Clara J. Fallone¹, Anthony G. Tessier^1,2, Catherine J. Field³, and Atiyah Yahya^1,2
1Department of Oncology, University of Alberta, Edmonton, AB, Canada, ²Department of Medical Physics, Cross Cancer Institute, Edmonton, AB, Canada, ³Department of Agricultural, Food and Nutritional Science, University of Alberta, Edmonton, AB, Canada

Omega-3 (ω-3) fat intake is important to disease and reflected in adipose tissue composition. Optimized PRESS (Point RESolved Spectroscopy) at 9.4T is used to quantify relative ω-3 fat content in adipose tissue of mice and rats fed a 16% ω-3 diet and in a control diet mouse. The actual ω-3 fat content in the mice was determined using gas chromatography and compared to that obtained using optimized MRS (R² = 0.98). On average, the % ω-3 content measured with MRS in rat adipose tissue was approximately 39% higher than that in mouse.

Gregory Simchick^1,2, Ruiyang Zhao^1,2, Gavin Hamilton³, Scott Reeder^1,2,4,5,6, and Diego Hernando^1,2,4
1Radiology, University of Wisconsin-Madison, Madison, WI, United States, ²Medical Physics, University of Wisconsin-Madison, Madison, WI, United States, ³Radiology, University of California-San Diego, San Diego, CA, United States, ⁴Biomedical Engineering, University of Wisconsin-Madison, Madison, WI, United States, ⁵Medicine, University of Wisconsin-Madison, Madison, WI, United States, ⁶Emergency Medicine, University of Wisconsin-Madison, Madison, WI, United States

The precision (repeatability and reproducibility) of quantitative stimulated echo acquisition mode (STEAM) MRS was evaluated in subjects with known or suspected liver iron overload at both 1.5T and 3T. The test-retest repeatability and reproducibility across acquisition modes (multi-TE vs multi-TE-TR) of STEAM-MRS multi-parametric quantification was evaluated using linear regression and Bland-Altman analyses. STEAM-MRS demonstrated high precision for several biomarkers, including R2_water and R2_fat, FWHM_water and FWHM_fat as surrogates for R2*, and PDFF, at both 1.5T and 3T. R1_water and R1_fat exhibited moderate repeatability. The precision of the water estimates (R1, R2, and FWHM) was relatively similar at both field strengths.

Sevgi Emin¹, Jonas Svensson^1,2, Martin Englund³, and Pernilla Peterson^1,2
1Medical Radiation Physics, Department of Translational Medicine, Lund University, Malmo, Sweden, ²Medical Imaging and Physiology, Skane University Hospital, Lund, Sweden, ³Orthopaedics, Department of Clinical Sciences Lund, Lund University, Lund, Sweden

Chemical shift-encoded magnetic resonance imaging may be used to estimate the fatty acid composition (FAC) of bone marrow adipose tissue (BMAT), and the method may benefit from the improved spectral resolution at ultra-high field strength. In this work, the in vivo feasibility of FAC quantification in BMAT at 7 T was investigated, and simulations were used to determine the optimal image acquisition parameters and reconstruction model. The study showed that using a short inter-echo time, FAC quantification in BMAT was feasible and robust at 7 T, and that the noise performance might be improved using a constrained reconstruction model.

Mingming Wu¹, Cora Held¹, Maximilian N. Diefenbach^1,2, Jakob Meineke³, Aashley S.D. Sardjoe Mishre^4,5, Kilian Weiss⁶, Hermien E. Kan⁵, Daniela Junker¹, Hans Hauner^7,8, and Dimitrios C. Karampinos^1
1Department of Diagnostic and Interventional Radiology, School of Medicine, Technical University of Munich, Munich, Germany, ²Division of Infectious Diseases and Tropical Medicine, University Hospital, LMU Munich, Munich, Germany, ³Philips Research Lab, Hamburg, Germany, ⁴Division of Endocrinology and Einthoven Laboratory for Experimental Vascular Medicine, Department of Medicine, Leiden University Medical Center, Leiden, Netherlands, ⁵Department of Radiology, C.J. Gorter Center for High Field MRI, Leiden University Medical Center, Leiden, Netherlands, ⁶Philips Healthcare, Hamburg, Germany, ⁷Else Kröner Fresenius Center for Nutritional Medicine, School of Life Sciences, Technical University of Munich, Munich, Germany, ⁸Institute for Nutritional Medicine, School of Medicine, Technical University of Munich, Munich, Germany

Proton density fat fraction and T2* mapping have been used to characterize fat tissue in the human supraclavicular fossa with the aim of detecting brown adipose tissue and its response to activation by changes of the two. However, chemical shift encoding-based water-fat separation in that region has been primarily performed in free-breathing mode. The present work reports on breathing-induced B₀ fluctuations in the human supraclavicular fossa, and the severe bias introduced on both PDFF and T2* quantification, as shown with simulated B₀ fluctuation effects.  The effect of respiratory triggering on artefact reduction is investigated in a cohort of 13 volunteers.

Juan Pablo Meneses^1,2, Cristobal Arrieta^1,2, Gabriel della Maggiora^1,2, Pablo Irarrazaval^1,2,3,4, Cristian Tejos^1,2,4, Marcelo Andia^1,2,5, Carlos Sing Long^1,2,3,6,7, and Sergio Uribe^1,2,5
1Biomedical Imaging Center, Pontificia Universidad Catolica de Chile, Santiago, Chile, ²ANID – Millennium Science Initiative Program – Millennium Nucleus for Cardiovascular Magnetic Resonance, Santiago, Chile, ³Institute for Biological and Medical Engineering, Pontificia Universidad Catolica de Chile, Santiago, Chile, ⁴Department of Electrical Engineering, Pontificia Universidad Catolica de Chile, Santiago, Chile, ⁵Radiology Department, School of Medicine, Pontificia Universidad Catolica de Chile, Santiago, Chile, ⁶Institute for Mathematical & Computational Engineering, Pontificia Universidad Catolica de Chile, Santiago, Chile, ⁷ANID – Millennium Science Initiative Program – Millennium Nucleus for Discovery of Structure in Complex Data, Santiago, Chile

Proton density fat fraction (PDFF) and $$$R_2^*$$$, key biomarkers associated to liver disease, can be obtained by solving the water/fat separation problem. Convolutional Neural Networks (CNN) have been proposed for solving this problem. However, proposed solutions have not achieved accurate $$$R_2^*$$$ mapping. We introduce MDWF-Net, a CNN model for computing high quality water-fat, $$$R_2^*$$$ and field mapping from abdominal acquisitions. The results were evaluated considering error and structural similarity and compared against a U-Net. We also evaluated ROIs in the liver for PDFF and $$$R_2^*$$$. The proposed MDWF-Net overperforms the original U-Net, especially for $$$R_2^*$$$ maps, even with fewer echoes.

Jong Bum Son¹, David Rauch¹, Bikash Panthi¹, Zijian Zhou¹, Benjamin Musall¹, Marion Scoggins², Mark Pagel³, and Jingfei Ma^1
1Imaging Physics Department, The University of Texas MD Anderson Cancer Center, Houston, TX, United States, ²Diagnostic Radiology Department, The University of Texas MD Anderson Cancer Center, Houston, TX, United States, ³Cancer Systems Imaging Department, The University of Texas MD Anderson Cancer Center, Houston, TX, United States

We applied a video processing deep-learning neural network (originally developed to synthesize high resolution photorealistic video from a time series of dancing poses, semantically segmented street-view labels, or human face outline sketches) to Dixon imaging to present combined benefits of 2D and 3D networks. The developed Dixon Video Domain Transfer Generative Adversarial Network (DixonVDTGAN) could create slice-to-slice consistent water images with reduced demand on GPU memory. It could also successfully correct deep-learning processing errors for robust water and fat signal separation under two assumptions that the deep-learning processing errors are localized and the image phase is spatially smooth.

Vadim Malis¹, Asako Yamamoto¹, Yoshimori Kassai², Won Bae¹, Marianne McDonald³, and Mitsue Miyazaki^1
1Radiology, UC San Diego, San Diego, CA, United States, ²Canon Medical, Tochigi, Japan, ³UC San Diego, San Diego, CA, United States

In general, orbit fat suppression is not achieved due to the presence of unsaturated fat signals near water resonance. We have developed PASTA with unsaturated and saturated fat opposed phase (USOP) like a Dixon method. Semi-quantitative analysis shows about 6 – 18% of unsaturated fat in the orbits. Interestingly, two subjects having lazy eye (amblyopia) with double visions (diplopia) shows decreased in unsaturated fat signal.

Chuanli Cheng¹, Zhiming Wang¹, Qian Wan¹, Yangzi Qiao¹, Changjun Tie¹, Hairong Zheng¹, Xin Liu¹, and Chao Zou^1
1Shenzhen Institutes of Advanced Technology,Chinese Academy of Sciences, Shenzhen, China

Over the past several decades, the worldwide obesity epidemic has become a significant public health. As a consequence, accurate measurement of obesity is critical for obesity management. In the present study, an automated algorithm is proposed to segment subcutaneous adipose tissue (SAT) and internal adipose tissue (IAT) from the total adipose tissue for whole-body fat distribution analysis using proton density fat fraction (PDFF) magnetic resonance images. The dice coefficient of the network achieved 97.6%, and the processing time was less than 0.1s/image.

Guanwu Li¹, Wenshen Zhou², Shixin Chang¹, Dongmei Wu³, and Yongming Dai^2
1Department of Radiology, Yueyang Hospital of Integrated Traditional Chinese and Western Medicine, Shanghai University of Traditional Chinese Medicine, Shanghai, China, ²Central Research Institute, United Imaging Healthcare, Shanghai, China, ³Shanghai Key Laboratory of Magnetic Resonance, School of Physics and Electronics Science, East China Normal University, Shanghai, China

Marrow proton density fat fraction (PDFF) and R2* at the lumbar spine in postmenopausal females of widely varying bone density were simultaneously measured with a 3D Fat Analysis & Calculation Technique (FACT) Sequence at 3.0T. Discrimination of osteopenia and osteoporosis can be improved with a combination of the PDFF and R2*.

Qinjia Bao¹, Xiaojun Li², Kewen Liu², Zhao Li³, Hongxia Xiong², Jingjie Yan⁴, Yalei Chen², and Chaoyang Liu^3
1Weizmann Institute of Science, Rehovot, Israel, ²School of Information Engineering, Wuhan University of Technology, Wuhan, China, ³State Key Laboratory of Magnetic Resonance and Atomic and Molecular Physics, Wuhan Center for Magnetic Resonance, Wuhan Institute of Physics and Mathematics, Innovation Academy for Precision Measurement Science and Technology, Wuhan, China, ⁴Huazhong University of Science and Technology, Wuhan, China

We proposed a novel deep learning network for water-fat separation from undersampled mGRE data. The network contains three components: The first is the reconstruction module, which can effectively take advantage of the similarity between different echoes to recover the fully sampled image from the undersampled data; the second is the feature extraction module, which learns the correlations between consecutive echoes; and the third is the water-fat separation module that processes the feature information extracted from the feature extraction module. The results show that the proposed network can effectively obtain high-quality water and fat images at 6 times acceleration.

Ruvini Navaratna^1,2, Ruiyang Zhao^1,2, Timothy J Colgan², Houchun Harry Hu³, Mark Bydder⁴, Takeshi Yokoo⁵, Mustafa R Bashir^6,7,8, Michael S Middleton⁹, Suraj D Serai¹⁰, Dariya Malyarenko¹¹, Thomas Chenevert¹¹, Mark Smith³, Walter Henderson⁹, Gavin Hamilton⁹, Yunhong Shu¹², Claude B Sirlin⁹, Jean Tkach¹³, Andrew T Trout^13,14, Jean H Brittain¹⁵, Diego Hernando^1,2, and Scott B Reeder^{1,2,16,17,18
1}Medical Physics, University of Wisconsin - Madison, Madison, WI, United States, ²Radiology, University of Wisconsin - Madison, Madison, WI, United States, ³Radiology, Nationwide Children's Hospital, Columbus, OH, United States, ⁴Radiological Sciences, University of California - Los Angeles, Los Angeles, CA, United States, ⁵Radiology, University of Texas - Southwestern Medical Center, Dallas, TX, United States, ⁶Radiology, Duke University Medical Center, Durham, NC, United States, ⁷Division of Hepatology, Duke University Medical Center, Durham, NC, United States, ⁸Center for Advanced Magnetic Resonance Development, Duke University Medical Center, Durham, NC, United States, ⁹Radiology, University of California - San Diego, San Diego, CA, United States, ¹⁰Radiology, Children's Hospital of Philadelphia, Philadelphia, PA, United States, ¹¹Radiology, University of Michigan, Ann Arbor, MI, United States, ¹²Radiology, Mayo Clinic, Rochester, MN, United States, ¹³Radiology, Cincinnati Children's Hospital and Medical Center, Cincinnati, OH, United States, ¹⁴Radiology, University of Cincinnati College of Medicine, Cincinnati, OH, United States, ¹⁵Calimetrix, LLC, Madison, WI, United States, ¹⁶Biomedical Engineering, University of Wisconsin - Madison, Madison, WI, United States, ¹⁷Medicine, University of Wisconsin - Madison, Madison, WI, United States, ¹⁸Emergency Medicine, University of Wisconsin - Madison, Madison, WI, United States

Chemical shift-encoded (CSE) MRI is a well-established technique to quantify proton density fat-fraction (PDFF) as a quantitative imaging biomarker of hepatic steatosis. However, temperature is known to affect the accuracy and precision of PDFF quantification in phantoms. Previous work has demonstrated the benefit of temperature-correction using magnitude-based CSE-MRI PDFF estimation. However, the effect of temperature correction on complex- and hybrid-based CSE-MRI PDFF estimation is unknown. In this study­­­, we aim to compare the effects of PDFF temperature-correction for magnitude-based, complex-based, and hybrid-based CSE-MRI using simulations, temperature-controlled phantom experiments, and a multi-center phantom study.

Ali Pirasteh¹, Srijyotsna Volety¹, Matthias R Muhler², Diego Hernando¹, and Scott B Reeder^3
1Radiology and Medical Physics, University of Wisconsin-Madison, Madison, WI, United States, ²Radiology, University of Wisconsin-Madison, Madison, WI, United States, ³Radiology, Medical Physics, Biomedical Engineering, Medicine, Emergency Medicine, University of Wisconsin-Madison, Madison, WI, United States

We evaluated the robustness of measuring liver and bone marrow PDFF by chemical shift encoded (CSE) MRI measured before and 1-3 days after intravenous (IV) administration of ultra-small superparamagnetic iron oxide (USPIO) particles, i.e., ferumoxytol, as IV contrast. Among the eight included subjects, there was strong linearity and no significant bias between pre- and post-ferumoxytol CSE-MRI liver PDFF (R2=0.99, slope=0.99 and intercept=0.38%; mean bias [limits of agreement]=0.4% [±1.7] at 1.5T, and 0.2% [±1.7] at 3.0T. However, CSE-MRI bone marrow PDFF suffered from bias and overestimation post-ferumoxytol (R2=0.85, slope=0.85, intercept=12.8%, mean bias [LOA]=5.4% [±6.2] at 1.5T and 4.4% [±6.5] at 3.0T).

Venkata Veerendranadh Chebrolu¹, Peter Kollasch¹, Daniel B Spence², Joel P Felmlee², Benjamin M Howe², Matthew A Frick², and Kimberly K Amrami^2
1Siemens Medical Solutions USA, Inc., Rochester, MN, United States, ²Department of Radiology, Mayo Clinic, Rochester, MN, United States

In this work, we evaluate the efficacy of Spectral-Heterogeneity-Adaptive Radiofrequency Pulses (SHARP)-based partial fat saturation for improving meniscal pathology conspicuity at 3T and 7T. Under the guidelines of an IRB, 11 subjects with knee meniscal pathology detected on a previous 3T MRI were imaged at 3T and 7T using six different turbo spin-echo–based MRI imaging options that included SHARP. Two fellowship-trained musculoskeletal radiologists evaluated meniscal pathology conspicuity and selected the top three preferred options for each case. SHARP improved conspicuity of meniscal pathology at 3T and 7T and was within top three preferred options for the highest number of times. 

Chuanli Cheng¹, Qian Wan¹, Yangzi Qiao¹, Changjun Tie¹, Xin Liu¹, Hairong Zheng¹, and Chao Zou^1
1Shenzhen Institutes of Advanced Technology,Chinese Academy of Sciences, Shenzhen, China

Magnetic Resonance thermometry (MRT) in activated brown adipose tissue (BAT) is able to measure the BAT function directly. A previous proposed fat-referenced 1H proton resonance frequency shift MRT method is adopted to compare the thermogenic capacity of rats injected with different dose of norepinephrine (NE). It is found that the rats with 2mg/kg and 1mg/kg NE injection had comparable maximum temperature rises after NE injection (7.1^oC vs. 6.7^oC), while temperature rises for rats with NE of 0.5mg/kg was significantly lower (3.8^oC).

Lijie Zhong¹, Mainwen Li¹, Yukun Ding¹, Xintao Zhang¹, Yanjun Chen¹, Yingjie Mei², and Xiaodong Zhang^1
1Department of Medical Imaging, The Third Affiliated Hospital of Southern Medical University, Guangzhou, China, ²China International Center, Philips Healthcare, Guangzhou, China

Infrapatellar fat pad (IPFP) is an adipose structure of knee joint, which plays an important role in the progression of knee osteoarthritis(KOA). In this paper, we evaluate the fat content and composition of the IPFP in normal subjects and different stages of knee osteoarthritis patients using hydrogen proton MR spectroscopy (¹H-MRS). In addition, we investigate the relationship between the fat content and composition of the IPFP and Hoffa-synovitis. The results showed that the quantitative parameters of the IPFP by MRS were correlated with the progression of KOA and Hoffa-synovitis.

Hahnsung Kim^1,2, Lisa C. Krishnamurthy ^3,4, Jinsuh Kim⁵, and Phillip Zhe Sun^1,2
1Yerkes Imaging Center, Yerkes National Primate Research Center, Emory University, Atlanta, GA, United States, ²Department of Radiology and Imaging Sciences, Emory University School of Medicine, Atlanta, GA, United States, ³Center for Visual and Neurocognitive Rehabilitation, Atlanta VA, Decatur, GA, United States, ⁴Department of Physics & Astronomy, Georgia State University, Atlanta, GA, United States, ⁵Department of Radiology, University of Alabama Birmingham, Birmingham, AL, United States

Single-shot EPI imaging has become a widely used acquisition method in amide proton transfer (APT) imaging due to its fast imaging readout. Because APT effects are relatively small, fat artifacts may significantly impact the asymmetry analysis. Recently, a slice selective gradient reversal (SSGR) method has been reported to effectively suppress fat signals. This work aims to introduce the SSGR method into spin-echo EPI APT imaging, which can be used for fat suppression in the non-human primate brain with substantial pericranial fat.

David M Morris¹, Chengjia Wang¹, Giorgos Papanastasiou², Calum Gray³, Scott IK Semple¹, Tom MacGillivray¹, and William Cawthorn^1
1Centre for Cardiovascular Science, University of Edinburgh, Edinburgh, United Kingdom, ²University of Essex, Colchester, United Kingdom, ³Edinburgh Imaging, University of Edinburgh, Edinburgh, United Kingdom

The UK Biobank is a  research tissue data bank that contains genetic, health and MRI data. This includes 2-point Dixon sequences allowing for the assessment of fat fraction to evaluate bone marrow adiposity. Bone marrow adipose tissue is a potential biomarker of diseases including osteoporosis and endocrine disorders. Manual processing of this big data is not feasible and automated processes must be developed and validated. Machine Learning has been developed for the automated segmentation of anatomically appropriate bone marrow regions of interest and the extracted Fat Fraction values validated against expected values across different subject groups.         

Nikolaos Panagiotopoulos¹, Thekla Helene Oechtering¹, David Rutkowski², Jean Brittain², Diego Hernando^1,3, and Scott B Reeder^1,3,4
1Department of Radiology, University of Wisconsin-Madison, Madison, WI, United States, ²Calimetrix LLC, Madison, WI, United States, ³Department of Medical Physics, University of Wisconsin-Madison, Madison, WI, United States, ⁴Departments of Biomedical Engineering, Medicine, and Emergency Medicine, University of Wisconsin-Madison, Madison, WI, United States

A phantom placed alongside a patient, i.e., “pocket phantom,” can enable quality assurance of proton density fat fraction (PDFF) measurements. To determine the optimal positioning, the spatial dependence of PDFF measurements must be understood. To address this unmet need, we filled the bore of a 3T MRI with pocket phantoms. Though PDFF bias was low, it increased with distance from isocenter. Absolute bias was <2.5% within 18cm (in-plane) and within 10cm (z-direction). We conclude that the optimal position for a PDFF pocket phantom is on the table, centered below the patient at the level of the examined organ.

Yanjun Chen¹, Xintao Zhang¹, Mianwen Li¹, Lijie Zhong¹, Jian Wang¹, Yingjie Mei², and Xiaodong Zhang^1
1The Third Affiliated Hospital, Southern Medical University, Guangzhou, China, ²China International Center, Philips Healthcare, Guangzhou, China

Knee osteoarthritis (KOA) is characterized by pain and limitation in the range of knee joint motion with resultant disability in a substantial proportion of elderly subjects, and infrapatellar fat pad (IPFP) alterations play an important role in the progress of KOA(1,2). IPFP damage causes a direct change in the proportion of fat in the tissue to water and could be evaluated by magnetic resonance imaging (MRI)(3,4).  Fat fraction (FF) and the T2* relaxation time can detect the water and fat changes in IPFP. In this study, IPFP was quantitatively estimated by FF and T2* using modified DIXON (mDIXON) technique.

Alessandro Scotti^1,2, Jin Gao³, Weiguo Li³, Mehran Shaghaghi¹, Fred Damen¹, and Kejia Cai^1,2
1Radiology, University of Illinois at Chicago, Chicago, IL, United States, ²Bioengineering, University of Illinois at Chicago, Chicago, IL, United States, ³Research Resource Center, University of Illinois at Chicago, Chicago, IL, United States

The measurement of temperature in fatty tissues remains a challenge, since the presence of fat protons modifies the phase difference and therefore introduces errors in the quantification from proton resonance frequency shift. By using Z-spectrum MRI, combined with a Jump-Return binomial pulse preparation, we create a sinusoidal variation of the signal as a function of the resonance frequency. The magnitude-based readout eliminates the risk of phase-related artifacts and allows to measure the subtle shifts of water resonance and therefore the temperature, even in fatty tissues. We tested the sequence in mice and cream phantom.