radiology, Uniklinikum Freiburg, Freiburg, Germany

NMR diffusion measurements are confounded by microscopic tissue heterogeneity. A known phenomenon is the underestimation of apparent diffusion coefficient (ADC) in the presence of susceptibility-induced microscopic magnetic field when the diffusion length is shorter than the field correlation length. This study demonstrates for the first time that the effect is opposite in the case of long diffusion length: The ADC overestimates the genuine diffusion coefficient. This effect is anomalously strong for effectively two-dimensional disordered objects, such as blood capillaries, which agrees with experimental data on hypercapnia and diffusion fMRI. Conversely, heterogeneous transverse relaxation rate results in a decrease of ADC.

Thursday May 12th
  13:30 - 15:30 Computer 78

13:30 3955.   Gene therapy evaluated using in vivo diffusion tensor imaging 
Joong Hee Kim1, Adarsh S. Reddy2, Mark S. Sands2, and Sheng-Kwei Song1
1Radiology, Washington University, St. Louis, Missouri, United States, 2Internal Medicine, Washington University, St. Louis, Missouri, United States

Globoid-cell Leukodystrophy (GLD) is an inherited demyelinating disease caused by a deficiency of the lysosomal enzyme Galactosylceramidase (GALC). A previous study demonstrated dramatic synergy between CNS-directed AAV2/5 gene therapy and myeloreductive bone marrow transplantation (BMT) in the murine model of GLD (twitcher). In the present study, in vivo DTI was employed to evaluate the effect of gene therapy of twitcher mice demonstrating the treatment induced recovery of white matter integrity.

14:00 3956.   Quantitative DTI of White Matter Abnormalities upon Early Postnatal Visual Impairments 
Kevin C. Chan1,2, Joe S Cheng1,2, Shu Juan Fan1,2, Matthew M Cheung1,2, and Ed X. Wu1,2
1Laboratory of Biomedical Imaging and Signal Processing, The University of Hong Kong, Pokfulam, Hong Kong, China, People's Republic of, 2Department of Electrical and Electronic Engineering, The University of Hong Kong, Pokfulam, Hong Kong, China, People's Republic of

This study examines in vivo the effect of 4 types of early postnatal visual impairments [binocular enucleation (BE), monocular enucleation (ME), monocular deprivation (MD), and dark-rearing (DR)] on the development and plasticity of the rat visual pathways using diffusion tensor imaging. Compared to age-matched control rats at 6 weeks old, results indicated a significantly lower fractional anisotropy (FA) in the visual pathways projected from both eyes in the BE rats, and from the enucleated right eye in the ME rats. A small but significantly lower FA could also be observed in the left posterior optic tract of MD rats monocularly deprived in the right eye, whereby a significantly higher FA was found in the left prechiasmatic optic nerve in the monocularly enucleated rats. No apparent difference was observed in DR rats along either visual pathway. The results of this study documented in vivo the varying degrees of microstructural alterations along the visual pathways in the 4 rat models. Future DTI studies are envisioned that measure the development and reorganization of the impaired visual pathways after early interventions in longitudinal studies.

14:30 3957.   Neuroregenerative effect of Mesenchymal Stem Cell following Hypoxia-Ischemia in the pup mouse brain assessed by Diffusion Tensor Imaging 
Yohan van de Looij1,2, Cindy T van Velthoven3, Rolf Gruetter2,4, Petra S Hüppi1, Annemieke Kavelaars3, Cobi J Heijnen3, and Stéphane V Sizonenko1
1Division of Child Growth & Development, University of Geneva, Geneva, Switzerland, 2Laboratory for Functional and Metabolic Imaging, Ecole Polytechnique Fédérale de Lausanne, Lausanne, Switzerland, 3Lab. for Neuroimmunology and Developmental Origins of Disease, University Medical Center Utrecht, Utrecht, Netherlands,4Department of Radiology, Universities of Geneva and Lausanne, Geneva and Lausanne, Switzerland

Premature infants are at risk of white matter injury and altered development resulting in a chronic disturbance of myelination. Cerebral Hypoxia-Ischemia/Reperfusion in the premature infant represents a major cause for perinatal white matter injury. Recently, it has been shown that Mesenchymal Stem Cell (MSC) treatment after neonatal HI had neuroregenerative effects. The goal of this study was to assess the neuroregenerative effect of MSC treatment in neonatal mouse brain following HI by DTI derived parameters. This study confirmed white matter damages following neonatal HI on mouse brain as well as neuroregenerative effect of MSC on HI induced white matter damages.

15:00 3958.   Can Diffusion Kurtosis Imaging Provide Better Ischemic Lesion Delineation? 
Edward S. Hui1, Fang Du1, Qiang Shen1, Shiliang Huang1, and Timothy Q. Duong1
1Research Imaging Institute, University of Texas Health Science Center San Antonio, San Antonio, Texas, United States

Conventionally, lesion core is defined as the region with reduced apparent diffusion coefficient. It has been reported that diffusion-weighted images (DWI) acquired with high b-value allowed better delineation of ischemic lesion. In other words, advanced diffusion techniques that make use of DWI with high b-value might potentially provides more reliable ischemic tissue staging. One of which is diffusion kurtosis imaging (DKI) model which measures non-Gaussianity of water diffusion. DKI was thus performed in an established transient middle cerebral artery occlusion (MCAO) rat model during the hyperacute, acute and subacute phases. Comparisons were made with perfusion, conventional diffusion techniques.