New Insights & Innovations in Cardiovascular MRI

Thursday 4 June 2015

Sébastien Roujol¹, Cory Tschabrunn¹, Tamer A. Basha¹, Kraig V. Kissinger¹, Warren J. Manning^1,2, Mark E. Josephson¹, Elad Anter¹, and Reza Nezafat^1
1Department of Medicine, Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, Massachusetts, United States, ²Department of Radiology, Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, Massachusetts, United States

A variety of animal models have been used in cardiac MRI (CMR) studies to help characterize the underlying mechanisms of several cardiomyopathies1. In-vivo and ex-vivo CMR often provide complementary information. In-vivo CMR can be used for hemodynamic, functional and viability evaluation while ex-vivo CMR can provide structural information with high resolution due to the absence of physiological motion. The combination of this information is however challenging since the excised heart generally shows substantial variations in shape, when compared to the same in-vivo heart. In this study, we sought to investigate the feasibility of integrating a 3D printed model of the in-vivo LV cavity into the ex-vivo heart to enable imaging of similar LV shape during in-vivo and ex-vivo CMR.

Irvin Teh¹, Penny L Hubbard Cristinacce^2,3, Feng-Lei Zhou^2,4, Geoffrey JM Parker^2,3, and Jürgen E Schneider^1
1Division of Cardiovascular Medicine, Radcliffe Department of Medicine, University of Oxford, Oxford, United Kingdom, ²Centre for Imaging Sciences, Manchester Academic Health Sciences Centre, The University of Manchester, Manchester, United Kingdom, ³Biomedical Imaging Institute, The University of Manchester, Manchester, United Kingdom, ⁴The School of Materials, The University of Manchester, Manchester, United Kingdom

There is increasing demand for phantoms to aid methods development in diffusion MRI, and to improve quality assurance of data. With the majority of phantoms designed to simulate brain tissue, we have designed a novel phantom, based on co-electrospinning, to simulate diffusion in the heart. The phantom was scanned with DTI over four weeks, and the average mean ADC and FA of 0.76 ± 0.02 x 10^-3 mm²/s and 0.38 ± 0.02 respectively, coincides with published ex-vivo preclinical data. Fibre tracking illustrates clearly the transition in helix angle, from -59 to 53 degrees, from the simulated subepicardium to subendocardium.

Cyonna Holmes¹, Sarah Brant¹, LaShondra Powell¹, Michael Erik Jessen¹, and Matthias Peltz^1
1Cardiovascular and Thoracic Surgery, University of Texas Southwestern Medical Center, Dallas, Texas, United States

Coronary artery disease remains the leading case of mortality in the United States. The diabetes drug metformin has been shown to reduce cardiovascular morbidity and mortality in population based studies. The impact of metformin on myocardial metabolism has not been well characterized. In this study, we investigated the role of metformin on myocardial metabolism by 13C MRS. We demonstrated that metformin reversed insulin associated alterations in myocardial substrate selection.

Robin Navest¹, Cornelis van den Berg¹, Alexander Raaijmakers¹, Peter Luijten¹, Jan Lagendijk¹, and Anna Andreychenko^1
1Imaging Division, UMC Utrecht, Utrecht, Netherlands

The standard sensor in MRI to detect cardiac activity is ECG, but this becomes unstable at higher magnetic field strength. Therefore it is investigated whether thermal noise variance is sensitive to cardiac activity.

Sébastien Roujol¹, Tamer A. Basha¹, Cory Tschabrunn¹, Kraig V. Kissinger¹, Warren J. Manning^1,2, Mark E. Josephson¹, Elad Anter¹, and Reza Nezafat^1
1Department of Medicine, Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, Massachusetts, United States, ²Department of Radiology, Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, Massachusetts, United States

ICD therapy is the first line therapy for the prevention of sudden cardiac death. Reduced incidence of ICD therapy can be achieved when combined with catheter based ventricular tachycardia (VT) ablation in patient with history of myocardial infarction. However, the recurrence rate of VT/VF after VT ablation is about ~50% in these patients and better technique for VT substrate identification are needed. Late gadolinium enhancement (LGE) CMR has the potential for non-invasive assessment of the VT substrate, which consists of reentry circuits near or within a chronic scar. In this study, we sought to utilize a high-resolution 3D LGE sequence with 1 mm3 isotropic spatial resolution to image a surrogate of the VT substrate in swine model of VT.

Gavin D Merrifield¹, Lindsay Gallagher¹, James Mullin¹, Carl S Tucker², Maurits A Jansen^2,3, William M Holmes¹, and Martin A Denvir^2
1Glasgow Experimental MRI Centre, University of Glasgow, Glasgow, Glasgow, United Kingdom, ²University of Edinburgh/British Heart Foundation Centre for Cardiovascular Science, University of Edinburgh, Edinburgh, Midlothian, United Kingdom, ³Edinburgh Preclinical Imaging, University of Edinburgh, Edinburgh, Midlothian, United Kingdom

As an emerging biomedical model the development of Magnetic Resonance Imaging techniques for use in the zebrafish would prove particularly useful to many streams of in vivo research. However the exceptionally small size of this species compared to rodent models and it's unique aquatic environment provide several practical challenges to be overcome. We present a custom MRI compatible flow cell equipped with an integrated solenoid coil for successful imaging that meet those challenges.

Victoria X Wang¹, Cheuk Tang², Maryann McLaughlin³, Edmund Wong¹, Johnny C Ng¹, Lazar Fleysher¹, Fayad A Zahi⁴, Maceda Cynara⁵, Heather N Beebe⁶, and Joseph Friedman^7
1Radiology, Mount Sinai School of Medicine, New York, NY, United States, ²Radiology & Psychiatry, Mount Sinai School of Medicine, New York, NY, United States, ³Cardiology & Medicine, Mount Sinai School of Medicine, New York, NY, United States, ⁴Radiology & Cardiology, Mount Sinai School of Medicine, New York, NY, United States, ⁵Cardiology, Mount Sinai School of Medicine, New York, NY, United States, ⁶Mount Sinai School of Medicine, New York, NY, United States, ⁷Psychiatry, Mount Sinai School of Medicine, New York, NY, United States

Early cognitive impairments associated with CV disease risk factors are subclinical, and are infrequently identified during routine medical care. Such mild cognitive impairments, even if not immediately obvious to the physician, may impact preventive care, medication adherence, clinical follow-up, and quality of life. Cognitive impairment in individuals at risk for vascular pathology, but who have not suffered clinical stroke, suggests the presence of a cognitive prodrome for vascular neurodegenerative disease. This cognitive impairment likely reflects subclinical cerebral disease. In this study we aim to investigate the severity of cognitive changes using fMRI and DTI in otherwise asymptomatic persons (no end organ damage) carrying CV disease risk factors such as hypertension, diabetes mellitus, hyperlipidemia, elevated BMI and smoking.

Sébastien Roujol¹, Tamer A. Basha¹, Cory Tschabrunn¹, Kraig V. Kissinger¹, Warren J. Manning^1,2, Mark E. Josephson¹, Elad Anter¹, and Reza Nezafat^1
1Department of Medicine, Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, Massachusetts, United States, ²Department of Radiology, Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, Massachusetts, United States

ICD therapy is the first line therapy for the prevention of sudden cardiac death. Reduced incidence of ICD therapy can be achieved when combined with catheter based ventricular tachycardia (VT) ablation in patient with history of myocardial infarction. However, the recurrence rate of VT/VF after VT ablation is about ~50% in these patients and better technique for VT substrate identification are needed. Late gadolinium enhancement (LGE) CMR has the potential for non-invasive assessment of the VT substrate, which consists of reentry circuits near or within a chronic scar. In this study, we sought to utilize a high-resolution 3D LGE sequence with 1 mm3 isotropic spatial resolution to image a surrogate of the VT substrate in swine model of VT.

Zhijun Li¹, Matthew Muller², Jianli Wang¹, Christopher Sica¹, Liang Han¹, Prasanna Karunanayaka¹, Jeffrey Vesek¹, Qing X. Yang^1,3, and Lawrence Sinoway^2
1Center for NMR Research, Department of Radiology, College of Medicine, The Pennsylvania State University, Hershey, Pennsylvania, United States, ²Heart and Vascular Institute, College of Medicine, The Pennsylvania State University, Pennsylvania, United States, ³3Department of Neurosurgery, College of Medicine, The Pennsylvania State University, Pennsylvania, United States

Unlike in human brain imaging, normalization to a common template during exercising is a difficult proposition in muscle-imaging studies. Here we propose a novel approach to use single subject Independent Component Analysis (ICA) to simultaneously identify and characterize the lower leg muscle behavior during rhythmic plantar-flexion. Since ICA requires no prior knowledge of the muscle hemodynamics, it is ideally suited for studies in which exploring muscle physiology (dynamics) is the main focus. ICA results showed good reproducibility in automatic segmentation of the calf muscle and highlighted a possible causal connection between muscle subgroups during exercising.

Simone Coppo¹, Jean Delacoste¹, Gabriele Bonanno¹, Davide Piccini^1,2, and Matthias Stuber^1
1Department of Radiology, University Hospital (CHUV), University of Lausanne (UNIL), Center for Biomedical Imaging (CIBM), Lausanne, Switzerland, ²Advanced Clinical Imaging Technology, Siemens Healthcare IM BM PI, Lausanne, Switzerland

Ultrashort echo time (UTE) has recently been demonstrated to have potential for the detection of plaque calcification ex-vivo. For this reason, a self-navigated 3D radial, ECG-triggered, whole-heart imaging sequence was adapted to perform free-breathing UTE acquisitions. This pulse sequence aims at taking advantage of the efficiency and ease of use of self-navigation techniques to enable UTE coronary imaging in vivo. A first in vivo human feasibility study was performed and initial results are discussed.

Christopher Nguyen¹, Zhaoyang Fan¹, Xiaoming Bi², and Debiao Li^1
1Biomedical Imaging Research Institute, Cedars Sinai Medical Center, Los Angeles, CA, United States, ²Siemens Healthcare, Los Angeles, CA, United States

A novel cardiac diffusion tensor imaging sequence is proposed that utilizes a widely-available clinical MRI system. Different levels of motion compensation are tested with second-order gradient moment nulling resulting with the least amount of signal loss. B1 inhomogeneity still persists as a challenge to overcome but preliminary results are promising.

Kévin Moulin^1,2, Pierre Croisille^1,3, Thorsten Feiweier⁴, Benedicte M.A. Delattre¹, Hongjiang Wei¹, Benjamin Robert², Olivier Beuf¹, and Magalie Viallon^1,3
1CREATIS; CNRS (UMR 5220); INSERM (U1044); INSA Lyon; Université de Lyon, Lyon, France, ²Siemens Healthcare France, Saint-Denis, France, ³Department of Radiology, Centre Hospitalier Universitaire de Saint- Etienne, Université Jean-Monnet, France, ⁴Healthcare, Siemens AG, Erlangen, Germany

In-vivo cardiac diffusion using either the Intra-Voxel Incoherent Motion (IVIM) model or the Diffusion Tensor Imaging (DTI) model shows promise to provide new insights into heart pathologies. However, due to the combined challenge of respiratory and heart motion, currently proposed acquisition methods are quite demanding for unhealthy patients. In this study, we propose to use the motion information provided by a navigator to prospectively update in real time the position of the diffusion-weighted slices in order to offer an efficient free-breathing strategy for rapid and improved cardiac diffusion acquisition using a single-shot Spin-Echo EPI sequence (SE-EPI).

Maelene Lohezic¹, Remi Peyronnet², Craig A. Lygate³, Debra McAndrew³, Irvin Teh¹, Peter Kohl^2,4, and Jurgen E. Schneider^1
1BMRU, Division of Cardiovascular Medicine, Radcliffe Department of Medicine, University of Oxford, Oxford, United Kingdom, ²National Heart and Lung Institute, Imperial College London, London, United Kingdom, ³Division of Cardiovascular Medicine, Radcliffe Department of Medicine, University of Oxford, Oxford, United Kingdom, ⁴Department of Computer Science, University of Oxford, Oxford, United Kingdom

Myocardial microstructure is closely related to cardiac function in health and disease, and can be assessed by diffusion tensor imaging (DTI) throughout the cardiac cycle in a non-destructive manner. We investigated changes in DTI parameters in live isolated chronically-infarcted rat hearts in three mechanical states. Significant increase in the mean apparent diffusion coefficient was observed in infarcted hearts in the contractured state, while significant decrease in fractional anisotropy was observed in infarcted hearts in all three states. Furthermore, differences in primary myocyte orientations were observed between healthy hearts and the remote zone of infarcted hearts in the relaxed state.

Zhijun Li¹, Prasanna Karunanayaka¹, Matthew Muller², Lawrence Sinoway², and Qing X. Yang^1,3
1Center for NMR Research, Department of Radiology, College of Medicine, The Pennsylvania State University, Hershey, Pennsylvania, United States, ²Heart and Vascular Institute, College of Medicine, The Pennsylvania State University, Pennsylvania, United States, ³Department of Neurosurgery, College of Medicine, The Pennsylvania State University, Pennsylvania, United States

Unlike in human brain imaging, normalization to a common template during exercising is a difficult proposition in muscle-imaging studies. Here we propose a novel approach to use single subject Independent Component Analysis (ICA) to simultaneously identify and characterize the lower leg muscle behavior during rhythmic plantar-flexion. Since ICA requires no prior knowledge of the muscle hemodynamics, it is ideally suited for studies in which exploring muscle physiology (dynamics) is the main focus. ICA results showed good reproducibility in automatic segmentation of the calf muscle and highlighted a possible causal connection between muscle subgroups during exercising.

Jyh-Miin Lin¹, Chengcheng Zhu², Hsiao-Wen Chung³, Martin Graves⁴, and Andrew Patterson^4
1Department of Radiolgoy, University of Cambridge, Cambridge, Cambridgeshire, United Kingdom, ²Department of Radiology, UCSF School of Medicine, San Francisco, California, United States, ³Department of Electrical Engineering, National Taiwan University, Taipei, Taiwan, ⁴Department of Radiology, Cambridge University Hospitals NHS Foundation Trust, Cambridge, United Kingdom

The application of cine fast-spin echo (FSE) in measuring abdominal aortic wall displacements is hampered by respiratory motions and flow turbulences. Free-breathing black-blood (FBBB) cine FSE is proposed to locally select the abdominal aorta without the respiratory artifacts from adjacent tissues. Flow artifacts are inhibited using the delay alternating with nutation for tailored excitation flow suppression pulse (DANTE). The transverse images and M-mode confirmed that respiratory triggering or navigation is not needed for the FBBB cine FSE. Clinical indexes can be derived from aortic wall displacements using the FBBB cine FSE.

Pierre-André Vuissoz^1,2, Christophe Meyer^1,2, Jacques Felblinger^3,4, and Laurent Bonnemains^1,2
1Imagerie Adaptative Diagnostique et Interventionnelle, Université de Lorraine, Nancy, France, ²U947, INSERM, Nancy, France, ³CIC-IT 1433, INSERM, Nancy, France,⁴University Hospital Nancy, Nancy, France

Due to heart rate variability, Cardiac Magnetic Resonance requires the determination of a cardiac phase model to reconstruct trigged or cine images. Calibration of such a model can be personalized on a patient basis using a 1D+t phase contrast real time acquisition. Benefit of an automatic calibration in term of temporal precision of the predicted systole phase has been tested on twenty-three healthy volunteers. Personalized adaptive cardiac model predicted end-systolic time better than global model on the whole population. Personalization of subject-specific cardiac model is feasible automatically in MRI and reduces the prediction error of systole.

Manivannan Jayapalan¹, Bhargav Bhatt², and Vijikumar N^3
1MR PSD & Applications, GE Healthcare, Bangalore, Karnataka, India, ²MR Systems, GE Healthcare, Bangalore, Karnataka, India, ³MR Applications, GE Healthcare, Bangalore, Karnataka, India

Cardiac Magnetic Resonance Imaging (MRI) requires synchronization of electrocardiogram (ECG) signal with the acquisition. As the complete acquisition might not be acquired in one heart cycle, its successive acquisitions have to be accurately combined with the cardiac phase motion. Such requirements depends on a reliable detection of the R-wave of the ECG to guarantee that consecutive image data collections always start at the same point of the cardiac cycle. However the interaction of blood flow with static magnetic field, known as Magnetohydrodynamic (MHD) effect, introduces special kind of artifact in ECG which is known as MHD artifact or flow artifact. The effect of MHD is directly proportional the field strength and complicates the detection of R wave peak from ECG acquired during MR acquisition especially at higher field strengths. This work presents a method to remove noises and to identify R wave peaks which is contaminated with MHD artifact using multilevel wavelet decomposition and multi-level thresholding.

Cardiovascular MR - Tissue Characterization

Thursday 4 June 2015

Yasuo Amano¹, Kumiko Mine¹, Fumi Yamada¹, and Shinichiro Kumita^1
1Radiology, Nippon Medical School, Tokyo, Tokyo, Japan

We sought to evaluate the distribution and significance of myocardial hyperintensity on T2-weighted imaging (T2-HI) in patients with asymmetrical septal HCM (ASH). On third of T2-HI was outside late gadolinium enhancement, and the presence of T2-HI was significantly related to unexpected syncope. T2-HI had longer T2 values than remote myocardium of ASH and normal myocardium on T2 mapping. T2-weighted MRI should be interpreted carefully to identify T2-HI related syncope in ASH patients.

Sébastien Roujol¹, Tamer A. Basha¹, Jihye Jang¹, Kraig V. Kissinger¹, Beth Goddu¹, Sophie Berg¹, Warren J. Manning^1,2, and Reza Nezafat^1
1Department of Medicine, Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, Massachusetts, United States, ²Department of Radiology, Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, Massachusetts, United States

Myocardial extracellular volume fraction (ECV) mapping requires co-registration of native and post-contrast T1 maps, generally acquired with a b-SSFP readout. However the b-SSFP signal is susceptible to B0 field inhomogeneity, T2 dependent, and sensitive to magnetization transfer. To overcome these limitations, slice interleaved inversion recovery spoiled gradient echo (GRE) imaging has been recently proposed for T1 mapping. However, the co-registration of native and post-contrast GRE T1 scans is challenging since short inversion time (TI) images have very different intensity/contrast between pre/post-contrast imaging and long TI images have low blood/myocardium contrast. In this study, we sought to develop and evaluate a combined T2-magnetization preparation imaging and slice interleaved inversion recovery spoiled gradient echo imaging to facilitate the co-registration process and improve ECV mapping.

Till Huelnhagen¹, Fabian Hezel¹, Andreas Pohlmann¹, Andreas Graessl¹, Jan Rieger², Darius Lysiak², Christof Thalhammer¹, Peter Kellman³, Marcel Prothmann⁴, Jeanette Schulz-Menger^4,5, and Thoralf Niendorf^1,5
1Berlin Ultrahigh Field Facility (B.U.F.F.), Max-Delbrueck Center for Molecular Medicine (MDC), Berlin, Germany, ²MRI.TOOLS GmbH, Berlin, Germany, ³National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, MD, United States, ⁴Dept. of Cardiology and Nephrology, HELIOS Klinikum Berlin-Buch, Berlin, Germany,⁵Experimental and Clinical Research Center, a joint cooperation between the Charite Medical Faculty and the Max-Delbrueck Center, Berlin, Germany

This work examines the capability of T₂^* mapping at 7.0 T to differentiate between normal myocardium and myocardial tissue affected by hypertrophic cardiomyopathy. This is the first patient study report on cardiac MR at 7.0 T. Results show, that high spatial resolution myocardial T₂^* mapping at 7.0 T reveals differences between healthy volunteers and patients with hypertrophic cardiomyopathy. This observation holds the promise to provide means for an MR based biomarker for the risk stratification of HCM.

Sébastien Roujol¹, Tamer A. Basha¹, Jihye Jang¹, Sophie Berg¹, Warren J. Manning^1,2, and Reza Nezafat^1
1Department of Medicine, Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, Massachusetts, United States, ²Department of Radiology, Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, Massachusetts, United States

Native myocardial T1 sequences generally use a balanced steady state free precession (b-SSFP) readout. However, the b-SSFP signal is sensitive to many factors including B0/B1 field inhomogeneities, T2, and magnetization transfer. We recently developed the slice-interleaved T1 (STONE) sequence which we extended to spoiled gradient echo (GRE) imaging. The GRE readout removes the T2 dependence and the magnetization transfer sensitivity of T1 estimates, and reduces the overall sensitivity to B0 field inhomogeneities. Two-point fit model provides improved robustness against artefact and provides higher T1 precision. However, this model is associated with reduced accuracy induced by its sensitivity to imperfect inversion efficiency and signal disturbances caused by the imaging pulses. In this study, we sought to develop and evaluate an alternative fitting approach where the data are fitted to the simulated signal of the entire pulse sequence using Bloch equations.

Julie Magat¹, Benjamin Berte¹, Hubert Cochet¹, Jérôme Naulin¹, Daniele Ghidoli², Pierre Jais¹, Stephen Henry Gilbert³, Olivier Bernus¹, and Bruno Quesson^1
1IHU-LIRYC U1045, University of Bordeaux, Pessac, France, ²Biosense Webster,R&D, Diamond Bar, California, United States, ³Mathematical Cell Physiology, Max Delbrück Center for Molecular Medicine, Berlin, Germany

Catheter ablation became standard practice for the treatment of ventricular tachycardia. However, radiofrequency catheter ablation is limited by lesion depth. The aim of this study was to evaluate the lesion site area on sheep infarct model and characteristics of intramural ablation using a new irrigated needle catheter by MRI. We present results obtained using T₁ mapping/weighted after contrast injection at 1.5T and high resolution images at 9.4T. Within vivo and ex vivo images, we were able to identify infarct area in the septum and lesions due to catheter ablation. Moreover, high field MRI allowed us (resolution of 80µm x 80µm x 235µm) to evaluate precisely ablations zones inside the muscle.

Laurence H Jackson¹, Thomas Roberts¹, Valerie Taylor¹, Josef Habib², Daniel J Stuckey¹, and Mark F Lythgoe^1
1Centre for Advanced Biomedical Imaging, University College London, London, United Kingdom, ²Imaging Sciences and Biomedical Engineering, Perinatal Imaging and Health, Kings College London, London, United Kingdom

In recent years it has become feasible to treat injured myocardium using regenerative cellular strategies. These techniques generally focus on the delivery of treatment cells to the myocardium and rely on their successful engraftment to existing tissue. The work presented here determines a method and the success of tracking bone marrow mononuclear cells suspended in gadolinium doped hydrogel within the myocardium. Correlating the presence of cells to changes to regional strain in the myocardium using tag-cine MRI then allows us to directly assess therapeutic success.

Rheal A. Towner¹, Nataliya Smith¹, Jorge Carrizales¹, Debra Sauners¹, Robert Silasi-Mansat², Florea Lupu², Marilyn Ehrenshaft³, and Ronald P. Mason^3
1Advanced Magnetic Resonance Center, Oklahoma Medical Research Foundation, Oklahoma City, OK, United States, ²Cardiovascular Biology, Oklahoma Medical Research Foundation, Oklahoma City, OK, United States, ³NIEHS, NC, United States

Cardiovascular disease is the primary cause of morbidity and mortality among the diabetic population. One of the characteristics associated with diabetic cardiomyopathy is the generation of free radicals. In this study we incorporated a spin trapping compound, DMPO, to trap free radicals in STZ-induced diabetic mice [2], and used a trapped free radical-targeted molecular MRI probe (anti-DMPO probe) to detect in vivo levels of free radicals in cardiac muscle of diabetic mice. There was a significant increase in the percent change in MRI signal intensity in diabetic mouse hearts (p<0.01) compared to non-diabetic mice, both administered DMPO and the anti-DMPO free radical-targeted probe. The biotin moiety on the anti-DMPO probe was used to confirm its presence in excised tissue with streptavidin-Cy3. In conclusion, a free radical-targeted molecular MRI can be used to detect in vivo heterogeneous levels of free radicals in a mouse model for diabetic cardiomyopathy.

Tamer Basha¹, Sébastien Roujol¹, and Reza Nezafat^1
1Department of Medicine, Beth Israel Deaconess Medical Center & Harvard Medical School, Boston, Massachusetts, United States

T₂ Mapping is an alternative technique to T₂W imaging, where multiple T₂W images are acquired and the voxel-wise T₂ values are estimated through a fitting model. One major limitation is the need for several rest cycles after each image to allow for full spin recovery, which reduces the scan time efficiency of the technique especially if more slice coverage is needed. In this work, we propose to use a slice-selective T₂prep pulse combined with interleaved slice acquisition to eliminate the need for the rest cycles, and provide more slice coverage within efficient scan time.

Esben Søvsø Szocska Hansen^1,2, Steen Fjord Pedersen³, Steen Bønnelykke Pedersen⁴, Uffe Kjærgaard¹, Nikolaj Hjort Schmidt⁵, Hans Erik Bøtker⁶, and Won Yong Kim^1,6
1The MR Research Centre, Aarhus University, Skejby, Aarhus, Denmark, ²Danish Diabetes Academy, Odense, Denmark, ³Dept. of Cardiothoracic and Vascular Surgery T, Aarhus University Hospital Skejby, Skejby, Aarhus N, Denmark, ⁴Dept. of Department of Endocrinology and Internal Medicine, Aarhus University Hospital THG, Skejby, Aarhus, Denmark, ⁵Department of Clinical Medicine - Comparative Medicine Laboratory, Aarhus University, Skejby, Aarhus, Denmark, ⁶Dept. of Cardiology, Aarhus University Hospital Skejby, Skejby, Aarhus, Denmark

Recent studies have used cardiovascular magnetic resonance (CMR) and T2-weighted short tau inversion recovery (T2-STIR) imaging to detect intramyocardial haemorrhage (IMH) as a measure of ischemic/reperfusion injury. We investigated the ability of T2-STIR to differentiate between microvascular obstruction (MVO) and IMH in an experimental pig model of ischemic/reperfusion injury. T2-STIR results were validated by histopathology showing a sensitivity of 100% and specificity of 29% respectively, for detection of IMH. T2-mapping showed no significant difference between T2 values from areas with MVO versus MVO and IMH. In conclusion, T2-STIR was unable to differentiate between IMH and MVO in myocardial ischemic/reperfusion injury.

Neil Chatterjee¹, Octavia Bane², Bruce Spottiswoode³, James Carr⁴, and Timothy Carroll^4
1Biomedical Engineering, Northwestern University, Chicago, IL, United States, ²Mount Sinai, NY, United States, ³Siemens Healthcare, Chicago, IL, United States, ⁴Radiology, Northwestern University, IL, United States

Quantitative measurement of myocardial blood volume (MBV) may have importance as a biomarker for cardiovascular disease, but measurement is complicated by errors introduced by water exchange between the intravascular and extravascular compartments. Here, we use ferumoxytol to measure water exchange rates in the myocardium and calculate quantitative MBV in a cohort of healthy volunteers. Our measured MBV and water exchange frequencies fall within the expected range seen in the literature from animal studies.

Nadja M Meßner¹, Lothar R Schad¹, and Frank G Zöllner^1
1Computer Assisted Clinical Medicine, Medical Faculty Mannheim, Heidelberg University, Mannheim, Germany

This abstract provides information on T1 mapping in the right ventricle (RV) for MR scientists and cardiac radiologists. Certain diseases such as arrhythmogenic right ventricular dysplasia (ARVD) affect predominantly the RV myocardium, however, its T1 quantification has been neglected so far, because the RV myocardium has a very thin and variable structure. Our goal was to show its feasibility at high-resolution. 6 healthy volunteers were investigated with a 3 T Siemens Skyra with a 5(3)3 MOLLI sequence and b-SSFP readout. Images were prospectively ECG triggered at systole instead of diastole, which lead to a larger RV ROI size. A spatial resolution of 1.17x1.17x5 mm³ could be achieved. T1 times in the RV were found to be 1,354±20 ms, LV T1 times were 1,223±37 ms. The significantly higher T1 time of the RV is likely to be explained by the higher collagen content of the RV. In conclusion, RV T1 mapping with the MOLLI sequence provides robust, high-resolution results at 3 T in short acquisition times and could therefore be included in clinical routine in the near future.

Christian Torben Stoeck^1,2, Constantin von Deuster^1,2, Martin Genet¹, David Atkinson³, and Sebastian Kozerke^1,2
1Institute for Biomedical Engineering, University and ETH Zurich, Zurich, Switzerland, ²Imaging Sciences and Biomedical Engineering, King's College London, London, United Kingdom, ³Centre for Medical Imaging, University College London, London, United Kingdom

Spin-echo based diffusion weighted sequences allow for single-shot acquisitions but are highly sensitive to cardiac motion. In this work second order motion compensated diffusion encoding was implemented and compared to first order motion compensation during systolic contraction of the heart. Second order motion compensation was found to yield an extended systolic window suitable for imaging when compared to first order motion compensation (210 vs. 90 ms) thereby enabling spin-echo DTI at various points during the cardiac cycle.

Eva Peper¹, Till Huelnhagen¹, Andreas Pohlmann¹, Min-Chi Ku¹, and Thoralf Niendorf^1,2
1Berlin Ultrahigh Field Facility (B.U.F.F.), Max Delbrück Center for Molecular Medicine, Berlin, Germany, ²Experimental and Clinical Research Center, a joint cooperation between the Charité Medical Faculty and the Max Delbrück Center, Berlin, Germany

This study investigates the tissue contrast for QSM of ex vivo rodent hearts at 9.4 T in comparison to T₂^* mapping. Results show that QSM can provide a higher contrast for myocardial fiber structures than magnitude images and T₂^* maps, without using contrast agents. Susceptibility maps showed good correlation with histological images from literature. This observation provides encouragement that QSM might be a useful tool for assessing myocardial microstructure and could provide further insights into myocardial fiber arrangement and myolaminar structure.

Maythem Saeed¹, Loi Do¹, Roland Krug¹, Steven W Hetts¹, and Mark W Wilson^1
1Radiology and Biomedical Imaging, University of California San Francisco, San Francisco, Ca, United States

The aim of this swine study was to longitudinally quantify microemboli effects on viability of previously ischemic myocardium using MRI (at 3 days and 5 weeks after intervention) and microscopy (5 weeks). Animals were subjected to 45min LAD coronary artery occlusion with and without microemboli. Delayed contrast enhanced MRI has the potential to detect additional loss of cellular integrity in ischemic myocardium exposed to microembolization, but with limited spatial resolution for quantifying true microinfarct compared to microscopy. Cardiac injury biomarkers suffer similar limitation. MRI may be useful in testing the efficacy of newer techniques designed to minimize microemboli during revascularization.

Jihye Jang^1,2, Tamer Basha¹, Sophie Berg¹, Cory Tschabrunn¹, Elad Anter¹, Sébastien Roujol¹, and Reza Nezafat^1
1Department of Medicine, Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, MA, United States, ²Computer Aided Medical Procedures, Technische Universität München, Munich, Bayern, Germany

Myocardial T₁ mapping is performed by pixel-wise curve fitting of single-shot images collected over cardiac cycles. Single-shot imaging often has limited spatial resolution, requires high acceleration factor and prone to cardiac motion that occurs over >200ms acquisition window during cardiac cycle. This results in partial voluming error and reduced measurement precision in T₁ mapping. We recently developed a free-breathing slice-interleaved T₁ (STONE) mapping sequence which removes the breath-holding constrain and allows efficient simultaneous imaging of multiple slices. In this study, we sought to further extend STONE imaging sequence to allow ECG segmented multi-shot data acquisition to improve spatial resolution.

Constantin von Deuster^1,2, Christian T. Stoeck^1,2, Lukas Wissmann², Georg Spinner², Thea Fleischmann^3,4, Maximilian Y. Emmert^4,5, Nikola Cesarovic⁴, and Sebastian Kozerke^1,2
1Division of Imaging Sciences and Biomedical Engineering, King's College London, London, United Kingdom, ²Institute for Biomedical Engineering, University and ETH Zurich, Zurich, Switzerland, ³Institute of Laboratory Animal Science, University of Zurich, Zurich, Switzerland, ⁴Department of Surgical Research, University Hospital Zurich, Zurich, Switzerland, ⁵Swiss Center of Regenerative Medicine, Zurich, Switzerland

Intra-voxel incoherent motion (IVIM) has the potential to measure perfusion without contrast agent administration. Its application to the heart however remains challenging due to cardiac and respiratory motion. In this work IVIM data was acquired with a second order motion compensated diffusion weighted spin echo sequence in a pig model with myocardial infarction. The IVIM theory was validated by pre- and post-mortem comparison in the same animal. Infarcted area identified by a low perfusion fraction in the IVIM model agreed with dynamic contrast-enhanced myocardial perfusion imaging.

Avinash Kali^1,2, Eui-Young Choi³, Behzad Sharif¹, Young Jin Kim³, Xiaoming Bi⁴, Bruce Spottiswoode⁵, Ivan Cokic¹, Hsin-Jung Yang^1,2, Mourad Tighiouart⁶, Debiao Li¹, Daniel S Berman^1,7, Byoung Wook Choi³, Hyuk-Jae Chang³, and Rohan Dharmakumar^1,8
1Biomedical Imaging Research Institute, Cedars-Sinai Medical Center, Los Angeles, California, United States, ²Department of Bioengineering, University of California, Los Angeles, CA, United States, ³Yonsei University College of Medicine, Seoul, Korea, ⁴Siemens Healthcare, Los Angeles, CA, United States, ⁵Siemens Healthcare, Chicago, IL, United States, ⁶Biostatistics and Bioinformatics Research Center, Cedars-Sinai Medical Center, Los Angeles, California, United States, ⁷Cedars-Sinai Heart Institute, Cedars-Sinai Medical Center, Los Angeles, CA, United States, ⁸Department of Medicine, University of California, Los Angeles, CA, United States

We investigated whether native T1 maps at 3T can reliably characterize chronic myocardial infarctions (cMIs) in two pilot patient populations with prior ST- elevation MI (STEMI) and non-ST elevation MI (NSTEMI). Using semi-automatic thresholding analysis (Mean + 5 standard deviations criterion), native T1 maps and LGE images were not different for measuring infarct size and transmurality in both STEMI and NSTEMI patients. With reference to the LGE measurements, threshold analysis showed a strong sensitivity and specificity for detecting cMIs on native T1 maps. Visual detection of cMI on native T1 maps showed high specificity, but modest sensitivity.

Anthony G. Christodoulou¹ and Zhi-Pei Liang^1
1Beckman Institute and Department of Electrical and Computer Engineering, University of Illinois at Urbana-Champaign, Urbana, IL, United States

T1 mapping of the myocardium shows great promise for quantitative characterization of myocardial tissue. The partial separability (PS) model has previously been shown to be effective for reducing the data acquisition requirements of static parameter mapping; here we extend the model to accelerate variable-flip-angle dynamic 3D T1 mapping using a time-varying subspace (or tensor-product subspace). We have successfully applied the proposed method to significantly accelerate variable-flip-angle dynamic 3D T1 mapping for diagnosis of ischemic reperfusion injury in rats.

Jing Liu¹ and David Saloner^1,2
1Radiology and Biomedical Imaging, University of California San Francisco, San Francisco, CA, United States, ²Radiology Service, VA Medical Center, San Francisco, CA, United States

In this study, we aim to achieve highly accelerated free-breathing 3D T1/T2/proton density mapping in the heart, by using incomplete inversion recovery with bSSFP acquisition as well as dictionary match of the acquired data to the evolution curves derived from Block simulations with the same acquisition scheme.

Chia-Ying Liu¹, David A Bluemke¹, Gary Gerstenblith², Stefan L Zimmerman², Ji Li², Hong Zhu², Shenghan Lai², and Hong Lai^2
1Radiology and Imaging Sciences, NIH, Bethesda, MD, United States, ²Johns Hopkins School of Medicine, MD, United States

The objectives were to estimate the reference value of noncontrast T1 time on 3T MR systems and explore the factors that are associated with altered T1 relaxation times in 92 healthy AAs aged 21 years or older, without symptoms or clinical evidence of cardiovascular disease. The median native T1 time of the myocardium was 1228 ms (IQR:1200-1263). Among the factors investigated, only myocardial fat and elevated CRP (CRP>2.0mg/dL) were independently associated with T1 relaxation time.

Yasuo Amano¹, Masaki Tachi¹, Keisuke Inui², Fumi Yamada¹, Makoto Obara³, Shogo Imai¹, and Shinichiro Kumita^1
1Radiology, Nippon Medical School, Tokyo, Tokyo, Japan, ²Cardiology, Nippon Medical School, Tokyo, Tokyo, Japan, ³Philips Asia Pacific, Tokyo, Japan

Extracellular volume fraction (ECV) measurement is expected to reflect the degree of fibrosis in dilated cardiomyopathy. The right ventricular (RV) blood T1 value could be measured in cases of flow artifacts or turbulence in the left ventricular (LV) cavity, but the RV blood T1 and the effect of blood T1 value on ECV remain unknown. The RV blood T1 was shorter than the LV blood T1, but the ECV of myocardium, including scarring, was not affected by the ventricle where blood T1 was measured. The measurement in the RV is allowed to estimate myocardial ECV in some patients.

James W Goldfarb^1,2, Kathleen Gliganic¹, and Nathaniel Reichek^1,2
1Research and Education, Saint Francis Hospital, Roslyn, New York, United States, ²Biomedical Engineering, Stony Brook University, Stony Brook, New York, United States

Blood and myocardial T1, T2 and T2* relaxation times were measured with room air and supplemental oxygen supplied by nasal cannula and a non-rebreather mask using current vendor protocols. Only left ventricular and atrial blood T1s were significantly different from room when oxygen was supplied using a non-rebreather mask. Use of supplemental oxygen can change measured cardiovascular relaxation values and subsequent derived values. If supplemental oxygen is used, one can measure blood relaxation times from the right side of the heart as they are unaffected.

Emeline Lugand^1,2, Jérôme Yerly^1,2, Hélène Feliciano^1,2, Jérôme Chaptinel^1,2, Matthias Stuber^1,2, and Ruud B van Heeswijk^1,2
1Department of Radiology, University Hospital (CHUV) and University of Lausanne (UNIL), Lausanne, Switzerland, ²Center for Biomedical Imaging (CIBM), Lausanne, Switzerland

While cardiac T₂ mapping through radial gradient echo imaging is highly robust to motion, it also results in a relatively low signal-to-noise ratio (SNR) due to the undersampling of the k-space periphery and the resulting density compensation function (DCF), which increases the weight of the k-space periphery. An undersampled k-space-weighted image contrast (KWIC) filter that shares the k-space periphery between images was therefore implemented to improve precision and to shorten the acquisition time. The technique was tested on phantoms and was successfully validated in healthy volunteers.

Tori A Stromp¹, Steve W Leung^2,3, Vincent L Sorrell^2,3, and Moriel H Vandsburger^1,2
1Department of Physiology, University of Kentucky, Lexington, KY, United States, ²Saha Cardiovascular Research Center, University of Kentucky, Lexington, Kentucky, United States, ³Gill Heart Institute, University of Kentucky, Lexington, KY, United States

Patients with chronic kidney disease are at high risk for adverse cardiac events with an emerging link to fibrosis. While such patients would benefit from diagnosis with LGE, they are contraindicated due to safety concerns. We compared two methods of non-contrast cardiac MRI to identify interstitial remodeling in a patient population on routine hemodialysis for CKD. Both native T1 mapping and our previously developed cine bSSFP method (termed 2-point bSSFP) showed greater enhancement patterns in patients with CKD compared to healthy controls, indicating areas of interstital remodeling.

Richard Dylan Lawless¹, Steve Leung^1,2, Tori Stromp¹, Katherine Thompson³, Vincent Sorrell^1,2, and Moriel Vandsburger^1,4
1Saha Cardiovascular Research Center, University of Kentucky, Lexington, KY, United States, ²Gill Heart Institute, University of Kentucky, KY, United States, ³Department of Statics, University of Kentucky, KY, United States, ⁴Department of Physiology, University of Kentucky, KY, United States

Cine bSSFP is routinely used to image left ventricular structure and function. Changes in signal intensity of images throughout the cardiac cycle are affected by changes in tissue composition and longitudinal shortening. We found that reduced longitudinal ventricular shortening and reduced peak changes in bSSFP signal intensity could be used to differentiate healthy patients from those with LGE enhancement.

Erik P. Skulborstad¹, Zachary S. Borden¹, Karl K. Vigen¹, Glenn S. Slavin², Kang Wang³, Mark L. Schiebler¹, Scott K. Nagle¹, Scott B. Reeder^1,4, Thomas M. Grist^1,4, and Christopher J. Francois^1
1Department of Radiology, University of Wisconsin-Madison, Madison, WI, United States, ²GE Healthcare, Bethesda, MD, United States, ³Global MR Applications and Workflow, GE Healthcare, Madison, WI, United States, ⁴Department of Medical Physics, University of Wisconsin-Madison, Madison, WI, United States

MOdified Look Locker Inversion recovery (MOLLI) is a commonly employed sequence for measuring cardiac T₁, but is subject to measurement error secondary to heart rate variability and acquisition of "apparent" T₁ (T₁*). Saturation Method using Adaptive Recovery Times for cardiac T₁ Mapping (SMART₁Map) is an emerging technique that is able to account for variations in heart rate, in conjunction with acquiring true T₁. We compared myocardial T₁ values obtained with SMART₁Map to those obtained with MOLLI at 3T and found that SMART₁Map offered improved variability and repeatability. Consequently, SMART₁Map should enable increased robustness and precision in myocardial tissue characterization.

Pauline Ferry^1,2, Glenn S. Slavin³, Anne Menini⁴, Anja Brau⁵, Damien Mandry^1,6, Laurent Bonnemains^1,6, Jacques Felblinger^1,7, and Marine Beaumont^7,8
1IADI, Université de Lorraine, Nancy, France, ²U947, INSERM, Nancy, France, ³GE Healthcare, Bethesda, MD, United States, ⁴GE Global Research, Munich, Germany, ⁵GE Healthcare, Munich, Germany, ⁶University Hospital, Nancy, France, ⁷CIC-IT 1433, INSERM, Nancy, France, ⁸CIC-IT, University Hospital, Nancy, France

T1 is an intrinsic tissue parameter. Moreover, myocardial extracellular volume, a surrogate for fibrosis extent estimation, can be derived from pre- and post-injection T1 measurements. This parameter is altered in most, if not all, cardiomyopathies, making it of paramount interest SMART1Map is a single-point method that has been proposed for a true myocardium T1 measurement. This method also provides insensitivity to heart rhythm variation since it adapts the longitudinal magnetization saturation recovery curve sampling to each subject. In this study, we applied SMART1map on healthy volunteers in order to assess the average normal true myocardium T1 value.

Jiaxin Shao¹, Stanislas Rapacchi¹, Kim-Lien Nguyen^1,2, and Peng Hu^1,3
1Department of Radiological Sciences, David Geffen School of Medicine, University of California, Los Angeles, CA, United States, ²Department of Medicine, Division of Cardiology, David Geffen School of Medicine, University of Calif, Los Angeles, CA, United States, ³Biomedical Physics Inter-Departmental Graduate Program, University of California, Los Angeles, CA, United States

We developed the FLASH-MOLLI sequence using proposed Bloch equation simulation with slice profile correction (BLESSPC) for T1 estimation. The FLASH-MOLLI sequence was evaluated against the bSSFP-MOLLI sequence using phantom study, and in 10 healthy volunteers at 3.0T. Phantom results show that the FLASH-MOLLI with BLESSPC fitting reduced average T1 estimation error from -58.5±50.6ms by bSSFP-MOLLI to -1.5±15.3ms for T1s from 440ms-1774ms and HRs from 40bpm-100bpm. The native myocardial T1 values by FLASH-MOLLI were significantly higher than that by bSSFP-MOLLI by 99.0±31.7ms (1454.9±23.6ms vs. 1355.8±23.9ms, p<0.001) at heart rates of 62.5±9.9 bpm, and FLASH-MOLLI is less sensitive to off-resonance artifacts.

Sébastien Roujol¹, Jihye Jang^1,2, Tamer A. Basha¹, Sebastian Weingärtner^1,3, Sophie Berg¹, and Reza Nezafat^1
1Department of Medicine, Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, Massachusetts, United States, ²Computer Aided Medical Procedures, Technische Universität München, Munich, Germany, ³Computer Assisted Clinical Medicine, University Medical Center Mannheim, Heidelberg University, Mannheim, Germany

Native myocardial T1 mapping sequences usually employed a balanced steady state free precession readout, which is susceptible to the B0 field inhomogeneity, T2 dependent and sensitive to magnetization transfer. These effects lead to regional T1 variations and potential reduced accuracy. Gradient recalled echo (GRE) imaging has been proposed for post-contrast T1 mapping but its feasibility for native myocardial T1 mapping remain to be investigated further. In this study, we sought to develop and evaluate a multi-slice T1 mapping sequence using GRE readout and characterize its accuracy, precision, and reproducibility for in-vivo native myocardial T1 mapping.

Tobias Wech¹, Felix Rützel^1,2, Johannes Tran-Gia¹, Andreas Schindele³, Theresa Reiter⁴, Thorsten Klink¹, Michael Braun^1,2, Alfio Borzi³, Walter H. Kullmann², Thorsten A. Bley¹, and Herbert Köstler^1
1Department of Diagnostic and Interventional Radiology, University of Wuerzburg, Würzburg, Germany, ²Institute of Medical Engineering, University of Applied Sciences Würzburg-Schweinfurt, Schweinfurt, Germany, ³Institute of Mathematics, University of Wuerzburg, Würzburg, Germany, ⁴Department of Internal Medicine I, University of Wuerzburg, Würzburg, Germany

Late Gd enhancement imaging is a well-established tool for the determination of infarcted myocardial tissue. However, the necessity for the operator to manually adjust the effective inversion time TI prior to a LGE acquisition makes the acquisition susceptible to sub-optimal contrast between vital and necrotic tissue. The aim of this study was to incorporate a forward-backward motion correction in the model based acceleration of parameter mapping (MAP) technique to reconstruct images of multiple contrasts out of a 2D dataset acquired after a single IR preparation. The presented method allows for retrospectively selecting the desired image contrast and thus improves the flexibility of diagnostic LGE imaging.

Archontis Giannakidis^1,2, Shouvik Haldar¹, Eva Nyktari¹, Jennifer Keegan^1,2, Irina Suman Horduna¹, Dudley J. Pennell^1,2, Raad Mohiaddin^1,2, Tom Wong¹, and David N. Firmin^1,2
1NIHR Cardiovascular Biomedical Research Unit, Royal Brompton Hospital, London, United Kingdom, ²National Heart Lung Institute, Imperial College London, London, United Kingdom

No algorithm has been deemed better than others for segmenting enhancement in late gadolinium enhancement magnetic resonance imaging of atrial fibrillation (AF). We propose a robust -though simple- method to classify enhanced tissue in AF patients studied at out center. We demonstrate the technique's potential for utility in clinical AF studies, and compare it with other related approaches.

Ian Gavin Murphy¹, Oisin Flanagan¹, Marcos J Botelho¹, Jeremy Collins¹, Bruce J Spottiswoode², Maria J Carr², Michael Markl², Robert R Edelman³, and James C Carr^1
1Cardiovascular Imaging, Feinberg School of Medicine, Northwestern Memorial Hospital, CHICAGO, ILLINOIS, United States, ²Cardiovascular Imaging, Northwestern University, CHICAGO, ILLINOIS, United States, ³Cardiovascular Imaging, Northshore Hospital, Evanston, Illinois, United States

Motion corrected SSFP is an investigational prototype sequence which includes fully automated inline non-rigid motion compensation. Our aim was to compare image quality and scar characterisation using this technique compared with gold standard (single shot SSFP and breath-held GRE) sequences. We found that overall image quality is significantly improved in the data analyzed.

Constantin von Deuster^1,2, Christian T. Stoeck^1,2, Martin Genet², Nicolas Toussaint³, David Atkinson⁴, and Sebastian Kozerke^1,2
1Division of Imaging Sciences and Biomedical Engineering, King's College London, London, United Kingdom, ²Institute for Biomedical Engineering, University and ETH Zurich, Zurich, Switzerland, ³Dept of Med Phys & Biomedical Eng, University College London, London, United Kingdom, ⁴Centre for Medical Imaging, University College London, London, United Kingdom

Computational cardiac models require a realistic description of left-ventricular cardiac fiber architecture. In this work, diffusion tensor data across the entire left ventricle of the human heart were successfully acquired in both systole and diastole including correction for myocardial strain. A comprehensive set of cardiac DTI data is presented and will be made available in combination with motion data. The data may serve as realistic input for computational heart modeling projects.

Antonella Meloni¹, Vincenzo Positano¹, Antonino Vallone², Paolo Preziosi³, Maria Chiara Resta⁴, Gennaro Restaino⁵, Maria Giovanna Neri¹, Roberta Renni⁶, Monica Benni⁷, Petra Keilberg¹, Cristina Salvatori⁸, and Alessia Pepe^1
1CMR Unit, Fondazione G. Monasterio CNR-Regione Toscana, Pisa, Italy, ²Istituto di Radiologia, Az. Osp. "Garibaldi" Presidio Ospedaliero Nesima, Catania, Italy, ³U.O.C. Diagnostica per Immagini e Interventistica, Policlinico “Casilino", Roma, Italy, ⁴Struttura Complessa di Radiologia, OSP. SS. Annunziata ASL Taranto, Taranto, Italy,⁵Dipartimento di Radiologia, Università Cattolica del Sacro Cuore, Campobasso, Italy, ⁶Day Hospital, Ospedale Civile “F. Ferrari”, Casarano (LE), Italy, ⁷Servizio di Immunoematologia e Centro Trasfusionale, Policlinico S. Orsola "L. e A. Seragnoli", Bologna, Italy, ⁸Unità Operativa Sistemi Informatici, Fondazione G. Monasterio CNR-Regione Toscana, Pisa, Italy

The left ventricular global function index (LVGFI) is a functional parameter integrating structural as well as mechanical behaviour. In the MESA study a LVGFI <37% was shown to be strongly predictive of cardiovascular events. In the present study, involving patients with thalassemia major, we found out that different patterns of myocardial iron overload evaluated by the T2* multislice technique were more strongly associated with a LVGFI <37% than with a LV dysfunction. Thus, a LVGFI<37% could better identify a significant higher risk of adverse cardiovascular events beyond heart failure in iron loaded patients.

Antonella Meloni¹, Michele Rizzo², Giovanni Carulli³, Esther Natalie Oliva⁴, Francesco Arcioni⁵, Sergio Storti⁶, Maria Giovanna Neri¹, Stefania Renne⁷, Emanuele Grassedonio⁸, Gennaro Restaino⁹, Vincenzo Positano¹, and Alessia Pepe^1
1CMR Unit, Fondazione G. Monasterio CNR-Regione Toscana, Pisa, Italy, ²Reparto di Ematologia, Azienda Sanitaria Provinciale Caltanissetta - Ospedale “Sant'Elia, Caltanisetta, Italy, ³Dip. di Oncologia, dei Trapianti e delle Nuove Tecnologie in Medicina – Divisione di Ematologia, Facoltà di Medicina e chirurgia – Università degli Studi di Pisa, Pisa, Italy,⁴Hematology Unit, A.O. Bianchi-Melacrino-Morelli, Reggio Calabria, Italy, ⁵Dipartimento di Medicina Clinica e Sperimentale, Università degli Studi di Perugia, Perugia, Italy, ⁶UOC di Onco-Ematologia, Università Cattolica del Sacro Cuore, Campobasso, Italy, ⁷Struttura Complessa di Cardioradiologia-UTIC, P.O. “Giovanni Paolo II”, Lamezia Terme, Italy,⁸Dipartimento di Radiologia, Policlinico "Paolo Giaccone", Palermo, Italy, ⁹Dipartimento di Radiologia, Università Cattolica del Sacro Cuore, Campobasso, Italy

The aim of our study was to assess the changes in cardiac and hepatic iron overload and the morpho-functional cardiac parameters by MRI in myelodysplastic patients who performed the follow-up (FU) MRI at 12 months. The new occurrences of cardiac iron, reduced cardiac function, increased LV and RV EDVI and myocardial fibrosis and the worsening in MRI LIC (liver iron concentration) values suggest the need of performing periodic MRI scans, in order to better manage these patients.

Christopher M. Sandino^1,2, Peter Kellman², Michael S. Hansen², Andrew E. Arai², and Hui Xue^2
1Ming Hsieh Department of Electrical Engineering, University of Southern California, Los Angeles, California, United States, ²Lab of Cardiac Energetics, National Heart, Lung, and Blood Institute, Bethesda, Maryland, United States

Pixel-wise T2* mapping is an emerging clinical technique for detection of myocardial iron overload and provides the benefit of having spatial context between tissues when compared to ROI-based T2* mapping. However, pixel-wise mapping is less precise than ROI-based mapping due to its use of low SNR data. We propose and validate a tool to estimate error (standard deviation) maps for evaluating the precision of pixel-wise T2* mapping techniques. This tool will be useful to researchers prototyping and developing more precise pixel-wise mapping techniques.

Junfei Lu¹, J Paul Finn^2,3, and Peng Hu^2,3
1Department of Bioengineering, UCLA, Los Angeles, California, United States, ²Department of Radiological Sciences, UCLA, Los Angeles, California, United States, ³Biomedical Physics Inter-Departmental Graduate Program, UCLA, Los Angeles, California, United States

We propose to use ferumoxytol, an intravascular contrast agent, to achieve flow-independent black blood imaging by taking advantage of its strong R2 relaxivity. As our technique eliminates the need for DIR preparation, we could achieve higher scan efficiency by interleaving the imaging slices. Meanwhile the sharpness of the septal wall is still maintained. It also provides higher SNR and CNR due to signal boost from ferumoxytol.

Matthew Firth¹, Marco Mingarelli¹, Hugh Seton¹, and Dana Dawson^1
1University of Aberdeen, Aberdeen, United Kingdom

In-vivo MRI was carried out on four healthy rats at 4.7 T to obtain a series of T2* measurements in the heart septum. A prospectively gated gradient echo pulse sequence was used with TE values covering the range 5-30 ms. The results were combined to provide a normal dataset for future comparison with rat studies of iron overload cardiomyopathy or cardiac contrast agents. The mean T2* value was 10.12 ± 4.40 ms. We discuss measures that could be taken to improve SNR and reduce signal variation in prospectively gated studies.

Jinnan Wang¹, Rene Bastkowski², Jeffrey H Maki³, Chun Yuan³, and Peter Boernert^4
1Philips Reserach North America, Seattle, WA, United States, ²Philips Reserach Europe, Hamburg, Germany, ³University of Washington, Seattle, WA, United States, ⁴Philips Research Europe, Hamburg, Germany

Referenceless Acquisition of Phase-sensitive Inversion-recovery with Decisive reconstruction (RAPID) imaging was recently proposed, as a time-efficient alternative to the Phase Sensitive Inversion Recovery (PSIR), to reconstruct phase-sensitive images without the need for reference acquisitions. RAPID achieves phase sensitive reconstruction by eliminating the background phase variations (largely introduced by B0 variations) and restoring the underlining phase change introduced by magnetization polarities. The RAPID algorithm, however, may be subject to errors when the image SNR is low and/or the underlining B0 change is severe. Encouraged by recent advancements in water-fat imaging, a geometry based B0 estimation algorithm is incorporated into RAPID to help mitigate these B0-related phase changes and potentially improve the robustness of RAPID algorithm.

Rebecca E Thornhill^1,2, Julie Robillard^3,4, Michael Gollob⁵, Carole Dennie^1,2, Alexander Dick^6,7, Edith Kolozsi⁶, and Elena Pena^1,2
1Medical Imaging, The Ottawa Hospital, Ottawa, ON, Canada, ²Radiology, University of Ottawa, Ottawa, ON, Canada, ³Radiology, Montreal Heart Institute, Montreal, PQ, Canada, ⁴Radiologie, Université de Montréal, Montreal, PQ, Canada, ⁵Electrophysiology, Peter Munk Cardiac Centre, Toronto, ON, Canada, ⁶Cardiology, University of Ottawa Heart Institute, Ottawa, ON, Canada, ⁷Medicine, University of Ottawa, Ottawa, ON, Canada

Evidence of late gadolinium enhancement (LGE) has been associated with the development of arrhythmias and sudden cardiac death (SCD) in patients with hypertrophic cardiomyopathy (HCM). One of the challenges is how to improve the predictive value of LGE in HCM as enhancement may appear heterogeneous. Texture analysis was applied to quantify enhancement heterogeneity in HCM patients considered at high risk of SCD and those who are not. These features were significantly elevated in patients considered at high risk of SCD, even in non-hypertrophic, non-fibrotic segments (ie, normal appearing). Quantitative textural features show potential for assisting in risk stratification in HCM.

Joseph S Soltys¹, Ibrahim M Saeed², Sanjaya Gupta², Piero Ghedin³, Anja C.S. Brau³, James A Case¹, and Timothy M Bateman^1,2
1Cardiovascular Imaging Technologies, Kansas City, MO, United States, ²Saint Luke's Mid America Heart and Vascular Institute, Kansas City, MO, United States, ³Global Research Center, GE Healthcare, Munich, Germany

It is the purpose of this study to visualize and quantitatively evaluate left atrial cryoablation lesions with late gadolinium enhanced cardiac MR using commercially available pulse sequences. The individual pulmonary veins of 10 subjects were included. In particular, consideration is given to acquisition parameters such as the timing of image acquisition post ablation. Significant LGE scar was observed on the majority of the 10 pulmonary veins with well visualized lesions.

Maryam Nezafat^1,2, Sébastien Roujol², Jihye Jang², Tamer Basha², and René M. Botnar^1
1King’s College London, London, UK, United Kingdom, ²Beth Israel Deacons Medical Center and Harvard Medical School, Boston, MA, United States

Quantitative T1 mapping provides myocardial tissue characterization for assessment of interstitial diffuse fibrosis in cardiomyopathies. Various T1 mapping methods with different acquisition scheme have been proposed to sample the magnetization recovery curve. Presence of fatty infiltration is common in many patients with cardiomyopathies. However, current T1 measurements are influenced by lipid contents in the myocardium. The aim of our study is to develop a T1 mapping sequence which eliminates the impact of fat presence on T1 measurements.

Eric Aliotta^1,2, Stanislas Rapacchi¹, Peng Hu¹, and Daniel Ennis^1,2
1Radiological Sciences, UCLA, Los Angeles, CA, United States, ²Biomedical Physics IDP, UCLA, Los Angeles, CA, United States

The purpose was to improve the robustness to bulk motion of spin-echo (SE) cardiac diffusion tensor MRI (DTI) by using ultrahigh maximum gradient amplitudes (Gmax=80mT/m, G80) and a biophysical reconstruction constraint algorithm (BRCA). We found that G80+BRCA offers more accurate ADC and FA quantification and produces a more homogeneous ADC map in accordance with the literature. Shorter diffusion preparation times made possible with ultrahigh gradients combined with a biophysical reconstruction constraint algorithm (BRCA) improved the robustness to bulk motion of cardiac DTI.

Davide Piccini^1,2, Simone Coppo², Giulia Ginami², Gabriele Bonanno², Tobias Rutz³, Gabriella Vincenti³, Juerg Schwitter³, and Matthias Stuber^2
1Advanced Clinical Imaging Technology, Siemens Healthcare IM BM PI, Lausanne, Switzerland, ²Department of Radiology, University Hospital (CHUV) and University of Lausanne (UNIL) / Center for Biomedical Imaging (CIBM), Lausanne, Switzerland, ³Division of Cardiology and Cardiac MR Center, University Hospital of Lausanne (CHUV), Lausanne, Switzerland

Respiratory self-navigation (SN) for whole-heart MRI provides 100% scan efficiency and increased ease of use for the operator. The SN technique was combined with a 3D radial acquisition for phase sensitive inversion recovery (3D-SN PSIR) imaging, featuring high motion and undersampling robustness. Free-breathing 3D-SN PSIR was compared to standard 2D breathhold PSIR imaging in patients with positive late gadolinium enhancement findings. Preliminary results show that free-breathing 3D-SN PSIR achieves robust nulling of healthy myocardium, while preserving high-signal in fibrotic tissue. Overall high diagnostic quality is preserved. Such high-resolution isotropic scar definition could allow for better planning of arrhythmia ablation procedures.

Avinash Kali^1,2, Ivan Cokic¹, Hsin-Jung Yang^1,2, Richard L Q Tang¹, Behzad Sharif¹, and Rohan Dharmakumar^1,3
1Biomedical Imaging Research Institute, Cedars-Sinai Medical Center, Los Angeles, California, United States, ²Department of Bioengineering, University of California, Los Angeles, CA, United States, ³Department of Medicine, University of California, Los Angeles, CA, United States

Native T1 mapping at 3T has been previously shown to be a reliable alternative to Late Gadolinium Enhancement imaging for characterizing chronic myocardial infarctions (cMIs). The purpose of this study is to establish the sensitivity and specificity of automatic, semi-automatic and visual methods; and identify the optimal approach for characterizing cMIs on the basis of native T1 maps at 3T. Semi-automatic approach using Mean + 5 standard deviations criterion showed the strongest diagnostic performance for detecting cMIs on the basis of native T1 maps, while visual delineation of hyperintense cMIs on native T1 maps showed the weakest diagnostic performance.

Jack Miller^1,2, Samira Lakhal-Littleton¹, Magda Wolna¹, Carolyn Carr¹, Ana Santos³, Rebeca Diaz³, Daniel Biggs³, Ben Davies³, Vicky Ball¹, Peter Robbins¹, and Damian Tyler^1
1Department of Physiology, Anatomy & Genetics, University of Oxford, Oxford, United Kingdom, ²Department of Physics, University of Oxford, Oxford, United Kingdom,³Wellcome Trust Centre for Human Genetics, University of Oxford, Oxford, United Kingdom

Iron is vital to mamalian life. We show that the homeostasis of iron in the heart occurs through the protein ferroportin, which is normally only considered in organs with a role in systemic iron homoestasis, such as the liver. Cardiac specific ferroportin knockout mice show severe cardiac dysfunction, as quantified by Cine imaging, and also show a significant decrease in myocardial T2* which correlates with an increased iron concentration, as quantified by mass spectrometry.

Sean Robison¹, Daniel Kim², Kyungpyo Hong², Emma Hornsey¹, Piyush Srivastava^3,4, Gerard Smith¹, Leighton Kearney³, and Ruth P Lim^1,4
1Department of Radiology, Austin Health, Melbourne, Victoria, Australia, ²UCAIR, Department of Radiology, University of Utah, Salt Lake City, Utah, United States,³Department of Cardiology, Austin Health, Melbourne, Victoria, Australia, ⁴The University of Melbourne, Melbourne, Victoria, Australia

We compare the performance of two cardiac T1 mapping pulse sequences, Modified Look-Locker inversion recovery (MOLLI) and arrhythmia insensitive rapid (AIR), in patients with cardiomyopathies. In ten patients referred for varying indications, we found significantly different cardiac T1 and partition coefficient values between MOLLI and AIR. These findings are consistent with previous studies which reported that different cardiac T1 mapping pulse sequences yield significantly different T1 and extracellular volume fraction measurements in normal hearts. Awareness of these differences is important for clinical application and interpretation.

Elias Ylä-Herttuala¹, Svetlana Laidinen¹, Maarit Pulkkinen¹, Hanne Hakkarainen¹, and Timo Liimatainen^1
1Biomedical Imaging Unit, University of Eastern Finland, A. I. Virtanen instute, Kuopio, Finland

We measured longitudinal rotating frame relaxation time (T₁) and late gadolinium enhanced (LGE) images in chronic infarct in mouse model. Infarct was induced by ligation of left anterior descending (LAD). Linear association (R²=0.94, p < 0.001) between the infarct volumes based on T₁-maps and LGE images were found. We conclude that chronic infarct size can be determined by T₁-mapping in mouse chronic infarct model.

Peter David Gatehouse¹, Peter Kellman², EeLing Heng¹, Michael Gatzoulis³, James C Moon⁴, Sonya Babu-Narayan¹, and David N Firmin^3
1Royal Brompton Hospital, London, UK, United Kingdom, ²National Institutes of Health, Bethesda, DC, United States, ³Royal Brompton Hospital, UK, United Kingdom, ⁴The Heart Hospital, University College Hospitals London, London, UK, United Kingdom

Optimisation of single-shot dark-blood fat-water separated imaging for right ventricular T1 measurements by SASHA.

Shingo Kato¹, Roujol Sébastien¹, Jihye Jang¹, Basha Tamer¹, Berg Sophie¹, Kissinger Kraig¹, Goddu Beth¹, Evan Appelbaum¹, Martin Maron², Warren J Manning³, and Nezafat Reza^3
1Department of Medicine, Beth Israel Deaconess Medical Center, Boston, Massachusetts, United States, ²Division of Cardiology, Tufts Medical Center, Boston, Massachusetts, United States, ³Department of Medicine and Radiology, Beth Israel Deaconess Medical Center, Boston, Massachusetts, United States

Myocardial T1 mapping in patients with hypertrophic cardiomyopathy (HCM) should be performed over the entire left ventricle due to regional variability of the location of the hypertrophic region. We have recently developed a free-breathing slice-interleaved T1 (STONE) mapping sequence which allows T1 mapping with complete coverage of LV during free-breathing in 1:30 min. Multi-slice T1 mapping by using STONE sequence could be advantageous to overcome limited cardiac coverage of conventional single-slice T1 mapping technique and to accurately detect the diffuse myocardial fibrosis in HCM patients.

Mingxi Liu¹, Wanshi Zhang², Ziheng Zhang³, Limin Meng², Jie Liu¹, and Wanfeng Gong^2
1The Fourth Military Medical Unviersity, Xi'an, Shannxi, China, ²Air Force General Hospital, Beijing, Beijing, China, ³GE Healthcare China, Beijing, Beijing, China

T1, T2 mapping and multi-b DWI are novel CMR technologies with different advantages on myocardium viability evaluation. The aim of this study is to evaluate myocardial viability using T1, T2 mappings, and multi-b DWI, and explore the characteristics of pre-contrast T1, ECV, T2 and ADC values for patients with recent, sub-acute and chronic myocardial infarction. From this preliminary study, it is not hard to quantitatively identify the infarcted myocardium from the normal, but hard for the reperfusion reversible myocardium around the infarction zone. It¡¯s been demonstrated ECV has a high correlation with the findings of LGE comparing with other parameters.

Jihye Jang^1,2, Cory Tschabrunn¹, Elad Anter¹, Tamer Basha¹, and Reza Nezafat^1
1Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, Massachusetts, United States, ²Computer Aided Medical Procedures, Technische Universität München, Munich, Bayern, Germany

Quantitative T₂ mapping allows non-invasive assessment of myocardial inflammation/edema. In T₂ mapping sequences, rest-cycles are needed between different T₂ weighted images to allow full magnetization recovery, which results in reduced scan time efficiency. In such sequence, multi-segment acquisition is difficult to be employed due to the long scan time. Recently, free-breathing multi-slice myocardial T₂ mapping sequence was proposed to increase scan time efficiency by combining a slice-selective T₂prep pulse and interleaved slice acquisition. In this study, we sought to further extend the multi-slice T₂ mapping sequence to allow segmented acquisition to achieve higher in-plane spatial resolution within acceptable scan time.

Lucia F Giles¹, Vicky Ball¹, Carolyn A Carr¹, Anne-Marie L Seymour², Lydia Le Page¹, Lucy Ambrose¹, and Damian J Tyler^1
1Department of Physiology, Anatomy and Genetics, University of Oxford, Oxford, Oxfordshire, United Kingdom, ²Department of Biological Sciences, University of Hull, Hull, United Kingdom

The development of left ventricular hypertrophy is an independent risk factor in the development of heart failure and is underpinned by metabolic remodelling. Reduced coupling between glycolysis and glucose oxidation has been observed with hypertrophic development in the abdominal aortic constriction (AC) rat model. Dichloroacetate (DCA) is a potent activator of pyruvate dehydrogenase enhancing coupling between glycolysis and glucose oxidation. AC rats were treated with DCA following hypertrophic development and cardiac function and metabolism assessed using cine MRI and hyperpolarized [¹³C]pyruvate. Despite increasing glucose oxidation and altering the myocardial metabolic profile with DCA failed to affect hypertrophic.

Katrien Vandoorne¹, Moriel H. Vandsburger², Yue Han¹, Igor Jacobs¹, Hagit Dafni³, Klaas Nicolay¹, and Gustav J. Strijkers^1,4
1Department of Biomedical Engineering, Eindhoven University of Technology, Eindhoven, Netherlands, ²Department of Physiology, University of Kentucky, Kentucky, United States, ³Weizmann Institute of Science, Israel, ⁴Academic Medical Center, Amsterdam, Netherlands

This study aimed to develop a method to quantify the dynamics of a blood pool contrast agent for mapping the myocardial fractional blood volume, a measure for microvascular density and permeability or the rate of extravasation from myocardial blood vessels using in vivo dynamic contrast enhanced (DCE) MRI. This method enabled noninvasive 3D quantitative mapping of microvascular density and initial rate of extravasation of high-molecular-weight contrast materials in the distinct regions of infarcted myocardium with altered endothelial function and has the potential to track endothelial dysfunction in models of myocardial healing.

Vessel Wall

Thursday 4 June 2015

Chengcheng Zhu¹, Henrik Haraldsson¹, Sinyeob Ahn², Jing Liu¹, Michael Hope¹, and David Saloner^1
1Radiology, UCSF, San Francisco, Califronia, United States, ²Siemens Healthcare, CA, United States

This study developed a 3D MRI technique of abdominal aorta aneurysm which can achieve high isotropic resolution and good blood suppression within a clinically acceptable scan time using DANTE prepared SPACE sequence. The feasibility of using DANTE-SPACE to differentiate thrombus components and the ability to identify inflammation with USPIO contrast agents was demonstrated. These methods can potentially improve patient risk stratification.

Niranjan Balu¹, Jie Sun¹, Jin Liu², Shuo Chen³, Huijun Chen³, and Chun Yuan^1
1Radiology, University of Washington, Seattle, Washington, United States, ²Bioengineering, University of Washington, Seattle, Washington, United States, ³CBIR, Tsinghua University, Beijing, China

Large calcification size and juxtaluminal status of calcification may be indicative of high-risk plaque. Calcification assessment using plaque MRI traditionally requires multicontrast MRI including black-blood and bright-blood weightings. We develop a new method using 3D isotropic gray-blood MRI contrast that is acquired simultaneously with image weightings for identifying intraplaque hemorrhage and stenosis. We validate the method against 2D multi-contrast plaque MRI for assessment of calcification size and juxtaluminal/intraplaque calcification status and show that our method provides accurate results comparable to established plaque MRI methods.

Iulius Dragonu^1,2, Thomas Wehrum², Christoph Strecker², Benjamin R. Knowles¹, Jürgen Hennig¹, and Andreas Harloff^2
1Radiology, Medical Physics, University Medical Center Freiburg, Freiburg, Germany, ²Neurology, University Medical Center Freiburg, Freiburg, Germany

Despite complete routine diagnostics, the etiology of stroke remains cryptogenic in ca. 30% of the patients. Here, we present a novel multi-contrast 3D MRI protocol comprising bright-blood T1-weighted; dark-blood T2-weighted and proton density-weighted images that allows the thorough assessment of the entire thoracic aorta in a clinically feasible time. Accuracy of this protocol was tested in both volunteers and stroke patients with aortic atherosclerosis.

Shuo Chen¹, Zechen Zhou¹, Huijun Chen¹, Bida Zhang², Rui Li¹, Jinnan Wang^3,4, Chun Yuan^1,3, and Xihai Zhao^1
1Center for Biomedical Imaging Research, Department of Biomedical Engineering, School of Medicine, Tsinghua University, Beijing, Beijing, China, ²Healthcare Department, Philips Research China, Shanghai, China, ³Department of radiology, University of Washington, Seattle, United States, ⁴Philips Research North America, Briarcliff Manor, NY, United States

Atherosclerosis is a systemic disease that affects different vascular beds simultaneously. Previously developed techniques are limited by the small coverage, low slice resolution, and long scan time in large coverage atherosclerosis imaging. In this study, we sought to develop a 3D large coverage Atherosclerotic Plaque Assessment with Single Scan (APASS) technique that can be used to image the plaque burden and compositions simultaneously in one scan in a short time at different vascular beds. APASS was validated on healthy volunteers for both carotid artery imaging (160mm, < 4min) and femoral artery imaging (250 mm, <8min).

Lei Zhang¹, Jaeseok Park², Jun Wu³, Xin Liu¹, and Yiu-Cho Chung^1
1Paul C. Lauterbur Center for Biomedical Imaging, Shenzhen Institutes of Advanced Technology, Chinese Academic of Sciences, Shenzhen, Guangdong, China, ²department of brain and cognitive engineering, Korea university, Seoul, Korea, ³Neurology, Peking University Shenzhen Hospital, Shenzhen, Guangdong, China

Intracranial atherosclerosis can be imaged using 3DTSE variants such as T1w-SPACE. T1w-SPACE needs high SNR for high resolution imaging. Meanwhile, the high CSF signal around the intracranial vessels must be low so that vessel wall can be reliably detected. This can be achieved using a previously proposed flip-down pulse. However, the pulse reduces both CSF signal and image SNR. In this study, an application specific signal profile for T1w-SPACE is proposed to compensate for the SNR loss caused by the flip-down pulse. The net result is improved contrast between vessel wall and CSF with little effect on overall image SNR.

Haining Liu¹, Niranjan Balu², Jinnan Wang³, and Chun Yuan^1,4
1Bioengineering Department, University of Washington, Seattle, WA, United States, ²Radiology Department, University of Washington, Seattle, WA, United States, ³Philips Research North America, NY, United States, ⁴Bioengineering Department, Tsinghua University, Beijing, China

Luminal stenosis level, IPH, calcification and lipid rich necrotic core are biomarkers for vulnerable plaques. Currently available 2D multi-contrast black-blood MRA schemes suffer from small FOV, low SNR and misregistration between sequences. In this study, we develop a new sequence which combines the advantages of 3D-MERGE and SNAP into a single 3D sequence that provides large-coverage plaque imaging to visualize MRA, lipid core, calcification and IPH simultaneously.

Abdallah G. Motaal^1,2, Wouter WV Potters¹, Huiming Dong², Luc M. J. Florack³, Klaas Nicolay², Aart J Nederveen¹, Gustav J Strijkers^2,4, and Bram F Coolen^1
1Department of Radiology, Academic Medical Center, Amsterdam, Netherlands, ²Department of Biomedical Engineering, Eindhoven University of Technology, Eindhoven, Netherlands, ³Mathematics and Computer Science, Eindhoven University of Technology, Eindhoven, North Brabant, Netherlands, ⁴Department of Biomedical Engineering and Physics, Academic Medical Center, Amsterdam, Netherlands

A common method to assess vessel wall stiffness is measuring the velocity of the pulse wave (PWV) created by cardiac contraction. PWV can be determined by measuring the time delay between the flow curves at two different slice locations. These measurements require sufficient spatial resolution for accurate flow quantification, as well as high temporal resolution with respect to the travel time of the pulse wave between the two slices. Here, we present a new MRI method for carotid PWV assessment, where we combine retrospective triggering and compressed sensing reconstruction, resulting in flow measurements with in-plane spatial resolution of 0.85 mm and temporal resolution of 200 frames/cardiac cycle.

Roos Blankena^1,2, Rachel Kleinloog¹, Pim van Ooij³, Bon Verweij¹, Bennie ten Haken², and Jaco Zwanenburg^4
1Neurology and Neurosurgery, Brain Center Rudolf Magnus, University Medical Center Utrecht, Utrecht, Utrecht, Netherlands, ²Technical Medicine, University of Twente, Enschede, Overijssel, Netherlands, ³Biomedical Engineering & Physics, Academic Medical Center, Amsterdam, Noord-Holland, Netherlands, ⁴Radiology, University Medical Center Utrecht, Utrecht, Utrecht, Netherlands

In-vivo assessment of the aneurysm wall thickness in combination with wall shear stress (WSS), may provide new insights in the development of the aneurysm. Using high-resolution 7T images of the vessel wall, in vivo aneurysm wall thickness was compared with WSS, obtained from PCA measurements. A preliminary analysis was performed, including 9 unruptured aneurysms of 7 patients. All analysed patients showed variations in wall thickness. A negative correlation between local wall thickness and WSS was found in almost all aneurysms. This is the first study presenting a correlation of local wall thickness with WSS from in vivo measurements.

Claudia Calcagno¹, Martin J Willemink², Bei Zhang¹, Hadrien Dyvorne¹, Philip M Robson¹, Olivier Lairez¹, Bram F Coolen³, Gustav J Strijkers⁴, Tim Leiner², Venkatesh Mani¹, Willem JM Mulder^1,3, and Zahi A Fayad^1
1Department of Radiology, Icahn School of Medicine at Mount Sinai, New York, NY, United States, ²Department of Radiology, University Medical Center Utrecht, Utrecht, The Netherlands, Netherlands, ³Department of Radiology, Academic Medical Center, Amsterdam, The Netherlands, Netherlands, ⁴Department of Biomedical Engineering and Physics, Academic Medical Center, Amsterdam, The Netherlands, Netherlands

Imaging of the atherosclerotic vessel wall requires acquisition with high in plane spatial resolution, high SNR and CNR. On 1.5T and 3T scanners these parameters can be optimized at the expense of acquisition time, which renders imaging protocols unfeasible. Here we present a feasibiltiy study for imaging the abdominal aorta of atherosclerotic rabbits at 7T, and compare SNR and CNR with 1.5 and 3T clinical scanners.

Olivia Viessmann¹, Linqing Li¹, and Peter Jezzard^1
1Nuffield Department of Clinical Neurosciences, Oxford Centre for Functional Magnetic Resonance Imaging of the Brain, Oxford, United Kingdom

DANTE is a low flip angle preparation module for black-blood and CSF suppression. In combination with a TSE with variable flip angle (SPACE) it is suitable for intracranial vessel wall imaging. Low flip angle methods are particularly important at 7 Tesla where SAR limits the application of sequences using large flip angles for blood and CSF nulling. We implemented DANTE-SPACE at 7 Tesla and analysed the signal loss due to brain movement, the effect of B1 inhomogeneity and relaxation time variations on the signal level. A clinically relevant protocol for intracranial vessel wall imaging is presented.

A.A. Harteveld¹, N.P. Denswil², J.C.W. Siero¹, J.J.M. Zwanenburg^1,3, A. Vink⁴, W.G.M. Spliet⁴, P.R. Luijten¹, M.J. Daemen², J. Hendrikse¹, and A.G. van der Kolk^1
1Department of Radiology, University Medical Center Utrecht, Utrecht, Utrecht, Netherlands, ²Department of Pathology, Academic Medical Center, Amsterdam, Netherlands,³Image Sciences Institute, University Medical Center Utrecht, Utrecht, Netherlands, ⁴Department of Pathology, University Medical Center Utrecht, Utrecht, Netherlands

In recent years, several intracranial vessel wall imaging techniques using (ultra)high-field MRI have emerged for the evaluation and characterization of atherosclerotic vessel wall lesions in vivo. However, a thorough validation of MRI results of intracranial plaques with histopathology is still lacking. The purpose of this study was to validate the signal characteristics of different intracranial plaque components with quantitative MRI at 7 tesla using post-mortem circle of Willis specimen. The presented ultrahigh-resolution MR imaging protocol shows that different atherosclerotic plaque components can be identified and characterized using multi-parametric quantitative T1, T2, T2* and PD images.

CE & Non-CE MRA

Thursday 4 June 2015

Andrew N Priest¹, Kristian H Mortensen¹, and David J Lomas^1
1Department of Radiology, Addenbrooke's Hospital and Cambridge University, Cambridge, United Kingdom

Patients with central venous obstruction or restricted venous access may benefit from venous imaging, but contrast agent administration is not possible in this group. We investigated four free-breathing (navigator-gated) and four breath-hold subtractive NCE-MRA sequences, assessing their utility in imaging the superior vena cava (SVC), innominate veins and subclavian veins, in 12 healthy volunteers. These sequences combine DANTE or dual iMSDE flow-suppression methods with either balanced SSFP or gradient echo readouts. The navigator-gated methods always gave diagnostic image quality in the SVC and innominate veins, which are hardest to visualise using ultrasound, with DANTE methods performing slightly better than iMSDE.

Peter Börnert^1,2, Jan Groen³, Christian Stehning¹, Jouke Smink³, and Kay Nehrke^1
1Philips Research, Hamburg, Germany, ²Radiology, LUMC, Leiden, Netherlands, ³Philips Healthcare, Best, Netherlands

In this study a fat-suppressed and T2 contrast-prepared whole heart Zero Echo Time imaging (ZTE) sequence has been implemented and was evaluated in healthy adult volunteers, thus combining optimal contrast, volumetric coverage and short echo times into one single volumetric scan with very low acoustic noise. ZTE shows an image contrast comparable to standard sequences with the additional potential to depict short T2 signal from e.g. plaque or fibrosis. However, SNR problems related to maximum B1 and RF bandwidth limitations have to be addressed.

Gabriele Bonanno¹, Davide Piccini^1,2, Bénédicte Marechal^2,3, Cristophe Sierro⁴, Juerg Schwitter⁵, and Matthias Stuber^1
1Radiology, University Hospital (CHUV) and University of Lausanne (UNIL) / Center for Biomedical Imaging, Lausanne, Switzerland, ²Advanced Clinical Imaging Technology, Siemens Healthcare IM BM PI, Lausanne, Switzerland, ³Radiology, CHUV - LTS5 - Ecole polytechnique Fédérale de Lausanne, Lausanne, Switzerland, ⁴Division of Cardiology and Cardiac MR Center, University Hospital (CHUV) and University of Lausanne (UNIL), Lausanne, Switzerland, ⁵Division of Cardiology and Cardiac MR Center, University Hospital (CHUV) and University of Lausanne (UNIL), Switzerland

In order to improve motion correction, a respiratory signal is extracted from independent component analysis (ICA) of fluctuations of the k-space center amplitude in all receiver coils throughout a 3D radial coronary MR angiography (MRA) acquisition. This signal is then used to bin data in under sampled 3D sub-images related to different respiratory phases. Image-based 3D motion correction is thus enabled, without the need of additional navigators or tracking of the heart from multiple 1D projections. In comparison to 1D self-navigation, the proposed 3D self-navigation method preliminarily demonstrated improved vessel delineation in a small cohort of 7 patients with suspected cardiovascular disease.

Yunlong Song¹, Dongmei Wang¹, Guangnan Quan², and Lizhi Xie^2
1Department of CT & MRI, Air Force General Hospital, Beijing, Beijing, China, ²GE Healthcare China, Beijing, China

Much attention has been drawn on the non-contrast-enhanced MR angiography (NCE-MRA). In this work, to evaluate the diagnostic value of Delta-Flow NCE-MRA for lower extremity arterial stenosis with 3.0T.

Nii Okai Addy¹, Jieying Luo¹, Bob S Hu², and Dwight G Nishimura^1
1Electrical Engineering, Stanford University, Stanford, CA, United States, ²Cardiovascular Disease, Palo Alto Medical Foundation, Palo Alto, CA, United States

Outer volume suppression was applied in coronary MR angiography to reduce the transverse field-of-view in a sub-millimeter resolution, whole-heart acquisition. The reduced field-of-view enabled lowering the required acceleration factor, facilitating iterative reconstruction with L1-ESPIRiT.

Guoxi Xie¹, Xiaoyong Zhang^1,2, Caiyun Shi¹, Xin Liu¹, Debiao Li³, and Zhaoyang Fan^3
1Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Shenzhen, Guangdong, China, ²University of Science and Technology of China, Hefei, Anhui, China,³Biomedical Imaging Research Institute, Cedars Sinai Medical Center, Los Angeles, CA, United States

Recent work shows that NC-MRA for venography (NC-MRV) can be realized by subtracting an image with all vessels suppressed from an image with only arterial vessel suppressed using motion-sensitized driven equilibrium (MSDE) preparations. However, MSDE is sensitive to the velocity of blood flow. It has difficulty to suppress the venous vessels due to the venous blood flow rather slow. A recent developed dark-blood preparation module, DANTE, has minimum flow sensitive for flow suppression. Thus, we used the MSDE and DANTE preparation together to realize NC-MRV. Preliminary in vivo studies demonstrate the feasibility of the proposed technique for NC-MRV.

Sishu Yuan¹ and Liming Xia^1
1Radiology, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology., Wuhan, Hubei, China

To evaluate detection of pulmonary embolism (PE) using conventional MR angiography and unenhanced MR angiography applying spatial labeling with multiple inversion pulses sequence (SLEEK) imaging

Kalina V Jordanova¹, Taehoon Shin², Adam B Kerr¹, and Dwight G Nishimura^1
1Electrical Engineering, Stanford University, Stanford, CA, United States, ²Diagnostic Radiology and Nuclear Medicine, University of Maryland, Baltimore, MD, United States

We design a new acceleration-selective (AS) magnetization preparation pulse sequence that generates arterial image contrast directly using a single acquisition and does not require subtraction. The method allows for nulling times to be chosen for optimal background suppression. Refocusing pulses were incorporated into the AS pulse to mitigate the effects of off-resonance frequency and B1 variations on the excitation profile. We demonstrate the method’s application for lower extremity MRA in healthy volunteers at 1.5T.

Li Jiang¹, Andy Jiang¹, Zhigang Wu¹, Allan Jin¹, Stephon Xu¹, and Feng Huang^1
1Philips Healthcare (Suzhou), Suzhou, Jiangsu, China

Recently, researchers have successfully applied a MSDE (or FSD) prepared non-contrast-enhanced (NCE) MRA technique to anatomies like distal lower extremities, hand and foot. However the proposed method requires excessive additional pre-scans to determine the optimal imaging parameters, which is not suitable for widespread use clinically, especially when it comes to imaging of the three stations of lower extremities. This study optimized and accelerated the iMSDE prepared NCE-MRA technique in three aspects to facilitate easier and faster scans in clinical practice.

Haining Liu¹, Niranjan Balu², Jinnan Wang³, Jie Sun², and Chun Yuan^4,5
1University of Washington, Seattle, WA, United States, ²Radiology Department, University of Washington, Seattle, WA, United States, ³Philips Research North America, NY, United States, ⁴Bioengineering Department, University of Washington, Seattle, WA, United States, ⁵Bioengineering Department, Tsinghua University, Beijing, China

SNAP can simultaneously detect lumen stenosis and intraplaque hemorrhage in a single scan. The accuracy SNAP lumen measurement relative to clinically used techniques such as contrast-enhanced MRA (CE-MRA) is not known. Lumen area measurements from 15 patients scanned with SNAP and CEMRA were compared. Our results show that that SNAP provides comparable measurement to CE-MRA for lumen area indicating that SNAP is a potential non-contrast alternative to CEMRA for lumen stenosis assessment.

Yunxia Li^1,2, Qiang Shen¹, Shiliang Huang¹, Wei Li¹, Eric R Muir¹, Justin Alexander Long¹, and Timothy Q Duong^1
1Research Imaging Institute, The University of Texas Health Science Center at San Antonio, San Antonio, TX, United States, ²Department of Neurology, Tongji Hospital, Tongji University, Shanghai, China

Chronic hypertension increases susceptibility to neurological disorders. The goal of this study was to evaluate the morphology of cerebral arteries using MRA and vessel-wall imaging in an established rat model of hypertension (SHR) at different stages of the disease progression. Comparisons were made with age-matched normotensive Wistar Kyoto (WKY) rats. CBF and CR were altered in early stage of chronic hypertension and worsen with disease progression, ultimately resulting hypoperfusion and compromised cerebrovascular reserve. MRI has the potential to be used to identify brain regions susceptible to hemodynamic compromise, improve understanding of disease pathogenesis, and guide treatments in hypertension.

Robert R. Edelman^1,2, Shivraman Giri³, Ian Murphy², and Ioannis Koktzoglou^1,4
1Radiology, NorthShore University HealthSystem, Evanston, Illinois, United States, ²Radiology, Feinberg School of Medicine, Northwestern University, Chicago, Illinois, United States, ³Siemens Healthcare, Chicago, Illinois, United States, ⁴Radiology, Pritzker School of Medicine, University of Chicago, Chicago, Illinois, United States

Inversion-based arterial spin labeling (ASL) techniques have been used for two decades to create nonenhanced MRA of the cerebrovascular circulation and other vascular systems. Potential limitations of inversion-based ASL MRA include signal loss within distal vessels due to T1 relaxation during transit of the labeled spins, and reduced scan efficiency due to the requirement for a long inflow time (typically ≈1 sec). We propose a new saturation-based approach for ASL MRA called TASSL that improves scan efficiency and reduces sensitivity to arterial transit time. The technique was tested in the circle of Willis and peripheral arterial system.

Robert R. Edelman^1,2, Shivraman Giri³, Ian Murphy², Kieran O'Brien⁴, Matthew D. Robson⁵, and Ioannis Koktzoglou^1,6
1Radiology, NorthShore University HealthSystem, Evanston, Illinois, United States, ²Radiology, Feinberg School of Medicine, Northwestern University, Chicago, Illinois, United States, ³Siemens Healthcare, Chicago, Illinois, United States, ⁴Siemens Healthcare, Switzerland, ⁵Department of Cardiovascular Medicine, Oxford University, Oxford, United Kingdom, ⁶Radiology, Pritzker School of Medicine, University of Chicago, Chicago, Illinois, United States

In patients with peripheral arterial disease and impaired renal function, quiescent-interval single-shot (QISS) MRA using a balanced steady-state free precession (bSSFP) readout has proven to be an accurate and useful alternative to contrast-enhanced MRA and CT angiography. However, the use of a bSSFP readout makes the technique sensitive to off-resonance effects, as may occur in the vicinity of orthopedic hardware such as fixation screws or metallic prostheses. Moreover, the bSSFP readout is potentially sensitive to artifacts from flow turbulence caused by a severe stenosis. In order to overcome these limitations, we tested the feasibility of incorporating an ultra-short echo (UTE) readout into QISS MRA.

Masaki Ishida¹, Ryohei Nakayama¹, Shinichi Takase¹, Katsuhiro Inoue¹, Yoshitaka Goto¹, Yasutaka Ichikawa¹, Motonori Nagata¹, Kakuya Kitagawa¹, and Hajime Sakuma^1
1Radiology, Mie University Hospital, Tsu, Mie, Japan

Target-volume coronary MRA can visualize coronary arteries within a predefined target volume within a shorter acquisition time compared to whole-heart coronary MRA. Relatively small SLAB volume of this approach permits acquisitions of multi-frame 3D data without prolonging scan duration. Recently, non-rigid image registration has been emerged as a technique which can merge images and improve SNR and CNR. Multi-frame 3D acquisitions and non-rigid image reregistration allow for acquisition of free-breathing target-volume 3T coronary MRA covering entire coronary arteries with the image quality that is superior to the single-frame acquisition, within a significantly shorter acquisition time compared to whole-heart coronary MRA.

Dongchan Kim¹, Changheun Oh¹, Hyunseok Seo¹, and HyunWook Park^1
1Electrical engineering, KAIST, Daejeon, Yuseong-Gu, Korea

Recently, non-contrast-enhanced MR angiography techniques have been proposed for the increased concern of nephrogenic systemic fibrosis caused by gadolinium-based contrast agents. Recently, Quiescent-interval single-shot (QISS) MRA technique was developed to resolve these problems. QISS uses saturation RF pulse and one-shot balanced SSFP technique to overcome these issues.2 However, QISS technique needs multiple RF pulses to saturate background and fat signal and these multiple RF pulses can cause the high SAR. In this work, we propose a new MRA technique using the Fourier velocity encoding and dynamic reconstruction, which could eliminate background signals without saturation pulses.

Davide Piccini^1,2, Gary R. McNeal³, W. James Parks^4,5, Michael O. Zenge⁶, and Tim C. Slesnick^4,5
1Advanced Clinical Imaging Technology, Siemens Healthcare, Lausanne, Switzerland, ²Department of Radiology, University Hospital (CHUV) and University of Lausanne (UNIL) / Center for Biomedical Imaging (CIBM), Lausanne, Switzerland, ³Customer Solutions Group, Siemens Medical Solutions USA, Inc, Malvern, PA, United States, ⁴Department of Pediatrics, Emory University, Atlanta, GA, United States, ⁵Children's Healthcare of Atlanta, Atlanta, GA, United States, ⁶MR Product Innovation and Definition, Siemens AG, Healthcare Sector, Erlangen, Germany

Free-breathing coronary MRI in pediatric patients is a valuable radiation-free alternative to coronary CT. Although navigator-gated inversion-recovery gradient echo (NAV IR GRE) imaging after administration of blood pool contrast agents has shown some advantages, compared to T2-prepared bSSFP imaging, unpredictable acquisition times and low scan efficiency due to respiratory gating remain a problem. In this work 3D radial whole-heart coronary MRI with respiratory self-navigation was adapted to IR GRE imaging. After protocol optimization, this approach was compared to NAV IR GRE in a first pediatric patient population. Equivalent vessel quality, sharpness and length were achieved with highly predicable scan times.

Tao Zhang^1,2, Ufra Yousaf¹, Albert Hsiao³, Joseph Y Cheng^1,2, Marcus Alley¹, Michael Lustig^2,4, John M Pauly², and Shreyas S Vasanawala^1
1Radiology, Stanford University, Stanford, CA, United States, ²Electrical Engineering, Stanford University, Stanford, CA, United States, ³Radiology, UC San Diego, San Diego, CA, United States, ⁴Electrical Engineering and Computer Sciences, UC Berkeley, Berkeley, CA, United States

Contrast-enhanced MR Angiography (CE-MRA) is ideal for pediatric vascular imaging due to the lack of ionizing radiation. However, pediatric CE-MRA is usually limited by motion and compromised spatiotemporal resolution. General anesthesia (GA) with periods of suspended respiration is often necessary. Recently, a fast free-breathing CE-MRA technique with high spatiotemporal resolution has been developed. It can significantly reduce the level of GA administrated without compromising image quality. In this work, we investigate the clinical performance of free-breathing pediatric time-resolved CE-MRA.

Hoesu Jung¹, Sohyun Han¹, Seokha Jin¹, Dongkyu Lee¹, and Hyungjoon Cho^1
1Department of Biomedical Engineering, UNIST (Ulsan National Institute of Science & Technology), Ulsan, Gyeongsangnam-do, Korea

MR angiogram is used to investigate vascular malfunction in brain. R1-weighted (T1-weighted) angiogram represents accurate vessel information with positive enhancements, but its sensitivity is low for smaller vasculature in deep brain. Transverse relaxation based ¥ÄR2 and ¥ÄR2* angiograms shows improved visibility in deep brain, while vulnerable to susceptibility artifacts in the vicinity of air-tissue interface and overestimation of vessel size. In this study, we systematically compared the strengths and weaknesses of ¥ÄR2, ¥ÄR2* and R1-weighted angiograms. Conclusively, proposed R1- ¥ÄR2- ¥ÄR2* combined angiogram presents the potential tool to visualize a wide range of vasculature with minimized susceptibility artifacts and enhanced sensitivity.

Naoharu Kobayashi¹, Jianing Pang², Steen Moeller¹, Pierre-Francois van de Moortele¹, Sebastian Schmitter¹, Kamil Ugurbil¹, Debiao Li², Michael Garwood¹, and Gregory J Metzger^1
1Center for Magnetic Resonance Research, Department of Radiology, University of Minnesota, Minneapolis, MN, United States, ²Biomedical Imaging Research Institute, Cedars-Sinai Medical Center, Los Angeles, CA, United States

We investigated the feasibility of using an ultra-short TE (UTE) imaging method combined with the benefits of self-gating to perform contrast enhanced CMRA applications at 7T. Self-gating was performed by PCA on the center k-space point from each acquired profile and validated by comparing with physiological monitoring. The self-gating accurately detected respiratory and cardiac motion. Reconstruction with high temporal resolution captured an initial first pass signal increase after contrast administration. Extremely short TE in UTE diminished the T2^* shortening effect due to the increased r2^* at 7T, and maximized enhancement of T1 contrast for CMRA.

Harald Kramer^1,2, Gregor Jost³, Hubertus Pietsch³, and Maximilian F Reiser^1
1Department of Clinical Radiology, University of Munich, Munich, Bavaria, Germany, ²Department of Radiology, University of Wisconsin - Madison, Madison, Wisconsin, United States, ³MR and CT Contrast Media Research, Bayer Healthcare, Berlin, Berlin, Germany

Although contrast enhanced (CE) techniques are well accepted as the standard of reference for most magnetic resonance angiography (MRA) applications, very little is known about the influence of contrast agent (CA) application schemes on CA bolus shape. Unlike iodinated CAs in computed tomography angiography (CTA) or conventional digital subtraction angiography (DSA), gadolinium (Gd) based CAs (GBCA) have a non-linear relationship between CA-concentration and MR signal. Published CA application approaches are highly variable between institutions and are based on anecdotal evidence only.

Bum-soo Kim¹, Jieun Back¹, Song Lee¹, Jinhee Jang¹, Hyun Seok Choi¹, So-Lyung Jung¹, and Kook-Jin Ahn^1
1Department of Radiology, Seoul St.Mary's Hospital, The Catholic University of Korea, Seoul, Seoul, Korea

Spinal cord MRA is useful in preangiographic evaluation of spinal vascular diseases, but limited FOV makes their detection in a single MRA session difficult. Combined low-dose time-resolved and single-phase high-resolution (HR) contrast-enhanced MRA (CE-MRA) at 3.0T was performed in 9 consecutive patients. Low-dose TR CE-MRA was useful in localization of spinal vascular disease and bolus timing of subsequent single phase high resolution CE-MRA. CE-MRA correctly diagnosed spinal AVM (n=5) and spinal dural arteriovenous fistula (DAVF) (n=4), and was useful as preangiographic tool providing information regarding level of shunt and origin of intercostal and lumbar arteries.

Cardiac Perfusion & Function

Thursday 4 June 2015

Behzad Sharif¹, Reza Arsanjani¹, Rohan Dharmakumar¹, Noel Bairey Merz¹, Daniel S. Berman¹, and Debiao Li^1
1Biomedical Imaging Research Institute, Dept. of Biomedical Sciences, Cedars-Sinai Medical Center, Los Angeles, California, United States

We have developed a non-ECG-gated myocardial perfusion imaging technique capable of imaging all slices at the same systolic phase. The method uses steady-state FLASH imaging with continuous radial sampling. The results in healthy volunteers and CAD patients demonstrate ability of the proposed method in visualizing the transmural extent of perfusion defects and compare its performance to the conventional ECG-gated method.

Yasir Q Mohsin¹, Sajan Goud Lingala², Edward DiBella³, and Mathews Jacob^1
1Electrical Engineering, University of Iowa, Iowa city, IA, United States, ²Electrical Engineering, University of Southern California, Los Angeles, California, United States,³Department of Radiology, University of Utah, S.L.City,UT, United States

We recover myocardial perfusion imaging (MPI) data from undersampled measurements. We are concerned to address the issues related to the motion compensation of the MRI images when the interframe motion is considerably high and that the current CS schemes often result in unacceptable spatio-temporal blurring and residual alias artifacts in the presence of respiratory motion and cardiac motion due to inaccurate gating. The current ME-MC schemes have been shown to improve the results but they are usually slow in implementation and come up with considerably increased computational complexity. A challenging problem when non-convex functions are involved in the recon while we use continuation to address the local minima issues. Our framework does not require the motion parameters to be explicitly estimated, and hence is considerably more efficient than explicit ME-MC schemes.

Ganesh Adluru¹, Liyong Chen², Eugene Kholmovski¹, John Roberts¹, and Edward V.R. DiBella^1
1Radiology, University of Utah, Salt Lake City, Utah, United States, ²Advanced MRI Technologies, CA, United States

Myocardial perfusion MRI is a unique imaging technique that offers information about myocardial health due to narrowed/blocked coronary arteries. Fast imaging methods that undersample k-space and use advanced reconstruction methods like compressed sensing have led to increased slice coverage or increased spatial resolution over conventional methods without sacrificing temporal resolution. Feasibility of ungated perfusion imaging with these advanced methods was also shown recently. An ungated acquisition continuously acquires perfusion data ignoring ECG gating and is highly efficient in data collection compared to standard ECG gated acquisitions, especially in case of arrhythmias. Here we present a framework that is complementary to k-space undersampling methods to increase slice coverage by using a simultaneous multi-slice radial acquisition. Promising results are shown in-vivo combining CAIPI with radial undersampling and compressed sensing reconstructions.

Daniel Stäb^1,2, Peter Speier³, Theresa Reiter⁴, Thorsten Klink², Henning Neubauer², Thorsten A Bley², Tobias Wech², Andreas Max Weng², and Herbert Köstler^2
1The Centre for Advanced Imaging, The University of Queensland, Brisbane, Queensland, Australia, ²Institute of Radiology, University of Würzburg, Würzburg, Bavaria, Germany, ³Siemens AG Healthcare Sector, Erlangen, Bavaria, Germany, ⁴Department of Internal Medicine I, University of Würzburg, Würzburg, Bavaria, Germany

Giving an example of high coverage cardiac perfusion MRI, we demonstrate how reformulating MS-CAIPIRINHA as a pure in-plane acceleration problem facilitates the integration of MS-CAIPIRINHA into the clinical workflow. By employing the multi-band excitation with phase oversampling in the imaging protocol, the multi-band acceleration problem reduces to a simple in-plane acceleration problem. Consequently the standard inbuilt parallel imaging functionality can be used for accelerating the measurement and reconstructing the images. This in turn renders mayor modifications to the sequence and the reconstruction chain unnecessary and provides the flexibility to attach available post-processing functionality like motion correction and up-slope map calculation.

Johannes Tran-Gia^1,2, Jesse Hamilton², David Lohr¹, Kestutis Barkauskas², Andreas M. Weng¹, Herbert Köstler¹, and Nicole Seiberlich^2
1Department of Diagnostic and Interventional Radiology, University of Würzburg, Würzburg, Germany, ²Biomedical Engineering, Case Western Reserve University, Cleveland, Ohio, United States

A technique is presented for the absolute quantification of myocardial perfusion in a 3D volume covering the whole ventricle using contrast-enhanced myocardial first-pass perfusion imaging. After the contrast agent injection, a highly accelerated 3D volume is acquired for each consecutive heartbeat using a saturation-recovery prepared FLASH sequence with a stack-of-spirals trajectory. After the acquisition, missing data are reconstructed by applying parallel imaging in all spatial dimensions. By applying a dual-bolus approach, an unsaturated arterial input function is determined, and absolute perfusion values are quantified in the entire myocardium.

Johannes Tran-Gia¹, David Lohr¹, Andreas M. Weng¹, Christian O. Ritter^1,2, Thorsten A. Bley¹, and Herbert Köstler^1
1Department of Diagnostic and Interventional Radiology, University of Würzburg, Würzburg, Germany, ²Department of Diagnostic and Interventional Radiology, University Medical Center Göttingen, Göttingen, Germany

A method for correcting signal variations due to coil sensitivity variations in quantitative myocardial first-pass perfusion imaging is presented. Using a model-based reconstruction technique, T1 mapping is possible even for the extremely short acquisition times of about 200ms in myocardial perfusion imaging. This requires T1 to be sufficiently short, which is the case during the first pass of the contrast agent bolus through the right and left ventricle. By quantifying T1 in these areas, the coil sensitivity can be estimated locally, and signal changes due to coil sensitivity variations in the myocardium can effectively be reduced in a perfusion image series without additional scans.

Srikant Kamesh Iyer^1,2, Tolga Tasdizen², Ganesh Adluru³, and Edward DiBella^3
1Electrical and Computer Engineering, University of Utah, Salt Lake City, Utah, United States, ²Scientific Computational Institute, University of Utah, Salt Lake City, Utah, United States, ³UCAIR/Radiology, University of Utah, Salt Lake City, Utah, United States

Acquiring dynamic contrast enhanced cardiac perfusion images with high spatio-temporal resolution in necessary to accurately diagnose diseases. Temporal resolution can be improved by acquiring less data in k-space, but this causes loss of image quality. Compresed sensing based methods can leverage sparsity constraints to achieve good quality reconstruction, but the use of traditional gradient descent based minimization schemes yield slow rate of convergence. We develop a multicoil reconstruction method based on Split Bregman and fast iterative shrinkage-thresholding algorithms (FISTA) that outperforms traditional gradient descent based minimization methods.

Graeme A Keith¹, Christopher T Rodgers¹, Michael A Chappell², and Matthew D Robson^1
1Oxford Centre for Clinical Magnetic Resonance Research, University of Oxford, Oxford, Oxfordshire, United Kingdom, ²Institute of Biomedical Engineering, University of Oxford, Oxford, Oxfordshire, United Kingdom

A method for the investigation of myocardial perfusion using Arterial Spin Labelling in conjunction with a Look-Locker style readout is presented. The effect of the order of the slice-selective and globally-selective inversion pulses on the apparent T1 values is investigated. Estimates of the Myocardial Blood Flow are calculated for the six standard myocardial segments.

Johannes Tran-Gia¹, David Lohr¹, Andreas M. Weng¹, Christian O. Ritter^1,2, Thorsten A. Bley¹, and Herbert Köstler^1
1Department of Diagnostic and Interventional Radiology, University of Würzburg, Würzburg, Germany, ²Department of Diagnostic and Interventional Radiology, University Medical Center Göttingen, Göttingen, Germany

In this work, an approach for the determination of unsaturated AIFs is presented and validated. For this purpose, a model-based reconstruction is applied to myocardial first-pass perfusion datasets acquired with a radial saturation-recovery prepared FLASH sequence to reconstruct a contrast agent dynamic for a recovery time of ~3ms. This extremely short recovery time promises the determination of an unsaturated AIF. To assess any saturation effects, the AIFs obtained in 6 healthy volunteers using the proposed method were compared to a dual-bolus AIF determination. Additionally, absolute perfusion values were determined to investigate the influence of saturated AIFs on the quantification.

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

The influence of atherosclerotic plaques on bolus dispersion in contrast-enhanced myocardial perfusion imaging has not yet been studied in detail. In this contribution, we use computational fluid dynamics simulations in a high-detailed coronary vascular model to study the impact of both stenosis grade and morphology on bolus dispersion and, consequently, on MBF quantification.

Maythem Saeed¹, Loi Do¹, Steven W Hetts¹, and Mark W Wilson^1
1Radiology and Biomedical Imaging, University of California San Francisco, San Francisco, Ca, United States

Myocardial perfusion and function was assessed in ischemic myocardium with and without microemboli using MRI and postmortem histomorphometry. Pigs (n=24) served as controls or were subjected to brief coronary artery occlusion with and without microemboli. MRI was used at 3 days and 5 weeks to acquire cine and perfusion images. It was found that perfusion MRI has the ability to differentiate mildly ischemic from microembolized ischemic myocardium, while cine MRI demonstrated the severity of injury at global level. The combination of cine and perfusion MRI showed persistent myocardial stunning and perfusion deficits in mildly ischemic myocardium with minor infarction.

Regine Schmidt¹, Dirk Graafen^1,2, Karsten Sommer¹, Hanns-Christian Breit¹, and Laura Maria Schreiber^1,3
1Section of Medical Physics, Department of Radiology, Johannes Gutenberg University Medical Center, Mainz, Germany, ²Department of Radiology, Johannes Gutenberg University Medical Center, Mainz, Germany, ³Department of Cellular and Molecular Imaging, Comprehensive Heart Failure Center (CHFC), Wuerzburg, Germany

The contrast agent bolus dispersion at myocardial perfusion MRI was examined for a stenotic coronary bifurcation geometry by using computational fluid dynamics simulations. Different flow conditions through the stenotic branch were considered. Simulation results were analyzed quantitatively with MMID4 (a) and the Fermi function model (b). A non-negligible underestimation of myocardial blood flow of up to -16.1% (a) and -11.4% (b) and an overestimation of the myocardial perfusion reserve of up to 7.5% (a) and 10.0% (b) were found. In additional simulations the influence of the erythrocytes in blood on the contrast agent bolus dispersion was found to be non-negligible.

Sarayu Parimal^1,2, Smita Sampath^1,2, Michael Klimas², Dai Feng³, Richard Baumgartner³, Elaine Manigbas⁴, Willy GSell⁴, Jeffrey L. Evelhoch², and Chin Chih-Liang^1,2
1Imaging, MSD, Singapore, ²Imaging, Merck & Co. Inc., WestPoint, Philadelphia, United States, ³Biometric Research, Biostatistics and Research Decision Sciences, Merck & Co. Inc., Rahway, New Jersey, United States, ⁴MRI department, Maccine Pte Ltd, Singapore

Nonhuman primate models of human diseases offer unique platforms to evaluate novel therapeutics with better translatability to patients, where imaging biomarkers revealing early disease progression or treatment responses can be critical. Cardiac dysfunction and perfusion abnormalities at stress are early hallmarks of underlying disease. Herein, the robustness of cardiac function/perfusion biomarkers in healthy nonhuman primates during inotropic stress was evaluated. We found good inter- and intra- observer reproducibility for perfusion and functional reserves. Inter-study repeatability was higher for perfusion reserve. Significant differences between rest and stress existed for peak circumferential strain and myocardial blood flow, but not for diastolic strain-rate.

Gregory J Wehner¹, Jonathan D Suever², Christopher M Haggerty², Linyuan Jing², David K Powell¹, Sean M Hamlet³, Jonathan D Grabau², Dimitri Mojsejenko², Xiaodong Zhong⁴, Frederick H Epstein⁵, and Brandon K Fornwalt^1,6
1Biomedical Engineering, University of Kentucky, Lexington, KY, United States, ²Pediatrics, University of Kentucky, Lexington, KY, United States, ³Electrical Engineering, University of Kentucky, Lexington, KY, United States, ⁴MR R&D Collaborations, Siemens Healthcare, Atlanta, GA, United States, ⁵Biomedical Engineering, University of Virginia, Charlottesville, VA, United States, ⁶Physiology and Medicine, University of Kentucky, Lexington, KY, United States

Displacement Encoding with Stimulated Echoes (DENSE) is a cardiac magnetic resonance technique that encodes tissue displacement into the phase of the MR signal. A spiral acquisition has been used at 1.5T to efficiently acquire data and increase SNR. This sequence has not been validated at 3T, where the SNR would be higher, but field inhomogeneities may lead to measurement errors. We developed a novel method for measuring displacement errors in vivo and applied it in humans at both 1.5T and 3T. Our primary finding is that the same spiral cine DENSE acquisition that has been used at 1.5T can be applied at 3T with equivalent accuracy.

Celal Oezerdem¹, Lukas Winter¹, Andreas Graessl¹, Katharina Paul¹, Antje Els¹, Dirk Voit², Jens Frahm^2,3, and Thoralf Niendorf^1,4
1Berlin Ultra-High Field Facility (B.U.F.F.), MDC, Berlin, Germany, ²Biomedizinische NMR Forschungs GmbH am Max-Planck-Institut für biophysikalische Chemie, Göttingen, Germany, ³DZHK (German Center for Cardiovascular Research), partner site Göttingen, Germany, ⁴Experimental and Clinical Research Center, a joint cooperation between Charité Medical Faculty and the Max Delbrueck Center, Berlin, Germany

In conventional cardiac MR (CMR) breath-held 2D CINE acquisitions segmented over 10-16 heartbeats are the clinical standard for left ventricular function assesment, making the conventional CMR constrained by physiological constraints. Offsetting this shortcoming provides a strong driving force for explorations into real-time imaging of the heart. To meet this goal this work examines the applicability of free breathing real time imaging of the heart and the aorta For this purpose a sixteen channel bow tie dipole transceiver array tailored for cardiac MR at 7.0 T is employed. For comparison traditional 2D CINE FLASH imaging of the heart and aorta is performed. Our findings demonstrate that the 16 channel bow tie antenna arrays supports an extended S-I coverage using a FOV as large as 35 cm. The spatial resolution of (1.2 x 1.2 x 6.0) mm³and the frame rate of 30 frames per second fully meets the requirements of standardized left ventricular structure and function assessment protocols.

Celal Oezerdem¹, Lukas Winter¹, Andreas Graessl¹, Katharina Paul¹, Antje Els¹, and Thoralf Niendorf^1,2
1Berlin Ultra-High Field Facility (B.U.F.F.), MDC, Berlin, Germany, ²Experimental and Clinical Research Center, a joint cooperation between Charité Medical Faculty and the Max Delbrueck Center, Berlin, Germany

Improvements in signal-to-noise-ratio (SNR)and B₁⁺ efficiency of the TX/RX array configurations hold the promise to increase the spatial resolution, which is promising in order to gain better insight into myocardial microstructures. For this reason this work investigates the capabilities of a 16 channel cardiac array for sub-millimeter spatial resolution of the heart at 7.0 T. For this purpose numerical field simulations are performed. A phase setting is derived to afford uniform and efficient excitation. 2D CINE FLASH imaging of the heart is performed in healthy subject using spatial resolutions up to (0.8x0.8x2.5)mm³.The results of this work demonstrated that high resolution CINE imaging of the heart at 7.0 T is feasible.

R Reeve Ingle¹, Kenneth O Johnson¹, Galen D Reed¹, Juan M Santos¹, William R Overall¹, and Bob S Hu^1,2
1HeartVista, Inc., Menlo Park, California, United States, ²Cardiology, Palo Alto Medical Foundation, Palo Alto, California, United States

A technique for free-breathing 2D cardiac cine MRI is proposed. Respiratory motion is monitored in real time using the respiratory bellows signal, and the diminishing variance algorithm is used to reacquire motion-corrupted data. The proposed technique is demonstrated in free-breathing volunteer scans, acquiring a stack of short-axis images covering the entire left ventricle.

Meng-Chu Chang¹, Ming-Ting Wu², Marius Menza³, Mao-Yuan Su⁴, Hung-Chieh Huang², and Hsu-Hsia Peng^5
1Interdisciplinary Program of Nuclear Science, National Tsing Hua University, Hsinchu, Taiwan, ²Department of Radiology, Kaohsiung Veterans General Hospital, Kaohsiung, Taiwan, ³Medical Physics, Department of Radiology, University Hospital Freiburg, Freiburg, Germany, ⁴Department of Medical Imaging, National Taiwan University Hospital, Taipei, Taiwan, ⁵Department of Biomedical Engineering and Environmental Sciences, National Tsing Hua University, Hsinchu, Taiwan

Tetralogy of Fallot (TOF) patients frequently encounter residual pulmonary regurgitation (PR) after repaired surgery. To prevent irreversible heart failure, repaired TOF (rTOF) patients have to undergo pulmonary valve replacement (PVR) to relieve PR and its pathological response to myocardium. This study aims to evaluate radial and longitudinal myocardial velocities and septal curvature for rTOF by phase-contrast MRI. Prior to significant impaired left ventricular ejection fraction, the quantified indices can provide satisfied information to assess regional abnormal myocardial function and flat curvature of septum in rTOF. Those indices might be helpful for patient managements regarding the timing of undergoing PVR surgery.

Tzu-Yu chou^1,2, Hsin-Hui Chiu³, Wen-Yih Isaac Tseng⁴, Marius Menza⁵, and Hsu-Hsia Peng^2
1Institute of Biomedical Engineering, National Taiwan University, taipei, taiwan, Taiwan, ²Department of Biomedical Engineering and Environmental Sciences, National Tsing Hua University, Hsinchu, Taiwan, ³Department of Pediatrics, Taipei Medical University Hospital, Taipei, Taiwan, Taiwan, ⁴Center for Optoelectronic Biomedicine, College of Medicine, National Taiwan University, Taiwan, ⁵Medical Physics, Department of Radiology, University Hospital Freiburg, Freiburg, Germany

Marfan syndrome is a multi-systemic connective disorder and an inherent mutation affecting fibrillin-1 gene. In this study, we used dark-blood tissue phase mapping with PC-MRI to quantify the left ventricular myocardial function and the myocardial dyssynchrony index. The indices were compared with age-matched normal controls so that to show the myocardial abnormality of MFS. MFS demonstrated dyssynchronized TTPr or TTPz either in systole or in diastole. In conclusion, the investigated LV myocardial function and the quantified myocardial DI can reflect the abnormal dyssynchronized conditions for MFS and provide helpful information to evaluate myocardial function before deteriorated cardiac function was shown.

Hyunseok Seo¹, Dongchan Kim¹, and HyunWook Park^1
1Electrical Engineering, KAIST, Daejeon, Korea

Cardiac and respiratory motions cause severe motion artifacts in cardiac MR (CMR) images. In order to avoid the motion artifacts, additional devices of ECG and respiratory-belt gating for detection of cardiac and respiratory motions, respectively, are commonly attached. However, preparation time to set up these additional gating devices and discomfort of patients during imaging are burden for clinical applications. In this work, a novel self-gating (SG) method using the phase information of MR signal in aorta for cardiac gating and in body for respiratory gating is proposed which successfully reconstructs the CMR images.

Meng-Chu Chang¹, Ming-Ting Wu², Marius Menza³, Mao-Yuan Su⁴, Hung-Chieh Huang², and Hsu-Hsia Peng^5
1Interdisciplinary Program of Nuclear Science, National Tsing Hua University, Hsinchu, Taiwan, ²Department of Radiology, Kaohsiung Veterans General Hospital, Kaohsiung, Taiwan, ³Medical Physics, Department of Radiology, University Hospital Freiburg, Freiburg, Germany, ⁴Department of Medical Imaging, National Taiwan University Hospital, Taipei, Taiwan, ⁵Department of Biomedical Engineering and Environmental Sciences, National Tsing Hua University, Hsinchu, Taiwan

After the repaired surgery for tetralogy of Fallot (rTOF) patients, residual pulmonary regurgitation becomes the most problematic sequel, which leads to heart failure and influences the long-term mortality. Therefore, early detection of restrictive myocardial function is important. In this study, our purpose is to evaluate circumferential myocardial motion velocity (V∅) for rTOF patients and to speculate the dyssynchrony by using dark-blood phase-contrast MRI. Patients with rTOF presented weaker rotation during systole and dyssynchrony during diastole. These parameters can provide satisfied information to characterize cardiac twist function of rTOF patients before the cardiac function is impacted.

Yuzo Yamasaki¹, Michinobu Nagao², Masato Yonezawa¹, Satoshi Kawanami², Takeshi Kamitani¹, Torahiko Yamanouchi¹, Kenichiro Yamamura³, Ichiro Sakamoto⁴, Hidetake Yabuuchi⁵, and Hiroshi Honda^1
1Clinical Radiology, Kyushu University, Fukuoka, Fukuoka, Japan, ²Molecular Imaging & Diagnosis, Kyushu University, Fukuoka, Fukuoka, Japan, ³Pediatrics, Kyushu University, Fukuoka, Fukuoka, Japan, ⁴Cardiovascular Medicine, Kyushu University, Fukuoka, Fukuoka, Japan, ⁵Health Sciences, Kyushu University, Fukuoka, Fukuoka, Japan

In this study, we clarified the relative value of left ventricular eccentricity index (EI) with regards to right ventricular (RV) volumetric indices, late gadolinium enhancement (LGE), and pressure indices in patients with congenital heart disease (CHD). A negative correlation between diastolic EI and RV ejection fraction (EF), and a strong correlation between mean pulmonary artery pressure (PAP) and systolic EI was seen. Both systolic and diastolic EIs were significantly higher in patients with LGE compared to without LGE. EI is a simple and valuable parameter to predict RV hemodynamics and myocardial fibrosis in CHD.

Rieko Ishimura¹, Kenich Yokoyama¹, Toshiya Kariyasu¹, Shigehide Kuhara², and Toshiaki Nitatori^1
1department of radiology, Kyorin University, Mitaka, Tokyo, Japan, ²Toshiba medical systems, Otawara, Tochigi, Japan

Severe pulmonary hypertension(PH) causes significant dilatation of the right ventricle(RV), and the compensatory hypertrophy of the trabeculae carneae in response to PH leads to severe deformation of the RV. This deformation makes it difficult to identify the anatomical positions of the tricuspid valve and pulmonary valve in CMR imaging. TAX image is more difficult to identify of the interface between the blood within the cardiac chambers and the myocardium due to partial volume effects. In patients with diseases associated with deformation of RV, such as CTEPH, SAX images may be more useful than TAX images for right heart functional analysis.

Helene Feliciano^1,2, Ruud B. van Heeswijk^1,2, Davide Piccini^3,4, Pierre Monney^5,6, Juerg Schwitter^5,6, Roger Hullin⁵, and Matthias Stuber^1,2
1Department of Radiology, University Hospital (CHUV) and University of Lausanne (UNIL), Lausanne, Switzerland, ²Center for Biomedical Imaging (CIBM), Lausanne, Switzerland, ³Department of Radiology, University Hospital (CHUV) and University of Lausanne (UNIL) / Center for Biomedical Imaging (CIBM), Lausanne, Switzerland,⁴Advanced Clinical Imaging Technology, Siemens Healthcare IM BM PI, Lausanne, Switzerland, ⁵Division of Cardiology, Department of Internal Medicine, University Hospital (CHUV) and University of Lausanne (UNIL), Lausanne, Switzerland, ⁶Cardiac MR Center (CRMC), University Hospital of Lausanne (CHUV), Lausanne, Switzerland

In this study, left ventricular tissue relaxometry and dynamics measurements (T₂ mapping and slice-followed CSPAMM myocardial tagging) were combined to longitudinally and quantitatively characterize the myocardium of heart transplant recipients in comparison to the gold standard myocardial biopsy. Initial results obtained in a small patient cohort are presented and discussed.

Osama Abdullah¹, Arnold David Gomez¹, Samer Merchant¹, Michael Heidinger², Steven Poelzing², and Edward W Hsu^1
1Bioengineering, University of Utah, Salt Lake City, Utah, United States, ²Cardiac Research and Training Institute, University of Utah, Utah, United States

Intravoxel incoherent motion (IVIM) was investigated in an animal model of isolated perfused heart as functions of myocardial blood flow (MBF) and diffusion encoding direction. The IVIM-derived mean blood velocity (D*) and volume fraction (VF) were correlated to independent measure of MBF using arterial spin labeling. The D* was faster in the direction parallel than perpendicular to myofibers at normal inflow pressures. The VF had a strong dependence on MBF but not myofiber orientation. Collectively, results indicate that IVIM parameters in the myocardium depend on microcirculation and myofiber orientation in consistent fashions with the heart’s known anatomy and circulation physiology.

Nivedita K. Naresh¹, Xiao Chen¹, Yikui Tian², Eric M. Moran¹, Brent A. French¹, and Frederick H. Epstein^1,3
1Biomedical Engineering, University of Virginia, Charlottesville, Virginia, United States, ²Surgery, University of Virginia Health System, Charlottesville, Virginia, United States,³Radiology, University of Virginia, Virginia, United States

Myocardial blood flow (MBF) imaging in mice can be used to study mechanisms underlying heart disease. We compared first-pass MRI and ASL for perfusion imaging in C57Bl/6 mice in terms of image quality, repeatability and user variability under different conditions: rest, stress and after myocardial infarction. Image quality was better with ASL under all conditions. Repeatability was better with first-pass MRI at low MBF conditions such as infarct imaging. User variability was better with ASL at high MBF conditions such as stress imaging. Our results may be useful in planning future studies investigating perfusion abnormalities in heart diseases.

Dana C Peters¹, Stephanie L Thorn², Alda Bregasi², Edgar J Diaz¹, Mitchel R Stacy², Christi Hawley², and Albert J Sinuas^2
1Radiology, Yale School of Medicine, New Haven, CT, United States, ²Cardiology, Yale School of Medicine, New Haven, CT, United States

Our study focused on determining the acute effects of myocardial infarction on left atrial (LA) function and myocardial injury, with the hypothesis that myocardial infarction may lead to changes in atrial mechanics, which will result in acute inflammation and later fibrosis. Furthermore, we investigated whether atrial fibrillation (AF) inducibility after myocardial infarction correlates to changes in the LA function and substrate. Atrial remodeling after myocardial infarction was studied in 6 control swine and 9 pigs post-MI. LA volumes were increased after infarction. The atrial fibrosis volume by LGE trended greater in pigs with AF inducibility (6/9 post-MI and 0/6 controls).

Emil Knut Stenersen Espe^1,2, Jan Magnus Aronsen^1,3, Kristine Skårdal^1,2, Lili Zhang^1,2, and Ivar Sjaastad^1,2
1Institute for Experimental Medical Research, Oslo University Hospital and University of Oslo, Oslo, OSLO, Norway, ²KG Jebsen Cardiac Research Center and Center for Heart Failure Research, University of Oslo, Oslo, Oslo, Norway, ³Bjørknes College, Oslo, Oslo, Norway

The rat model of congestive heart failure (CHF) as an outcome of myocardial infarction (MI) plays an irreplaceable role in preclinical cardiovascular research. However, not all post-MI rats develop CHF. It is therefore imperative to establish robust in vivo diagnostic criteria for CHF in post-MI rats. In this study, we investigated whether MRI can constitute a noninvasive tool for detecting of CHF in post-MI rats. We found that MRI is able to accurately identify rats with CHF by measuring RV mass. MRI thus constitute a noninvasive tool for longitudinal evaluation of CHF in the postinfarction rat.

Dan Mu¹, Wei Bo Chen², Bin Zhu³, and Biao Xu^4
1Drum Tower Hospital, Nanjing, Jiangsu, China, ²Philips Healthcare, Shanghai, China, ³Radiology, Drum Tower Hospital, Nanjing, Jiangsu, China, ⁴Cardiology, Drum Tower Hospital, Nanjing, Jiangsu, China

The current study investigate the effects of transplanted autologous mesenchymal stem cells overexpressing ILK on myocardial perfusion, myocardial fibrosis and cardiac function in the swine AMI model by MR imaging.

Tobias Wech¹, Nicole Seiberlich², Andreas Schindele³, Michael L. Gyngell⁴, Valentina Davidoiu⁵, Alfio Borzi³, Herbert Köstler¹, and Jürgen E. Schneider^6
1Department of Diagnostic and Interventional Radiology, University of Wuerzburg, Würzburg, Germany, ²Biomedical Engineering, Case Western Reserve University, Cleveland, OH, United States, ³Institute of Mathematics, University of Wuerzburg, Würzburg, Germany, ⁴Perspectum Diagnostics Ltd, Oxford, United Kingdom, ⁵Division of Imaging Sciences & Biomedical Engineering, King's College London, London, United Kingdom, ⁶Division of Cardiovascular Medicine, University of Oxford, Oxford, United Kingdom

The feasibility of performing real time imaging to assess cardiac function in mice at 9.4 T was explored. A radial gradient echo sequence was applied to acquire highly undersampled data without cardiac and/or respiratory gating. Radial GRAPPA and Compressed Sensing were combined to obtain fully sampled data. The method was first optimized in simulations and then applied to three mice in vivo. Left-ventricular volumes and ejection fractions quantified from a mid-ventricular slice agreed well with corresponding fully sampled segmented cine acquisitions. Our work indicates that it is possible to accurately measure LV function in mice with real-time MRI.

Thomas A Roberts¹, Anthony N Price², Anna L David³, Valerie Taylor¹, Daniel J Stuckey*¹, and Mark F Lythgoe*^1
1Centre for Advanced Biomedical Imaging, Division of Medicine, University College London, London, United Kingdom, ²Division of Imaging Sciences and Biomedical Engineering, London, United Kingdom, ³Institute for Women’s Health, University College London, London, United Kingdom

Common measures of systolic cardiac function often fail to detect heart failure. Many patients present with preserved ejection fraction despite underlying pathology. Therefore, a more sensitive measure of heart failure is required. In this study, a high-temporal resolution (HTR-) CINE sequence (approximately one frame every millisecond) is presented for the assessment of diastolic function in mice. The performance of the sequence is compared against Doppler ultrasound, which is the established method for assessing diastolic function. Repeatability of the HTR-CINE sequence is shown to be comparable to ultrasound in a cohort of naïve mice. In a cohort of infarcted mice, HTR-CINE MRI is found to be more sensitive at detecting heart failure compared to ultrasound.

Let It Flow

Thursday 4 June 2015

Flora A. Kennedy McConnell¹, Thomas W. Okell², Michael A. Chappell¹, and Stephen J. Payne^1
1Institute of Biomedical Engineering, Department of Engineering Science, University of Oxford, Oxford, Oxfordshire, United Kingdom, ²FMRIB Centre, Nuffield Department of Clinical Neurosciences, University of Oxford, Oxford, Oxfordshire, United Kingdom

Conventional angiography techniques only provide qualitative information about cerebrovascular disease and collateral blood flow. Here a novel mathematical model for the quantification of blood flow rates from dynamic MR angiography data is proposed. Fitting the model to vessel-encoded pseudo-continuous arterial spin labeled signals from a flow phantom allowed accurate estimation of water flow rates in short vessel segments. Applying the method to healthy volunteer data produced brain-feeding artery flow rate estimates within physiological norms. Also demonstrated was the potential of the technique to identify and estimate flow through often unseen, collateral vessels by fitting the model to signals detected downstream.

Lennart J. Geurts¹, Willem H. Bouvy², Hugo J. Kuijf³, Peter R. Luijten¹, L. Jaap Kappelle², Geert Jan Biessels², and Jaco J.M. Zwanenburg^1
1Radiology, UMC Utrecht, Utrecht, Netherlands, ²Neurology, UMC Utrecht, Utrecht, Netherlands, ³Imaging Sciences Institute, UMC Utrecht, Utrecht, Netherlands

Reproducibility study at 7T for blood flow velocity measurements in cerebral perforating arteries.A single slice phase contrast sequence with 0.3x0.3x2.0mm voxels was used at three locations (medial cerebral artery:M1, basal ganglia:BG and semioval centre:CSO) in six healthy volunteers.Mean velocities (in cm/s) and their coefficients of repeatability (CoR) were 44(10) for M1, 4.6(1.9) for BG, 0.63(0.46) for CSO.The pulsatility index and its CoR was 0.56(0.24) for M1, 0.40(0.15) for BG and 0.28(0.14) for CSO.Blood flow velocity and pulsatility in cerebral perforating arteries may be an interesting new metric to study hemodynamic changes in ageing and cerebral small vessel disease.

Julio Garcia¹, Michael Markl¹, Jeremy Collins¹, James Carr¹, and Alex J Barker^1
1Radiology, Northwestern University, Chicago, Illinois, United States

Elevated helical flow in the thoracic aorta may result from hemodynamic and anatomic alterations and valve phenotype. Helical flow is often qualitatively estimated by visual methods. A quantitative approach to assess the degree of helicity can be achieved by the computation of localized normalized helicity (LNH). This study aimed to use the full volume of the 4D flow measured velocity fields to demonstrate that LNH quantification may differentiate helical flow alterations in the aorta between healthy controls and BAV subjects and the correlation of elevated LNH with BAV aortic dilation.

Mohammed S.M. Elbaz¹, Patrick J.H. de Koning¹, Jos J.M. Westenberg¹, Emmeline E. Calkoen², Boudewijn P.F. Lelieveldt^1,3, Arno A.W. Roest², and Rob R.J. van der Geest^1
1Division of Image Processing, Radiology, Leiden University Medical Center, Leiden, Netherlands, ²Paediatric Cardiology, Leiden University Medical Center, Leiden, Netherlands,³Intelligent Systems, Delft University of Technology, Delft, Netherlands

Vortex formation in the left ventricle during diastolic filling has been confirmed by many in vivo and in vitro studies. Although many studies have postulated a potentially important role of vortex formation on the efficiency of cardiac pumping function, only limited knowledge is currently available on the actual contribution of the intrinsically three-dimensional vortex formation during diastole to the normal intra-cardiac blood flow in vivo. In this work, using 4D Flow MRI from ten healthy controls, we applied 3D vortex core analysis in combination with particle tracing to quantify the contribution of early and late filling 3D vortex rings to the normal LV flow over the cardiac cycle.

Ke Ma¹, Zechen Zhou¹, Aiqi Sun¹, Shuo Chen¹, and Rui Li^1
1Center for Biomedical Imaging Research, Department of Biomedical Engineering, School of Medicine, TsingHua university, Beijing, China

4D flow MRI has been widely used for the assessment of in vivo blood Flow .Blood flow and pressure differences are an important clinical marker for the severity of cardiovascular disease such as pulmonary hypertension . Previous studies found high reproducibility of blood flow and WSS in aorta and pulmonary artery (PA),but no report on the reliability of pressure difference (PD) .The aim of this study was to investigate the scan-rescan reproducibility of 4D MRI in the PA for blood flow and PD.

Ying Wang^1,2, Jing Jiang^1,3, Paul Kokeny¹, Yi Zhong⁴, and E. Mark Haacke^1,4
1Department of Biomedical Engineering, Wayne State University, Detroit, MI, United States, ²College of Information Science and Engineering, Northeastern University, Shenyang, Liaoning, China, ³Department of Radiology, Wayne State University, Detroit, MI, United States, ⁴MR Innovations, Inc., Detroit, MI, United States

Quantifying flow from 2D phase contrast MRI data requires that the vessels of interest be identified and segmented. Doing so manually is time consuming and depends on the skill level of the processor. Here, a tissue similarity mapping (TSM) based automatic segmentation and labeling method for use in the neck is proposed. Magnitude and phase information is utilized through TSM to extract and classify vessels as arteries or veins. A priori knowledge about vessel locations are used to identify ten major vessels found at the C6 level. Accuracy of the method is demonstrated on in vivo human data.

Vinicius Rispoli¹, Jon-Fredrik Nielsen², Krishna Nayak³, and Joao Luiz Carvalho^1
1University of Brasilia, Brasilia, DF, Brazil, ²University of Michigan, Ann Arbor, MI, United States, ³University of Southern California, Los Angeles, CA, United States

Fourier velocity encoding (FVE) is a promising MRI method for assessment of cardiovascular blood flow. FVE provides considerably higher SNR than phase contrast (PC) imaging, is robust to partial-volume effects, and can be rapidly acquired using spiral readouts. On the other hand, FVE data do not directly provide velocity maps. CFD driven by PC velocity maps has been previously demonstrated. This work introduces a method for using FVE data (rather than PC data) to guide CFD simulations. We show that FVE-driven CFD achieves better agreement with a PC-measured velocity map than pure CFD solutions.

Alex J Barker¹, Pim van Ooij¹, David Guzzardi², S. Chris Malaisrie³, Patrick M. McCarthy³, James Carr¹, Jeremy Collins¹, Michael Markl^1,4, and Paul W. M. Fedak^2,3
1Radiology, Northwestern University, Chicago, IL, United States, ²Department of Cardiac Sciences, University of Calgary, Calgary, AB, Canada, ³Division of Surgery-Cardiac Surgery, Northwestern University, Chicago, IL, United States, ⁴Biomedical Engineering, Northwestern University, Chicago, IL, United States

Aggressive aortic resection strategies for bicuspid aortic valve (BAV) patients with significant aortopathy are sometimes warranted. Recent evidence has shown that 4D flow MRI can identify regions of the aorta with elevated wall shear stress (WSS) that may be at risk of disease progression and thus may require resection during aneurysm repair. This study assessed the efficacy of standard aortic resection practices at our instition to include tissue areas at risk as determined by pre- and post-operative imaging. In selected patients with BAV, aggressive resection using open distal/hemi-arch repair acheived complete resection of tissue at risk of disease progression.

Teodora Chitiboi^1,2, Susanne Schnell², Jeremy Collins², James Carr², Horst Hahn¹, and Michael Markl^2
1Fraunhofer MEVIS, Bremen, Bremen, Germany, ²Radiology, Northwestern University, Chicago, IL, United States

This work presents a novel, robust approach to compute myocardial torsion from Tissue Phase Mapping (TPM) MRI, capable of capturing transmural twisting patterns. A high resolution temporal velocity field of the heart muscle is acquired for three short-axis TPM slices, from which the average rotational velocity is computed. Torsion is quantified as the difference in rotational motion between heart slices over time, normalized by slice distance. Torsion parameter is compared with domain expert assessment of ejection fraction, delayed enhancement and wall motion abnormalities for cardiomyopathy patients and healthy volunteers showing potential for predicting cardiac contraction problems.

Julio Sotelo^1,2, Israel Valverde^3,4, Philipp Beerbaum⁵, Heynric B. Grotenhuis⁶, Gerald Greil⁷, Tobias Schaeffter⁷, Reza Razavi⁷, Daniel E. Hurtado², Sergio Uribe^1,8, and C. Alberto Figueroa^7,9
1Biomedical Imaging Center, Electrical Engineering Department, Pontificia Universidad Catolica de Chile, Santiago, RM, Chile, ²Structural and Geotechnical Engineering Departement, Pontificia Universidad Catolica de Chile, Santiago, RM, Chile, ³Pediatric Cardiology Unit, Hospital Virgen del Rocio, Seville, Spain, ⁴Cardiovascular Pathology Unit, Institute of Biomedicine of Seville (IBIS), Seville, Spain, ⁵Hannover Medical University, Hannover, Niedersachsen, Germany, ⁶Child Cardiology Department, Leiden University, Leiden, Netherlands, ⁷Division of Imaging Sciences and Biomedical Engineering, King’s College London, London, United Kingdom, ⁸Radiology Department, School of Medicine, Pontificia Universidad Catolica de Chile, Santiago, Chile, ⁹Department of Surgery and Biomedical Engineering, University of Michigan, Michigan, United States

We predict the pressure gradient using CFD and MRI in seven patients with AoCo, who had a previous MRI and cardiac catheterization study at rest and stress condition. We obtained a good agreement of pressure gradients measurement with the mean and peak value of the pressure curves. The average difference of pressure gradient using the mean values was 0.09mmHg and 1.34mmHg for rest and stress conditions, the mean difference of pressure gradient using the peak value was 0.6mmHg and 15.3mmHg for rest and stress conditions respectively. We are able to predict non-invasively the pressure gradient using CFD based on MRI.

Victoria Stoll¹, Aaron Hess¹, Malenka Bissell², Jonatan Eriksson³, Petter Dyverfeldt³, Andrew Lewis², Tino Ebbers³, Saul Myerson², Carl-Johan Carlhäll³, and Stefen Neubauer^2
1Division of Cardiovascular Medicine, OCMR, Oxford, United Kingdom, ²OCMR, Oxford, United Kingdom, ³Division of Cardiovascular Medicine and Center for Medical Imaging Science and Visualization (CMIV), Linköping University, Linköping, Sweden

Left ventricular failure results in inefficient intra-cardiac blood flow; 4D flow imaging visualises this blood flow and allows calculation of the kinetic energy (KE) of the blood. 15 healthy volunteers were scanned twice to assess the reproducibility and variability of the KE values of the intra-cardiac blood flow components within the healthy heart, in order to understand the changes in diseased states. The end diastolic KE values were reproducible with no significant physiological variability over time. Calculation of the end diastolic KE values of the different left ventricular flow components is a robust technique, with good intra and inter-scan reproducibility.

Thekla Oechtering¹, Julian Haegele¹, Peter Hunold¹, Michael Scharfschwerdt², Markus Huellebrand³, Hans-Hinrich Sievers², Jörg Barkhausen¹, and Alex Frydrychowicz^1
1Clinic for Radiology and Nuclear Medicine, University Hospital Schleswig-Holstein, Lübeck, Germany, ²Department of Cardiac and Cardiothoracic Vascular Surgery, University Hospital Schleswig-Holstein, Lübeck, Germany, ³Fraunhofer MEVIS, Bremen, Germany

The anatomically shaped sinus prosthesis (Uni-Graft®W SINUS, Braun) implies near-physiological hemodynamics. Therefore, we sought to assess their flow characteristics in comparison to straight grafts and volunteers by use of 4D Flow MRI. 15 patients (13 with sinus prosthesis, 3 straight grafts) and 15 age-matched healthy volunteers were examined at 3T (Philips Achieva). Secondary flow patterns in four segments of the thoracic aorta and geometry were analyzed. We found near-physiological flow in the sinuses of sinus prostheses in contrast to straight grafts. Patients typically presented with angular geometries and increased vortex formation in the ascending aorta, presumably related to prosthesis implantation.

Sarah Kefayati¹, Henrik Haraldsson², Belén Casas Garcia³, Jonas Lantz³, Tino Ebbers³, and David Saloner^2
1University of California, San Francisco, San Francisco, California, United States, ²University of California, San Francisco, California, United States, ³Linköping University, Sweden

Due to clinical significance, estimation of pressure changes are normally desired across a region with restricted blood flow that can favor turbulence production. Recent advances in magnetic resonance velocimetry enable velocity-based estimation of relative pressure without the effects of turbulence. However, laminar assumption for turbulent flow results in significant errors in solving the pressure equations. This study investigates the importance of including the full-turbulent tensor components (both normal and shear Reynolds stresses) in calculating pressure. Our findings show the significance of taking into account the turbulent property of the flow resulting in notably lower estimation of pressure recovery.

Jan Paul¹, Peter Bernhardt¹, Heiko Neumann², and Volker Rasche^1
1Internal Medicine II, University Hospital Ulm, Ulm, Germany, ²Institute of Neural Information Processing, University of Ulm, Ulm, Germany

Due to the distinct ghosting artifacts, in Tissue Phase Mapping (TPM), inflowing blood causes phase errors corrupting the velocity information in Cartesian velocity-encoded acquisitions. Suppression of these phase errors demand blood suppression [1], causing a reduction of the temporal fidelity of the cine data. The different point-spread function in radial data acquisition causes less distinct artifacts and may lead to less velocity corruption enabling white-blood TPM imaging. This contribution compares the degree of corruption in black- and white-blood TPM for Cartesian and radial velocity measurements.

Jennifer Keegan¹, Hitesh Patel¹, Robin Simpson², Raad Mohiaddin^1,3, and David Firmin^1,3
1Royal Brompton Hospital, London, United Kingdom, ²University of Freiburg, Freiburg, Germany, ³Imperial College, London, United Kingdom

Renal resistive index (RI) and pulsatility index (PI) are reliable measures of downstream renal resistance which correlate with the severity of renal disease. To date, breath-hold MR phase velocity mapping studies have lacked the temporal resolution required to accurately determine these pulsatility parameters. We have developed a high temporal resolution (19 ms) breath-hold spiral phase velocity mapping technique for the assessment of the temporal flow patterns in the renal arteries and show high inter-observer and inter-study reproducibility. We conclude that high temporal resolution spiral phase velocity mapping allows reproducible assessment of renal pulsatility indices.

Pei-Hsin Wu¹, Hsiao-Wen Chung¹, Cheng-Chieh Cheng¹, Ming-Ting Wu², and Cheng-Wen Ko^3
1Institute of Biomedical Electronics and Bioinformatics, National Taiwan University, Taipei, Taiwan, ²Department of Radiology, Kaohsiung Veterans General Hospital, Kaohsiung, Taiwan, ³Department of Computer Science and Engineering, National Sun Yat-Sen University, Kaohsiung, Taiwan

Qualitative investigation of repaired Tetralogy of Fallot (ToF) shows that the forward and regurgitant flow could occur in the pulmonary arteries and be recorded simultaneously in any single cardiac phase. In this study, we use the index of eccentricity and onset of regurgitant flow to quantify the spatial flow profile and the degree of flow inhomogeneity for both branch pulmonary arteries. The results show that these spatial indices could reflect the heterogeneous flow profile pattern quantitatively, and thus may find useful applications in post-surgical evaluations.

Emmeline Calkoen¹, Patrick de Koning², Rob van der Geest², Albert de Roos², Arno Roest¹, and Jos Westenberg^2
1Pediatric Cardiology, LUMC, Leiden, Netherlands, ²Radiology, LUMC, Leiden, Netherlands

We aimed to quantitatively describe left ventricular (LV) blood flow patterns and evaluate ejection efficiency using 4-dimensional velocity-encoded cardiac magnetic resonance imaging (4DFlow MRI) and particle tracing in healthy volunteers and corrected AVSD patients. Patients presented more lateral and apical inflow, which resulted in decreased direct flow and increased retained flow percentage for apical and lateral LV cavity segments, which may contribute to a decreased cardiac pumping efficiency.

T. Stan Gregory¹, John Oshinski², Ehud J. Schmidt³, Mikayel Dabaghyan³, Raymond Y. Kwong⁴, William G. Stevenson⁴, and Zion Tsz Ho Tse^1
1College of Engineering, The University of Georgia, Athens, Georgia, United States, ²Department of Radiology, Emory University Hospital, Atlanta, Georgia, United States,³Department of Radiology, Brigham and Women's Hospital, Boston, Massachusetts, United States, ⁴Department of Cardiology, Brigham and Women's Hospital, Boston, Massachusetts, United States

Development of a technique to assess Beat-to-Beat Stroke Volume and Aortic Blood Flow using Magnetohydrodynamic Voltages Found in 12-lead Electrocardiograms.

Jan Paul¹, Stefan Wundrak¹, Heiko Neumann², and Volker Rasche^1
1Internal Medicine II, University Hospital Ulm, Ulm, Germany, ²Institute of Neural Information Processing, University of Ulm, Ulm, Germany

We investigated transmural (endocardial–epicardial) differences of LV motion by means of velocities measured with high resolution tissue phase mapping (TPM). The results show that some transmural differences are lost by simple averaging of multiple respiratory states, but are preserved with applied motion correction. In conclusion, high resolution TPM allows detailed analysis of LV motion. Motion correction enables improved SNR while preserving image sharpness and velocity information.

Suliman Barhoum¹, Michael C. Langham¹, Jeremy F. Magland¹, Chamith S. Rajapakse¹, Cheng Li¹, and Felix W. Wehrli^1
1Radiology, University of Pennsylvania, Philadelphia, Pennsylvania, United States

The cerebral metabolic rate of oxygen consumption (CMRO₂) is an important, tightly regulated, physiologic parameter. A recently reported quantitative MRI method denoted OxFlow has been shown to be able to quantify whole-brain CMRO₂ by simultaneously measuring oxygen saturation (SvO₂) and cerebral blood flow (CBF) in the major vessels draining and feeding the brain in 30 seconds, which is typically adequate for measurement at baseline but not necessarily in response to neuronal activation. Here, we present an improved version of the pulse sequence that quantifies CMRO₂ in 10 seconds scan time and compared it with the parent sequence in eight subjects. Results indicate good agreement between both sequences, with mean differences of 2%, 3%, and 1% for SvO₂, CBF, and CMRO₂, respectively.

Patrick Magrath^1,2, Stanislas Rapacchi², Fei Han^1,2, Peng Hu², J. Paul Finn², and Daniel B. Ennis^1,2
1Bioengineering, University of California, Los Angeles, California, United States, ²Radiology, University of California, Los Angeles, California, United States

The purpose of this study was an initial evaluation of a 4D-flow workflow consisting of: 1) a ventilator-gated 4D-flow acquisition that leverages the regularity of respiratory motion provided in these cases; and 2) 4D flow segementation using a recently developed co-registered, high resolution, ventilator gated, CE-MRA technique made possible through the use of an intravascular contrast agent (Ferumoxytol). These initial results demonstrate that a workflow consisting of ventilator gated 4D flow acquisition coupled with segmentation using the MUSIC MRA sequence made possible by an intravascular contrast agent is feasible and provides a straightforward workflow.

Anthony G. Christodoulou¹, Rebecca Ramb², Marius Menza², Jürgen Hennig², and Zhi-Pei Liang^1
1Beckman Institute and Department of Electrical and Computer Engineering, University of Illinois at Urbana-Champaign, Urbana, IL, United States, ²Department of Radiology, Medical Physics, University Medical Center, Freiburg, Baden-Württemburg, Germany

This work presents a method to accelerate 4D flow imaging using a physics-based image model. This model is generated by integrating computational fluid dynamics into image reconstruction: we solve the Navier-Stokes equations with boundary conditions reconstructed from limited (k,t)-space data, and we reconstruct 4D velocity-encoded images using the Navier-Stokes solution as a constraint. This physics-based constraint complements existing image models that enforce mathematical properties of cardiovascular images (e.g., sparsity, low-rankness) to further enhance the speed and reconstruction quality of 4D flow MRI.

Michael Langham¹, Yongxia Zhou¹, Erica N Chirico¹, Erin K Englund¹, Emile R Mohler², Jeremy F Magland¹, Wensheng Guo³, and Felix W Wehrli^1
1Radiology, University of Pennsylvania, Philadelphia, Pennsylvania, United States, ²Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, United States,³Biostatistics and Epidemiology, University of Pennsylvania, Philadelphia, Pennsylvania, United States

The long-term effects of chronic cigarette smoking on cardiovascular health are well known to reduce smoker’s lifespan by upwards of 10 years. In this study we examined various surrogate markers of endothelial function and vascular reactivity in 169 asymptomatic subjects of varying age, divided into smokers and nonsmokers. Methods included a cuff-occlusion paradigm applied to the thigh to quantify femoral artery and vein hyperemia, the latter quantified in terms of the dynamics of venous oxygen saturation. Central and peripheral pulse-wave velocity was quantified also. The results show functional impairment in smokers independent of age in terms of metrics assessed.

Da Wang^1,2 and Peng Hu^1,2
1Department of Radiological Sciences, David Geffen School of Medicine, University of California Los Angeles, Los Angeles, California, United States, ²Biomedical Physics Interdepartmental Graduate Program, University of California Los Angeles, Los Angeles, California, United States

Phase-contrast MRI (PC-MRI) is a well-established technique for quantification of blood flow and velocity. Quite often, for through-plane Flow Encoding (FE) 2D dynamic PC-MRI, the imaging slice may not be perpendicular to the blood vessel of interest and result in underestimation of peak velocity. Three-directional FE acquisitions may mitigate the underestimation but necessitate a lower temporal-resolution. We assume that the blood flow velocity direction does not change significantly between adjacent cardiac phases. We propose a fast 2D dynamic PC-MRI technique by using a constrain term of velocity direction to calculate flow compensation data using three-directional FE acquisitions Only (FEsO).

Ruyi Bao¹, Qingwei Song¹, Ailian Liu¹, and Zhiyong Li^1
1Radiology department, The First Affiliated Hospital of Dalian Medical University, DaLian, LiaoNing, China

Hemodynamic assessment of pulmonary artery on smokers with 3.0T phase-contrast MR imaging: initial experience To investigate pulmonary blood flow of main pulmonary artery and its changes induced by smoking with phase-contrast MR imaging (PC-MRI).Ten smokers and ten non-smokers, underwent PC-MRI of main pulmonary artery. Main measurement parameters included peak positive velocity, average flow, average positive flow and distensibility. Average flow and average positive flow of main pulmonary artery in smokers group were significantly lower than these and in non-smokers group. Peak positive velocity and distensibility of main pulmonary artery in smokers group were also lower than these in non-smokers group, but there were no statistical differences between them. The findings of this study suggest that PC-MRI may reflect pulmonary blood flow change induced by smoking, which is useful to quantitatively evaluate and monitor pulmonary blood flow change in smokers.

Samir Sarda¹, Joshua J. Chern¹, Nilesh K. Desai², and John Oshinski^2,3
1Pediatric Neurosurgery Associates, Children's Healthcare of Atlanta, Atlanta, Georgia, United States, ²Department of Radiology and Imaging Sciences, Emory University School of Medicine, Atlanta, Georgia, United States, ³Department of Biomedical Engineering, Georgia Institute of Technology, Atlanta, Georgia, United States

In the pediatric population with Type I Chiari malformation (CM-I), cerebrospinal fluid (CSF) flow is known to be obstructed due to the herniation of cerebellar tissue at the level of the foramen magnum. This study examined differences in CSF flow between pediatric patient subjects, adult subjects, and pediatric controls. By utilizing ECG-gated, phase contrast MR to quantify flow, we found that pediatric subjects exhibit higher peak systolic and diastolic CSF flow velocities compared to adults. Additionally, CSF flow dynamics appear to be disrupted only in the pediatric CM-I population with syringomyelia when compared to controls.

Ana Beatriz Solana Sánchez¹, Piero Ghedin², Ek Tsoon Tan³, Christopher J. Hardy³, and Anja Brau^2
1GE Global Research, Garching bei Muenchen, Bayern, Germany, ²GE Healthcare, Garching bei Muenchen, Bayern, Germany, ³GE Global Research, Niskayuna, New York, United States

Phase Contrast MRI provides quantitative blood velocity and net flow measurements, but its accuracy is affected by residual phase errors, limiting the clinical use of this technique in both congenital and acquired heart diseases. Background correction fitting techniques have been proposed to eliminate these errors. However, the appearance of aliased static tissue in the Field of View (FOV) severely impacts the accuracy of any static-tissue fitting algorithm. Here, we quantify this effect for different levels of FOV aliasing and we demonstrate that the use of parallel imaging prevents phase-encoding aliasing and hence improves the robustness of background phase-based correction techniques.

Francesco Santini¹, Oliver Bieri¹, and Tilman Schubert^2
1Radiological Physics, University of Basel Hospital, Basel, Switzerland, ²Department of Radiology, University of Basel Hospital, Basel, Switzerland

We used Doppler Ultrasound and Phase Contrast MRI to quantitatively characterize the frequency content of the blood velocity waveform in different locations of the body. The peripheral vessels had higher frequency content with respect to the central vessels, and we determined that a temporal resolution of 30 to 40 ms is both necessary and sufficient to correctly characterize the signal.

Hui Xue¹, Peter Kellman², Kendall O'Brien³, and Michael Schacht Hansen^1
1Magnetic Resonance Technology Program, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, MARYLAND, United States, ²Medical Image and Signal Processing Program, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, MARYLAND, United States, ³Children's National Medical Center, Washington, DC, United States

Phase contrast measurements can either be acquired with a breath-held segmented acquisition or a free-breathing acquisition with multiple averages to mitigate respiratory motion artifacts. The breath-held techniques require long breath-holds that are poorly tolerated by patients and the multiple average acquisitions are time consuming and suffer from reduced vessel edge sharpness. This work proposes a new approach for free breathing phase contrast velocity measurements where motion correction is explicitly integrated into the reconstruction along with non-linear parallel imaging. The technique provides faster high-resolution free breathing flow quantification.

Satoshi Yatsushiro¹, Akihiro Hirayama², Naokazu Hayashi², Mitsunori Matsumae², Nao Kajihara³, Afnizanfizal Abdullah⁴, and Kagayaki Kuroda^1
1Course of Information Science and Engineering, Tokai University, Hiratsuka, Kanagawa, Japan, ²Department of Neurosurgery, Tokai University School of Medicine, Isehara, Kanagawa, Japan, ³Department of Radiology, Tokai University Hospital, Isehara, Knagawa, Japan, ⁴Faculty of Computer Science and Information Systems, Universiti Teknologi Malaysia, Johor, Malaysia

A novel technique called correlation mapping technique was proposed to characterize pulsatile cerebrospinal fluid (CSF) motion obtained by four dimensional velocity mapping. This technique comprise two functionalities, delay time mapping, which reflects the delay time in the velocity wave propagation, and maximum correlation mapping, which exhibits the similarity of the velocity wave shape. Application of this technique to 5 healthy and 5 normal pressure hydrocephalus (NPH) subjects demonstrated that the percentiles of the area with high correlation coefficients (> 0.7) in the CSF space was significantly (p < 0.05) larger in the normal group.

Can Wu^1,2, Susanne Schnell², and Michael Markl^1,2
1Biomedical Engineering, Northwestern University, Chicago, Illinois, United States, ²Radiology, Northwestern University, Chicago, Illinois, United States

The purpose of this study was to systematically investigate the effects of temporal resolution and different velocity encoding strategies on aortic flow quantification. 2D phase-contrast MRI was performed in 15 volunteers with three different velocity encoding schemes (i.e. single-directional through-plane, interleaved and sequential three-directional velocity encoding). The results demonstrated that high temporal resolution (e.g. ~5ms) may be required for accurately measuring aortic peak velocities and relatively low temporal resolution (~20ms) can be used for aortic blood flow quantification. In addition, aortic peak velocities were underestimated with single-directional velocity encoding compared with 3-directional velocity encoding.

Masataka Sugiyama¹, Yasuo Takehara², Naoki Oishi², Marcus Alley³, Tetsuya Wakayama⁴, Atsushi Nozaki⁴, Hiroyuki Kabasawa⁴, Shuhei Yamashita¹, and Harumi Sakahara^1
1Radiology, Hamamatsu University School of Medicine, Hamamatsu, Shizuoka, Japan, ²Radiology, Hamamatsu University Hospital, Shizuoka, Japan, ³Radiology, Stanford University School of Medicine, California, United States, ⁴Applied Science Laboratory Asia Pacific, GE Healthcare Japan, Tokyo, Japan

To test the hypothesis that initial trigger for atherosclerosis in the lower aortic wall is abnormal hemodynamics, flow dynamics within the non-dilated aorta were measured with 3D cine PC MR imaging (4D Flow) and the data were compared to CT grades of aortic atherosclerosis. Statistical analysis in relation to the aortic wall atheroma, only the oscillatory shear index (OSI) was the significant determinant. Streamline analysis depicted prominent backflow and turbulent flow in the region with higher OSI. The abnormal flow dynamics reflected by high OSI might be a potential trigger to the atherosclerosis of non-dilated lower aorta.

Abdallah G. Motaal¹, Verena Hoerr², Huiming Dong¹, Luc M. J. Florack³, Klaas Nicolay¹, and Gustav J. Strijkers^1
1Biomedical NMR, Department of Biomedical Engineering, Eindhoven University of Technology, Eindhoven, North Brabant, Netherlands, ²Department of Clinical Radiology, University Hospital of Muenster, Muenster, Germany, ³Mathematics and Computer Science, Eindhoven University of Technology, Eindhoven, North Brabant, Netherlands

Signal from short T2 components cannot be detected with conventional MR pulse sequences. Ultrashort echo time MRI makes echo times in the range of 0.01-0.5 ms possible. Here, we introduce a MRI protocol for time resolved 3D UTE Cine Imaging. The sequence involves a retrospectively triggered 3D radial UTE MRI acquisition and compressed sensing reconstruction. 3D time resolved movies with isotropic spatial resolution of 250 µm and 8 frames per cardiac cycle were achieved in 25 minutes. By incorporating flow encoding gradients, flow velocity could be estimated using phase contrast MRI with a very short TE of 0.68 ms.

Gwenael Page¹, Roger Bouzerar¹, Dominique Haye², Dong-Joo Kim³, Hack-Jin Lee³, Anne-Virginie Salsac⁴, and Olivier Baledent^1
1BioFlow Image, CHU Amiens, Amiens, France, ²PFT Innovaltech, France, ³Department of Brain and Cognitive Engineering, Korea, ⁴Laboratoire de Biomecanique et Bioengenierie, CNRS, France

The aim of this study is to compare a high spatial resolution 2D PC-MRI protocol with a 4D PC-MRI protocol in a phantom with calibrated thin tubes and physiological oscillating flows. Protocols were applied on the phantom in a 3T, Achieva dStream, Philips using 32 head coils channels. Accuracy of the flow measurement was then evaluated by calculating the difference between expected and measured flow in each branches. For both sequences we had an error less than 10%, however 2D PC-MRI measurement was more accurate. It is possible to have 4D PC-MRI flow measurement consistent with 2D PC-MRI measurement

Amir Awwad¹, Daniel Rodrieguez¹, Marcus Alley², Shane MacSweeney³, Sebastian Kozerke⁴, and Dorothee P Auer^1
1Sir Peter Mansfield Imaging Centre (SPMIC), University of Nottingham, Nottingham, United Kingdom, ²Radiological Sciences Laboratories, Lucas Centre for Imaging, Stanford University, Palo Alto, California, United States, ³Vascular & Endovascular Surgery Dept., Nottingham University Hospitals NHS Trust, Nottingham, United Kingdom, ⁴Institute of Biomedical Engineering, University and ETH Zurich, Zurich, Switzerland

An in-vitro experimental study comparing derived relative pressure gradients obtained using a 4D flow MRI sequence (3 Tesla) with those measured using an intraluminal fluid-filled pressure catheter readings. Study utilises a custom-made whole body elastic vascular phantom with continuous (stead-flow) pumping of a slippery blood-analogue. Phantom-catheter setup is assembled to be bubble-free, pre-calibrated (zero-level) with dynamic 2nd-order (sinusoidal) response. Derivation of relative pressure change is based on Navier-Stokes Equations of velocity vector-fields (post-processing) in a higher special/temporal resolutions. Experiment results demonstrates the concordant potential in 4D flow MRI to derive non-invasively intravascular relative pressure gradients in continuous-flow dynamics.

Charlotte E Buchanan^1,2, Claire Grant², Eleanor F Cox¹, Nick M Selby^2,3, Chris W McIntyre^2,4, Maarten W Taal², and Susan T Francis^1
1SPMIC, University of Nottingham, Nottingham, Nottinghamshire, United Kingdom, ²Division of Medical Sciences and Graduate Entry Medicine, Royal Derby Hospital, Nottingham, United Kingdom, ³Department of Renal Medicine, Royal Derby Hospital, Derby, United Kingdom, ⁴Schulich School of Medicine and Dentistry, University of Western Ontario, London, Ontario, Canada

This study evaluates cardiac function in patients with chronic kidney disease (CKD) and chronic liver disease (CLD) in response to a handgrip challenge, which may be more effective at discriminating normal from abnormal cardiovascular function. Phase contrast imaging of the aorta was performed during a handgrip challenge to determine aortic strain, aortic velocity, cardiac output, aorta area and stroke volume. Aortic strain decreased significantly in CKD and HC on exercise. Stroke volume was found to significantly increase on exercise for patients with CLD.

Merih Cibis¹, Kelly Jarvis^2,3, Michael Markl^2,3, Michael Rose^2,4, Cynthia Rigsby^2,4, Alex J Barker², and Jolanda J Wentzel^1
1Biomedical Engineering, Erasmus MC, Rotterdam, Netherlands, ²Radiology, Northwestern University, Chicago, Illinois, United States, ³Biomedical Engineering, Northwestern University, Chicago, Illinois, United States, ⁴Medical Imaging, Ann& Robert H Lurie Children’s Hospital of Chicago, Chicago, Illinois, United States

Viscous dissipation inside Fontan circuit might be associated with the exercise performance of Fontan patients. Information on viscous dissipation might therefore be used in clinical assessment of these patient and can be calculated in-vivo by using flow MRI data. However the magnitude of viscous dissipation might depend on the spatial resolution of the MRI data. We investigated the influence of resolution on the estimated viscous dissipation. By using computational fluid dynamics (CFD) simulations, we showed that decreased spatial resolution results in underestimated viscous dissipation. However we were able to capture the differences in viscous dissipation between patients also at lower spatial resolution.

Hojin Ha¹, Guk Bae Kim², Jihoon Kweon², Young-Hak Kim³, Namkug Kim^4,5, Dong Hyun Yang⁴, and Sang Joon Lee^1
1Department of Mechanical Engineering, Pohang University of Science and Technology, Pohang, Gyeongbuk, Korea, ²Asan Institute of Life Science, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Korea, ³Department of Cardiology, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Korea, ⁴Department of Radiology, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Korea, ⁵Department of Convergence Medicine, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Korea

The present study aims to improve accuracy and noise level of four-dimensional velocity encoding (VENC) MRI by employing multiple VENC parameters. A 4D PC-MRI sequence with five different VENC parameters was employed to measure three dimensional flow field in a stenosis flow phantom. The velocity field from larger VENC value was combined with that of smaller VENC value unless the velocity data are lost by phase-aliasing and phase dispersion. Results showed that multi-VENC overlapping significantly increased accuracy of the flow measurement with reduction of the noise, and wide velocity dynamics range.

Henrik Haraldsson¹, Sarah Kefayati¹, Belén Casas Garcia², Jonas Lantz², Tino Ebbers², and David Saloner^1
1University of California, San Francisco, San Francisco, California, United States, ²University of Linkoping, Sweden

Turbulence is an important factor in hemodynamics, and has been associated with several pathological conditions including energy losses and thrombogenic platelet deformations. Quantification of the full turbulence tensor allows for more complex hemodynamic parameters such as pressure losses and shear stresses that has been linked to platelet activation. In this work we show that using an ICOSA6 flow encoding and determining the turbulence stress tensor by solving the lease-square-problem we are able to estimate turbulent shear components with lower noise sensitivity and without compromising the result of the normal turbulence components.

Hanieh Mirzaee¹ and Anja Hennemuth^1
1Fraunhofer MEVIS, Bremen, Bremen, Germany

Cardiovascular pressure gradients are an important clinical marker for the evaluation of the severity of cardiovascular diseases. The pressure estimation techniques based on the MRI velocity data demand accurate segmentation in order to set the boundary conditions. Segmentation is difficult to perform on these images due to minimal contrast between blood and vascular anatomy. Alternatively, imaging sequences such as 3D whole heart (anatomy) are more suitable for geometry extraction. In this work, we propose to enhance noninvasive pressure measurements through a fused anatomy and flow image data.

Stanislas Rapacchi^1,2, Yutaka Natsuaki³, Paul J Finn², Gerhard Laub⁴, Daniel Ennis², and Peng Hu^2
1CRMBM, Aix-Marseille University, Marseille, France, ²Radiology, UCLA, los angeles, CA, United States, ³Siemens, Los Angeles, CA, United States, ⁴Siemens, CA, United States

4D Phase-Contrast MRI (“4D Flow”) has become a choice modality for assessment of flow patterns in complex vasculatures, such as congenital heart diseases. For thoracic imaging, the technique relies on ECG-triggering and an imaging navigator to compensate for breathing motion. The acquisition of the navigator requires switching off the 4D flow acquisition, thus breaking the steady-state and missing a phase in the cardiac cycle. We propose to use double gating from ECG and a pressure-driven belly-belt to monitor breathing motion while maintaining continuous steady-state.

Lizhen Cao¹, Zhiyuan Dong¹, Aur¨¦lien F. Stalder², Xiangying Du¹, Tianjing Zhang³, Andreas Greiser², and Kuncheng Li^1
1The Department of Radiology, Xuanwu Hospital of Capital Medical University, Beijing, China, China, ²Siemens AG Healthcare Sector, Erlangen, Germany, ³Siemens MR Northeastern Collaboration, Beijing, China, China

Thoracic aorta flow sensitive 4D MR imaging in hypertension