Myocardial Tissue Differentiation

Thursday 4 June 2015

Bhairav Bipin Mehta¹, Michael Salerno^1,2, and Frederick H Epstein^1,3
1Department of Biomedical Engineering, University of Virginia, Charlottesville, Virginia, United States, ²Department of Medicine, Cardiology Division, University of Virginia, Charlottesville, Virginia, United States, ³Department of Radiology and Medical Imaging, University of Virginia, Charlottesville, Virginia, United States

Assessment of fibrosis in thin structures such as the right ventricular (RV) and left atrial walls would be valuable in various disorders. However, current cardiac T1 mapping techniques have limited spatial resolution for these applications. We previously developed a 2D sequence, ANGIE, which enables high-resolution T1 mapping. However, high resolution 2D imaging is limited by low SNR and time efficiency. In the present study, we extend ANGIE to perform high-resolution three-dimensional (3D) T1 mapping of the heart within a clinically acceptable scan time. 3D ANGIE was evaluated by performing comparisons with 2D ANGIE and 2D MOLLI in six healthy volunteers.

Wei Li¹, Eugene Dunkle², Claire Feczko³, Shivraman Giri⁴, and Edelman R Robert^1
1Northshore University HealthSystem, Evanston, IL, United States, ²Northshore University HealthSystem, IL, United States, ³Northshore University HealthSystem, Evanston, IL, United States, IL, United States, ⁴Siemens Healthcare, Chicargo, IL, United States

To assess the effects of T1 mapping and ECV calculating by adding MOLLI T1 mapping sequence to our routine cardiac MR work up protocol, eighty-eight patients, who had pre- and post- Gd contrast MOLLI acquisitions for T1 mapping and ECV calculation in addition to routine sequences for cardiac work up, were reviewed. Only mid-short axis and mid-4chember views were acquired for T1 mapping to avoid increasing the exam time too much. We observed that adding MOLLI T1 mapping sequence to a cardiac work up protocol is a simple and effective way to catch T1 and myocardial ECV message even only limited planes were applied.

Kyle Erjin Jeong^1,2, Kyungpyo Hong^1,2, and Daniel Kim^2,3
1Bioengineering Department, University of Utah, Salt Lake City, Utah, United States, ²Utah Center for Advanced Imaging Research, University of Utah, Salt Lake City, Utah, United States, ³Department of Radiology, University of Utah, Utah, United States

Arrhythmia insensitive rapid (AIR) cardiac T1 mapping is particularly useful for imaging patients with arrhythmia and/or rapid heart rates. The original AIR pulse sequence was designed with saturation recovery time delay (TD) of 600 ms. We sought to perform numerical simulation to derive an optimal TD to achieve a good balance between precision and scan time. Our numerical experiments show that optimal TD is 780 ms for expected T1 range of 400-2,000 ms at 3T. This simulation framework is adaptable for patients with different heart rates.

Ivan Cokic¹, Avinash Kali¹, Hsin-Jung Yang¹, Raymond Yee², Richard Tang¹, Mourad Tighiouart³, Xunzhang Wang⁴, Warren M. Jackman⁵, Sumeet S. Chugh⁴, James A. White⁶, and Rohan Dharmakumar^1
1Biomedical Sciences - BIRI, Cedars-Sinai Medical Center, Los Angeles, California, United States, ²Department of Medicine - Division of Cardiology, London Health Sciences Centre, London, ON, Canada, ³Biostatistics and Bioinformatics Research Center, Cedars-Sinai Medical Center, Los Angeles, California, United States, ⁴Cedars-Sinai Heart Institute, Cedars-Sinai Medical Center, Los Angeles, California, United States, ⁵Heart Rhythm Institute, University of Oklahoma, Oklahoma City, OK, United States, ⁶Department of Cardiac Sciences, University of Calgary - Stephenson Cardiac Imaging Centre, Calgary, AB, Canada

This study will help to improve the risk stratification for ICD implantation.

Markus Henningsson¹, Rene Botnar¹, and Tobias Voigt^1,2
1Division of Imaging Sciences and Biomedical Engineering, King's College London, London, United Kingdom, ²Clinical Research Europe, Philips Research, Hamburg, United Kingdom

Free-breathing whole-heart myocardial T1 mapping is technically challenging due to the difficulty of respiratory motion compensation. Conventional respiratory gating techniques are impractical due to the prolonged scan time associated with these approaches. In this work we propose an image-based navigator approach which allows for direct tracking of the respiratory motion of the heart. Although this means navigators at different inversion times have different contrast, potentially leading to registration errors, we solve this problem by using a separate navigator reference for each inversion time. We evaluate this free-breathing three-dimensional T1-mapping method in 7 healthy volunteers and compared it to 2D T1-mapping.

Eugene G. Kholmovski¹, Ravi Ranjan², Joshua Silvernagel², and Nassir F. Marrouche^2
1UCAIR, Department of Radiology, University of Utah, Salt Lake City, Utah, United States, ²CARMA Center, University of Utah, Salt Lake City, Utah, United States

Catheter radio-frequency (RF) and cryo-ablations are being increasingly used for treatment of atrial fibrillation and ventricular tachycardia. However, reported success rate of the procedures is moderate. The main causes of procedure failure are tissue recovery and gaps in ablation. The extent of ablations and lesion permanency cannot be accurately evaluated by conventional electro-physiological measurements. MRI can be used to assess lesions dimension and to predict their permanency. In this study, we have compared visibility of acute cryo and RF ablation lesions using non-contrast and contrast enhanced MRI techniques. Cryo and RF lesions have similar appearance in post-contrast T1w and LGE MRI images. However, the lesions have significantly different appearance in native T1w and T2w images.

Jie Zheng¹, Qian Yin¹, David Muccigrosso¹, Ridong Chen², and Dana Abendschein^3
1Radiology, Washington University School of Medicine, Saint Louis, Missouri, United States, ²APT Therapeutics, Saint Louis, Missouri, United States,³Cardiology Division, Washington University School of Medicine, Saint Louis, Missouri, United States

A non-contrast MRI fibrosis index method was developed to consecutively assess myocardial fibrosis in a canine model after induced myocardial infarction, reperfusion, and an experimental drug APT102 treatment. In comparison with controlled study, the fibrosis index demonstrated similar trend from baseline to 24 hours after APT102 treatments, but dramatically decreased at 7 days. This was confirmed by histopathological stains in myocardial tissues.

Joep van Oorschot¹, Hamza El Aidi¹, Fredy Visser², Peter Luijten¹, Tim Leiner¹, and Jaco Zwanenburg^1
1University Medical Center Utrecht, Utrecht, Utrecht, Netherlands, ²Philips Healthcare, Best, Noord-Brabant, Netherlands

T1ρ-mapping shows promising results for the endogenous detection of myocardial fibrosis. A T1ρ-map can be obtained by multiple T1ρ-weighted images with different spin-lock (SL) preparation times. We propose a cardiac T1ρ-mapping method in a single breath-hold. Five healthy volunteers and 2 patients with chronic myocardial infarction were scanned. T1ρ-mapping was performed by acquiring 5 images with different SL preparation times (SL = 0, 10, 20, 30, 40 ms). High resolution T1ρ-maps were successfully acquired in all subjects. The mean T1ρ-relaxation time was 52.7±3 ms. Coefficient of repeatability in the healthy subjects was 3.74 ms, and in infarct patients 4.59 ms.

Andrea S. Dell'Aquila¹, Sofia A. Papadopoulou¹, Sanjay Sharma¹, Lisa J. Anderson¹, and Taigang He^1
1Cardiovascular Sciences Research Centre, St George's, University of London, London, Greater London, United Kingdom

Increased left ventricle trabeculations (LVT) are frequently seen in various cardiac diseases. LVT can be accurately delineated from cardiovascular magnetic resonance (CMR) images, but methods to quantify them remain limited. Fractal analysis of CMR images was recently proposed and proved to be an effective quantitative measure in distinguishing normal from hypertrophic cardiomyopathy. In this study, we quantified LVT using fractal dimension in a large cohort of normal controls, athletes with and without athlete heart syndrome, patients with non-compaction, hypertrophic and non-ischaemic dilated cardiomyopathy. Of clinical importance, this study demonstrates a wide spectrum of LVT patterns between normal cardiac conditions and abnormal ones, such as hypertensive heart disease, hypertrophic cardiomyopathy and left ventricular non-compaction.

Satoshi Kawanami¹, Michinobu Nagao¹, Masato Yonezawa², Yuzo Yamasaki², Takeshi Kamitani², Torahiko Yamanouchi², Tomomi Ide³, Ryohei Funatsu⁴, Hidetake Yabuuchi⁵, and Hiroshi Honda^2
1Molecular Imaging & Diagnosis, Kyushu University, Graduate School of Medicine, Fukuoka, Fukuoka, Japan, ²Clinical Radiology, Kyushu University, Graduate School of Medicine, Fukuoka, Fukuoka, Japan, ³Cardiovascular Medicine, Kyushu University, Graduate School of Medicine, Fukuoka, Fukuoka, Japan, ⁴Radiological Technology, Kyushu University Hospital, Fukuoka, Fukuoka, Japan, ⁵Health Sciences, Kyushu University, Graduate School of Medicine, Fukuoka, Fukuoka, Japan

In this study, we analyzed T2* value in the mid-left ventricular septum avid both normoxia and hyperoxia among clinical cases with non-ischemic heart disease. The oxygen-enhanced T2* cardiac magnetic resonance (CMR) was promising to evaluate the myocardial blood-oxygen dependent (BOLD) response to hyperoxia. The development of quantitative evaluation technique for the oxygen metabolism in human myocardium in vivo has opened up new avenues for the study of the cardiac pathophysiology. To our knowledge, this is the first clinical study that has assessed the myocardial ĢT2* response to hyperoxic respiratory challenge by BOLD-CMR.

David A. Reiter¹, Mustapha Bouhrara¹, and Richard G Spencer^1
1Laboratory of Clinical Investigation, NIH/National Institute on Aging, Baltimore, MD, United States

The creatine kinase (CK) reaction is considered to be important for maintaining ATP supply to demand in the myocardium over the wide operating range of cardiac output. The CUPS method has been previously applied to measurements of chemical exchange rates, such as CK flux, under preclinical experimental conditions. The current work uses Cramér-Rao lower bound analysis and Monte Carlo simulations, targeting in vivo human MRS conditions for studies of the myocardium at clinical field strength, demonstrating the feasibility of this approach for reliable measurement of CK fluxes under a wide range of experimental and physiological conditions.

Pan-Ki Kim¹, Joonsung Lee¹, and Byoung Wook Choi^1
1Yonsei University, Seoul, Korea

The quantification of T1 relaxation time has become an important indicator for diffuse cardiomyopathies. In small animal studies, such as mouse and rat, fast heart beats and respiratory rates are major obstacles to use clinical T1 mapping methods. For small animal T1 mapping, SALLI and mCINE-IR had been reported using the Look-Locker scheme. Since the Look-Locker recovery evolution has to be consistently maintained, the respiratory gating is particularly challenging. In general, the multiple averages were applied to avoid motion artifacts. In this study, respiratory motion navigated Look-Locker imaging (NALLI) was proposed to overcome respiratory motion artifacts for small animal myocardial T1 mapping. To evaluate feasibility, the proposed method was performed for phantoms and a normal mouse.

Keigo Kawaji¹, Akiko Tanaka², Mita Patel¹, Sui-Cheng Wang³, Hui Wang⁴, Takeyoshi Ota², Roberto M. Lang¹, and Amit R. Patel^1
1Medicine, Section of Cardiology, The University of Chicago, Chicago, Illinois, United States, ²Surgery, The University of Chicago, Chicago, Illinois, United States, ³Biomedical Engineering, Northwestern University, Evanston, Illinois, United States, ⁴Philips Medical Systems, Cleveland, Ohio, United States

In this study, we propose a novel ECG-triggered and navigator-gated 3D LGE acquisition method that uses CENTRA-PLUS profile/view ordering that is designed to acquires central and peripheral k-space with a weighted navigator gating window. Six pigs were surgically implanted with Right Ventricular (RV) patches, and MRI was performed 7-9 weeks after surgery. RV patch volume quantification was performed using both 2D PSIR LGE and the proposed 3D PSIR LGE images, and volume measurement errors were compared to the gold standard direct measurements made on the surgically excised RV patch region.

Martijn Froeling^1,2, Gustav J Strijkers³, Aart J Nederveen², and Peter R Luijten^1
1Radiology, UMC Utrecht, Utrecht, Netherlands, ²Radiology, AMC, Amsterdam, Netherlands, ³Biomedical engineering and physics, AMC, Amsterdam, Netherlands

The aim of this study was to develop SE-based cardiac diffusion MRI protocol with second order moment nulling, thus also compensating for acceleration, and to compare its performance to that of Stejskal-Tanner and bipolar gradients waveforms. Using this approach we have shown that it is feasible to quantify the transmural helix angle for the entire heart in five subjects.

Arvin Arani¹, Shivaram Poigai Arunachalam¹, Phillip Rossman¹, Armando Manduca², David S. Lake¹, Joshua D. Trzasko¹, Kiaran P. McGee¹, Kevin J Glaser¹, Richard L. Ehman¹, and Philip Araoz^1
1Radiology, Mayo Clinic, Rochester, Minnesota, United States, ²Physiology and Biomedical Engineering, Mayo Clinic, Rochester, Minnesota, United States

Normal cardiac function is dependent on the mechanical properties of the myocardium, which may serve as a valuable predictor of disease. Several groups have investigated the use of MR Elastography to measure stiffness in vivo. However, a thorough evaluation of the accuracy of the reported stiffness measurements and the optimal experimental settings has not been well established. The purpose of this work, was to evaluate the accuracy of 3D MRE direct inversion at multiple frequencies, on curl wave fields in a morphologically accurate left ventricular phantom using dynamic mechanical testing (DMA) as the reference standard.

Kyungpyo Hong^1,2, Eun-Kee Jeong¹, and Daniel Kim^1
1UCAIR, Department of Radiology, University of Utah, Salt Lake City, Utah, United States, ²Department of Bioengineering, University of Utah, Salt Lake City, Utah, United States

Standard cardiac T₁ mapping generates image artifacts in patients with implantable cardioverter-defibrillator (ICD). We propose to perform successful cardiac T₁ mapping in patients with ICD by incorporating a saturation radio-frequency pulse with wide frequency bandwidth (8.9 kHz). In 11 subjects with ICD, compared with T₁ measurements produced by standard T₁ mapping without ICD as the control, T₁ measurements by standard T₁ mapping with ICD were significantly different (p < 0.05), whereas T₁ measurements by wideband T₁ mapping with ICD were not significantly different. This study demonstrates the feasibility of wideband cardiac AIR T₁ mapping for suppressing artifacts induced by ICD.

Adrian Lam¹, Ankit Parikh², Michael Lloyd², and John Oshinski^1,3
1Department of Biomedical Engineering, Georgia Institute of Technology, Atlanta, Georgia, United States, ²Department of Medicine, Emory University, Georgia, United States, ³Department of Radiology and Imaging Science, Emory University, Georgia, United States

Optimal patient response to Cardiac Resynchronization Therapy is associated with left ventricular lead implantation in the latest contracting segment of the myocardium that is not predominantly myocardial scar. However, transvenous implantation through the coronary sinus limits lead implantation locations to regions that can be accessed by the coronary veins. A 3D, contrast-enhanced, free-breathing inversion recovery (IR) FLASH sequence can visualize the coronary veins and can be used for late gadolinium enhanced (LGE) imaging. The purpose of this study was to evaluate the potential of 3D, LGE, free-breathing, IR-prepared FLASH compared to the clinical standard, 2D LGE Phase Sensitive Inversion-Recovery (PSIR).

Kyungpyo Hong^1,2 and Daniel Kim^1
1UCAIR, Department of Radiology, University of Utah, Salt Lake City, Utah, United States, ²Department of Bioengineering, University of Utah, Salt Lake City, Utah, United States

Arrhythmia-insensitive-rapid (AIR) cardiac T₁ mapping pulse sequence is a promising method for imaging patients with rapid heart rates and/or arrhythmia. We sought to improve further AIR cardiac T₁ mapping through k-space ordering and flip angle optimization. In human volunteers and patients, centric k-space ordering acquisitions yielded image artifacts arising from eddy currents, whereas centric-pair k-space ordering acquisitions suppressed image artifacts. For flip angles ranging from 25-65°, standard deviation decreased with flip angles, with significant difference in mean T₁ values for any of the 4 types of measurement (native myocardial T₁, native blood T₁, post-contrast myocardial T₁, and post-contrast blood T₁).

Stephen Henry Gilbert¹, Julie Magat², Mark Trew³, Valery Ozenne², Fanny Vaillant², Jérôme Naulin², Olivier Bernus², and Bruno Quesson^2
1Mathematical Cell Physiology, Max Delbrück Center for Molecular Medicine, Berlin, Germany, ²L'Institut de rythmologie et modélisation cardiaque LIRYC, Pessac, France, ³Auckland Bioengineering Institute, Auckland, New Zealand

We present first high-spatial resolution ex vivo MRI of myocardial sheetlet structure from a human sized heart (the sheep heart), obtained at 9.4T/30cm, using a novel 7 elements array coil. This image data allowed us to reconstruct the ventricular sheetlet and myocyte orientations using structure tensor image analysis. We demonstrate that this methodology that has only previously been applied to small mammalian hearts is scalable to large animal clinical models. This methodology allows correlation of explanted myocardial structure to in vivo and myocardial strain/shear distributions in large animal clinical models and is directly applicable to normal/pathologic human hearts after transplantation.

Nilesh R Ghugre^1,2, Xiuling Qi¹, Jennifer Barry¹, Bradley H Strauss³, and Graham A Wright^1,2
1Physical Sciences Platform, Sunnybrook Research Institute, Toronto, ON, Canada, ²Department of Medical Biophysics, University of Toronto, Toronto, ON, Canada, ³Schulich Heart Program, Sunnybrook Health Sciences Centre, Toronto, ON, Canada

Myocardial hemorrhage is a frequent complication in acute myocardial infarction; however its impact on the remote tissue is currently unknown. The aim of our study was to investigate whether hemorrhage contributes to a remote myocardial response using a novel porcine model of myocardial hemorrhage. To this end, a blood-oxygen-level-dependent (BOLD) approach with T2 contrast was employed to evaluate vasodilator function using a stress agent. We noted an elevated remote T2 in the rest state that was associated with a suppressed stress response under hemorrhagic conditions. Our study demonstrates that hemorrhage not only promotes cellular and microvascular damage and inflammation but may also further be responsible for remote myocardial remodeling post-infarction.

Yijen Lin Wu^1,2, Fang-Cheng Yeh³, and Chien Ho^4
1Developmental Biology, University of Pittsburgh, Pittsburgh, PA, United States, ²Rangos Research Center Imaging Core, Children's Hospital of Pittsburgh of UPMC, Pittsburgh, PA, United States, ³Psychology, Carnegie Mellon University, Pittsburgh, PA, United States, ⁴Biological Sciences, Carnegie Mellon University, Pittsburgh, PA, United States

Coronary heart disease is the leading cause of death in US. Timely re-circulation has greatly decreased 30-day in-hospital death, and increased survival rate during the acute phase of myocardium infarction (MI). However, even with successful blood-flow restoration, ischemia reperfusion injury (IRI) can result in greater tissue damage and adverse remodeling. The rate of developing post-MI heart failure increases as the acute mortality decreases. Long-term post-MI ventricular remodeling and the consequent potential heart failure remains a challenge. We report here that Intralipid can preserve heart function after IRI can reduce long-term post-MI ventricular remodeling.

Byoung Wook Choi¹, Yoo Jin Hong¹, Chul Hwan Park¹, and Panki Kim^1
1Radiology, Yonsei University, Seoul, Korea

Cardiotoxicity of doxorubicin in patients with cancer causes irreversible myocardial damage leading to worse prognosis. To detect early stage of cardiotoxicity, T1 mapping and extracellular volume fraction was tested to determine the feasibility for detection of diffuse myocardial fibrosis in an early stage of cardiotoxicity. There was a good correlation between myocardial ECV measurement and the degree of fibrosis on histology.

Dana C Peters¹, Daniel Cornfeld¹, Albert J Sinusas², James S Duncan¹, Xenios Papademetris¹, Karl Grunseich¹, and Sudhakar Chelikani^1
1Radiology, Yale School of Medicine, New Haven, CT, United States, ²Cardiology, Yale School of Medicine, New Haven, CT, United States

We tested the hypothesis that left atrial strain was inversely correlated to atrial fibrosis. Left atrial strain was assessed using 2-chamber and 4-chamber cine images in sixteen patients with AF, and 13 control patients, using atrial contouring and a point-matching algorithm. Left atrial late gadolinium enhancement (LGE) was also assessed to obtain a fibrosis score. Average 4-chamber and 2-chamber strains were inversely correlated to the extent of atrial fibrosis in AF subjects, but not in controls. Regional strains also showed strong correlations to LGE.

Lei Zhao¹, Xiaohai Ma¹, Songnan Li², Tianjing Zhang³, Jing An³, Greiser Andreas⁴, and Zhanming Fan^1
1Radiology, Anzhen Hospital, Capital Medical University, Beijing, Beijing, China, ²Cardiology, Anzhen Hospital, Capital Medical University, Beijing, Beijing, China, ³MR Collaborations NE Asia, Siemens Healthcare, Beijing, China, Beijing, China, ⁴Siemens Healthcare, Erlangen, Germany

Unlike focal fibrosis, diffuse myocardial fibrosis is not visualized on delayed enhancement magnetic resonance imaging (MRI), but can be quantified with extracellular volume fraction (ECV) mapping. In atrial fibrillation (AF), it may be induced by arrhythmia or reflect pre-existing cardiomyopathy. Ten subjects (8 persistent AF patients and 2 controls) underwent cardiac MR examination which included T1 mapping sequence. Our results reveal that ECV mapping identifies diffuse LV fibrosis in patients with AF.

Cardiac Perfusion & Function

Thursday 4 June 2015

Shokoufeh Golshani¹ and Abbas Nasiraei Moghaddam^1,2
1BME, Amirkabir University of Technology (Tehran Polytechnic), Tehran, Tehran, Iran, ²Radiology, UCLA, Los Angeles, CA, United States

In radial tagging, the information of taglines required for strain imaging is located on a donut shaped region in k-space that can be collected efficiently through radial sampling sequence. In this study, we investigated the efficiency of radial data acquisition for strain imaging in conjunction with an adapted PFT reconstruction algorithm; a Hankel-based transform used for polar data. Both phantom and human images showed high robustness of PFT reconstruction against reduced number of radial data lines. This can be exploited towards real-time imaging, which is necessary in some cardiac assessments such as stress test. In addition, its implementation is adequately fast for in-line reconstruction.

Abram Makram¹, Ayman Khalifa¹, Hossam El-Rewaidy², Ahmed Fahmy², and El-Sayed H. Ibrahim^3
1Helwan University, Cairo, Egypt, ²Nile University, Cairo, Egypt, ³University of Michigan, Ann Arbor, MI, United States

Quantification of the global cardiac function from the tagged images is challenging due to the anatomy obstruction by the taglines and low myocardium-to-blood contrast. In this work, a method is presented for customized tagline removal and improving myocardium-to-blood contrast. The proposed method is applied on images from 12 patients to calculate global cardiac functional parameters and compare them to those from cine images. The results show that the proposed method can be used to accurately estimate ventricular volume and mass, which would allow for reduced scan-time through measuring both regional and global cardiac function parameters from one set of images.

Kai Lin¹, Robert A Gordon², Keith H Benzuly², Clyde W Yancy², Jon W Lomasney², Vera H Rigolin², Allen S Anderson², Michael Markl¹, and James C Carr^1
1Radiology, Northwestern University Feinberg School of Medicine, Chicago, IL - Illinois, United States, ²Cardiology, Northwestern University Feinberg School of Medicine, Chicago, IL - Illinois, United States

Our findings suggest that tissue phase mapping with low temporal resolution may serve as an alternative in quantifying cardiac motion for subjects who can not hold the breath for a long time.

Zhe Wang^1,2, Ziwu Zhou^1,2, Yi Wang³, Peng Hu^1,2, and Daniel B. Ennis^1,2
1Radiological Science, University of California, Los Angeles, CA, United States, ²Bioengineering, University of California, Los Angeles, CA, United States,³Neurology, University of California, Los Angeles, CA, United States

Current clinical protocols for measuring LV twist require two separate breath holds to acquire tagged images at the LV apex and LV base, which subjects estimates of LV twist to breath hold differences. Multi-slice excitation (CAIPIRINHA) permits combining these two acquisitions into a single scan. We implemented the CAIPIRINHA technique with a tagging sequence, acquired data in five healthy volunteers, and compare the results with traditional two breath-hold LV twist results. Linear regression shows a high correlation between the two (r=0.99) and LV twist Bland-Altman result shows a small bias of 0.21° with narrow 95% confident intervals of [-0.47°, 0.89°].

Tomoyoshi Kimura^1,2, Hideki Ota¹, Koichiro Sugimura³, Kazuomi Yamanaka¹, Tatsuo Nagasaka¹, Hiroaki Shimokawa³, Kei Takase¹, and Haruo Saito^2
1Radiology, Tohoku University Hospital, Sendai, Miyagi, Japan, ²Graduate School of Medicine, Tohoku University, Sendai, Miyagi, Japan, ³Cardiology, Tohoku University Hospital, Sendai, Miyagi, Japan

We enrolled 11 patients with chronic thromboembolic pulmonary hypertension and one with idiopathic pulmonary arterial hypertension. Biventricular function and left ventricular (LV) strain analyses were performed before and after treatment. MR-derived parameters of cardiac functions were compared with mean pulmonary arterial pressure (mPAP). All patients demonstrated normal LV ejection fraction at baseline. Right ventricular (RV) end-systolic and end-diastolic volume index were correlated with mPAP at baseline; none of the strain parameters were correlated. However, improvements of circumferential and radial strain in inferolateral segment as well as RV end-systolic volume index were independently correlated with reduction of mPAP.

Xiaoxia Zhang^1,2, Nikhil Jha^1,2, Himanshu Gupta³, Nouha Salibi^2,4, and Thomas Jr. Denney^1,2
1Department of Electrical and Computer Engineering, Auburn University, Auburn, AL, United States, ²AU MRI Research Center, Auburn University, Auburn, AL, United States, ³Department of Medicine, Division of Cardiovascular Disease, University of Alabama at Birmingham, Birmingham, AL, United States, ⁴MR R&D, Siemens Healthcare, Malvern, PA, United States

Cardiac MRI provides important structural and functional information. However routine quantitative analysis of cine MRI is often restricted to left and right ventricular ejection fraction measurements. Functional assessment of left and right atria is frequently not performed. In this abstract, we present an image acquisition and analysis framework for comprehensive evaluation of all cardiac chambers simultaneously, which leverages limited human input to optimize quantitative functional output and allows for evaluating bi-atrial and biventricular interactions simultaneously that may be important in various disease states.

Mai Wael¹, El-Sayed H. Ibrahim², and Ahmed Fahmy^1
1Nile University, Cairo, Egypt, ²University of Michigan, Ann Arbor, MI, United States

A method is presented for detecting regional wall motion abnormality based on capturing the variation in myocardial thickness during the cardiac cycle from standard cine MRI images. The extracted wall thickness patterns are normalized relative to the average epicardial radius, mapped to lower dimensions using principal component analysis, and then classified into normal or abnormal using the maximum likelihood criterion with leave-one-out method. The developed method provides automatic assessment of regional abnormality for each segment in each slice; therefore, it could be a valuable tool for automatic and fast determination of regional wall motion abnormality from conventional untagged cine images.

Jiming Zhang¹, Amol Pednekar², Jie Chen¹, Claudio Arena¹, Debra Dees¹, Benjamin Cheong¹, and Raja Muthupillai^1
1Diagnostic and Interventional Radiology, CHI St Luke's Health, Houston, TX, United States, ²Philips Healthcare, Houston, TX, United States

Fifty percent of the heart failure patients have diastolic dysfunction with preserved ejection fraction (HFpEF), and traditionally echocardiography is used for evaluating diastolic dysfunction due to its superior temporal resolution. We propose LV volume based metrics as a MR based index that is comparable to conventional echo based index for the evaluation of diastolic dysfunction. The results from this pilot study suggest that it is feasible to evaluate both systolic and diastolic function with a single set of high-temporal resolution cine SSFP images acquired during free breathing.

Subashini Srinivasan^1,2, Randall M Kroeker³, Adam Plotnik¹, Simon Gabriel¹, Nancy Halnon⁴, Peng Hu¹, J. Paul Finn¹, and Daniel B Ennis^1,2
1Department of Radiological Sciences, University of California, Los Angeles, California, United States, ²Department of Bioengineering, University of California, Los Angeles, California, United States, ³Siemens Healthcare, Malvern, Pennsylvania, United States, ⁴Department of Pediatrics, University of California, Los Angeles, California, United States

Breath-hold, segmented, constant flip angle (BH-CFA) bSSFP cardiac cine imaging can be challenging at 3T in patients with poor breath holding capacity which results in motion artifacts. Real-time, free-breathing imaging with retrospective reconstruction can be SAR intensive, which may limit image quality. The objective of this study was to develop and evaluate a free breathing variable FA (FB-VFA) cardiac cine imaging technique with reduced SAR at 3T for improved comfort and compliance. FB-VFA reduced the SAR by 25% compared to BH-CFA and maintained blood-myocardium contrast and image quality sufficient to perform global and regional cardiac functional analysis.

Peng Lai¹, Joseph Y Cheng², Shreyas S Vasanawala², and Anja CS Brau^3
1Global MR Applications & Workflow, GE Healthcare, Menlo Park, CA, United States, ²Radiology, Stanford University, CA, United States, ³Global MR Applications & Workflow, GE Healthcare, Munich, Germany

Conventional 2D cine requires long scan time and is not reformattable. Breathhold 3D cine suffers from reduced contrast and increased SSFP banding artifacts. This work developed a novel technique for free-breathing whole-heart cine imaging. Our results show that this new method is more robust than breathhold 3D cine and enables 3D imaging with isotropic resolution and thus offline reformatting.

Daniel James Stuckey¹, Thomas A Roberts¹, Laurence H Jackson¹, Rajiv Ramasawmy¹, Valerie Taylor¹, Anna L David², Bernard Siow*¹, and Mark F Lythgoe*^1
1Centre for Advanced Biomedical Imaging, UCL - University College London, London, United Kingdom, ²Institute for Women’s Health, UCL - University College London, London, United Kingdom

We show for the first time that 1T “benchtop” MRI can assess myocardial viability as well as contraction in a rat model of myocardial infarction. Results compared well with high field MRI and ultrasound. Although the speed with which ultrasound acquisitions can be acquired make it an effective choice for rapidly measuring cardiac contractility, the ability of MRI to rapidly quantify infarct size using late gadolinium enhancement, as well as RV function, demonstrates that low field MRI systems can provide essential information beyond what is possible from ultrasound, making it a valuable tool for studying experimental myocardial infarction and therapy.

Azza Hassanein¹, Ayman Khalifa¹, and El-Sayed H. Ibrahim^2
1Helwan University, Cairo, Egypt, ²University of Michigan, Ann Arbor, MI, United States

HARP is a commonly used technique for analyzing tagged images. Nevertheless, HARP tracking of the material points fails near the myocardial boundaries. Band-Pass Optical Flow (BPOF) is a motion analysis technique that, similar to HARP, is based on k-space analysis, but enforces optical flow constraints, which could provide a remedy to the HARP limitation. The purpose of this work is to use BPOF for calculating strain from tagged images, and compare the results to HARP at different myocardial regions based on numerical phantom and in-vivo images. The results show that BPOF is superior to HARP in tracking boundary taglines.

Weina Cui¹, Lei ye¹, Albert Jang¹, Pengyuan Zhang¹, Qiang Xiong¹, and Jianyi Zhang^1
1Department of Medicine/cardiology, University of Minnesota, minneapolis, MN, United States

We investigated the functional impact of combined intramyocardial transplantation of cardiomyocytes, endothelial cells, and smooth muscle cells derived from hiPSCs to a porcine model of ischemia reperfusion (I/R). It was demonstrated that ATP turnover rate could be measured with two components, the creatine reaction rate and ATP hydrolysis rate, even without measuring Pi levels. MI injury has a heterogeneous effect on myocardial bioenergetics. The border zone PCr/ATP ratio is improved in the cell treated group compared with open patch group. The KpcrATP in the border zone is slightly increased in cell treated group compared with patch only group.

Michael Salerno^1,2, Yang Yang³, Peter Shaw⁴, Angela Taylor⁴, Craig Meyer³, Fred Epstein³, and Christopher Kramer^4,5
1Medicine, Cardiology, University of Virginia, Charlottesville, VA, United States, ²Radiology, University of Virginia, Charlottesville, VA, United States,³Biomedical Engineering, University of Virginia, VA, United States, ⁴Medicine, Cardiology, University of Virginia, VA, United States, ⁵Radiology, University of Virginia, VA, United States

Patients with non-obstructive coronary artery disease may have abnormal myocardial perfusion reserve (MPR) resulting from microvascular disease (MVD). Reduced MPR as demonstrated by PET is a significant predictor of cardiac death. CMR has the unique ability to quantify transmural differences in perfusion given its high spatial resolution providing new insights into MVD. We developed a high resolution (1.5mm) quantitative spiral perfusion pulse sequence and performed vasodilator stress in patients with a high likelihood of MVD. We demonstrate an abnormal stress endocardial to epicardial perfusion gradient as well as abnormal MPR in these patients.

Xiao Chen¹, Michael Salerno^2,3, Christopher M. Kramer^3,4, Bhairav B. Mehta¹, Yang Yang¹, Peter Shaw⁴, and Frederick H. Epstein^1
1Biomedical Engineering, University of Virginia, Charlottesville, VA, United States, ²Radiology, University of Virginia, Charlottesville, VA, United States,³Cardiology, University of Virginia, Charlottesville, VA, United States, ⁴Medicine, Cardiovascular Medicine, University of Virginia, Charlottesville, VA, United States

First-pass cardiac perfusion MRI demands fast imaging techniques to achieve high spatial resolution and coverage within a small acquisition window. We recently developed a CS method BLOSM to accelerate first-pass perfusion with respiratory motions, and demonstrated the advantages of BLOSM using retrospectively-undersampled first-pass data. In the present study, we aimed to evaluate prospectively-accelerated BLOSM first-pass images in patients with suspected heart disease. Using prospectively accelerated data, BLOSM showed improved reconstruction quality compared to k-t SLR. BLOSM may provide clinically acceptable image quality at higher acceleration rates such as 6, even with the presence of respiratory motion.

Sandeep Kaushik¹, Dattesh Shanbhag¹, Anne Menini², Sheshadri Thiruvenkadam¹, Stephanie Reiter³, Tobias Heer³, Günter Pilz³, and Anja Brau^4
1Medical Image Analysis Lab, GE Global Research, Bangalore, Karnataka, India, ²GE Global Research, Garching, Bavaria, Germany, ³Department of Cardiology, Clinic Agatharied Academic Teaching Hospital, University of Munich, Hausham, Bavaria, Germany, ⁴GE Healthcare, Garching, Bavaria, Germany

Cardiac perfusion MRI (PWI) scans are typically riddled with heart motion due to scan times too long for breath hold. Non-rigid registration (NRR) algorithms are often used for motion correction. In this work, a group-wise NRR algorithm for motion correction of cardiac PWI images is proposed. Signal intensity variations due to contrast flow can potentially distort and introduce undesirable intensity changes in NRR output. This has been addressed using an intensity normalization scheme. Temporal filter post-processing has been used to eliminate the residual motion. Resulting contrast curves demonstrate time course conformity as validated against ground truth contrast curves.

Sonia Nielles-Vallespin¹, Peter Kellman¹, Li-Yueh Hsu¹, and Andrew E Arai^1
1National Institutes of Health, Bethesda, MD, United States

Perfusion scans preceded by both FLASH and SSFP-PD frames from 10 patients with no myocardial infarction were analyzed semi-quantitatively and quantitatively. Low flip-angle SSFP protocol exhibited severe ghosting artifacts in 6 of 10 subjects. These artifacts lead to errors in quantitative perfusion maps and semi-quantitative perfusion indexes. They can be severe enough to cause false positives diagnoses. Using FLASH-PD frames as input for the surface coil intensity correction (SCIC) improved the homogeneity of perfusion measurements. Thus, FLASH PD images are recommended for SCIC of SSFP perfusion images in place of low flip angle SSFP PD images.

Haonan Wang¹, Neal Kepler Bangerter¹, Liyong Chen², Ganesh Adluru³, and Edward V.R DiBella^3
1Department of Electrical & Computer Engineering, Brigham Young University, Provo, Utah, United States, ²Advanced MRI Technologies, Sebastopol, California, United States, ³Utah Center for Advanced Imaging Research, University of Utah, Salt Lake City, Utah, United States

First-pass myocardial perfusion imaging offers a valuable method for characterizing the blood flow of the myocardial tissue. Much of the work in myocardial perfusion imaging has used saturation preparation and an ECG trigger. An alternative approach is to use a spoiled gradient echo (SPGR) steady-state acquisition that acquires images rapidly enough so that ECG gating is not needed. In 2D simultaneous multi-slice acquisition, we utilize the CAIPI technology to achieve whole heart coverage and combined that with interleaved slice acquisition, which increases the effective TR. In this study, we demonstrate that the ungated golden ratio radial CAIPI with interleaved acquisition provides reasonable CNR while providing whole heart coverage.

Roman Wesolowski^1,2, Eva Sammut², Niloufar Zarinabad Nooralipour², Eike Nagel², and Amedeo Chiribiri^2
1University of Birmingham, Birmingham, West Midlands, United Kingdom, ²King's College London, London, United Kingdom

Coil-bias most persistently contaminates cardiac magnetic resonance (CMR) imaging despite the vendors’ efforts in reducing its effects. Due to the heart’s shape and its oblique position in reference to the coil’s receiving element, perfusion abnormalities can often be disguised in the areas highly affected by this effect, which can reduce CMR’s diagnostic capabilities. Although proton density-based solutions have been in practice, we propose a data-driven dynamic coil-bias correction (DCBC) algorithm for segmented myocardial perfusion. DCBC does not require additional scans and can be applied retrospectively. We show that it significantly reduces the effect of coil-bias, superseding proton density-based technique.

Mou anna¹, Li zhiyong², Zhang ziheng³, Song qingwei², and Liu ailian^2
1The First Affiliated Hospital of Dalian Medical University, China, Liaoning, China, ²The First Affiliated Hospital of Dalian Medical University, Liaoning, China,³GE Healthcare China, Beijing, China

Microvascluar perfusion abnormality and ischemia condition could be an important risk factor for the clinical assessment of hypertrophic cardiomyopathy(HCM). We preliminarily investigate the difference of myocardial microvascluar perfusion between HCM patients and normal volunteers with low-multi-b values(0,20,50,100,200 s/mm2) diffusion weighted imaging. We found that the measurement parameters reflected microvasluar perfusion (Standard ADC, Slow ADC, Fast ADC and fraction of Fast ADC value)in HCM patients were significant lower than that of normal volunteers. We conclude that low-multi-b values diffusion weighted imaging could quantitatively and noninvasively evaluate the perfusion status in HCM patients .

Alexander Gotschy^1,2, Lukas Wissmann¹, Datta Singh Goolaub¹, Markus Niemann³, Sebastian Kozerke¹, and Robert Manka^1,3
1Institute for Biomedical Engineering, University and ETH Zurich, Zurich, Switzerland, ²Department of Internal Medicine, University Hospital Zurich, Zurich, Switzerland, ³Department of Cardiology, University Hospital Zurich, Switzerland

The most relevant parameters for the assessment of coronary artery disease are myocardial perfusion and the status of the coronary arteries. Therefore, the aim of our study was to investigate the feasibility and potential added value of MR-based hybrid imaging by the combined assessment and fusion of 3D-MR coronary angiography (MRCA) with a 3D-CMR-perfusion sequence. We found that the combined evaluation had superior sensitivity at the cost of a loss in specificity when compared to CMR-perfusion alone in a vessel-based approach. Furthermore in the fused images, coronary stenosis could be matched with perfusion deficits in different myocardial layers.

Hsin-Jung Yang¹, Damini Dey², Jane Sykes³, John Butler³, Avinash Kali², Ivan Cokic², Behzad Sharif², Sotirios Tsaftaris⁴, Debiao Li², Piotr Slomka², Frank Prato³, and Rohan Dharmakumar^2
1Cedars Sinai Medical Center, Los angeles, CA, United States, ²Cedars Sinai Medical Center, CA, United States, ³Lawson Health Research Institute, ON, Canada, ⁴IMT Lucca Institute, Lucca, Italy

Myocardial BOLD MRI is an emerging non-contrast approach for assessing ischemic heart disease. Current myocardial BOLD methods are limited by poor spatial coverage, heart-rate dependency and image artifacts (e.g. coil bias, B1 and B0 inhomogeneities). To overcome these limitations, we developed a heart-rate independent, fast, free breathing 3D T2 mapping technique at 3T that utilizes near perfect imaging efficiency. This quantitative BOLD approach, which can be performed within 3 minutes, permits full LV coverage during pharmacological stress. In this study, we tested our approach in a canine model and validated our findings with simultaneously acquired 13N-ammonia PET perfusion data in a clinical PET-MR system.

Zhengwei Zhou^1,2, Xiaoming Bi³, Hsin-Jung Yang^1,2, Rohan Dharmakumar¹, Reza Arsanjani^1,4, C Noel Bairey Merz⁴, Daniel Berman^1,4, Debiao Li^1,2, and Behzad Sharif^1
1Biomedical Imaging Research Institute, Cedars-Sinai Medical Center, Los Angeles, CA, United States, ²Department of Bioengineering, University of California, Los Angeles, Los Angeles, CA, United States, ³MR R&D, Siemens Healthcare, Los Angeles, CA, United States, ⁴Cedars-Sinai Heart Institute, Los Angeles, CA, United States

Conventional first-pass myocardial perfusion MRI methods are prone to dark-rim artifacts. We developed a free-breathing DRA-reduced FPP scheme with Cartesian k-space sampling and instantaneous “online” image reconstruction. This can be implemented into clinical routine, enhancing its clinical accessibility. Preliminary results of canine with myocardial infarction showed high accuracy compared to late gadolinium enhancement imaging.

Merlin J Fair^1,2, Peter D Gatehouse^1,2, and David N Firmin^1,2
1NHLI, Imperial College London, London, United Kingdom, ²NIHR Cardiovascular BRU, Royal Brompton Hospital, London, United Kingdom

Whilst dark-rim artefacts have been studied in 2D first-pass perfusion, they could also have the potential to be produced in the through-plane direction of 3D first-pass perfusion due to the typically coarse resolution employed in the second phase-encoding direction. An examination of the potential for this Gibbs truncation effect, and its relationship to the also troubling partial-volume effect, is studied in conical and anatomical numerical phantoms as well as a representative in-vivo dataset.

CE & Non-CE NRA

Thursday 4 June 2015

Jinnan Wang¹, Haining Liu², Zechen Zhou³, Niranjan Balu², Thomas S Hatsukami², Jin Liu², Peter Boernert⁴, and Chun Yuan^2
1Philips Reserach North America, Seattle, WA, United States, ²University of Washington, Seattle, WA, United States, ³Tsinghua University, Beijing, China,⁴Philips Research Europe, Hamburg, Germany

Simultaneous Non-contrast Angiography and intraPlaque hemorrhage (SNAP) imaging was proposed as a technique for joint MRA and high risk atherosclerotic plaque feature (intraplaque hemorrhage, or IPH) detection. For carotid artery imaging, to achieve ideal MRA, an inversion coverage of 60cm is desired but can’t be afforded by most scanners due to the bore length limitation. As a result, flow artifacts can be occasionally observed on the carotid SNAP MRA images. To address this issue, a new SNAP acquisition scheme is proposed to improve the MRA performance by optimizing k-space acquisition ordering so that flow artifacts can be minimized.

Taehoon Shin¹, Seth J Kligerman¹, Robert S Crawford², Sanjay Rajagopalan³, and Rao P Gullapalli^1
1Radiology, University of Maryland, Baltimore, MD, United States, ²Vascular Surgery, University of Maryland, MD, United Kingdom, ³Cardiovascular Medicine, University of Maryland, Baltimore, MD, United States

A non-contrast-enhanced peripheral MR venography method is proposed which generates high-contrast 3D venograms directly using a single acquisition. The proposed method combines velocity-selective magnetization preparation for suppressing arterial signal and transient balanced SSFP readout for suppressing muscle signal. The sequence parameters are optimized to maximize vein-to-background contrast and validated in vivo. The feasibility of visualization of the entire peripheral venous systems using multi-station is demonstrated in healthy subjects.

Christopher J Hanrahan¹, Marc Lindley¹, Michelle Mueller², Daniel Sommers¹, Marta E Heilbrun¹, Glen Morrell¹, Daniel Kim¹, and Vivian S Lee^1
1Radiology, UCAIR, University of Utah School of Medicine, Salt Lake City, Utah, United States, ²Vascular Surgery, University of Utah School of Medicine, Salt Lake City, United States

Non-contrast MRA techniques using QISS, QIR, and ECG-gated FSE were compared among 3 stations in peripheral arterial disease patients at 3 Tesla using gadolinium enhanced MRA as the standard.

Ruth P Lim^1,2, Adrienne CY Lam¹, Matthew Lukies¹, Dinesh Ranatunga¹, Emma K Hornsey¹, Brenden McColl¹, Yuliya Perchyonok^1,2, Jason Chuen^2,3, Jason Heidrich¹, Pei-Heng Ko³, and Robert R Edelman^4
1Radiology, Austin Health, Melbourne, Victoria, Australia, ²The University of Melbourne, Melbourne, Victoria, Australia, ³Vascular Surgery, Austin Health, Melbourne, Victoria, Australia, ⁴Radiology, NorthShore University Health System, Chicago, IL, United States

Imaging peripheral arterial disease (PAD) in diabetic patients is challenging. We evaluate a combined QISS MRA and QISS arterial spin labeled MRA approach (cQISS MRA) for imaging from the infrarenal aorta to the pedal vessels in 15 diabetic patients with symptomatic PAD, using digital subtraction angiography as the reference standard. High diagnostic confidence was observed for cQISS-MRA, however this was significantly lower in the foot compared to more proximal regions. 74.7% sensitivity and 86.8% specificity for hemodynamically significant (≥50%) stenosis was observed overall, with highest sensitivity and specificity for the pelvic region, and poor specificity (50%) for the pedal region.

Na Zhang^1,2, Zhaoyang Fan³, and Xin Liu^1,2
1Lauterbur Research Center for Biomedical Imaging, Shenzhen Institutes of Advanced Technology of Chinese Academy of Sciences, Shenzhen, Guangdong, China, ²Shenzhen Key Laboratory for MRI, Shenzhen, Guangdong, China, ³Biomedical Imaging Research Institute, Cedars-Sinai Medical Center, Los Angeles, CA, United States

Three-dimensional (3D) non-contrast enhanced MR angiography (NCE-MRA) using flow-sensitive dephasing (FSD) prepared steady-state free precession (SSFP) had proved a promising strategy for evaluation of peripheral vasculature without the use of gadolinium-based contrast agent. The magnetization preparation with FSD gradients was used for the blood signal suppression with advantages of flow-independence. It could be flexibly configured in both direction and strength so that multidirectional flow suppression was feasible. The purpose of this study was to compare the image quality of pedal arteries acquired from the NCE-MRA technique with single directional and two-directional FSD modules preparation, respectively.

Qin Qin^1,2, Taehoon Shin³, Michael Schar¹, Hua Guo⁴, and Ye Qiao^1
1Radiology, Johns Hopkins University, Baltimore, Maryland, United States, ²Kirby Center, Kennedy Krieger Institute, Baltimore, Maryland, United States,³Radiology, University of Maryland, Baltimore, Maryland, United States, ⁴Center for Biomedical Imaging Research, Biomedical Engineering, Tsinghua University, Beijing, China

The new velocity-selective saturation (VSS) pulse train, by embedding a pair of composite pulses for refocusing in each k-segment and phase cycling applied over consecutive k-segments, offers improved robustness to B0/B1 inhomogeneity in the brain at 3T. Velocity-selective magnetization-prepared non-contrast-enhanced cerebral MR angiography was demonstrated on healthy volunteers. Compared with TOF MRA, VSS prepared MRA depicted more distal branches of cerebral arteries and allowed for the whole-brain coverage without the requirement of section positioning orthogonal to the direction of flow.

Taehoon Shin¹, Qin Qin², Jang-Yeon Park³, and Sanjay Rajagopalan^4
1Diagnostic Radiology, University of Maryland, Baltimore, MD, United States, ²Radiology, Johns Hopkins University, Baltimore, MD, United States, ³Biomedical Engineering, Sungkyunkwan University, Suwon, Gyeonggi-do, Korea, ⁴Cardiology, University of Maryland, Baltimore, MD, United States

While velocity-selective (VS) magnetization-prepared non-contrast-enhanced MR angiography has shown great promise at 1.5T with advantages of non-subtractiveness and large 3D FOV, its feasibility at 3T remains questionable due to the sensitivity of VS excitation to B0 and B1 inhomogeneities. In this study, we present an updated version of VS-MRA pulse sequence which employs paired refocusing for reducing the sensitivity to B0 and B1 offsets and double VS preparations for removing stripe artifacts caused by imperfect refocusing. The feasibility of the proposed method is shown in a healthy subject.

Naoyuki Takei¹, Kevin F. King², Adriana Kanwischer², and Hiroyuki Kabasawa^1
1GE Healthcare, Hino, Tokyo, Japan, ²GE Healthcare, WI, United States

3D TOF MR Angiography is an established clinical routine application. Scan time saving is important for brain routine that is frequently used in MRI scan in the world. Compressed sensing (CS) is a promising technique to accelerate scanning for high imaging contrast that has sparse image. In this work, the combination of CS and parallel imaging (PI) was applied to 3D TOF and demonstrated feasible. Volunteer scan was performed compared with PI only in brain. The CS +PI sequence reduced scan time by 43 % identifying the main and distal arteries. In conclusion, CS provides equivalent image quality with less scanning time than PI only.

Elizabeth M Hecht¹, Firas Ahmed¹, Anuradha Shenoy-Bhangle¹, Guillermo Jimenez¹, Stuart Bentley-Hibbert¹, and Martin Prince^1
1Columbia University, New York, NY, United States

This study was performed to assess the additive value of non-contrast enhanced MRA (NC-MRA) using a respiratory gated 3D SSFP based sequence for evaluation of liver arterial anatomy in patients under going contrast enhanced MRA (Gd-MRA) for preoperative planning for potential liver donation. Liver donor arterial anatomy evaluation is not significantly improved by the addition of NC-MRA but can be useful when GD-MRA is suboptimal and further improvements in sequence design would be helpful to limit contrast exposure in this population and increase efficiency of patient throughput.

Andrew Peter Aitken¹, Mehmet Akçakaya², Rene Botnar¹, and Claudia Prieto^1
1Division of Imaging Sciences and Biomedical Engineering, King's College London, London, London, United Kingdom, ²Department of Medicine, Beth Israel Deaconess Medical Center, Boston, MA, United Kingdom

A highly accelerated 100% scan efficiency reconstruction approach for whole-heart coronary MRA is proposed. This approach achieves 100% scan efficiency by estimating the motion from an interleaved golden radial image navigator. Further acceleration is achieved by undersampling the MRA acquisition using a golden-step spiral-like Cartesian acquisition and the low dimensional structure self-learning and thresholding (LOST) reconstruction. This LOST-motion corrected approach leads to slightly reduced image quality in comparison to that of an ~8 times longer scan time navigator-gated approach.

Yutaka Nattsuaki¹, Xiaoming Bi¹, Aurelien F Stalder², and Gerhard Laub^1
1Siemens Healthcare, Los Angeles, CA, United States, ²Siemens Healthcare, Erlangen, Germany

Robust but slow 3D Time-of-flight (TOF) MR angiography has its practical limitation in slice FOV coverage. Recent advancements in MRI acceleration with TOF sequences (e.g. TOF with sparse undersampling) can resolve this practical limitation, and large slice FOV coverage TOF may be realized. Not only for the larger coverage, extra time savings can be reinvested in slice resolution for better delineation of small arteries. The current work introduces the variable slice resolution TOF, a novel concept that adjusts slice resolutions according to anatomical needs for the optimal large FOV TOF within clinically acceptable scan time (<10 min.).

Giuseppe Muscogiuri^1,2, Akos Varga-Szemes¹, U. Joseph Schoepf¹, Carlo N De Cecco^1,2, Davide Piccini^3,4, Wolfgang G Rehwald^5,6, Anthony M Hlavacek¹, and Arni C Nutting^1
1Medical University of South Carolina, Charleston, SC, United States, ²University of Rome Sapienza, Rome, Italy, ³Siemens Healthcare IM BM IP, Lausanne, Switzerland, ⁴University of Lausanne, Lausanne, Switzerland, ⁵Siemens Medical Solutions, Chicago, IL, United States, ⁶Duke Cardiovascular Magnetic Resonance Center, Durham, NC, United States

In this project we compared a prototype non-contrast free-breathing self-navigated 3D (SN3D) whole-heart MRA acquisition and coronary CTA for the evaluation of the proximal course of coronary arteries in a pediatric patient population. Evaluation of image quality according to a 4-grade scale (1, non-diagnostic; 2, sufficient; 3, good; 4, excellent) and the visualization of the left main, left anterior descending, circumflex, first diagonal, posterior descending, and right coronary arteries were performed, in both techniques, by two experienced observers. Our preliminary results indicate that the non-contrast free-breathing SN3D MRA acquisition allows for the evaluation of coronary artery anatomy with similar image quality to cCTA without any radiation or administration of contrast agent.

Xiangzhi Zhou¹, Cheng Ouyang¹, Aiming Lu¹, and Mitsue Miyazaki^1
1Toshiba Medical Research Institute USA, Vernon Hills, IL, United States

In conventional subtraction based non-contrast ECG gated 3D single shot fast spin echo (SSFSE) technique, each slice encoding and the followed echo train are played in a fixed TR (TR=n*RR). Ideally, TR should be sufficiently long for the recovery of the longitudinal magnetization of blood and background tissue, which led to a long scan time. In this work, a variable TR (vTR) method is proposed to reduce the scan time while still maintain the blood signal. The results showed the vTR method can significantly reduce scan time by 20-40% and had comparable arterial image quality as the fixed TR scan.

Tomohisa Okada¹, Koji Fujimoto¹, Yasutaka Fushimi¹, Akira Yamamoto¹, Kei Sano², Toshiyuki Tanaka², Naotaka Sakashita³, and Kaori Togashi^1
1Dept. of Diagnostic Imaging and Nuclear Medicine, Kyoto University Graduate School of Medicine, Kyoto, Kyoto, Japan, ²Department of Informatics, Kyoto University Graduate School of Informatics, Kyoto, Kyoto, Japan, ³Toshiba Medical Systems, Otawara, Tochigi, Japan

Visualization of the lenticulostriate artery (LSA) using TOF-MRA requires high resolution and takes long acquisition time, which might be mitigated by under-sampling and compressed sensing (CS) reconstruction. However, application of CS to small structures such as LSA is yet to be investigated. CS-reconstructed images of 25, 50 and 75% data were compared with the full-sampled images in 11 volunteers. Under-sampling of 75% data was comparable to the control, but visualization was impaired at 50% and 25%. Higher contrast between LSA and background, as well as improved CS reconstruction, would be required for LSA visualization.

Shivraman Giri¹, Eugene Dunkle², Wei Li², Ian Murphy^2,3, Ioannis Koktzoglou^2,4, and Robert R Edelman^2,3
1Siemens Healthcare, Chicago, IL, United States, ²Radiology, NorthShore University HealthSystem, IL, United States, ³Radiology, Northwestern University Feinberg School of Medicine, IL, United States, ⁴Radiology, The University of Chicago Pritzker School of Medicine, IL, United States

Quiescent-interval single-shot (QISS) has proven to be an accurate and robust alternative to contrast-enhanced MRA and CT angiography for the evaluation of peripheral arterial disease, but scan time is longer than with these other imaging tests. In this work, we demonstrate a pulse sequence modification that cuts QISS scan time in half while maintaining image quality.

Allison Grayev¹, Utaroh Motosugi^1,2, Peter Bannas^1,3, Naoyuki Takei⁴, Kevin King⁵, Kang Wang⁶, James Holmes⁷, Scott Reeder^8,9, and Aaron Field^1
1Department of Radiology, University of Wisconsin, Madison, WI, United States, ²Department of Radiology, University of Yamanashi, Yamanashi, Japan,³Department of Radiology, University Hospital Hamburg-Eppendorf, Hamburg, Germany, Germany, ⁴Global MR Applications and Workflow, GE Healthcare, Hino, Japan, ⁵Global MR Applications and Workflow, GE Healthcare, Waukesha, WI, United States, ⁶Global MR Applications and Workflow, GE Healthcare, Madison, WI, United States, ⁷Department of Medical Physics, University of Wisconsin, Madison, WI, United States, ⁸Department of Radiology; Department of Medical Physics, University of Wisconsin, Madison, WI, United States, ⁹Department of Biomedical Engineering and Medicine; Department of Emergency Medicine, University of Wisconsin, Madison, WI, United States

Vascular imaging is an important step in the assessment of stroke; however, optimal evaluation of the neck vasculature often requires multiple sequences. The purpose of this study was to develop a non-contrast method of evaluating the carotid arteries from the aortic arch to skull base with excellent resolution in acceptable time parameters. Inflow-enhanced inversion recovery with fast spin echo readout (IFIR-FSE) was combined with parallel imaging and compressed sensing and compared to both contrast enhanced MRA and time-of-flight MRA in volunteer patients. After the image quality was optimized, patients were recruited to demonstrate feasibility in the clinical setting. 

Hiroki Matoba¹, Akiyoshi Yamamoto¹, Yuji Shintani¹, Daiji Uchiyama¹, Seigo Yoshida¹, Katsumi Nakamura^1,2, and Mitsue Miyazaki^3
1Radiology, Tobata Kyoritsu Hospital, Kitakyusyu, Fukuoka, Japan, ²Radiology, Hikari Central Hospital, Hikari, Yamaguchi, Japan, ³Toshiba Medical Research Institute USA, Vernon Hills, Illinois, United States

We evaluate the visualization of sequential MR portography by using the Time-SLIP flow-in technique after food intake to determine an appropriate examination timing. MR portography was performed on 5 healthy volunteers at pre-meal, just after meal, 30 minutes after, 1 hour after, 2 hours after, 3 hours after, and 4 hours after. The visualization of extra- and intra-hepatic portal veins was significantly improved after food intake which increased portal venous flow, among which especially prominent at 1 to 2 hours after meal, which were thought to the appropriate timing to perform Time-SLIP flow-in portogrpahy rather than during the fasting period

Marc D Lindley^1,2, Daniel Kim¹, Glen Morrell¹, Marta E Heilbrun¹, Christopher J Hanrahan¹, and Vivian S Lee^1
1UCAIR, Radiology, University of Utah, Salt Lake City, Utah, United States, ²Physics, University of Utah, Salt Lake City, Utah, United States

Fresh blood imaging (FBI) is based on subtraction of two acquisitions done at two different cardiac phases, by which the background signal cancels out. In clinical practice, however, we observe residual background signal after subtraction, which makes it challenging to visualize the vessels. In this study we sought to add fat saturation to FBI, in order to suppress better suppress background signal. We imaged eight patients with peripheral arterial disease using FBI with and without fat saturation, and our results show that FBI with fat saturation improves background suppression compared with FBI without fat saturation.

Xinzeng Wang¹, Joshua S Greer^1,2, Shu Zhang¹, and Ananth J Madhuranthakam^1,3
1Radiology, UT Southwestern Medical Center, Dallas, Texas, United States, ²Bioengineering, UT Dallas, Dallas, Texas, United States, ³Advanced Imaging Research Center, UT Southwestern Medical Center, Dallas, Texas, United States

Velocity selective preparation (VSP) allows non-contrast enhanced MR angiography with minimal background signal in a single acquisition. However, the empirical velocity encoding used with VSP leads to venous signal contamination. In this work, we present a phase sensitive reconstruction combined with VSP that can determine the flow direction and can potentially differentiate arterial and venous signal with background signal suppression in a single acquisition. The proposed method successfully determined the flow direction in a flow phantom with good background suppression. The phase sensitive reconstruction also determined the flow direction in the popliteal artery of the lower leg in normal volunteers.

Andrew J. Walsh¹, Derek J. Emery², Ken Warren³, Ingrid Catz³, and Alan H. Wilman^1
1Biomedical Engineering, University of Alberta, Edmonton, Alberta, Canada, ²Radiology and Diagnostic Imaging, University of Alberta, Edmonton, Alberta, Canada, ³Neurology, University of Alberta, Edmonton, Alberta, Canada

First pass, 3D gadolinium-enhanced venography of the jugular veins is difficult to perform when seeking both large spatial coverage and high spatial resolution. Particularly when considering routine clinical 1.5 T systems that may not be equipped with dense phased arrays for parallel imaging, capturing the venous phase without overwhelming arterial enhancement becomes difficult. We test first pass Gd-enhanced venography encompassing a large 3D coronal field-of-view at 0.64 mm3 voxel dimensions using a single 36 sec acquisition. High resolution enables proper assessment of venous narrowing. We report on the feasibility in application to 60 subjects, 30 with multiple sclerosis (MS) and 30 healthy controls to examine cerebral venous drainage.

Aurélien Julien Trotier¹, William Lefrancois¹, Kris Van Renterghem¹, Jean-Michel Franconi¹, Eric Thiaudière¹, and Sylvain Miraux^1
1RMSB-UMR5536, CNRS - Université de Bordeaux, Bordeaux, Aquitaine, France

To show that 3D sequences with ultra-short echo times (UTEs) can generate a positive contrast whatever the magnetic field (4.7, 7 or 9.4 T) and whatever Ultra Small Particles of Iron Oxide (USPIO) concentration injected and to use it for 3D time-resolved imaging of the murine cardiovascular system with high spatial and temporal resolutions. This approach might be useful to measure the functional cardiac parameters or to assess anatomical modifications to the blood vessels in cardio-vascular disease models.

Jeffrey H Maki¹, Guenther Schneider², Alexander Massmann², Matthias Leist², Diane Wagner-Jochem², and Gregory J Wilson^1
1Radiology, University of Washington, Seattle, WA, United States, ²Radiology, University Hosptial of Saarland, Homburg, Germany

Eight juvenile swine were imaged with high temporal resolution (1.7 s) CE-MRA 3x each during the bolus passage of one of four Gd agents at 1,2,3 ml/sec (2 animals/agent, all single dose, 24 total injections). Simultaneous aortic blood sampling was performed (every 2 s) to measure actual Gd concentration. Ex vivo arterial blood was doped with each contrast agent to determine R1 and R2* vs. [Gd]. Predicted and actual signal intensity (SI) vs. time as well as bolus duration were compared, demonstrating a relatively flat SI response vs. injection rate, validating the benefit of slower Gd injections for CE-MRA.

Evan Kao^1,2, Farshid Faraji¹, Sarah Kefayati¹, Van Halbach¹, Matthew Amans¹, and David Saloner^1
1Radiology, UCSF, San Francisco, CA, United States, ²Bioengineering, UC Berkeley, Berkeley, CA, United States

We investigated the flow patterns in the jugular veins of two subject groups, one normal and one with suspected anamolies, using MRI-based CFD. Our results show that there are two types of venous geometries and flow patterns: (1) One which has helical flow in the jugular bulb (the junction between the sigmoid sinus and jugular vein), which results in a relatively straight flow in the jugular vein; (2) and another in which the jugular bulb is shaped such that it redirects flow at a right angle to the jugular vein, which creates pronounced helical flow pattern.

Todd C. Soesbe^1,2, Ketan B. Ghaghada³, S. James Ratnakar¹, Chandreshkumar Patel³, Mark Milne¹, A. Dean Sherry^1,4, and Robert E. Lenkinski^2
1Advanced Imaging Research Center, UT Southwestern Medical Center, Dallas, Texas, United States, ²Department of Radiology, UT Southwestern Medical Center, Dallas, Texas, United States, ³Texas Children's Hospital, Houston, Texas, United States, ⁴Department of Chemistry, University of Texas at Dallas, Dallas, Texas, United States

Synopsis goes here.

Vessel Wall & Cardiovascular Image Processing

Thursday 4 June 2015

Ilkay Oksuz¹, Anirban Mukhopadhyay¹, Marco Bevilacqua¹, Hsin-Jung Yang^2,3, Rohan Dharmakumar^2,3, and Sotirios A. Tsaftaris^1,4
1IMT Institute for Advanced Studies, Lucca, Tuscany, Italy, ²Biomedical Research Institute, Cedars Sinai Medical Center, Los Angeles, California, United States,³Medicine, University of California, Los Angeles, California, United States, ⁴Electrical Engineering and Computer Science, Northwestern University, Illinois, United States

Cardiac phase-resolved Blood-Oxygen-Level-Dependent (CP-BOLD) MRI is a new approach for detecting ischemia at rest. Currently disease assessment relies on segmental analysis and uses only a few images in the phase-resolved acquisition. It is expected that using all phases can permit pixel-level characterization of CP-BOLD MRI. In this study, state-of-the-art image registration techniques are evaluated on cardiac BOLD MRI data for the first time. The results show that cardiac phase-dependent variations in myocardial BOLD contrast in CP-BOLD images creates a statistically significant decrease in the accuracy compared to standard Cine MR images acquired under conditions of health and myocardial ischemia.

Ryohei Nakayama¹, Masaki Ishida¹, Yasutaka Ichikawa¹, Yoshitaka Goto¹, Motonori Nagata¹, Kakuya Kitagawa¹, and Hajime Sakuma^1
1Department of Radiology, Mie University School of Medicine, Tsu, Mie, Japan

A 3D SR technique was developed to improve image resolution and quality of whole-heart coronary MRA (WHCMRA) by utilizing GPUs. WHCMRA images with 0.6x0.6x0.75 mm resolution were reconstructed from the down-sampled WHCMRA images (1.2x1.2x1.5 mm) by using 3D SR technique, 2D SR technique and 3D bicubic interpolation. The computational time with 3D SR technique per patient considerably reduced from 4 h and 41.1 min to 45.5 min by the parallel implementation in GPUs. Root mean square error, structural similarity index and signal-to-noise ratio for 3D SR technique were 2.85, 0.984 and 65.8, being significantly greater than those for other two methods. The 3D SR approach would improve the diagnostic performance of WHCMRA for detection of stenosis.

Haris Saybasili¹, Marie-Pierre Jolie², and Bruce Spottiswoode^1
1Siemens Healthcare, Chicago, Illinois, United States, ²Imaging and Computer Vision, Siemens Corporation, Corporate Technology, NJ, United States

Real-time free-breathing imaging methods are viable alternatives to conventional segmented cine imaging for patients that cannot hold their breath. However, heart-rate changes during real-time acquisitions result in a different number of phases per beat/slice. Methods of obtaining a predefined number of cardiac phases per beat/slice from real-time images has been shown previously. However, cardiac phases acquired during inconsistent respiratory phases are hard to analyze due to respiratory motion. In this work, we propose an unsupervised motion correction scheme to generate high SNR, single heart-beat breath-held (end expiration) cine images from any real-time free-breathing acquisition covering multiple heart-beats.

Andreia S. Gaspar^1,2, David J. Cox¹, Alan M. Groves^1,3, and Anthony N. Price^1
1Centre for the Developing Brain, King's College London, London, London, United Kingdom, ²Instituto de Biofisica e Engenharia Biomedica, Faculdade de Ciencias, Universidade de Lisboa, Lisboa, Lisboa, Portugal, ³Department of Pediatrics, Weill Cornell Medical College, New York, United States

Neonatal cardiac MRI presents significant challenges associated with subject size; cine sequences require multiple signal averages to increase SNR. However, when scanning unsedated, free-breathing infants, motion artefacts compromise individual averages and the resultant averaged dataset. We present a tool that rigidly registers averages, correcting motion displacement, before using intensity analysis to remove averages of poor image quality. Applying the software on neonatal cohort improved image quality leading to clearer definition of endo- and epicardial borders. Such post-processing of multi-average acquisitions could prove valuable in future trials using neonatal cardiac MRI.

Jens Wetzl^1,2, Christoph Forman³, Andreas Maier^1,2, Joachim Hornegger^1,2, and Michael O. Zenge^3
1Pattern Recognition Lab, Department of Computer Science, Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, Germany, ²Erlangen Graduate School in Advanced Optical Technologies (SAOT), Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, Germany, ³Siemens AG, Healthcare, Imaging & Therapy Systems, Magnetic Resonance, Erlangen, Germany

Respiratory motion represents a challenge in free-breathing whole-heart coronary MR angiograpy. For respiratory motion compensation, weighted iterative reconstruction aims to reconstruct a consistent sub-set of the acquired data. However, this may lead to increased sub-sampling artifacts. Motion-compensated (MoCo) reconstruction promises to overcome this by incorporating all acquired data using a motion model. Unfortunately, computation times are longer, and the resulting signal-to-noise ratio (SNR) improvement may not always justify this effort. This work proposes a method to predict the benefit of MoCo over weighted reconstruction directly after data acquisition. This prediction method was evaluated with in-vivo experiments in 15 volunteers.

Rui Li¹, Shujing Cao¹, Feng Huang², and Chun Yuan^1,3
1Center for Biomedical Imaging Research, Tsinghua University, Beijing, Beijing, China, ²Philips Research China, Shanghai, China, ³University of Washington, Seattle, Washington, United States

Carotid vessel wall magnetic resonance imaging (MRI) has been validated to discriminate different components in plaques and evaluate their vulnerability. Patient motion such as breathing and swallow is inevitable during acquisition. Navigator echo or self-gating techniques were explored by several previous works to monitor swallowing. However all these techniques are prospective methods and adopted accept/reject-reacquisition as their data management strategy, which interrupted data acquisition and prolonged scan duration. The target of this work is to develop a retrospective motion compensation method iteratively using a GRAPPA like convolution and optimized coil element data combination to preserve SNR during iteration.

Huijun Chen¹, Jin Liu², Zechen Zhou³, Chun Yuan², Peter Boernert⁴, and Jinnan Wang^5
1Tsinghua University, Beijing, Beijing, China, ²University of Washington, Seattle, WA, United States, ³Tsinghua University, Beijing, China, ⁴Philips Research Europe, Hamburg, Germany, ⁵Philips Reserach North America, Seattle, WA, United States

Motion induced artifact is a common cause of unsatisfactory image quality in carotid artery imaging: previous study found that motion was the most significant contributor to low image quality. Monitoring and quantifying motion in the neck area is more challenging than other body parts due to the complex non-rigid motion pattern. Structured light based motion detection system has been recently developed for MR. It was selected for neck based motion tracking in this study because of its capability in accurately measuring distance changes.

Anirban Mukhopadhyay¹, Marco Bevilacqua¹, Ilkay Oksuz¹, Rohan Dharmakumar^2,3, and Sotirios Tsaftaris^1,4
1IMT Institute for Advanced Studies Lucca, Lucca, LU, Italy, ²Biomedical Imaging Research Institute, Cedars-Sinai Medical Center, Los Angeles, CA, United States, ³Medicine, University of California, Los Angeles, Los Angeles, CA, United States, ⁴Electrical Engineering and Computer Science, Northwestern University, Evanston, Il, United States

Cardiac phase-dependent variations of myocardial signal intensities in Cardiac Phase-resolved Blood-Oxygen-Level-Dependent (CP-BOLD) MRI can be exploited for the identification of ischemic territories. This technique requires segmentation to isolate the myocardium. However, spatio-temporal variations of BOLD contrast, prove challenging for existing automated myocardial segmentation techniques, because they were developed for acquisitions where contrast variations in the myocardium are minimal. Appropriate feature learning mechanisms are necessary to best represent appearance and texture in CP-BOLD data. Here we propose and validate a feature learning technique based on multiscale dictionary model that learns to sparsely represent effective patterns under healthy and ischemic conditions.

Marco Bevilacqua¹, Anirban Mukhopadhyay¹, Ilkay Oksuz¹, Cristian Rusu², Rohan Dharmakumar^3,4, and Sotirios A. Tsaftaris^1,5
1IMT Institute for Advanced Studies, Lucca, LU, Italy, ²University of Vigo, Vigo, Galicia, Spain, ³Biomedical Imaging Research Institute, Cedars-Sinai Medical Center, Los Angeles, CA, United States, ⁴Medicine, University of California, Los Angeles, CA, United States, ⁵Electrical Engineering and Computer Science, Northwestern University, Evanston, IL, United States

Cardiac Phase-resolved Blood-Oxygen-Level-Dependent (CP-BOLD) MRI has been recently demonstrated to detect an ongoing myocardial ischemia at rest, taking advantage of spatio-temporal patterns in myocardial signal intensities, which are modulated by the presence of disease. However, this approach does require significant post-processing to detect the disease and to this day only a few images of the acquisition are used coupled with fixed thresholds to establish biomarkers. We propose a threshold-free unsupervised approach, based on dictionary learning and one-class support vector machines, which can generate a probabilistic ischemia likelihood map.

Ming Li^1,2, Himanshu Gupta³, Steven G. Lloyd³, Louis J. Dell'Italia³, and Thomas S. Denney Jr.^1,2
1Auburn University MRI Research Center, Auburn University, Auburn, AL, United States, ²Electrical and Computer Engineering, Auburn University, AL, United States, ³Division of Cardiovascular Disease, University of Alabama at Birmingham, AL, United States

In this abstract we propose an integer optimization based method for unwrapping harmonic phase from tagged cardiac MR images and use this method to compute 3D biventricular cardiac strain through the cardiac cycle. Compared to the traditional quality guided phase unwrapping method, the proposed method was considerably less sensitive to the phase inconsistencies and therefore required less manual intervention. 5 min of automated phase unwrapping and 15 min of manual corrections were required for each study. Biventricular strains from the proposed method were compared to both a feature-based and a phase-based method.

Yun-Jung Jack Tsai¹, Yingmin Liu¹, Andreas Greiser², Carmel Hayes², Helen Lam¹, Chris Occleshaw¹, Alistair Young¹, and Brett Cowan^1
1University of Auckland, Auckland MRI Research Group, Auckland, New Zealand, ²Siemens Healthcare, Erlangen, Germany

Cardiac strain quantification is important in the assessment of regional myocardial function, and has many different clinical applications. The current gold standard for non-invasive myocardial strain measurement is cardiac magnetic resonance (CMR) tagging. It is also possible to estimate global and regional strain from standard (non-tagged) steady-state free precession (SSFP) cines using non-rigid registration , thereby removing the need to acquire additional sequences. We report an ‘in-line’ implementation providing numerical and graphical strain measurements without user interaction as part of the standard SSFP acquisition.

Junmin Liu¹, Dana C Peters², and Maria Drangova^1,3
1Imaging Research Laboratories, Robarts Research Institute, Schulich School of Medicine & Dentistry, University of Western Ontario, London, Ontraio, Canada,²Department of Diagnostic Radiology, Yale Medical School, New Haven, Connecticut, United States, ³Department of Medical Biophysics, Schulich School of Medicine & Dentistry, University of Western Ontario, London, Ontario, Canada

Dixon techniques to suppress fat signal are sometimes required for assessment of fibrosis/scar with late gadolinium enhancement (LGE). Among Dixon LGE techniques, the bipolar dual-echo technique is intrinsically the fastest, and is thus optimal for high-resolution LGE imaging. However, perfect fat-water separation with bipolar dual-echo acquisition is still very difficult to obtain. We extended a recently published (mono-polar multi-echo) non-iterative phase-unwrapping-based field mapping method (B0-NICE) to this challenging case. The method is tested in three human subjects with left atrial LGE and the results are compared with those obtained using an established region-growing algorithm.

Akos Varga-Szemes¹, Rob J van der Geest², Bruce Spottiswoode³, Giuseppe Muscogiuri^1,4, Carlo N De Cecco^1,4, Pal Suranyi¹, Wolfgang G Rehwald^3,5, and U. Joseph Schoepf^1
1Medical University of South Carolina, Charleston, SC, United States, ²Leiden University Medical Center, Leiden, Netherlands, ³Siemens Medical Solutions, Chicago, IL, United States, ⁴University of Rome Sapienza, Rome, Italy, ⁵Duke Cardiovascular Magnetic Resonance Center, Durham, NC, United States

We evaluated the feasibility of myocardial infarct quantification using synthetic phase-sensitive inversion recovery (PSIR) images. Patients underwent conventional PSIR and prototype modified look-locker IR (MOLLI)-based T1 acquisitions. Synthetic IR images were retrospectively calculated based on the voxel-by-voxel T1 values using an in-house developed application integrated in the Mass Research Software. Late gadolinium enhancement (LGE) measured by conventional and synthetic PSIR techniques showed no statistical difference indicating that LGE quantification using synthetic PSIR images is feasible. With the increasing availability of T1-mapping, the need for conventional PSIR images could be omitted, resulting in a significant reduction in scanner time.

Kyungpyo Hong^1,2, Ganesh Adluru¹, Edward VR. DiBella¹, and Daniel Kim^1
1UCAIR, Department of Radiology, University of Utah, Salt Lake City, Utah, United States, ²Department of Bioengineering, University of Utah, Salt Lake City, Utah, United States

Cardiac T₁ mapping sequences are conducted with breath-holding and ECG-gating to acquire multiple T₁-weighted images for calculation of pixel-by-pixel T₁map. This approach may fail in patients with arrhythmia and/or poor breath-holding. One approach to address this problem is to perform ungated, free-breathing T₁ mapping with motion correction during post processing. We sought to evaluate the performance of ungated, free-breathing arrhythmia-insensitive-rapid (AIR) cardiac T₁ mapping with diffeomorphic registration. In 11 patients with atrial fibrillation, compared with ECG-gated breath-holding AIR as the control, ungated, free-breathing AIR with motion correction yielded T₁ values that are less than 5% different from control values.

Kyungpyo Hong^1,2, Edward VR. DiBella¹, Akram Shaaban³, Daniel Sommer³, Leif Jensen³, Eugene G. Kholmovski¹, Ravi Ranjan⁴, and Daniel Kim^1
1UCAIR, Department of Radiology, University of Utah, Salt Lake City, Utah, United States, ²Department of Bioengineering, University of Utah, Salt Lake City, Utah, United States, ³Department of Radiology, University of Utah, Salt Lake City, Utah, United States, ⁴Cardiology, Internal Medicine, University of Utah, Salt Lake City, Utah, United States

Late gadolinium-enhanced (LGE) image could be derived from a post-contrast cardiac T₁ map using the Bloch equation describing inversion recovery. However, this requires T₁ of the normal myocardium to null it. We sought to automatically generate seven synthetic LGE images based on k-means clustering of T₁values, without prior knowledge of T₁ of the normal myocardium. Data from canines with radio-frequency ablation lesions in the left ventricle show that both standard and synthetic LGE images (derived from cardiac T₁ maps) produce 96 and 90% accuracy in lesion detection, respectively.

Sébastien Roujol¹, Tamer A. Basha¹, Sebastian Weingärtner¹, Mehmet Akcakaya¹, Sophie Berg¹, Warren 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

Quantitative myocardial T2 mapping sequences generally use a breath-hold ECG-triggered T2-prepared steady-state free precession acquisition. Due to limitations imposed by breath-hold duration, these sequences are typically restricted to the acquisition of four T2-weighted images. Free breathing myocardial T2 mapping sequences remove the time constraint and enable the acquisition of more samples along the T2 decay curve, which may result in improved precision and reproducibility of T2 estimates. We recently developed the ARCTIC motion correction technique which we have evaluated for myocardial T1 mapping. In this study, we sought to evaluate the performance of ARCTIC for motion compensation in free breathing myocardial T2 mapping and to evaluate its impact on in-vivo reproducibility and spatial variability of myocardial T2 estimates.

Merih Cibis¹, Kelly Jarvis^2,3, Alex J Barker², Michael Rose^2,4, Cynthia Rigsby^2,4, Michael Markl^2,3, 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

Previous studies suggested that the hemodynamics of the Fontan connection may play a role in the outcome of Fontan patients. However, the reported hemodynamic parameters in the literature, such as flows and power losses inside Fontan circuit show large discrepancies and the segmentation of the Fontan circuit might be one of the causes of the discrepancy. In this study, our aim was to investigate the influence of segmentation on the estimated geometrical and hemodynamic parameters. We showed that lumen segmentation on different image types might lead to significant differences in the estimated lumen diameters, flows and subsequent power losses.

Thekla Oechtering¹, Carl Frederik Hons¹, Julian Haegele¹, Peter Hunold¹, Michael Scharfschwerdt², Anja Hennemuth³, 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 physiologically shaped sinus prosthesis (Uni-Graft® W SINUS, Braun) should preserve near-normal pressure gradients across the aortic root. By means of 4D Flow MRI we analyzed transvalvular pressure gradients and their time course in 6 patients with sinus prosthesis in comparison to 12 healthy volunteers and pressure gradients derived from follow-up echocardiography. Peak pressure gradients did not differ significantly between patients and volunteers or between MRI and echocardiography in patients. Temporal evolution of pressure gradients was consistent with normal pressure gradients. The Venturi effect could be demonstrated across the aortic valve with a pressure drop in the bulb in peak systole.

Seong-Eun Kim¹, John A. Roberts¹, J. Scott Mcnally¹, Bradley D. Bolster, Jr.², Gerald S. Treiman^3,4, and Dennis L. Parker^1
1UCAIR, Department of Radiology, University of Utah, Salt Lake City, Utah, United States, ²Siemens Healthcare, Salt Lake City, Utah, United States,³Department of Surgery, University of Utah, Salt Lake City, Utah, United States, ⁴Department of Veterans Affairs, VASLCHCS, Salt Lake City, Utah, United States

Cardiac pulsations cause blurring of wall morphology. Gated acquisition increases scan time and results in image degradation due to non-constant TR. We have developed a cardiac or respiratory ciné reconstruction method using retrospective ordering and compressed sensing (ciné-ROCS) to minimize, characterize, and eliminate the artifacts. A Radial based k-space trajectory may offer reduced motion sensitivity, reduced artifacts and more robust ciné-ROCS reconstructions due to its inherent oversampling of central k-space. The goal in this study was to incorporate ciné-ROCS into 3D Stack of Stars sequence to demonstrate improvements for reliable plaque imaging and individual plaque component identification.

Zhaoyang Fan¹, Qi Yang^1,2, and Debiao Li^1
1Biomedical Imaging Research Institute, Cedars-Sinai Medical Center, Los Angeles, California, United States, ²Radiology, Xuanwu Hospital, Beijing, China

High resolution dark-blood MRI has been used to evaluate the intracranial arterial wall and provide findings indicative of wall thickening, intraplaque hemorrhage, or inflammation. T1-weighted 3D variable-flip-angle TSE has gain popularity in this research area but involves limitations including obscured outer wall boundary by CSF and limited spatial coverage. This work developed a 3D TSE technique that provides CSF attenuated 0.5-mm isotropic resolution whole-brain wall imaging within a reasonable scan time. With enhanced T1 weighting, the technique would be useful in the assessment of both plaque burden and high-T1-signal wall abnormalities in the incracranial artery wall.

M.G.M. van de Steeg^1,2, M. Henningsson², A. Noorani², K. Nicolay¹, and R. Botnar^2
1Division of Molecular Bioengineering and Molecular Imaging, Eindhoven University of Technology, Eindhoven, Netherlands, ²Division of Imaging Sciences and Biomedical Engineering, King's College London, London, United Kingdom

T2prepared Phase Sensitive Inversion Recovery (T2prep-PSIR) shows an enhanced distinction between vessel wall and surrounding tissue in black-blood MRI. This is especially useful in patients with aortic dissection. However, significant flow artifacts may occur when the T2prep is performed in a cardiac phase with high blood flow, even when performed in early systole. We developed an adapted T2prep-PSIR, in which a T2prep is performed during slow blood flow in end-diastole and compared it with conventional T2prep-PSIR. Assessment was based on visual score and compared to blood flow curves. The adapted sequence shows an improved visualization of the aortic vessel wall.

Shan Gao¹, Bram F. Coolen², Rob J. van der Geest¹, Dirk H.J. Poot^3,4, and Aart J. Nederveen^2
1Division of Image Processing, Department of Radiology, Leiden University Medical Center, Leiden, Netherlands, ²Radiology, Academic Medical Center, Amsterdam, Netherlands, ³Biomedical Imaging Group Rotterdam, Erasmus MC Rotterdam, Rotterdam, Netherlands, ⁴Imaging Science and Technology, Delft University of Technology, Delft, Netherlands

Carotid artery T1 and T2 mapping requires careful registration to deal with motion between the individual quantitative scans. Furthermore T1 and T2 estimates rely on accurate vessel wall segmentation. In this study we investigated whether incorporating an additional 3D black-blood reference scan in the registration and segmentation pipeline improves carotid artery T1 and T2 quantification.

Olivia Palmer¹, Amos Cao², Ulrich Scheven², Jose A Diaz³, and Joan M Greve^2
1Biomedical Engineering, University of Michigan, Ann Arbor, MI, United States, ²Biomedical Engineering, University of Michigan, MI, United States, ³Surgery, Section of Vascular Surgery, Conrad Jobst Vascular Research Lab, University of Michigan, MI, United States

Cardiovascular disease is the number one killer in the United States and worldwide. The venous system is understudied, a discrepancy which must be corrected as the US population with deep vein thrombosis is projected to double by 2050. Flow and pressure are known to differ in the venous system compared to the arterial system. Wall shear stress and wall motion are not well-defined. This work focuses on the last, comparing lumen expansion along the venous system of murine models. Lumen expansion was independent of gender and maximum values were less than half the expansion typically seen in the arterial system.

Raf H.M. van Hoof^1,2, Evelien Hermeling^1,2, Julie Salzmann³, Judith C. Sluimer^2,4, Sylvia Heeneman^2,4, Arnold P.G. Hoeks^2,5, Harry A.J. Struijker-Boudier^2,6, Jérôme Roussel³, Joachim E. Wildberger^1,2, and M. Eline Kooi^1,2
1Radiology, Maastricht University Medical Center, Maastricht, Netherlands, ²Cardiovascular Research Institute Maastricht (CARIM), Maastricht University, Maastricht, Netherlands, ³Institut de Recherches Internationales Servier, Suresnes, France, ⁴Pathology, Maastricht University Medical Center, Maastricht, Netherlands, ⁵Biomedical Engineering, Maastricht University Medical Center, Maastricht, Netherlands, ⁶Pharmacology, Maastricht University Medical Center, Maastricht, Netherlands

The effect of Ivabradine, a heart rate lowering drug, on the formation of atherosclerosis in cholesterol-fed rabbits was investigated. Plaque size and microvasculature was measured using (Dynamic Contrast-Enhanced) MRI and plaque biomechanics using ultrasound imaging. Results showed no difference in plaque size and beat-to-beat biomechanics, but a decrease in heart rate and plaque microvasculature determined using semi-quantitative (Area-Under-the-Curve) and quantitative (K^trans) parameters. Thus, use of Ivabradine led to decreased plaque microvasculature, which is thought to be an important determinant of reduced plaque vulnerability.

Let It Flow

Thursday 4 June 2015

Susanne Schnell¹, Can Wu^1,2, Ian G Murphy¹, Julio Garcia¹, and Michael Markl^1,2
1Radiology, Northwestern University, Chicago, Illinois, United States, ²Biomedical Engineering, Northwestern University, Evanston, Illinois, United States

A newly developed dual-venc 4D flow MRI sequence was applied in healthy volunteers. The k-t GRAPPA accelerated dual-venc 4D flow MRI sequence was developed using a shared reference scan for both low- and high-venc acquisition within a single scan. The resulting dual-venc data was combined using the low- and high-venc data, whereby the high-venc scan was solely used for anti-aliasing purposes. The findings of this feasibility study showed that dual-venc 4D flow MRI can provide improved visualization and quantification of venous and arterial hemodynamics across a wide range of the velocity spectrum. Velocity noise was reduced compared to single-venc implementations.

Claudio Santelli^1,2, Michael Loecher³, Julia Busch², Oliver Wieben^3,4, Tobias Schaeffter¹, and Sebastian Kozerke^1,2
1Imaging Sciences and Biomedical Engineering, King's College London, London, United Kingdom, ²Institute for Biomedical Engineering, University and ETH Zurich, Zurich, Switzerland, ³Department of Medical Physics, University of Wisconsin-Madison, Wisconsin, United States, ⁴Department of Radiology, University of Wisconsin-Madison, Wisconsin, United States

Iterative image reconstruction in which magnitude and phase are regularized separately was implemented to improve velocity vector field reconstruction from undersampled 3D flow MRI data by penalizing divergence of the measured flow field. Velocity data was regularized to reduce divergence, using either a finite difference (FD) or divergence-free Wavelets (DFW) method. They were tested on simulated and in-vivo data and compared to standard Compressed Sensing (CS) reconstruction. Phase regularization led to reduced directional error and divergence, increased streamline length and improved vector field visualization.

Michael J Rose¹, Kelly Jarvis^2,3, Varun Chowdhary², Alex J Barker², Bradley D Allen², Joshua D Robinson^4,5, Michael Markl^2,3, Cynthia K Rigsby^1,2, and Susanne Schnell^2
1Medical Imaging, Ann & Robert H. Lurie Children's Hospital of Chicago, Chicago, IL, United States, ²Radiology, Northwestern University, Chicago, IL, United States, ³Biomedical Engineering, Northwestern University, Chicago, IL, United States, ⁴Pediatrics, Northwestern University, Chicago, IL, United States, ⁵Pediatric Cardiology, Ann & Robert H. Lurie Children's Hospital of Chicago, Chicago, IL, United States

Currently, standard methods for measuring systolic peak velocity, an important clinical measure of aortic valve stenosis severity, are Doppler echocardiography and 2D phase contrast MRI. Both these methods, due to single-direction velocity encoding and reliance on manually placed 2D analysis planes, can underestimate true peak velocity in the case of complex flow jets as often seen in patients with aortic valve disease. 4D flow MRI, having three-directional velocity encoding, can better accommodate complex blood flow. In this study, we employ a new approach to measuring peak velocity using systolic velocity maximum intensity projections based on aortic 4D flow MRI data.

Susanne Schnell¹, Julio Garcia¹, Can Wu^1,2, and Michael Markl^1,2
1Radiology, Northwestern University, Chicago, Illinois, United States, ²Biomedical Engineering, Northwestern University, Evanston, Illinois, United States

A new Cartesian dual-venc phase-contrast MRI sequence was developed applying one reference scan and 6 velocity encoding gradients in three orthogonal directions using two different velocity encoding strengths. All 7 encodings are played within one cardiac phase and the shortest possible TR and TE. The gradient’s Maxwell terms are calculated and corrected during reconstruction. Rotation phantom experiments were performed to evaluate the combined anti-aliased high dynamic range velocity data with low velocity to noise ratio.

Kelly Jarvis^1,2, Susanne Schnell¹, Alex J Barker¹, James Carr¹, Joshua D Robinson^3,4, Cynthia K Rigsby^1,4, and Michael Markl^1,2
1Radiology, Northwestern University, Chicago, IL, United States, ²Biomedical Engineering, Northwestern University, Chicago, IL, United States, ³Pediatrics, Northwestern University, Chicago, IL, United States, ⁴Medical Imaging and Cardiology, Ann & Robert H Lurie Children’s Hospital of Chicago, IL, United States

Non-uniform distribution of caval blood flow to the left and right lung is suspected to cause complications in patients with Fontan circulation. Caval flow distribution can be comprehensively assessed using 4D flow MRI. However, 3D flow visualization of 4D flow data by traditional single pathline methods does not take into account the effects of velocity noise that may propagate over time-steps in the cardiac cycle. The aim of this study was to systematically analyze 3D probabilistic flow connectivity mapping of 4D flow data for taking into account the influence of noise in flow distribution quantification for patients with Fontan circulation.

Joseph Y Cheng^1,2, Kate Hanneman², Tao Zhang^1,2, Marcus T Alley², Peng Lai³, Jonathan I Tamir⁴, Martin Uecker⁴, Michael Lustig⁴, John M Pauly¹, and Shreyas S Vasanawala^2
1Electrical Engineering, Stanford University, Stanford, CA, United States, ²Radiology, Stanford University, Stanford, CA, United States, ³Global MR Applications & Workflow, GE Healthcare, Menlo Park, CA, United States, ⁴Electrical Engineering and Computer Sciences, University of California, Berkeley, CA, United States

Volumetric time-resolved phase contrast MRI (4D flow) permits evaluation of congenital heart disease (CHD). The ease of prescribing the scan and the ability to quantify function and flow make the technique compelling for CHD. However, a comprehensive CHD exam must also assess anatomy. Therefore, we evaluated the performance of a soft-gated compressed-sensing 4D flow technique by assessing anatomy. With IRB approval and informed consent, 23 consecutive patients were scanned using 4D flow with ferumoxytol enhancement. Images from the developed view-ordering scheme with soft-gating were determined to have improved morphologic diagnostic quality with no loss in flow quantification accuracy.

Michael Loecher¹, Kevin M Johnson¹, Patrick Turski², and Oliver Wieben^1,2
1Medical Physics, University of Wisconsin Madison, Madison, Wisconsin, United States, ²Radiology, University of Wisconsin Madison, Madison, Wisconsin, United States

Visualizing 4D flow MRI data with streamlines has become an important tool for assessing vascular hemodynamic. However streamlines can be severely degraded due to displacement artifacts from radial acceleration and misrepresent the actual flow field. This work tests the effectiveness of various correction techniques in reducing the effects of displacement artifacts. Results show significant improvements for iterative streamline based methods, as well as combing this with divergence-free algorithms.

Andreas Greiser¹, Christoph Forman¹, Jens Wetzel², Christoph Tillmanns³, Aurelien F. Stalder¹, Michaela Schmidt¹, Michael Zenge⁴, and Edgar Mueller^1
1Siemens AG, Healthcare, Imaging & Therapy Systems, Magnetic Resonance, Erlangen, Bavaria, Germany, ²Department of Computer Science, Friedrich-Alexander-Universität Erlangen-Nuernberg, Pattern Recognition Lab, Erlangen, Bavaria, Germany, ³Diagnostikum Berlin, Berlin, Germany, ⁴Siemens Healthcare, NY, United States

We validated the method of an optimized regularization in 2D k-t-sparse cine phase contrast in volunteers and patients and 4D PCI lasting 6 minutes. For 2D, average absolute difference between reference and accelerated was 3.10 ml for net forward volume (NFV) and 8.58 ml for vmax, systematic difference was 1.19%/-0.65% for vmax/NFV. In 4D average relative difference was -3.75%/-3.32% for vmax/NFV, average absolute difference was 7.91% and 5.23% for vmax/NFV. Short-breathhold single-slice and fast 4D flow acquisitions are feasible without apparent drawbacks. The inline reconstruction on a standard clinical scanner allows the use of iterative reconstruction techniques in clinical practice.

Merih Cibis¹, Wouter V Potters², Mariana Selwaness³, Frank J Gijsen¹, Andres M Arias Lorza⁴, Aad van der Lugt³, Aart J Nederveen², and Jolanda J Wentzel^1
1Biomedical Engineering, Erasmus MC, Rotterdam, Netherlands, ²Radiology, AMC, Amsterdam, Netherlands, ³Epidemiology, Erasmus MC, Rotterdam, Netherlands, ⁴Radiology and Medical Informatics, Erasmus MC, Rotterdam, Netherlands

Wall shear stress inside vessels is associated with atherosclerosis and calculated generally by computational fluid dynamics (CFD) and more recently by flow MRI data. It is of vital importance to know if CFD and MRI based WSS associate similarly with the early markers of atherosclerosis, such as wall thickening. We therefore compared CFD and MRI based WSS in elderly healthy population and also their association with wall thickness. We showed that both CFD and MRI based WSS results show an inverse relationship with WT in internal carotid artery but WSS in common carotid artery must be interpreted more carefully.

Aurélien Julien Trotier¹, Charles Castets¹, William Lefrancois¹, Jean-Michel Franconi¹, Eric Thiaudière¹, and Sylvain Miraux^1
1RMSB-UMR5536, CNRS - Université de Bordeaux, Bordeaux, Aquitaine, France

Time-Resolved Phase Contrast sequences is mostly performed in 2D or in thin 3D volumes with cartesian sequences. This strategy is not applicable on the whole mouse cardiovascular system due to the saturation of the spin into this volume. To overcome this problem we proposed to use a new 3D time-resolved Phase Contrast UTE sequence in combinaison with an injection of Ultra Small Particles of Iron Oxide to obtain a positive signal in blood. Method was optimized (Phase error correction, TE < 0.6 ms) to generate 4D low images at 7T with a spatial resolution of 156 µm. Flow measurements can be performed in numerous cardiac an pulmonary vessels of mice models.

Jennifer Keegan¹, Claire Raphael¹, Kim Parker², Robin Simpson³, Ranil de Silva¹, Carlo Di Mario¹, Julian Collinson⁴, Rod Stables⁵, Stephen Strain¹, Sanjay Prasad¹, and David Firmin^1,2
1Royal Brompton Hospital, London, United Kingdom, ²Imperial College, London, United Kingdom, ³Radiological Physics, Freiburg, Germany, ⁴Chelsea and Westminster Hospital, United Kingdom, ⁵Liverpool Heart and Chest Hospital, United Kingdom

The temporal pattern of coronary flow velocity through the cardiac cycle provides important information about coronary haemodynamics and atherosclerotic disease state. In this study, we assess the reproducibility of breath-hold high temporal resolution (19ms) spiral phase velocity mapping of coronary flow velocity and validate it against invasive ‘gold-standard’ Doppler flow wire measurements. As expected, MR velocities are lower than Doppler velocities but the temporal flow patterns are highly similar (R2 = .60 - .97). Inter breath-hold reproducibility of MR velocities is excellent. We conclude that spiral phase velocity mapping can assess temporal patterns of coronary flow velocity non-invasively.

Christian Binter¹, Alexander Gotschy^1,2, Robert Manka^1,3, Simon H. Sündermann⁴, and Sebastian Kozerke^1,5
1Institute for Biomedical Engineering, University and ETH Zurich, Zurich, Switzerland, ²Dept. of Internal Medicine, University Hospital Zurich, Switzerland,³Dept. of Cardiology, University Hospital Zurich, Switzerland, ⁴Division of Cardiovascular Surgery, University Hospital Zurich, Switzerland, ⁵Imaging Sciences and Biomedical Engineering, King's College London, United Kingdom

Phase-Contrast MRI allows for time-resolved quantification of velocities in the aortic arch and the assessment of the energy stored in turbulent flow. This information potentially offers a more direct assessment of the severity of aortic stenosis than parameters derived from Doppler echocardiography. The present study compares turbulence levels in patients before and after aortic valve replacement as well as age-matched controls and relates the findings to disease severity markers as determined by echocardiography. Significant differences between groups were found for both Turbulent Kinetic Energy and Mean Pressure Gradient.

Prashob Porayette¹, Christopher Macgowan², Sujana Madathil¹, Edgar Jaeggi¹, Lars Grosse-Wortmann¹, Shi-Joon Yoo³, John Kingdom⁴, Greg Ryan⁵, Steven Miller⁶, and Mike Seed^1
1Pediatric Cardiology, The Hospital for Sick Children, Toronto, ON, Canada, ²Physiology & Experimental Medicine, The Hospital for Sick Children, Toronto, ON, Canada, ³Diagnostic Imaging, The Hospital for Sick Children, Toronto, ON, Canada, ⁴Obstetrics & Gynaecology, Mount Sinai Hospital, Toronto, ON, Canada, ⁵Maternal-Fetal Medicine, Mount Sinai Hospital, Toronto, ON, Canada, ⁶Neurology, The Hospital for Sick Children, Toronto, ON, Canada

We have recently published the reference ranges for blood flow in normal human fetal circulation. In this study, we demonstrated the flow as well as the oxygen saturation measurements using novel MRI techniques in major vessels of fetuses with common cyanotic congenital heart disease.

Jan Paul¹, Stefan Wundrak¹, Peter Bernhardt¹, Wolfgang Rottbauer¹, Heiko Neumann², and Volker Rasche^1
1Internal Medicine II, University Hospital Ulm, Ulm, Germany, ²Institute of Neural Information Processing, University of Ulm, Ulm, Germany

Tissue Phase Mapping (TPM) allows analysis of global and regional LV motion and calculation of motion quantification parameters. Due to the long scan time of high-resolution or volumetric TPM acquisitions, respiratory motion has to be considered. In this study, image-based self-gating (SG) is compared to conventional pencil beam (RNAV) gating for TPM acquisitions. Additionally, the influence of the regularization strength  in the radial SPARSE SENSE reconstruction on the velocities is investigated.

Wouter V Potters¹, Merih Cibis², Frank JH Gijsen², Henk A Marquering^1,3, Ed vanBavel³, Jolanda J Wentzel², and Aart J Nederveen^1
1Radiology, Academic Medical Center, Amsterdam, Netherlands, ²Biomedical Engineering, Erasmus Medical Center, Rotterdam, Netherlands, ³Biomedical Engineering & Physics, Academic Medical Center, Amsterdam, Netherlands

Wall shear stress (WSS) and oscillatory shear stress (OSI) depend on the spatial and temporal resolution. We assessed these spatio-temporal dependencies using an in vitro carotid bifurcation phantom based on the geometry and the flow profile from a healthy volunteer. 36 2D PC MRI measurements with varying spatio-temporal resolution were analyzed using fully-automatic segmentation and a WSS calculation algorithm to retrieve WSS and OSI values. The results show that both OSI and WSS magnitude are heavily influenced by spatio-temporal resolution. OSI tends to be more effected by temporal resolution, whereas WSS magnitude tends to be more effected by spatial resolution.

Martin Fasshauer^1,2, Alexander L. Powell³, Alex J. Barker³, Susanne Schnell³, Joachim Lotz^1,2, and Michael Markl^3,4
1Institute for Diagnostic And Interventional Radiology, University Medical Center Goettingen, Goettingen, Lower saxony, Germany, ²German Center for Cardiovascular Research, DZHK, partner site Goettingen, Germany, ³Department of Radiology, Nortwestern University, Chicago, IL, United States,⁴Department of Biomedical Engineering, Northwestern University, Chicago, IL, United States

Thoracic aortic aneurysms are a life threatening disease due to the common absence of symptoms in most patients. Our study applies 4D flow MRI for assessment of complex flow patterns and comparison of 2D peak systolic velocity in 109 patients with dilated ascending aorta. Our study shows that 4D PC-MRI has the potential to yield more accurate peak velocities in complex flow environments.

Arun Antony Joseph¹, Martin Fasshauer², Klaus-Dietmar Merboldt³, and Jens Frahm^3
1Biomedizinsche NMR Forschungs GmbH am Max Planck Institut fuer biophysikalische Chemie, Goettingen, Niedersachsen, Germany, ²Abteilung Diagnostische Radiologie, Universitätsmedizin Goettingen, Niedersachsen, Germany, ³Biomedizinsche NMR Forschungs GmbH am Max Planck Institut fuer biophysikalische Chemie, Niedersachsen, Germany

Real-time PC imaging with highly undersampled radial FLASH and regularized nonlinear inversion reconstruction was used for the analysis of pulse wave velocity. Access to flow parameters from multiple cardiac cycles by real-time PC MRI provides accurate analysis of pulse wave velocity at short acquisition times. Pulse wave velocity values obtained from real-time PC MRI were further compared with Cine PC MRI.

Mehmet Akif Gulsun¹, Arne Littmann², Timothy Slesnick³, Ning Jin⁴, Andreas Greiser², Marie-Pierre Jolly¹, Gary McNeal⁴, and Aurelien F Stalder^2
1Imaging and Computer Vision, Siemens Corporate Technology, Princeton, NJ, United States, ²Siemens Healthcare, Erlangen, Germany, ³Emory University School of Medicine, Children's Healthcare of Atlanta, GA, United States, ⁴Siemens Healthcare, IL, United States

This work presents an automatic inline flow processing method for analysis of 2D PC MRI data. Flow images were automatically processed at the scanner for background phase correction, dynamic vessel segmentation, flow quantification and advanced visualization. The proposed method substantially reduces the complexity and time required for flow analysis and provides an effective way to review the flow analysis results just after the acquisition so as to repeat the acquisition while the patient is still in the scanner in case of inconsistency. Inline flow processing could contribute to making application of 2D PC in the clinical routine easier, faster and more accurate.

Peng Lai¹, Ann Shimakawa¹, Joseph Y Cheng², Marcus T Alley², Shreyas S Vasanawala², and Anja CS Brau^3
1Global MR Applications & Workflow, GE Healthcare, Menlo Park, CA, United States, ²Radiology, Stanford University, CA, United States, ³Global MR Applications & Workflow, GE Healthcare, Munich, Germany

Cardiac 4D flow with conventional parallel imaging and respiratory gating requires long scan time, substantially prolonging cardiac exam time. This work developed a new free-breathing 4D flow method with high scan efficiency and under 7min scan time. Our results show that the proposed approach can effectively suppress respiratory motion and achieve high acceleration and potentially enables pre-LGE 4D flow imaging with high blood signal.

Julio Garcia¹, Daniel Hirtler², Alex J Barker¹, and Julia Geiger^2,3
1Radiology, Northwestern University, Chicago, Illinois, United States, ²Congenital Heart Defects and Pediatric Cardiology, University Hospital Freiburg, Freiburg, Germany, ³Radiology, University Childrens’ Hospital Zurich, Zurich, Switzerland

Tetralogy of Fallot (TOF) is the most common cyanotic congenital heart defect. TOF patients usually undergo corrective surgery early in life. This evaluated flow disturbances due to vortex formation in the right atrium (RA) and right ventricle (RV) hemodynamics of TOF patients after repair (rTOF) using qualitative and quantitative vorticity.

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

Coil arrays with many elements are applied to enable higher acceleration factors in parallel imaging. However, the increased amount of data slows down the reconstruction process. Coil array compression has been introduced to combine the data into fewer ‘virtual coils’, thus reducing the amount of data while preserving most of the information. In this contribution, the influence of coil array compression on reconstructed velocities for Tissue Phase Mapping (TPM) is investigated.

Hideki Ota¹, Koichiro Sugimura², Haruka Sato², Kotaro Nochioka², Shunsuke Tatebe², Saori Yamamoto², Masanobu Miura², Kimio Satoh², Yuta Urushibata³, Yoshiaki Komori³, Aurelien F. Stalder⁴, Andreas Greiser⁴, Hiroaki Shimokawa², and Kei Takase^1
1Diagnostic Radiology, Tohoku University Hospital, Sendai, Miyagi, Japan, ²Cardiology, Tohoku University Hospital, Sendai, Miyagi, Japan, ³Siemens Japan K.K, Tokyo, Japan, ⁴Siemens Healthcare, Erlangen, Germany

In chronic thromboembolic pulmonary hypertension (CTEPH), percutaneous transluminal pulmonary angioplasty improves patients�f prognosis. There is a need to monitor mean pulmonary arterial pressure (mPAP) in long-term follow-up of patients who undergo angioplasty. Parameters of 4D flow and cardiac cine MR imaging were evaluated for a potential estimator of mPAP in 16 patients with CTEPH. Vortex flow in the pulmonary trunk was observed on 4D flow MR images in all patients. Backward flow area ratio in the end-systolic phase in the pulmonary trunk was the strongest estimator for both baseline mPAP before angioplasty and change of mPAP after angioplasty.

Eric Mathew Schrauben¹, Kevin M Johnson¹, Aaron Field², and Oliver Wieben^1,2
1Medical Physics, University of Wisconsin - Madison, Madison, Wisconsin, United States, ²Radiology, University of Wisconsin - Madison, Madison, Wisconsin, United States

Blood flow measurements using 4D flow MRI are compared between MS patients and healthy controls in cerebrospinal veins. No significant difference in flow pattern between the groups was found.

Jesús Urbina^1,2, Julio Sotelo^2,3, Cristian Montalba², Cristián Tejos^2,3, Pablo Irarrázaval^2,3, Marcelo Andía^2,4, Israel Valverde^5,6, and Sergio Uribe^2,4
1School of Medicine, Pontificia Universidad Católica de Chile, Santiago, Chile, ²Biomedical Imaging Center, Pontificia Universidad Católica de Chile, Santiago, Chile, ³Electrical Engineering Department, Pontificia Universidad Católica de Chile, Santiago, Chile, ⁴Radiology Department, Pontificia Universidad Católica de Chile, Santiago, Chile, ⁵Pediatric Cardiology Unit, Hospital Virgen del Rocio, Seville, Spain, ⁶Institute of Biomedicine of Seville, Universidad de Sevilla, Seville, Spain

Peak to peak pressure gradient through the aortic coarctation is the clinical standard to determine the severity of this disease and to refer patients to surgery repair. Pressure gradient can be measured with echocardiography and catheterization. During the last years, 4D flow has emerged as a MRI technique capable to measure hemodynamic parameters. The purpose of this work is to evaluate the accuracy of pressure gradients derived from 4D flow compared with catheterization in controlled experiments using a physiological healthy and aortic coarctation phantoms at rest and stress conditions.

New Insights & Innovations in Cardiovascular MRI

Thursday 4 June 2015

Dingxin Wang^1,2, Edward Auerbach², Gary McNeal³, Peter Kollasch¹, Uma Valeti⁴, Vibhas Deshpande⁵, Kamil Ugurbil², and Greg Metzger^2
1Siemens Healthcare, Minneapolis, Minnesota, United States, ²CMRR, Department of Radiology, University of Minnesota, Minneapolis, Minnesota, United States, ³Siemens Healthcare, Dallas, Texas, United States, ⁴Departments of Medicine and Radiology, University of Minnesota, Minneapolis, Minnesota, United States, ⁵Siemens Healthcare, Austin, Texas, United States

Our study demonstrates, for the first time, the feasibility of simultaneous multi-slice cardiac dark blood imaging using double-inversion recovery multiband slice accelerated TSE. Multiband slice acceleration improves the acquisition efficiency of TSE dark blood imaging beyond what is possible with standard in-plane parallel imaging acceleration thus allowing more slice coverage per breath hold and reducing the number of breath holds required to obtain whole heart coverage.

Choukri Mekkaoui¹, Timothy G Reese², Stephen F Cauley², Kawin Setsompop², Himanshu Bhat³, William J Kostis², Marcel P Jackowski⁴, and David E Sosnovik^2
1Harvard Medical School - Massachussetts General Hospital, Boston, MA, United States, ²Harvard Medical School-Massachusetts General Hospital, Boston, MA, United States, ³Siemens, Boston, MA, United States, ⁴University of São Paulo, São Paulo, Brazil

Successful clinical translation of cardiac Diffusion Tensor Imaging (DTI) will require efficient free-breathing techniques to be developed. We combined simultaneous multislice (SMS) excitation and a diaphragmatic navigator (NAV) to perform free-breathing DTI of the entire human heart in vivo in less than 25 minutes. DTI of the heart was performed using a NAV-based blipped-CAIPI stimulated echo sequence with rate 3 SMS excitation and spatiotemporal registration (STR), with minimal loss of image quality. The elimination of breath-holds in favor of a free-breathing NAV approach makes cardiac DTI accessible to patients with a broad range of cardiovascular conditions.

Karen Holst¹, Martin Ugander¹, and Andreas Sigfridsson^1
1Department of Clinical Physiology, Karolinska Institutet and Karolinska University Hospital, Stockholm, Sweden

Respiratory variation in ventricular shape may provide clinically important information, which can be measured using respiratory resolved imaging. Golden angle radial data were acquired during free breathing, and cardiac and respiratory self-gating signals were extracted. All radial spokes were sorted into combinations of respiratory and cardiac phases and left ventricular (LV) area was measured from end-diastolic images over 10 respiratory phases. Self-gating signals and LV area measurements showed good correspondence with similar images from real time cine. Respiratory induced changes in LV area could be detected and the golden angle method provided higher image quality and flexibility.

Fabian Kording¹, Bjoern Schoennagel¹, Friedrich Ueberle², Gunnar Lund¹, Gerhard Adam¹, and Jin Yamamura^1
1Department of Diagnostic and Interventional Radiology, University Medical Center Hamburg-Eppendorf, Hamburg, Hamburg, Germany, ²Faculty of Life Sciences, University of Aplied Sciences, Hamburg, Germany

Doppler Ultrasound and its potential to provide more details for cardiac MRI triggering was evaluated at 3T in comparison to VCG and pulse oximetry. Trigger signals of each method were acquired simultaneously for short axis cine and phase contrast sequences. Trigger sensitivity and RR interval length were compared between each method and showed high accuracy. Motion blurring was assess by endocardial border sharpness and showed a significant decrease for DUS triggered images during diastole. The appearance of DUS trigger signals correlated strong with quiescent heart phases which was successfully used for coronary MRA

Eric Mathew Schrauben¹, Christopher J François², Oliver Wieben^1,2, and Alejandro Roldán-Alzate^2
1Medical Physics, University of Wisconsin - Madison, Madison, Wisconsin, United States, ²Radiology, University of Wisconsin - Madison, Wisconsin, United States

4D flow MRI, gated to both respiratory and cardiac cycles, is used to assess respiratory plateau induced waveform changes in Great Vessels of healthy controls.

Helmar Waiczies¹, Jan Rieger¹, Armin M. Nagel², Andreas Graessl³, Lukas Winter³, and Thoralf Niendorf^3
1MRI.Tools GmbH, Berlin, Germany, ²Division of Medical Physics in Radiology, Cancer Research Center (DKFZ), Heidelberg, Germany, ³Berlin Ultrahigh Field Facility (B.U.F.F.), Max Delbrück Center for Molecular Medicine, Berlin, Germany

Sodium MRI (²³Na-MRI) is an emerging approach for gaining metabolic insights in Cardiovascular MRI. It has been shown to be suitable for the detection and assessment of acute and chronic heart disease due to increased sodium concentration after myocardial infarction. Therefore it is conceptually appealing to use the sensitivity gain intrinsic to ultrahigh fields for cardiac ²³Na-MRI. This work proposes a multi-channel transmit and receive (TX/RX) radiofrequency (RF) coil that supports four TX/RX channels (loop elements) for ²³Na and two 1H TX/RX channels (loopole elements) for cardiac imaging at 7.0T

CHARLES ANTHONY MISTRETTA¹, CHARLES STROTHER², and OLIVER WIEBEN^1
1MEDICAL PHYSICS AND RADIOLOGY, U OF WISCONSIN-MADISON, MADISON, WI, United States, ²RADIOLOGY, U OF WISCONSIN-MADISON, MADISON, WI, United States

We present a new imaging modality that embeds velocity or velocity-derived information from 4D flow MRI into time resolved 4D DSA volumes. This is done using a 2nd order constrained reconstruction in which the 4D DSA frames are formed using a 3D DSA constraint and the time-dependent angular projections acquired during iodine inflow. Then, a time averaged velocity map provided by PC-VIPR is used as a constraining image that is modulated by the 4D DSA temporal volumes. SNR and spatial resolution are provided by the 4D DSA data while color-coded velocity information is provided by the MR data.

Joep van Oorschot¹, Sanne Jansen of Lorkeers¹, Fredy Visser², Pieter Doevendans¹, Johannes Gho¹, Steven Chamuleau¹, Peter Luijten¹, and Jaco Zwanenburg^1
1University Medical Center Utrecht, Utrecht, Utrecht, Netherlands, ²Philips Healthcare, Best, Noord-Brabant, Netherlands

Native cardiac T2*- and T1ρ-mapping was performed in vivo in a porcine animal model of chronic myocardial infarction (MI), and correlated with in vivo gold standard LGE and ex vivo histology. A significant higher T1ρ relaxation time was found in the infarct region (61±11 ms), compared to healthy myocardium (36±4 ms), p<0.001. T2* was significantly lower in the infarct region (17±5 ms), compared to remote area (24±7 ms) (p<0.002). Iron staining showed an abundance of iron in the infarct area compared to remote myocardium, explaining the decrease in T2*. T1ρ-mapping is most promising for non-contrast enhanced detection of myocardial fibrosis.

Tsung-Lun Wu¹, Ching-Lung Cheng², Ming-Ting Wu^3,4, Ming-Long Wu^1,2, and Tzu-Cheng Chao^1,2
1Department of Computer Science and Information Engineering, National Cheng-Kung University, Tainan, Taiwan, ²Institute of Medical Informatics, National Cheng-Kung University, Tainan, Taiwan, ³Department of Radiology, Kaohsiung Veterans General Hospital, Kaohsiung, Taiwan, ⁴School of Medicine, National Yang-Ming University, Taipei, Taiwan

Cardiac CINE imaging has been extensively used in routine exams and it typically requires breath-hold for better image quality. Therefore a free breathing protocol may help to ease the patient¡¦s burden in this exam. In the present work, a strategy integrating GRAPPA and PROPELLER was proposed to improve self-gated free breathing CINE imaging, which enables to acquire a larger blade for better image-based motion tracing while retaining temporal resolution and reducing scan time. And the results show that the present work can accomplish cardiac CINE imaging for a slice with acceptable quality in less than 1 minute.

Rajiv G Menon¹, G Wilson Miller², Jean Jeudy¹, Sanjay Rajagopalan³, and Taehoon Shin^1
1Diagnostic Radiology and Nuclear Medicine, University of Maryland, Baltimore, Baltimore, MD, United States, ²Department of Radiology and Medical Imaging, University of Virginia, Charlottesville, Virginia, United States, ³Division of Cardiovascular Medicine, University of Maryland, Baltimore, Baltimore, MD, United States

The purpose of this study was to develop a free-breathing 3D Late Gadolinium Enhancement (LGE) Cardiac MR sequence and to test it on cardiac patients. The sequence consists of an Inversion pulse, Fat saturation, outer volume suppression as magnetization preparation followed by 1D projection navigators for motion correction, and 3D stack of spirals acquisition. The 3D LGE sequence and conventional 2D multi-slice breath-hold LGE were tested on 5 cardiac patients. Compared to conventional 2D LGE, the 3D LGE sequence gives better spatial coverage, less discomfort for patients, and takes less time, with comparable results. Testing on larger clinical cohort is under investigation.

Jacob Macdonald¹, Omid Forouzan², Jared Warczytowa², Oliver Wieben^1,3, Naomi Chesler², and Christopher Francois^3
1Medical Physics, University of Wisconsin - Madison, Madison, WI, United States, ²Biomedical Engineering, University of Wisconsin - Madison, Madison, WI, United States, ³Radiology, University of Wisconsin - Madison, Madison, WI, United States

The effect of exercise on cardiac function is important in the diagnosis of cardiovascular disease, but is clinically unavailable for MRI. In this pilot study, changes to cardiac function and aortic flow were investigated using specialized MRI-compatible exercise equipment that allowed healthy subjects to exercise in a supine position in the bore of the magnet. It was found that pulse wave velocity and cardiac output in the aorta, as well as ejection fraction and cardiac output in the left ventricle, showed statistically significant increases under exercise stress. Future studies will investigate the significance of these parameters in a diseased population.

Edward DiBella^1,2, Elwin Bassett³, Kyungpyo Hong^2,3, Ganesh Adluru³, Devavrat Likhite³, Promporn Suksaranjit⁴, Brent Wilson⁴, Chris McGann⁴, and Daniel Kim^2,3
1University of Utah, Salt Lake City, UT, United States, ²Bioengineering, University of Utah, Salt Lake City, Utah, United States, ³Radiology, University of Utah, Utah, United States, ⁴Cardiology, University of Utah, Utah, United States

Gating and breath-holding complicate acquisition of cardiac MRI data. A simpler streamlined pushbutton approach would significantly increase the clinical utility of cardiac MR. Methods for ungated acquisitions and retrospective self-gating were recently introduced. Here we propose and develop methods for a full CMR protocol (cine, perfusion, LGE, and T1/ECV mapping) that is ungated and free-breathing.

Amir Moussavi¹, Philipp R. Bovenkamp², Verena Hoerr², Cornelius Faber², and Susann Boretius^1
1Section Biomedical Imaging, Department of Radiology and Neuroradiology, Christian-Albrechts-University, Kiel, Germany, ²Institute of Clinical Radiology, University Hospital of Muenster, Muenster, Germany

Truly real-time imaging of mouse heart (temporal resolution = 41.25 ms, spatial resolution = 0.234 × 0.234 × 1.5 mm3) was accomplished by combining high field MRI and cyrocoil-technology with radial k-space encoding and nonlinear inverse reconstruction.

Mahdi Salmani Rahimi¹, Dominik Fleischmann¹, Anne Chin^1,2, and Roland Bammer^1
1Radiology, Stanford University, Stanford, CA, United States, ²Radiology, Centre hospitalier de l'Université de Montréal, Montreal, QC, Canada

Diffusion-weighted imaging of the heart is a challenging task due to signal losses associated with cardiac and respiratory movements. In this work, a restricted field of view RF excitation approach is combined with an automated timing mechanism to synchronize the diffusion encoding gradients with the most quiescent part of the cardiac cycle. Three small b-values and four large b-values were acquired during a single breath-holding period in healthy volunteers to probe the fast diffusing and slow diffusing compartments in the myocardium, respectively.

Sohae Chung^1,2, Pippa Storey^1,2, and Leon Axel^1,2
1Center for Advanced Imaging Innovation and Research (CAI2R), Department of Radiology, New York University School of Medicine, New York, NY, United States, ²Bernard and Irene Schwartz Center for Biomedical Imaging, Department of Radiology, New York University School of Medicine, New York, NY, United States

T1-mapping and the use of T1 changes from contrast enhancement for extracellular volume (ECV) quantification are clinically promising techniques for absolute quantification of myocardial fibrosis. However, whole-heart T1-mapping is rarely performed in clinical practice, due to the associated time-consuming data acquisition. Instead, it is usually performed at a single mid-ventricular level; this can lead to sampling error when the fibrotic process is not homogenous. Here, we present a rapid whole-heart T1-mapping in a single breath-hold of, e.g., 6 heartbeats (typically, 5-7 seconds for total 9 T1 maps) for the quantitative assessment of myocardial fibrosis.

Rachel Clough¹, Ondrej Holub², Henry Fok², Nick Gaddum², Jordi Alastruey², and Ralph Sinkus^2
1King's College London, London, London, United Kingdom, ²King's College London, London, United Kingdom

Aortic stiffness is an important risk factor for cardiovascular disease. Tonometry is the current standard but only provides spatially-averaged measurements. Assessment of specific aortic locations may be more important, particularly of the ascending aorta due the direct interaction this has with the heart. Magnetic resonance elastography measures material properties by studying the way shear waves travel through biological tissues. The aim of this study was to develop a new transducer-free MRE sequence to measure aortic stiffness in patients, using aortic valve closure, an intrinsic source for elastography, to generate shear waves in the aortic wall.

Anou Sewonu^1,2, Ramiro Moreno^1,2, Olivier Meyrignac³, and Hervé Rousseau^3
1I2MC, INSERM/UPS UMR 1048, Toulouse, France, ²ALARA Expertise, Strasbourg, France, ³Pôle imagerie, CHU Toulouse, Toulouse, France

The Multilayer Flow Modulator (MFM) has been introduced as an alternative minimal invasive treatment for arterial and aortic aneurysms. The concept of flow modulation is to reduce flow velocity and vorticity in the aneurysmal circulating volume, while improving laminar flow in collateral arteries. Although the concept had been proven, treatment in certain clinical cases was not conclusive. In this work, we demonstrate that a combination of pre-operatory velocity-encoded MRI and Computed Fluid Dynamics is an excellent tool for forecasting the becoming of a MFM stent graft.

Niema M. Pahlevan^1,2, Thao T. Tran³, Peyman M. Tavallali⁴, Derek G. Rinderknecht⁵, Marie Csete⁶, and Morteza M. Gharib^4
1Medical Engineering, California Institute of Technology, Pasadena, California, United States, ²Magnetic Resonance Spectroscopy, Huntington Medical Research Institute, Pasadena, California, United States, ³Magnetic Resonance Spectroscopy, Huntington Medical Research Institutes, Pasadena, CA, United States,⁴Graduate Aerospace Laboratory, California Institute of Technology, Pasadena, California, United States, ⁵Aerospace, California Institute of Technology, Pasadena, California, United States, ⁶Huntington Medical Research Institutes, Pasadena, CA, United States

We recently developed mechanical biomarkers of cardiovascular function, intrinsic frequencies (ω1 and ω2), calculated only from a carotid pulse waveform captured using an iPhone. We compare left ventricle ejection fraction (LVEF) derived from intrinsic frequencies vs. from cardiac MRI as a gold standard. LVEF computed from intrinsic frequency parameters and compared with LVEF measured from MRI. There was strong agreement between the LVEF measurements with average error of 9% and SD of 8%.

Yifei Liu¹, Thomas J Royston^1,2, and E Douglas Lewandowski^3,4
1Department of Mechanical & Industrial Engineering, University of Illinois at Chicago, Chicago, Illinois, United States, ²Department of Bioengineering, University of Illinois at Chicago, Chicago, Illinois, United States, ³Center for Cardiovascular Research, University of Illinois at Chicago, Chicago, Illinois, United States, ⁴Department of Physiology & Biophysics and Medicine (Cardiology), University of Illinois at Chicago, Chicago, Illinois, United States

Increased stiffness of the left ventricle (LV) wall is a contributing factor to the abnormal function associated with both impaired systolic and diastolic function during the progression of cardiomyopathies to overt heart failure. The mouse is a common animal model for studying the progression of cardiac pathologies and is readily for the application of MR Elastography, a non-invasive method to estimate stiffness of tissue. This study demonstrates the feasibility of in vivo cardiac MR Elastography (MRE) on the mouse heart to monitor myocardial stiffness, and the stiffness ratio between end-diastole and end-systole.

Daniel Beauchamp^1,2, Steven G Lloyd^3,4, Michael Allon³, Timmy Lee³, Nouha Salibi^1,5, and Thomas S Denney Jr.^1,2
1AU MRI Research Center, Auburn University, Auburn, Alabama, United States, ²Electrical and Computer Engineering, Auburn University, Auburn, Alabama, United States, ³Department of Medicine, University of Alabama at Birmingham, Birmingham, Alabama, United States, ⁴VA Medical Center, Birmingham, Alabama, United States, ⁵MR R&D, Siemens Healthcare, Malvern, Pennsylvania, United States

Arteriovenous grafts (AVGs) used for hemodialysis vascular access have a high failure rate due to thrombosis arising from stenosis around the venous anastomosis. It is thought that this stenosis may result from patterns of low wall shear stress (WSS) in the anastomosis region. This study used CFD analysis and phase-contrast MRI scans of flow phantoms to examine WSS in a geometry resembling an AVG, in order to determine whether low WSS is in fact present in the region of frequent problematic stenosis formation. These techniques can potentially be used to predict and prevent graft failure due to thrombosis.

Chia-Ying Liu¹, David A Bluemke¹, Gary Gerstenblith², Stefan L Zimmerman², Ji li², hong zhu³, Shenghan Lai², and Hong Lai^2
1Radiology and Imaging Sciences, NIH, Bethesda, MD, United States, ²Johns Hopkins School of Medicine, MD, United States, ³johns Hopkins School of Medicine, MD, United States

The objectives were to explore the factors that are associated with myocardial triglyceride, and to examine whether myocardial triglyceride is associated with ventricular function in healthy African Americans (AAs). Proton magnetic resonance spectroscopy was performed to noninvasively quantify myocardial triglyceride content on 92 AAs using a 3T MRI scanner. Obesity was shown to be associated with cardiac steatosis, and myocardial triglyceride content was shown to be associated with LV regional function in women. These findings suggest a link between obesity, cardiac steatosis, and subclinical cardiac dysfunction.

Michael S Hansen¹, Russel R Cross², Laura J Olivieri^1,2, Kendall O'Brien^1,2, Hui Xue¹, Matthew R DiPrimio³, Paul Taylor³, Tsinghua Zheng³, Xiaoyu Yang³, Matthew Finnerty³, and Peter Kellman^1
1National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, MD, United States, ²Children's National Medical Center, Washington, D.C., United States, ³Quality Electrodynamics, Mayfield Village, OH, United States

Neonatal cardiac MRI is typically done with general-purpose coils that are not optimized for cardiac imaging. With the aim of improving pediatric cardiac MRI, a twelve channel dedicated neonatal cardiac coil has been developed and is presented along with preliminary testing in phantoms and an in vivo imaging example. The coil enables parallel imaging with high parallel imaging factors and good g-factor performance. Real-time cine imaging with a temporal resolution of 40ms per frame is demonstrated in a small infant.

Jianing Pang¹, Behzad Sharif¹, Zhaoyang Fan¹, Xiaoming Bi², Reza Arsanjani¹, Daniel S Berman¹, and Debiao Li^1,3
1Biomedical Imaging Research Institute, Cedars-Sinai Medical Center, Los Angeles, CA, United States, ²MR R&D, Siemens Healthcare, Los Angeles, CA, United States, ³Medicine and Bioengineering, University of California, Los Angeles, California, United States

In this work, we perform preliminary comparisons between a recently proposed self-gated 4D coronary MRA technique, which offers simultaneous cardiac function and coronary imaging from a single measurement, with conventional cine and coronary MRA protocols. The cardiac function parameters obtained from the proposed technique are in good agreement with those from the conventional protocols. The coronary visualization quality of the proposed technique also compares favorably to the ECG-gated protocols.

Ian Gavin Murphy¹, Alex J Barker², Michael Markl², Chris memorial Malaisrie³, Patrick M McCarthy³, Colleen memorial Clennon⁴, James C Carr¹, and Jeremy Collins^1
1Cardiovascular Imaging, Feinberg School of Medicine, Northwestern Memorial Hospital, CHICAGO, ILLINOIS, United States, ²Cardiovascular Imaging, Northwestern University, CHICAGO, ILLINOIS, United States, ³Cardiothoracic Surgery, Feinberg School of Medicine, Northwestern Memorial Hospital, CHICAGO, ILLINOIS, United States, ⁴Cardiothoracic Specialist Nurse, Feinberg School of Medicine, Northwestern Memorial Hospital, CHICAGO, ILLINOIS, United States

Bicupsid aortic valve has many phenotypes and the resulting haemodynamic disturbance may be associated with differing aortopathies. Using in plane cine bSSFP and PC imaging at the aortic valve, we present a robust and reproducible classification system for subcategorization of the aorti