Sheng Fang¹, Wenchuan Wu², and Hua Guo^2
1Institute of nuclear and new energy technology, Tsinghua University, Beijing, China, ²Center for Biomedical Imaging Research, Department of Biomedical Engineering, School of Medicine, Tsinghua University, Beijing, China

A highly accelerated parallel imaging technique based on variable density spiral (VDS) acquisition and spatial adaptive CORNOL reconstruction is proposed. With an optimized undersampling and a reconstruction tailored for VDS, the proposed method can make the full use of MR image features by exploiting both the incoherence of aliasing artifacts and the coherence of image structures. The simulation and in vivo VDS experiments results demonstrate that this method can achieve a very high reduction factor while maintaining high SNR and well-preserved image structure

Florian Knoll¹, Gerrit Schultz², Kristian Bredies³, Daniel Gallichan², Maxim Zaitsev², Jürgen Hennig², and Rudolf Stollberger^1
1Institute of Medical Engineering, Graz University of Technology, Graz, Styria, Austria, ²Dept. of Radiology, Medical Physics, University Medical Center Freiburg, Freiburg, Germany, ³Institute of Mathematics and Scientific Computing, University of Graz, Graz, Austria

This work presents a new iterative reconstruction method for undersampled radial PatLoc imaging based on the Total Generalized Variation (TGV). Results from numerical simulations and in-vivo PatLoc measurements with as few as 16 radial projections for the reconstruction of a 256x256 matrix are presented, which demonstrate significant improvements in image quality.

Michael Abram Ohliger¹, Simon Hu¹, Peder E. Z. Larson¹, Robert Bok¹, Peter Shin¹, James Tropp², Lucas Carvajal¹, Sarah J. Nelson¹, John Kurhanewicz¹, and Daniel Vigneron^1
1Radiology and Biomedical Imaging, University of California San Francisco, San Francisco, CA, United States, ²General Electric Healthcare, Fremont, CA, United States

Autocallibrated parallel MR strategies generally rely on a densely sampled center of k-space to obtain coil array reference information. This dense sampling can be a large burden when applied to spectroscopic imaging. We present a new strategy for MR spectroscopic imaging with parallel reconstruction that eliminates the need for extra auto-calibrating lines. Combining spatial and spectral undersampling, select chemical species alias to unused portions of the spectrum, and can be used as sensitivity references. For an undersampling factor of 3, 38% fewer phase encode lines are acquired. Theoretical background, numerical simulations, and in vivo examples are presented using hyperpolarized 13C.

Yu Li¹, Feng Huang², Wei Lin², Randy Duensing², and Charles L. Dumoulin^1
1Imaging Research Center, Radiology Department, Cincinnati Children's Hospital Medical Center, Cincinnati, Ohio, United States, ²Invivo Diagnostic Imaging, Philips HealthCare, Gainesville, Florida, United States

Parallel imaging provides a generic solution to accelerating single-scan MRI. However, a clinical MRI protocol needs a series of MRI scans for acquiring a number of images with different contrast and geometry. In the presented work, we propose a new high-speed MRI framework, correlation imaging, that can accelerate multi-scan MRI by combining multi-channel acceleration mechanisms underlying parallel imaging and the shared information of a multi-scan acquisition. In a three-scan anatomical MRI protocol, we demonstrated that correlation imaging optimized for multi-scan MRI outperforms conventional parallel imaging techniques optimized for single-scan MRI.

Jerome Yerly^1,2, Michel Louis Lauzon^2,3, and Richard Frayne^2,3
1Electrical and Computer Engineering, University of Calgary, Calgary, AB, Canada, ²Seaman Family MR Research Centre, Foothills Medical Centre, Calgary, AB, Canada, ³Departments of Radiology and Clinical Neurosciences, University of Calgary, Calgary, AB, Canada

Recent advances in image reconstruction from sparsely sampled k-space data, such as CS and parallel MR imaging, provide potential solutions to enable visualization of small cerebral vessels without increasing the total acquisition time. In this study, we investigated and compared SENSE, GRAPPA, CS, SPIRiT, CS-SENSE, and L1-SPIRiT techniques to accelerate time-of-flight 3-T MR imaging. The reconstructions involving an L1-norm regularization procedure (i.e., CS, CS-SENSE, and L1-SPIRiT) resulted in lower aliasing interference, but also less conspicuous small cerebral vessels due to blurring. The auto-calibrating techniques (i.e., GRAPPA, SPIRiT, and L1-SPIRiT) exhibited less sensitivity artifacts and most reliably depicted small vessels.

Kawin Setsompop^1,2, and Lawrence L Wald^1,3
1Dept. of Radiology, A. A. Martinos Center for Biomedical Imaging, MGH, Charlestown, Massachusetts, United States, ²Harvard Medical School, Boston, MA, United States, ³Harvard-MIT Division of Health Sciences and Technology, Cambridge, Massachusetts, United States

Simultaneous multi-slice EPI using parallel imaging allows for an increase in sampling rate of fMRI and reduced acquisition time in diffusion imaging. Nonetheless, the unaliasing problem is difficult for closely spaced simultaneously excited slices. A blipped-CAIPI acquisition scheme was introduced to mitigate this issue by creating an inter-slice image shift to increase the distance between aliased voxels. In this work, we show that spatially varying field inhomogeneities and eddy currents can cause significant aliasing artifact in the form of inter-slice ghost leakage which is particularly problematic for blipped-CAIPI acquisitions. A multi-kernel slice-GRAPPA reconstruction technique is proposed to mitigate this issue.

Yuchou Chang¹, and Leslie Ying^1
1Electrical Engineering, University of Wisconsin - Milwaukee, Milwaukee, Wisconsin, United States

IIR GRAPPA incorporates recursive terms to improve conventional GRAPPA, but has the limitation that outliers and noise lead to poor estimation in the recursive coefficients. A novel method using nonlinear ARMA (NLARMA) model is proposed to address the issue in IIR GRAPPA reconstruction. The proposed nonlinear AMRA model improves over the linear MA model used in conventional GRAPPA by incorporating both recursion and nonlinearity. The experimental results demonstrate that the proposed method is able to significantly improve the reconstruction quality of the conventional GRAPPA and IIR GRAPPA in suppressing noise and artifacts.

Yu Ding¹, Hui Xue², Ti-chiun Chang³, Christoph Guetter³, and Orlando Simonetti^1
1Dorothy M. Davis Heart and Lung Research Institute, The Ohio State University, Columbus, OH, United States, ²siemens corporate research, ³Siemens Corporate Research

Several new algorithms have been proposed to take advantage of k-space correlations, such as SPIRiT and PRUNO. However, these methods have no closed-form solutions, and can only be solved using computationally-intensive iterative methods. We propose a new k-space based pMRI technique, self-consistent GRAPPA by including an extra set of linear equations utilizing the intrinsic correlation between skipped k-space points. SC-GRAPPA combines the linear equations of traditional GRAPPA with these additional equations to solve for the missing k-space data. SC-GRAPPA utilizes a least-square solution of the linear equations, and therefore has a closed-form solution without any free parameters.

Yu Ding¹, Hui Xue², Ti-chiun Chang³, Christoph Guetter³, and Orlando Simonetti^1
1Dorothy M. Davis Heart and Lung Research Institute, The Ohio State University, Columbus, OH, United States, ²siemens corporate research, ³Siemens Corporate Research

GRAPPA reconstructs the missing k-space by applying a convolution kernel which is estimated from ACS lines using linear regression. Intuitively, ACS lines with higher SNR should boost the accuracy of the kernel estimation, and increase the SNR of GRAPPA reconstruction. Paradoxically, Sodickson and his colleagues pointed out that the higher SNR in ACS lines may lead to lower SNR in GRAPPA reconstruction. We study quantitative about how the noise in the ACS lines affects the SNR of the GRAPPA reconstruction, and proposes a simple solution to improve the SNR of TGRAPPA.

Johannes M Peeters¹, and Miha Fuderer^1
1Philips Healthcare, Best, Netherlands

In this work, an extension of the SENSE parallel imaging framework is proposed. In this new framework called Minimal Artifact Factor SENSE (MAF-SENSE), also artifact probability due to incorrect knowledge of the receiver coil sensitivities is taken into account. The latter is realized by adding an uncertainty in this knowledge in order to enable weighting of residual artifact level and SNR in the inversion problem. Results show that MAF-SENSE is well capable of removing residual aliasing artifacts caused by usage of incorrect receiver coil sensitivities.

Daniel S Weller¹, Jonathan R Polimeni^2,3, Leo Grady⁴, Lawrence L Wald^2,3, Elfar Adalsteinsson¹, and Vivek Goyal^1
1EECS, Massachusetts Institute of Technology, Cambridge, MA, United States, ²A. A. Martinos Center, Dept. of Radiology, Massachusetts General Hospital, Charlestown, MA, United States, ³Dept. of Radiology, Harvard Medical School, Boston, MA, United States, ⁴Dept. of Image Analytics and Informatics, Siemens Corporate Research, Princeton, NJ, United States

When applying GRAPPA at high accelerations, it is not always feasible to acquire sufficiently many auto-calibration signal (ACS) lines to properly calibrate the interpolation kernels. The proposed calibration method employs regularization promoting joint sparsity of the coil images that would be reconstructed. This method improves reconstruction quality and increases the total acceleration that is achievable with GRAPPA.

Nolwenn Caillet¹, Jason Stockmann², Leo Tam², Gigi Galiana¹, and R. Todd Constable^1
1Diagnostic Radiology, Yale University, New Haven, CT, United States, ²Biomedical Engineering, Yale University, New Haven, CT, United States

With a traditional cylindrical strip array coil, one cannot accelerate a 3D pulse sequence along z since all the coils present the same sensitivity profile in this direction. To overcome this limitation, we propose to twist the striplines along the substrate, such that the aliased data for a given coil element do not overlap in z in undersampled acquisitions. This permits use of SENSE or GRAPPA in two phase-encoding directions, achieving higher acceleration factors without increasing the number of receiving channels.

Kay Nehrke¹, and Peter Börnert^1
1Philips Research Laboratories, Hamburg, Germany

In the present work a new B1 mapping approach dubbed DREAM (Dual Refocusing Echo Acquisition Mode) is proposed for free-breathing abdominal B1 mapping. The DREAM approach acquires a 2D B1 map in a fraction of a second, making it possible to freeze respiratory motion efficiently. The approach is used to study dynamical changes of B1 in the liver during free breathing at 3T.

Yulin V Chang^1,2, and Philip V Bayly^2
1Radiology, Washington University, Saint Louis, MO, United States, ²Mechanical Engineering and Materials Science, Washington University, St. Louis, MO, United States

Most of the phase-based flip angle (FA) or B₁ mapping methods are sensitive to the background field (B₀) inhomogeneity. In this work we focus on the recently proposed orthogonal- method and investigate the effect of B₀ inhomogeneity on the FA maps. We demonstrate that by acquiring an addition image with only a single RF pulse the FA-map quality can be maintained in the presence of B₀ inhomogeneity, provided that the inter-pulse delay is short compared to T₂^*.

Alexis Amadon¹, Martijn Anton Cloos¹, Nicolas Boulant¹, Marie-France Hang¹, Christopher John Wiggins¹, and Hans-Peter Fautz^2
1I2BM/Neurospin/LRMN, CEA, Gif-sur-Yvette, France, ²Siemens Healthcare, Erlangen, Germany

Fast rather than accurate B1-mapping may be seeked for Tx-array calibration. We evaluate a 2D technique which is 6 times faster than a gold standard method. We compare the Flip Angle maps issued from tailored pulses designed with both calibration methods. These pulses target whole human brain excitation homogenization at 7T. Their outcome is very similar, which shows the benefit of using the faster sequence.

Alessandro Sbrizzi¹, Hans Hoogduin², Jan J Lagendijk², Peter R Luijten², and Cornelis A van den Berg^3
1Imaging Division, UMC Utrecht, Utrecht, Utrecht, Netherlands, ²UMC Utrecht, ³UMC Utrecht, Netherlands

In this work, a model based reconstruction method for accelerated 3D B1+ mapping is illustrated. Only a limited number of transverse slices is needed as input for the reconstruction problem, reducing the scan time up to a factor 5. Positive results from simulations and in vivo measurements confirm the validity of the method.

Narae Choi¹, Joonsung Lee¹, and Dong-Hyun Kim^1
1Electrical and Electronic Engineering, Yonsei University, Sinchon dong, Seoul, Korea

The measurement of the transmitted radiofrequency (RF) field is useful for many MR applications. Conventionally, the amplitude of the B1 was measured by B1 mapping methods and the phase of the active magnetic RF field component H+ was estimated by using spin echo (SE) sequence. Using SE sequence has one drawback which is long scan time due to relatively long TR compared to GRE sequence. This study focuses on the estimation of B1 magnitude and phase simultaneously using much faster acquisition, double TR multi-echo gradient echo sequence. This approach is validated in phantom and in-vivo experiments.

Mohammad Mehdi Khalighi¹, Michael Zeineh², and Brian K Rutt^2
1Global Applied Science Lab, GE Healthcare, Menlo Park, California, United States, ²Radiology Deaprtment, Stanford University, Stanford, California, United States

High field MRI image signal is distorted by receive sensitivity variations. With the added complication of highly non-uniform transmit (B₁⁺) at high field, receive sensitivity correction becomes even more challenging. Here we used Bloch-Siegert (B-S) B₁⁺ mapping method to measure B₁⁺ field and thus transmit sensitivity. Then the B-S magnitude images were used to estimate the receive sensitivity map for each individual channel over the whole brain, which was used to correct subsequent structural brain images. These receive sensitivity maps could be also used in parallel imaging reconstruction such as ASSET or SENSE.

Marcin Jankiewicz^1,2, John C. Gore^1,2, and William A. Grissom^1,3
1Institute of Imaging Science, Vanderbilt University, Nashville, TN, United States, ²Radiology and Radiological Sciences, Vanderbilt University, Nashville, TN, United States, ³Biomedical Engineering, Vanderbilt University, Nashville, TN, United States

We propose a novel way of designing saturation pulses for Bloch-Siegert B₁⁺ mapping, which directly maximizes Bloch-Siegert sensitivity subject to a constraint requiring negligible on-resonant excitation.

Hans-Peter Fautz¹, Rene Gumbrecht¹, Patrick Gross¹, and Franz Schmitt^1
1Siemens Healthcare, Erlangen, Germany

The typically large dynamic range of B1 fields generated by local transmit array coils are a limiting factor for fast and robust B1 mapping. This works extends the sensitivity range of quantitative B1 mapping techniques towards low B1 amplitudes incorporating the information gained from relative B1 maps into the B1 map reconstruction.

Volker Jörg Friedrich Sturm¹, Thomas Christian Basse-Lüsebrink¹, Thomas Kampf¹, and Peter Michael Jakob^1
1Experimental Physics 5, University of Würzburg, Würzburg, Germany

A new method for phase-based B₁⁺ mapping based on the Bloch-Siegert (BS) shift (ϕ_BS) was presented in the past. Being a phase-based method, the maximal ϕ_BS can be restricted since phase wraps cannot easily be removed for all given situations. This can be critical for BS B₁ mapping using surface coils. Thus, to also obtain high quality B₁ data in insensitive areas phase wraps in more sensitive areas cannot always be prevented. To reduce this problematic, a method is proposed which allows variation of ϕ_BS in a single echo train. Consequently, by combining the information of the various obtained ϕ_BS it is possible to remove phase wraps.

Alexander Gotschy^1,2, Uvo C. Hölscher³, Thomas C. Basse-Lüsebrink^1,4, André Fischer⁵, Morwan Choli³, Thomas Kampf¹, Volker Sturm¹, Daniel Neumann³, Volker Herold¹, Herbert Köstler⁵, Dietbert Hahn⁵, Guido Stoll⁴, Wolfgang R. Bauer², and Peter M. Jakob^1,3
1Department of Experimental Physics 5, University of Würzburg, Würzburg, Germany, ²Department of Internal Medicine I, University of Würzburg, Würzburg, Germany, ³Research Center Magnetic Resonance Bavaria (MRB), Würzburg, Germany, ⁴Department of Neurology, University of Würzburg, Würzburg, Germany, ⁵Institute of Radiology, University of Würzburg, Würzburg, Germany

An accurate knowledge of the B1+-field distribution becomes increasingly important for many MR applications like B1+ shimming, Spatially Selective Excitation algorithms or guaranteeing patient safety at ultrahigh fields. To meet the requirements of high field applications we used the lately introduced Bloch-Siegert-Shift based B1+-mapping method in a BURST sequence. The new sequence combines robustness against T2* effects with reduced SAR. In the presented study we found that a centric encoded BS-BURST sequence allows highly accelerated B1+-mapping with superior quality. The B1+ map of a whole head could be acquired in 36s which is adequate for applications in clinical routine.

Galen D. Reed¹, Peder E.Z. Larson¹, James Tropp², Albert P. Chen³, Adam B. Kerr⁴, Mark van Criekinge¹, Douglas A. C. Kelley⁵, John M. Pauly⁴, Kayvan R. Keshari¹, John Kurhanewicz¹, Sarah J. Nelson¹, and Daniel B. Vigneron^1
1Radiology and Biomedical Imaging, University of California San Francisco, San Francisco, California, United States, ²GE Healthcare, Fremont, California, United States, ³GE Healthcare, Toronto, Ontario, Canada, ⁴Electrical Engineering, Stanford University, Stanford, California, United States, ⁵GE Healthcare, San Francisco, California, United States

With hyperpolarized ¹³C magnetic resonance imaging in the early stages of clinical translation, methods of characterizing transmitter inhomogeneity and calibration of transmit gain become increasingly important. This study presents double angle spectra acquired from a hyperpolarized ¹³C scan of a human prostate as well as a field mapping using the Bloch Siegert shift with ethylene glycol phantoms for the characterization of transmitter nonuniformity. This nonuniformity was incorporated into the patient study as a 2 dB attenuation adjustment, and the efficacy of this approach was validated in the in vivo patient exam.

Thomas Christian Basse-Lüsebrink¹, Volker Jörg Friedrich Sturm¹, Anna Vilter¹, Thomas Kampf¹, Volker Christian Behr¹, and Peter Michael Jakob^1
1Experimental Physics 5, University of Würzburg, Würzburg, Bavaria, Germany

Due to ¹⁹F properties such as high sensitivity, unambiguous localization of labeled cells and direct quantification, the MR community has regained great interest in ¹⁹F MRI. Surface coils, however, are often used due to the normally low SNR in ¹⁹F images. Since an inhomogeneous B₁ profile is inherent with surface coils quantification of the ¹⁹F signal is hampered and thus strategies to map the ¹⁹F B₁ profile are of great interest. The present study investigates whether or not fast ¹H CPMG-based Bloch-Siegert B₁ mapping of a double-resonant surface coil can enable assessment of the ¹⁹F B₁ profile.

Weitian Chen¹, Mai LH Nguyen², and Garry E Gold^2
1Global MR Applied Science Laboratory, GE Healthcare, Menlo Park, CA, United States, ²Radiology, Stanford University, Palo Alto, CA, United States

Recently multi-exponential relaxation analysis was reported to have improved specificity of matrix assessment in cartilage. However, quantification of more than one component of T1rho or T2 significantly increases scan time and is impractical in clinical applications. We investigate mono-exponential relaxation methods to quantify relaxation property of a single relaxation component by suppressing signal from other components. Such methods are feasible in terms of clinical scan time, and have potential to achieve improved specificity to matrix component in cartilage compared to conventional mono-exponential relaxation analysis.

Marshall Stephen Sussman¹, and Walter Kucharczyk^2
1Medical Imaging, University Health Network, Toronto, Ontario, Canada, ²Medical Imaging, University of Toronto, Toronto, Ontario, Canada

Knowledge of a tissue’s full T₂ spectrum may provide additional information on tissue pathology beyond current clinical techniques. Unfortunately, methods used to measure full T₂ spectra require impractically long scan times. As a possible alternative, linear combination filtering (LCF) provides information on a specified region of the T₂ spectrum in clinically reasonable scan times. However, the effectiveness of LCF depends on the specific filter design. The purpose of this project is to optimize the LCF filter parameters for the myelin component of the white matter T₂ spectrum.

Anup Singh¹, Mohammad Haris¹, Kejia Cai¹, Victor Babu¹, Feliks Kogan¹, Hari Hariharan¹, and Ravinder Reddy^1
1Radiology, University of Pennsylvania, Philadelphia, Pennsylvania, United States

The sensitivity of gagCEST in human knee cartilage is evaluated at 3T and 7T field strengths. Calculated gagCEST values without accounting for B0 inhomogeneity (~0.6 ppm) were > 20%. After B0 inhomogeneity correction, calculated gagCEST values were negligible at 3T and ~6% at 7T. Results obtained with varying saturation pulse durations and amplitudes as well as the consistency between numerical simulations and our experimental results indicate that the negligible gagCEST observed at 3T is due to direct saturation effects and fast exchange rate. At high fields such as 7T, this method holds promise as a viable clinical technique.

Adrienne N Dula¹, Pooja Gaur^1,2, Catherine E Frame^1,3, Jane Anne Trapp Hirtle¹, Blake E Dewey¹, Richard D Dortch¹, and Seth A Smith^1
1Vanderbilt University Institute of Imaging Science, Vanderbilt University Medical Center, Nashville, TN, United States, ²Chemical and Physical Biology, Vanderbilt University Medical Center, Nashville, TN, United States, ³Radiation Oncology, Vanderbilt University Medical Center

We apply CEST to healthy cervical spinal cord at 7T and compare the results with our previous studies at 3T. We hypothesize that development of spinal cord CEST could reveal the health of the spinal cord and its relationship to disease earlier than conventional methods. Results include calculation of the amide proton asymmetry effect by integrating the area between a Lorentzian fit and the acquired CEST spectra. APT asymmetry maps at 7T reveal white/gray matter delineation but the mean values of 5 healthy controls were not significant.

Kimberly L Desmond¹, and Greg J Stanisz^1,2
1Medical Biophysics, University of Toronto, Toronto, Ontario, Canada, ²Imaging Research, Sunnybrook Research Institute, Toronto, Ontario, Canada

The nature of the endogenous CEST Z-spectrum was examined at 7T in six anatomical regions / tissue types in mice: muscle (in vivo and ex vivo), grey matter, white matter, adipose, bladder and tumour. CEST images were acquired for a range of offset frequencies from -2000 to 2000 Hz (6.71ppm) and the shape of the spectrum was observed as a function of amide proton exchange, hydroxyl proton exchange and the presence of mobile lipids.

Guang Jia¹, Wenbo Wei¹, Xiangyu Yang¹, David C Flanigan², Jochen Keupp³, Jinyuan Zhou⁴, and Michael V Knopp^1
1Department of Radiology, The Ohio State University, Columbus, OH, United States, ²Department of Orthopedics, The Ohio State University, Columbus, OH, United States, ³Innovative Technologies, Research Laboratories, Philips Research Europe, Hamburg, Germany, ⁴Department of Radiology, Johns Hopkins University, Baltimore, MD, United States

Amide-proton-transfer MRI (APT-MRI) has recently emerged as a new protein-based molecular imaging technique. In fat-containing tissue, one of the challenges is that fat gives signal drop at -3.5ppm deteriorating the calculation of asymmetric magnetization transfer ratio (MTR) at 3.5 ppm in APT-MRI, which cannot be resolved using a regular fat-suppression method. This study is to evaluate a multi gradient-echo DIXON (mDIXON)-based APT-MRI in bone marrow edema. The proposed technique has the potential to be applicable to fat-containing tissue tumors, such as osteosarcoma, breast cancer, and fatty liver lesion.

Leanne Katherine Bell¹, Davina Honess¹, Dominick McIntyre¹, David Tuveson², and John Griffiths^1
1Magnetic Resonance Imaging and Spectroscopy, Cambridge Research Institute, Cambridge, Cambridgeshire, United Kingdom, ²Tumour Modelling and Experimental Medicine (Pancreatic Cancer), Cambridge Research Institute, Cambridge, Cambridgeshire, United Kingdom

Pancreatic Ductal Adenocarcinoma (PDA) is a lethal disease with minimally effective treatments available. The genetically engineered KPC mouse develops spontaneous PDA in situ, recapitulating the genetic, molecular and pathological aspects of the human disease, including its dense desmoplastic tumour stroma. In contrast, transplanted mouse PDA has minimal stroma. Comparing these tumour types shows that Magnetisation Transfer MRI, specifically Magnetisation Transfer Ratio (MTR), can distinguish between differing stromal contents. We conclude that MTR is an effective in vivo biomarker for assessing macromolecular content and preliminary data suggest it may be valuable in assessing tumour stromal changes in a therapeutic setting.

Kim Desmond¹, Firas Moosvi¹, and Greg J Stanisz^1,2
1Medical Biophysics, University of Toronto, Toronto, Ontario, Canada, ²Imaging Research, Sunnybrook Research Institute, Toronto, Ontario, Canada

Parametric maps of endogenous CEST Z-spectrum properties (amide, OH-, and mobile lipid peaks) were obtained in five mice with Lewis lung carcinoma xenografts at 7T. Maps were generated by decomposing the spectrum for each voxel into the sum of four Lorentzian functions of variable amplitude and width representing the major endogenous proton pools. The growth of the tumour was followed for five days, and an increase in mean peak amplitude was observed for each pool.

Adrienne N Dula^1,2, Lori R Arlinghaus^1,2, Richard D Dortch^1,2, Blake E Dewey¹, Jennifer G Whisenant¹, Dan Ayers³, Thomas E Yankeelov^1,2, and Seth A Smith^1,2
1Vanderbilt University Institute of Imaging Science, Vanderbilt Medical Center, Nashville, TN, United States, ²Radiology and Radiological Sciences, Vanderbilt Medical Center, Nashville, TN, United States, ³Biostatistics, Vanderbilt Medical Center, Nashville, TN, United States

The goal of this study is to determine the repeatability of CEST MRI applied to human fibroglandular (FG) of the breast at 3T. We were able to reliably produce maps of the CEST effect caused by the presence of amide protons in healthy FG tissue. The results from this study indicate that a change in this quantified amide proton transfer effect larger than 1.321 for an individual or 0.418 for a group of 10 patients would be statistically significant (= 0.05).

Kejia Cai¹, Anup Singh¹, Mohammad Haris¹, Ravi Prakash Reddy Nanga¹, Hari Hariharan¹, and Ravinder Reddy^1
1CMROI, Department of Radiology, University of Pennsylvania, Philadelphia, PA, United States

Glutamate (Glu) is the major excitatory neurotransmitters in the brain, and is likely involved in nearly all signal processing functions of the central nervous system (CNS). Based on Glu amine proton exchange saturation transfer (GluCEST) effect, recently, we demonstrated the feasibility of mapping relative changes in Glu concentration as well as pH in vivo. GluCEST provides markedly increased spatial and temporal resolution than ¹H-MRS. In the current study, brain regional variation of GluCEST is investigated and compared to existing MRS and positron emission tomography (PET) studies.

Kejia Cai¹, He N. Xu², Mohammad Haris¹, Anup Singh¹, Ravinder Reddy¹, and Lin Z. Li^2
1CMROI, Department of Radiology, University of Pennsylvania, Philadelphia, PA, United States, ²Dapartment of Radiology, University of Pennsylvania, Philadelphia, PA, United States

Predicting breast tumor metastatic potential remains a challenge in cancer research and clinical practice. Previous studies indicate that tumor redox ratio heterogeneity (rim-core difference) serves as a biomarker of tumor metastatic potential. In this study, we demonstrated that CEST MRI showed tumor core-rim heterogeneity correlating with the mitochondrial redox state whereas T₁ and T₂ weighted MRI did not. CEST MRI of breast tumor metabolites (especially Glu) may be used as a potential non-invasive imaging biomarker of mitochondrial redox state for predicting breast cancer metastatic potential.

Chien-Yuan Lin^1,2, Nirbhay N Yadav^2,3, S. James Ratnakar¹, A. Dean Sherry^1,4, and Peter C. M. van Zijl^2,3
1Advanced Imaging Resarch Center, University of Texas Southwestern Medical Center, Dallas, Texas, United States, ²F.M. Kirby Center for Functional Brain Imaging, Kennedy Krieger Institute, Baltimore, Maryland, United States, ³Division of MR Research, Russell H. Morgan Dept. of Radiology and Radiological Science, Johns Hopkins University, Baltimore, Maryland, United States, ⁴University of Texas at Dallas, Dallas, Texas, United States

Frequency labeled exchange transfer (FLEX) MRI is a pulse sequence that uses excitation instead of saturation to detect chemical exchange saturation transfer (CEST) agents. It has previously been shown to allow detection of rapidly exchanging water groups on paramagnetic (paraCEST) agents as well as separation of magnetization transfer (MT) effects from exchange effects using time domain analysis. In the present study, we implemented the FLEX method on a preclinical MRI system and show that FLEX imaging of paraCEST agents is possible. This should have potential for imaging the paraCEST agents in vivo where MT effects often obscure such agents due to the need for using high B1 for rapidly exchanging groups.

Marcus Björk¹, R. Reeve Ingle², Joëlle K. Barral³, Erik Gudmundson^4,5, Dwight G. Nishimura², and Petre Stoica^1
1Department of Information Technology, Uppsala University, Uppsala, Sweden, ²Electrical Engineering, Stanford University, Stanford, California, United States, ³HeartVista, Inc., Los Altos, California, United States, ⁴Centre for Mathematical Sciences, Lund University, Lund, Sweden, ⁵Signal Processing Lab, ACCESS Linnaeus Center, KTH – Royal Institute of Technology, Stockholm, Sweden

Banding artifacts are a major issue in bSSFP. Several images with equally spaced phase increments are typically acquired to mitigate this issue. However, to the best of our knowledge, no theoretical justification exists for such a distribution of phase increments. We derived the Cramér-Rao bound (CRB) from the bSSFP signal model and optimized the phase increments distribution to minimize the maximum CRB over all possible off-resonances. We show that the worst-case CRB is minimized when using equally spaced phases, and that the resulting performance is relatively close to the optimal, had the off-resonance been known a priori.

Sarah Eskreis-Winkler^1,2, Pascal Spincemaille¹, Nanda Deepa Thimmappa¹, Martin R. Prince¹, and Yi Wang^1
1Weill Medical College of Cornell University, New York, NY, United States, ²New York Hospital Queens, Flushing, NY, United States

Sliding window 3D spiral contrast enhanced MRI of the liver has been shown to significantly increase the ability to visualize the arterial phase. However, spiral imaging is prone to off-resonance artifacts. In this abstract, a time-efficient approximate algorithm for off-resonance correction was used by acquiring a single off-resonance frequency for each coil and each slice. This did not increase acquisition and reconstruction time and significantly improved image quality (p<0.0001), particularly during the arterial phase.

Kristin L Granlund^1,2, Weitian Chen³, Dawei Gui⁴, Donglai Huo⁴, Patrick LeRoux⁵, and Yuval Zur^6
1Radiology, Stanford University, Stanford, California, United States, ²Electrical Engineering, Stanford University, Stanford, CA, United States, ³Applied Science Laboratory, GE Healthcare, Menlo Park, CA, United States, ⁴MR PSD and Applications, GE Healthcare, Waukesha, WI, United States, ⁵Applied Science Laboratory, GE Healthcare, Europe, ⁶GE Healthcare, Israel

The use of 3D FSE sequences is limited off-isocenter due to gradient non-idealities, such as gradient-induced eddy currents. By measuring the errors during prescan, we are able to adjust the imaging gradients to improve image quality. We use a phase cycling approach to measure the phase error due to the readout gradients and a z phase encoding approach to measure the phase error due to the slice encoding gradients. The correction removes banding artifacts for off-isocenter 3D FSE acquisitions in short prescan times.

Wei Lin¹, Feng Huang¹, Enrico Simonotto¹, and Randy Duensing^1
1Invivo Corporation, Philips Healthcare, Gainesville, Florida, United States

A rapid EPI distortion correction method is proposed by off-resonance artifacts correction with convolution in k-space (ORACLE). A basis kernel is first computed by k-space data fitting between two EPI datasets with slightly different echo times. A B0 map is then computed from the inverse Fourier Transform of this basis kernel, resulting in robust estimations even in low signal intensity regions. Then different convolution kernels are applied to different phase-encoding lines to correct for EPI distortion directly in the k-space. In vivo brain DWI study demonstrates excellent correction results. An additional advantage of ORACLE is the reusability of correction kernels for images acquired with the same sequence but different contrasts.

Weitian Chen¹, Lai Peng¹, Anja CS Brau¹, Mai LH Nguyen², and Garry E Gold^2
1Global MR Applied Science Laboratory, GE Healthcare, Menlo Park, CA, United States, ²Radiology, Stanford University

Fast spin echo acquisition (FSE) plays a central role in clinical MRI. A major limitation to this approach is that T2 decay during multiple echoes acquisition can result in image blurring. Methods based on flip angle modulation have been reported to address this problem so that long echo train can be used to achieve high SNR efficiency in 3D imaging without excessive image blurring. In this work, we reported an efficient approach to further reduce the blurring in such advanced 3D FSE imaging methods based on deconvolution of k-space.

Martin Krämer¹, and Jürgen R Reichenbach^1
1Medical Physics Group, Department of Diagnostic and Interventional Radiology I, Jena University Hospital, Jena, Germany

Shifts of the resonance frequency will with an EPI readout lead to a shift of the corresponding image in phase encoding direction. With a PROPELLER-EPI sampling scheme the periodically changing phase encoding direction causes the low resolution images to be shifted in different directions, ultimately leading to a strong image blurring. In the present work we characterize the impact of frequency shifts on PROPELLER-EPI measurements while discussing different solutions for measuring and compensating occurring frequency shifts.

Rebecca Sostheim¹, Julian Maclaren¹, Frederik Testud¹, and Maxim Zaitsev^1
1University Medical Center Freiburg, Freiburg, Baden Württemberg, Germany

Susceptibility-induced field distortions in the human brain lead to image artifacts in MRI data, especially when motion of the head occurs. To predict and correct for these distortions using a susceptibility model it is necessary to accurately distinguish between bone and air, which is difficult in normal MRI images. In this work we compare two methods (a UTE-based method and an atlas-based method) of generating a head model that can be used to calculate the induced magnetic field. We show that it is sufficient to use a model that includes only the head and consists of only two susceptibility components.

Yu Cai¹, Mark Woods¹, John Grinstaed², William Rooney¹, Xin Li¹, Qingwei Liu³, Craig Hamilton⁴, and Hongyu An^5
1Oregon Health&Science University, Portland, OR, United States, ²Siemens Healthcare, U.S.A, ³Barron Neurological Institure, Phonenix, AZ, United States, ⁴Wake Forest University, ⁵University of North carolina

Point spread function (PSF) mapping method has been proposed to correct for the geometric distortion. We further developed the PSF mapping technique using the regularized inverse solution and dual echo acquisition strategy for EPI distortion correction at 3.0T, in which the first EPI readout serve the PSF mapping and the second EPI readout is the actual EPI acquisition Its advantage is no need of the extra scan time to acquire the PSF map. With the rapidly expanding interest of fMRI study in utilizing ultra high field system, it becomes urgent to develop a robust EPI distortion correction scheme at 7.0T. We attempted to apply PSF mapping technique developed in 3.0T to test its correction efficacy at 7.0T.

Maarten J. Versluis^1,2, Bradley P. Sutton³, Paul W. de Bruin¹, Peter Boernert^2,4, Andrew G. Webb^1,2, and Matthias J.P. van Osch^1,2
1Radiology, Leiden University Medical Center, Leiden, Netherlands, ²C.J. Gorter Center for high field MRI, Leiden University Medical Center, Leiden, Netherlands, ³Department of Bioengineering, University of Illinois at Urbana-Champaign, Urbana, Champaign, IL, United States, ⁴Research Europe, Philips, Hamburg, Germany

Spatio-temporal magnetic field changes in the brain caused by breathing or body movements can lead to image artifacts. This is especially a problem in high magnetic field T₂^*-weighted sequences. By acquiring a navigator echo using multiple receive channels and their respective coil sensitivity profiles the spatially varying magnetic field was estimated, termed SENSE navigators. A retrospective reconstruction framework was implemented in Matlab using an iterative conjugate gradient solver. Compared to a simple zeroth order navigator the SENSE navigator technique lead to a significant reduction of image artifacts.

Daniel Giese^1,2, Christoph Kolbitsch¹, Andrew Aitken¹, Martin Buehrer², Tobias Schaeffter¹, and Sebastian Kozerke^1,2
1Division of Imaging Sciences, King's College London, London, United Kingdom, ²Institute for Biomedical Engineering, University and ETH Zurich, Zurich, Switzerland

Radial k-space trajectories are widely used in motion correction applications and for undersampled acquisitions given the repeated sampling of the centre of k-space. Radials are, however, prone to eddy currents and gradient delays effects. In this work, we analyze the miscentering of k-space using a 3rd order magnetic field camera, measuring k-space and quantifying gradient delays and eddy currents. Single slice, dual slice and cardiac cine images are acquired in-vivo and corrected using the measured trajectories reducing the level of artifacts. The method uses a single preparation scan to acquire k-space and can be used for any angular spoke distribution and level of undersampling.

Christopher Sica¹, Matt Restivo², Chris Kramer³, Craig Meyer², and Michael Salerno^3,4
1Radiology, Pennsylvania State University, Hershey, PA, United States, ²Biomedical Engineering, University of Virginia, Charlottesville, VA, United States, ³Cardiology, University of Virginia, Charlottesville, VA, United States, ⁴Biomedical Engineering, University of Virginia

Off-resonance correction removes spiral-related blurring artifacts. Usually a collected field map is required, lowering the efficiency of data collection. Time sensitive applications would benefit from an automatic deblurring algorithm that does not require an acquired field map. This work compares the performance of a field map based semi-automatic (SA) deblurring technique with full automatic (FA) technique and a standard gridding reconstruction (UC), in the context of spiral stress perfusion imaging. A blinded evaluation of images from 8 subjects yielded mean scores of 3.9 (UC), 4.2 (SA) and 4.3 (FA) (p=0.2), with a trend towards automatic reconstruction (FA/SA) being superior to UC (p=0.14).

Dongyeob Han¹, Yoonho Nam¹, Sung-Min Gho¹, and Dong-Hyun Kim^1
1Electrical & Electronic Engineering, Yonsei University, Seoul, Korea

Accurate R2* value quantification is required for many applications. Multi-echo GRE imaging method is generally used to evaluate R2* value. However, the main drawback of this method is the influence of macroscopic B0 inhomogeneity which occurs as unintended phase dispersion. In this study, we have presented a volumetric R2* map including which corrects for macroscopic B0 inhomogeneity. The methods use a multi-echo 3D GRE approach with increasing field inhomogeneity compensation gradient and does not require additional scan time.

Hima Patel¹, Rajat Vikram Singh², Vidit Aatrey³, Ramasubramanian Sundararajan¹, and Vivek Vaidya^1
1Global Research, General Electric, Bangalore, Karnataka, India, ²Information Technology, Indraprastha Institute of Information Technology - Delhi, New Delhi, India, ³Electrical Engineering, Indian Institute of Technology - Delhi, New Delhi, India

Manual numbering & plane prescription along spinal MRI image is a tedious and time consuming task. This paper describes a technology for automatic annotation, numbering, and oblique axial plane prescription along the vertebral column. A robust solution was developed for cervical, thoracic, and lumbar regions.

Christian Ros¹, Hellmuth Ricardo Muller-Moran², Daniel Güllmar¹, Martin Stenzel³, Hans-Joachim Mentzel³, and Jürgen Rainer Reichenbach^1
1Medical Physics Group, Department for Diagnostic and Interventional Radiology I, Jena University Hospital, Jena, Thuringia, Germany, ²McGill University, Montreal, Quebec, Canada, ³Pediatric Radiology, Department for Diagnostic and Interventional Radiology I, Jena University Hospital, Jena, Thuringia, Germany

With this contribution we present a new tractography-based estimation technique to determine the error of co-registration techniques. Acquired datat sets are used as gold standard (GS). Displacement fields are then utilized to generate prototype data sets (PDS). These PDS are co-registered to the GS and transformed back to the original space. Due to the one-to-one correspondance of the fiber tracts, that were transformed as well, the error can be exactly determined. The technique was used to evaluate the ANTs co-registration framework, in order to determine optimal registration parameter as well as suitable contrasts for the co-registration of white matter regions.

Witaya Sungkarat¹, Jiraporn Laothamatas¹, Boonlert Lumlertdacha², Supaporn Wacharapluesadee³, and Thiravat Hemachudha^4
1Advanced Diagnostic Imaging and Image-Guided Minimal Invasive Therapy Center and Radiology Dept., Ramathibodi Hospital Faculty of Medicine, Mahidol University, Rajataewe, Bangkok, Thailand, ²Queen Saovabha Memorial Institute, Thai Red Cross Society, Bangkok, Thailand, ³WHO Collaborating Center for Research and Training on Viral Zoonoses, Faculty of Medicine, Chulalongkorn University, Bangkok, Thailand, ⁴Department of Medicine and WHO Collaborating Center for Research and Training on Viral Zoonoses, Faculty of Medicine, Chulalongkorn University, Bangkok, Thailand

A new normalization technique using whole brain probabilistic DTI tractography maps is presented. This technique can address immense brain variability, even in dog’s brains. Both FSL and SPM2 were used. The usefulness of the technique was demonstrated in conducting voxel-based group analyses of FA and MD maps of rabid and normal dogs.

Sushmita Datta¹, Xiaojun Sun¹, and Ponnada A. Narayana^1
1Department of Diagnostic and Interventional Imaging, The University of Texas Health Science Center at Houston, Houston, TX, United States

Automated classification of lesions in multiple sclerosis (MS) is often hindered by the presence of false classifications (FCs). These FCs occur due to presence of some regions mimicking lesions. We have developed and implemented a false classification probability (FCP) map for improved lesion classification using the knowledge of false classifications obtained from automated segmented and validated lesion classifications. The application of FCP map significantly improved the lesion classification in 57 MS subjects as assessed by the Dice similarity indices.

Ke Gan¹, Jianli Wang², Sica Christopher², and Daisheng Luo^1
1College of Electronics and Information Engineering, Sichuan University, Chengdu, Sichuan, China, ²Department of Radiology, College of Medicine, Pennsylvania State University, Hershey, PA, United States

A new method is presented for automatic localization of the anterior and posterior commissure in MR images of human brain. Experimental results demonstrated the accuracy and effectiveness of the method when compared with a neuroradiologist’s manual delineation.

Takao Goto¹, and Hiroyuki Kabasawa^1
1Global Applied Science Laboratory, GE Healthcare, Hino-shi, Tokyo, Japan

We present a new method of the automatic scan prescription for MRI liver scans. This method is robust for the variety of liver shape due to individuals, disease and post-surgery. 3D dataset acquired using fast T1 sequence is preprocessed and converted into 2D coronal projection images as well as previous method. To deal with the shape of deformed live, MAP estimate using local inhomogeneity and liver signal distribution was applied without any shape constraints. 45 volunteers and 12 datasets simulating deformed liver were tested and showed the satisfactory results in the position accuracy and the computation time.

Ahmet Murat Bagci¹, Sudarshan Ranganathan¹, Juan S Gomez¹, Byron Lam², and Noam Alperin^1
1Radiology, University of Miami, Miami, FL, United States, ²Ophthalmology, University of Miami, Miami, FL, United States

The amount of cerebrospinal fluid (CSF) and its distribution with the central nervous system are central to our understanding of CSF related brain and spinal cord disorders. While reliable automated segmentation methods are available for brain, no such method is available for spinal CSF segmentation. Current segmentation of spinal CSF spaces relies on manual delineation which is time consuming and operator dependent. A methodology to delineate the volume of CSF within the whole central nervous system is presented and preliminary results were obtained from scans acquired before and after lumbar puncture.

Tong Zhu¹, Rui Hu², Xing Qiu², Wei Tian¹, Sven Ekholm¹, and Jianhui Zhong^1
1Imaging Sciences, University of Rochester, Rochester, NY, United States, ²Biostatistics and Computational Biology, University of Rochester, Rochester, NY, United States

Diffusion tensor imaging (DTI) technique has been widely applied to study white matter (WM) abnormality longitudinally. For many neurodegenerative diseases, such as multiple sclerosis (MS), the WM abnormality has strong subject- or time-dependent heterogeneous patterns towards which most group-based analyses for DTI, including VBM and TBSS, are less sensitive. In this study, we propose a novel statistical analysis approach for DTI based on a nonparametric spatial regression fitting of DTI data among neighboring voxels. Statistical inference can be made for both group comparison among individuals and longitudinal comparison within the same individual. Effectiveness of this approach on group comparison was compared with the VBM approach while the proof of concept for longitudinal analysis of a single subject was evaluated through a MS patient with progressive lesions. Results show that the new method is comparable to the VBM approach and performs better than VBM when the number of scans per group is small than 5. Moreover, it detected longitudinally decreased FA from only one DTI scan per time point for a MS patient with progressive lesions. Our method provides a potentially better way to assess individual abnormality to determine more directly how the brain has changed as a result of disease/injury.

Usha Sinha¹, Ali Moghaddasi², and Shantanu Sinha^2
1Physics, San Diego State University, San Diego, CA, United States, ²Radiology, University of California at San Diego, San Diego, CA, United States

Strain rate describes the rate of regional deformation and does not require 3dal tracking, or a reference state since strain rate is an instantaneous measure of kinematic properties. Our focus was to extract strain rate from spatial gradients of the velocity vector calculated from velocity encoded PC images. A challenge is the noise in the gradient maps which was addressed with a novel application of an anisotropic diffusion filter. Strain rate maps were calculated in 5 normal subjects during ankle plantarflexion excursion. Compressive and expansive strain tensors could be visualized and revealed complex spatial patterns with temporal changes.

Usha Sinha¹, Alec Biccum¹, and Shantanu Sinha^2
1Physics, San Diego State University, San Diego, CA, United States, ²Radiology, University of California at San Diego, San Diego, CA, United States

The study focuses on the application of a non-linear warping algorithm based on optical flow to recover muscle tissue displacements from dynamic magnitude MR images. These were acquired on a 1.5-T GE scanner with a gated VE-PC imaging sequence while the subject exerted periodic ankle plantarflexion excursions. Tissue displacements were extracted from the magnitude images by warping the reference (first) image to images acquired at other phases. The deformation vectors recovered by the algorithm are visualized on a 2D grid; the grid deformation agreed with anticipated tissue displacements. The study shows the potential for deformations to be recovered directly from magnitude images.

Takao Goto¹, and Hiroyuki Kabasawa^1
1Global Applied Science Laboratory, GE Healthcare, Hino-shi, Tokyo, Japan

We present a new method of automated positioning of Navigator Tracker for MRI liver scans. This method analyzes 2D images acquired by usual scout scan for scan planning and the dome peak of the liver is used for the Navigator Tracker position. Following recognition of body outline and lung location, edge-base detection of the dome peak with organ-dependent constraint enabled a few seconds computation and no additional scan only for this method, resulting in workflow improvement. 51 volunteers�f dataset from both SSFSE and FSPGR-base scout scan were tested offline and obtained satisfactory results.

Bradley P Sutton¹, Andrew Naber¹, Jason Wang¹, Jamie L. Perry², and David P Kuehn^3
1Bioengineering Department, University of Illinois at Urbana-Champaign, Urbana, IL, United States, ²Department of Communication Sciences and Disorders, East Carolina University, Greenville, NC, United States, ³Department of Speech and Hearing Sciences, University of Illinois at Urbana-Champaign, Champaign, IL, United States

As the frame rate of dynamic speech imaging with MRI increases, automated extraction of frame-by-frame soft tissue movements becomes critical for evaluating large studies of pathology or cultural differences in movement. This is a challenging task as dynamic MRI suffers from noisy images and lack of contrast in structures. We present a semi-automated algorithm to extract two tongue positions (tip and dorsum) and compare the tracking results with three trained speech scientists. The semi-automated algorithm performs well in correlating with the manual tracings on data from four study participants.

Dan Rettmann¹, Xiaodong Tao², Jun Xie³, and Ajit Shankaranarayanan^4
1Global Applied Science Laboratory, GE Healthcare, Rochester, MN, United States, ²GE Healthcare, Beijing, China, ³MR Engineering, GE Healthcare, Waukesha, WI, United States, ⁴Global Applied Science Laboratory, GE Healthcare, Menlo Park, CA, United States

Prospective motion correction techniques have been shown to be beneficial in less compliant patient populations. The image-based PROspective MOtion (PROMO) correction performs best if navigators are optimally placed with respect to the anatomy. This work describes a new technique to automatically position the navigators based on an automatic scan planning algorithm that uses a rapid implicit symmetry axis determination.

Jens-Peter Kühn^1,2, James H Holmes³, Diego Hernando¹, and Scott B Reeder^1,4
1Department of Diagnostic Radiology, University of Wisconsin-Madison, Madison, Wisconsin, United States, ²Department of Diagnostic Radiology, University Greifswald, Greifswald, MV, Germany, ³Global Applied Science Laboratory, GE Healthcare, Madison, Wisconsin, United States, ⁴Medical Physics, Biomedical Engineering and Medicine, University of Wisconsin-Madison, Madison, Wisconsin, United States

This work optimizes the excitation flip angle of a navigator pencil-beam to maximize optimal liver/lung contrast while avoiding saturation artifacts in the liver. We used an investigative navigator-gated T1-weighted 3D GRE sequence under clinical conditions for gadoxetic acid-enhanced liver MRI in the hepatobiliary phase. The presence of saturation artifacts and the optimal navigator flip angle are highly dependent on the imaging flip angle and the presence of gadolinium in the liver. Using an imaging flip angle of 30o, 20-60 minutes after gadoxetic acid, the optimal navigator flip angle is 90o, without any appreciable saturation artifact.

Trong-Kha Truong^1
1Brain Imaging and Analysis Center, Duke University, Durham, NC, United States

Multi-shot spiral imaging is a promising alternative to echo-planar imaging for high-resolution diffusion tensor imaging. However, subject motion in the presence of diffusion-weighting gradients causes phase inconsistencies among different shots, resulting in signal loss and aliasing artifacts in the reconstructed images. Such artifacts can be reduced by using a variable-density spiral trajectory or a navigator echo, however at the cost of a longer scan time. Here, we propose a novel iterative phase correction method that can inherently correct for these motion-induced phase errors with no scan time penalty.

Anja Brau¹, and Yuji Iwadate^2
1Global Applied Science Lab, GE Healthcare, Menlo Park, CA, United States, ²Global Applied Science Lab, GE Healthcare, Hino, Tokyo, Japan

Respiratory navigator techniques rely on the processing algorithm to accurately extract motion information from the navigator echo. When using a multi-channel receive coil, the selection of which channel to use can influence the success of the navigator processing algorithm. By selecting a subset of channels, the coil sensitivity modulation of the navigator signal can be exploited to enhance the prominence of features of interest while also improving real-time processing efficiency. In this work, we evaluate the suitability of different coil selection methods for use with navigator processing and propose a method based on a matched filter.

Zarko Celicanin¹, Oliver Bieri¹, Klaus Scheffler^2,3, and Francesco Santini^1
1Division of Radiological Physics, Department of Radiology and Nuclear Medicine, University of Basel Hospital, Basel, Switzerland, ²MRC Department, MPI for Biological Cybernetics, Tübingen, Germany, ³Dept. Neuroimaging and MR-Physics, University of Tübingen, Tübingen, Germany

A spectrally selective pencil-beam navigator method has been recently published for motion compensation of MRgHIFU therapy of abdominal organs. The suggested spectral navigator consisted of a spiral 2D pencil-beam using 1-[-1]-1 binomial pulses and was proposed for the tracking of abdominal organs, such as the liver or kidney, based on the surrounding adipose tissue signal. Here, we present a novel spectrally selective navigator technique based on a spin echo with orthogonal planes for the excitation and refocusing pulses , also called crossed-pair navigator, but using 1-[-1] binomial pulses for excitation and refocusing.

Gabriele Beck¹, John Penatzer², Vincent Denolin³, Kenneth Coenegrachts⁴, and Gwenael Herigault^1
1Philips Healthcare, Best, Netherlands, ²Philips Healthcare, Cleveland, United States, ³Philips Healthcare Benelux, Brussels, Belgium, ⁴Department of Radiology, AZ St.-Jan Brugge-Oostende AV, Bruges, Belgium

In oncology there is a high interest in investigating suspicious lesions over different MR contrasts. In the abdomen breathing motion and breath hold variations are the main issues that restrict the direct comparison of a lesion across different scans which makes the classification of those lesions more difficult. In this work, prospective interscan alignment using a fast spiral Navigator acquisition allowed for real time adaptation of the scan position and registration of the volume of interest. This allows a direct comparison of lesions across breath hold or free breathing scans within an examination. End expiration triggered and breath held scans show a good consistency over the different contrast scans which were tested. While for inspiration breath holds the large feet head difference in the dome level is corrected, variations in the appearance can be larger in some subjects and can be explained by rotation and deformation of part of volume.

Feng Huang¹, Wei Lin¹, Chiel den Harder², Gabrielle Beck², Clemens Bos³, George Randy Duensing¹, and Arne Reykowski^1
1Invivo Corporation, Gainesville, FL, United States, ²Advanced Solutions, MRI, Philips Healthcare, Best, Netherlands, ³MR Clinical Science, Philips Healthcare, Best, Netherlands

Most existing motion compensation techniques for brain MRI assume that the motion is rigid. In fact, many kinds of non-rigid motion such as eye movement (eye ball rolling), skin movement (frowning), and jaw movement (swallowing, yawning), can also cause serious spatially local artifacts in brain imaging. These inevitable problems have not been carefully addressed. To remove these artifacts, data rejection and reconstruction with remaining unpolluted data can be used. However, the reconstruction with partial data could result in potential artifacts, such as reduced SNR and loss of contrast. In this work, it is proposed to use the image reconstructed with the full k-space, which is locally artifact corrupted but with high SNR, as a regularization in reconstruction to achieve an image with low artifact level and high SNR.

James H Holmes¹, Philip J Beatty^2,3, Howard A Rowley⁴, Ann Shimakawa⁵, and Jean H Brittain^1,6
1Global Applied Science Laboratory, GE Healthcare, Madison, WI, United States, ²Global Applied Science Laboratory, GE Healthcare, Toronto, ON, Canada,³Physical Sciences, Sunnybrook Research Institute, Toronto, ON, Canada, ⁴Radiology, University of Wisconsin-Madison, Madison, WI, United States, ⁵Global Applied Science Laboratory, GE Healthcare, Menlo Park, CA, United States, ⁶Canada

In this work, we describe a method for generating motion corrected T1w and PDw (Proton Density) weighted images using the PROPELLER acquisition. Traditionally, PROPELLER relies on wide image blades for motion correction and blade combination methods to maximize the SNR of the overall image. We demonstrate two modifications. First, a centric acquisition is performed. Motion correction parameters are then determined using the full ETL and associated blade width of 35-29 echoes. However, the data used for image generation is taken from only the center 3-5 echoes, representing the shortest TE times. The motion correction parameters from the wide blade acquisition are then used to correct the narrow blade image data.

Candice A. Bookwalter¹, Nicole Seiberlich², Michael W. Harrell³, Philipp Ehses^4,5, Mark A. Griswold^1,2, and Vikas Gulani^1,2
1Department of Radiology, University Hospitals Case Medical Center\Case Western Reserve University, Cleveland, OH, United States, ²Department of Biomedical Engineering, Case Western Reserve University, Cleveland, OH, United States, ³School of Medicine, Case Western Reserve University, Cleveland, OH, United States, ⁴Max Planck Institute for Biological Cybernetics, Tübingen, Germany, ⁵Department for Neuroimaging, University Hospital Tübingen, Tübingen, Germany

While radial k-space trajectories are known to be inherently insensitive to motion due to the oversampling of the center of k-space, patient motion causes image degradation by object distortion and streaking artifacts. An algorithm called Motion Artifact Removal by Retrospective Resolution Reduction (MARs) for rectilinear trajectories has been previously described, which automatically and retrospectively identifies a transition between breath hold to free breathing and subsequently removes the corrupted data for a motion artifact free, yet lower resolution image. This work expands the MARs method to radial sequences demonstrated through volunteer and patient data.

Andre Jan Willem van der Kouwe¹, and Himanshu Bhat^2
1A. A. Martinos Center for Biomedical Imaging, Department of Radiology, Massachusetts General Hospital, Charlestown, MA, United States, ²Siemens Healthcare US, Charlestown, MA, United States

By ordering the spokes of a 3D radial encoded scan in such a way that images may be reconstructed from smaller windows of the acquisition, rigid body motion detection and correction is possible. A method is presented for uniformly ordering the spokes across a hierarchy of window lengths, along with a hierarchical image reconstruction approach that results in a complete and consistent k-space representation and final reconstruction. The method is demonstrated in an ultrashort TE scan of a volunteer’s brain.

Ghislain Vaillant¹, Claudia Prieto², Christoph Kolbitsch², Graeme Penney², and Tobias Schaeffter^2
1Division of Imaging Sciences and Biomedical Engineering, King's College London, London, United Kingdom, ²King's College London

Motion is a dramatic factor of image quality degradation in MRI. Dynamic MR can be used for limiting the effect of motion during the acquisition. We propose to combine a conventional sliding window reconstruction with a Fourier-based registration technique in order to correct for rigid motion in highly segmented radial acquisitions. The proposed method has been compared to image-based registration and was successfully applied in-vivo on 6 volunteers for 2D brain imaging.

Ashley Gould Anderson III¹, Julia Velikina¹, Oliver Wieben^1,2, and Alexey Samsonov^2
1Medical Physics, University of Wisconsin, Madison, Wisconsin, United States, ²Radiology, University of Wisconsin, Madison, Wisconsin, United States

Robust 3D rigid-body motion correction technique using 3D radial trajectories. The method was adapted for additional challenges of cardiac-gated phase contrast flow imaging.

Hernan Jara¹, Stephan William Anderson¹, and Osamu Sakai^1
1Radiology, Boston University Medical Center, Boston, Massachusetts, United States

Purpose: To develop an algorithm for synthesizing T1rho maps using standard qMRI relaxometry instead of using direct spin-lock imaging acquisition techniques. Methods: A variant of the mixed turbo spin echo pulse sequence which is multispectral in PD, T1, and T2 was used at 3T. Images were processed with qMRI algorithms to generate T1 maps at varying spin-lock phase angles. Results: The resulting T1rho maps vary in appearance as a function of the spin-lock phase angle. Conclusion: A qMRI relaxometry-based technique for mapping the longitudinal relaxation time in the rotating frame T1rho has been developed and tested in the human head.

Kotaro Sekiya^1,2, Memi Watanabe¹, Osamu Sakai¹, Rohini N. Nadgir¹, Adham A. Mottalib¹, Mingxin Zheng³, and Hernan Jara^1
1Boston Medical Center, Boston University School of Medicine, Boston, MA, United States, ²Nihon University Graduate School of Dentistry at Matsudo, Chiba, Japan, Japan, ³Graduate Student of Biomedical Engineering Department, Boston University, Boston, MA, United States

Purpose: To develop a technique for mapping the qMRI T1/T2 ratio and to study the hierarchical order of T1/T2 ratio values of the normal human head. Methods: A variant of the mixed turbo spin echo pulse sequence which is multispectral in PD, T1, and T2 was used at 3T. Images were processed with qMRI algorithms to generate T1/T2 ratio maps and whole head histogram. Results: The resulting T1/T2 ratio maps exhibit dark CSF and bright muscles. The cerebral tissues appear in the hierarchy of signal intensities of a T2-weighted fluid attenuated with inversion recovery (T2W-FLAIR) image. Conclusion: A technique for mapping the T1/T2 relaxometric ratio has been developed.

Umesh Rudrapatna¹, Annette van der Toorn¹, and Rick Dijkhuizen^1
1Biomedical MR Imaging and Spectroscopy Group, Image Sciences Institute, University Medical Center Utrecht, Utrecht, Netherlands

Quantitative MRI is an indispensable tool for the research community, the goal of which is to find unbiased and minimum variance estimates of critical MR parameters of interest. Improvements in bias and variance minimization in these estimates can lead to gainful tradeoffs between imaging time, SNR and resolution. Here, we present one such possible improvement in T₁ and T₂ parameter estimation which holds promise by pushing the limits of confidence in these estimations to low SNR regions, where conventional approaches fail. This technique, based on a separable least-squares approach to fit complex MR data has the ability to overcome the usual difficulties associated with using complex data for T₁ and T₂ (T₂*) mapping.

Weitian Chen¹, Patrick D Koon², and Ajit Shankaranarayanan^1
1Global MR Applied Science Laboratory, GE Healthcare, Menlo Park, CA, United States, ²GE Healthcare, San Francisco, CA, United States

Quantification of T2 or T1rho in brain can provide additional diagnostic information to anatomy images. A major challenge to such quantification methods is very long scan time in order to achieve high resolution with 3D coverage of the brain. In this work, we investigated fast 3D quantitative T1rho or T2 imaging of whole brain based on highly SNR efficient 3D fast spin echo acquisition. We reported a simple method to reduce adversary eddy current effect in the developed pulse sequence. The T2-weighted image contrast can be utilized for simultaneous anatomical imaging besides T1rho or T2 quantification.

Dushyant Kumar¹, Thanh Nguyen², Susan Gauthier³, and Ashish Raj^2
1Neuroradiology, University of Hamburg, Hamburg, Hamburg, Germany, ²Radiology, Weill Cornell Medical College, Newyork, NY, United States,³neurology, Weill Cornell Medical College, Newyork, NY, United States

Problem: The adoption of conventional T2-relaxometry approach in clinical practice is impeded by long acquisition time and challenging T2-data analysis. Methods: The voxelwise conventional regularization is performed, then the spatial smoothness constraint over local neighborhood is implemented using a Bayesian spatial approach. Results: Our algorithm has better white matter tissue detection ability and better lesions detection compared to other methods. Conclusions: We extract consistent entire brain MWF map by imposing the spatial constraints on noisy T2-prep spiral data acquired within clinically feasible scan time (3T: 10 minutes; 1.5T: 20 minutes).

Mengchao Pei^1,2, Thanh D. Nguyen¹, Lijia Wang², Tian Liu¹, Mitch Anthony Cooper¹, Jianqi Li², and Yi Wang^1,2
1Weill Cornell Medical College, New York, NY, United States, ²Shanghai Key Laboratory of Magnetic Resonance, East China Normal University, Shanghai, Shanghai, China

A novel T2* mapping method based on the definite integral of the T2* signal decay curve (DISC) is proposed. Numerical simulations show that this method is more robust against noise effect than the widely used Levenberg-Marquardt (LM) algorithm, particularly at low SNR. DISC can reduce T2* data fitting time for a brain data set from 20 hours by LM to 5 minutes. This method is easy to implement and does not require an initial guess.

Arvin Arani^1,2, Michael Da Rosa³, Elizabeth Ramsay¹, Donald Plewes^1,2, Masoom Haider¹, and Rajiv Chopra^1,2
1Sunnybrook Research Institute, Toronto, Ontario, Canada, ²Medical Biophysics, University of Toronto, Toronto, Ontario, Canada, ³Institute of Medical Science, Sunnybrook Research Institute, Toronto, Ontario, Canada

Problem: The objectives of this study were to investigate the tolerability and technical feasibility of performing endorectal magnetic resonance elastography (eMRE) in human volunteers within the representative age group commonly affected by prostate cancer. Methods: Dynamic endorectal magnetic resonance elastography was conducted on 12 volunteers at oscillation frequencies and displacement amplitudes of 100-300Hz and 1-40µm, respectively. Results: Volunteers were able to tolerate eMRE without experiencing any pain. Wave propagation was observed throughout the entire prostate at frequencies as high as 300Hz. Conclusions: These results suggest eMRE is tolerable and technically feasibility, motivating future evaluation in patients.

Philippe Garteiser¹, Ramin Sahebjavaher², Ralph Sinkus¹, Leon Ter Beek³, Bernard E Van Beers⁴, Septimian E Salcudean², and Ralph Sinkus^1
1CRB3 U773 Université Paris Diderot, Sorbonne Paris Cite, INSERM, Paris, 75018, France, ²University of British Columbia, Vancouver, Canada, ³Philips Medical Systems, Netherlands, ⁴Service of Radiology, Hopital Beaujon, Clichy, France

A novel magnetic resonance elastography pulse sequence is introduced that strikes the balance between imaging speed, motion-encoding efficiency and high quality phase images. This pulse sequence was experimentally validated as capable of acquiring full 3D motion-encoding in a volume in less than a minute, and as suitable for mono- and multifrequency MR rheology experiments.

Maxime Ronot^1,2, Mathilde Wagner¹, Simon Lambert², Sabrina Doblas², Ralph Sinkus², Valerie Vilgrain^1,2, and Bernard E Van-Beers^1,2
1Radiology, Beaujon Hospital, Clichy, France, ²CRB3 INSERM U773, France

The diagnosis, staging and quantification of fibrosis rely on liver biopsy, which is invasive with significant risks. MR Elastography (MRE) is an emerging technique that allows fibrosis assessment by measuring the viscoelastic properties of the liver Our study presents the preliminary results of a high-frequency MRE-assessed fibrous liver analysis in an ex vivo rat model. Fibrosis was induced using CCl4 intoxication. We showed that the mean value of the elasticity shear modulus (Gd) in normal livers was significantly lower than in fibrous ones after 3 weeks of intoxication (p=.01) and correlated to the histological findings. It is expected that Gd will significantly increase in the next 5 weeks.

Joshua D. Trzasko¹, and Armando Manduca^1
1Mayo Clinic, Rochester, MN, United States

In steady-state magnetic resonance elastography (MRE), the delivery of precise quantitative information about tissue stiffness inherently depends on accurate estimation of the invoked harmonic phase signal. In this work, we propose a novel, statistically-motivated strategy for estimating the first harmonic signal directly from raw, complex, multichannel MRI data, and discuss the incorporation of signal prior models for prospective noise suppression. As a result, the challenge associated with handling the complex and signal-dependent noise distribution of MRE phase images during retrospective denoising is completely mitigated.

Matthew C Murphy¹, John Huston, III¹, Kevin J Glaser¹, Armando Manduca², Joel P Felmlee¹, and Richard L Ehman^1
1Department of Radiology, Mayo Clinic, Rochester, MN, United States, ²Department of Physiology and Biomedical Engineering, Mayo Clinic, Rochester, MN, United States

MR elastography (MRE) is an MR technique for noninvasively measuring tissue stiffness. Some applications, such as the brain, require a volumetric acquisition and 3D inversion because of through-plane wave propagation. An EPI pulse sequence is useful for such an application because it can acquire a volume of data with short acquisition times. However, slice-to-slice discontinuities in the phase of the EPI images can arise will bias the final inversion results. The purpose of this work is to present a filtering method that removes these slice-to-slice discontinuities, and as a result improves MRE repeatability.

Bogdan Dzyubak¹, Kevin Glaser², Meng Yin², Armando Manduca², and Richard Ehman^2
1Mayo Graduate School, Mayo Clinic, Rochester, Minnesota, United States, ²Radiology, Mayo Clinic, Rochester, Minnesota, United States

Measurements of liver stiffness based on MR Elastography images have an inherent variability due to differences in ROI definition between readers. This study was aimed at developing an automatic algorithm for segmenting the liver in MRE images and measuring stiffness in an ROI with reliable wave propagation. The algorithm was shown to perform as well as an experienced MRE reader even in images with high artifact. This technique may streamline the measurement of liver stiffness while reducing time, variability, and cost.

Melvyn B. Ooi¹, Murat Aksoy¹, Ronald D. Watkins¹, and Roland Bammer^1
1Department of Radiology, Stanford University, Stanford, CA, United States

The ability to track the positions of multiple micro RF-coil “active markers� in the MRI scanner has been the foundation of several recent advances in MR-guided intervention, as well as motion correction applications. The current work demonstrates that un-tuned markers, while simplifying coil design and reducing unwanted RF-related effects, provide sufficient signal for accurate/precise position tracking measurements in a temporal resolution suitable for use in real-time applications. A series of position measurements using un-tuned and tuned markers are performed on a well-defined grid-phantom for comparison.

David Rotenberg^1,2, Mark Chiew^1,2, Fred Tam^1,3, Shawn Ranieri¹, and Simon Graham^2,3
1Rotman Research Institute, Toronto, Ontario, Canada, ²Department of Medical Biophysics, University of Toronto, Toronto, Ontario, Canada, ³Physical Sciences Sunnybrook Research Institute, Toronto, Ontario, Canada

Head motion artifacts are a substantial source of error in Blood Oxygen Level Dependent fMRI that limits its use in neuroscience research and clinical settings. Real-time scan-plane correction by optical tracking has been shown to suppress artifacts due to slice-misalignment and non-linear spin-history effects, however, residual artifacts due to dynamic magnetic field non-uniformity may remain in the data. We demonstrate a novel correction approach that integrates volume by volume, geometric distortion correction by PLACE into a real-time scan-plane update system by optical tracking, applied to an fMRI finger tapping experiment with overt head motion to induce dynamic field non uniformity.

Murat Aksoy¹, Melvyn Ooi¹, Ronald D Watkins², Daniel Kopeinigg², Christoph Forman³, and Roland Bammer^1
1Center for Quantitative Neuroimaging, Department of Radiology, Stanford University, Stanford, CA, United States, ²Department of Radiology, Stanford University, Stanford, CA, United States, ³Computer Science, Friedrich-Alexander-University Erlangen-Nuremberg, Erlangen, Germany

Active markers and optical tracking systems (i.e. cameras) have been used for prospective correction of motion artifacts in head MRI. Optical tracking systems have the advantage that they require no additional MR data acquisition (i.e. navigators) to detect motion, whereas active markers can detect motion without the need to perform cross-calibration between the scanner and the tracking system. In this study, we combined active marker and optical tracking to benefit from the advantages of both systems. The active marker system was used to perform cross-calibration of the optical tracking system in a very short time with no discomfort to the patient so that the actual tracking can be done with the camera.

Eric K Gibbons¹, Samantha J Holdsworth², Melvyn B Ooi², Murat Aksoy², and Roland Bammer^2
1Department of Bioengineering, Stanford University, Palo Alto, California, United States, ²Center for Quantitative Neuroimaging Department of Radiology, Stanford University

Echo-Planar Imaging (EPI) is a fast acquisition technique that can reduce the effect of patient motion but it is still prone to motion that occurs between volumes. The real-time prospective approach can correct for head rotation by rotating the gradients. However, gradient hardware delays can result in different ghosting parameters for each rotation. Here we test whether the ghost correction parameters estimated from the first EPI scan can be applied to oblique angles that are limited to the expected extent of patient head motion, in order to test whether this extra reference scan is necessary for real-time prospective motion correction applications.We found that even in small rotations of 2.5 degrees ghosting artifacts appeared in the image and required a separate correction parameter to be calculated. Additionally, in the presence of gradient delays, such ghosting correction is equally unfeasible with or without the additional scans.

Benedikt Andreas Poser¹, Kazim Gumus¹, Brian Robert Keating¹, Brian Armstrong², Todd P Kusik², Julian Maclaren³, Thomas E Prieto⁴, Oliver Speck⁵, Maxim Zaitsev³, V Andrew Stenger¹, and Thomas Ernst^1
1UH-QMC Neuroscience and MR Research Program, University of Hawaii, Honolulu, Hawaii, United States, ²Department of Electrical Engineering and Computer Science, University of Wisconsin-Milwaukee, United States, ³Magnetic Resonance Development and Application Center, University Hospital Freiburg, Germany, ⁴Medical College Wisconsin, United States, ⁵Dept Biomedical Magnetic Resonance, Otto-von-Guericke-University, Magdeburg, Germany

fMRI data quality critically depends on subject motion. In EPI, its leads to head pose dependent distortions that cannot be corrected by post-hoc image realignment and hamper the detection of task activation. ‘Prospective’ motion correction approaches have been developed to track head position in real-time and continuously update slice positions during the scan. However, changes in absolute head orientation relative to B0 may alter the field distribution and hence EPI image distortions. To address this concern, we explored a dynamic distortion correction approach. We used IDEA EPI (interleaved dual echo with acceleration EPI) which allows field maps to be extracted from and applied to every fMRI volume, combined with real-time motion correction by means of an retro-grade-reflector (MPT) based tracking system.

Jessica Schulz¹, Thomas Siegert¹, Enrico Reimer¹, Maxim Zaitsev², Julian Maclaren², Michael Herbst², and Robert Turner^1
1Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany, ²University Medical Center Freiburg, Freiburg, Germany

We have designed an embedded optical tracking system that provides real-time motion information for prospective head motion correction for MRI at 7T. This system enables the acquisition of motion-corrected very high resolution in vivo MR images of the human brain (0.4 mm isotropic 3D FLASH, 0.5 mm isotropic 2D TSE). Uncorrected and corrected 3D FLASH images were acquired in four volunteers. Quantitative analysis demonstrated that the image quality was reliably improved by the motion correction.

Feng Huang¹, Wei Lin¹, Chiel den Harder², Clemens Bos³, George Randy Duensing¹, and Arne Reykowski^1
1Invivo Corporation, Gainesville, FL, United States, ²Advanced Solutions, MRI, Philips Healthcare, Best, Netherlands, ³MR Clinical Science, Philips Healthcare, Best, Netherlands

Multi-channel coils have been widely used in clinical practice for larger field of view (FOV) and higher signal to noise ratio (SNR). Due to the coil geometry, different coil elements have different sensitivity to some kinds of artifacts, such as flow, pulsation, breathing, hardware imperfections, etc. Conventional artifact reduction algorithms treat all coil elements equally, which could result in lower sensitivity of motion detection, reduced SNR, or residual artifacts. In this work, it is proposed to detect and compensate artifacts for each channel individually. Examples with in-vivo data show that higher SNR and lower residual artifact level can be achieved by using a channel by channel approach than by using a combined channel approach.

Feng Huang¹, Wei Lin¹, Chiel den Harder², Gabrielle Beck², Clemens Bos³, George Randy Duensing¹, and Arne Reykowski^1
1Invivo Corporation, Gainesville, FL, United States, ²Advanced Solutions, MRI, Philips Healthcare, Best, Netherlands, ³MR Clinical Science, Philips Healthcare, Best, Netherlands

Typical clinical MR examinations are composed of several sets of scans to acquire images with different contrast, such as T1w, T2w and DWI. Currently, the acquisition and reconstruction of these images are kept separate. Since the same subject is scanned in the same system using the same RF coil, there is sharable common information among these images. It is shown that the data correlation among channels from motion insensitive sequence can be shared for a motion compensation. Preliminary results with in-vivo data sets show that rigid motion artifacts can be corrected using sharable information from images with different contrast.

Dan Rettmann¹, Jun Xie², Nathan S White³, Anders Dale^4,5, and Ajit Shankaranarayanan^6
1Global Applied Science Laboratory, GE Healthcare, Rochester, MN, United States, ²MR Engineering, GE Healthcare, Waukesha, WI, United States,³Dept. of Cognitive Science, University of California, San Diego, La Jolla, CA, United States, ⁴Dept. of Neuroscience, University of California, San Diego, La Jolla, CA, United States, ⁵Dept. of Radiology, University of California, San Diego, La Jolla, CA, United States, ⁶Global Applied Science Laboratory, GE Healthcare, Menlo Park, CA, United States

The use of double inversion recovery (DIR) sequences for improved lesion detection in multiple sclerosis and diseases involving abnormal grey matter has been increasing in popularity. DIR sequences are inherently long due to necessary T1 recovery and as such are more susceptible to patient motion. In this work we investigate the application of a prospective motion correction (PROMO) technique applied to a 3D FSE DIR sequence.

Freddy Odille^1,2, Anne Menini^1,2, Pierre-André Vuissoz^1,2, Laurent Bonnemains^1,3, Damien Mandry^1,2, and Jacques Felblinger^1,4
1IADI, INSERM U947 - Nancy Université, Nancy, France, ²Pôle Imagerie, CHU de Nancy, Nancy, France, ³Pôle Cardiologie, CHU de Nancy, Nancy, France, ⁴CIC-IT 801, CHU de Nancy, Nancy, France

The combination of partial Fourier acquisition with motion-compensated reconstruction schemes is investigated. Similar to SENSE, such generalized reconstructions can be modified to include a phase constraint and therefore deal with asymmetric k-space sampling. However partial Fourier is sensitive to the construction of a “good” low resolution phase map from the central part of k-space, which may also be corrupted by motion. Here we propose to build a motion-compensated phase map using the GRICS motion correction framework. The technique was tested on volunteers with a free-breathing high resolution 3D sequence used clinically for myocardial viability assessment.

Yuji Iwadate¹, Anja C.S. Brau², and Hiroyuki Kabasawa^1
1Global Applied Science Laboratory, GE Healthcare Japan, Hino, Tokyo, Japan, ²Global Applied Science Laboratory, GE Healthcare, Menlo Park, CA, United States

Respiratory-gated fat-suppressed 3D spoiled gradient-recalled echo with navigator echo (Navigated LAVA) enables 3D T1-weighted imaging of free breathing patients with considerably reduced motion artifacts. In this work, enhanced navigated LAVA was developed to address the spin saturation effect on navigator signal, utilizing a wait insertion and a variable flip angle scheme. Variable flip angle scheme was realized by combination of ramp-up, ramp-down, and attenuation strategies in order to optimize the point spread function. In volunteer imaging, enhanced navigated LAVA not only increased navigator signal but also improved image quality of LAVA itself.

Suchandrima Banerjee¹, and Ajit Shankaranarayanan^1
1Global Applied Science Lab, GE Healthcare, Menlo Park, California, United States

Motion remains a major source of artifact in day-today MR scans. In recent years there has been several works on utilization of additional information from multiple coils to reduce motion artifacts. Most of these methods use parallel imaging reconstruction algorithms to detect and resynthesize motion-corrupted acquisition segments. In this work we take a different approach and propose to detect individual coil data those are most affected by the motion. Our reconstruction strategy is based on the assumption that coil sensitivities being spatially varying, motion artifacts will appear differently relative to mean signal in the various coil images.

Joon-Sup Jeong¹, Jong-Ho Lee², Se-Hong Oh¹, Taek-Hyun Ryu¹, Dae-Hyuk Kwon¹, Young-Bo Kim¹, and Zang-Hee Cho^1
1Neuroscience Research Institute, Inchon, Korea, ²Department of Radiology, University of Pennsylvania, Pennsylvania, Armenia

Quantitative Susceptibility Mapping whose signal intensity is proportional to the underlying tissue magnetic susceptibility. The method provides a novel contrast in MRI and allows us to visualize different magnetic susceptibility components. In this abstract, we applied MEDI method to calculation magnetic susceptibility at 1.5T, 3T and 7T and compared the susceptibility values across the field strengths. In general, high-field-strength provides a better magnitude image than low-field-strength. In MEDI method, this magnitude method affects two aspects of processing QSM. This magnitude method affects two aspects of processing QSM. Therefore, 7T QSM image tends to have more accurate anatomical information than low-field-strength QSM image.

Sung-Min Gho¹, Yoonho Nam¹, Dongyeob Han¹, Chunlei Liu², and Dong-Hyun Kim^1
1Electrical and Electronic Engineering, Yonsei University, Sinchon-dong, Seoul, Korea, ²Brain Imaging and Analysis Center, Duke University, Durham, NC, United States

Image phase has unique information about tissue composition such as gray and white matter structures. Phase imaging is used for various imaging methods such as susceptibility weighted imaging and quantitative susceptibility mapping. To obtain phase image, long echo time is needed to acquire sufficient phase shift. However, there is limitation due to the B0 inhomogeneity artifacts (i.e. signal loss in the frontal and sinus regions). One method to solve this problem in 3D imaging is the 3D z-shimming method. We propose to solve the above problem for obtaining the phase image by using a single-scan 3D multi-echo z-shimming sequence and image reconstruction method.

Obaidah Anees Abuhashem¹, Berkin Bilgic¹, and Elfar Adalsteinsson^1,2
1EECS, Massachusetts Institute of Technology, Cambridge, MA, United States, ²Harvard-MIT Division of Health Sciences and Technology, Cambridge, MA, United States

Quantitative Susceptibility Mapping (QSM) is used to quantify tissue magnetic susceptibility, leading to applications such as tissue contrast enhancement, venous blood oxygenation, and iron quantification. Quantification of the susceptibility distribution χ involves removal of background effects on the MRI signal phase and the solution of an ill-posed inverse problem describing the mapping from the phase to the tissue susceptibility. In this work, background removal is achieved by using the effective dipole fitting algorithm and susceptibility inversion is performed via imposing ℓ1 norm regularization on the spatial gradients of χ. As both algorithms are computationally demanding, it is crucial to increase the computational throughput and make regularized QSM a feasible and real-time methodology. Herein, the computational power Graphics Processing Cards (GPUs) is utilized to greatly accelerate the processing times, and both MATLAB and GPU libraries of the regularized QSM method are made available online for reproducibility.

Wei Li¹, Bing Wu¹, and Chunlei Liu^1,2
1Brain Imaging & Analysis Center, Duke University, Durham, North Carolina, United States, ²Radiology, Duke University, Durham, North Carolina, United States

Magnetic susceptibility typically has high contrast and SNR, and is related to the fiber angle with a simple sine-squared relationship, thus provides a promising candidate for extracting the white matter fiber orientation information from gradient echo MRI. The application of susceptibility tensor imaging, however, can be hampered by the imperfect registration due to different image distortion at different head orientations. In this work, we developed a regularized k-space-based method for susceptibility tensor reconstruction, which can effectively reduced the artifacts caused by imperfect registration and enhance the robustness of susceptibility tensor imaging.

Cynthia Wisnieff^1,2, Tian Liu^1,2, Pascal Spincemaille², and Yi Wang^1,2
1Cornell University, New York City, NY, United States, ²Weill Cornell Medical College, New York City, NY, United States

Susceptibility tensor imaging is an ill-posed inverse problem that requires sampling at many impractical orientations. We investigate the ability to detect susceptibility anisotropy in the human brain using structural prior information and reducing the number of orientations to as few as three. The prior information includes 1) the susceptibility tensor shape is cylindrically symmetric (CS); and 2) the susceptibility tensor shares its orthonormal basis with the diffusion tensor. Use of CS was validated in carbon fibers. It is observed here that the susceptibility anisotropy pattern detected in the brain appears to be similar between the 13 and 3 orientation reconstructions.

Cynthia Wisnieff^1,2, Tian Liu^1,2, and Yi Wang^1,2
1Cornell University, New York City, NY, United States, ²Weill Cornell Medical College, New York City, NY, United States

Susceptibility tensor imaging (STI) is an ill posed problem that suffers from noise propagation from non-ideal data. Assuming cylindrical symmetry in the shape of the tensor may improve the condition of this inverse problem. Anatomic prior information may be used to reduce noise through its incorporation into reconstructions of apparent susceptibilities at all orientations, from which the susceptibility tensor can be computed. These ideas were approximately confirmed in our phantom data and initial human data.

Shuai Wang^1,2, Weiwei Chen³, Tian Liu⁴, A. John Tsiouris², Jianlei Liu⁵, and Yi Wang^2,6
1University of Electronic Science and Technology of China, Cheng Du, Si Chuan, China, ²Radiology, Weill Cornell Medical College, New York, New York, United States, ³Radiology, Tongji Hospital, Tongji Medical College, Huazhong University of Science&Technology, Wuhan, Hubei, China, ⁴MedImageMetric LLC, New York, New York, United States, ⁵XiDian University, Xi'an, ShaanXi, China, ⁶Biomedical Engineering, Cornell University, Ithaca, New York, United States

In this study, we retrospectively evaluated 114 consecutive patients imaged with Quantitative Susceptibility Mapping (QSM) by examining image quality under different situations, including presence of diverse pathology, various brain regions, and a wide range of age. The consistent image quality observed in this study suggested the potential of QSM as a tool for clinical diagnosis.

Jin Tang¹, Saifeng Liu¹, Jaladhar Neelavalli², Yu-Chung Norman Cheng², and E Mark Haacke^1,2
1Biomedical Engineering, McMaster University, Hamilton, ON, Canada, ²Academic Radiology, Wayne State University, Detroit, Michigan, United States

Mapping susceptibility from field perturbation data often uses a high pass filter to remove the low spatial frequency phase, however, using high pass filter will result in a concomitant loss of important local phase information and lead to decreases of susceptibility values inside vessels, especially for large vessels. To solve this problem and to improve the accuracy of susceptibility quantification of veins, we propose a new method which uses correcting factor (CF) to automatically adjust the underestimated susceptibility value inside the vein. The underestimated susceptibility values will be completely compensated by the CF.

Meng-Chi Hsieh^1,2, San-Chao Hwang³, Hsu Chang³, and Jyh-Horng Chen^1,2
1Graduate Institute of Biomedical Electronics and Bioinformatics, National Taiwan University, Taipei, Taiwan, ²Interdisciplinary MRI/MRS Lab, Department of Electrical Engineering, National Taiwan University, Taipei, Taiwan, ³Division of Medical Engineering Research, National Health Research Institutes, Zhunan, Taiwan

The purpose of this study is to compare four quantitative susceptibility map (QSM) methods, including threshold, Tikhonov, MEDI and Total Variation (TV), by using numerical analysis on simulated magnetic field maps. To quantitatively evaluate the effects, the noise propagation and contrast-to-noise ratio (CNR) were calculated. Theoretical and numerical analysis were used to be fundamental justification for the result of each approach. The threshold method was fast and simple, but those non-linear regularization methods could potentially reduce the noise propagation. In summary, by quantitative comparison, our results suggested that TV regularization method could be a robust method in susceptibility mapping and potentially helpful to further applications.

Jianlei Liu^1,2, Tian Liu², Shuai Wang³, Keigo Kawaji^2,4, and Yi Wang^2,4
1XiDian University, Xi'an, Shaanxi, China, ²Radiology, Weill Cornell Medical College, New York, New York, United States, ³University of Electronic Science and Technology of China, ⁴Biomedical Engineering, Cornell University, Ithaca, New York, United States

The morphology enabled dipole inversion with L1_norm (MEDI_L1) method use the structural information of magnitude image as prior information to uniquely determine the susceptibility distribution.The method can get desired results with high resolution of scanning. For low resolution of scanning, the MEDI_L1 is sensitive to the prior information. However, there are inconsistencies between the magnitude gradient and the gradient of the true susceptibility distribution.In this case, we use the structural information which comes from both magnitude image and relative difference field (RDF) image as prior information to improve the performance of MEDI_L1.

Weili Zheng¹, Yu-Chung Norman Cheng¹, Saifeng Liu², Helen Nichol³, and E. Mark Haacke^1
1Radiology, Wayne State University, Detroit, MI, United States, ²School of Biomedical Engineering, McMaster University, Hamilton, Ontario, Canada,³Department of Anatomy and Cell Biology, University of Saskatchewan, Saskatoon, Saskatchewan, Canada

Iron is an important endogenous biomarker for many neurological diseases as well as for normal aging. An iron-loaded ferritin phantom was used here to investigate the correlation of iron with susceptibility and the other commonly used iron predictor, R2* (1/T2*). It was found susceptibility mapping predicts iron more reliably than does R2*. Ferritin gelatin phantoms may are a feasible model for human brain iron susceptibility studies. The effect of myelin and chemical exchange may not be negligible when predicting iron using susceptibility mapping and this needs to be further explored in order to accurately predict ferritin iron concentrations in vivo.

Christoph Leussler¹, Philipp Karkowski¹, and Ulrich Katscher^1
1Philips Research Europe, Hamburg, 22457, Germany

Thermal energy is used to treat tumors in liver, kidney and other body organs. Radiofrequency (RF) ablation, microwave ablation, and hyperthermia therapy use the penetration and absorption of electromagnetic waves, which depend on tissue conductivity. The conductivity of tissue depends on frequency and temperature. Relative changes of conductivity of the lesion during treatment with RF energy depend on biochemical changes and on tissue temperature. In this study, we report on the temperature-related change of conductivity, which will need to be distinguished and separated during treatment from biochemical conductivity changes during treatment.

Yi-Cheng Hsu¹, I-Liang Chern¹, and Fa-Hsuan Lin^2,3
1National Taiwan University, Taipei, Taiwan, ²National Taiwan University, Taiwan, ³Massachusetts General Hospital, United States

Using one single transmitter and quadratic magnetic fields, we propose the spatially selective RF quadratic field excitation (QEF) to 1) efficiently excite small FOV and 2) to generate a distribution of Mxy to compensate the B1 inhomogeneity in high field MRI. Specifically, compared to using linear gradients and 25 spokes, QFE is more than 2 times faster to complete the 1/3 FOV excitation with the improved accuracy of the excitation profile by 2.2-fold. In addition, QFE is 2.5 times faster than fast-Kz to excite a slice with in-plane circular symmetric flip angle distribution complementary to high field B1 inhomogeneity.

Priti Balchandani¹, Daniel M. Spielman¹, and John M. Pauly^2
1Radiology, Stanford University, Stanford, California, United States, ²Electrical Engineering, Stanford Universiy, Stanford, California, United States

An improved Slice-selective Tunable-flip AdiaBatic Low peak-power Excitation (STABLE) pulse with shorter duration and increased off-resonance immunity was designed. The new pulse, STABLE-2, utilizes a more uniform spectral pulse envelope generated using the adiabatic SLR method for pulse design. The improved pulse properties for STABLE-2 make it suitable for use in more pulse sequences and at higher field strengths. The pulse performance was validated in a phantom and in vivo.

Rainer Schneider^1,2, Dieter Ritter¹, Jens Haueisen², and Josef Pfeuffer^1
1Siemens Healthcare, Erlangen, Germany, ²Institute of Biomedical Engineering and Informatics, TU Ilmenau, Ilmenau, Germany

Echo-planar imaging with a reduced FOV has been often used to increase imaging speed or acquire high-resolution images. However, to date nearly all of the studies use analytically designed pulses neglecting off-resonance effects and B1 inhomogeneity. In this work a 2DRF pulse optimization framework is proposed, which takes individual B0 and B1 maps into account. In addition, a novel trajectory design incorporating off-resonance information is introduced to further tackle geometric distortions. The proposed methods are evaluated with simulated and experimental data. It is shown that optimized 2DRF pulses outperform analytically designed pulses in regard to the overall excitation.

Rene Gumbrecht^1,2, and Hans-Peter Fautz^1
1Siemens Healthcare, Erlangen, Germany, ²Department of Physics, University of Erlangen, Erlangen, Germany

The power of 2D selective RF pulses is to excite any spatial pattern within the field of view. However, the length of these pulses causes artifacts that especially affect background suppression. It can be observed that conventional 1D slice selective RF pulses have a superior sharpness and background suppression compared to 2D selective pulses. The goal of this study is to transfer these qualities of a 1D slice selective RF pulse to a 2D selective excitation

Alessandro Sbrizzi¹, Cornelis A van den Berg², Peter R Luijten³, Jan J Lagendijk³, and Hans Hoogduin^3
1Imaging Division, UMC Utrecht, Utrecht, Utrecht, Netherlands, ²UMC Utrecht, Netherlands, ³UMC Utrecht

In this work, we introduce the Tip Angle Doubling (TAD) principle, which makes possible to use small tip angle design methods to achieve large tip angles at no extra computation costs. Bloch equation simulations show how the method works in practice. The methods allows fast large tip angle pulse design and is applicable to multi transmit systems with B1+ inhomogeneities correction and (local) SAR optimization.

Tingting Shao¹, Ling Xia¹, Feng Liu², and Stuart Crozier^2
1Biomedical Engineering, Zhejiang University, Hangzhou, Zhejiang, China, ²School of Information Technology & Electrical Engineering, University of Queensland, Brisbane, Australia

This work presents a novel approach for designing RF pulses to achieve excellent whole-brain excitation homogeneity in high field MRI. Based on the parallel transmission technology, an optimized 3D tailored RF (TRF) pulse has been proposed to account for the severe RF(B1) inhomogeneity at 11.7T. The pulse is designed with an adaptive stack-spiral trajectory tailored according to the trajectory container, whose interior is mostly responsible for the desired excitation pattern. An iterative RF pulse design method is employed to ensure the excitation accuracy. Test simulations show that the proposed scheme provides homogeneous excitation over the entire brain volume in spite of the inhomogeneous field and insufficient longitudinal coverage.

Ulrich Katscher¹, Hanno Homann¹, and Peter Börnert^1
1Philips Research Europe-Hamburg, Hamburg, Germany

Transmit SENSE is usually applied to improve 2D or 3D RF pulses. This study applies Transmit SENSE to 1D pulses, possible in case of large B1 variations across the slice or slab to be excited. Typically, such large B1 variations are found across the slabs excited for 3D imaging or for REST. 1D Transmit SENSE can improve excitation profile and particularly RF power / SAR behavior. The resulting pulses have the same duration as standard pulses, and can easily be incorporated in standard sequences maintaining sequence timing. The approach was tested using synthetic and realistic coil sensitivity profiles.

Rainer Schneider^1,2, Dieter Ritter¹, Jens Haueisen², and Josef Pfeuffer^1
1Siemens Healthcare, Erlangen, Germany, ²Institute of Biomedical Engineering and Informatics, TU Ilmenau, Ilmenau, Germany

Parallel transmission (pTx) has been shown to enable B1 and B0 inhomogeneity mitigating spatially selective pulses. The majority of studies uses multichannel spatially selective pulses based on a spiral trajectory design and analyze the potential of pulse acceleration due to the additional rf transmit channels. Here, spatially-selective pulses based on a rectilinear EPI trajectory design (2DRF) are successfully implemented on a parallel transmission system. Further, the differences in excitation performance compared to one channel pulses are analyzed and demonstrated in simulations and phantom experiments. In conclusion, a second rf transmit channel significantly increases the excitation accuracy of optimized 2DRF pulses.

Rene Gumbrecht^1,2, and Hans-Peter Fautz^1
1Siemens Healthcare, Erlangen, Germany, ²Department of Physics, University of Erlangen, Erlangen, Germany

For ultra-high field strengths of 7T and above and especially for parallel transmission, tissue heating due to RF field exposure is a dominant limiting factor for high performance human imaging. However, most current pulse design methods use a constraint on forward RF energy to find a trade-off between excitation quality and RF energy deposition. The goal of this study is to optimize the flip-angle distribution of non-linear high flip-angle RF pulses, such as composite pulses for parallel transmission, and to minimize the peak local RF energy at the same time in one run of a non-linear solver.

Davut I. Mahcicek¹, Haldun Ozgur Bayindir², Taner Demir¹, and Ergin Atalar^3
1National Magnetic Resonance Research Center (UMRAM), Bilkent University, Ankara, Turkey, ²Department of Mathematics, Middle East Technical University, Ankara, Turkey, ³National Magnetic Resonance Research Center (UMRAM), Bilkent University, Department of Electrical and Electronics Engineering, Ankara, Turkey

Specific Absorption Rate (SAR) and B1 inhomoeneity are two main problems of high field Magnetic Resonance Imaging. Multi-channel transmit systems may be used to fight againts inhomogeneity problems. Variable Rate Selective Excitation (VERSE) is an useful method to deal with SAR problems. Application of VERSE method to multi-channel transmit system gives chance to design SAR efficient RF pulses with minimum B1 inhomogeneity. To obtain maximum SAR reduction with VERSE method RF pulse design should be optimized.

Mohammad Mehdi Khalighi¹, Priscilla Chan², and Brian K Rutt^3
1Global Applied Science Lab, GE Healthcare, Menlo Park, California, United States, ²Electrical Engineering Department, Stanford University, Stanford, California, United States, ³Radiology Deaprtment, Stanford University, Stanford, California, United States

Present methods of measuring SAR are insufficient to verify safety in MR scans at high field especially with the use of parallel transmit. An FDTD-modeling software package with realistic body models was used to model SAR using either quadrature (qTx) or 2-channel parallel (pTx) transmit with a birdcage head coil. These numerical simulations were validated by comparison to experimentally-measured in-vivo B₁⁺ fields. SAR and B₁⁺ uniformity were compared between qTx and different pTx pulses; results show that pTx pulses with greater number of spokes create better uniformity and lower average SAR with similar SAR hot spot distribution compared to qTx.

Weiran Deng¹, Benedikt Poser¹, and V Andrew Stenger^1
1Department of Medicine, University of Hawaii JABSOM, Honolulu, HI, United States

Optimal Control Joint Design of Large-Tip-Angle RF Pulses and Gradient Waveforms for Parallel Transmission

Dong-Hyun Kim¹, Sung-Min Gho¹, Narae Choi¹, and Chunlei Liu^2
1Electrical and Electronic Engineering, Yonsei University, Sinchon dong, Seoul, Korea, ²Brain Imaging and Analysis Center, Duke University, Durham, NC, United States

MRI has shown to be able to quantify both magnetic and electrical property. Current studies normally perform one of the above methods individually. Being able to simultaneously quantify both susceptibility () and conductivity () mapping can be useful since misregistration can be alleviated due to separate measurements. Here, we introduce a simultaneous susceptibility and conductivity quantification method. Using a multiecho GRE sequence, the phase of the spins at TE=0 can be retrieved providing conductivity information while the subsequent phase evolution can be used for susceptibility mapping.

Hyung Joong Kim¹, Zijun Meng¹, Saurav ZK Sajib¹, Woo Chul Jeong¹, Young Tae Kim¹, Rosalind J Sadleir², and Eung Je Woo^1
1Biomedical Engineering, Kyung Hee University, Yongin, Gyeonggi, Korea, ²University of Florida, Gainesville, Florida, United States

Imaging of cell membrane conductivity change may provide a truly direct method of locating neural activity compared to fMRI. The advantage of MREIT as a basis for direct neural activity imaging is that the imaged quantity, membrane conductivity, is a scalar and therefore is not subject to cancellation errors such as those inherent in neural current imaging techniques. In this work, we use a realistic head model to compute signal levels produced as a consequence of a predicted 5% conductivity change occurring within gray matter. We reconstructed conductivity images, showing that these small conductivity differences can be detected and imaged.

Astrid L.H.M.W. van Lier¹, Ulrich Katscher², Alexander Raaijmakers¹, and Cornelis A.T. van den Berg^1
1Radiotherapy, UMC Utrecht, Utrecht, Netherlands, ²Philips Research Europe, Hamburg, Germany

Recently, it was shown that dielectric properties can be mapped using MRI. Electrical properties mapping relies on measurements of the B₁⁺ amplitude and phase. This phase, however, cannot be measured directly; therefore, the assumption that the transceive phase is twice the B₁⁺ phase is used in a transceiver setup. This assumption is acceptable at low field strengths, however, leads to significant reconstruction errors at 7T. Here we show, that the wave-number (|k|) is less susceptible to these errors, and can be mapped more reliably than the permittivity and conductivity at 7T.

Astrid L.H.M.W. van Lier¹, Cornelis A.T. van den Berg¹, and Ulrich Katscher^2
1Radiotherapy, UMC Utrecht, Utrecht, Netherlands, ²Philips Research Europe, Hamburg, Germany

The complex permittivity (*) of biological tissues depends on their biochemical composition and the applied frequency. In this study we investigate a method that measures the tissue conductivity in the biologically interesting low frequency (LF,=Hz-kHz) range. A first implementation of this method – LF-EPT– is based on similar reconstruction principles as MR-CDI, but it employs the imaging gradient instead of electrodes to induce (eddy) currents. First experiments show that this method is feasible in phantoms.

E. Balidemaj¹, A. L. van Lier², A.J. Nederveen³, J. Crezee¹, and C.A.T. van den Berg^2
1Radiotherapy, Academic Medical Center, Amsterdam, Netherlands, ²Radiotherapy, UMC Utrecht, Utrecht, Netherlands, ³Radiology, Academic Medical Center, Amsterdam, Netherlands

Feasibility of EPT in the human pelvic at 3T for use in Hyperthermia Treatment Planning, SAR determination or tumour detection and characterization.

Selaka Bandara Bulumulla¹, Seung-Kyun Lee¹, and Teck Beng Desmond Yeo^1
1GE Global Research, Niskayuna, New York, United States

Extracting electrical properties (relative permittivity and conductivity) from B1+ maps is a promising method that has applications in local SAR estimation, RF hyperthermia treatment planning and diagnosis of tissue malignancy. In this work, we compare the two primary calculation methods, Laplacian and Integral based, under the constraint that both methods use a constant number of B1+ samples to calculate permittivity and conductivity. Our results indicate that under the constraint, (a) both methods predict accurate values, (b) Laplacian based method is advantageous for rapid estimates and (c) Integral based method is more robust with noisy B1+ maps.

Emre Kopanoglu^1,2, Yildiray Gokhalk², Ugur Yilmaz^1,2, Volkan Acikel^1,2, and Ergin Atalar^1,2
1Electrical and Electronics Engineering, Bilkent University, Ankara, Turkey, ²UMRAM, Bilkent University, Ankara, Turkey

To correct for B1 inhomogeneity artifacts, multi-dimensional pulses are widely used. However, such pulses increase SAR significantly. In this study, we compare, and combine two SAR reduction methods, the variable rate selective excitation and excitation using nonlinear gradient fields in order to observe their effect on SAR. 5-spoke pulses are designed for both methods to correct for a central brightening inhomogeneity, which is widely seen at the UHF regime. When two methods are used separately, SAR reductions around 75% were obtained whereas when combined, the methods yielded an SAR reduction of 91%.

Emre Kopanoglu^1,2, Ugur Yilmaz^1,2, Burak Akin², Volkan Acikel^1,2, and Ergin Atalar^1,2
1Electrical and Electronics Engineering, Bilkent University, Ankara, Turkey, ²UMRAM, Bilkent University, Ankara, Turkey

For numerous applications, the region of interest is merely a portion of the imaged volume. Conventional one-dimensional RF pulses have relatively low SAR, but they cannot excite local regions and therefore causes long scan times. Although small volumes can be excited using multi-dimensional excitation pulses, such pulses are long and/or have relatively high SAR. It is shown that, using gradient fields with nonlinear variation in space, excitation regions can be confined in three dimensions without increasing pulse duration and SAR. In this abstract we show that a 60% decrease in required phase-encoding steps is obtained while imaging the occipital lobe of the brain. Initial in-vivo results are presented.

Sebastian Schmitter¹, Gregor Adriany¹, Edward J Auerbach¹, Kamil Ugurbil¹, and Pierre-Francois Van de Moortele^1
1Center for Magnetic Resonance Research, University of Minnesota, Minneapolis, MN, United States

At 7T the short transmit B1 (B1+) RF wavelength leads to spatial variations of |B1+| including possible B1+ nodes of fully destructive interference. RF phase shimming using multi-channel TX coils can mitigate this problem. However, a 'good starting-phase set' without B1+ nodes and with reasonably high B1+ efficiency is not necessarily known before starting a scanning session. Yet, such “CP-like” mode is indispensable for whole brain anatomical imaging, B0 mapping and RF power calibration. In this work we present a simple but efficient method to determine a good starting set of “CP-like” RF phases for multi-channel head coils at 7T.

Yuval Zur^1
1GE Healthcare, Tirat Carmel, Israel

Spectral spatial (spsp) RF pulses are used for simultaneous fat suppression and slice selective excitation. Reduction of the minimum slice width to less than 2 mm at high field strength (3T) requires very powerful gradients. The purpose of this work is to design spsp RF pulses with lower gradients, such that thin slices ¡Ü 1.5 mm become feasible. The drawback is that spsp pulses are very sensitive to RF gradient delay and eddy currents. We present a calibration method and design a new spsp RF pulse that is less sensitive to system imperfections. The combination of the two results in robust 1.3 mm spsp RF pulse.

Mirjam Holbach¹, Joerg Lambert², and Dieter Suter^3
1Physics, Technical University Dortmund, Dortmund, Germany, ²Leibniz-Institute for Analytical Sciences – ISAS, Dortmund, Germany, ³Technical University Dortmund, Dortmund, Germany

Selective excitation of metabolite signals in in vivo MRS is important for the correct and robust quantification of the content of some key metabolites, which is of particular relevance in low field strength MRS, where signal overlap and low intensities impede quantification of MRS signals. In this study a Krotov-based optimal control theory approach is used to develop optimized pulse shapes for the selective excitation of individual metabolites. The resulting pulses were investigated by simulation of the corresponding spectra, which were then verified using experimental data. Further simulations of the performance of such pulses show an enhanced robustness to experimental imperfections compared to standard pulses.

Mathies Breithaupt¹, Moritz Cornelius Berger¹, Ann-Kathrin Homagk¹, Wolfhard Semmler¹, and Michael Bock^1,2
1Dept. of Medical Physics in Radiology, German Cancer Research Center (DKFZ), Heidelberg, Germany, ²Radiology - Medical Physics, University Hospital Freiburg, Freiburg, Germany

Adiabatic pulses are commonly used to overcome B₁ inhomogeneity which is particularly important at ultra-high magnetic fields due to standing wave effects. In slice-selective adiabatic pulses, the frequency sweep causes B_eff,z-dependent off-resonances leading to an asymmetric evolution of the slice profile. Here, we present a detailed temporal analysis of the magnetization manipulation during slice-selective adiabatic pulses by comparing measured and simulated data.

Martin Haas¹, Jeff Snyder¹, Stefanie Buchenau¹, Denis Kokorin^1,2, Johannes T. Schneider^1,3, Peter Ullmann³, Jürgen Hennig¹, and Maxim Zaitsev^1
1Department of Radiology, Medical Physics, University Medical Center Freiburg, Freiburg, Germany, ²International Tomography Center, Novosibirsk, Russian Federation, ³Bruker BioSpin MRI GmbH, Ettlingen, Germany

Spatially selective excitation of an arbitrary three-dimensional target pattern is demonstrated in vivo in the head of a volunteer at 3T for the first time. A parallel transmit system with eight independent RF channels is used in combination with a slew-optimized single-shot 3D shells trajectory. The good excitation fidelity allows for the reduction of the field of view below the extent of the head and consequently for a shorter acquisition with higher spatial resolution.

Martin Haas¹, Jeff Snyder¹, Jürgen Hennig¹, and Maxim Zaitsev^1
1Department of Radiology, Medical Physics, University Medical Center Freiburg, Freiburg, Germany

The duration of multi-dimensionally spatial selective excitation (SSE) RF pulses usually implies high excitation fidelity of a given target pattern only for a narrow range of frequencies. Segmentation of an SSE pulse across several repetitions results in short RF pulses and broadband excitation of the target pattern in the summed signal. In this work, a generalized nonlinear large tip angle RF optimization algorithm is presented which jointly optimizes the RF segments and thus extends previous approaches using small tip angles or segment-wise large tip angle optimization.

Laura Sacolick¹, Mika W. Vogel¹, William A. Grissom², Guido Kudielka¹, Theodor Vetter¹, and Ileana Hancu^3
1GE Global Research, Garching b. Munchen, Bayern, Germany, ²Biomedical Engineering, Vanderbilt University, Nashville, TN, United States, ³GE Global Research, Niskayuna, NY, United States

B1 homogeneity and relative SAR is compared in the 3T breast for several configurations of parallel transmit system. Active RF shimming with 1, 2, and 8 transmit channels is compared to passive shimming- where one off-tuned coil element is placed around the right breast, weakly coupling to the main transmit field. B1 homogeneity improves with increasing channel count, with the passive transmit coil giving B1 homogeneity similar to a two-channel system, and significantly lower SAR than any active parallel transmit configuration.

Cem Murat Deniz^1,2, Ryan Brown¹, Riccardo Lattanzi^1,2, Leeor Alon^1,2, Daniel K. Sodickson^1,2, and Yudong Zhu^1,2
1Department of Radiology, Bernard and Irene Schwartz Center for Biomedical Imaging, New York University School of Medicine, New York, NY, United States, ²Sackler Institute of Graduate Biomedical Sciences, New York University School of Medicine, New York, NY, United States

Radiofrequency shimming with multiple channel excitation has been proposed to increase the transverse magnetic field uniformity and reduce specific absorption rate at high magnetic field strengths (≥ 7 Tesla), where wavelength effects can make traditional single channel volume coils unsuitable for transmission. In the case of deep anatomic regions and power-demanding pulse sequences, optimization of transmit efficiency may be a more critical requirement than homogeneity. This work introduces a new method to maximize transmit efficiency using multiple channel excitation and radiofrequency shimming. Shimming weights are calculated in order to obtain the highest possible transverse magnetic field using the lowest possible net radiofrequency power deposition into the subject.

Hai Zheng¹, Tiejun Zhao², Yongxian Qian³, Claudiu Schirda³, Tamer Ibrahim^1,3, and Fernando Boada^1,3
1Bioengineering, University of Pittsburgh, Pittsburgh, Pennsylvania, United States, ²Siemens Medical Solutions USA, Pittsburgh, Pennsylvania, United States, ³Radiology, University of Pittsburgh, Pittsburgh, Pennsylvania, United States

T2* weighted fMRI in ultra high field is hampered by susceptibility-induced signal loss near air/tissue interfaces of the brain. We demonstrate a robust RF pulse design, based on parallel transmission (PTX) and 3D tailored RF (3DTRF), that is capable of simultaneously and precisely recovering signal loss at multiple locations. An additional scheme for PTX 3DTRF based on time-interpolation approach, is also presented. All schemes were observed to significantly and reproducibly improve signal recovery across multiple slices in human subjects. Time-interpolation method proved to be more robust at regions with large frequency offsets, albeit at the expense of increased computational demand.

Christopher Sica¹, Sukhoon Oh¹, and Christopher Collins^1
1Radiology, Pennsylvania State University, Hershey, PA, United States

An implementation of an array optimized composite pulse (ACP) for parallel transmit applications is presented. The pulse design utilizes two RF pulses, with independently adjustable magnitude and phase per channel and per pulse. A non-selective version of the ACP pulse is compared with RF shimming and a quadrature drive configuration in a water phantom, and with quadrature drive in the human brain at 3T. The ACP pulse offers improved excitation uniformity relative to RF shimming in phantoms, and significant improvement over a quadrature drive configuration in-vivo. Future potential for a slice-selective version of the ACP pulse is also demonstrated.

Ulrich Katscher¹, Karim Djamshidi¹, Tobias Voigt², Marko Ivancevic³, Hiroyuki Abe⁴, Gillian Newstead⁴, and Jochen Keupp^1
1Philips Research Europe-Hamburg, Hamburg, Germany, ²Philips Research Europe-Aachen, Aachen, Germany, ³Philips Healthcare, Cleveland, United States, ⁴University of Chicago, Chicago, United States

Ex vivo studies exhibited significantly altered electric conductivity of breast tumors, opening the chance to increase the specificity of breast tumor characterization. Conductivity can be measured in vivo using “Electric Properties Tomography” (EPT), which has shown its potential in phantom, volunteer, and initial clinical studies. However, the complex frayed structure of fat and ductile tissue in the breast hampers the straight-forward application of EPT, based on the second derivative of the RF TX phase. In this study, a new EPT reconstruction via fitting local parabolic functions on the TX phase is developed and applied to an example breast tumor.

Tobias Voigt¹, Andreas Schuster², Masaki Ishida², Christian Stehning³, Ulrich Katscher³, Amedeo Chiribiri², Eike Nagel², and Tobias Schaeffter^2
1Philips Research, London, United Kingdom, ²King’s College London, London, United Kingdom, ³Philips Research, Hamburg, Germany

In this work we present the application of EPT conductivity mapping in two isolated perfused pig hearts. Conductivity values of normally perfused heart tissue were compared to values in ischemic regions after a blockade of the left anterior descending artery (LAD). It could be shown that ischemic ventricular tissue is less conductive than healthy myocardium.

Astrid L.H.M.W. van Lier¹, Anja G. van der Kolk², Manon Brundel³, Jeroen Hendrikse², Peter R. Luijten², Jan J.W. Lagendijk¹, and Cornelis A.T. van den Berg^1
1Radiotherapy, UMC Utrecht, Utrecht, Netherlands, ²Radiology, UMC Utrecht, Netherlands, ³Neurology, UMC Utrecht, Netherlands

In ischemic stroke, [Na⁺] increases due to ischemia and subsequent disturbance of ion homeostasis. Recently, electrical properties tomography (EPT) was introduced as a new contrast in MRI. This contrast enables us to visualize the local electrical properties of tissue, being the conductivity and permittivity; the electrical conductivity is related to [Na⁺]. In this study, it was shown that electrical conductivity was locally elevated in the tissue affected by the infarction.

Hyung Joong Kim¹, Saurav ZK Sajib¹, Woo Chul Jeong¹, Young Tae Kim¹, Tong In Oh¹, and Eung Je Woo^1
1Biomedical Engineering, Kyung Hee University, Yongin, Gyeonggi, Korea

Conductivity values of cancerous tissues in the breast are significantly higher than those of surrounding normal tissues, breast MREIT may provide a new noninvasive way of detecting early stage of breast cancer. As a step toward clinical application, we present results of experimental and numerical simulation studies of breast MREIT. From phantom experiments, we evaluated practical amounts of noise in measured Bz and built a realistic three-dimensional model. Simulation results are promising to show that we can detect a cancerous anomaly in the breast while restricting the maximal current density inside the heart below the level of nerve excitation.

Jiaen Liu¹, Xiaotong Zhang¹, and Bin He^1
1Biomedical Engineering, University of Minnesota, Minneapolis, MN, United States

In the present study, we investigate the feasibility of a complex B1 mapping technology using a multi-channel transceiver coil at ultrahigh field for the purpose of imaging electrical properties of realistic human head by computer simulation. Complex B1 mapping and electrical properties reconstruction were carried out on both normal symmetric head and pathological asymmetric head with a tumor at 7 T and 9.4 T. The simulation results show that on realistic asymmetric geometry, the method performs equally well as on symmetric geometry and suggest its practical value for diagnostic purpose and SAR calculation at UHF.

Ulrich Katscher¹, Astrid L.H.M.W. van Lier², Cornelis A.T. van den Berg², and Jochen Keupp^1
1Philips Research Europe-Hamburg, Hamburg, Germany, ²University Medical Center Utrecht, Utrecht, Netherlands

In the framework of “Electric Properties Tomography” (EPT), approximate conductivity imaging is possible by analyzing the B1 phase, assuming constant B1 amplitude. The more this assumption is violated, the less accurate the reconstructed conductivity. This study analyzes the influence of modifying the B1 amplitude by parallel RF transmission on the precision of phase-based EPT. It turns out that phase-based EPT benefits significantly from optimizing B1 amplitude homogeneity via with RF shimming.