Caroline Hoad¹, Carolyn Costigan¹, Luca Marciani², Philip Kaye³, Robin Spiller², Penny Gowland¹, Guru Aithal², and Susan Francis^1
1School of Physics and Astronomy, University of Nottingham, Nottingham, Nottinghamshire, United Kingdom, ²Nottingham Digestive Diseases Centre, NIHR Biomedical Research Unit, University Hospitals NHS Trust, Nottingham, Nottinghamshire, United Kingdom, ³Department of Cellular Pathology, University Hospitals NHS Trust, Nottingham, Nottinghamshire, United Kingdom

Phase contrast angiography of the portal vein and multi-slice True-FISP FAIR arterial spin labelling were used to measure hemodynamic parameters in 36 chronic liver disease patients. Reproducibility of these parameters was assessed in 5 healthy volunteers. Perfusion values were determined on a voxel-by-voxel basis. Portal vein mean flux and velocity across the cardiac cycle were also determined. A large inter-subject variability of these hemodynamic parameters was found across the liver patients, whereas the intra-subject variability was found to be much smaller. These techniques have the potential to monitor non-invasively the hemodynamic changes in the liver associated with liver fibrosis and cirrhosis.

Sara Aristizabal¹, Meng Yin¹, Kevin J. Glaser¹, Arunark Kolipaka¹, Armando Manduca¹, and Richard L Ehman^1
1Mayo Clinic, Rochester, Minnesota, United States

Portal hypertension is an important complication of chronic liver disease, and estimates of status of portal blood pressure elevation are regarded as important for management. This study evaluated the hypothesis that splenic stiffness assessed with MR Elastography is systematically related to splenic pulp pressure. The mechanical properties of seven resected ex-vivo pig spleens under varying splenic vein pressures were evaluated using MRE. A systematic increase in splenic stiffness with increase in pulp pressure was observed, providing preliminary evidence that MRE may provide a unique non-invasive method for estimating portal venous pressure.

Jonathan Scalera¹, Hernan Jara¹, Jorge A Soto¹, James A Hamilton², Michael O'Brien³, and Stephan William Anderson^1
1Radiology, Boston University Medical Center, Boston, MA, United States, ²Physiology and Biophysics, Boston University Medical Center, ³Pathology and Laboratory Medicine, Boston University Medical Center

Purpose: To characterize the multiexponential T2 (MET2) relaxation of liver using a murine model of hepatic fibrosis. Methods: 3,5-dicarbethoxy-1,4-dihydrocollidine(DDC) was utilized to induce hepatic fibrosis. Ex vivo liver specimens were imaged using 11.7T MRI and MET2 analyses were carried out using two algorithms. Findings were compared to degrees of liver fibrosis. Results: Two distinct peaks were seen, a dominant, short T2 and a minor, long T2 component. Moderate correlation was seen between the long T2 peak and degrees of fibrosis. Conclusion: MET2 relaxation offers potential for evaluating the microenvironment in liver fibrosis with implications for noninvasive characterization of liver disease.

Heiko G. Niessen¹, Michael Neumaier¹, Thomas Kaulisch¹, Ingolf Sack², Dieter Klatt³, Thomas Klein⁴, Juergen Braun³, and Detlef Stiller^1
1In-Vivo Imaging, Target Discovery Research, Boehringer Ingelheim Pharma GmbH & Co. KG, Biberach, Germany, ²Dept. of Radiology, Charite-University Medicine Berlin, ³Dept. of Medical Informatics, Charite-University Medicine Berlin, ⁴CardioMetabolic Diseases Research, Boehringer Ingelheim Pharma GmbH & Co. KG, Biberach, Germany

Magnetic Resonance Elastography (MRE) is a promising tool to distinguish steatohepatosis from fibrosis and to differentiate various stages of fibrosis. Here, rats were fed with a combined methionine-choline-deficient (MCD) and high-fat diet (HFD) to induce liver fibrosis. Age-dependent alterations were investigated with MRE, imaging (T2, ADC), and spectroscopy. For MRE a newly developed hydraulic-based actuator was used. MRE yields a significantly different storage and loss modulus in the rat liver after MCD/HF-diet. In addition, T2 is significantly increased, the ADC reduced compared to control animals. However, the magnitude of the latter effects is smaller than those obtained with MRE.

Zoran Stankovic¹, Zoltan Csatari¹, Peter Deibert², Wulf Euringer¹, Susanne Eggerking², Philipp Blanke¹, Zahra Abdullah Zadeh¹, Mathias Langer¹, and Michael Markl^1
1Radiology, University Hospital Freiburg, Freiburg, Ba.-Wü., Germany, ²Gastroenterology, University Hospital Freiburg, Freiburg, Ba.-Wü., Germany

Flow-sensitive 4D MRI was applied for the qualitative and quantitative analysis of portal venous hemodynamics of liver cirrhosis patients (n=20) and volunteers (n=41). Our results revealed a significant correlation of 4D MRI with the reference standard Doppler Ultrasound for maximum and mean velocities and flow for the intrahepatic vessels. Quantitative flow analysis could be performed retrospectively at any location of interest and revealed significant alterations in velocities and flow in patients compare to age matched and younger control groups.

Zoltan Csatari¹, Zoran Stankovic², Peter Deibert², Wulf Euringer², Julia Geiger², Wolfgang Kreisel², Mathias Langer², and Michael Markl^2
1University Hospital Freiburg, Freiburg, Baden Württemberg, Germany, ²University Hospital Freiburg

Flow-sensitive 3-dimensional magnetic resonance imaging was successfully used for simultaneous visualization and quantification of flow characteristics in the complete hepatic arterial und portal venous system in healthy volunteers and in patients with liver cirrhosis before and after treatment with transjugular intrahepatic portosystemic shunt (TIPS). As expected, considerable changes in blood flow characteristics after treatment with TIPS were evident. Although agreement to the reference standard Doppler ultrasound was limited, our results demonstrate the feasibility of time-resolved 3D phase contrast MRI for evaluation of normal and pathological vascular hemodynamics of the great hepatic vessels.

Greg Colin Brown¹, David James Taylor¹, and Donald McRobbie^2
1Radiology, Royal Adelaide Hospital, Adelaide, South Australia, Australia, ²Radiological Sciences Unit, Imperial College Healthcare NHS Trust, London, United Kingdom

Accurate measurement of liver iron concentration (LIC) is critical in the management of iron overload states, drug development, and understanding iron loading physiology. R2 and R2* relaxometry are the predominant non invasive methods. We investigate the performance of R2* relaxometry using FerriScan R2 relaxometry as the reference standard. Results indicate that R2* relaxometry, in its current form, has a problematic “saturation threshold” that has not been described previously in liver R2* relaxometry. When used with published calibration equations the R2* method substantially underestimates LIC above moderate levels, thus restricting its clinical utility.

Anou Sewonu^1,2, Marine Beaumont^3,4, Fanny Carbillet¹, Maélène Lohezic², René Anxionnat⁴, Jacques Felblinger^2,4, and Gabriel Hossu^3,4
1Alara-Solutions, Strasbourg, France, ²IADI Lab., Nancy-Université, Nancy, France, ³CIT801, INSERM, Nancy, France, ⁴IADI Lab., CHU Nancy, Nancy, France

Quality assurance of liver iron assessment requires knowledge of a connected but inactive coil effects on the exam. Four volunteers were investigated on a 1.5T MR unit in body coil with an inactive torso coil plugged and then in body coil only. A signal-based method and a R2* method for liver iron measurement were used. Image quality was also assessed. Liver to muscle signal ratio and contrast were significantly decreased for body coil only measurement. R2* did not show any significant difference. Further investigation involving more subjects is required for confirming these promising outcomes.

Rexford D Newbould¹, and Giulio Gambarota^1
1GSK Clinical Imaging Centre, Hammersmith Hospital, London, United Kingdom

T2* mapping of the liver can be used to assess iron loading, which is relevant in a number of diseases including thalassemia, hereditary hemochromatosis, and sickle cell disease. However, with no innate flow suppression, multi-echo spoiled gradient echo (SPGR) sequences suffer from artifacts originating from flowing blood spins. This work combines a motion-sensitized driven equilibrium (MSDE) preparation with multi-echo SPGR to suppress the blood signal intensity in breath-hold T2* mapping of the liver. Flow suppression is tested in a healthy volunteer, and T2* maps are acquired in five volunteers.

Cemil Kirbas¹, Eric Zalusky¹, Stefan Czerwinski², Miryoung Lee², Ke Cheng Liu³, and Jason G Parker^1
1Innovation Center, Kettering Health Network, Kettering, OH, United States, ²Department of Community Health, Wright State University, Kettering, OH, United States, ³Siemens Medical Solutions, United States

Assessment of fat content in the liver is of interest in the evaluation of a spectrum of diseases. Quantification of liver fat fraction using magnetic resonance imaging (MRI) has become popular recently. One of the popular MRI methods to quantify liver fat fraction is the two-point Dixon method using in-phase (IP) and out-phase (OP) images. In this study we used a Dixon-based method and evaluated the difference between using individual correction factor calculated for each subject and an average correction factor. We also evaluated the accuracy of fat fraction calculation using a ROI’s from a single slice versus multiple slices.

Thomas David Reed¹, Rashmi Agni², Catherine Hines¹, Richard Bruce¹, Mona Ranade¹, Benjamin Soriano², Kiyarash Mohajer¹, and Scott B Reeder^1
1Radiology, University of Wisconsin, Madison, WI, United States, ²Pathology, University of Wisconsin, Madison, WI, United States

Over the last several years, quantitative MR methods for accurate measurement of hepatic steatosis have been developed. The majority of studies validated quantitative MR methods using single voxel MRS as the reference standard. These methods have demonstrated excellent correlation and agreement between proton density fat-fraction and MRS so long as all confounding factors (eg. T1, T2*, spectral complexity of fat, noise bias and eddy current correction) are addressed. However, there is a paucity of tissue validation in these studies. In this retrospective study involving 26 patients, we demonstrate good correlation (r2=0.77) between MRI fat-fraction and histological grading of steatosis.

Scott Brian Reeder¹, Catherine D Hines¹, Huanzhou Yu², Charles A McKenzie³, and Jean H Brittain^4
1Radiology, University of Wisconsin, Madison, WI, United States, ²Global Applied Science Laboratory, GE Healthcare, Menlo Park, CA, United States, ³Medical Biophysics, University of Western Ontario, London, Ontario, Canada, ⁴Global Applied Science Laboratory, GE Healthcare, Madison, WI, United States

Proton density fat-fraction (PDFF) is a useful metric for fat quantification, providing a platform- and protocol-independent metric of tissue fat concentration. PDFF is the ratio of unconfounded signal from mobile fat protons, normalized by the total unconfounded signal from mobile fat protons and mobile water protons. Unfortunately, reference assays that measure concentrations of triglycerides do not account for NMR invisible species, and therefore will correlate, but may not agree directly with, PDFF measured with MRI. In this work, the relationship between true tissue fat concentration and PDFF is described and validated using a fat-water-deuterium oxide phantom.

Puneet Sharma¹, Hiroumi D Kitajima¹, Xiaodong Zhong², Bobby Kalb³, Alton B Farris⁴, Miriam B Vos⁵, and Diego R Martin^3
1Radiology, Emory Healthcare, Atlanta, GA, United States, ²MR R&D Collaborations, Siemens Healthcare, Atlanta, GA, United States, ³Radiology, Emory University, Atlanta, GA, United States,⁴Pathology, Emory University, Atlanta, GA, United States, ⁵Hepatology, Children's Hospital of Atlanta, Atlanta, GA, United States

Therapy development for early disease has required accurate and reproducible measurement for the longitudinal monitoring of hepatic lipid. Recently, a multi-echo breath hold T2-corrected MR spectroscopy technique (HISTO-MRS) has demonstrated clinical feasibility, while proving robust in phantom validation of lipid content. This investigation extends these initial observations for a more broad evaluation of accuracy and reproducibility in a clinical setting. The results showed that the HISTO technique was highly reproducible, while significantly correlating with biopsy measures, allowing effective longitudinal assessment of hepatic lipid. Combined with efficient and automated post-processing steps, HISTO-MRS demonstrates clinical applicability as a biopsy surrogate.

Pascal Spincemaille¹, Doug Brylka¹, Martin R Prince¹, and Yi Wang^1,2
1Radiology, Weill Cornell Medical College, New York, NY, United States, ²Biomedical Engineering, Cornell University, Ithaca, NY, United States

Dynamic contrast enhanced liver imaging is used for the detection and characterization of liver lesions. Current clinical protocols rely on a multiple phase approach acquired over multiple breath-hold, limiting spatial and temporal resolution. In a considerable subset of patients, decreased breath-holding ability leads to a dramatic decrease in image quality. In this work, a 3D hybrid Cartesian-radial acquisition is combined with respiratory gating to suppressed motion related artifacts

Feng Zhang¹, Jiakai Li¹, Yanfeng Meng¹, Jihong Sun¹, Stephanie San Juan Soriano¹, Huidong Gu¹, Patrick Willis¹, and Xiaoming Yang^1
1Image-Guided Bio-Molecular Intervention Section, Department of Radiology, University of Washington School of Medicine, Seattle, WA, United States

This study was to develop a new technique of intrabiliary MRI-guided local agent delivery. We first confirmed the capability of motexafin gadolinium(MGd) entering cholangiocarcinoma cells. Then, we established surgical protocol for intrabiliary balloon-mediated MGd delivery into pig common bile duct (CBD) walls ex vivo. We finally validated the feasibility of using MR to monitor, in vivo, intrabiliary delivery of MGd into the pig CBD wall of a pig, which was confirmed by subsequent MRI-histology correlation. This new technique may open new revenues for MR-guided intrabiliary local delivery of therapeutics, such as genes and drugs, to treat malignant pancreatobiliary diseases.

Eleanor F Cox¹, Caroline L Hoad¹, John J Totman¹, Carolyn Costigan¹, Luca Marciani², Robin C Spiller², and Penny A Gowland^1
1SPMMRC, School of Physics & Astronomy, University of Nottingham, Nottingham, Nottinghamshire, United Kingdom, ²NDDC NIHR BRU, Nottingham University Hospitals, Nottingham, United Kingdom

Different food ingredients can trigger gallbladder contraction but which food ingredients are most effective in contracting the gallbladder and at which dose level is not well understood. Here we used serial MRI to test the ability of several food ingredients to stimulate gallbladder contraction. High fat emulsion and semi-skimmed milk stimulated the highest gallbladder contraction (42% and 41% contraction respectively) of 10 test drinks, which also had the highest fat content. A fat-dose response of the gallbladder was subsequently shown using a milk-based drink, varying only the fat content, and correlated with changes in plasma cholecystokinin.

Ekkehard Küstermann¹, Anke Meyer², Amod Godbole², Wolfgang Dreher³, and Kathrin Maedler^2
1ZKW, University of Bremen, Bremen, Germany, ²CBIB, University of Bremen, ³FB2, University of Bremen

Diabetes research strongly benefits from non-invasive methods of analysing pancreatic beta-cell mass. A promising strategy is to measure Manganese uptake in pancreatic tissue by T1-weighted in-vivo MR imaging. Here, a “3D ir-Snapshot-FLASH” protocol for robust, least-invasive T1-weighted MR imaging of mouse abdomen in spontaneously braething mice is presented.

Bobby Kalb¹, Juan M Sarmiento², N Volkan Adsay³, James Costello¹, Hiroumi Kitajima¹, Puneet Sharma¹, Christina Lurie¹, and Diego R Martin^1
1Radiology, Emory University School of Medicine, Atlanta, GA, United States, ²Surgery, Emory University School of Medicine, Atlanta, GA, United States, ³Pathology, Emory University School of Medicine, Atlanta, GA, United States

Paraduodenal pancreatitis (PDP) is a clinicopathologically distinct form of focal chronic pancreatitis (CP) thought to be related to obstruction of the pancreatic accessory duct. Differentiation between PDP and pancreatic duct adenocarcinoma (CA) is challenging and represents an important unmet clinical need. Our results demonstrate PDP may be distinguished from CA with contrast-enhanced MRI even with non-experienced readers, with a sensitivity of 88.2% and specificity of 86.7%, when strict diagnostic criteria are followed. Our study contributes to optimized therapeutic management of patients with a pancreatic head mass, supporting a primary diagnostic role for MRI.

Yogesh kannan Mariappan¹, Kevin Glaser¹, Naoki Takahashi¹, Phillip Young¹, and Richard L Ehman^1
1Department of Radiology, Mayo Clinic, Rochester, MN, United States

Magnetic Resonance Elastography (MRE) is a phase-contrast MR elasticity imaging technique that can noninvasively measure the shear stiffness of soft tissues. The hypothesis of this work was that the stiffness of the pancreas in patients with pancreatitis is significantly different from the stiffness of healthy pancreas and that MRE can used to detect this difference. To test this hypothesis, an MRE acquisition was implemented to acquire multislice vector shear wave displacement data to measure the shear stiffness of the pancreas. This approach was applied to healthy subjects and to patients suspected of having pancreatic disease. Preliminary data obtained from this study indicates that the stiffness of the pancreas in patients with chronic pancreatitis is higher than in healthy volunteers.

Lidia S Szczepaniak¹, Edward W Szczepaniak¹, Qi Peng², and Ildiko Lingvay^3
1The Heart Institute, Cedars-Sinai Medical Center, Los Angeles, California, United States, ²Radiology, University of Texas, Health Science Center, San Antonio, Texas, United States,³Endocrinology, University of Texas, Southwestern Medical Center, Dallas, Texas, United States

Ethnic minorities – African American and Latinos – are disproportionally affected by obesity and type 2 diabetes for reasons not completely understood. We present here results of our study designed to provide new insights into understanding the mechanism of beta cell dysfunction. We focused on the role of fat deposition in the pancreas (pancreatic steatosis). We used localized proton MRS to measure pancreatic steatosis in vivo and frequently sampled IV glucose tolerance test to measure insulin secretion. We detected diverse patterns of pancreatic fat storage that may contribute to ethnic variations in pathogenesis of insulin secretion.

Saeed Kiani¹, Isky Gordon², Iosif Mendichovszky³, Marica Cutajar², and Kevin Wells^1
1CVSSP, Faculty of Engineering & Physical Sciences, University of Surrey, Guildford, Surrey, United Kingdom, ²UCL Institute of Child Health, London, United Kingdom, ³Wolfson Molecular Imaging Centre, University of Manchester, Manchester, United Kingdom

Robustness of DCE-MRI renography in the presence of diminished renal function needs to be addressed. Renal impairment may result in uptake reduction of the contrast agent (Gd-DTPA) and subsequently signal-to-noise ratio (SNR) reduction in the DCE-MR image data. This work aims to evaluate the robustness of DCE-MRI renography in the presence of synthetic SNR variation on the renal data. The synthetic SNR variation is achieved by degrading the data in both spatial and temporal domains. The results demonstrated substantial variations (i.e. >10%) in estimated kidney filtration rates (GFR) for a slight variation of SNR, about 0.05-0.1, from the original data.

Andreas D Merrem¹, Frank G Zoellner¹, and Lothar R Schad^1
1Computer Assisted Clinical Medicine, Medical Faculty Mannheim, Heidelberg University, Mannheim, Germany

DCE-MRI is an emerging technique for a more accurate assessment of local renal function. Measured time-intensity-curves used for quantification of kidney function are hampered by respiratory motion. We analysed the feasibility of a variational approach to image registration in DCE-MRI. Evaluation was performed by checkerboards, measuring the vertical displacement of an anatomical reference point, and fitting a 2-compartment model to a ROI in the renal cortex. Registration was successful for 4 out of the 5 data sets. The motion could be reduced to below the in-plane resolution of 1.5mm. The Akaike error of fitting the 2-compartment model is reduced up to 24%.

Yutong Duan^1,2, Ralf Berthold Loeffler¹, Ruitian Song¹, Aaryani Tipirneni¹, Sheri Spunt³, Niels Oesingmann⁴, Anne Viano², and Claudia Maria Hillenbrand^1
1Radiological Sciences, St. Jude Children's Research Hospital, Memphis, TN, United States, ²Physics, Rhodes College, Memphis, TN, United States, ³Oncology, St. Jude Children's Research Hospital, Memphis, TN, United States, ⁴Siemens Medical Solutions USA, Inc., New York, NY, United States

This study includes techniques that smooth the curve by removing the spikes from the curve, and a simulation test that validates this optimization with multiple sets of noise reproducing the inflow effect.Four sets of artificial noise were extracted from real patient data and used to simulate the abrupt spikes due to the inflow effect. After noise was applied the GFR values were calculated and compared with the reference GFR. The correlation between the original AIF and the two gamma variate function varied from 0.95 to 0.99. The smoothing method was able to reduce the inflow artifact on the AIF signal.

Florian Lietzmann¹, Frank G. Zoellner¹, Ulrike Attenberger², Henrik J. Michaely², Stefan Haneder², Stefan O. Schoenberg², and Lothar R. Schad^1
1Computer Assisted Clinical Medicine, Heidelberg University, Mannheim, Germany, ²Institute of Clinical Radiology and Nuclear Medicine, University Medical Centre Mannheim, Mannheim, Germany

Dynamic contrast enhanced MRI (DCE-MRI) provides an alternative technique to evaluate kidney function such as perfusion or glomerular filtration parameters. The self-navigating BLADE-sequence offers an alternative approach for a motion corrected DCE-examination without the need of respiratory triggering. The aim of this study was to evaluate whether kidney functional parameters can be reliably derived from a pharmacokinetic model while providing improved motion correction based on the BLADE algorithm.

Jeff D Winter^1,2, Keith S St. Lawrence^3,4, and Hai-Ling Margaret Margaret Cheng^1,5
1Physiology and Experimental Medicine, The Hospital for Sick Children, Toronto, Ontario, Canada, ²Research and Development, IMRIS, Winnipeg, Manitoba, Canada, ³Imaging Division, Lawson Research Institute, London, Ontario, Canada, ⁴Medical Biophysics, University of Western Ontario, London, Ontario, Canada, ⁵Medical Biophysics, University of Toronto, Toronto, Ontario, Canada

The injury or loss of renal microvessels is a determinant of renovascular disease severity. Assessment of renal microvasculature may be achieved using MR perfusion techniques including quantitative DCE and ASL. This study compared renal perfusion estimates for ASL and DCE MRI. DCE MRI was performed with a dual-bolus contrast injection for improved arterial input sampling. We found DCE estimates of renal cortex perfusion (3.57 ± 0.96 ml/g/min) were in agreement with ASL (3.28 ± 0.59 ml/g/min). Moreover, this study showed feasibility of the dual bolus approach for quantitative DCE for the first time outside of cardiac imaging.

Mao-Yuan Su¹, Chin-Chen Chang¹, Kao-Lang Liu¹, Ting-Fang Tiffany Shih¹, and Wen-Chau Wu^1,2
1Department of Medical Imaging, National Taiwan University Hospital, Taipei, Taiwan, ²Graduate Institute of Oncology, National Taiwan University, Taipei, Taiwan

To the best of our knowledge, cross-validation between ASL and DCE remains absent for the kidneys. In the present study, we examined the compatibility of ASL and DCE at the field strength of 3T by measuring renal blood flow with both techniques on a sample of healthy volunteers and one patient with renal artery stenosis. Correlation was found between ASL and DCE measurements (r = 0.9, p < 0.005). Both ASL and DCE were able to distinguish the reduced RBF in the patient with renal artery stenosis. We conclude that ASL and DCE provide compatible measurement of renal perfusion at 3T.

Eleanor F Cox¹, Caroline L Hoad¹, and Susan T Francis^1
1SPMMRC, School of Physics & Astronomy, University of Nottingham, Nottingham, Nottinghamshire, United Kingdom

A respiratory triggered variant of multiphase TrueFISP arterial spin labelling (ASL) is introduced to measure renal perfusion and transit time across the kidney at high spatial resolution. The increase in spatial resolution reduces partial volume effects allowing the visualisation of progressive inflow of labelled blood to the kidney and assessment of heterogeneity across the kidney, important in the study of renal physiology and disease. From the high resolution data the mean perfusion across the cortex was 329 ± 85 ml/100g/min and the mean transit time was 472 ± 178 ms.

Mark Stephen Dobbs^1,2, Neil Woodhouse³, Geoff J.M Parker^1,2, and Josephine H Naish^1,2
1Imaging Science and Biomedical Engineering, The University of Manchester, Manchester, Greater Manchester, United Kingdom, ²The Biomedical Imaging Institute, The University of Manchester, Manchester, Greater Manchester, United Kingdom, ³AstraZeneca, Macclesfield, Cheshire, United Kingdom

There has been increasing interest in applying perfusion measurement techniques, such as Arterial Spin Labelling (ASL), to the kidneys in light of concerns over the link between gadolinium-bearing contrast agents and nephrogenic systemic fibrosis (NSF). Correction or avoidance of breathing-induced motion artefacts in the kidneys is beneficial. Recent work has shown that multiple-time point and multiple phase approaches for ASL can aid renal perfusion quantification by extracting the bolus arrival time (the time for labelled blood to reach the renal tissue), which itself may have clinical relevance. When combining a multiple-time point approach with presaturation, there are large contrast differences in the raw ASL images at different inversion times (TIs), relative to the ASL signal, which makes registration problematic. This is also a problem for inversion recovery (IR) sequences applied to the kidneys. In this abstract we demonstrate the feasibility of a model-based registration approach, based on the T1-recovery of signal intensities, for registration of multiple-time point ASL and IR sequences applied to the kidneys.

Yin Huang¹, Nathan Hanson¹, Elizabeth Sadowski², David Niles¹, Nathan Artz¹, Arjang Djamali³, Thomas Grist^1,2, and Sean Fain^1,2
1Medical Physics, University of Wisconsin Madison, Madison, WI, United States, ²Radiology, University of Wisconsin Madison, Madison, WI, United States, ³Nephrology, University of Wisconsin Madison, Madison, WI, United States

The k-means segmentation method was implemented to semi-automatically segment kidney cortex and medulla for MR BOLD images of 6 subjects. By acquiring an extra T1 weighted image, k-means segmentation was performed based on two kidney feature values -- T1 and T2* weighted signal intensities. Manual segmentation results on the same subjects were used as reference and sensitivity and specificity measures were calculated to evaluate the quality of the k-means segmentation.

David Joseph Niles¹, Sean B Fain^1,2, Nathan S Artz¹, Yin Huang¹, Karl K Vigen², Arjang Djamali³, Thomas M Grist^1,2, and Elizabeth A Sadowski^2
1Medical Physics, University of Wisconsin, Madison, WI, United States, ²Radiology, University of Wisconsin, Madison, WI, United States, ³Nephrology, University of Wisconsin, Madison, WI, United States

Routine monitoring of renal function following transplantation is important for prolonging allograft viability. In this study, BOLD MRI was used to measure renal oxygenation in matched donor-recipient pairs up to one year post-transplantation. An 11 % reduction of R2* was observed in the medulla of transplanted kidneys 3 months post-transplantation (p < 0.01), indicating increased oxygen availability. The cortex of the transplanted kidney and the donors’ remaining kidneys showed no change of R2* after 3 months. Additional measurements in this ongoing study will be used to assess oxygenation one and two years post-transplantation.

Xiang He¹, Chan-Hong Moon¹, Jung-Hwan Kim¹, and Kyongtae Ty Bae^1
1Department of Radiology, University of Pittsburgh, Pittsburgh, Pennsylvania, United States

In this study, we improved the T1ρ sequence to compensate for both B0 and B1 field imperfections and applied this technique in the first time for T1ρ-weighted imaging and T1ρ mapping of in vivo human kidneys. Our result demonstrated that T1ρ signal of the kidney may provide additional information regarding the underlying structure of the kidney beyond the traditional T1, T2 and T2* measurement, and may serve as an important imaging biomarker for quantitative functional MRI of kidney. Additionally, the sensitivity of T1ρ to tissue compositions and interactions may broaden our understanding and detection of renal pathologies.

Eleanor F Cox¹, Caroline L Hoad¹, and Susan T Francis^1
1SPMMRC, School of Physics & Astronomy, University of Nottingham, Nottingham, Nottinghamshire, United Kingdom

Quantification of T₁ in the kidneys is subject to breathing induced movement between acquisitions of differing contrast. Here we use a modified respiratory triggered inversion recovery TrueFISP acquisition to measure T₁ in the kidneys, requiring no breath hold or registration techniques. The mean T₁ was 950 ± 40 ms in the renal cortex and 1210 ± 60 ms in the medulla. High resolution T₁ maps show differentiation in T₁ between the renal cortex and outer and inner medulla that could not be resolved at lower resolution.

Moritz Bernhard Mie¹, Frank Gerrit Zoellner¹, and Lothar Rudi Schad^1
1Computer Assisted Clinical Medicine, Heidelberg University, Mannheim, Germany

Susceptibility weighted imaging (SWI) is not only a valuable tool for brain imaging, but also for renal imaging. In this study, the sensitivity of SWI to kidney-function induced changes of blood flow is investigated. Because the blood oxygenation level dependent (BOLD) contrast of blood depends on the local blood flow, the SWI contrast has been investigated while the kidney function has been increased. SWI has been applied before and after water uptake. In healthy volunteers, the SWI contrast decreases after water uptake and then increases again to the equilibrium value of the resting state.

Frank G Zoellner¹, Holger Best¹, Simon Konstandin¹, Stefan Haneder², Stefan O Schoenberg², Henrik J Michaely², and Lothar R Schad^1
1Computer Assisted Clinical Medicine, Heidelberg University, Mannheim, Germany, ²Institute of Clinical Radiology and Nuclear Medicine, University Medical Center Mannheim, Heidelberg University, Mannheim, Germany

One main purpose of the human kidneys is the maintenance and the regulation of the fluid and electrolyte homeostasis. About 80% of the filtrated substances are sodium. The combination of relative low in-vivo concentration and MR sensitivity of 23Na compared to 1H results in relative low MR signal. A logical step forward is to enhance the images by applying post-processing filters. We compared the performance of different filters in sodium imaging of the human kidney at 3T. SNR before and after filtering sodium images obtain in 4 healthy volunteers is improved by a factor of 3.6 for the Fermi, 3.2 for the Hamming and 3.2 for the Gaussian filter.

Jochen Keupp¹, Ivan Dimitrov^2,3, Sander Langereis⁴, Osamu Togao³, Masaya Takahashi³, and A Dean Sherry^3
1Philips Research Europe, Hamburg, Germany, ²Philips Healthcare, United States, ³Advanced Imaging Research Center, UT Southwestern Medical Center, Dallas, Texas, United States, ⁴Philips Research Europe, Eindhoven, Netherlands

Clinical MR techniques to report on local pH would be of high interest, because several pathologies are associated with pH changes like in renal diseases. Fortuitously, a clinically approved CT agent, Iopamidol, has been shown to have exchangeable protons that generate pH-dependent CEST (chemical exchange saturation transfer)-MRI contrast. It provides two proton pools, which allow a ratiometric pH measure independent from the local agent concentration, as recently demonstrated pre-clinically. In the present study it is shown, that Iopamidol can be used for in vivo human pH measurement in the renal pelvis using breathing triggered CEST-MRI on a clinical 3T scanner.

Cory R Wyatt¹, Brian Dale², Elmar Merkle¹, James MacFall¹, and Brian Soher^1
1Radiology, Duke University, Durham, NC, United States, ²Siemens Medical Solutions, Inc., Morrisville, NC, United States

In-phase and-opposed phase chemical shift magnetic resonance imaging has been extensively used for differentiating between an adenoma and metastatic growth in the adrenal glands. However, due to the range of scan parameters used in these studies, the thresholds reported differ significantly. In this work, we present a method for normalizing signal intensity indices used to differentiate adrenal adenomas from metastases regardless of the clinical parameters used to acquire the IP/OP images. It was found that the method was able to correlate signal intensity index values to the true fat/water signal ratio for phantoms containing small amounts of fat.

Tomohiro Namimoto¹, Kosuke Morita², Toshinori Hirai², Shinichi Nakamura², Seitaro Oda², Daisuke Utsunomiya², Yasuyuki Yamashita², and Makoto Obara^3
1Radiology, Kumamoto University, Kumamoto, Kumamoto, Japan, ²Kumamoto University, ³Philips Medical Systems

3D high resolution 3-point Dixon techniques offer an alternative to 2D gradient echo chemical shift imaging for fat quantification of adrenal masses with higher accuracy.

Geoff Charles-Edwards^1,2, Robert Johnstone¹, Sarah Natas¹, and Audrey Jacques^1
1Guy's & St Thomas' NHS Foundation Trust, London, United Kingdom, ²King's College London, London, United Kingdom

Endometrial carcinoma is the most common gynaecological malignancy in Western countries. Conventional grading and staging by biopsy and imaging does not always correspond with the final histological grade determined post-hysterectomy. This results in some patients having less aggressive initial surgery, while others will have unnecessary extensive lymph node dissection. Consequently there is a need for additional information to improve preoperative assessment. This work describes some initial results from MR spectroscopy at 3T providing metabolic information about endometrial cancer that may aid in preoperative assessment of disease.

Lucy Elizabeth Kershaw¹, Yuexi Huang¹, Hallie Taylor^2,3, Elizabeth David², Kullervo Hynynen^1,4, Greg Stanisz^1,4, and Laurent Milot^2,3
1Imaging Research, Sunnybrook Research Institute, Toronto, Ontario, Canada, ²Radiology, Sunnybrook Research Institute, Toronto, Ontario, Canada, ³Medical Imaging, University of Toronto, Toronto, Ontario, Canada, ⁴Medical Biophysics, University of Toronto, Toronto, Ontario, Canada

Uterine fibroids are benign masses affecting ~20% of women of reproductive age, resulting in poor quality of life due to pelvic pain and dysmenorrhea. The success of treatments such as high intensity focused ultrasound depends on the perfusion of the fibroid, therefore a baseline assessment of the fibroid properties is important in the selection of the most appropriate treatment on a case-by case basis. Although the appearance on T₂w images has been documented, quantitative assessment of fibroid properties has been rare. In this study, ADC, T₁ measurements and DCE-MRI parameters derived from two different models are presented.

John P. Mugler, III¹, Talissa A. Altes¹, Wilhelm Horger², and Berthold Kiefer^2
1Radiology, University of Virginia, Charlottesville, VA, United States, ²Siemens Healthcare, Erlangen, Germany

An adiabatic T2-contrast preparation was implemented for T2-weighted (T2W) single-slab 3D-TSE (SPACE) imaging of the prostate and uterus, and the resulting contrast properties were compared to those for the standard SPACE method and conventional 2D TSE in five healthy subjects (2 males, 3 females). T2-prepared SPACE provided image contrast that was different than that for either conventional T2W 2D TSE or standard T2W SPACE, and, in particular, yielded improved contrast of structures within the prostate. This observation, coupled with the ability to reconstruct multiple image planes from high-resolution 3D acquisitions, suggests that T2-prepared SPACE warrants evaluation in subjects with disease.

Tammar Kushnir¹, Chen Hoffmann¹, Lisa Raviv-zilka¹, Yishay Salem¹, Eli Konen¹, and Orly Goitein^1
1Dept. of Diagnostic Imaging, MRI Unit, The Chaim Sheba Medical Center, Tel Hashomer, Israel

T2* multi-echo gradient echo is a robust validated non-invasive method for the accurate evaluation of tissue iron. Low T2* values serve as an early marker of iron deposition in the target organs. The purpose of this study was to use the method to evaluate the standard liver T2* value in healthy fetus. To our knowledge this is the first documentation of T2* values for fetus. In our small cohort, average liver T2* values at the third trimester were found to be 21.5 ± 6.3 ms (range 12.5 -30), shorter than those of normal adult subjects.

Yeji Han¹, Sandra Huff², HyunWook Park¹, and Ute Ludwig^2
1Department of Electrical Engineering, Korea Advanced Institute of Science and Technology, Daejeon, Daejeon, Korea, Republic of, ²Medical Physics, University Hospital Freiburg, Freiburg, Germany

To overcome the difficulties of multistation whole-body DWI (wbDWI), the authors have proposed a continuously moving table (CMT) wbDWI method in the previous work. However, more aspects of the CMT-wbDWI have to be validated to utilize this method in a clinical setting. Thus, the influence of continuous table motion on diffusion data was examined in this study by calculating apparent diffusion coefficient (ADC) values using a spherical water phantom. The experimental results show that the general tendency of ADC changes is preserved in all cases.

Alan John Stone^1,2, Jacinta E Browne³, Brian Lennon⁴, James F. Meaney¹, and Andrew J Fagan^1,5
1Centre for Advanced Medical Imaging (CAMI), St. James’s Hospital / Trinity College, University of Dublin, Ireland, ²Cardiff University Brain Research Imaging Centre (CUBRIC), School of Psychology, Cardiff, United Kingdom, ³School of Physics, Dublin Institute of Technology, Dublin, Ireland, ⁴Dept. Medical Physics and Bioengineering, St. James’s Hospital, Dublin, Ireland, ⁵School of Medicine, Trinity College, University of Dublin, Ireland

The use of DWI techniques to categorise malignancy or therapy efficacy in tumours has led in recent years to the use of the DWIBS technique for ADC quantification. A comparative study of DWIBS with an accurate ADC measurement technique found no difference in mean ADC values measured by the two techniques in a moving phantom and when static. However, a noticeable spread in ADC values was noted, particularly in small tumours, which may be of significance when monitoring therapy efficacy where slight changes in tumour heterogeneity may be inferred from subtle changes in ADC histograms

Yoshimitsu Ohgiya¹, Jumpei Suyama¹, Syouei Sai¹, Masaaki Kawahara¹, Jirou Munechika¹, Makoto Saiki¹, Noritaka Seino¹, Masanori Hirose¹, and Takehiko Gokan^1
1Showa University School of Medicine, Tokyo, Japan

The purpose of this study was to correlate urinary bladder cancer having a stalk observed on 3-Tesla MRI with histopathological T staging and cystoscopic findings and to compare T2-weighted imaging (T2WI) and diffusion-weighted imaging (DWI) in stalk detectability. All of tumors with a stalk on MRI were papillary tumors noted on cystscopy. Non-papillary tumors did not show stalks on MRI. Most of the bladder tumors with a stalk on 3-Tesla MR imaging were in T1 or lower pathologic stages. DWI had superior detectability in stalks of papillary bladder tumors to T2WI, particularly in tumors more than 10 mm.

Moti Freiman¹, Stephan Voss², and Simon K Warfield^1
1Computational Radiology Laboratory, Dept. of Radiology, Children's Hospital, Harvard Medical School, Boston, MA, United States, ²Dept. of Radiology, Children's Hospital, Harvard Medical School, Boston, MA, United States

Diffusion Weighted Imaging and the derived Apparent Diffusion Coefficient (ADC) map can be use to quantify abdominal tumor response to therapy. Current techniques use only the tumor region mean or median ADC value. However, the actual response is more complex due to inherent tumor heterogeneity. We present a new approach to quantify the tumor's response based on entire cumulative histogram analysis. Our approach provides a single number that encapsulates the overall diffusivity changes over time. Representative results on three cases show that our measure is comparable with the radiologist qualitative evaluation.

Tetsuo Ogino^1,2, Tomohiko Horie³, Hayato Takano⁴, Thomas Kwee⁵, Taro Takahara^6,7, Marc Van Cauteren⁸, and Tosiaki Miyati^9
1Healthcare division, Philips Electronics Japan, LTD, Minato-ku, Tokyo, Japan, ²Graduate School of Medical Sciences, Kanazawa University, Kanazawa-shi, Ishikawa-ken, Japan, ³Dept. of Radiology, Tokai Univiersity Hospital, Isehara-shi, Kanagawa, Japan, ⁴Dept. of Radiology, Tokai University Hospital, Isehara-shi, Kanagawa, Japan, ⁵Department of Radiology, University Medical Center Utrecht, Utrecht, Netherlands, ⁶Biomedical Engineering, Tokai University School of Engineering, Hiratsuka-shi, Kanagawa, Japan, ⁷Dept. of Radiology, UMC Utrecht, Utrecht, Netherlands,⁸Philips Healthcare, ⁹Department of Quantum Medical Technology, Kanazawa University, Kanazawa, Ishikawa, Japan

Cardiac and Respiratory motion affects image quality and ADC measurement on abdominal DWI. Combined method of “PPU” Peripheral Pulse Unit cardiac triggering and "TRON" Respiratory motion compensation on 3.0T DWI is demonstrated. Signal uniformity and ADC uniformity on whole liver is improved. Averaged among 5 healthy volunteers, ADC value ratio between left and right lobe of liver is reduced to 1.04 with PPU TRON from 1.5 with conventional Respiratory Triggering DWI. Average scan time among 5 volunteers was 3:48. PPU TRON enables more uniform image quality and reproducible ADC measurement in clinically feasible scan time.

Moritz Florian Kircher^1,2, Alan Xu¹, Anja C. Brau³, Martin Laufik¹, Yuji Iwadate⁴, Jarrett Rosenberg¹, Bruce L Daniel¹, and Robert J Herfkens^1
1Radiology, Stanford University, Stanford, CA, United States, ²Radiology, Memorial Sloan-Kettering Cancer Center, New York, NY, United States, ³Global Applied Science Laboratory, GE Healthcare, Menlo Park, CA, United States, ⁴Global Applied Science Laboratory, GE Healthcare, Hino, Japan

Diffusion-weighted imaging (DWI) has become a routine part of abdominal MRI examinations at many institutions. Several different DWI techniques are currently in use, including breath-hold DWI and non-breathhold DWI techniques such as free-breathing, respiratory-triggered DWI and Navigator-triggered DWI. In this study, we aimed to prospectively evaluate and compare the subjective image quality of these four commonly used DWI techniques in 30 patients with liver lesions. The results demonstrate superiority of navigator-triggered and respiratory-triggered DWI over breath-hold and free-breathing DWI for most image quality parameters.

Mamak Eatesam¹, Susam M Noworolski^1,2, Phyllis C Tien³, Michelle Nystrom¹, Jennifer L Dodge⁴, Raphael B Merriman⁵, and Aliya Qayyum^1
1Radiology and Biomedical Imaging, UCSF, San Francisco, CA, United States, ²Graduate Group in Bioengineering, UC San Francisco and Berkeley, San Francisco and Berkeley, CA, United States,³Department of Medicine, UCSF and San Francisco Veterans Affairs Medical Center, San Francisco, CA, United States, ⁴Department of Internal Medicine, UCSF, Fresno, CA, United States,⁵Department of Medicine, California Pacific Medical Center, San Francisco, CA, United States

Liver apparent diffusion coefficient (ADC) values were compared between breath-hold and free-breathing DWI in 12 healthy volunteers and 16 patients, (9 NAFLDs and 7 HCVs). BH ADC was lower than FB ADC, (p<0.01). BH and FB liver ADC showed a moderate correlation at best such that BH and FB techniques should not be used interchangeably. A higher correlation between FB and BH was observed in NAFLD patients which may reflect differences in disease distribution or reduced range of respiratory effort in NAFLD patients due to a higher BMI.

Andrew James Gilman¹, Susan Moyher Noworolski¹, Mamak Eatesam¹, Jennifer Lynne Dodge¹, Raphael Brendan Merriman², and Aliya Qayyum^1
1Radiology & Biomedical Imaging, University of California, San Francisco, San Francisco, CA, United States, ²Division of Gastroenterology, California Pacific Medical Center, San Francisco, CA, United States

The relationship between liver ADC and steatosis was evaluated for high (125, 500), low (0, 125), and conventional (0, 500) b-value DWI at 3T in 6 healthy volunteers and 12 patients with NAFLD or HCV. Subjects with steatosis grade 1 or 2 (n=5) had significantly lower high b-value ADC compared to those with grade 0 (n=13) for both large (0.88 versus 1.09, p<0.04) and small (0.82 versus 1.12, p<0.02) ROIs. The effect of steatosis on liver ADC should be taken into account when using DWI to discriminate between features of diffuse liver disease.

Mamak Eatesam¹, Michelle Nystrom¹, Susan M Noworolski^1,2, Jennifer L Dodge³, Raphael B Merriman⁴, and Aliya Qayyum^1
1Radiology and Biomedical Imaging, UCSF, San Francisco, CA, United States, ²Graduate Group in Bioengineering, UC San Francisco and Berkeley, San Francisco and Berkeley, CA, United States,³Department of Internal Medicine, UCSF, Fresno, CA, United States, ⁴Department of Medicine, California Pacific Medical Center, San Francisco, CA, United States

The repeatability of liver ADC measurement was compared across different image analysis techniques designed to reduce the impact of vessels on the ADC. Two successive breath-hold DWI scans were performed in 14 subjects; 6 NAFLD and 8 healthy volunteers. ADC was derived from 6 small and 3 large ROIs and with 4 different image threshold applications. Results indicated that large ROI ADC measurements were more repeatable than small ROI measurements. Vessel correction was associated with the narrowest range of ADC difference between scans, suggestive of highest repeatability.

Thomas L. Chenevert¹, Craig J. Galbán¹, Frank J. Londy¹, Charles R. Meyer¹, Timothy D. Johnson², Alnawaz Rehemtulla³, and Brian D. Ross^1
1Radiology - MRI, University of Michigan, Ann Arbor, Michigan, United States, ²Biostatistics, University of Michigan, Ann Arbor, Michigan, United States, ³Radiation Oncology, University of Michigan, Ann Arbor, Michigan, United States

Prerequisites for use of ADC, as well as any image-based biomarker in multi-center trials is standardization of data acquisition/analysis, and certification of sites/systems via quantitative measurements on known test objects. The objective of this study is to assess reproducibility of ADC measurements performed at multiple sites on various clinical platforms using icewater as a simple universal temperature-controlled diffusing medium. In this study system outlier(s) were identified, and 3-5% reproducibility in diffusion coefficient measurement exclusive of outliers was demonstrated. Systematic spatial-dependent error was also demonstrated with its source under investigation.

Hai-Ling Margaret Cheng^1,2, Yasir Loai², and Walid A Farhat^2
1Medical Biophysics, University of Toronto, Toronto, Ontario, Canada, ²Hospital for Sick Children, Toronto, Ontario, Canada

Quantitative MRI parameters such as T1 and T2 relaxation times provide a means to probe tissue microstructure and composition. However, specificity is often limited, as is the case when interrogating complex systems such as regenerating tissue where concurrent biophysical and biochemical changes exist. This study explores multiexponential diffusion and effective transverse relaxation time T2* at 1.5 Tesla for probing cell growth and tissue composition. The slow diffusion fraction from multiexponential diffusion analysis yielded the best correlation with cellularity, with minimal influence from underlying matrix degradation. T2* measurements were sensitive to macromolecular content without confounding influence from tissue hydration.

Anne-Sophie Glantenay¹, Chiel J Den Harder¹, Johan S Van Den Brink¹, Gwenael Herigault², and Jos Koonen^3
1Advanced Development, Philips Healthcare, Best, Netherlands, ²MR Cliniclal Science, Philips Healthcare, Best, Netherlands, ³MR Development, Philips Healthcare, Best, Netherlands

Diffusion-weighted EPI has become an indispensable tool in body MRI. However geometric distortions due to field inhomogeneities may hamper comparison of the diffusion images with T2W_TSE images. Here we implement and evaluate an efficient geometry correction method based on the non phase–encoded EPI reference acquisition used for Nyquist ghost removal. The method provides accurate and robust global geometry correction in the absence of strong, local phase offsets. It does not require additional time for calibrations, and is directly compatible with parallel imaging methods.

Aziz H Poonawalla¹, Catherine DG Hines¹, Diego Hernando¹, Pablo Irarrazaval^1,2, and Scott Brian Reeder^1
1Radiology, University of Wisconsin, Madison, WI, United States, ²Biomedical Engineering, Pontificia Universidad Catolica de Chile, Santiago, Chile

Fully-automated measurement of total adipose tissue using quantitative chemical shift MRI is demonstrated using a quantitative volumetric chemical-shift fat/water imaging method in phantoms. Fat, water and quantitative fat-fraction images were acquired at 3T with a 32-channel phased array in 26 sec of a series of peanut oil-filled phantoms of known volume and increasing surface area via inclusion of agar balls, glass rods, and empty vials. Robust fat volumes were measured in excellent agreement (r2 = 0.99) with known volumes of oil, with no user intervention, and immunity to partial volume effects.

Henriette Bertram¹, Gregor Thörmer¹, Florian Dazinger¹, Matthias Raschpichler¹, Nikita Garnov¹, Thomas Kahn¹, Matthias Blüher², and Harald Busse^1
1Department of Diagnostic and Interventional Radiology, University Hospital of Leipzig, Leipzig, Saxony, Germany, ²Department of Endocrinology and Nephrology, University Hospital of Leipzig, Leipzig, Saxony, Germany

Abdominal fat is associated with an increased risk for various diseases and can be quantified with MRI. Manual segmentation of subcutaneous (SAT) and visceral adipose tissue (VAT) is time consuming while automatic analyses are prone to errors. Therefore, an automatic technique with manual correction option was developed. Good agreement was observed between two readers for manual VAT and SAT quantification in 10 obese patients. Automatic volumes were 17±8% (VAT, p<0.001) and 1±3% (SAT, p=0.397) higher than the manually derived. These preliminary results suggest that the combined approach holds great promise for a fast (6 min) and sufficiently accurate fat quantification.

Anqi Zhou¹, Horacio Murillo¹, and Qi Peng^1
1Radiology, UT Health Science Center at San Antonio, San Antonio, TX, United States

Fast and accurate abdominal fat quantification has been more challenging on non-water-saturated (NWS) than on water-saturated (WS) MR images. We propose a general approach based on a combination of fuzzy c-means (FCM) clustering and signal intensity thresholding. The novel FCM algorithm utilizes the signal intensity of the neighboring voxels in addition to that of each voxel to eliminate full-volume water voxels from the original image. Statistically consistent results were obtained when the same algorithm was applied to both WS and NWS MR images. It is therefore a general approach to non-subjective, reproducible abdominal fat quantification on MR images.

Jesper Lundbom¹, Antti Hakkarainen², Sanni Söderlund³, Jukka Westerbacka³, Nina Lundbom¹, and Marja-Riitta Taskinen^3
1HUS Medical Imaging Center, Helsinki, Finland, ²HUS Medical Imaging Center, ³Department of Medicine, Helsinki University

We used long TE 1H-MRS to characterize visceral adipose tissue saturated fat. The CH2/CH3 ratio correlated positively with liver fat content in 14 male subjects. This is in contrast to a previously reported negative correlation between subcutaneous CH2/CH3 and liver fat content.

Aziz H Poonawalla¹, Brett P Sjoberg¹, Michael Schroeder¹, Diego Hernando¹, Pablo Irarrazaval^1,2, and Scott Brian Reeder^1
1Radiology, University of Wisconsin, Madison, WI, United States, ²Biomedical Engineering, Pontificia Universidad Catolica de Chile, Santiago, Chile

Quantitative measurement of visceral adipose tissue (VAT) is demonstrated using a volumetric chemical-shift fat/water imaging method with corrections for relaxation effects and accurate spectral modeling of fat. Fat, water and quantitative fat-fraction images were acquired over the entire abdomen and pelvis at 3T with a 32-channel phased array in a single 26-sec breath-hold. Robust VAT volumes were measured semi-automatically with <30 min post-processing, with strong correlation (r2 = 0.99) to manually-segmented VAT requiring ~4 hours.

Douglas E Befroy^1,2, Kitt Falk Petersen², Peter B Brown¹, Douglas L Rothman^1,3, and Gerald I Shulman^2,4
1Diagnostic Radiology, Yale University School of Medicine, New Haven, CT, United States, ²Internal Medicine, Yale University School of Medicine, New Haven, CT, United States, ³Biomedical Engineering, Yale University School of Medicine, New Haven, CT, United States, ⁴Howard Hughes Medical Institute, New Haven, CT, United States

Monitoring liver TCA cycle metabolism in vivo using classic ¹³C labeling strategies (infusing 1-¹³C glucose or 2-¹³C acetate) is unfeasible since intrahepatic lipid obscures the detection of C₄-glutamate enrichment. Recently, alternative labeling schemes have been demonstrated in brain whereby enrichment at C₅-glutamate, which resonates in a region free from overlapping lipid peaks, was monitored following the infusion of 2-¹³C glucose or 1-¹³C acetate. By adopting this strategy, we established that C₅-glutamate enrichment can be detected in human liver without interference from intracellular lipid and observed hepatic substrate oxidation via the TCA cycle for the first time.

Peter Edward Thelwall¹, Fiona Elizabeth Smith¹, Matthew Clemence², Kieren G Hollingsworth¹, Roy Taylor¹, and Michael P Gamcsik^3
1Newcastle Magnetic Resonance Centre, Newcastle University, Newcastle upon Tyne, Tyne and Wear, United Kingdom, ²Philips Healthcare - Clinical Science, Guildford, United Kingdom, ³Joint Department of Biomedical Engineering, University of North Carolina / NC State University, United States

Glutathione is a tripeptide that is a central component of cellular defences against oxidative stress. We have shown that a ¹³C label can be introduced into glutathione in humans by oral administration of [2-¹³C]-glycine, where this amino acid is incorporated into glutathione. We have shown that ¹³C-labelled glutathione can be detected by 1H-decoupled 13C spectroscopy in human liver, allowing measurements of glutathione concentration and synthesis rate and providing an insight into cellular defences against oxidative stress.

Shing-Ru Chen¹, Yi-Ru Lin², Posse Stefan^3,4, and Shang-Yueh Tsai^5
1Graduate Institute of Biomedical Engineering, National Taiwan University of Science and Technology, Taipei, Taiwan, ²Department of Electronic Engineering, National Taiwan University of Science and Technology, Taipei, Taiwan, ³Department of Neurology, School of Medicine, University of New Mexico, Albuquerque, New Mexico, United States, ⁴Department of Electrical And Computer Engineering, University of New Mexico, Albuquerque, New Mexico, United States, ⁵Department of Electrical Engineering, Chang Gung University, Taoyuan, Taiwan

Fatty liver as a common disease in modern life can be associated with a variety of disorders. Quantification of liver fat content relies on liver biopsy. Recently it has been shown that MRS can be used as an alternative method to measure liver fat. In this preliminary report, proton echo-planar spectroscopic imaging (PEPSI) is proposed to quantify spatial distribution of fat content in liver. Feasibility and reproducibility of liver MRSI were investigated. Results showed that PEPSI is able to detect liver fat content within short acquisition time. And the measured lipid fat fraction is highly reproducible.

Gavin Hamilton¹, Michael S Middleton¹, Takeshi Yokoo¹, Lisa G Clark¹, and Claude B Sirlin^1
1Department of Radiology, University of California, San Diego, San Diego, California, United States

We examine the repeatability of liver triglyceride characterization by ¹H MR Spectroscopy using a theoretical triglyceride model. Estimates of triglyceride unsaturation obtained from ¹H MRS respiratory gated and free breathing acquisitions were compared. ¹H MR Spectroscopy can repeatably measure the degree of liver fat unsaturation, and spectra collected with respiratory gating give a more repeatable characterization of liver fat composition than spectra collected with free breathing.

Kristen Zakian¹, Jing Qi¹, Lawrence Schwartz², Yuman Fong³, Leonard Saltz⁴, Nancy Kemeny⁴, Michael D'Angelica³, William Jarnagin³, and Jason Koutcher^1
1Medical Physics, Memorial Sloan-Kettering Cancer Center, New York, NY, United States, ²Radiology, Columbia University Medical School, ³Surgery, Memorial Sloan-Kettering Cancer Center,⁴Medicine, Memorial Sloan-Kettering Cancer Center

Liver toxicity related to chemotherapy is a growing concern in the combination use of chemo agents. To evaluate the chemotherapy related liver steatosis, serial ¹H MRS was applied to monitor hepatic triglyceride contents changes in 27 colorectal cancer patients during FOLFOX or FUDR/irinotecan treatment. Fat/fat+water ratio (FFW) measured from baseline and post treatment were compared in 17 patients who finished 12 cycle chemotherapy. 9 patients showed increased FFW greater than 15% compared to baseline, indicating chemotherapy associated steatosis. Serial¹H-MRS quantitatively evaluates the mobilization of liver fat, providing a more accurate way to detect chemo-associated steatosis.

Jan-Bernd Hövener¹, Uta Dahmen², Bernd Merkel³, Olaf Dirsch⁴, Jürgen Hennig¹, and Dominik von Elverfeldt^1
1Department of Radiology, Medical Physics, University Medical Center, Freiburg, Germany, ²Experimentelle Transplantationschirurgie, Klinik für Allgemein-, Viszeral- und Gefäßchirurgie, Jena, Germany, ³Fraunhofer MEVIS, Germany, ⁴Institut für Pathologie, Jena, Germany

test

Jing Qi¹, Mithat Gönen², Jinru Shia³, Lawrence Schwartz⁴, Nancy Kemeny⁵, Michael D'Angelica⁶, Yuman Fong⁶, Jason Koutcher¹, and Kristen Zakian^1
1Medical Physics, Memorial Sloan-Kettering Cancer Center, New York, NY, United States, ²Epidemiology-Biostatistics, Memorial Sloan-Kettering Cancer Center, ³Pathology, Memorial Sloan-Kettering Cancer Center, ⁴Radiology, Columbia University Medical School, ⁵Medicine, Memorial Sloan-Kettering Cancer Center, ⁶Surgery, Memorial Sloan-Kettering Cancer Center

To assess the reproducibility of hepatocellular triglyceride content measurements on a 1.5T scanner using ¹H MRS during a single breath-hold, 16 heathy adults were studied twice with a 2 week interval. The intra- and inter- examination coefficients of variance were 6% and 15% respectively. To evaluate the accuracy of ¹H MRS in quantitation of HTGC, fat/(fat+water) (FFW) from 16 colorectal cancer patients was correlated with liver histology. A positive correlation (r = 0.78, P < 0.001) was observed between FFW and histology steatosis grade. A FFW cutoff value of 0.025 results in 86% sensitivity and 89% specificity for detection of clinically relevant steatosis.

Marinette van der Graaf^1,2, and Geert J Wanten^3
1Clinical Physics Lab of Dept of Pediatrics, Radboud University Nijmegen Medical Centre, Nijmegen, Gelderland, Netherlands, ²Dept of Radiology, Radboud University Nijmegen Medical Centre, Nijmegen, Netherlands, ³Dept of Gastroenterology and Hepatology, Radboud University Nijmegen Medical Centre, Nijmegen, Netherlands

Patients with severe intestinal problems who receive Home Parenteral Nutrition (HPN) are at risk for developing hepatic dysfunctioning due to steatosis. In the present study, hepatic lipid content was determined by MRS in 13 patients on HPN. To allow for correction for T2 relaxation, single voxel STEAM measurements were performed at 3T using multiple echo times. Five patients had high hepatic lipid contents (12.8%±3.5%, mean±SD), while the other eight had normal liver fat concentration (1.1%±0.6%). In addition, most patients showed increased T2 relaxation behavior probably due to elevated hepatic iron or manganese content.

Jennifer Leigh Rehm¹, Vanessa A Curtis¹, Catherine DG Hines², Ellen L Connor¹, Aaron L Carrel¹, David B Allen¹, and Scott B Reeder^2,3
1Pediatrics, University of Wisconsin Hospital and Clinics, Madison, WI, United States, ²Radiology, University of Wisconsin Hospital and Clinics, Madison, WI, United States, ³Biomedical Engineering, University of Wisconsin Hospital and Clinics, Madison, WI, United States

Non-alcoholic fatty liver disease is a growing problem in the pediatric population and is anticipated to be the leading cause of liver failure in the future. Traditional methods of detecting fatty liver, such as ultrasound or liver enzyme blood screening, are insensitive to early stages of disease. We have demonstrated the feasibility of a rapid, clinically relevant, non-invasive method for early detection and quantitative staging of hepatic steatosis. While hepatic steatosis does correlate with elevated ALT, it is only moderately correlated with traditional markers of metabolic risk such as insulin resistance and hyper-triglyceridemia, and does not correlate with BMI. We have demonstrated that a novel quantitative MRI fat-fraction technique holds promise as a method for early detection, staging and treatment monitoring of NAFLD in obese adolescents.

Marek Chmelik¹, Stephan Gruber¹, Siegfried Trattnig¹, and Wolfgang Bogner^1
1MR Centre of Excellence, Department of Radiology, Medical University Vienna, Vienna, Austria

The feasibility of hepatic in vivo localized ³¹P-MRS at 7T in clinically acceptable measurement time was tested using three different localization strategies (1D-ISIS, 3D-ISIS and hybrid 1D-ISIS/2D-CSI). The quality of acquired spectra was compared in terms of SNR, linewidths and PCr contamination from surrounding muscle. Liver ³¹P MRS at 7T provides improved data quality in acceptable measurement time. 3D-ISIS and 1D-ISIS/2D-CSI provide excellent localization whereas 1D-ISIS provides highest SNR in shortest time but higher contamination.

Qiong Ye¹, Alexander Fuchs¹, and Markus Rudin^1,2
1University and ETH Zürich, Institute for Biomedical Engineering, Zürich, 8093, Switzerland, ²University of Zürich, Institute of Pharmacology and Toxicology, Zürich, Switzerland

Both hepatic lipids amount and their composition were accessed non-invasively with proton magnetic resonance spectroscopy on ob/ob, a well-established murine model of obesity, and their heterozygous ob/+ mice at 9.4T. In this work, the results from both ob/ob and ob/+ mice were compared. From the results, ob/ob mice showed significant higher hepatic lipid amount than ob/+ mice. Hepatic lipid from ob/ob mice were characterized with much more saturated components and longer molecular chain length. But no significant difference was obtained in the total unsaturated index, polyunsaturated bond index and total unsaturated bond index at the age of 24 weeks.

Nirilanto Ramamonjisoa¹, Helene Ratiney¹, Elodie Mutel², Herve Guillou³, Gilles Mithieux², Frank Pilleul^1,4, Fabienne Rajas², Olivier Beuf¹, and Sophie Cavassila^1
1CREATIS, CNRS UMR 5220, Inserm U1044, INSA-Lyon, Université Lyon 1, Université de Lyon, Lyon, France, ²Inserm U855, Université Lyon 1, Université de Lyon, France, ³INRA ToxAlim – Integrative Toxicology & Metabolism, Toulouse, France, ⁴Imagerie Digestive - CHU, Hospices Civils de Lyon, France

Glycogen storage disease type I (GSD1) is an autosomal recessive metabolic disorder resulting in severe impairment of glucose production and large accumulation of liver fatty acids. In this paper, short-TE in vivo MRS and gas chromatography (GC) analysis were used to evaluate the fatty liver content and the fat fraction composition in the mouse model of GSD1a under two different diets, a chow and a high fat high sucrose diet. Strong agreements between the quantitative lipid profiles measured by MRS and GC were demonstrated. Our results confirm that 1H MRS is a suitable non invasive tool for hepatic lipid quantification and lipid fraction estimation.

Gaspard d'Assignies^1,2, Ghislaine Fontés^3,4, Louis Gaboury⁵, Yvan Boulanger^6,7, Gilles Soulez³, Vincent Poitout^3,4, and An Tang^8
1Department of Medical Imaging, Hôpital Saint-Luc, Montreal, Quebec, Canada, ²Department of Radiology, Beaujon Hospital, Université Paris VII, Paris, France, ³CRCHUM, Canada, ⁴Montréal Diabetes Research Center, Canada, ⁵CHUM, Canada, ⁶Hôpital Saint-Luc, University of Montreal, ⁷CRCHUM, ⁸Department of Medical Imaging, CRCHUM, Montreal, Quebec, Canada

This study compares quantification of liver steatosis by MRI and histopathology by using MRS as the reference standard for the discrimination of three rat phenotypes assigned to an experimental glucolipotoxic steatosis model or a control group. A significant correlation was found between dual-echo MRI and MRS. A weaker correlation was found between histopathology and MRS. MRI and MRS accurately distinguished the rats receiving the infusion of glucose + Intralipid from those receiving the saline control, whereas histology did not. Although many authors use histopathology as the gold standard, vacuolar degeneration and glycogen may limit its accuracy for liver fat quantification.

Gavin Hamilton¹, Daniel L Smith², Mark Bydder¹, Krishna S Nayak³, and Houchun H Hu^3
1Department of Radiology, University of California, San Diego, San Diego, California, United States, ²Department of Nutrition Sciences, University of Alabama at Birmingham, Birmingham, Alabama, United States, ³Ming Hsieh Department of Electrical Engineering, University of Southern California, Los Angeles, California, United States

We explored the MR signatures of brown adipose tissue (BAT) compared to white adipose tissue (WAT) using single-voxel STEAM ¹H MR spectroscopy on a 3 Tesla clinical whole body scanner from excised murine adipose tissue samples. Spectra were acquired at multiple TEs and TIs to measure the T1, T2, and T2-corrected individual fat and water peak areas. A theoretical triglyceride model characterized the multi-peak fat spectra. The water fraction in BAT was higher compared to WAT and the T1 of water was lower in BAT compared to WAT. BAT also contained lower levels of unsaturated triglycerides in comparison to WAT.

Ulrich Flögel¹, Sara Tucci², Jürgen Schrader¹, and Ute Spiekerkoetter^2
1Institute for Cardiovascular Physiology, Heinrich Heine University, Düsseldorf, NRW, Germany, ²Department of General Pediatrics

Therapeutical approaches to limit metabolic derangement in patients with long chain fatty acid oxidation disorders (LCAD) include fat-modified diets, in which long-chain triglycerides are replaced by medium-chain triglycerides (MCT). The present study investigated the effects of a long-term MCT therapy on body lipid homeostasis using ¹H MRI and ¹H/¹³C MRS in a murine model of LCAD. One year MCT-based diet resulted in massive visceral fat infiltration, impaired body fat composition, and steatohepatitis. Although MCT diet has been reported to prevent development of cardiomyopathy and the myopathic phenotype, according to our data its use and dose has to be well considered.

Jürgen Machann¹, Christian Billich², Kathrin König¹, Christian Würslin¹, Fritz Schick¹, and Uwe Schütz^2
1Section on Experimental Radiology, University Hospital Tübingen, Tübingen, Germany, ²Department of Diagnostic and Interventional Radiology, University Hospital Ulm, Ulm, Germany

Twelve endurance athletes underwent follow-up MR examinations during the Transeuropean Footrace 2009 from Bari to the North Cape on a mobile MR-scanner for determination of whole-body tissue composition. A T1-weighted TSE sequence was applied and different adipose tissue (AT) compartments as well as muscle mass in lower and upper extremities were quantified and analyzed. All AT compartments showed a strong reduction during the race with the strongest decline for visceral AT (-70% compared to baseline). Muscle mass of lower extremities was also reduced. Comparison of these MRI-findings with biometrical data and laboratory findings will present further details regarding tissue metabolism and energy balance during long distance running.

Arunima Pola¹, Sandra Tan², Terry Yew Shze Keong¹, Zhihong Zhou², Mika Murabayashi³, Yoshihide Nakano³, Naoki Furuyama³, Nicholas Hird³, and Sendhil Sambashivam Velan^1
1Laboratory of Molecular Imaging, Singapore Bioimaging Consortium, Singapore, Singapore, ²Takeda Singapore Pte Ltd, Singapore, ³Takeda Pharmaceutical Company Ltd, Japan

Obesity is a medical condition in which excess body fat accumulates in areas such as abdomen, liver and skeletal muscle. In this study, experimental rodent models of obesity were used to investigate fat accumulation in these organs by longitudinal in vivo MRS/MRI measurements. Fat estimation in the abdomen was based on axial image segmentation. Spectral analysis by LCModel was used for fat calculation in liver and skeletal muscle. All our measurements are correlated with body weight, food intake and blood chemistry. The results demonstrate that MRI/MRS approach can be used to evaluate the potential of novel drugs in rodent models of obesity and diabetes.

Tedros Bezabeh¹, Omkar B Ijare¹, Nils Albiin², Matthias Lohr², Annika Bergquist², Urban Arnelo², and Ian C.P. Smith^1
1National Research Council Institute for Biodiagnostics, Winnipeg, MB, Canada, ²Karolinska University Hospital, Karolinska Institutet, Huddinge, Stockholm, Sweden

Pancreatic cancer is the fourth most common cause of cancer-related deaths with a very low overall 5-year survival rate (<5%). Proteomics-based analyses have detected molecular alterations (K-ras mutations, telomerase reactivation, or methylation of the tumor-suppressor genes) in pancreatic juice samples obtained from patients with pancreatic cancer. However, no definitive molecular markers have been identified yet for the diagnosis of pancreatic cancer. In this study, we have performed ¹H-MRS-based biochemical analysis of pancreatic juice samples from different pancreatic diseases including adenocarcinoma. Such analysis will serve as a basis for a metabolomics-based diagnostic approach for the detection of pancreatic cancer.

Sanaz Mohajeri^1,2, Tedros Bezabeh¹, Scott B. King¹, Omkar B. Ijare¹, M. A. Thomas³, Gerald Minuk², Jeremy Lipschitz², Iain Kirkpatrick², and Ian CP. Smith^1
1National Research Council Institute for Biodiagnostics, Winnipeg, Manitoba, Canada, ²University of Manitoba, Winnipeg, Manitoba, Canada, ³University of California, Los Angeles, California, United States

Magnetic Resonance Spectroscopy (MRS) is a method useful for the detection of mobile molecular species in a sample that have diagnostic utility. In vitro ¹H MRS of human bile samples has been successful in detecting various hepatopancreatobiliary disorders. Therefore, there is a great interest in performing such analysis in vivo. We performed 1-D and 2-D ¹H MRS of human gallbladder bile using a 3T clinical scanner and an optimized home-built receive array coil. Our preliminary efforts in obtaining good quality spectra are promising.

Qing Yuan¹, Ivan Dimitrov², Naim M Maalouf³, Khashayar Sakhaee³, and Paul T Weatherall^1
1Radiology, University of Texas Southwestern Medical Center at Dallas, Dallas, TX, United States, ²Philips Medical Systems, Cleveland, Ohio, United States, ³Internal Medicine, University of Texas Southwestern Medical Center at Dallas, Dallas, TX, United States

Noninvasive measurement of renal cortical fat content can potentially determine the degree of renal fat accumulation in various kidney disorders. The goal of this study was to develop a technique to perform in vivo 1H MRS in human kidney at 3T and to establish nominal values of fat content in normal volunteers. A respiratory-triggered non-water suppressed single-voxel PRESS sequence with TE-averaging was used to reduce motion and sideband artifacts. No fat content was detected from kidney parenchyma in healthy subjects. Voxel placement was crucial since including perirenal or hilar fat within the kidney resulted in fat contamination of the spectra.

Jacqueline T. Jonker¹, Ralph L. Widya², Sebastiaan Hammer², Linda D. van Schinkel¹, Rutger W. van der Meer², Eelco J.P. de Koning³, Henk J.G. Bilo⁴, Andrew Webb², Hermien E. Kan², and Hildo J. Lamb^2
1Endocrinology, Leiden University Medical Center, Leiden, Zuid-Holland, Netherlands, ²Radiology, Leiden University Medical Center, Leiden, Zuid-Holland, Netherlands, ³Nephrology, Leiden University Medical Center, Leiden, Zuid-Holland, Netherlands, ⁴Internal medicine, Isala klinieken (Sophia), Zwolle, Netherlands

Ectopic fat accumulation in type 2 diabetes is associated with insulin resistance and increased risk for cardiovascular disease. Although changes in diet have beneficial effects on ectopic fat accumulation, this is less clear for effects of exercise training. The purpose of the present study was to assess effects of a 7-month exercise program on ectopic fat accumulation in heart, liver and skeletal muscle measured by 1H-MRS at 1.5T and 7T. 7-month exercise decreases hepatic TG content. Myocardial and myocellular lipid content remained unchanged. Therefore, exercise intervention in patients with type 2 diabetes mellitus induces tissue-specific changes in ectopic fat distribution.