Traditional Poster Session - Pulse Sequences & Reconstruction B
  Off-Resonance Artifact Correction 2405-2429
  Metal Artifact Correction 2430-2437
  Artifacts & Correction: Misc. 2438-2456
  Motion Correction 2457-2472
  Intensity/Distortion Correction 2473-2479
  Water/Fat Imaging 2480-2492
  Pulse Sequence 2493-2503
  B1/B0 Mapping 2504-2513
  Elastography 2514-2528
  ESR 2529-2531
  Electromagnetic Tissue Property Mapping 2532-2534
  Image Reconstruction 2535-2549
  Computational Implementation & Tools 2550-2559
  Image Analysis 2560-2570
  Evaluation 2571-2574

Off-Resonance Artifact Correction

Click on to view the abstract pdf. Click on to view the poster (Not all posters are available for viewing.)

Wednesday 9 May 2012
Exhibition Hall  16:00 - 18:00

2405.   Fast, Variable System Delay Correction for Spiral Trajectories
Payal Sharad Bhavsar1, Nicholas Ryan Zwart1, and James Grant Pipe1
1Neuroimaging Research, Barrow Neurological Institute, Phoenix, Arizona, United States

Spiral encoding methods may exhibit image artifacts due to system delays and eddy current. This papers extends a previous approach to estimate time-varying system delays for stack-of-spirals based trajectories, reducing the computational time by a factor of 30. The proposed method is very fast, easy to implement, includes gradient coupling effects, is robust to off-resonance, and estimates delays for each gradient channel over the data acquisition time.

2406.   k-Space Shift Correction Using an Alternating Gradient Readout Acquisition for Improved Radial Fat-Water MRI
Xuelin Cui1, John C. Gore2, and E. Brian Welch2
1Biomedical Engineering, Vanderbilt university, Nashville, TN, United States, 2Radiology and Radiological Sciences, Vanderbilt University, Nashville, TN, United States

A primary challenge associated with the radial MRI sampling trajectory is the imperfect alignment of each sampled diagonal with the center of k-space. Gradient imperfections and timing delay errors cause the sampled trajectory to shift from the intended trajectory. Methods exist to align miscentered k-space samples for alternating radial readouts to the center of k-space by comparing neighboring radial lines that have nearly opposite readout direction. In this work, an alternating readout direction phase correction algorithm for radial MRI is tested on a conventional single gradient echo acquisition and a multiple fast field echo (mFFE) acquisition used for fat-water imaging.

2407.   Evaluation of chemical shift based fat suppression using 2DRF pulses
Rainer Schneider1,2, Thorsten Speckner1, Jens Haueisen2, and Josef Pfeuffer1
1Siemens Healthcare, Erlangen, Germany, 2Institute of Biomedical Engineering and Informatics, TU Ilmenau, Ilmenau, Germany

Custom-designed spatial spectral pulses (SPSP) have been shown to suppress fat signal effectively by taking advantage of chemical shift along the fast direction. However, the proper design of such SPSP pulses depends on many variables and their performance was never evaluated so far. In this work, a simplified implementation of this approach and a detailed performance evaluation is done. Reduced FOV diffusion-weighted images subject are acquired in phantom and human and compared to two other state-of-the-art fat suppression techniques. In conclusion, the proposed approach demonstrated here can offer the best compromise regarding residual fat signal and acquisition speed.

2408.   Through-Plane Chemical Shift Correction in Rf-Power Reduced Sequences at High Field Strength
Mathias Nittka1, Vladimir Jellus1, and Lars Lauer1
1Siemens Healthcare, Erlangen, Germany

Application of low power, low bandwidth rf-pulses in slice selective sequences at high field strength may lead to severe through-plane chemical shift artifacts of fat signal. The presented methods uses a fat/water separating DIXON turbo spin echo sequence to generate through-plane chem. shift corrected images.

2409.   Chemical-shift selective multislice imaging using the gradient reversal technique
Yeji Han1, Yoojin Lee1, ChangHyun Oh1, and HyunWook Park1
1Department of Electrical Engineering, Korea Advanced Institute of Science and Technology, Daejeon, Korea

The chemical shift phenomenon can produce severe artifacts of geometric shift or blurring in MRI. Instead of treating the chemical shift phenomenon as a source of artifact, many imaging technologies, including the gradient reversal (GR) technique, have been developed to acquire separate sets of images from tissues of different resonance frequencies. Although the GR method has not been widely used, recent reports have confirmed the effectiveness of this method for fat suppression at 3.0T and higher field. In this abstract, we expand the concept of the GR technique to acquire multislice chemical-shift selective images.

2410.   An Investigation of Fundamental Pitfall in SSGR
Axel Hartwig1, and Stefan Skare1
1Clinical Neurosience, Karolinska Institute, Stockholm, Sweden

SSGR is a time efficient and robust fat suppression technique that achieves fat suppression by displacing fat during slice-selection. Fat suppression is achieved by displacing excitation and refocusing pulses in opposite directions. This work shows how, depending on slice-selection direction, displaced excitations and refocusing pulses may interact and form an unwanted echo during acquisition of latter slices. The birth of the echo is investigated and two methods to eliminate it are discussed.

2411.   Global and spatially varying B0 drifts due to gradient system heating
Karl-Heinz Herrmann1, Martin Krämer1, and Jürgen R Reichenbach1
1IDIR I, Medical Physics Group, Jena University Hospital, Jena, Germany

EPI is sensitive to B0 drifts and especially if the phase encoding direction changes as in PROPELLER EPI, images can suffer from severe blur if B0 shifts are not compensated. However, heating up the gradients during DTI or fMRI scans causes not only global drifts of B0, which are easy to compensate once known, but also spatially varying B0changes of up to 30Hz.

2412.   An automatic real-time feedback calibration of RF phase cycling by off-resonance weighted imaging sequence
Yu-Wei Tang1, Teng-Yi Huang1, Ming-Long Wu2, and Cheng-Wen Ko3
1Electrical Engineering, National Taiwan University of Science and Technology, Taipei, Taiwan, 2Computer Science and Information Engineering, National Cheng Kung University, Tainan, Taiwan, 3Computer Science and Engineering, National Sun Yat-sen University, Kaohsiung, Taiwan

The adjustment of system frequency or RF phase-cycling has demonstrated able to avoid dark-band artifact of bSSFP imaging. The optimal frequency can be detected by a sweep scan of bSSFP imaging. In this study, we proposed a new sequence named off-resonance weighted imaging (OWI) for accomplishing the frequency calibration. For higher accuracy and efficiency, an automatic real-time feedback optimization system was introduced. According to the results, the automatic real-time feedback system shows able to calibrate frequency accurately and rapidly. It is potential to serve as a pre-scan for bSSFP imaging or blood oxygenation sensitive SSFP (BOSS) fMRI.

2413.   Deblurring of spiral images in the presence of fat and rapidly varying B0
Eric Aboussouan1, and James G. Pipe1
1Neuroimaging Research, Barrow Neurological Institute, Phoenix, Az, United States

Spiral imaging with long acquisition windows is efficient but highly susceptible to off-resonance blurring artifacts. Iterative approaches to deblurring in the presence of fat and/or rapidly varying field are presented. Deblurring results of real and synthesized images are compared favorably to results of a common non-iterative deblurring approach.

2414.   Off-resonance Correction for 3D Cones Imaging Using Multifrequency Interpolation
Wendy W Ni1,2, Holden H Wu1,3, and Dwight G Nishimura1
1Department of Electrical Engineering, Stanford University, Stanford, California, United States, 2Department of Radiology, Stanford University, Stanford, California, United States, 3Department of Cardiovascular Medicine, Stanford University, Stanford, California, United States

In this study, we extended the multifrequency interpolation (MFI) algorithm to three dimensions (3D) and performed off-resonance correction for the 3D cones trajectory. Using an emulated low-resolution field map of isotropic 5 mm resolution, we have demonstrated the feasibility of MFI in both phantom and in vivo data of isotropic 1 mm resolution, achieving significant improvements in image contrast and detail.

2415.   A Simple Acquisition Strategy to Remove Off-Resonance Blurring in Spiral Imaging
Samuel Fielden1, Xue Feng1, and Craig Meyer1,2
1Biomedical Engineering, University of Virginia, Charlottesville, Virginia, United States, 2Radiology, University of Virginia

Spiral imaging has been hampered by off-resonance blurring since its inception. Here we show that redundant spiral-in/spiral-out trajectories naturally correct for the most severe off-resonance artifacts in a simple way.

2416.   An XS-Guided Solution for bSSFP Banding Artifact Correction with Reduced Scan Time
Michael Nicholas Hoff1, and Qing-San Xiang1,2
1Physics, University of British Columbia, Vancouver, British Columbia, Canada, 2Radiology, University of British Columbia, Vancouver, British Columbia, Canada

bSSFP imaging is plagued by banding and general signal modulation stemming from off-resonance-induced phase accumulation. The geometric cross-solution (XS) eliminated all these artifacts using four phase cycled datasets; here a weighted solution of two phase cycled datasets coupled with weighting guide data is proposed. Weights are found by minimizing the regional differential energy of the solution from a low resolution XS guide. Weighting guided by 25% of the XS k-space phase encodes yields excellent signal demodulation, and uses a scan time comparable to the acquisition of only 2.5 regular bSSFP images.

2417.   Signal Demodulation of bSSFP Imaging with a Two-Point Algebraically Weighted Solution
Michael Nicholas Hoff1, and Qing-San Xiang1,2
1Physics, University of British Columbia, Vancouver, British Columbia, Canada, 2Radiology, University of British Columbia, Vancouver, British Columbia, Canada

Banding and general signal modulation disturb the diagnostic viability of bSSFP imaging. Previous algebraic and geometric techniques have shown that complete signal demodulation is possible with four phase cycled bSSFP images. Here a two-image algebraic solution is employed to demodulate signal and to guide an improved Algebraically Weighted Solution (AWS). The AWS uses regional least squares processing to reduce residual banding. Relative to a gold standard reference image, the AWS exhibits 60% or less error than a two-image complex sum in all scenarios tested. AWS represents a viable bSSFP demodulation solution which is especially valuable for time-sensitive imaging.

2418.   View Angle Tilting for Distortion Compensated EPI: Effects of RF Pulse Width on Image Blurring and Slice Profile
Cheng Li1, Felix Wehrli1, and Hee Kwon Song1
1Radiology, University of Pennsylvania, Philadelphia, PA, United States

This work describes the implementation of the VAT methodology for EPI sequences as a means to eliminate in-plane distortions. Results demonstrate that a relatively long RF pulse, on the order of the duration of the total EPI readout period, is necessary to minimize image blurring. However, the reduced in-plane blurring is achieved at the cost of increased slice profile distortion.

2419.   Combination of VAT and Z-shimming in Echo Planar Imaging for Distortion Correction and Signal Recovery
Sinyeob Ahn1, and Xiaoping Hu1
1Biomedical Engineering, Georgia Institute of Technology/Emory University, Atlanta, GA, United States

Image distortion and signal loss are detrimental artifacts caused by field inhomogeneity in gradient-echo EPI (GE-EPI). This work describes the combination of view angle tilting (VAT) with z-shimming to address both artifacts in GE-EPI. VAT was implemented with parallel imaging to make it practical for EPI. VAT corrects in-plane distortion in the phase-encoding direction while z-shimming reduces signal loss. The combined sequence was used to generate images with improved quality, particularly in the frontal and inferior orbito-frontal regions of human brain.

2420.   Comparison of EPI geometric distortion correction using Field mapping and forward/reverse phase encoding directional EPI scans
Wanyong Shin1, Erik Beall1, Ken Sakaie1, and Mark Lowe1
1Radiology, Cleveland Clinic, Cleveland, OH, United States

Since EPI images with forward (for.)/reverse (rev.) phase encoding (PE) directions provide voxel shifts in opposite directions, several approaches have been proposed to calculate voxel displacement (VD) map in PE direction using the two, in short, 2PE method here. Since both field map (FM) and 2PE methods generate VD maps, the direct voxel-wise comparison of VD maps between FM and 2PE method would provide the comprehensive insight of pros/cons of 2PE. In this study, we compare VD maps from FM and 2 PE methods, and demonstrate their performance and limitation.

2421.   Distortion correction using the susceptibility based field map estimation in echo planar imaging reconstruction
Hiroyuki Takeda1, and Boklye Kim1
1Radiology, University of Michigan, Ann Arbor, MI, United States

Reconstructed EPI images suffer from geometric distortion often due to the magnetic field inhomogeneity and significantly undermine the performance of activity analyses. In this work, we focus on a fundamental approach of the susceptibility-induced magnetic field inhomogeneity map to retrospectively recover the spin density map for EPI image reconstruction. By modeling the acquisition process, we identify the effect of the field inhomogeneity to the EPI reconstruction, and then obtain distortion-free density images using a regularized least-square method. Result shows that our approach works effectively for recovering the original undistorted image with an accurate estimate of the field map.

2422.   Application of k-space energy spectrum analysis to artifact correction in PROPELLER EPI
Zhengguo Tan1, and Nan-kuei Chen2
1Department of Biomedical Engineering, Duke University, Durham, North Carolina, United States, 2Brain Imaging and Analysis Center, Duke University, Durham, NC - North Carolina, United States

The sequence that integrates PROPELLER and EPI has proven valuable for studies that require high throughput and tolerance to subject motion. However, the PROPELLER-EPI quality is usually degraded by distortions resulting from background susceptibility field gradients. Although the distortion can potentially be corrected through B0 mapping, it is not easy to use a conventional B0 mapping procedure to quantify the field inhomogeneities that may change from blade to blade due to subject motion. The goal of this study is to characterize the B0 field inhomogeneity patterns directly from each blade of the PROPELLER-EPI data using the k-space energy spectrum analysis.

2423.   Identification of over-estimated diffusion coefficients obtained with very high b-values in diffusion MRS
Vaclav Brandejsky1, Roland Kreis1, and Chris Boesch1
1Depts Clinical Research and Radiology, University of Bern, Bern, Switzerland

Eddy currents are prohibitive for high b-value diffusion measurements in vivo. A method for quantitation of the eddy current effects on balanced diffusion gradient pairs is presented. Both the remaining eddy currents and the signal drop due to suboptimal rephasing can be assessed.

2424.   Correction of in-plane intra-voxel dephasing effects in gradient echo images.
Peter van Gelderen1, Jacco A de Zwart1, and Jeff H Duyn1
1Advanced MRI Section, LFMI, NINDS, National Institutes of Health, Bethesda, MD, United States

Macroscopic B0 field gradients may result in signal dropouts in gradient echo images obtained at high field, and obscure the information reflecting local tissue structure and composition. While several methods exist to correct for these dropouts, we demonstrate that improved correction is possible with a simple method that better takes into account the effects of in-plane dephasing effects.

2425.   Correction of gradient echo images for first and second order macroscopic signal dephasing using phase derivative mapping
Hendrik de Leeuw1, and Chris J.G. Bakker1
1Image Sciences Institute, Utrecht, Utrecht, Netherlands

Gradient Echo MRI is frequently applied for its inherent sensitivity to mesoscopic field inhomogeneities, which reflect local tissue properties. Unfortunately, macroscopic field inhomogeneities, such as inhomogeneities invoked by air cavities, interfere with the mesoscopic effects. Compensation methods typically lack flexibility (need precise, object dependent, scan adjustments), lengthen data acquisition (additional reconstructions, acquisitions) or is not generally applicable (multi-echo only). Herein a generally applicable (single, multi-echo) post-processing method is presented which compensates for first and second order macroscopic phase incoherences, while preserving mesoscopic effects. The method is demonstrated on a phantom containing mesoscopic field disturbers and the head of a volunteer.

2426.   A comprehensive Gaussian Process framework for correcting distortions and movements in diffusion images
Jesper L. R. Andersson1, Junquian Xu2, Essa Yacoub2, Edward Auerbach2, Steen Moeller2, and Kamil Ugurbil2
1FMRIB, Oxford, Oxfordshire, United Kingdom, 2Center for Magnetic Resonance Research (CMRR), University of Minnesota, Minnesota, Minneapolis, United States

Registration based approaches to correcting for eddy current distortions and movements are complicated by the images containing different information. Possible solutions to this is i) To collect pairs of dwis with opposite polarity diffusion gradients, ii) To collect pairs of dwis with opposite polarity phase-encoding or iii) To register observed dwis to model based predictions. We present a method that include any and all of those sources of information depending on how the data was collected. It utilises Gaussian Processes to predict dwis and to predict eddy currents for one acquisition given other acquisitions.

2427.   Segmented Trajectory Correction for Non-Cartesian Imaging with Ramp-up Acquizaiton
Zhang Qiong1, Weng De he1, and Liu Wei1
1Siemens, Shen Zhen, Guang Dong, China

Eddy currents can be a significant impediment for application of oblique non-Cartesian techniques In permanent MRI system, In our work, eddy current distorted k-space was divided into two segments, and an effective approach to align peripheral k-space segment has been presented, which use physical gradient delays to counteract trajectory errors accumulated in oblique period. After that, eddy current induced trajectory errors were restricted in center non Lineal area, and the aligned peripheral k-space segment shows isotropic density, which makes the entire image reconstruction more stable and further saving time.

2428.   Static Magnetic Field Inhomogeneity Correction of Radial MRI Using an Alternating Gradient Readout Acquisition
Xuelin Cui1, John C. Gore2,3, and E. Brian Welch2,3
1Biomedical Engineering, Vanderbilt university, Nashville, TN, United States, 2Radiology and Radiological Sciences, Vanderbilt University, Nashville, TN, United States, 3Institute of Imaging Science, Vanderbilt University, Nashville, TN, United States

Static magnetic field (B0) inhomogeneity is one of the major causes of artifacts in MRI. Although numerous correction methods have been implemented for Cartesian k-space, relatively little investigation has been conducted for radial MRI. B0 inhomogeneity can cause signal displacement in image space along the frequency encode direction. In this work, a previously described reverse gradient method (Fitzpatrick JM et al., IEEE TMI, 1992) is applied to correct radial MRI data without prior knowledge of the B0 inhomogeneity.

2429.   Field-Corrected MP-SWIRLS for 3D Isotropic High-Resolution T1-weighted Brain Imaging
Yunhong Shu1, Joshua D Trzasko2, Armando Manduca2, and Matt Bernstein2
1Radiology, Mayo Clinic, Rochester, MN, United States, 2Mayo Clinic

We demonstrate that off-resonance correction (ORC) can help mitigate image blurring that arises from the long readout of the MP-SWIRLS technique. An in vivo experiment suggests that MP-SWIRLS with ORC can provide robust and efficient 3D isotropic high-resolution T1-weighted brain imaging.
Traditional Poster Session - Pulse Sequences & Reconstruction B

Metal Artifact Correction

Click on to view the abstract pdf. Click on to view the poster (Not all posters are available for viewing.)

Wednesday 9 May 2012
Exhibition Hall  16:00 - 18:00

2430.   Pure phase encode MRI in the vicinity of metal structures
Hui Han1,2, Igor Mastikhin1, Bryce MacMillan1, and Bruce Balcom1
1MRI Centre, Department of Physics, University of New Brunswick, Fredericton, NB, Canada, 2Brain Imaging and Analysis Center, Duke University Medical Center, Durham, NC, United States

The presented work shows the robustness of pure phase encode MRI to B0 field inhomogeneity by showing 3D SPRITE phantom images of samples with significant metal content. The paper also outlines the robustness of SPRITE to the gradient eddy currents induced by switching pulsed field gradients, which has enabled new high pressure MRI study inside metallic containers at variable temperature and pressure condition. An eddy current self-compensated SPRITE sequence allows us to acquire geometrically correct images inside highly conductive cylindrical metal containers.

2431.   Spectrally Resolved Fully Phase-Encoded 3D Fast Spin-Echo for Metal Artifact Reduction and Spectroscopic Imaging
Nathan S Artz1, Diego Hernando1, Valentina Taviani1, Kevin Johnson2, Jean H Brittain3, and Scott B Reeder1,2
1Department of Radiology, University of Wisconsin, Madison, WI, United States, 2Department of Medical Physics, University of Wisconsin, Madison, WI, United States, 3Global Applied Science Laboratory, GE Healthcare, Madison, WI, United States

The purpose of this work was to develop a spectrally-resolved, fully phase-encoded 3D fast spin-echo technique that eliminates all frequency-encoding-related shift artifacts and permits spectral data sampling at each spin-echo. A sphere containing gadolinium was placed in water and imaged with both this method and conventional 3D-FSE. Signal modeling produced distortion-free proton density and B0 maps with the proposed method, in contrast to the conventional 3D-FSE images. Dipole effects were also visible both in the estimated B0 map and the spectroscopic images. Lastly, a hip prosthesis was well visualized without distortion while conventional 3D-FSE demonstrated severe distortion artifacts.

2432.   MR Imaging near metallic implants using Selective Multi-Acquisition with Variable Resonances Image Combination
Chiel J. den Harder1, Ulrike A. Blume1, and Clemens Bos2
1MR Advanced Solutions, Philips Healthcare, Best, Netherlands, 2MR Clinical Science, Philips Healthcare, Best, Netherlands

With a substantial part of the population having metal implants, there is a clear need for MR imaging resistant to metal artifacts. SEMAC and MAVRIC have been shown to significantly reduce susceptibility artifacts, at the cost of increased scan time. MAVRIC, especially, has been described as spatially non-selective with clinically unacceptable long scan times depending on anatomy size. This work proposes a spatially selective MAVRIC variant (sMAVRIC) that retains the well-defined spectral selectivity of spectral bins, and enables scan time reduction, without the risk of back-folding. Spatial and spectral selectivity are determined by sequence parameters and independent of implant material.

2433.   Comparison of 2D Spin-Echo, Spin-Echo Multi-Spectral Imaging, and Ultra-wide Bandwidth 3D Radial Techniques for Imaging near Metal
Kevin M Koch1, Kevin F King1, Michael Carl2, and Graeme C McKinnon1
1Applied Science Laboratory, GE Healthcare, Milwaukee, WI, United States, 2Applied Science Laboratory, GE Healthcare, San Diego, CA, United States

Imaging near metal hardware has become a routine clinical need in the MR community. There are a variety of factors that dictate which techniques can succeed in performing this task with acceptable levels of image artifact. Here, we compare standard high-bandwidth 2D spin echo approaches with two alternative methods, spin-echo Multi-Spectral Imaging (MSI), and ultra-high-bandwidth zero-TE 3D-radial imaging. Images were acquired on a gridded total hip replacement phantom at 3T. It is shown that spin-echo MSI techniques are able to substantially reduce artifacts compared to 2D spin-echo, but the 3D-radial techniques suffer from substantial image artifacts due to the severe warping of encoded projections

2434.   MR Imaging Near Metal With 3D UTE-MAVRIC Sequences
Michael Carl1
1GE Healthcare, San Diego, CA, United States

The potential of using a MAVRIC-like combination of images obtained at different spectral frequencies with UTE sequences rather than FSE sequences for imaging in the presence of metal was investigated. Phantom and in-vivo studies were performed. UTE-MAVRIC sequences were able to significantly reduce typical artifacts near metal implants and detect very short T2 signals not seen with clinical pulse sequences. Thus, UTE-MAVRIC may further improve the diagnostic capability MR near metal implant, by visualizing short T2 tissues such as tendons, ligaments and cortical bone.

2435.   Local Gradient Effects on Spectral Binning of 3D Multi-Spectral Images Near Metal Implants
Kevin M Koch1, Pauline Worters2, and Brian A Hargreaves2
1Applied Science Laboratory, GE Healthcare, Milwaukee, WI, United States, 2Department of Radiology, Stanford University, Stanford, CA, United States

3D Multi-Spectral Imaging (MSI) methods can substantially reduce MR susceptibility artifacts commonly found near metal implants. MSI techniques excite and independently encode a multitude of spectral "bins" or, spectral-spatial volumes, and then add these bins together to form a composite image. MSI techniques must carefully choose spectral bin windowing strategies to provide a smooth composite image. Near metal, regions are often encountered that possess strong gradients that disrupt encoding processes applied even in MSI techniques. Here, we address the effect of these strong gradients on MSI spectral binning strategies.

2436.   Multi-Spectral Imaging Near Metal: Understanding Performance Differences Between 1.5T and 3.0T
Kevin M Koch1, Matthew F Koff2, and Hollis G Potter2
1Applied Science Laboratory, GE Healthcare, Milwaukee, WI, United States, 2Department of Radiology and Imaging, Hospital for Special Surgery, New York, NY, United States

Initial clinical investigations have focused on applying Multi-Spectral Imaging (MSI) techniques near metal at 1.5T. Recent clinical probes applying MAVRIC-SL MSI technique at 3.0T have also shown encouraging results. Here, we demonstrate the performance impact of MAVRIC SL at 1.5T vs 3.0T. The observed performance difference between the two field strengths is explained, and a modification to MAVRIC-SL is proposed to improve performance at 3T. The results indicate that MAVRIC-SL substantially reduces susceptibility artifacts and can be applied effectively at both field strengths. However, the image quality of MAVRIC-SL, or any MSI technique, will consistently be better at 1.5T compared to 3T

2437.   Quantification of Image Distortion of Orthopedic Materials in Magnetic Resonance Imaging
Matthew F Koff1, Parina Shah1, Kevin M Koch2, and Hollis G Potter1
1Department of Radiology and Imaging - MRI, Hospital for Special Surgery, New York, New York, United States, 2Applied Science Laboratory, General Electric Healthcare, Waukesha, Wisconsin, United States

Susceptibility artifacts occur when performing MRI around orthopedic hardware. This study evaluated the magnitude of in-plane and through plane image distortion produced by a variety of orthopaedic implant materials in standard-of-care 2D fast-spin-echo (FSE) images and multi-acquisition variable-resonance image combination (MAVRIC) images. Metal test samples produced the largest distortion fields, with the largest distortions closest to the sample. MAVRIC scans reduced the in-plane and the through plane image distortions as compared to FSE scans. This study quantitatively evaluated the significant image distortion reduction capabilities of MAVRIC imaging.
Traditional Poster Session - Pulse Sequences & Reconstruction B

Artifacts & Correction: Misc.

Click on to view the abstract pdf. Click on to view the poster (Not all posters are available for viewing.)

Wednesday 9 May 2012
Exhibition Hall  16:00 - 18:00

2438.   A 3D Parametric Model for Imaging Normalization
Tiejun Zhao1, and Kwan-Jin Jung2
1Siemens Healthcare USA; Siemens Medical Solutions USA, Inc., Pittsburgh, PA, United States, 2Psychology, Carnegie-Mellon University, Pittsburgh, PA, United States

Multi-channel receiving (Rx) coils have become a standard asset of routine MR imaging due to its improved signal-to-noise ratio and the parallel imaging capabilities. However, the image intensity variations from the receiving profile could be problematic for tissue segmentations and various quantification analyses. While the pre-scan normalization that acquires additional data using body coil provided a popular approach for removing this imaging shading artifact, a retrospective normalization can still be invaluable especially when the extra data is not available due to various reasons (e.g., the original protocol did not include a pre-scan for normalization or a uniform body coil is not available for some head only scanners or current most 7T scanners.) In this abstract, we proposed and demonstrated a simple 3D parametric model for modeling and removing the smooth image intensity variations presented in images acquired with multi-channel Rx coil.

2439.   Evaluation of MR Image Intensity Inhomogeneity Correction Algorithms
Jinghua Wang1, Lili He2, and Zhong-lin Lu3
1Center for Cognitive and Behavioral Brain Imaging, The Ohio State Univeristy, Columbus, Ohio, United States, 2Center for Perinatal Research, Nationwide Children's Hospital, Columbus, Ohio, United States, 3Center for Cognitive and Behavioral Brain Imaging, The Ohio State University, Columbus, Ohio, United States

MR image intensity inhomogeneity has rendered quantitative MRI analysis in anatomical studies a major challenge. Various methods for performing inhomogeneity correction have been proposed. Evaluation of these methods has often relied on subjective assessment because of the lack of the ground truth. Here, we present a new method to evaluate four popular inhomogeneity correction methods based on both uniform phantom and in vivo brain images. We found that the field map method outperformed the others. This method can be used to guide parameter optimization for existing correction methods, improve bias field modeling, and evaluate and optimize new correction methods.

2440.   3D Variable Flip Angle Fast-Low-Angle-Shot Experiments in the Presence of B1 Inhomogeneity and Slab-Select Gradient
Kelly C McPhee1, and Stefan A Reinsberg1
1Physics and Astronomy, University of British Columbia, Vancouver, BC, Canada

Variable Flip Angle (VFA) Fast-Low-Angle-Shot (FLASH) experiments are commonly used for T1 and B1 mapping. Due to the effect of non-ideal pulse shape on signal intensities, 3D experiments are often used in place of 2D experiments. We show the effects of this assumption through both simulations and experimental data. The shape of the VFA curve is altered, and a zero-crossing will only be observed in the real part of the signal – not in the magnitude – of a FLASH VFA curve. We compare T1 and flip angle map results to a VFA experiment with no slab-select-gradient, and find large T1 errors.

2441.   Software-based automated measurement of susceptibility artifacts on magnetic resonance images
Andreas Heinrich1, and Felix Guettler1
1Department of Radiology, University Hospital Jena, Jena, Thuringia, Germany

A platform independent software system (SAM) for rapid, objective and reproducible measurement of susceptibility artifacts, complying with ASTM F2119-07, was developed. The artifact of a defined titanium cylinder was imaged in a Siemens Magnetom Avanto. For evaluation purposes 13 probands determined the greatest artifact, manually with a ruler (Osirix) and with the SAM software, which could reduce the dispersion of the results by more than 80%. Furthermore, the SAM software safes 50% of MR measurement time, since the required ASTM reference images without test object are no longer necessary.

2442.   FreeSurfer Parcellation of Brains Containing Large Infarcts
Niharika Gajawelli1, Sinchai Tsao1, Darryl Hwang1, Bryce Wilkins1, Stephen Kriger2, and Manbir Singh1
1Radiology and Biomedical Engineering, University of Southern California, Los Angeles, California, United States, 2Center for Imaging of Neurodegenerative Diseases, VA Medical Center, San Francisco, California, United States

FreeSurfer parcellation of the T1 weight human brain often fails in the presence of cerebral infarcts. This study presents a method of using voxels from the contralateral hemisphere to fill-in the cerebral infarct. The technique relied on SPM normalization including the normal and the hemispherically flipped versions of the brain, without any manual intervention other than an initial masking. A novel validation methodology was developed based on comparing ratios of FreeSurfer parcellated regions in a simulation study where the ground truth was known. Results show successful processing of 16 stroke subjects after filling, whereas processing in half failed before filling.

2443.   A semi-automatic k-space despiking algorithm for the removal of striping artefacts in MR images
Adrienne E Campbell1,2, Oliver Josephs3, Mark F Lythgoe1, Roger J Ordidge4, and Dave L Thomas5
1Centre for Advanced Biomedical Imaging, Division of Medicine and Institute of Child Health, University College London, London, United Kingdom, 2Department of Medical Physics and Bioengineering, University College London, London, United Kingdom, 3University College London and Birkbeck College, London, United Kingdom, 4Centre for Neuroscience, University of Melbourne, Melbourne, Australia, 5Department of Brain Repair and Rehabilitation, UCL Institute of Neurology, University College London, London, United Kingdom

Striping artefacts are a common in MR images and are caused by RF noise spikes in k-space. We present a semi-automatic post-processing algorithm for the removal of RF spikes from k-space, while preserving quantitative information. The algorithm detects spikes by taking advantage of their brief duration and random appearance. Then, removal of this anomalous data and replacement with suitable substitute data takes place, while protecting the centre of k-space. This algorithm was found to effectively remove RF spikes, improving image quality and retaining quantitative information.

2444.   Simple Method for Attenuation of Streaking Artifacts from Peripheral Intensity Accumulation
Kai Tobias Block1, and Matthias Fenchel2
1Department of Radiology, NYU Langone Medical Center, New York, NY, United States, 2Siemens AG Healthcare Sector, Erlangen, Germany

Images acquired with radial k-space sampling frequently suffer from mild or severe streaking artifacts even if the Nyquist sampling requirements are fulfilled. Often, the streaking artifacts arise from high-intensity signal spots in the periphery of the field-of-view. This work describes a simple method to attenuate such long-distance streaks by constraining the image contributions from each coil element to only those areas where the element is known to have sufficient sensitivity. The method works fully automatic, does not affect the SNR or prolong the scan time, and achieved clear improvement of the image quality in volunteer studies.

2445.   Towards real-time 4D field shift predictions: optimizing Fourier-based calculations of the susceptibility induced perturbation of the magnetic field
J. G. Bouwman1, J. S. van Gorp2, P. R. Seevinck3, and C. J.G. Bakker3
1Image Sciences Institute - Radiology, UMC Utrecht, Utrecht, Netherlands, 2UMC Utrecht, Netherlands, 3UMC Utrecht

Upspeeding the Fourier-Based calculation of the suceptibility induced field. A higher level of efficiency can be gained if the alias prevention by zero-padding is decoupled from the actual field calculation. This can be done by afterwards estimating this aliasing in a lower resolution. This 'virtual' zero-padding is more than three times faster compared to conventional zero-padding, with only minor accuracy loss. This algorithm is especially applicable in dynamic field shift calculations.

2446.   Improved diffusion-weighted body imaging using high-order eddy current correction and smart combination
ZHIQIANG LI1, Dan Xu2, and Kenichi Kanda1
1MR Engineering, GE Healthcare, Waukesha, WI, United States, 2Applied Science Laboratory, GE Healthcare, Waukesha, WI, United States

Single-shot DW-EPI of the liver has the drawback in signal loss due to cardiac motion. Smart combination has been proposed to mitigate this issue, but requires double spin-echo diffusion to avoid mis-registration between the individual source images acquired with single spin echo (SSE) diffusion, with the trade-off in SNR. High-order eddy-current (HOEC) correction has been presented to minimize the mis-registration in data acquired with SSE. In this work, we apply both HOEC correction and smart combination on SSE DW-EPI to demonstrate the feasibility of generating liver image with high SNR, minimal signal loss and mis-registration.

2447.   PROPELLER-EPI improved by 2D phase cycled reconstruction
Hing-Chiu Chang1,2, Nan-Kuei Chen3, Tzu-Chao Chuang4, Chun-Jung Juan5, Ming-Long Wu6,7, and Hsiao-Wen Chung2,5
1Global Applied Science Laboratory, GE Healthcare, Taipei, Taiwan, 2Institute of Biomedical Electronics and Bioinformatics, National Taiwan University, Taipei, Taiwan, 3Brain Imaging and Analysis Center, Duke University, Durham North Carolina, United States, 4Electrical Engineering, National Sun Yat-sen University, Taiwan, 5Department of Radiology, Tri-Service General Hospital, Taipei, Taiwan, 6Institute of Medical Informatics, National Cheng Kung University, Taiwan, 7Department of Computer Science and Information Engineering, National Cheng Kung University, Taiwan

PROPELLER-EPI consists of EPI signal readout, where the phase inconsistencies between odd and even echoes generate oblique N/2 ghost artifact in each rotating blade. A 2D phase-cycled reconstruction for inherent correction of N/2 ghost without any extra reference scan has been reported recently. In this work, we integrate an improved version of the 2D phase-cycled reconstruction to PROPELLER-EPI, and compare the result to that reconstructed using the 2D reference-based method. 2D phase-cycled reconstruction is particularly suitable for correcting oblique ghost in PROPELLER-EPI without any need of extra reference scan, and thus may broaden the clinical use of the PROPELLER-EPI technique.

2448.   PROPELLER with Echo Stabilization
ZHIQIANG LI1, Dawei Gui1, Ajeet Gaddipati1, and Xiaoli Zhao1
1MR Engineering, GE Healthcare, Waukesha, WI, United States

PROPELLER has been widely used in the clinic, but also experiences the violation of the Carr-Purcell-Meiboom-Gill condition as in FSE, which degrades the signal. FSE employs phase correction to measure the phase error and adjust the data acquisition accordingly. However, its direct application in PROPELLER is not feasible since measurement at all blade angles leads to unacceptable increase in scan time. In this work, a model based on physical gradient is proposed so that the phase error can be predicted from measurement at a few blade angles. This reduces scan time, helps the application of PROPELLER in body/MSK imaging.

2449.   Rician Noise Removal in Diffusion Kurtosis Imaging
Nanjie Gong1, Chun Sing Wong1, Sau May Wong2, and Bart ter Haar Romeny2
1Diagnostic Radiology, The University of Hong Kong, Hong Kong, China, 2Biomedical Engineering, University of Technology Eindhoven, Netherlands

In the past decade, many denoising algorithms have been proposed for MR images. Most often the Gaussian filter is used for this purpose. However it has already been proven that noise on the magnitude MR images is not Gaussian, but Rician distributed. Gaussian filters are used in general for the reason that the noise distribution of magnitude MR images with high signal-to-noise ratio (SNR) can be well approximated with a Gaussian distribution. Diffusion Kurtosis Imaging (DKI), an extension of Diffusion Tensor Imaging (DTI), generates images with a lower SNR compared to images obtained with DTI due to the need of larger and more b-values. Subsequently the noise distribution in obtained DKI MR images can no longer be treated as Gaussian distributed. In this study, two Rician filters, namely the non-local mean Rician filter (NLM) and the non-local maximum likelihood filter (NLML), and Gaussian filter are investigated on their denoising performance on DKI MR images. Results obtained from the simulations show more accurate derived parameters from DKI images denoised with the Rician filters than those from images denoised with the Gaussian filter. In addition, regarding the parameters� value derived from real human data, there are significant differences between those from data filtered with the Rician filters and those from data filtered with Gaussian filter.

2450.   MRI-Based Attenuation and Background Correction in Nuclear Projection Imaging
Mark Jason Hamamura1, Seunghoon Ha1, Werner W Roeck1, James Hugg2, Dirk Meier3, Bradley E Patt2, and Orhan Nalcioglu1,4
1Tu & Yuen Center for Functional Onco-Imaging, University of California, Irvine, CA, United States, 2Gamma Medica, Inc., Northridge, CA, United States, 3Gamma Medica, Inc., Fornebu, Norway, 4Department of Cogno-Mechatronics Engineering, Pusan National University, Pusan, Korea

Accurate radiotracer quantification in nuclear projection imaging requires both attenuation and background corrections. In this study, we acquired MR and nuclear projection images of a phantom using a novel MR-SPECT system. The MRI was then used to facilitate correction of the nuclear projection data. The results demonstrate improved accuracy for radiotracer quantification using data acquired from simultaneous imaging.

2451.   A data-driven framework for removing physiological noise in fMRI
Nathan Churchill1, and Stephen Strother2
1Medical Biophysics, University of Toronto, Toronto, Ontario, Canada, 2University of Toronto / Rotman Research Institute, Toronto, Ontario, Canada

We have developed a multivariate, data-driven framework to estimate and control physiological noise in functional MRI. This model (1) identifies and down-weights sources of pulsatile flow (e.g. arteries, sinuses and ventricles), based on a map of high-frequency spectral power, correlated with an atlas of potential artifact regions. It then (2) regresses out physiological noise present in grey matter, using an adaptation of Canonical Correlation Analysis and the spatial weights derived from Step (1). This denoising procedure consistently reduces activation false positives, and increases both prediction accuracy and reproducibility of activation maps in subsequent analyses.

2452.   A Robust Algorithm Framework for Small DTI Samples
Ivan Maximov1, Farida Grinberg1, and N. Jon Shah1,2
1Institute of Neuroscience and Medicine 4, Forschungszentrum Juelich GmbH, Juelich, Germany, 2RWTH Aachen University, Department of Neurology, Faculty of Medicine, JARA, Aachen, Germany

Low redundancy DTI data sets are a rather complicated problem for diffusion tensor estimation, especially in a clinical human brain imaging. The diffusion signal attenuations are frequently corrupted by a physiological noise such as a cardiac pulsation, bulk head motion, respiratory motion, etc. As a consequence, diffusion tensor estimation becomes unstable and very poor. We have developed fitting algorithms based on the least trimmed squares and the median absolute deviation robust estimators in order to improve the tensor assessment in small DTI samples.

A modified variable fip angle using a predefiend slice profile in a consecutive interleaved EPI
Dae-Hun Kang1, Jun-Young Chung1, Da-Eun Kim1, Young-Bo Kim1, and Zang-Hee Cho1
1Neuroscience Research Institute, Gachon University of Medicine and Science, Incheon, Korea

We re-calculated a VFA considering a predefined slice excitation profile without changing RF excitation pulse by a computer simulation. By applying a modified VFA to consecutive interleaved multi-shot EPI (ciEPI) schemes, a intersegment magnitude variation was reduced and a SNR of images was improved.

2454.   Fat Fraction Bias Correction using T1 estimates and Flip Angle Mapping
Issac Y. Yang1, Yifan Cui2, Curtis N. Wiens2, Trevor P. Wade1, and Charles A. McKenzie1,2
1Medical Biophysics, University of Western Ontario, London, Ontario, Canada, 2Physics and Astronomy, University of Western Ontario, London, Ontario, Canada

Fat fraction quantification with IDEAL requires the use of small flip angles to minimise T1 bias and therefore has reduced SNR. We employed rapid flip angle mapping to improve T1 bias correction for high flip angle acquisitions in the estimation of spinal bone marrow fat fraction in volunteers. We compared fat fraction estimation with 2 and 8 degree flip angles, with and without bias correction based on flip angle mapping. The results show the fat fractions corrected for high flip angle induced T1 bias have improved accuracy and precision compared to a low flip angle acquisition.

2455.   Spike noise removal in MR images using over sampling
Yongchuan Lai1, and Xiaocheng Wei1
1GE Healthcare, Beijing, China

MR images are reconstructed from K space raw data. Sometimes K space raw data can be corrupted with high amplitude value noise, resulting gross striation artifact. This noise is referred as spike noise. The causes of these spikes can be arcing, loose connection or other electrical/mechanical problems in system. Spike noise is difficult to diagnosis and usually it is cost and time consuming job to fix a spike noise issue. Over-sampling is widely used in current MR products. Usually, over-sampled data are directly filtered to avoid alias artifact in frequency direction. In method of this abstract, over-sampled data are used not only to avoid alias but also to find spike noise.

2456.   Characterization of NIST/ISMRM MRI System Phantom
Stephen E Russek1, Michael Boss1, Edward F Jackson2, Dominique L Jennings3, Jeffrey L Evelhoch4, Jeffrey L. Gunter5, and A. Gregory Sorensen6
1NIST, Boulder, CO, United States, 2M D Anderson Cancer Center, Houston, TX, United States, 3Massachusetts General Hospital, United States, 4Merck Research Laboratories, 5Mayo Clinic, Rochester, MN, United States, 6Siemens Healthcare, United States

An MRI system phantom has been developed through collaboration between the ISMRM ad-hoc committee on Standards for Quantitative Magnetic Resonance and the National institute of Standards and Technology (NIST). This will be the first system phantom with SI-traceable components and the first to be monitored by NIST for stability and accuracy. The system phantom consists of the following elements: fiducial , T1, T2, proton density arrays; resolution, slice profile, and SNR insets. The system phantom is designed to assess scanner performance and quantitative mapping protocols and to easily compare performance with other scanners across the world. The system phantom has been imaged extensively at Mass General Hospital and M D Anderson Cancer Center in a variety of 1.5 and 3T scanners.
Traditional Poster Session - Pulse Sequences & Reconstruction B

Motion Correction

Click on to view the abstract pdf. Click on to view the poster (Not all posters are available for viewing.)

Wednesday 9 May 2012
Exhibition Hall  16:00 - 18:00

2457.   Enhanced Motion Correction Combining PROPELLER and Parallel Imaging
Yi Guo1, Xiaoke Wang2, Sheng Fang3, Kui Ying4, Hua Guo5, and Shi Wang4
1Dept.of Electrical Engineering, Tsinghua University, Beijing, Beijing, China, 2Dept.of Biomedical Engineering, Tsinghua University, 3Institute of nuclear and new energy technology, Tsinghua University, 4Dept.of Engineering Physics, Tsinghua University, 5Center for Biomedical Imaging Research, Tsinghua University

In this study we proposed a method to combine parallel imaging (PI) with PROPELLER to correct non-rigid motion in addition to rigid motion either within a blade or between blades. Instead of using PI techniques to speed up MR acquisition, here the data redundancy in fully sampled PI is exploited to detect and eliminate corrupt k-space lines, thus to correct a variety of motion. We use CG-PROPELLER to reconstruct the image from multi-coil PROPELLER data, and then split back to original k-space blades. By comparing the difference between the regenerated and original lines for all blades, corrupt lines are detected and discarded. The final image is then reconstructed from the under-sampled data using CG-PROPELLER. Both phantom and in-vivo results show the feasibility of enhanced non-rigid motion correction.

2458.   Motion Correction of a new T1-w Propeller Sequence (SE-prop)
Stefan Skare1
1Dept of Clinical Neuroscience, Karolinska Institutet, Stockholm, Sweden

A new propeller sequence is presented for motion-robust imaging with T1-w contrast, SE-prop. In SE-prop, the core sequence is the same as the Cartesian Spin-Echo, with one phase encoding line of a given blade read out per TR. While motion between the blades can be corrected using image realignment similar to other propeller sequences, motion may now also occur within a blade. Means to overcome this is presented.

2459.   High Temporal Resolution 3D Motion Correction of MP-RAGE with Cylindrical Sampling and Parallel Reconstruction
Wei Lin1, Qin Qin2,3, Feng Huang1, and Randy Duensing1
1Invivo Corporation, Philips Healthcare, Gainesville, Florida, United States, 2Department of Radiology, The Johns Hopkins University, Baltimore, Maryland, United States, 3F.M. Kirby Research Center, Kennedy Krieger Institute, Baltimore, Maryland, United States

A self-navigated method for motion correction in multicoil imaging applications, involving both data collection and reconstruction, is presented. The cylindrical k-space is collected using multiple pairs of orthogonal view planes with a bit-reversed radial angle ordering scheme. A rapid self-calibrated parallel imaging method, generalized GRAPPA Operator for Wider readout Lines (GROWL), facilitates both motion detection and image reconstruction. Full 3D rigid-body motion detection is achieved at a temporal resolution of every two consecutive view planes (2-6 seconds). The efficacy of the proposed technique is demonstrated in healthy volunteers using a magnetization-prepared rapid gradient echo (MP-RAGE) sequence.

2460.   Respiratory Gated VIBE Sequence
Alto Stemmer1, MunYoung Paek2, Jeong Min Lee3, and Berthold Kiefer1
1Healthcare Sector, Siemens AG, Erlangen, Germany, 2Siemens Ltd. Seoul, Korea, 3Seoul National University, Korea

The aim of this work is to describe a recently developed navigator gated 3D spoiled gradient echo sequence (VIBE) for high resolution liver imaging during the uptake phase of hepatobiliary contrast media. Radial phase encoding allows a flexible choice of the number of TR intervals per navigator/fat suppression pulse and hence temporal resolution of the respiratory curve. An extended PAWS gating algorithm is used to ensure robustness in the case of a varying breathing pattern.

2461.   Blind Retrospective Motion Correction of MR Images
Alexander Loktyushin1, Hannes Nickisch2, Rolf Pohmann3, and Bernhard Schölkopf1
1Max Planck Institute for Intelligent Systems, Stuttgart, Germany, 2Philips Research Laboratories, Hamburg, Germany, 3Max Planck Institute for Biological Cybernetics, Tübingen, Germany

Patient motion in the scanner is one of the most challenging problems in MRI. We propose a new retrospective motion correction method for which no tracking devices or specialized sequences are required. We seek the motion parameters such that the image gradients in the spatial domain become sparse. We then use these parameters to invert the motion and recover the sharp image. In our experiments we acquired 2D TSE images and 3D FLASH/MPRAGE volumes of the human head. Major quality improvements are possible in the 2D case and substantial improvements in the 3D case.

Non-rigid Motion Correction in 3D Using Autofocusing with Localized Linear Translations
Joseph Y. Cheng1, Marcus T. Alley2, Charles H. Cunningham3,4, Shreyas S. Vasanawala2, John M. Pauly1, and Michael Lustig5
1Electrical Engineering, Stanford University, Stanford, California, United States, 2Radiology, Stanford University, Stanford, California, United States, 3Medical Biophysics, University of Toronto, Toronto, Ontario, Canada, 4Imaging Research, Sunnybrook Health Sciences Centre, Toronto, Ontario, Canada, 5Electrical Engineering and Computer Science, University of California, Berkeley, California, United States

MR scans are sensitive to motion effects due to the scan duration. On a sufficiently small spatial-scale, the complex non-rigid motion can be well approximated as simple linear translations. This formulation allows for a practical autofocusing algorithm that locally minimizes a given motion metric -- more specifically, the proposed localized gradient-entropy metric. To reduce the vast search space for an optimal solution, possible motion paths are limited to motion measured from multi-channel Butterfly navigation data. The correction scheme is applied to free-breathing abdominal patient studies. A reduction in artifacts from non-rigid motion is observed.

2463.   Reduction of Breathing Motion Artifacts in Shoulder Imaging using an Orbital Navigator as Motion Sensor
Tim Nielsen1, Chiel den Harder2, Clemens Bos3, and Peter Börnert1
1Philips Research, Hamburg, Germany, 2Philips Healthcare, Best, Netherlands, 3University Medical Centre, Utrecht, Netherlands

A breathing motion artifact reduction method is demonstrated which can be applied to a clinical shoulder imaging protocol and 1.) uses an orbital navigator (onav) signal to sense the breathing motion and 2.) uses a reconstruction from a reduced dataset to improve image quality. The motion signal from the navigator is used to control which data are used in the reconstruction. Using only the sub-set of data where motion is minimal can significantly reduce ghosting and blurring.

2464.   On moving coil encoding and calibration in dynamic imaging
Johannes F.M. Schmidt1, Peter Boesiger1, and Sebastian Kozerke1
1Institute for Biomedical Engineering, University and ETH Zurich, Zurich, Switzerland

Image reconstruction of multi-channel coil data is based on sensitivity maps for receive coils retrieved from either a pre-scan or by autocalibration of sensitivities from the scan itself. The implications of moving objects and moving coils on reconstruction performance with static coil sensitivity maps are shown on basis of simulated free-breathing contrast enhanced myocardial perfusion imaging.

2465.   Adaptive respiratory triggering for high spatial and temporal resolution 3D DCE-MRI in the mouse.
Alexandr Khrapichev1, Veerle Kersemans1, Danny Allen1, and Sean Smart1
1Oxford University, Oxford, Oxfordshire, United Kingdom

A method for performing high temporal and spatial resolution imaging for DCE-MRI in the mouse is presented. Respiration triggering is used to reduce motion artefact, and dummy scans are used to regenerate the T1 steady state following each breath. Both the rate and depth of the breathing change over the duration of a typical DCE-MRI scan so the numbers of dummy and imaging acquisitions per breath may need to vary in order to maximise data capture rate whilst avoiding motion. 3D gradient echo scans with an isotropic resolution of ca. 400 micron and a temporal resolution of 5 s (untriggered) and 8-9 s (triggered) is achieved. Respiratory gated and non gated scans of the mouse abdomen and thorax were acquired sequentially during the uptake from a single bolus injection of Gd contrast agent. Improvements to signal-to-noise and image fidelity are shown when using triggering.

Multimodal non-rigid motion artifact correction with concurrent ultrasound
Jihan Kim1, and Dosik Hwang1
1School of Electrical and Electronic Engineering, Yonsei University, Seoul, Korea

This abstract provides a novel multi-modal imaging framework to compensate complicated motion artifacts in MR. Results prove the feasibility of our proposed method. This method can correct non-rigid and non-periodic motion even in a single k-space dataset (without repetitive acquisition for multiple cycles, without gating).

2467.   Self-adjusting multi-channel UWB radar for cardiac MRI
Olaf Kosch1, Steffen Schneider1, Bernd Ittermann1, and Frank Seifert1
1Physikalisch-Technische Bundesanstalt (PTB), Berlin, Germany

We have extended our motion detection by contactless multi-channel UWB radar for cardiac MRI navigator to a 4x8 channel system to make the application more robust. The new system is insensitive to a not well placed antenna configuration related to the position of the monitored person. Additionally a method for rejecting noisy channels was implemented to avoid average effects.

2468.   Prospective motion correction using NMR probes.
Saikat Sengupta1, John C. Gore1, and E. Brian Welch1
1Radiology and Radiological Sciences, Vanderbilt University, Nashville, TN, United States

Motion is one of the primary sources of artifacts in MRI and improved motion correction is especially useful in high-resolution imaging. Nuclear magnetic resonance probes have recently been employed successfully in rigid body motion correction at 1.5 Tesla. In this work, real time prospective head motion correction based on proton NMR probes at ultra high field (7T) is presented. Simulated motion on a phantom is used for system validation and testing of real time motion correction algorithm.

2469.   Motion-tracking: Fast high-resolution 2D motion quantification and depiction of shivering
Eberhard Eduard Munz1, Andreas Johannes Hopfgartner1, Sairamesh Raghuraman2, Titus Lanz3, and Peter Michael Jakob1
1Experimental Physics 5, University of Würzburg, Würzburg, Germany, 2MRB Research Centre, Würzburg, Germany, 3Rapid Biomedical GmbH, Rimpar, Germany

In hand imaging the "superman position" is stressful for the patient and therfore prone to image artifacts. In this study we used the method of cross-correlation in order to detect a contrast agent-filled sphere attached to the patient's hand. Detecting the marker very accurately and hence quantificating the amount of motion caused by three different hand-positions we identified the least stress-causing position for the patient. Herefrom it is possible to reduce the patient's shivering to a minimum and furthermore enhance image quality in high-field (7T) MR-Imaging.

2470.   Subject specific respiratory motion correction accounting for hysteresis
Ian Burger1, Ernesta Meintjes1, David Firmin2,3, and Jennifer Keegan2,3
1Human Biology, University of Cape Town, Cape Town, Western Cape, South Africa, 2Imperial College, London, United Kingdom, 3CMR Unit, Royal Brompton Hospital Trust, London, United Kingdom

In this study we investigated a non-rigid subject specific respiratory motion model in cardiac MRI that accounts for hysteresis. The model is constructed of an ellipse superimposed on a straight line. Nine healthy volunteers participated in a study in which a navigator sample and a 2D image of the heart were acquired in each cardiac cycle for 90 cycles. The results show that this model performed significantly better than a linear affine model. Further, we have demonstrated that a pre-scan of about 25 seconds (~25 cardiac cycles) is sufficient to construct the model.

2471.   Motion Correction: MRI-based ultra-fast high-resolution 3D tracking
Andreas Johannes Hopfgartner1, Eberhard Munz1, Florian Fidler2, and Peter Michael Jakob1
1Experimental Physics 5, University of Würzburg, Würzburg, Germany, 2Research Center Magnetic-Resonance-Bavaria, Würzburg, Deutschland, Germany

For high-resolution MR imaging a simple, flexible, fast and robust 3D motion tracking method was searched. Optical devices applications in motion correction suffer from the drawback that the devices have to be synchronized in time and space with the MRT hardware. The proposed method tracks marker spheres without additional equipment and uses scanner information only. Two alternatingly and rapidly measured “2D projection images” from small objects include the position information. The method allows a high-resolution (accuracy of plus-or-minus sign40µm and plus-or-minus sign0.1°) and ultra-fast (170ms) measurement for 3D motion correction and can be built in in every sequence as navigator block.

2472.   Fast Inter-scan Motion Detection and Compensation
Tim Nielsen1, Peter Börnert1, and Julien Sénégas1
1Philips Research, Hamburg, Germany

In a typical MRI exam multiple scans are performed. The geometry of these scans is planned on a survey. If the patient moves after the acquisition of the survey, the scanned location is different from the planned location. We present a fast method to detect and correct for motion between individual scans of an exam using sets of three orthogonal slices with low resolution to determine and track the head position. This tracking method is a simple, reliable and efficient approach to cope with inter-scan motion. It requires no user interaction and does not interfere with the clinical workflow.
Traditional Poster Session - Pulse Sequences & Reconstruction B

Intensity/Distortion Correction

Click on to view the abstract pdf. Click on to view the poster (Not all posters are available for viewing.)

Wednesday 9 May 2012
Exhibition Hall  16:00 - 18:00

2473.   Estimating Phase Maps from Partial K-Space Data
Tom Depew1, and Qing-San Xiang1,2
1Physics & Astronomy, University of British Columbia, Vancouver, BC, Canada, 2Radiology, University of British Columbia, Vancouver, BC, Canada

Phase can be used to encode information about numerous MRI properties such as magnetic field strength, motion or flow, and temperature. Phase maps are typically obtained by comparing two complex images in which the relative phase is sensitized to the physical quantity of interest. In situations where the phase can be assumed to change smoothly, the two images can be reconstructed from central k-space data to reduce scan time. Here we investigate the effect of truncation artifact associated with this approach and the feasibility of obtaining phase maps by fitting a convolution kernel to various partial k-space coverage patterns.

2474.   Efficient concomitant field artifacts reduction using a hybrid space-frequency domain formulism
Yi-Cheng Hsu1, Panu T. Vesanen2, Jaakko O. Nieminen2, Koos C. J. Zevenhoven2, Juhani Dabek2, I-Liang Chern3, Risto J. Ilmoniemi4, and Fa-Hsuan Lin3,5
1National Taiwan University, Taipei, Taiwan, 2Aalto University, Finland, 3National Taiwan University, Taiwan, 4Aalto University, Taiwan, 5Massachusetts General Hospital, United States

To mitigate the challenge of reconstructing MRI data deteriorated by concomitant fields, we develop a hybrid space-frequency (x-f) domain method to i) accurately describe the dynamics of spatially encoded magnetization allowing arbitrary directions of initial magnetization and phase/frequency encoding magnetic fields, and (ii) reconstructed images using combined fast Fourier transform and a linear equation solver. We demonstrate this method in ultra-low-field MRI reconstruction when the largest concomitant field is as strong as B0.

2475.   Selfcalibrated DCE MRI using Multi Scale Adaptive Normalized Averaging (MANA)
Thord Andersson1,2, Thobias Romu1,2, Bengt Norén2,3, Mikael Fredrik Forsgren2,4, Örjan Smedby2,3, Stergios Kechagias5,6, Sven Almer6,7, Peter Lundberg2,4, Magnus Borga1,2, and Olof Dahlqvist Leinhard2,4
1Dept of Biomedical Engineering (IMT), Linköping University, Linköping, Sweden, 2Center for Medical Image Science and Visualization (CMIV), Linköping University, Linköping, Sweden, 3Dept of Medical and Health Sciences (IMH), Div of Radiological Sciences, Linköping University, Linköping, Sweden, 4Dept of Radiation Physics, Linköping University and Radiation Physics, UHL County Council of Ostergotland, Linköping, Sweden, 5Dept of Medical and Health Sciences (IMH), Div of Cardiovascular Medicine, Linköping University, Linköping, Sweden, 6Dept of Endocrinology and Gastroenterology, UHL County Council of Ostergotland, Linköping, Sweden, 7Dept of Clinical and Experimental Medicine (IKE), Div of Gastroenterology and Hepatology, Linköping University, Linköping, Sweden

In conventional analysis of DCE MRI time series, constant sensitivity of the MR-scanner during the experiment usually is assumed. However, patient movement and other effects may affect the sensitivity during the time series. The Multi Scale Adaptive Normalized Averaging (MANA) method has been proposed to correct intensity inhomogeneity in fat-water MRI using Dixon imaging, and to provide reference scaling in DCE MRI time series. In this work, we validate statistically the correctness of MANA compared to conventional scaling. The results show that MANA provides more consistent intensities and successfully can recreate scaling information directly from the data.

2476.   Guided Multi-channel Blind Deconvolution Coil Intensity Correction for Improved Uniformity and Contrast Fidelity
Dominic M Graziani1, Stephen J Garnier2, Ek T Tan1, and Christopher J Hardy1
1GE Global Research, Niskayuna, NY, United States, 2GE Healthcare, Waukesha, WI, United States

We report on an iterative reconstruction technique for image-based intensity correction of phased-array images using multi-channel blind deconvolution (MBD) and homomorphic filtering. This combined method produces images with better bias field correction and higher contrast fidelity than either technique on its own. The algorithm is relatively insensitive with respect to the choice of adjustable parameters, and reasonably fast, taking around 2 seconds per slice in Matlab. We demonstrate the effectiveness of this technique by correcting spine images acquired with an 8 channel single-sided array with good resulting uniformity and contrast.

Nicolas Vibet1,2, François De Guio2,3, Aurélien Monnet4, Nadine Girard5, Stéphane Lehéricy2,6, Sophie Pérusat7, Carole Dufouil8, Christine Delmaire2,4, Cyril Poupon2,9, and Marie Chupin1,2
1CRICM, UPMC URMR_S975, INSERM U975, CNRS UMR7225, ICM, Paris, France, 2CATI, multicentre neuroimaging platform, Paris, France, 3LNAO, Neurospin, CEA, Gif-sur-Yvette, France, 4Department of Neuroradiology, AP–HP, Lille, France, 5Department of Neuroradiology, Hopital Timone, Marseille, France, 6CRICM & CENIR, UPMC/CNRS/INSERM/ICM, Paris, France, 7U593, INSERM, Bordeaux, France, 8U897, INSERM, Bordeaux, France, 9LRMN, Neurospin, CEA, Gif-sur-Yvette, France

Intensity non-uniformity is often largely visible in 3DT1 acquistions; it can lead to difficulties in advanced data analysis. Several methods are available, from manufacturers or post processing tools, but their use in multicentre study has yet to be evaluated; we compared the performances of manufacturers intensity non uniformity correction tools with two available post processing methods (BrainVISA and SPM) for three scanners (Siemens 1,5T and 3T, Philips 3T). This analysis is based on four quality indices computed on the image and its itensity histogram. In the framework of multicentre analyses, post-processing methods appear to be better adapted and perform satisfactorily.

2478.   Correcting geometric distortions of Echo Planar Imaging using demons and reversed phase encoding
M. Lyksborg1,2, H. Lundell2, N. Reislev2, H. R. Siebner2, R. Larsen1, and T. B. Dyrby2
1Technical University Of Denmark, Kgs. Lyngby, Denmark, 2Danish Research Centre for Magnetic Resonance, Hvidovre, Denmark

Correcting for geometric distortions of Echo Planar Imaging (EPI) is extremely important for anatomical fidelity. Based on the acquisition of 2 EPI’s with reversed gradient polarity, we suggest using an efficient non parametric image registration method to estimate the geometric distortion field. We demonstrate the method on 5 subjects and compare it with a minimally distorted structural image. The method achieves results similar to that of the field map and point spread function method.

2479.   Correcting slice selectivity in hard pulse sequences
David Manuel Grodzki1,2, Peter Jakob1, and Bjoern Heismann2,3
1EP5, University of Wuerzburg, Wuerzburg, Bavaria, Germany, 2Magnetic Resonance, Siemens AG, Erlangen, Germany, 3Friedrich-Alexander-University, Erlangen, Germany

Many MRI sequences use non-selective hard pulse excitation in the presence of imaging gradients. In this work, we investigate possible artefacts due to unwanted slice-selectivity in hard pulse sequences. A correction algorithm is proposed that eliminates the influence of the excitation profile. Phantom as well as in-vivo measurements prove that enhanced image quality can be obtained as long as the first minimum of the excitation profile lies outside the imaged object.
Traditional Poster Session - Pulse Sequences & Reconstruction B

Water/Fat Imaging

Click on to view the abstract pdf. Click on to view the poster (Not all posters are available for viewing.)

Wednesday 9 May 2012
Exhibition Hall  16:00 - 18:00

2480.   Iterative Decomposition of Water and Fat with Echo Asymmetric and Least-Squares Estimation (IDEAL) Compared to T1-weighted Spin Echo in the Evaluation of Vertebral Body Lesions
Behroze Adi Vachha1, Subhendra Sarkar1, Robert Greenman1, and David Hackney1
1Radiology, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA, United States

We compared the effectiveness of Iterative Decomposition with Echo Asymmetric and Least-Squares Estimation (IDEAL) fat-selective images to conventional T1-weighted spin-echo (SE) images for depiction of vertebral body lesions with reduced or increased fat content. We computed signal intensity ratios of lesions to adjacent vertebrae with both sequences. IDEAL showed significantly higher contrast between vertebral lesions and adjacent vertebral bodies than T1 SE images, ratios were 0.58 (T1 weighted), and 0.25 (IDEAL fat-selective) for reduced fat lesions and 1.47 (T1 weighted) and 1.83 (IDEAL fat-selective) for increased fat lesions; p<0.05 for both lesion types. IDEAL also displays significantly higher fat/water contrast, p<0.001.

2481.   Joint Inference of Field Inhomogeneities with Fat Likelihood Estimation from Three Echoes
Wenmiao Lu1, and Yi Lu2
1Beckman Institute, University of Illinois, Urbana-Champaign, Illinois, United States, 2Electrical and Computer Engineering, University of Illinois, Urbana-Champaign, Illinois, United States

This work presents a joint inference algorithm which effectively combines the field map smoothness constraint and the fat likelihood estimated from three echoes. This technique is potentially useful for motion-sensitive applications, such as liver fat quantification.

2482.   Field Map Error Location and Correction for Fat/Water Separation Methods
Sreenath Narayan1, Satish Kalhan2, and David Wilson1
1Department of Biomedical Engineering, Case Western Reserve University, Cleveland, OH, United States, 2Department of Molecular Medicine, Case Western Reserve University, Cleveland, OH, United States

We have shown that error correction can be used to reliably and automatically decompose fat and water at 7T, with minimal human interference. When combined with FLAWLESS, a previously published technique, we were able to avoid complete swaps of fat and water, while preventing partial swaps from appearing in 99.7% of our 134 images.

2483.   Image-based determination of the fat signal model for Dixon water and fat imaging
Tze Yee Lim1, and Jingfei Ma1
1University of Texas MD Anderson Cancer Center, Houston, TX, United States

Accurate modeling of the fat signal is important for Dixon water and fat imaging. However, many confounding factors (e.g., pulse sequence, scan parameters) of the fat signal cannot be accounted for even when the complex fat spectrum is accurately known. In this work, we present a method that can extract the amplitude and phase dependence of the fat signal directly from the different images by the Dixon acquisition. These results can then be used for processing data of the same pulse sequence and similar scan parameters from different patients.

2484.   Fast and Robust Separation of Multiple Chemical Species from Arbitrary Echo Times with Complete Immunity to Phase Wrapping
Yun Jiang1, Mark Griswold1,2, and Jeffrey Tsao3
1Department of Biomedical Engineering, Case Western Reserve University, Cleveland, Ohio, United States, 2Department of Radiology, Case Center for Imaging Research, University Hospitals of Cleveland and Case Western Reserve University, Cleveland, Ohio, United States, 3McKinsey & Company, Boston, Massachusetts, United States

We report a fast and robust method, hierarchical IDEAL, which has been extended to separate multiple chemical species from images acquired at three or more arbitrary echo times. The new method is completely immune to phase wrapping issues, and the optimized algebraic formulation allows for high computational efficiency. Representative results from human ankle and knee images are shown to demonstrate its performance.

2485.   Phase Combination for Multi-channel RF Coils with a Dual-echo Scan in Water-Fat Imaging
Zhongxu An1,2, Rui Li1, Haining Liu1, Chun Yuan1,3, and Hua Guo1
1Center for Biomedical Imaging Research, Department of Biomedical Engineering, School of Medicine, Beijing, China, 2Department of Electronic Engineering, Beihang University, Beijing, China, 3Department of Radiology, University of Washington, Seattle, WA, United States

Owing to the phase offset existing among different channels, phase images from multi-channel coils are difficult to combine. In this work, a new Multi-Channel Phase Combination with In-Out phase time (MCPC-IO) is proposed and verified by phantom studies. It is also applied in six-point IDEAL water-fat separation. The mean value of fat fraction calculated with MCPC-IO is closer to the true value than with the conventional method. Phase unwrapping is not needed in MCPC-IO, and it can also be simply applied to phase combination in susceptibility weighted imaging.

2486.   Multi-channel T2* IDEAL for Water Fat Separation with Prior Optimized Phase
Haining Liu1, Zhongxu An2, Xinwei Shi3, Chun Yuan4, and Hua Guo1
1Center for Bio-Medical Imaging Research, Tsinghua University, Beijing, Beijing, China, 2Beihang University, Beijing, Beijing, China, 3Tsinghua University, Beijing, Beijing, China,4Department of Radiology, University of Washington, Seattle, WA, United States

Since traditional multicoil water-fat separation methods process data for each channel individually, they are computation intensive and have poor noise resistance performance. This work aims to investigate a new method which can combine multichannel phase images optimally before performing the complex-fitting based water-fat separation. Compared with the traditional method, our method can not only significantly shorten the calculation time but also improve the quantification accuracy and noise resistance. And this method is not limited by in- or out-phase echo times.

2487.   Correction of slice profile deformations and estimation of optimal flip angles to enable accurate T1/T2 mapping using 2D variable flip angle techniques
Matthias Alexander Dieringer1,2, Michael Deimling1,3, Florian von Knobelsdorff-Brenkenhoff1,2, Andreas Greiser3, Jeanette Schulz-Menger1,2, and Thoralf Niendorf1,2
1Berlin Ultrahigh Field Facility, Max-Delbrueck Center for Molecular Medicine, Berlin, Germany, 2Experimental and Clinical Research Center (ECRC), Charité Campus Buch, Humboldt-University, Berlin, Germany, 3Siemens Healthcare, Erlangen, Germany

DESPOT1/2 uses 3D acquisitions for T1/T2 quantification. For 2D acquisitions, however, short repetition times evoke T1 and flip angle dependent saturation phenomena that deform the slice profile and hence bear the potential to deem T1/T2 quantification inaccurate. Therefore, this study examines the impact of slice profile deformations on 2D T1/T2 quantification using variable flip angles in simulations, phantoms, and in-vivo in the brain. A correction method to facilitate accurate quantification of T1/T2 using 2D DESPOT1/2 along with a flip angle optimization algorithm are presented.

2488.   Calalyzing the Refocused TSE Echo Train Facilitates Phase-Detection for Fat-Water Separation
Samuel Fielden1, John Mugler, III1,2, and Craig Meyer1,2
1Biomedical Engineering, University of Virginia, Charlottesville, Virginia, United States, 2Radiology, University of Virginia

Catalyzation techniques are common for bSSFP sequences. Here we apply a linear catalyzation ramp to the refocused TSE sequence in order to force the phase behavior of the spins rapidly into a steady state in order to support phase detection for fat/water separation based on simple phase detection.

2489.   Water Fat Separation with BLADE Based on Two Points Dixon Technique
Dehe Weng1,2
1Beijing MRI Center for Brain Research, Institute of Biophysics, Chinese Academy of Sciences, Beijing, China, 2Siemens Shenzhen Magnetic Resonance Ltd, Shenzhen, Guangdong, China

BLADE based on two points Dixon is proposed, two echoes are acquired to create time dependent phase shifts caused by water fat chemical shift, they are called in-phase and out-of-phase images. The in-phase image is reconstructed with re-gridding, while the out-of-phase image is reconstructed by using all blades from in-phase image for phase correction. Sign detection based on region growing is used to correct the out-of-phase image before the calculation of water and fat images, making this method insensitive to inconsistent error or eddy current.

2490.   Making SENSE of Chemical Shift: Separating Species in Single-Shot EPI using Multiple Coils
Martin Uecker1, and Michael Lustig1
1Electrical Engineering and Computer Science, University of California Berkeley, Berkeley, CA, United States

We present a method for separating chemical shift in single-shot EPI using multiple coils. Chemical shift causes severe artifacts in long echo-train EPI. This artifact is inconsistent with the coil sensitivities. For example, fat would have shifted sensitivity maps with respect to water. By finding a solution consistent with these two sets of maps, separation is achieved. Results are presented using maps obtained from a pre-scan showing high quality images without the need of knowing the field inhomogeneity. In addition, we present a fully automated joint estimation scheme based nonlinear inversion that does not require prior knowledge of the sensitivities.

2491.   Simultaneous EPI-based T1, T1, and T2' mapping
Daniel L Shefchik1, Andrew S Nencka1, Andrzej Jesmanowicz1, and James S Hyde1
1Biophysics, Medical College of Wisconsin, Milwaukee, WI, United States

The accelerated gradient-recalled echo, asymmetric spin-echo pulse sequence was developed to produce relaxivity maps with the same EPI readouts used in functional studies. The sequence had the ability to calculate T1, T2 and T2* voxel-wise relaxivity maps from a single excitation utilizing echo pairs. These echo pair calculations led to systematic errors due to low SNR for T2 and T2* echo pairs and the inaccuracy of the flip angles. A new non-linear model was developed which greatly increase the accuracy of the EPI collected relaxivity maps, while maintaining the ability to provide precise dynamic time series T2 and T2* maps.

2492.   A general analytical solution for optimized fat-suppression calculation
Caixia Fu1, Stephan Kannengiesser2, Dominik Nickel2, Xiaodong Zhou1, Berthold Kiefer2, and Dehe Weng1
1Siemens (Shenzhen) Magnetic Resonance Ltd., Shenzhen, Guangdong, China, 2MR Applications Development, Siemens AG, Healthcare Sector, Erlangen, Germany

This abstract proposes a general analytical solution for optimized fat-suppression calculation: Firstly, the magnetization expressions of fat with spoiled GRE sequence are derived from Bloch equation. And two boundary conditions including steady state of fat signal and optimized fat suppression condition are described. Secondly, a closed analytical equation for fat suppression calculation is derived Based on the magnetization expression of fat and 2 conditions, and the solution to calculate the flip angle for FatSat and TI fill time or TR fill time for SPAIR is described. Finally, In-vivo test result shows that the proposed solution is working well for both FatSat and SPAIR.
Traditional Poster Session - Pulse Sequences & Reconstruction B

Pulse Sequence

Click on to view the abstract pdf. Click on to view the poster (Not all posters are available for viewing.)

Wednesday 9 May 2012
Exhibition Hall  16:00 - 18:00

2493.   Spin Echo and Adiabatic Spin Echo Motion Compensated Spiral functional MRI
Priti Balchandani1, John M. Pauly2, Daniel M. Spielman1, and Gary H. Glover1
1Radiology, Stanford University, Stanford, California, United States, 2Electrical Engineering, Stanford Universiy, Stanford, California, United States

We have developed spin echo and adiabatic spin echo motion compensated spiral fMRI sequences to isolate the small blood vessel contribution to the BOLD signal. The sequences provide improved immunity to B0-inhomogeneity and motion artifacts. The adiabatic version of the sequence utilizes matched phase adiabatic SLR pulses to provide a B1-insensitive spin echo. 3T phantom and in vivo functional scans show more uniform signal-to-fluctuation-noise ratio and higher resolution functional maps are achieved by the spin echo sequences when compared to a conventional gradient echo spiral fMRI sequence.

2494.   Further acceleration of Partial Fourier-FOCUSS using Bunched Phase Encoding
Hisamoto Moriguchi1, Kohki Yoshikawa2, Morio Shimada2, Shin-ichi Urayama3, Yutaka Imai1, Manabu Honda4, and Takashi Hanakawa4
1Radiology, Tokai University, Isehara, Kanagawa, Japan, 2Radiological Sciences, Komazawa University, Tokyo, Japan, 3Human Brain Research Center, Kyoto University, Kyoto, Japan, 4Functional Brain Research, National Center of Neurology and Psychiatry, Tokyo, Japan

Bunched Phase Encoding (BPE) has significant potential for fast data acquisition in MRI. In BPE, k-space data are acquired along zigzag trajectories to reduce the scan time. Partial Fourier-FOCUSS (PF-FOCUSS) has been proposed as a new reconstruction method for PF imaging in standard Cartesian data acquisition. Images reconstructed using PF-FOCUSS are generally of high quality. In this study, BPE acquisition method is taken advantage of in PF-FOCUSS to further accelerate data acquisition while quality of the image is maintained. The newly proposed method is referred to as eBPE-PF-FOCUSSf. BPE-PF-FOCUSS can usually achieve reduction factor 4 with no significant noise amplification.

2495.   FMRI Using 3D PRESTO-CAN - A Novel Method Based on Golden Angle Hybrid Radial-Cartesian Sampling of K-Space
Maria Magnusson1,2, Olof Dahlqvist Leinhard2,3, Helene van Ettinger-Veenstra3,4, and Peter Lundberg2,3
1Electrical Engineering, Linköping University, Linköping, Sweden, 2Radiation Physics, Linköping University, Linköping, Sweden, 3Center for Medical Image Science and Visualization (CMIV), Linköping University, Linköping, Sweden, 4Medical Radiology, Linköping University, Linköping, Sweden

We have shown that it is possible to detect neural activity using fMRI and PRESTO-CAN, a previously developed MRI reconstruction method based on golden angle hybrid radial-Cartesian sampling of k-space. The neural activity appears to be comparable with what can be obtained using a 3D PRESTO Cartesian sequence. However, we have not yet utilized all the facilities of PRESTO-CAN, such as the hourglass filter. The PRESTO-CAN sequence investigated here is able cover the whole brain for fMRI examinations with sufficient temporal resolution.

2496.   Indirect Measurement of Echo Peak Shift for Half Echo Ultrashort TE (UTE) Imaging with Radial Sampling
Masahiro Takizawa1, Hikaru Hanada1, Kuniharu Oka1, and Tetsuhiko Takahashi1
1MRI system division, Hitachi Medical Corporation, Kashiwa, Chiba, Japan

UTE sequence based on half echo radial sampling is very sensitive to small k-space trajectory errors. Many kind of gradient waveform approximation were developed for correcting these k-space trajectory errors. The remaining echo peak shift after applying the gradient waveform approximation causes image degradation. Because UTE sequence acquires only half the echo signal, it is difficult to detect echo peak position directly. We developed indirect echo peak detection algorithm dedicated for half echo signal by using pre-scan. The results shows that this indirect method works well and makes it possible to effectively minimize the error in the k-space trajectory.

2497.   Selective TOF MRA using Beam Saturation pulse
Takashi Nishihara1, Hiroyuki Itagaki1, Chikako Moriwake2, David A. Lampman3, Tetsuhiko Takahashi1, Yosuke Hirata4, Kohsuke Kudo5, and Makoto Sasaki5
1MRI System Division, Hitachi Medical Corporation, Kashiwa, Chiba, Japan, 2Marketing Division, Hitachi Medical Corporation, Tokyo, Japan, 3Development Center of America, Hitachi Medical Systems America, Inc., Twinsburg, OH, United States, 4Center for Radiological Sciences, Iwate Medical University Hospital, Morioka, Iwate, Japan, 5Division of Ultrahigh Field MRI, Institute for Biomedical Sciences, Iwate Medical University, Morioka, Iwate, Japan

We investigated use of a 2D beam excitation pulse (Beam Sat pulse) to saturate the ICA, to realize selective TOF MRA. Beam shape stability was evaluated for B0 inhomogeneity. When B0 inhomogeneity was less than 50Hz, an expansion of FWHM from specification was less than 10%. The position and direction of the Beam Sat pulse were evaluated. As the results, one side ICA was saturated selectively. Saturation effect of signal intensity in the artery was evaluated. For 7 volunteers, the signal intensity from ICA to MCA-M4 was saturated. The Beam Sat pulse is able to saturate the ICA selectively.

2498.   Simultaneously FAT and WATER suppression by interleaved VAPOR, SPAIR and OVS for 1H Spectroscopy at 7T
Baolian Yang1
1BU MR, Philips Healthcare, Cleveland, OH, United States

Water and/or fat suppression is an essential part of pulse sequence in order to get high quality 1H spectroscopy of brain. But it is very difficult to get high water and fat suppression factors in the same time for 1H spectroscopy using the tradition techniques such as dual-band or sequential suppression due to non-uniform B1 field, especially at ultra-high field (>=7T). In this work, an interleaved VAPOR and SPAIR sequence is developed to accomplish simultaneously water and fat suppression with improved suppression factors. OVS pulses can also be added into the sequence to further improve water or fat suppression. The sequence is insensitive to B1 distribution and applicable to both short echo and long echo spectroscopy sequences.

2499.   Multiband Velocity EPI
David Feinberg1,2, Liyong Chen1,2, and An Vu1,2
1Helen Wills Neuroscience Institute, University of California, Berkeley, CA, United States, 2Advanced MRI Technologies, Sebastopol, CA, United States

A new velocity imaging pulse sequence is evaluated by incorporating velocity phase encoding into the multiband simultaneous multi-slice EPI sequence. Six simultaneous images at different levels through the brain were acquired in time series at 100ms TR frame rate to measure CSF, brain motion and arterial velocities without cardiac gating and without combining signals recorded in different cardiac cycles. The technique can be further extended with greater slice coverage for 3D spatial maps in each time frame.

2500.   Very low SAR imaging of the lower leg using variable angle for uniform signal excitation (VUSE) and balanced SSFP without RF phase cycling
Subashini Srinivasan1,2, J Paul Finn1,3, and Daniel B Ennis1,2
1Department of Radiological Sciences, University of California, Los Angeles, California, United States, 2Biomedical Engineering Interdepartmental Program, University of California, Los Angeles, California, United States, 3Biomedical Physics Interdepartmental Program, University of California, Los Angeles, California, United States

Variable flip angle for uniform excitation (VUSE) has been used recently for bSSFP with RF phase alternation to generate constant transverse magnetization during transient imaging. Here, we present the calculation of the flip angle train and simulation for bSSFP without RF phase cycling (VUSEnoalt). The simulation results show that the required flip angle train is very low in magnitude with high off-resonance sensitivity. 2D multi-slice interleaved, segmented images of the lower leg were acquired and the SNR of different tissues were compared with bSSFP and VUSEbSSFP. The results show that the low SAR property of VUSEnoalt may be useful especially when scanning patients with implanted devices or for non-contrast arterial angiography due to the good artery-vein contrast.

2501.   Low power adiabatic T1rho imaging
Ovidiu Cristian Andronesi1, Himanshu Bhat2, Helena Leitao3, Alexander Guimaraes4, and Peter Caravan5
1Martinos Biomedical Imaging Center, MGH, Charlestwon, MA, United States, 2Martinos Center, Siemens Medical Solutions USA, Charlestwon, MA, United States, 3Coimbra University, Portugal, 4Radiology, MGH, 5MGH


2502.   Midbrain nuclei visualization improved by susceptibility-enhanced 3D inversed Double Echo Steady State (iDESS) imaging
Ming-Long Wu1,2, Hing-Chiu Chang3,4, Tzu-Cheng Chao1,2, Pei-Hsin Wu5, and Nan-Kuei Chen6
1Institute of Medical Informatics, National Cheng Kung University, Tainan, Taiwan, 2Department of Computer Science and Information Engineering, National Cheng Kung University, Tainan, Taiwan, 3Global Applied Science Laboratory, GE Healthcare, Taipei, Taiwan, 4Institute of Biomedical Electronics and Bioinformatics, National Taiwan University, Taipei, Taiwan, 5Department of Electrical Engineering, National Taiwan University, Taipei, Taiwan, 6Brain Imaging and Analysis Center, Duke University, Durham, North Carolina, United States

Imaging anatomical structures in human midbrain is important for investigating Parkinson¡¦s Disease (PD). Conventionally, T2* gradient echo method with long echo time (> 20ms) can be used to visualize small nuclei in midbrain. However, clear visualization of midbrain nuclei remains challenging using SPGR due to limited contrast and suboptimal efficiency. In this study, we show that susceptibility-enhanced 3D inversed Double Echo Steady State (iDESS) can be used to efficiently obtain high contrast images for visualization of midbrain nuclei. Results from 1.5 Tesla MRI also demonstrate that iDESS has high potential to improve Deep Brain Stimulation (DBS) guidance in PD.

2503.   Highly Accelerated Single-Slab 3D GRASE with Phase-Independent Image Reconstruction
Hahnsung Kim1, Dong-Hyun Kim1, and Jaeseok Park2
1Electrical and Electronic Engineering, Yonsei University, Seoul, Korea, 2Brain and Cognitive Engineering, Korea University, Seoul, Korea

Single-slab 3D turbo/fast spin echo imaging with short non-selective pulses and variable flip angles in the refocusing pulse train has been recently introduced to increase echo train length and thus enhance imaging efficiency. It has been efforts to further prolong echo train length for high-resolution 3D brain imaging. However, this approach is highly energy-intensive and drops out SNR and CNR within permissible ranges of SAR. To enhance imaging efficiency while retaining SAR limit at high field, in this work we develop highly efficient single-slab 3D GRASE imaging, incorporating: 1) multiple echo planar imaging readouts into a framework of turbo/fast SE imaging and thereby reducing the number of refocusing pulses and 2) parallel imaging assisted phase-independent image reconstruction to remove phase discontinuity related artifacts.
Traditional Poster Session - Pulse Sequences & Reconstruction B

B1/B0 Mapping

Click on to view the abstract pdf. Click on to view the poster (Not all posters are available for viewing.)

Wednesday 9 May 2012
Exhibition Hall  16:00 - 18:00

2504.   Fast Simultaneous B0/B1 Mapping by Bloch-Siegert Shift with Improved Gradient Scheme and Pulse Design
Qi Duan1, Peter van Gelderen1, and Jeff H. Duyn1
1Advanced MRI section, LFMI, NINDS, National Institutes of Health, Bethesda, MD, United States

This abstract investigates fast simultaneous B0/B1 mapping by Bloch-Siegert shift. Based on investigation of the source of the SAR limitation, it proposes several modifications from the original scheme to reduce such limitation while maintaining same sensitivity to the B1 field. Simultaneous B0 mapping is added into the sequence to further improve the accuracy by compensating the increased B0 sensitivity due to reduced frequency offset. The newly proposed scheme is compared to the original scheme both via simulation and in vivo demonstration.

2505.   Simultaneous B1+ and B0 mapping using Dual-Echo Bloch-Siegert (DEBS) sequence.
Paulo Loureiro de Sousa1,2, Alexandre Vignaud3, and Jean-Paul Armspach1
1Université de Strasbourg, CNRS, Laboratoire d'Imagerie et de Neurosciences Cognitives (LINC), Strasbourg, France, 2Institut de Myologie, Paris, France, 3Siemens Healthcare, Saint Denis, France

There is increasing interest in using quantitative magnetization transfer imaging (qMTI) for investigation of the myelin integrity. Because qMTI results obtained from pulsed saturation experiments rely on an accurate knowledge of the power and offset frequency of RF saturation, inhomogeneities of the transmit (B1+) and/or the static field (B0), such as those observed in high field NMR imaging (≥ 3T), will lead to systematic errors in qMTI indices. In this work we investigate the feasibility of fast and simultaneous assessment of B0 and B1+ using a dual-echo Bloch-Siegert sequence (DEBS). In vivo results of field mapping using DEBS are presented.

2506.   Accurate B0 mapping with an adaptive algorithm integrating KESA, PRELUDE, and time-domain phase unwrapping
Pei-Hsin Wu1, Hsiao-Wen Chung1,2, and Nan-Kuei Chen3
1Department of Electrical Engineering, National Taiwan University, Taipei, Taiwan, 2Institute of Biomedical Electronics and Bioinformatics, Taipei, Taiwan, 3Brain Imaging and Analysis Center, Duke University Medical Center, Durham, NC, United States

The accuracy of B0 mapping in dual-TE data is dominated by the decision of the phase accumulation time (ΔTE). Specifically, the pronounced phase wrap-around effects and the susceptibility-induced intravoxel signal loss lead to degradation of B0 map with large ΔTE. In this study, an improved approach is proposed for high quantitative accuracy B0 mapping, where the phase unwrapping for phase accumulation map is accomplished by exploiting information from k-space energy spectrum analysis (KESA) accompanying with the appropriate use of the mask in the initial KESA estimation.

2507.   Shim component analysis for dynamic shimming of the breast
Eunhae Joe1, Yoonho Nam1, Min-Oh Kim1, and Dong-Hyun Kim1
1Electrical and Electronic Engineering, Yonsei University, Seoul, Korea

Respiratory motion can induce fluctuation of magnetic field causing artifacts in magnetic resonance spectroscopy (MRS) and proton resonance frequency shift (PRFS) MR thermometry in the breast. It has been reported that the magnitude of respiratory-induced B0 shift in the breast is ten times greater than in the brain and varies in location. It is therefore important to determine the regional and temporal variation of B0 fluctuation during respiration. The aim of this study is to estimate and analyze the respiratory-induced temporal B0 fluctuation in different location of breast in a three dimensional volume.

2508.   Robust Field Map Estimation using VARPRO and Multi-labeling Continuous Max-Flow
Abraam S. Soliman1,2, Jing Yuan2, Karl Vigen3, James A. White2,4, Terry M. Peters1,2, and Charles A. McKenzie1,5
1Biomedical Engineering, University of Western Ontario, London, Ontario, Canada, 2Imaging Research Laboratories, Robarts Research Institute, University of Western Ontario, London, Ontario, Canada, 3Department of Radiology, University of Wisconsin-Madison, Madison, WI, United States, 4Division of Cardiology, Department of Medicine, University of Western Ontario, London, Ontario, Canada, 5Medical Biophysics, University of Western Ontario, London, Ontario, Canada

Field map estimation has been the primary obstacle for obtaining robust water/fat separation. A novel two-stage approach is proposed to estimate the field inhomogeneities. The first stage applies a convex optimization based algorithm called Continuous Max-Flow that can reliably achieve a global minimum solution. The output is used in the second stage to initialize the well-known IDEAL method to estimate the field map, and hence, water and fat components. Our approach was tested on cardiac images of normal subjects, and has shown to be faster and more robust that existing field map estimation methods.

2509.   Analytical examination on asymmetric distributions of transmission and reception RF fields by a quadrature coil at high magnetic field
Hidehiro Watanabe1
1Center for Environmental Measurement and Analysis, National Institute for Environmental Studies, Tsukuba, Ibaraki, Japan

Maps of the transmission field B1t+ and the reception filed B1r- by a quadrature coil differ and show mirror symmetry in human brain at high field. This is similar to a feature in a linear coil. We analytically examined the source of this mirror symmetry in a quadrature coil. We found that B1t+ and B1r- differ in the sign of the sinusoidal phase difference. This relationship is similar to that in a linear coil. We also examined maps of magnitude and phase components in the laboratory frame computed from B1t+ and B1r- maps of human brains at 4.7 T.

2510.   Experimental Validation of FDTD Magnetic Field Modeling in the Human Head at 7T
Priscilla Chan1, Mohammad Mehdi Khalighi2, and Brian K Rutt3
1Electrical Engineering Department, Stanford University, Stanford, California, United States, 2Global Applied Science Lab, GE Healthcare, Menlo Park, California, United States,3Stanford University, Stanford, California, United States

The trend toward higher performance computing means that numerical EM modeling of coil/patient interactions is becoming much more practical and useful, especially for high field. Using realistic body models of the Virtual Family, provided by IT’IS Foundation, the magnetic and electric fields induced by a quadrature birdcage head coil were simulated using the commercial FDTD package SEMCAD X throughout the human head at 7T. B1+ fields created by driving the coil in single-channel or quadrature modes were compared to experimental measurements. The close match between these B1+ field distributions provides strong evidence of the accuracy of these simulation tools.

2511.   Experimental demonstration of the proportionality of RF reception field to B1-* in a complex vector space
Hidehiro Watanabe1
1Center for Environmental Measurement and Analysis, National Institute for Environmental Studies, Tsukuba, Ibaraki, Japan

The proportionality of the RF reception field to B1-* was demonstrated straightforwardly in the NMR experiments at 4.7 T. Maps of the reception field and B1- of a saline phantom were examined in magnitude and phase. After the reception field map was measured, the RF coil with the sample was rotated 180 degrees around the vertical axis and we measured the transmission field map, which corresponds to B1- in the original configuration. We also examined phases of the reception field maps in the original and inverted configurations. From these results we conclude the proportionality of the reception field to B1-*.

2512.   Regularized Estimation of Magnitude and Phase of Multiple-Coil B1 Field via Bloch-Siegert B1 Mapping
Feng Zhao1, Jeffrey A. Fessler2, Jon-Fredrik Nielsen1, Daehyun Yoon2, and Douglas C. Noll1
1Biomedical Engineering, University of Michigan, Ann Arbor, MI, United States, 2Electrical Engineering, University of Michigan, Ann Arbor, MI, United States

Bloch-Siegert (BS) B1 mapping is a promising method for multi-coil B1 mapping, but it may produce B1 maps with low SNR in low magnitude regions. Furthermore, estimation of relative B1 phase maps of multi-coil systems needs another set of scans. In this work, we propose a regularized method to jointly estimate the magnitude and (relative) phase of multi-coil B1 maps from BS B1 mapping data without using additional scans for the phase estimation. By utilizing the prior knowledge of B1 maps, the regularization terms for magnitude and phase can help improve quality of the B1 maps in low magnitude regions.

2513.   Practical SAR Constraints of the Bloch-Siegert B1 Mapping Method at 3T
Daniel J. Park1, Ahsan Javed1, Neal K. Bangerter1,2, Joshua Kaggie3, and Glen R. Morrell2
1Department of Electrical and Computer Engineering, Brigham Young University, Provo, UT, United States, 2Department of Radiology, University of Utah, Salt Lake City, UT, United States, 3Department of Physics, University of Utah, Salt Lake City, UT, United States

The recently-introduced Bloch-Siegert shift (BSS) method of B1 mapping has several advantages over existing B1 mapping methods. However, a major limitation of the BSS method is the high specific absorption rate (SAR) caused by the BSS pulse. The BSS pulse generates phase which is proportional to B12, and the accuracy of the method increases with increasing amplitude of the BSS pulse. Since SAR of this pulse is proportional to B12, SAR constraints place limits on the performance of the BSS method. This study investigates the tradeoff between SAR limits and BSS pulse area.
Traditional Poster Session - Pulse Sequences & Reconstruction B


Click on to view the abstract pdf. Click on to view the poster (Not all posters are available for viewing.)

Wednesday 9 May 2012
Exhibition Hall  16:00 - 18:00

2514.   Assessment of stiffness changes in the ex vivo porcine aortic wall using magnetic resonance elastography
Lei Xu1,2, Jun Chen1, Meng Yin1, Kevin J Glaser1, Qingshan Chen3, David A Woodrum1, and Richard L Ehman1
1Radiology Department, Mayo Clinic, Rochester, MN, United States, 2Radiology Department, Beijing Anzhen Hospital, Capital Medical University, Beijing, China, 3Orthopedics Biomechanics Lab, Mayo Clinic

MRE is a noninvasive phase-contrast technique for estimating the mechanical properties of tissues. In this study, we hypothesize that changes in arterial wall stiffness, experimentally induced by formalin fixation, can be measured using MRE in ex vivo porcine aortas. In agreement with our hypothesis, the significant stiffness increase after sample fixation were clearly demonstrated by MRE and confirmed by mechanical testing. The results indicate that MRE can be used to examine the stiffness changes of the aorta. This study has provided evidence of the effectiveness of using MRE to directly assess the stiffness change in aortic wall.

2515.   Rayleigh Damped based Brain Magnetic Resonance Elastography
Andrii Petrov1, Elijah E. W. Van Houten2, Matthew D. J. McGarry3, Peter Latta4, and Marco Gruwel4
1Centre for Bioengineering, University of Canterbury, Christchurch, New Zealand, 2Mechanical Engineering, University of Canterbury, Christchurch, New Zealand, 3Thayer School of Engineering, Dartmouth College, Hanover, NH, United States, 4Institute for Biodiagnostics, National Research Council of Canada, Winnipeg, Canada

This research study focuses on application of the subzone based Magnetic Resonance Elastography (MRE) using Rayleigh damped (RD) material model to quantify shear stiffness, damping behavior and elastic energy attenuation mechanism of the intracranial tissue in the in vivo healthy brain. The octahedral shear strain (OSS) SNR calculation confirmed significant attenuation of the shear strain waves in the deeper brain region. The measurement of brain viscoelastic properties revealed that ventricles exhibits much lower elasticity (0.8 kPa) than the surrounding white and gray matter (2.6 kPa). We conclude that RD MRE show promise for potential in vivo determination of different brain tissue types, and the possibility of providing additional diagnostic tools.

2516.   Repeatability of healthy subjects using intrinsically-activated MR elastography
Adam J Pattison1, Matthew DJ McGarry1, John B Weaver1,2, and Keith D Paulsen1,3
1Thayer School of Engineering, Dartmouth College, Hanover, NH, United States, 2Radiology, Dartmouth-Hitchcock Medical Center, Lebanon, NH, United States, 3Norris Cotton Cancer Center, Lebanon, NH, United States

Intrinsically-activated MR elastography has been shown to give reasonable results for brain tissue and detect differences between normal and hydrocephalic patients. However, natural changes in metabolism and blood pressure over time can change the mechanical property signature of a healthy subject. The aim here is to see if intrinsic activation is sensitive to changes in a healthy subject over time and, also, sensitive to changes amongst a healthy population. Results indicate that this technique detects a slightly larger mechanical property variation over time and has a much smaller variation for a single subject than among a group of healthy subjects.

2517.   High-Resolution Multi-Shot MR Elastography of the Brain with Correction for Motion-Induced Phase-Errors
Curtis L. Johnson1,2, Matthew D.J. McGarry3, Armen A. Gharibans1, John B. Weaver4, Keith D. Paulsen3,4, Bradley P. Sutton2,5, and John G. Georgiadis1,2
1Mechanical Science and Engineering, University of Illinois at Urbana-Champaign, Urbana, IL, United States, 2Beckman Institute for Advanced Science and Technology, University of Illinois at Urbana-Champaign, Urbana, IL, United States, 3Thayer School of Engineering, Dartmouth College, Hanover, NH, United States, 4Radiology, Dartmouth-Hitchcock Medical Center, Lebanon, NH, United States, 5Bioengineering, University of Illinois at Urbana-Champaign, Urbana, IL, United States

Recently, MR Elastography (MRE) sequences using multi-shot readouts have been used for higher resolution acquisitions. However, multi-shot MRE is prone to significant phase errors between shots due to inconsistencies in applied motion, resulting in phase cancellation and signal loss. In this work, a multi-shot, variable-density spiral sequence is used, which allows for phase offsets to be estimated for each shot and registered across shots. Results are presented which demonstrate the improvement in image quality, octahedral shear strain-based SNR, and estimated shear modulus estimates. We present 2mm isotropic high-resolution brain MRE data with mechanical estimates obtained with the iterative NLI algorithm.

2518.   Brain MR Elastography using Optimized Mechanical Vibration of the Patient Table
Mikio Suga1,2, Takayuki Obata2, Hajime Ikeda1, Shinya Ozawa1, Atsuhisa Koyama1, Tetsuya Wakayama3, and Riwa Kishimoto2
1Chiba University, Chiba, Chiba, Japan, 2National Institute of Radiological Sciences, Chiba, Japan, 3GE Healthcare, Tokyo, Japan

According to a previous study, the vibrations of the patient table that result from the impulse of an imaging gradient lobe could be used as a mechanical driving mechanism for MR elastography (MRE). However, the vibration frequency must be optimized so that high spatial resolution and accurate storage-modulus maps (elastogram) can be obtained. In this study, we add sinusoidally switching gradients before a conventional MRE pulse sequence and measure the amplitude of the patient table as it vibrates during acquisition. To evaluate the elasticity of the brain using the specific mechanical resonance frequencies of the patient table, heterogeneous phantom and in-vivo brain experiments are performed. The results suggest that the proposed method will enable accurate measurement of the brain elasticity.

2519.   Analysis of Viscoelastic and Poroelastic Behavior in MR elastography
Matthew Mcgarry1, Adam Pattison1, Elijah van Houten2, Curtis Johnson3, Bradley Sutton3, John Georgiadis3, John Weaver4, and Keith Paulsen1,4
1Thayer School of Engineering, Dartmouth College, Hanover, NH, United States, 2Université de Sherbrooke, Sherbrooke, QC, Canada, 3University of Illinois at Urbana-Champaign, Urbana, IL, United States, 4Dartmouth-Hitchcock Medical Center, Lebanon, NH, United States

The applicability of using viscoelastic and poroelastic material models as a basis for MR elastography (MRE) reconstruction algorithms is investigated for two variants of in vivo brain MRE: Standard external actuation, where motion is provided at ~50Hz by an external source, and intrinsic actuation, where motion from the natural pulsatility of the brain due to the cardiac pressure cycle is measured and used to calculate the stiffness distribution. Numerical experiments using viscoelastic and poroelastic finite element models offer an explanation for the observed performance of MRE using each of the two material models for each actuation type.

2520.   Multi-Resolution Reconstruction of Mechanical Properties using Non-Linear Inversion MR Elastography.
Matthew Mcgarry1, Elijah van Houten2, Curtis Johnson3, Bradley P Sutton3, John G Georgiadis3, John Weaver4, and Keith Paulsen1,4
1Thayer School of Engineering, Dartmouth College, Hanover, NH, United States, 2Université de Sherbrooke, Sherbrooke, QC, Canada, 3University of Illinois at Urbana-Champaign, Urbana, IL, United States, 4Dartmouth-Hitchcock Medical Center, Lebanon, NH, United States

MR elastography has proven useful for diagnosing various diseases (such as liver fibrosis), through quantitative images of the storage modulus. Images of alternative parameters such as the viscoelastic loss modulus or level of anisotropy are also available; however, their utility is hampered by higher image noise levels. An important factor determining noise in an MRE parameter image is the sensitivity of the displacements to changes in the parameter value. Low sensitivity will lead to more image noise. This work investigates using different resolutions for each reconstructed parameter type to improve reconstructions of low sensitivity parameters.

2521.   3D MR Elastography of in-vivo prostate cancer and correlation with histology: preliminary results
Ramin Sahebjavaher1,2, Philippe Garteiser2, Ralph Sinkus2, Louis O. Gagnon3, Ali Baghani1, Silvia Chang3, Simon Chatelin2, Edward C. Jones4, Chris Nguan3, Larry Goldenberg3, Piortr Kozlowski5,6, Mehdi Moradi7, and Septimiu Salcudean1
1Electrical and Computer Engineering, University of British Columbia, Vancouver, BC, Canada, 2Hôpital Beaujon, Centre de Recherche Biomédicale Bichat Beaujon (CRB3), Paris, France, 3Department of Urologic Sciences, University of British Columbia, Vancouver, BC, Canada, 4Department of Pathology & Laboratory Medicine, University of British Columbia, Vancouver, BC, Canada, 5Faculty of Medicine, University of British Columbia, Vancouver, British Columbia, Canada, 6The Prostate Centre, Vancouver General Hospital, Vancouver, British Columbia, Canada, 7Department of Radiology, Harvard Medical School, Cambridge, MA, United States

We present our first in-vivo prostate cancer case using the trans-perineal MR elastography (MRE) method. An improved rapid imaging sequence is used to acquire waves in 3D in the entire prostate gland. The mechanical waves efficiently transferred to the entire prostate via the perineum at 70Hz. The results show that the tumours are clearly distinguishable in the viscoelastic maps and that they correlate closely with histopathology. To the best of our knowledge, this is the first in-vivo prostate cancer MRE study where the viscoelastic maps are compared to whole-mount histopathology. These results show the potential for MRE to improve staging of prostate cancer tumours.

2522.   MR Elastography of Hepatic Fibrosis: Elastogram Analysis Strategies
David John Lomas1, Edmund Godfrey1, Andrew Patterson1, Richard Black1, Susan Davies2, Ilse Joubert1, Ashley Shaw1, Anant Krishnan1, Michael Allison3, Graeme Alexander3, Andrew Priest1, and Martin Graves1
1Radiology, Addenbrooke's Hospital & University of Cambridge, Cambridge, Cambridge, United Kingdom, 2Pathology, Addenbrooke's Hospital & University of Cambridge, Cambridge, Cambridge, United Kingdom, 3Hepatology, Addenbrooke's Hospital & University of Cambridge, Cambridge, Cambridge, United Kingdom

MR elastography is a promising new biomarker for quantifying liver fibrosis. This work investigates several ROI placement strategies for the analysis of MRE elastogram data, including a wave detection confidence method, and whether they improve correlation of the stiffness estimates with established histopathological grading schema and semiautomated collagen analysis with Sirius Red staining. These data indicate that using more sophisticated analysis strategies based on mean stiffness region of interest values makes no discernible difference to the correlation with liver biopsy grading.

2523.   Effect of stretching and compression on the dynamic shear modulus using MRE
Remy Blanchard1, Yogesh K. Mariappan2, Richard L. Ehman2, Denis Grenier1, Elisabeth Brusseau1, and Olivier Basset1
1Université de Lyon, CREATIS ; CNRS UMR5220 ; Inserm U1044 ; INSA-Lyon ; Université Lyon 1, Lyon, France, 2Department of Radiology, Mayo Clinic, Rochester, Minnesota, United States

The goal of this study was to analyze the effect of controlled compression and tension on the dynamic shear modulus for the two directions of wave propagation orthogonal to the particle motion. These mediums were a 100% PVC phantom and an ex vivo specimen of bovine liver. For both the phantom and the specimen, the measured shear modulus in both directions increases with the absolute value of the axial strain under both compression and tension. This data provide motivation to further investigate the effects of pre-strain and their clinical utility to improve the information obtained from MRE.

2524.   MR Elastography with Improved Coverage Using 3D Gradient-Echo Based EPI
WeiRui Cai1,2, WenHui Huang1, Hui Wang2, Chao Zou1, ChengZhi Zeng1, LiJuan Zhang1, HaiRong Zheng1, and Yiu-Cho Chung1
1Paul C. Lauterbur Research Center for Biomedical Imaging, Shenzhen Institutes of Advanced Technology, shenzhen, guangdong, China, 2College of Electrical and Information Engineering, Hunan University, Changsha, Hunan, China

MRE is useful in the diagnosis of various diseases such as liver cirrhosis. Multiple breathholds are needed for coverage when 2D gradient echoes are clinically used. EPI techniques previously proposed shorten acquisition time but reduce image SNR. Here, we propose 3D gradient echo based EPI for speed and coverage improvement and compared it with 2D gradient echo MRE sequence. Using the same scan time, we found that, the 3D technique gives slightly higher SNR than the 2D technique; it covers 6 slices instead of one; and there was no obvious difference between the stiffness maps from the two methods.

2525.   MR elastography of the prostate using an endorectal coil for actuation: feasibility in a phantom and porcine prostate
Gregor Thörmer1, Martin Reiss-Zimmermann1, Josephin Otto1, Nikita Garnov1, Michael Moche1, Thomas Kahn1, and Harald Busse1
1Dept. of Diagnostic and Interventional Radiology, Leipzig University Hospital, Leipzig, Saxony, Germany

MRE is an emerging technique that measures the propagation of mechanical waves to noninvasively determine the viscoelastic tissue properties. As such, it provides complimentary information to common morphological and functional MRI data, has already been evaluated in the breast and liver, and holds promise to improve prostate diagnostics as well. Previous prostate MRE work was partially limited by a suboptimal coupling of the mechanical waves into the prostate gland. This work describes a new concept for endorectal MRE at 3T and presents preliminary results in a phantom and a porcine prostate.

2526.   Design and Fabrication of a Driver for MR Elastography of the Head and Neck: Initial Results
David K.W. Yeung1, Kunwar Bhatia2, Yolanda Lee1, Ann D King2, Ralph Sinkus3, and Anil T Ahuja2
1Imaging and Interventional Radiology, Prince of Wales Hospital, Shatin, HKSAR, Hong Kong, 2Imaging and Interventional Radiology, The Chinese University of Hong Kong, Shatin, HKSAR, Hong Kong, 3Department of Radiology and IPMA, Beaujon University Hospital, Clichy, France

MR elastography is a technique that allows the assessment of mechanical properties of tissues that may be affected by disease processes. We designed and tested a new driver to allow MRE to be performed in the head and neck using a commercially available transducer, headrest moulds and piston extensions. This new driver design was simple to implement and comfortable to use. Our initial results showed that the performance of the driver was satisfactory and provided reproducible results for the normal parotid glands. This simple driver may allow a more widespread application of this technique to study head and neck pathologies.

2527.   Finite Element Simulations of acoustic radiation contrast in magnetic resonance using open source software
Anna-Lisa Kofahl1, Sebastian Theilenberg1, Deniz Ulucay1, Judith Wild1, Saskia Paul1, Karl Maier1, and Carsten Urbach1
1HISKP, University of Bonn, Bonn, NRW, Germany

Investigating the displacement caused by the ARF via MR phase images offers a new access to the viscoelastic properties of tissue and has been successfully tested on phantoms. For a better understanding of the measurements finite element (FE) simulations have been developed with special regard to the influence of boundary conditions. These show an immense influence on the simulation and the measurements and have to be considered when carrying out more quantitative investigations. The open source FE software that is used performs as satisfactorily as commercial software.

2528.   Measured elasticity and its frequency dependence are sensitive to tissue microarchitecture in MR Elastography
Simon CHATELIN1, Simon LAMBERT1, Lauriane JUGE1, Xing CAI2, Sven Peter NASHOLM2, Valérie VILGRAIN1, Bernard E Van BEERS1, Xavier MAITRE3, Lynne E BILSTON4, Bojan GUZINA5, Sverre HOLM2, and Ralph SINKUS1
1U773-CRB3, INSERM, Paris, France, 2Dept. of Informatics, University of Oslo, Oslo, Norway, 3UMR 8081-CNRS, Université Paris-Sud, Orsay, France, 4Neuroscience Research Australia, University of New South Wales, New South Wales, Australia, 5Dept. of Civil Engineering, University of Minnesota, Minneapolis, MN, United States

Understanding the effects of micro-obstacles on wave propagation is an essential part when trying to extract micro-structural information from MR-Elastography (MRE) data from tissue abnormalities such as small metastases or neovascularization. To date, the effects of wave scattering on mechanical properties measurements remain poorly understood. Nevertheless, scattering plays a major role in linking the architectural properties of a biological tissue to its mechanical properties measured by MRE. In this study, the influence of microparticle size distribution on wave scattering frequency dependence is investigated from a theoretical and numerical approach. This study shows for the first time the feasibility of linking size distribution and stiffness of micro-particles to the macroscopically observed elasticity.
Traditional Poster Session - Pulse Sequences & Reconstruction B


Click on to view the abstract pdf. Click on to view the poster (Not all posters are available for viewing.)

Wednesday 9 May 2012
Exhibition Hall  16:00 - 18:00

2529.   High Resolution EPR Imaging and T2-based Oximetry Using a Combination of Spin-Echo and Single Point Imaging
Sankaran Subramanian1, Nallathamby Devasahayam1, Shingo Matsumoto1, and Cherukuri M Krishna1
1National Cancer Institute, National Institutes of Health, Bethesda, Maryland, United States

Time-domain Electron Spin Resonance Imaging at RF frequency (300 MHz) has been developed for mapping oxygen distribution in mouse models of tumor. Oximetry is based on the relaxivity of in vivo pO2 on trityl radicals infused intravenously into the animals. A combination of spin echo and single point (constant-time) imaging (SPI) leads to highly resolved images with T2-weighting and generates quantitaive sptailly resolved oxygen ditribution maps.

2530.   Fast Gated EPR Imaging of the Beating Heart: Spatiotemporally-Resolved 3D Imaging of Free Radical Distribution during the Cardiac Cycle
Zhiyu Chen1, Levy A. Reyes1, David H. Johnson1, Murugesan Velayutham1, Changjun Yang1, Alexandre Samouilov1, and Jay L. Zweier1
1Davis Heart and Lung Research Institute, Ohio State University Medical Center, Columbus, OH, United States

In vivo or ex vivo electron paramagnetic resonance imaging (EPRI) is a powerful technique for determining the spatial distribution of free radicals and other paramagnetic species in living organs and tissues. However, applications of EPRI have been limited by long projection acquisition times and the consequent fact that rapid gated EPRI was not possible. Hence in vivo EPRI typically provided only time-averaged information. In order to achieve direct gated EPRI, a fast EPR acquisition scheme was developed to decrease EPR projection acquisition time down to 10 – 20 ms, along with corresponding software and instrumentation to achieve fast gated EPRI of the beating heart with submillimeter spatial resolution in as little as 2 to 3 minutes. Reconstructed images display the temporal and spatial variations of the free radical distribution, anatomical structure, and contractile function of a rat heart during the cardiac cycle.

2531.   Temporal Tomographic Imaging By Greedy Feteke and Golden Means Sampling
David H Johnson1, Zhiyu Chen1, Rizwan Ahmad1, Alexandre Samouilov1, and Jay L Zweier1
1Davis Heart and Lung Research Institute, Ohio State University, Columbus, OH, United States

A new sampling strategy is proposed and evaluated for tomographic imaging when temporal information is critical. Typical Electron Paramagnetic Resonance Imaging (EPRI) sampling strategies assume a static object and require a priori knowledge about its spatial distribution to produce reconstructions with few streak artifacts and accurate intensities. The new sampling strategy, Greedy Feteke (GF), was compared to a sampling strategy from the MRI projection acquisition literature, Golden Means (GM). A dynamic "washout" phantom was imaged by EPRI using both GF and GM, and GF was found to have superior temporal resolution to GM. The extension to 4D spectral-spatial EPRI is straightforward and compelling for GF sampling, and we anticipate a significant improvement in temporal cardiac EPRI.
Traditional Poster Session - Pulse Sequences & Reconstruction B

Electromagnetic Tissue Property Mapping

Click on to view the abstract pdf. Click on to view the poster (Not all posters are available for viewing.)

Wednesday 9 May 2012
Exhibition Hall  16:00 - 18:00

2532.   Breast Permittivity Imaging
Selaka Bandara Bulumulla1, and Ileana Hancu1
1GE Global Research, Niskayuna, New York, United States

Malignant breast tissue has been found to have higher permittivity and conductivity relative to healthy tissue ex-vivo, generating interest in non-invasive methods to image breast tissue electrical properties. While non-invasive imaging of electrical properties using B1+ maps has been demonstrated for the brain, the method has not been extended to other anatomical parts, partly due to lack of transmit-receive coils to obtain accurate B1+ maps. We investigate the feasibility of breast permittivity imaging using an eight channel, receive only breast coil array and demonstrate permittivity image consistent with adipose dominated breast tissue as well as fat suppressed anatomical image.

2533.   Quantification Error in MREPT due to B1 Map Inaccuracy
Jaewook Shin1, Joonsung Lee1, Jin Keun Seo2, and Dong-Hyun Kim1
1Electrical & Electronic Engineering, Yonsei University, Seoul, Korea, 2Computational Science & Engineering, Yonsei University, Seoul, Korea

Magnetic Resonance Electrical Property Tomography (MREPT) is recently introduced to noninvasively image the distribution of electric conductivity and permittivity in the human body at the Larmor frequency. Over homogeneous regions, the relationship between the electric properties and magnetic fields can be simplified as a Helmholtz equation of B1+. In this work, the methods using B1+ maps generated by EM simulations on numerical phantoms with two tissues were proposed to investigate and evaluate how the error in the conductivity estimates using the Helmholtz equation of B1+ is related to the SNR of B1+ and the size of the tissue.

2534.   Feasibility of Submillimeter Resolution MREIT Conductivity Imaging: Preliminary Phantom Study
Hyung Joong Kim1, Woo Chul Jeong1, Young Tae Kim1, Munish Chauhan1, and Eung Je Woo1
1Biomedical Engineering, Kyung Hee University, Yongin, Gyeonggi, Korea

Recent experimental MREIT studies demonstrated conductivity image reconstructions of in vivo animal and human subjects with a few millimeter pixel size. Such a resolution could be sufficient to distinguish different human tissues for certain applications. In order to visualize a conductivity distribution within a tissue or a small animal, it is necessary to develop a high-resolution MREIT technique. The experimental results using a 4.7 T research MRI scanner with a multi-echo ICNE pulse sequence and high-performance RF coils demonstrate that we can distinguish two different anomalies in reconstructed conductivity images with a pixel size as small as 400 ¥ìm.
Traditional Poster Session - Pulse Sequences & Reconstruction B

Image Reconstruction

Click on to view the abstract pdf. Click on to view the poster (Not all posters are available for viewing.)

Wednesday 9 May 2012
Exhibition Hall  16:00 - 18:00

2535.   A Novel Method for Coil Array Synthesis and Application to Breast MRI
Anderson Nnewihe1,2, Bruce L Daniel1, Deqiang Qiu1, Jafi Lipson1, Jarrett Rosenberg1, Debra Ikeda1, Sung H Kim1,3, Jennifer Kao1, and Brian A Hargreaves1
1Radiology, Stanford University, Stanford, California, United States, 2Bioengineering, Stanford University, Stanford, California, United States, 3Radiology, Seoul St. Mary's Hospital, The Catholic University of Korea, Seoul, Korea

The primary purpose of this work is to introduce and validate a method of synthesizing image data for an RF coil array given data from a second array. This would enable future studies comparing two coil arrays (source and reference arrays) for situations where it is not practical to acquire data with both such as breast contrast-enhanced MRI. For the future studies, the synthesized images would be used as a proxy for the reference coil array data. In a survey of experienced breast radiologists, we showed equivalence of the overall preference between the synthesized and reference coil array images.

2536.   Vessel-Mimicking Mask for Improved Unfolding in SENSE-Accelerated CE-MRA
Eric G. Stinson1, Casey P. Johnson1, and Stephen J. Riederer1
1MR Research Laboratory, Mayo Clinic, Rochester, Minnesota, United States

Masking in SENSE unfolding is known to reduce noise amplification and preserve SNR. Further, in contrast-enhanced MR angiography (CE-MRA), the final volume of interest is often a difference volume where, in the ideal case, signal is only present in the vasculature. The purpose of this work is to investigate the feasibility of improving the performance of 2D SENSE in 3D CE-MRA by masking all non-vascular voxels prior to performing the SENSE unfolding of a difference angiogram. Thigh CE-MRA exams were unfolded with conventional and vessel-mimicking masks and compared with respect to maximum intensity projection contrast-to-noise ratio and mean g-factor.

2537.   An Inexpensive Iterative Reconstruction for Under-sampled PROPELLER MRI
Jyh-Miin Lin1,2, Tzu-Chao Chuang2, Hing-Chiu Chang3,4, Wen-Chau Wu5, and Hsiao-Wen Chung4,6
1Department of Radiology, Duke University Medical Center, Durham, NC, United States, 2Department of Electrical Engineering, National Sun Yat-sen University, Kaohsiung, Taiwan, 3Global Applied Science Laboratory, GE Healthcare, Taipei, Taiwan, 4Institute of Biomedical Electronics and Bioinformatics, National Taiwan University, Taipei, Taiwan,5Graduate Institute of Oncology, National Taiwan University, Taipei, Taiwan, 6Department of Electrical Engineering, National Taiwan University, Taipei, Taiwan

Previously researchers had shown the possibility to use iterative reconstruction to process PROPELLER images. In this study, an economic algorithm named image space reconstruction algorithm (ISRA) is implemented for under-sampled PROPELLER MRI. Based on generalized inverse, the algorithm is fast and inexpensive, which is useful for dynamic or volumetric imaging. We performed phantom study with under-sampled PROPELLER images, in which the image qualities were restored by ISRA.

2538.   Automatic Regularization Parameter Selection for Iterative Nonlinear MRI Reconstruction
Sathish Ramani1, Jon-Fredrik Nielsen2, and Jeffrey A. Fessler1
1EECS, University of Michigan, Ann Arbor, Michigan, United States, 2fMRI Laboratory, University of Michigan, Ann Arbor, Michigan, United States

MRI reconstruction from undersampled sampled k-space data requires regularization to reduce artifacts. Nonquadratic regularization (e.g., l1 or TV) can be used to restore image quality, but its successful application depends on proper selection of the regularization parameter. In this work, we demonstrate the applicability of generalized cross validation (GCV) and an estimate of a mean-squared error (MSE)-type measure for quantitative selection of the regularization parameter for MRI reconstruction using the iterative split-Bregman algorithm with nonquadratic regularization. GCV and the MSE-type estimate require the Jacobian matrix of the reconstruction with respect to the data that we evaluate analytically for the SB algorithm. We illustrate with experiments on real MR (phantom) data that GCV and the MSE-type estimate lead to near-optimal reconstruction results.

2539.   Faster and More Robust MRI Using Sharable Information Among Images With Different Contrasts
Feng Huang1, Yu Li2, Wei Lin1, George Randy Duensing1, and Arne Reykowski1
1Invivo Corporation, Gainesville, FL, United States, 2Radiology Department, Cincinnati Children's Hospital Medical Center

A typical clinical MR examination is composed of a number of scans to acquired images with different contrasts, such as T1w, T2w and DWI. Currently, the acquisition and reconstruction schemes of these scans are independent to each other. In the presented work, we propose to jointly optimizing the acquisitions and reconstructions of a multi-scan MRI examination for faster and more robust MRI. The theory behind this concept is that the information such as B0 field, optimized acquisition trajectory, reconstruction parameters, etc, can be shared among all scans for different contrasts since the same subject is scanned in the same system using the same RF coil. This idea is fundamentally different from previous method that using one set of acquisition to artificially produce images with different contrast.

2540.   Improving SNR and Sensitivity of 3D Fluorine Imaging With a Side-Information-Constrained Regularized Reconstruction
Samuel Fielden1, Alexander Klibanov1,2, Yaquin Xu1, Frederick Epstein1,3, Brent French1,3, and Craig Meyer1,3
1Biomedical Engineering, University of Virginia, Charlottesville, Virginia, United States, 2Medicine, University of Virginia, 3Radiology, University of Virginia

Second color, or hot spot, fluorine imaging benefits from 3D acquisition from both a signal localization point of view and a desire to maximize signal sensitivity. Here we show fluorine nanoparticle-labeled macrophage infiltration into infarcted myocardium using a gated 3D turbo spin echo sequence, and improve the SNR and sensitivity using a regularized reconstruction technique which utilizes side information obtained from proton scout images.

2541.   Vessel Adapted Regularization for Iterative Reconstruction in MR Angiography
Jana Hutter1,2, Robert Grimm1, Christoph Forman1,2, Joachim Hornegger1,2, and Peter Schmitt3
1Pattern recognition lab, Universität Erlangen-Nürnberg, Erlangen, Germany, 2Erlangen graduate school in avanced optical technologies, Erlangen, Germany, 3MR Application & Workflow Development, Siemens AG, Healthcare Sector, Erlangen, Germany

Iterative methods such as Compressed Sensing involve regularization terms to stabilize and accelerate the optimization. Smoothness and sparsity assumptions are widely used as regularization, but the special structure of angiographic data is not yet part of the regularization. Our new method takes the localization and brightness of vessels, adapted to each individual vessel, into account. The information is extracted using an ellipsoid-based segmentation approach and included into the reconstruction with a modified Gaussian penalty map. This approach proved to be beneficial for the reconstruction time as well as for the general image quality.

2542.   GRAPPA Operator Shift Correction for Non-Cartesian Imaging Trajectories
Anagha Deshmane1, Nicole Seiberlich1, Jeffrey Duerk1,2, and Mark A. Griswold1,2
1Dept. of Biomedical Engineering, Case Western Reserve University, Cleveland, OH, United States, 2Dept. of Radiology, University Hospitals of Cleveland, Cleveland, OH, United States

Gradient timing delays can cause deviations from the ideal trajectory in non-Cartesian imaging. The proposed method aims to correct for trajectory errors and determine the signal at the center of k-space. Paths between points acquired along mis-centered radial projections and the true center of k-space are traced by a gradient ascent algorithm which iteratively applies a GRAPPA operator to shift k-space data while continually evaluating an estimate of the echo peak signal. This method can be used to perform trajectory measurement without field monitoring or extra acquisitions, and signal at the true k-space center can be used for retrospective self-gated imaging.

2543.   Inter-slice artifact reduction for slice-GRAPPA reconstruction of simultaneous multi-slice (SMS) acquisitions
Stephen F Cauley1, Kawin Setsompop1,2, Jonathan R Polimeni1,2, and Lawrence L Wald1,3
1A. A. Martinos Center for Biomedical Imaging, Dept. of Radiology, MGH, Charlestown, MA, United States, 2Harvard Medical School, Boston, MA, United States, 3Harvard-MIT Division of Health Sciences and Technology, MIT, Cambridge, MA, United States

For full-brain coverage simultaneously acquiring multiple slices can significantly improve acquisition time. Controlling the unaliasing process is extremely important given the presence of noise and artifacts in fMRI and diffusion. In this work we test a constrained optimization technique that reduces inter-slice artifacts by more than 10 fold. The improved kernels can be used in both fMRI and diffusion weighting studies for increased accuracy. The convex model ensures optimal solutions given the constraints and is implemented easily using readily available optimization packages.

2544.   Partial Fourier accelerated spiral SENSE imaging using magnetic field monitoring
Bertram Jakob Wilm1, Christoph Barmet1, Lars Kasper1, Max Haeberlin1, Benjamin Emanuel Dietrich1, and Klaas Paul Pruessmann1
1Institute for Biomedical Engineering, ETH Zurich and University of Zurich, Zurich, Switzerland

A combination of spiral MRI with partial Fourier phase constrained imaging could allow decreasing the acquisition duration, and thus to increase the achievable temporal resolution in MRI. However, accurate knowledge about the image phase that is required for partial Fourier imaging is hard to establish, since spiral imaging is very sensitive to any encoding deficiencies, such as gradient delays, eddy currents, concomitant fields and static B0 inhomogeneity. To overcome this problem, an accurate model of the dynamic field encoding is obtained by using dynamic field monitoring. The performance and the sensitivity against inaccuracy of the phase estimate are evaluated in vivo.

2545.   Multi-Coil Imaging with Algebraic Reconstruction
Christoph Juchem1, Terence W Nixon1, and Robin A de Graaf1
1Yale University School of Medicine, New Haven, CT, United States

Today’s MR imaging methods use low-order SH shapes or combinations thereof. Here we show that MR imaging does not rely on SH shapes and can also be achieved with random magnetic field shapes when analyzed by algebraic reconstruction. Magnetic fields for spatial encoding were generated with the multi-coil (MC) approach by a matrix of simple, circular coils.

2546.   A Density Compensation Function using Kaiser-Bessel Regridding
Lawrence Dougherty1, and Gamaliel Isaac1
1Radiology, University of Pennsylvania, Philadelphia, PA, United States

A rapid density compensation function (DCF) was developed and tested for reconstruction of non-Cartesian sampled data. Sampling locations were gridded onto a map in the same manner as the MR data. This creates a density map that is used to compensate for the irregular sampling. Phantom images were reconstructed using the gridded DCF from data acquired with a Golden-angle radial sequence and showed good performance as compared to a Ram-Lak filter or density computed from K-space area. The method is rapid and flexible, requiring only the knowledge of each sample’s position in K-space and is simple to implement.

2547.   Sampling Density Compensation Function Estimation by Regularized Conjugate Gradient Iteration with a Reduced Oversampling Ratio
Sen Jia1, and Ran Yang1
1School of Information Science and Technology, Sun Yat-Sen University, Guang Zhou, Guang Dong, China

The iterative convolution based estimation of non-Cartesian sampling density compensation function (DCF) typically employs an intermediate Cartesian grid oversampled by a ratio of two to achieve high convergence accuracy and this increases the memory and computational burden heavily when processing data of large size or of high dimension. And the iteration converges stably but saturates easily which limits the capacity to achieve higher estimation accuracy. In this work, an iterative procedure based on the Conjugate Gradient (CG) method is introduced to reduce the oversampled grid size while achieving higher DCF estimation accuracy without increasing computational and storage burden. Further the DCF estimated by the original method is employed as a regularization term to form a bi-criterion optimization problem to stabilize the CG iteration to achieve fast convergence with a reduced oversampling ratio.

2548.   Generalized Conjugate Gradient Image Reconstruction – Combination of SENSE/SPIRiT, Chemical Shift Imaging and Multi-Frequency Interpolation
Jonathan I. Sperl1, Florian Wiesinger1, and Rolf F. Schulte1
1GE Global Research, Garching n. Munich, BY, Germany

Image Reconstruction in MRI is addressed by a variety of techniques depending on the specific data acquisition scheme and the information one wants to resolve. This work proposes a general Conjugate Gradient reconstruction framework that can be applied to a wide range of MRI data sets. In more detail, we propose a generalized SENSE or SPIRiT type of reconstruction combining chemical shift imaging, e.g. to resolve different metabolites in vivo, and Multi-Frequency Interpolation for B0-correction. The method is applied to a severely B0-distorted 1H data set as well as to a 13C phantom experiment.

2549.   Generating synthetic DIR images using multi-parameter maps: application to lesion detection in MS
Riddhi Rajgor1, Nils Muhlert2, Antoine Lutti3, Matteo Atzori4,5, Nikolaus Weiskopf3, Claudia AM Wheeler-Kingshott2, Xavier Golay1, Alan J Thompson4, Olga Ciccarelli4, and David L Thomas1
1Department of Brain Repair and Rehabilitation, UCL Institute of Neurology, London, United Kingdom, 2NMR Research Unit, Department of Neuroinflammation, UCL Institute of Neurology, London, United Kingdom, 3Wellcome Trust Centre for Neuroimaging, UCL Institute of Neurology, London, United Kingdom, 4NMR Research Unit, Department of Brain Repair and Rehabilitation, UCL Institute of Neurology, London, United Kingdom, 5Department of Neurology, University of Padova, Padova, Italy

DIR brain imaging is widely applied to patients with multiple sclerosis, to suppress the signals from white matter and CSF and facilitate the visualisation of lesions. However, DIR is an inherently inefficient technique, since long inversion and repetition times are required to allow the desired contrast to evolve. The aim of this work was to investigate the potential of using FLASH-based multi-parameter mapping to generate synthetic DIR images with equivalent image SNR and contrast. It was found that synthetic DIR images suppress unwanted tissue signals well, but do not demonstrate grey matter lesions as clearly as acquired DIR images.
Traditional Poster Session - Pulse Sequences & Reconstruction B

Computational Implementation & Tools

Click on to view the abstract pdf. Click on to view the poster (Not all posters are available for viewing.)

Wednesday 9 May 2012
Exhibition Hall  16:00 - 18:00

2550.   More IMPATIENT: A Gridding-Accelerated Toeplitz-based Strategy for Non-Cartesian High-Resolution 3D MRI on GPU
Jiading Gai1, Joseph Lee Holtrop2, Xiao-Long Wu3, Fan Lam3, Maojing Fu3, Justin P. Haldar4, Wen-mei W. Hwu3, Zhi-Pei Liang3, and Bradley P. Sutton2
1Beckman Institute, University of Illinois at Urbana-Champaign, Urbana, IL, United States, 2Department of Bioengineering, University of Illinois at Urbana-Champaign, Urbana, IL, United States, 3Electrical and Computer Engineering, University of Illinois at Urbana-Champaign, Urbana, IL, United States, 4Electrical Engineering, University of Southern California, Los Angeles, CA, United States

We further accelerate the Illinois Massively Parallel Acceleration Toolkit for Image reconstruction with ENhanced Throughput in MRI (IMPATIENT MRI) package to approach clinically- acceptable times while still taking advantage of a variety of advanced image acquisitions and reconstruction techniques. The improved IMPATIENT implemented a faster Toeplitz-based iterative image reconstruction method, whose computation time is further reduced by an optimally tuned, GPU- accelerated gridding implementation. We demonstrate that the Toeplitz code running on a NVIDIA Tesla C1060 (field- corrected, SENSE) can reduce a one-week long, non-Cartesian 3D 1mm3 high-resolution, whole brain DTI reconstruction (4-channel acquisition) to 4.3 hours. These improvements will enable advances in 3D non-Cartesian sequences, such as cones and stacks of spirals.

2551.   High-Performance Gridding on Modern x86-based Multi-core Systems for 3D Non-Cartesian MRI
Dhiraj D. Kalamkar1, Joshua D. Trzasko2, Srinivas Sridharan1, Mikhail Smelyanskiy3, Daehyun Kim3, Yunhong Shu4, Matt A. Bernstein4, Bharat Kaul1, Pradeep Dubey3, and Armando Manduca2
1Parallel Computing Lab, Intel Labs, Bangalore, KA, India, 2Mayo Clinic, Rochester, MN, United States, 3Parallel Computing Lab, Intel Labs, Santa Clara, CA, United States,4Department of Radiology, Mayo Clinic, Rochester, MN, United States

With increasing usage of higher-resolution acquisitions, more receiver channels, and iterative reconstruction strategies, the ability to quickly and accurately transform an image to and from k-space, known as “reverse gridding” and “gridding”, is crucial for non-Cartesian MRI applications. In practice, both of these operations are typically realized via the non-uniform fast Fourier transform (NUFFT). In this work, we propose a novel preprocessing and parallelization strategy for both the forward and adjoint NUFFT targeted for x86 architectures. We demonstrate that this implementation strategy, which is based on a variable-size geometric partitioning along with a barrier-free task queue, and selective privatization, is substantially faster than contemporary x86 implementations, and computationally competitive with state-of-the-art GPU implementations.

2552.   Parallel-Computing Reconstruction for Parallel Encoding with Parallel Imaging
Ethan M Johnson1, and John M Pauly1
1Electrical Engineering, Stanford University, Stanford, CA, United States

Recent innovation in MRI has tended to realise improvements to image quality or reduction of acquisition time by 'supplementing' the innate Fourier encoding of magnetic resonance imaging. This typically complexifies the data-image relationship, such that reconstruction can become computationally expensive. Within this class of innovations are techniques which attempt to 'parallelise' the acquisition, with multiple receiver coils and non-bijective encoding fields. In these cases, reconstruction computations naturally exhibit parallelism that can be exploited to reconstruct images very quickly.

2553.   GPU-Accelerated Radial Image Reconstruction with an Improved Parallel Gridding Method
Jiangsheng Yu1, Yiqun Xue1, Xia Zhao1, Ping Wang1, Walter Witschey1, and Hee Kwon Song1
1University of Pennsylvania, Philadelphia, PA, United States

Graphic processing units (GPU) have recently been utilized to accelerate radial MR image reconstruction. One of the challenges of gridding using GPU is the issue of synchronization, and several approaches have been reported to address this issue. This work presents a simple and effective solution for avoiding this synchronization problem by utilizing a divide-and-conquer strategy that ensures that the gridded points are non-overlapping. Performance tests on the reconstruction of a DCE-MRI data set demonstrates that an acceleration factor of 67 can be achieved on an Nvida Tesla C2050 GPU system.

2554.   Low-Latency Radial GRAPPA Reconstruction using Multi-Core CPUs and General Purpose GPU Programming
Haris Saybasili1,2, Daniel A. Herzka2, Nicole Seiberlich3, and Mark Griswold1,3
1Department of Radiology, Case Western Reserve University, Cleveland, OH, United States, 2Biomedical Engineering, Johns Hopkins University School of Medicine, Baltimore, MD, United States, 3Biomedical Engineering, Case Western Reserve University, Cleveland, OH, United States

Real-time imaging using non-Cartesian trajectories permits very high acceleration rates with parallel imaging. However, image reconstruction from undersampled non-Cartesian datasets is computationally demanding, and leads to long reconstruction times .We present a hybrid (CPU- and GPU-based), fully auto-calibrated, fast reconstruction implementation for radial GRAPPA that uses RT-GROG to grid radial data prior to FFT. Radial GRAPPA and RT-GROG calibrations were performed on the CPU, while image reconstruction was performed asynchronously on the GPU. Our implementation was tested on healthy volunteer cardiac data for different parameters. Images from 12 coil, 144x256, R=9 data were reconstructed in 33ms/frame (acquisition time 42ms/frame).

2555.   Gridding: Exploring an Efficient Numerical Algorithm for 0-th Order Prolate Spheroidal Wave Function Evaluation as Convolution Kernel
Laurent Lamalle1
1RMN biomédicale et Neurosciences — SFR UJF, Inserm, Grenoble, Isère, France

An algorithm recently described for the efficient numerical evaluation of Prolate Spheroidal Wave Functions (PSWFs) was implemented in order to explore the possibility of using the 0-th order PSWF as convolution kernel in gridding reconstruction, instead of its Kaiser-Bessel approximation. The problem of evaluating the compensation function necessary after FFT of the samples interpolated by convolution with the kernel is addressed.

2556.   Fast l1-SPIRiT Compressed Sensing Parallel Imaging MRI: Scalable Parallel Implementation and Clinically Feasible Runtime
Mark Murphy1, and Miki Lustig2
1EECS UC Berkeley, Berkeley, CA, United States, 2EECS UC Berkeley

We describe the algorithmic and parallelization implementation decisions that lead to fast runtimes in l1-SPIRiT combined Parallel Imaging and Compressive Sensing reconstruction.

2557.   System for Real-time Reconstruction and Interactive Table Motion in Multi-Station 3D CE-MRA Exams
Eric A. Borisch1, Casey P. Johnson1, Roger C. Grimm1, and Stephen J. Riederer1
1MR Research Laborary, Mayo Clinic, Rochester, MN, United States

A system is described that enables multi-station interactive bolus-chase CE-MRA examinations. The system consists of: a custom pulse sequence with network connectivity, interactive graphical interface for the operator, and high-performance reconstruction software and hardware. Reconstructions of the imaged 3D time-series (acquired with 2.5s update times) are completed and displayed to the operator within 110ms while the examination is proceeding, enabling sufficient time for the operator to visualize bolus transit to determine if the current station is complete. The system has been used successfully for three-station lower extremity examinations.

2558.   StimFit: A toolbox for robust T2 mapping with stimulated echo compensation
R. Marc Lebel1
1Department of Electrical Engineering, University of Southern California, Los Angeles, CA, United States

Quantitative T2 mapping provides an absolute measure of transverse relaxation that is independent of other confounding factors; however, heterogeneous transmit fields and shaped RF pulses preclude the use a mono- or multi-exponential fit to measure T2. We present a freely available toolbox for robust quantitative T2 mapping with stimulated echo compensation based on recently published work. The toolbox enables accurate mono- or multi-component analysis despite heterogeneous transmit fields and slice-selective refocusing; the toolbox can be obtained from

2559.   A Digital Perfusion Phantom for T1-weighted DCE-MRI
R. Grimm1, J. Churt2, A. Fieselmann1, K. T. Block3, B. Kiefer4, and J. Hornegger1
1Pattern Recognition Lab, University of Erlangen-Nuremberg, Erlangen, Germany, 2Technical University Munich, Munich, Germany, 3Department of Radiology, NYU Langone Medical Center, New York, NY, United States, 4MR Application & Workflow Development, Siemens AG, Healthcare Sector, Erlangen, Germany

Novel reconstruction techniques such as Compressed Sensing promise high acceleration factors at a low noise level but do not guarantee to preserve the correctness of image intensities. This, however, is a prerequisite for quantitative evaluations in Dynamic Contrast-Enhanced MRI. To study these effects, an analytical phantom is proposed that allows the simulation of contrast agent kinetics in T1-weighted imaging. It is based on the Shepp-Logan phantom and supports sampling along arbitrary k-space trajectories. Contrast enhancement in tissue is computed according to the Tofts model, while for the arterial enhancement, the Parker and Weinmann functions or user-defined time curves can be used.
Traditional Poster Session - Pulse Sequences & Reconstruction B

Image Analysis

Click on to view the abstract pdf. Click on to view the poster (Not all posters are available for viewing.)

Wednesday 9 May 2012
Exhibition Hall  16:00 - 18:00

2560.   Artificial neural networks in radiological predictive models
Nikolaos Dikaios1, Taiki Fujiwara2, David Atkinson3, and Shonit Punwani2
1Department of Medical Physics and Bioengineering, University College London, London, Greater London, United Kingdom, 2Department of Radiology, University College London Hospital, 3University College London, Centre for Medical Imaging

Predictive models are being increasingly employed in radiology as diagnostic aids for cancer detection. A variety of model types exist. Linear discriminant analysis (LDA) models assume linearity, normality and that the input variables are independent, assumptions which may affect classification accuracy. Neural networks (NN) whilst less intuitive, do not make these assumptions and can detect complex non-linear relationships between the input variables. Both LDA and NN are prone to over-fitting. In this work we compared the performance of multilayer perceptron (MLP) artificial NN and LDA models for prediction of transition zone (TZ) prostate cancer (based on quantitative multi-parametric MRI variables) using a leave-one-out (LOO) and a 2-fold cross validation analysis.

2561.   Method for Improving Segmentation of Multispectral brain MRI by a Supervised Hybrid Classifier
Jyh-Wen Chai1, Clayton Chi-Chang Chen1, Hsian-Min Chen2, Yaw-Jiunn Chiou3, Shih-Yu Chen4, Yi-Ying Wu1, Chih-Ming Chiang1, Ching-Wen Yang5, Yen-Chieh Ouyang6, San-Kan Lee1, and Chein_I Chang4
1Department of Radiology, Taichung Veterans General Hospital, Taichung, Taiwan, 2Department of Biomedical Engineering, HungKung university, Taichung, Taiwan,3Department of Electrical Engineering, National Chung Hsing University, 4Department of Computer Science and Electrical Engineering, University of Maryland, Baltimore County, Baltimore, MD, 5Computer Center, Taichung Veterans General Hospital, Taichung, Taiwan, 6Department of Electrical Engineering, National Chung Hsing University, Taichung, Taiwan

A hybrid classifier, derived from iterative Fisher¡¦s linear discriminant analysis coupled with the volume sphering analysis and support vecter machine, was developed to effectively segment multi-slice data of multispectral brain MRI by using only one set of training samples. The proposed method has several advantages. One was a reduction of computational cost in data processing since it only needs one set of training samples to process the entire multislice images. Besides, the same saving is also applied in minimizing operator burden. The uppermost benefit is to avoid operator interferences from selecting training samples and improve the reproducibility.

2562.   Performance of an automated segmentation algorithm for MR renography
Artem Mikheev1, Jeff L Zhang2, Tariq Gill1, Marta Heilbrun2, Stella Kang1, Hersh Chandarana1, Henry Rusinek1, and Vivian S Lee2
1Radiology, NYU School of Medicine, New York, NY, United States, 2Radiology, University of Utah School of Medicine, Salt Lake City, UT, United States

A key prerequisite for analysis of MR renography (MRR) data is the ability to segment MRI images. We have developed and validated a new semi-automated renal segmentation technique based on edge-constrained region growing. The segmentation error is 7.6 ± 6.5 cm3 and the interobserver disparity 5.4 ± 4.5 cm3, a significant improvement over graph-cut method. The new algorithm achieves a ten-fold improvement in user processing time, from >20 min to 2.1 ± 0.7 min per kidney. With expedited image processing, MRR has the potential to expand our knowledge of renal function and to help diagnose different types of renal insufficiency.

2563.   A fully automated, hierarchical classification method for detecting white matter lesions in Multiple Sclerosis
Marco Battaglini1, Nicola De Stefano1, and Mark Jenkinson2
1Neurological and behavioral sciences, University of Siena, Siena, Tuscany, Italy, 2FMRIB Centre, Nuffield Department of Clinical Neurosciences, University of Oxford

A novel hierarchical classification method for segmenting lesions in Multiple Sclerosis is described. It uses two stages: (1) a standard voxel-wise classifier followed by (2) a novel cluster-wise classifier. Features used for the cluster-wise classifier consist of ratios of statistics extracted from within the first-level clusters (across several different image modalities), to those from the exterior borders of the clusters or from all Grey Matter or White Matter voxels. Results on images from a multi-site clinical dataset showed large reductions in False Positives (for voxel-based and lesion-based metrics), with minimal reductions in True Positives, demonstrating great potential for this approach.

2564.   Multi-Atlas Corpus Callosum Segmentation with Adaptive Atlas Selection
Babak A. Ardekani1,2, Toshikazu Ikuta3,4, Alvin Bachman1, and Philip R. Szeszko3,4
1Center for Advanced Brain Imaging, Nathan Kline Institute for Psychiatric Research, Orangeburg, NY, United States, 2Department of Psychiatry, New York University School of Medicine, New York, NY, United States, 3Feinstein Institute for Medical Research, Manhasset, NY, United States, 4Psychiatry Research, The Zucker Hillside Hospital, Glen Oaks, NY, United States


2565.   Automatic segmentation of breast lesions in dynamic contrast-enhanced MR images
Jonathan Arvidsson1, Fredrik Johansson1, Andrew Mehnert1,2, Darryl McClymont3, and Dominic Kennedy4
1Signals and Systems, Chalmers University of Technology, Gothenburg, Sweden, 2MedTech West, Gothenburg, Sweden, 3ITEE, The University of Queensland, Brisbane, Queensland, Australia, 4Queensland X-Ray, Greenslopes, Queensland, Australia

A novel method for automatically segmenting 3D lesions in dynamic contrast-enhanced breast MRI data is proposed. It is based on assigning a “suspiciousness” score to each voxel using features extracted from its time series, and then computing the spatial co-occurrence of this score in a 3D neighborhood about the voxel. In this way both the spatial and temporal variation in contrast enhancement are characterized. An empirical evaluation of the efficacy of this technique versus a competing method based on multispectral co-occurrence is also presented.

2566.   Fusion of electrical impedance tomography data with MRI
Ross A Little1, John L Davidson2, Josephine H Naish1, Paul Wright2, Ron Kikinis3, Hugh McCann2, and Geoff J Parker1
1Imaging Sciences, Manchester Academic Health Sciences Centre, University of Manchester, Manchester, United Kingdom, 2School of Electrical and Electronic Engineering, University of Manchester, United Kingdom, 3Surgical Planning Laboratory, Brigham and Women's Hospital, Harvard Medical School, Boston, MA, United States

We are interested in fusing electrical impedance tomography (EIT) data with MRI, to give temporally and anatomically resolved functional lung imaging. We have progressed this using gel-based phantoms using the Confeitir (CONverter of Functional Electrical Impedance Tomography Images for Registration) software and the 3D Slicer package. EIT data were captured using fEITER, a biomedical EIT instrument operating at 100 frames per second. 3D Slicer was used to import and align the images, which show good correspondence between the two modalities. We intend to use similar methods to jointly visualise human lung data acquired using both EIT and MRI.

2567.   A rigid Registration Method for automated Scan Planning in Follow-up Examinations: Retrospective Analysis from Volunteer and Patient Neuro Scans
Julien Sénégas1, Torbjoern Vik1, Jens von Berg1, Peter Koken1, Juergen Gieseke2,3, Vincent Denolin4, Roy König3, and Stefan Sunaert5
1Philips Research Laboratories, Hamburg, Germany, 2Philips Healthcare, Best, Netherlands, 3Radiology, University of Bonn, Bonn, Germany, 4Philips Healthcare, Brussels, Belgium, 5Medical Imaging Center, Radiology, Catholic University of Leuven, UZ Leuven, Leuven, Belgium

MRI based monitoring of cerebral diseases such as brain tumors and multiple-sclerosis requires a precise comparison of the signal intensity and structure between baseline and follow-up exams. To minimize visual changes due to variations in patient positioning and scan planning, an accurate replication of the baseline acquisition geometry is desired. If done manually by visual inspection of the baseline images, this is a very tedious, time-consuming procedure with limited accuracy. In this work, we propose a general approach for automated scan planning in follow-up exams that only requires the baseline diagnostic images as input, with no further restrictions on the baseline exam as far as types of MR sequences, of MR scanner (vendor, field strength) or of planning (automated or manual) are concerned. Results from a preliminary study based on a retrospective analysis of volunteer and patient scans are presented.

2568.   A Model-Based Method for Registration of Ex Vivo to In Vivo Prostate MRI Using Elastography
Guy Nir1, Ramin S. Sahebjavaher1, Piotr Kozlowski2, Ralph Sinkus3, and Septimiu E. Salcudean1
1Electrical and Computer Engineering, University of British Columbia, Vancouver, BC, Canada, 2MRI Research Centre, University of British Columbia, Vancouver, BC, Canada,3Centre de Recherche Biomédicale Bichat-Beaujon (CRB3), Paris, France

We propose a novel method for registration of an ex vivo 3D model of a prostate to an in vivo MRI volumetric image. Such method may be employed to generate probability maps for prostate cancer distribution. The in vivo scan is followed by an MR elastography scan, from which the true elasticity parameters can be extracted and assigned to each voxel of the in vivo image. This allows a realistic regularization of a non-rigid registration method in order for it to produce physical and accurate deformation maps. Experiments on both synthetic and clinical data show promising results.

2569.   Automatic Template-based Breast Segmentation on MRI Using Nonrigid Registration Algorithms
Muqing Lin1, Jeon-Hor Chen1,2, Orhan Nalcioglu1, and Min-Ying Lydia Su1
1Tu & Yuen Center for Functional Onco Imaging, University of California, Irvine, California, United States, 2Department of Radiology, China Medical University Hospital, Taichung, Taiwan

The purpose is to develop an automatic breast segmentation method in MRI. The new algorithm is based on template matching of chest region using nonrigid registration. The body landmarks defined on the template can be transformed to subject’s space for performing v-shape cut to determine the posterior boundary of the lateral breasts. The chest wall mussel is excluded using Bezier curve fitting. The results show that this template-based segmentation method can accurately segment different types of breasts. This tool can help developing computer-aided-diagnosis method for detection of breast cancer, as well as quantitative analysis of breast density for risk management.

2570.   Population-based human brain MRI atlas with sharp contrast and its application in image registration
Yajing Zhang1, Jiangyang Zhang2, Michael I. Miller1,3, and Susumu Mori2
1Department of Biomedical Engineering, Johns Hopkins University, Baltimore, MD, United States, 2Department of Radiology, Johns Hopkins University School of Medicine, Baltimore, MD, United States, 3Center of Imaging Science, Johns Hopkins University, Baltimore, MD, United States

An MR based brain atlas is a key component in modern image analysis process. In this study, a population-based brain atlas with preserved image resolution was generated from normal aged control and age-matched AD datasets, using a continuous fluid dynamic model. This so-called VTE atlas presents a group averaged shape that minimize its anatomical bias while preserves sharp contrast for accurate image mapping. The characteristics of the new atlas were examined with respect to single subject atlases and group averaged atlases. The registration accuracy by using the new atlas was compared with commonly used single subject atlas through kappa analysis.
Traditional Poster Session - Pulse Sequences & Reconstruction B


Click on to view the abstract pdf. Click on to view the poster (Not all posters are available for viewing.)

Wednesday 9 May 2012
Exhibition Hall  16:00 - 18:00

2571.   Investigation of the Resolution Dependence of White Matter Structure Analysis by Means of Variography
Fabian Keil1, Daniel Brenner1, Ana-Maria Oros-Peusquens1, and N. Jon Shah1,2
1Institute of Neuroscience and Medicine - 4, Forschungszentrum Jülich, Jülich, Germany, 2JARA - Faculty of Medicine, RWTH Aachen University, Aachen, Germany

Variography is a novel method for the quantification of heterogeneities and spatial correlations in 3D MP-RAGE images of white matter. This abstract investigates the resolution dependence of variographic analyses applied to MR images. The change of the variogram is demonstrated and analysed by means of a digitised, high-resolution photograph of a human brain slice. A method for the correction of resolution-based changes is introduced and applied to the low resolution variograms. The estimated correlation parameters of corrected, uncorrected and native variograms are compared and evaluated.

2572.   Quantitative comparison of the informational content of T2*-weighted magnitude, phase and SWI 7T MR data by means of texture analysis of the cortical ribbon in elderly subjects
N. T. Doan1, S. van Rooden2,3, M. J. Versluis2, A. G. Webb2, J. van der Grond3, M. A. van Buchem3, J. H. C. Reiber1, and J. Milles1
1LKEB - Department of Radiology, Leiden University Medical Center, Leiden, Zuid Holland, Netherlands, 2CJ Gorter Center - Department of Radiology, Leiden University Medical Center, Leiden, Zuid Holland, Netherlands, 3Department of Radiology, Leiden University Medical Center, Leiden, Zuid Holland, Netherlands

The aim of this work is to investigate and compare the information content of T2*-weighted magnitude, phase and SWI data by means of texture analysis. Textural features of the cerebral cortex on magnitude, phase or SWI images were derived based on Gray Level Co-occurrence Matrix. Subsequently, classification using respectively magnitude-based, phase-based and SWI-based features was carried out with a Fisher linear discriminant classifier and a leave-one-out cross-validation approach. The results suggest that, from a texture information perspective, phase images have higher information content than either magnitude or SWI images.

2573.   Comparing Anisotropic Diffusion Filters: Enhancement of 23Na MRI of Human Kidney
Frank G Zoellner1, Christoph M Decker1, Simon Konstandin1, and Lothar R Schad1
1Computer Assisted Clinical Medicine, Heidelberg University, Mannheim, Germany

In-vivo MR imaging of sodium is limited due to its electro-physiological characteristics. A logical step forward is to enhance image intensity by applying post-processing filters while Gibb’s ringing could also be suppressed. The increase in signal-to-noise ratio (SNR) by linear filters as previously proposed is paid by a loss in resolution (blurring). Filters that preserve contours while removing noise are anisotropic diffusion filters (AND). The aim of this study was to compare different anisotropic filters for enhancing sodium MR images of the kidney.

2574.   Comparison of single ROI Vs whole liver determination of liver steatosis
A F Alsalihi1,2, G J Cowin1, J R Jonsson3, E E Powell3, A D Clouston3, M Benson4, and G J Galloway1
1Centre for Advanced Imaging- The University of Queensland, Brisbane, QLD, Australia, 2Pysics Department, Education College, Basrah University, Basrah, Basrah, Iraq,3School of Medicine, The University of Queensland, Brisbane, QLD, Australia, 4Department of Radiology, Princess Alexandra Hospital, Brisbane, QLD, Australia

Purpose This study compares whole liver estimation of liver steatosis with defined region of interest methods. Methods Liver fat was measured by liver biopsy, MRS and MRI using IP/OP and ±fat sat. Analysis methods: Correlations were investigated between the following: whole liver IP/OP and ±fat sat, single ROI liver IP/OP and ±fat sat, MRS, biopsy results. Results All methods tested for the determination of liver fat were highly correlated. Conclusion These results suggest a single ROI does not increase the error in the estimation of liver steatosis using imaging methods