Enrico De Vita¹, Marta Vidorreta², Xavier Golay³, Maria Fernandez-Seara², David L Thomas³, and Evelyne Balteau^4
1Lysholm Department of Neuroradiology, National Hospital for Neurology and Neurosurgery, London, United Kingdom, ²Neuroimaging Laboratory, Center for Applied Medical Research, University of Navarra, Spain, ³Academic Neuroradiological Unit, Department of Brain Repair and Rehabilitation, UCL Institute of Neurology, London, United Kingdom, ⁴Cyclotron Research Centre, University of Liège, Liege, Belgium

Pseudo-continuous arterial spin labelling (pCASL) with 3DGRASE readout meets the recommendations of the AID network for clinical research and ASL applications to dementia. Typical spatial resolution in ASL is around 4x4x6mm^3. Single-shot 3DGRASE suffers from severe blurring in the partition/slice direction due to the excessively long echo train with respect to typical brain T2, especially when requiring higher spatial resolution and whole brain coverage. Multi-shot 3D-GRASE reduces echo train duration and related blurring. We propose a novel segmentation scheme for isotropic 3.2mm spatial-resolution acquisition combining in-plane and through-plane segmentation, improving through-plane point spread function and signal-to-noise ratio of 3DGRASE pCASL.

Xiufeng Li¹, Dingxin Wang², Steen Moeller¹, Kamil Ugurbil¹, and Gregory Metzger^1
1Center for Magnetic Resonance Research, University of Minnesota, Minneapolis, MN, United States, ²Siemens Medical Solutions USA Inc, Minneapolis, MN, United States

To achieve comprehensive insight into g-factor, leakage contamination and non-sequential slice acquisition order effects of MB EPI on ASL perfusion imaging, both imaging studies and theoretical evaluations were performed, indicating that even in the presence of inter-slice signal leakeage, MB EPI greatly benefits high-resolution whole brain ASL perfusion imaging with respect to both SNR efficiency and CBF quantification.

Enrico De Vita^1,2, Evelyne Balteau³, Xavier Golay², and David L Thomas^2
1Lysholm Department of Neuroradiology, National Hospital for Neurology and Neurosurgery, London, United Kingdom, ²Academic Neuroradiological Unit, Department of Brain Repair and Rehabilitation, UCL Institute of Neurology, London, United Kingdom, ³Cyclotron Research Centre, University of Liège, Liege, Belgium

3DGRASE is an extremely efficient acquisition strategy for ASL. Blurring in slice/partition direction, related to excessively long echo trains, is an issue in single-shot acquisitions. This blurring can be decreased with multi-shot acquisitions. In-plane geometric distortions related to the GE-EPI readout caused by magnetic susceptibility variations across the brain are still a nuisance. We test the effectiveness of a susceptibility distortion correction based on B0-field maps, similarly to that routinely used for fMRI. An M0 acquisition repeated by inverting the in-plane phase-encoding direction can be used to generate a usable B0 map in alternative to the conventional double-echo GE-2DFT acquisition.

Thomas W Okell¹, Michael A Chappell^1,2, Michael E Kelly^1,3, and Peter Jezzard^1
1FMRIB Centre, Nuffield Department of Clinical Neurosciences, University of Oxford, Oxford, Oxfordshire, United Kingdom, ²IBME, Department of Engineering Sciences, University of Oxford, Oxford, Oxfordshire, United Kingdom, ³Centre for Core Biotechnology Services, University of Leicester, Leicester, Leicestershire, United Kingdom

There is some debate regarding the ability of arterial spin labeling (ASL) to measure perfusion in white matter (WM). Here a multi-postlabeling delay vessel-encoded ASL sequence, capable of producing CBF and BAT estimates for each major brain-feeding artery separately, was used to demonstrate a significant delay in blood arrival within a restrictive WM mask compared to that in grey matter, showing that the WM signal does not arise simply from partial volume effects. In addition, WM CBF estimates arise almost entirely from the expected feeding artery, showing that these measurements are not dominated by noise.

Paul Kyu Han¹, Jong Chul Ye¹, Seung Hong Choi², and Sung-Hong Park^1
1Department of Bio and Brain Engineering, Korea Advanced Institute of Science and Technology, Daejeon, Korea, ²Department of Radiology, Seoul National University College of Medicine, Seoul, Korea

pCASL is a non-invasive technique that measures blood perfusion. Recently, pCASL with bSSFP has been proposed to overcome the limitations of EPI, however, the low temporal resolution may lead to limited spatial coverage. One potential approach to improve the spatial coverage of pCASL-bSSFP is to use compressed sensing (CS). We implemented the first accelerated in-vivo imaging of 3D pCASL-bSSFP with k-t FOCUSS at 3T for human brain. Experimental results show that CS acceleration by a factor of 4 works well for pCASL-bSSFP, which confirms that combination of CS and pCASL-bSSFP may be a good solution for distortion-free blood perfusion imaging.

Ze Wang^1
1Departments of Psychiatry and Radiology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, United States

Arterial spin labeling (ASL) MRI has a low signal-to-noise-ratio (SNR). Meanwhile, it has abundant spatial correlation and temporal prior information that can be exploited to improve the subsequent cerebral blood flow (CBF) quantification quality. To use those priors, we proposed a novel spatially constrained and temporally adapted method for ASL CBF quantification. The method demonstrated significantly improved SNR, spatial image quality, and clinical sensitivity of ASL CBF using normal healthy subjects’ data and data from patients with Alzheimer’s Disease.

Wouter M Teeuwisse^1,2, Sophie Schmid¹, Michael Helle³, and Matthias van Osch^1
1C.J. Gorter Center for High Field MRI, Leiden University Medical Center, Leiden, ZH, Netherlands, ²Leiden Institute for Brain and Cognition, Leiden, ZH, Netherlands, ³Innovative Technologies, Research Laboratories, Philips Technologie GmbH, Hamburg, Germany

In pCASL, shimming the B0 field for optimal homogeneity is indispensable for optimal labeling efficiency and image quality. In this study the implementation of dual-shim pCASL, with separate shim settings for imaging and labeling, is evaluated. This approach is compared with pCASL with identical shim settings for imaging and labeling. Mean gray matter CBF and temporal SNR are calculated for left-, right- and posterior flow territories and left-right asymmetry is assessed. It is concluded that, when 1st order shimming is applied, application of separate shim settings during labeling does not result in significant changes in quantitative CBF or temporal SNR

Giacomo Tarroni¹, Marco Castellaro¹, Carlo Boffano², Maria Grazia Bruzzone², Alessandra Bertoldo¹, and Enrico Grisan^1
1Department of Information Engineering, University of Padova, Padova, Italy, ²Neuroradiology Department, IRCCS Foundation Neurological Institute "C. Besta", Milano, Italy

Motion correction in Arterial Spin Labelling (ASL) is essential to accurately assess brain perfusion. The combination of ASL with background suppression determines a non-uniform magnetization, potentially hindering the application of conventional intensity-based registration methods. Accordingly, a motion correction technique based only on brain contour points is presented and tested on control and tagged acquisitions. The technique relies on image segmentation (to extract brain contour points) and on the iterative closest point algorithm (to estimate the roto-translation required to align them). Error metrics between contours manually traced on reference, moving and corrected volumes show the accuracy of the proposed registration technique.

Ovidiu Cristian Andronesi¹, Uvo C. Hoelscher², Himanshu Bhat³, Keith Heberlein³, and Bruce R. Rosen^1
1Martinos Center for Biomedical Imaging, Department of Radiology, Massachusetts General Hospital, Harvard Medical School, Boston, MA, United States,²Siemens Healthcare, AG, Erlangen, Bayern, Germany, ³Siemens Healthcare USA, Boston, MA, United States

Improving SNR is critical for the ASL methods. Inversion of magnetization of the blood spins is a main factor responsible for the SNR in ASL. With the ongoing trend in the development of MR hardware towards large-bore scanners and the active control of the RFPA duty cycle there is an increased need for efficient and low-power adiabatic inversion. Here we present the use of Gradient Offset Independent Adiabaticity GOIA-W(16,4) pulses for low power adiabatic tagging of blood in ASL sequences. GOIA pulses require only half of the B1max of the FOCI pulses that are typically used in ASL. Hence, the B1max of GOIA pulses is not clipped and can be increased above the adiabatic threshold allowing to compensate for B1 inhomogeneity. Perfusion maps obtained in human brain show that SNR is improved overall with GOIA pulses. Largest SNR gain (50%) is noticed especially in the middle of the brain as expected from simulations.

Makoto Obara¹, Morio Nagahata², Tomoyuki Okuaki³, Nobuko Minagawa⁴, Masanobu Nakamura¹, and Marc Van Cauteren^3
1Philips Electronics Japan, Minato-ku, Tokyo, Japan, ²Stroke Center, Yamagata City Hospital SAISEIKAN, Yamagata, Japan, ³Philips Healthcare, Asia Pacific, Tokyo, Japan, ⁴Department of Radiology, Yamagata City Hospital SAISEIKAN, Yamagata, Japan

The multiple time points territorial arterial spin labelling (tASL) was conducted in healthy volunteers and patients with cerebrovascular occlusive disease. We propose an accumulated-area map, accumulating the perfusion pixels picked up by t-test, of all phases in each vessel territory. We also calculate the Percentage Perfusion-Territory Area (%PT-A) from the accumulated-area map. We discuss the validity and clinical usefulness of this unique approach to the tASL.

Zungho Zun¹, R. Marc Lebel², Ajit Shankaranarayanan³, and Greg Zaharchuk^1
1Stanford University, Stanford, CA, United States, ²GE Healthcare, Calgary, AB, Canada, ³GE Healthcare, Menlo Park, CA, United States

A recent arterial spin labeling (ASL) white paper recommends pseudocontinuous ASL (PCASL) with labeling duration (LD)/post-labeling delay (PLD) of 1.8/2.0 s for clinical applications. Compared to PLD, there has been little exploration in optimization of LD. In this study, CBF measurement was performed in volunteers using PCASL with different LD, and ASL signal-to-noise ratio (SNR) efficiency was compared among scans with different LD. Measured SNR efficiency showed good agreement with simulation and was maximized when LD was 3-4 s, achieving about 30% gain compared to LD = 1.5 s. Potential reduction in apparent CBF measurement with prolonged LD was not found.

Jason Kaig Mendes¹, Christopher J Hanrahan¹, Jeff L Zhang¹, Gwenael Layec², Corey Hart², and Vivian S Lee^2
1Radiology, University of Utah, Salt Lake City, UT, United States, ²Medicine, University of Utah, Salt Lake City, UT, United States

The severity and treatment success of peripheral arterial disease (PAD) has been correlated with the integrity of peripheral vasculature. Muscle perfusion and tissue oxygenation are indicators of microvascular function and can be non-invasively measured using blood-oxygen-level dependent (BOLD) and pulsed arterial spin labeling (PASL) techniques. However, there is compelling evidence that quantification of multiple parameters may improve diagnosis and treatment response of PAD. In particular, muscle spin lattice relaxation (R2) can be used to more accurately quantify perfusion and tissue oxygenation in conjunction with BOLD and PASL data. We propose an adaptation of a current BOLD-PASL sequence which combines R2, BOLD and ASL techniques (RBASL) to simultaneously measure R2, R2* and perfusion.

Henri JMM Mutsaerts¹, Rebecca ME Steketee², Dennis FR Heijtel¹, Joost PA Kuijer³, Matthias JP van Osch⁴, Charles BLM Majoie¹, Marion Smits², and Aart J Nederveen^1
1Academic Medical Center, Amsterdam, Noord-Holland, Netherlands, ²Erasmus MC Medical Center, Rotterdam, Zuid-Holland, Netherlands, ³VU Medical Center, Amsterdam, Noord-Holland, Netherlands, ⁴Leiden University Medical Center, Leiden, Zuid-Holland, Netherlands

The current study compares 3T pseudo-continuous arterial spin labeling cerebral blood flow measurements acquired in two centers, at a General Electric (GE) and Philips scanner. The main difference between the GE and Philips acquisitions is the readout module: 3D vs. 2D. This study shows that arterial spin labeling results from different vendors can be compared for the total gray matter region of interest. However, for the white matter or for smaller gray matter regions standardization of readout modules between vendors is warranted.

Henri JMM Mutsaerts¹, Matthias JP van Osch², Fernando O Zelaya³, Mikjel J Skurdal⁴, Wibeke Nordhøy⁵, Yi Wang⁶, Oliver Geier⁵, Aart J Nederveen¹, Atle Bjørnerud⁵, Danny JJ Wang⁶, and Inge R Groote^5
1Academic Medical Center, Amsterdam, Netherlands, ²Leiden University Medical Center, Leiden, Netherlands, ³Kings College London, London, United Kingdom, ⁴Akershus University Hospital, Oslo, Norway, ⁵Oslo University Hospital, Oslo, Norway, ⁶UCLA, Los Angeles, California, United States

The three major MRI vendors have chosen different implementations of ASL, hampering the validity of cross-vendor comparisons and multi-site studies. The current study compares 3T pseudo-continuous ASL (pCASL) CBF-measurements acquired at two major vendors MRI systems: Philips and Siemens. For this study, pCASL sequences were created with near complete similarity between the two vendor systems. These data indicate that if an identical sequence is used, cross-vendor CBF differences are small. Currently, we are including a third major vendor (GE) in the comparison. Data from the three-vendor comparison will be available by January 2014.

Xiufeng Li¹, Lynn E. Eberly², Gulin Oz¹, Elizabeth R. Seaquist³, and Silvia Mangia^1
1Center for Magnetic Resonance Research, University of Minnesota, Minneapolis, MN, United States, ²Division of Biostatistics, School of Public Health, University of Minnesota, MN, United States, ³Division of Endocrinology and Diabetes, Department of Medicine, University of Minnesota, MN, United States

While many studies assessed the reproducibility of ASL, very few studies investigated pCASL reproducibility. More importantly, pCASL reproducibility has not been reported in specific sub-cortical brain regions such as the thalamus and the cerebellum. These regions-of-interest are of particular interest for applications in diabetes9 and neurodegenerative diseases. Here we obtained test-retest data on healthy human subjects to evaluate data variability within and across sessions using a whole-brain pCASL protocol at 3T.

Arnaud Guidon¹, Weiying Dai², and David C. Alsop^2
1GE Healthcare Global MR Applications and Workflow, Boston, MA, United States, ²Department of Radiology, Beth Israel Deaconess Medical Center, Boston, MA, United States

Arterial Spin Labeling is well-suited for evaluating patients with renal insufficiency for whom risk of gadolinium-based contrast administration is undesirable. However it is vulnerable to misalignment errors predominantly due to abdominal displacements during respiration. Strategies have previously been proposed either relying on controlled-breathing or navigators. The latter is incompatible with background suppression and the former still requires active patient¹s cooperation. Here, we propose an alternative approach allowing background suppressed Free-Breathing Renal Perfusion imaging with retrospective elastic registration of individual tag and control images and report on its relative robustness for 3 different background suppression levels.

Randall B Stafford^1,2, Michael C Langham¹, Scott V Davis³, Jongho Lee^1,2, and John A Detre^2,4
1Radiology, University of Pennsylvania, Philadelphia, PA, United States, ²Center for Functional Neuroimaging, University of Pennsylvania, Philadelphia, PA, United States, ³Psychiatry, Duke University, NC, United States, ⁴Neurology, University of Pennsylvania, Philadelphia, PA, United States

The purpose of this work is to develop a standardization phantom for quantitative perfusion with arterial spin labeling. A two-chamber phantom was constructed consisting of glass beads and synthetic fiber mesh . The phantom was tested using pCASL labeling with a FLASH-based readout at various post-label delays and flow rates. As expected, increasing the flow rate resulted in corresponding increases in the percent difference images. Such a phantom could be used for QA and sequence standardization across ASL sequences, vendors, and sites.

Lirong Yan¹, Collin Liu², Robert Smith¹, John Ringman¹, and Danny JJ Wang^1
1University of California Los Angeles, Los Angeles, CA, United States, ²University of Southern California, CA, United States

Both arterial stiffening (or reduced vascular compliance) and cerebral hypoperfusion have been found to be associated with aging and dementia. The purpose of this study is to investigate the relationship between intracranial vascular compliance and perfusion measured using a novel dynamic arterial-spin labeling (ASL) technique and pseudo-continuous ASL 3D GRASE MRI respectively. The results show that vascular compliance and perfusion are correlated with each other and both decrease with aging and vascular risks. This study supports arterial stiffening and cerebral hypoperfusion as markers of vascular risks and dementia.

Esther A.H. Warnert¹, Judith E. Hall², and Richard G. Wise^1
1CUBRIC. School of Psychology, Cardiff University, Cardiff, United Kingdom, ²Department of Anaesthetics and Intensive Care Medicine, School of Medicine, Cardiff University, United Kingdom

We investigate the possibility of using pulsed arterial spin labelling (PASL) to asses intracranial arterial compliance, a potential marker of cerebrovascular disease. To do so, we have acquired short inversion time PASL images (TI < 900ms) and retrospectively grouped these into systolic and diastolic phases of the cardiac cycle. Changes in arterial blood volume are calculated whilst taking the differences between label kinetics in systole and diastole into account that arise due to pulsatility of blood flow. Our results indicate that short inversion time PASL has the potential to assess arterial compliance of cerebral blood vessels.

Kaining Shi¹, Xin Lou², and Lin Ma^2
1Global MR Applications and Workflow (China), GE HealthCare, Beijing, Beijing, China, ²Radiology Department, PLA General Hospital, Beijing, Beijing, China

Arterial spin labeling (ASL) can provide cerebral perfusion images noninvasively but only obtain the cerebral blood flow (CBF), which is sensitive to the transit time effect, compared to the dynamic susceptibility contrast imaging (DSC-MRI). The combination of two ASL sequences with different post label delay time is an efficient way to evaluate both blood flow and transit time. In this work, the CBF difference between 2 PLDs was compared to DSC-MRI parameters. The CBF difference has high correlation with TTP in DSC-MRI.

Esther A. H. Warnert¹, Andrew Lansdown², Kevin Murphy¹, Judith E. Hall³, D. Aled Rees², and Richard G. Wise^1
1CUBRIC, School of Psychology, Cardiff University, Cardiff, United Kingdom, ²Centre for Endocrine and Diabetes Sciences, Institute of Experimental and Molecular Medicine, School of Medicine, Cardiff University, United Kingdom, ³Department of Anaesthetics and Intensive Care Medicine, School of Medicine, Cardiff University, United Kingdom

This abstract contains preliminary results of a clinical study investigating brain function and physiology in women suffering from Poly-Cystic Ovary Syndrome (PCOS). Here, we focus on the brainstem (and the medulla in particular), because of the recent finding that a potential marker of PCOS is hyperactivity of the sympathetic nervous system, in which nuclei in the medulla play a central role. Arterial spin labelling is used to assess differences in brainstem cerebral blood flow between PCOS patients and healthy controls.

Zungho Zun¹, R. Marc Lebel², Deqiang Qiu³, Ajit Shankaranarayanan⁴, and Greg Zaharchuk^1
1Stanford University, Stanford, CA, United States, ²GE Healthcare, Calgary, AB, Canada, ³Emory University, Atlanta, GA, United States, ⁴GE Healthcare, Menlo Park, CA, United States

Cerebral blood flow (CBF) reactivity to vasodilation provides better prognosis for vascular disease, and arterial spin labeling (ASL) may be ideal for this because of no burden of ionizing radiation or contrast agents. In this study, four different ASL methods were performed in subsets of sixteen moyamoya disease patients before and after acetazolamide (ACZ) injection and were compared to the reference standard xenon CT CBF with ACZ. Multi-delay PCASL showed the highest correlation with xenon CT in measuring CBF. For measuring CBF augmentation, however, all three PCASL methods showed similar correlation and VS-ASL showed lower correlation.

Renata Ferranti Leoni^1,2, Icaro Agenor Ferreira de Oliveira¹, Octavio Marques Pontes Neto², and João Pereira Leite^2
1Department of Physics - FFCLRP, University of Sao Paulo, Ribeirao Preto, Sao Paulo, Brazil, ²Department of Neuroscience and Behavioral Sciences - FMRP, University of Sao Paulo, Ribeirao Preto, Sao Paulo, Brazil

Altered cerebral perfusion has been observed in normal aging. To understand these alterations it is important to comprehend age-related neurological disorders. In this study, pulsed arterial spin labeling and hypercapnia were used to investigate regional cerebral blood flow (CBF) and cerebrovascular reactivity (CVR) in elderly healthy volunteers. Decreased CBF and CVR were observed across the entire cerebral cortex in elderly subjects compared to young subjects. This reduction may be related to arteriosclerosis and small vessel disease. Further studies may use the same protocol to investigate brain areas vulnerable to neurological disorders due to impaired vascular reserve.

Lindsey M Dethrage¹, Carlos C Faraco¹, Megan K Strother¹, and Manus J Donahue^1
1Vanderbilt University, Nashville, Tennessee, United States

The purpose of this study is to quantify the change in blood water R₁ that results from hypercarbic hyperoxia (i.e., carbogen) administration in humans. Owing to the small-to-negligible effect of hyperoxia on CBF, it is possible to fit for the blood water R₁ change during carbogen administration by performing sequential experiments with 5%CO₂/21%O₂/74%N₂ and 5%CO₂/95%O₂. Results reveal that blood water R₁ increases from 0.62 at baseline (21% O₂) to 0.73+/-0.07s^-1 during carbogen breathing.

Guido Hugo Jajamovich¹, Wei Huang², Cecilia Besa¹, Xin Li², Aneela Afzal², Hadrien Dyvorne¹, and Bachir Taouli^1
1Icahn School of Medicine at Mount Sinai, New York, NY, United States, ²Oregon Health & Science University, Portland, OR, United States

This study quantifies hepatocellular carcinoma (HCC) and background parenchyma perfusion parameters with DCE-MRI using two pharmacokinetic models: the Tofts Model (TM) and the Shutter-Speed Model (SSM). Parameters ve and kep from the TM and Ktrans and τi from the SSM showed significant differences between liver parenchyma and HCC, with better parameter reproducibility in liver parenchyma compared to HCC for both models and with the SSM showing worse reproducibility overall. The differences in τi may possibly reflect differences in metabolic activity between the two tissues, suggesting potential utility of this parameter for HCC characterization.

Amy Kuceyeski¹, Mitchell Cooper¹, Shan Hu¹, Martin R Prince¹, Yi Wang¹, and Pascal Spincemaille^1
1Radiology, Weill Cornell Medical College, New York, NY, United States

Chronic damage to the liver can cause fibrosis, or scarring of the tissue that can cause portal hypertension. Non-invasive gadolinium (Gd) enhanced MRI is used to study liver blood flow. Here, the liver’s uptake of Gd is modeled using a system of ordinary differential equations. A Bayesian framework that incorporates data and prior information is implemented and the resulting posterior density explored using Markov Chain Monte-Carlo methods. These parameter estimates are used to investigate the relative contribution of the hepatic artery to the total blood flow in the liver; we show that this value is significantly higher in fibrosis.

Bio Joo¹, Mi-Suk Park¹, Yeo-Eun Kim¹, and Sang-Joon Shin^2
1Department of Radiology and Research Institute of Radiological Science, Yonsei University College of, Seoul, Korea, ²Cancer Metastasis Research Center, Yonsei University College of Medicine, Seoul, Korea

We assessed the clinical outcomes of colorectal cancer liver metastasis (CRCLM) patients treated with Bevacizumab plus FOLFOX by Dynamic Contrast Enhanced (DCE)-MR imaging and plasma cytokines and angiogenic factors (CAFs). Significant changes in DCE-MRI and CAFs on 3 day after bevacizumab monotherapy proved anti-angiogenic effect of bevacizumab monotherapy. Higher baseline iAUC60 and early reduction of Ktrans could predict longer PFS in the patients with CRCLM.

David K Woolf¹, N. Jane Taylor², Nina Tunariu³, Andreas Makris¹, Sonia P Li¹, Mei-Lin W Ah-See¹, Mark J Beresford⁴, J. James Stirling², James A d'Arcy³, David J Collins³, and Anwar R Padhani^2
1Academic Department of Oncology, Mount Vernon Hospital, Northwood, Middlesex, United Kingdom, ²Paul Strickland Scanner Centre, Mount Vernon Hospital, Northwood, Middlesex, United Kingdom, ³CR-UK-EPSRC Cancer Imaging Centre, Institute of Cancer Research & Royal Marsden Hospital, Sutton, Surrey, United Kingdom, ⁴Royal United Hospital Bath, Bath, United Kingdom

We evaluated the performance of six modelled Arterial Input Functions (AIFs) used with the standard Toft’s model in the setting of primary breast cancer treatment with neoadjuvant (NAC) chemotherapy. Quantitative parameters were calculated for 27 patients at baseline and after 2 cycles of NAC. We assessed the models’ abilities to fit patient DCE-MRI data, provide physiologically-plausible results, and to predict pathologic complete response (pCR). The modified Fritz-Hansen model performed best with area under the ROC curve 0.74 and a sensitivity of 87.5% for pCR prediction. The Weinmann and femoral artery AIFs performed least well.

John Carr¹, Daniel Margolis¹, Steven Raman¹, and Kyung Sung^1
1Department of Radiological Sciences, UCLA, Los Angeles, California, United States

We assess pharmacokinetic modeling of DCE-MRI in prostate cancer where histo-pathology is available to investigate the ability of DCE-MRI to delineate between cancerous and normal prostate tissues. We show measuring dispersion which does not require an AIF can gives a dramatic increase in specificity compared to Ktrans. We use the AATH model to quantify perfusion and capillary permeability thus decoding Ktrans and allowing a more complete understanding of tissue physiology to be had. We believe combining dispersion mapping with physiological measurements derived from DCE-MRI using the AATH model will offer a powerful method for prostate cancer detection and active surveillance.

Junyu Guo¹, Wilburn E. Reddick¹, John Glass¹, Jianrong Wu¹, M. Beth McCarville¹, Barry L. Shulkin¹, Najat C. Daw¹, and Fariba Navid^1
1St Jude Children's Research Hospital, Memphis, TN, United States

This study investigated DCE-MRI and 18F-FDG PET for evaluating response to antiangiogenic and neoadjuvant chemotherapy and prognosis of event-free survival (EFS) and overall survival in 42 patients with Osteosarcoma. Ktrans and vp significantly dropped after 24 hours of the first bevacizumab administration. SUVmax strongly correlated with several DCE-MRI parameters at the different time points. Ktrans, tumor volume (TV), and SUVmax at week10 were significantly different between responder and nonresponders. Ktrans at week10 and TV on Day-2 and at week 5 were significantly associated with EFS and overall survival, and could serve as potential prognostic factors for these clinical outcomes.

Xiaoping Zhu¹, Sha Zhao², Ka-Loh Li¹, and Alan Jackson^1
1Wolfson Molecular Imaging Centre, University of Manchester, Manchester, United Kingdom, ²Imaging Science and Biomedical Engineering, University of Manchester, Manchester, United Kingdom

In dual-bolus-DCE-MRI, R1 values prior to the main-dose (R1init) were estimated in NF2 patients. R1init are significantly larger than native R1 (R1N) in vascular schwannomas and meningiomas, but not in normal brain tissue. R1init estimated from prebolus kinetic analysis showed close to those derived from variable flip angle (VFA)-GRE sequences. The R1N and R1init maps from VFA were then used in kinetic analysis of prebolus and main-dose DCE data respectively. The Ktrans from prebolus (one-fifth dose) were comparable to those from high spatial resolution main-dose data, achieving the same detection of changes in the brain tumors induced by VEGF-blockade.

Milica Medved¹, Aytekin Oto¹, Xiaobing Fan¹, Federico D Pineda¹, Russell Z Szmulewitz², and Gregory S Karczmar^1
1Radiology, University of Chicago, Chicago, Illinois, United States, ²Medicine, University of Chicago, Chicago, Illinois, United States

The determination of the arterial input function (AIF) is critical to applications of the two compartment model to tissue contrast agent uptake and washout. As the AIF cannot always be measured directly, approximate methods are sometimes used to derive it, such as using muscle – with known K^trans and v_evalues – as a reference tissue. Published values for skeletal muscle, such as the calf, are often used, but we show significant differences in K^trans and v_ebetween gluteal and deep pelvic muscles. We conclude that skeletal muscles should be characterized individually.

Meike Weidner¹, Frank G. Zöllner², Claudia Hagelstein¹, Thomas Schaible³, Katrin Zahn⁴, Lothar R. Schad², Stefan O. Schoenberg¹, and K. Wolfgang Neff^1
1Institute of Clinical Radiology and Nuclear Medicine, University Medical Center Mannheim, Mannheim, Germany, ²Computer Assisted Clinical Medicine, Medical Faculty Mannheim, Heidelberg University, Mannheim, Germany, ³Department of Pediatrics, University Medical Center Mannheim, Mannheim, Germany, ⁴Department of Pediatric Surgery, University Medical Center Mannheim, Mannheim, Germany

2-year old children after congenital diaphragmatic hernia (CDH) repair show significantly reduced lung perfusion values on the ipsilateral side in quantitative MR imaging. A temporal resolution of 1.5 s with an isotropic voxel size of 2 mm3 brings a gain of peak contrast to noise ratio (PCNR) and leads to higher perfusion values, when compared to a higher spatially resolved protocol. Consequently, these image parameters should be selected in routine follow-up programs.

Leon C. Ho^1,2, Ian P. Conner³, Seong-Gi Kim^1,4, Ed X. Wu², Chi-Wai Do⁵, Gadi Wollstein³, Joel S. Schuman³, and Kevin C. Chan^1,3
1Neuroimaging Laboratory, Department of Radiology, University of Pittsburgh, Pittsburgh, Pennsylvania, United States, ²Department of Electrical and Electronic Engineering, The University of Hong Kong, Pokfulam, Hong Kong, China, ³Department of Ophthalmology, University of Pittsburgh, Pittsburgh, Pennsylvania, United States, ⁴Center for Neuroscience Imaging Research, Institute for Basic Science (IBS), Dept. of Biological Sci, SKKU, Suwon, Korea,⁵School of Optometry, The Hong Kong Polytechnic University, Hung Hom, Hong Kong, China

Balanced aqueous humor flow dynamics is crucial to maintain healthy ocular physiology. Imbalanced aqueous humor flow dynamics would lead to altered intraocular pressure and retinal damage or visual loss in glaucoma disease. To date, the relationships between eye pressure, aqueous humor flow and glaucoma are not fully evaluated. Gd-enhanced MRI may non-invasively visualize flow dynamics of aqueous humor. In this study, dynamic Gd-enhanced MRI was employed to evaluate in vivo the ocular physiology and biotransport in a rat model of microbead-induced ocular hypertension and in healthy, normotensive rats after topical applications of 3 different ophthalmic hypotensive eye drops.

Christopher Larsson^1,2, Jayashree Kalpathy-Cramer³, Atle Bjørnerud^1,4, Patrick Y. Wen⁵, Tracy T. Batchelor^6,7, Rakesh K. Jain⁶, A. Gregory Sorensen^3,8, and Kyrre E. Emblem^1,3
1The Intervention Centre, Oslo University Hospital, Rikshospitalet, Oslo, Oslo, Norway, ²Faculty of Medicine, University of Oslo, Oslo, Oslo, Norway,³Department of Radiology and Athinoula A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts, United States, ⁴Department of Physics, University of Oslo, Oslo, Norway, ⁵Center for Neuro-Oncology, Dana-Farber/Brigham and Women’s Cancer Center and Harvard Medical School, Massachusetts, United States, ⁶Department of Radiation Oncology, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts, United States, ⁷Department of Neurology, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts, United States, ⁸Siemens Healthcare, Malvern, PA, United States

The role of DCE-MRI for assessment of tumor permeability in clinical decision making remains controversial. The usefulness of this technique is limited by the low reproducibility of parameters such as the capillary permeability transfer constant Ktrans in longitudinal analysis. In this study we demonstrate substantial improvements in reproducibility of Ktrans by re-use of patient-specific arterial input functions (AIFs) compared to determining the AIF at each scan. This suggests that patient-specific AIFs adjust for variations in systemic circulation while minimizing scan-rescan variability.

Siamak P. Nejad-Davarani^1,2, Hassan Bagher-Ebadian^3,4, Wilson B. Chwang³, James R. Ewing^1,4, Ashley VanSlooten⁵, Lonni Schultz⁵, Ali S. Arbab³, and Rajan Jain^6
1Department of Neurology, Henry Ford Hospital, Detroit, MI, United States, ²Department of Biomedical Engineering, University of Michigan, Ann Arbor, MI, United States, ³Department of Radiology, Henry Ford Hospital, Detroit, MI, United States, ⁴Department of Physics, Oakland University, Rochester, MI, United States, ⁵Department of Biostatistics, Henry Ford Hospital, Detroit, MI, United States, ⁶Department of Radiology, NYU Langone Medical Center, New York, NY, United States

We have compared the results of using a Standard Radiological Arterial Input Function (SRAIF) vs. the Manually Selected Arterial Input Function (MSAIF) for estimation of permeability parameters in Dynamic Contrast Enhanced (DCE)-MR images of rats using a nested model selection approach. DCE experiments were performed using Gd-DTPA and Gadofosveset contrast agents which have different molecular sizes. The results show that when estimation in done using the two AIFs, there is less variation in the estimated value of the forward vascular transfer constant (Ktrans) compared to the estimated values of the extracellular-extravascular space volume (ve) and plasma volume (vp).

Greg O Cron^1,2, Steven Sourbron³, Daniel P Barboriak⁴, Rhys Abdeen², Matthew Hogan^2,5, and Thanh B Nguyen^1,2
1Medical Imaging, Ottawa Hospital Research Institute, Ottawa, Ontario, Canada, ²University of Ottawa, Ottawa, Ontario, Canada, ³Institute of Genetics, Health, and Therapeutics, University of Leeds, Leeds, United Kingdom, ⁴Radiology, Duke University Medical Center, Durham, North Carolina, United States,⁵Neuroscience, Ottawa Hospital Research Institute, Ottawa, Ontario, Canada

Software packages (SPs) for analysis of DCE-MRI data abound. However, it is unclear how consistent they are with each other, which impedes standardization. The purpose of this work was to compare three DCE-MRI SPs, using identical simulated data, to see if the SPs have different bias and precision. We observed differences between the SPs, especially in terms of how they react to noise. These results support the notion that the bias and precision of a DCE-MRI analysis may be affected by the particular SP used. In conclusion, only a single SP should be used for a given DCE-MRI study.

Ashish A Rao¹, Dattesh D Shanbhag², Venkata V Chebrolu², Sandeep N Gupta³, and Rakesh Mullick^4
1Biomedical Signal Analysis Laboratory, GE Global Research, Bangalore, Karnataka, India, ²Medical Image Analysis Laboratory, GE Global Research, Bangalore, Karnataka, India, ³Clinical Systems and Signal Processing, GE Global Research, Niskayuna, NY, United States, ⁴Diagnostics and Biomedical Technologies, GE Global Research, Bangalore, Karnataka, India

We describe ICA analysis on DCE-MRI concentration data in prostate and its utility in separating the blood volume and permeability components and removing artifacts, thereby improving fidelity of pK maps. The method can also be used for appropriate pK model selection and gain a better understanding of tumor response to therapy.

Gilad Liberman^1,2, Guy Nadav^1,3, Yoram Louzoun^2,4, Moran Artzi^1,5, and Dafna Ben Bashat^1,5
1Functional Brain Center, Tel Aviv Sourasky Medical Center, Tel Aviv, Israel, ²Gonda Multidisciplinary Brain Research Center, Bar Ilan University, Ramat Gan, Israel, ³School of Electrical Engineering, Tel Aviv University, Tel Aviv, Israel, ⁴Mathematics Department, Bar Ilan University, Ramat Gan, Israel, ⁵Sackler Faculty of Medicine and Sagol School of Neuroscience, Tel Aviv University, Tel Aviv, Israel

DCE is commonly acquired in a temporal-resolution of ~6sec, lower than inter-tissue differences (~1-3sec). This study proposes a super temporal-resolution analysis method for DCE, based on the Extended-Tofts Model (ETM) by including a delay accounting for the bolus arrival-time (BAT). The method was (1) tested on simulations (2) tested on data from patients with brain tumors, acquired with high(~2sec) and standard temporal-resolution and (3) BAT values were compared with DSC-TTP values. This study shows that accounting for the delay in DCE analysis, can result in a better AIF estimation and provides additional important parameter relating to the tissue vasculature.

Xin Li¹, Yu Cai¹, Brendan Moloney¹, Wei Huang¹, Mark G. Garzotto^2,3, Mark Woods¹, and Charles S Springer^1
1Advanced Imaging Research Center, Oregon Health & Science University, Portland, Oregon, United States, ²Portland VA Medical Center, Portland, Oregon, United States, ³Urology, Oregon Health & Science University, Portland, Oregon, United States

In Dynamic Contrast Enhanced Magnetic Resonance Imaging (DCE-MRI), contrast reagent (CR) concentration is linearly related to water R1 change for a single compartment. The longitudinal relaxivity quantifies the CR potency to change water proton T1. It is current practice to assume that the longitudinal relaxivities are the same in blood plasma and all interstitial compartments. However, there is evidence suggesting a potential increase in the interstitium longitudinal relaxivity. Based on human prostate data, we demonstrate the implications of unequal interstitium and plasma relaxivity values on DCE-MRI pharmacokinetic parameters.

Vandoorne Katrien¹, Moriel H Vandsburger², Hagit Dafni³, Michal Neeman⁴, Gustav J Strijkers¹, and Klaas Nicolay^1
1Biomedical NMR, Biomedical Engineering, Eindhoven University of Technology, Eindhoven, Netherlands, ²Physiology, University of Kentucky, Lexington, Kentucky, United States, ³Veterinary Resources, Weizmann Institute of Science, Rehovot, Israel, ⁴Biological Regulation, Weizmann Institute of Science, Rehovot, Israel

This pilot study aimed to develop a method to quantify the dynamics of a blood pool contrast agent for mapping the myocardial microvascular density and the rate of extravasation from blood vessels at the myocardium using in vivo dynamic contrast enhanced (DCE) MRI. Therefore, we quantify the characteristics and dynamics of the macromolecular contrast agent in blood vessels of control murine myocardium. In the future, this method could be used to quantify permeability in diseased myocardium. Therefore, this method using high-molecular weight contrast agent might allow quantification of permeability in the future and has the potential to determine disease progression.

Won-Joon Do¹, Dong-Wook Lee¹, Paul Kyu Han¹, Seung Hong Choi², Jong Chul Ye¹, and Sung-Hong Park^1
1Department of Bio and Brain Engineering, Korean Advanced Institute of Science and Technology, Daejeon, Korea, ²Seoul National University Hospital, Seoul, Korea

We proposed and demonstrated a new approach for acquiring 3D T1 and T2* weighted images simultaneously and dynamically using a dual-echo sequence combined with a compressed sensing (CS) reconstruction, k-t FOCUSS. The conflicting aspect of T1 and T2* weighting could be satisfied by using a special K-space sub-sampled reordering scheme. The implemented CS scheme on the dual-echo imaging with downsampling factor of 4 efficiently improved temporal resolution of the 3D dual-echo imaging, confirmed by detection of CSF signal changes. The technique can potentially be applied for simultaneous DCE and DSC imaging, which can significantly improve clinical information and patient comfort.

Jeiran Jahani¹, Glyn Johnson¹, Valerij G Kiselev², and Dmitry S Novikov^1
1Bernard and Irene Schwartz Center for Biomedical Imaging, Department of Radiology, New York University School of Medicine, New York City, New York, United States, ²Department of Radiology, Medical Physics, University Medical Center Freiburg, Freiburg, Germany

Quantitative perfusion analysis using bolus passage is compromised by tracer recirculation. As an alternative to the commonly used first bolus selection, we focus on relating the recirculation boluses in the brain to the topology and quantitative properties of the circulation in the whole organism. The classical central volume principle is generalized for the recirculation and the presence of contrast elimination from the blood pool. Our analytical approach makes use of treating each organ’s response in the frequency domain, similar to complex admittance in linear circuits. The simplest among realistic models of circulation is selected according to the data fitting quality.

Karsten Sommer¹, Regine Schmidt¹, Hanns-Christian Breit¹, Dirk Graafen¹, and Laura M. Schreiber^1
1Department of Diagnostic and Interventional Radiology, University Medical Center Mainz, Mainz, Rhineland-Palatinate, Germany

In this contribution we present the CFD simulation of contrast agent transport in a high detailed 3D model of the cerebral arteries. We used the results of the simulations to determine the errors induced by contrast agent dispersion in cerebral blood flow quantification. Contrary to recent studies, our results do not show substantial contrast agent bolus dispersion. Dispersion-related cerebral blood flow errors vary between the different outlets of the 3D model, but are on the whole much smaller than previously predicted.

William D. Rooney¹, Xin Li¹, John W. Grinstead², Edward A. Neuwelt³, and Charles S. Springer, Jr.^1
1Advanced Imaging Research Center, Oregon Health & Science University, Portland, Oregon, United States, ²Siemens Healthcare, Portland, Oregon, United States, ³Blood-Brain-Barrier Program, Oregon Health & Science University, Portland, Oregon, United States

Monomeric Gd(III) chelate contrast agents (CA) do not extravasate in normal brain. In that case, shutter-speed pharmacokinetic analysis of DCE-MRI data allows mapping of the mean capillary water lifetime, tau_b. The magnitude of this biomarker is inversely proportional to neuronal Na⁺/K⁺ATPase [NKA] activity. However, in glioblastoma multiforme brain tumors these CAs extravasate rapidly. Here, we use an intravascular SPIO CA, Ferumoxytol, which does not extravasate during the study. It reveals that tumor tau_b is dramatically elevated: NKA activity is drastically reduced.

M. Ethan MacDonald^1,2, Estee Lee^1,2, Ting Lee^2,3, Jordan Woehr⁴, Chris d'Esterre^2,5, Michael R Smith^4,5, and Richard Frayne^2,5
1Biomedical Engineering, University of Calgary, Calgary, AB, Canada, ²Seaman Family Magnetic Resonance Research Centre, Hotchkiss Brain Institute, Foothills Medical Centre, Calgary, AB, Canada, ³Imaging Research Labs, Robarts Research Institute, London, ON, Canada, ⁴Electrical & Computer Engineering, University of Calgary, Calgary, AB, Canada, ⁵Radiology & Clinical Neurosciences, University of Calgary, Calgary, AB, Canada

In this work we use dynamic susceptibility contrast MR imaging and fit a duel compartmental model to the residue function. The proposed duel compartmental model has been used in CT perfusion with great success, consisting of a rectangular function and exponential decay. The model is interactively fit to the residue function with an iterative least squares conjugate gradient algorithm. Parametric maps and goodness of fit maps were produced for an acute ischemic stroke patient. Fit quality is high in regions of normal flow, but where flow is low, <10 ml/100 g/min, the quality of the fit is diminished.

Yikang Liu¹, Weiping Ding¹, and Bensheng Qiu^1
1Department of Electronic Science and Technology, University of Science and Technology of China, Hefei, Anhui, China

The unclear T1 and T2* effect dominance caused by contrast agent (CA) leakage makes the explanation of R2* curve difficult. However, through a numeric simulation where cells and vessels were randomly distributed while EES fraction and vascular permeability controlled, a linear relationship between EES fraction and ratio of the two parameters K1 and K2 in Weisskoff model has been demonstrated, which on the other hand varied little with vascular permeability. As a consequence, a potential approach to measure relative Ve by DSC-MRI is feasible and clinical data is expected to prove the theory.

Bo Xu^1,2, Pascal Spincemaille², Tian Xu³, Martin Prince², Silvina Dutruel², Ajay Gupta², Nandadeepa Thimmappa², and Yi Wang^1,2
1Cornell University, New York, NY, United States, ²Weill Cornell Medical College, NY, United States, ³MedImageMetric LLC, NY, United States

Quantitative susceptibility mapping (QSM) has the potential to quantify contrast agent bio-distribution. The purpose of this study is to develop a QSM technique with sufficient temporal resolution for mapping first pass contrast agent concentration in cerebral perfusion imaging.Initial in-vivo brain imaging data demonstrated the feasibility of mapping time resolved contrast agent quantification using the proposed method. Cerebral blood flow agrees with that obtained from arterial spin labeling technique.