Unique Acquisition Strategies 2

Wednesday 14 May 2014

Jie Wen¹, Anne H. Cross², and Dmitriy A. Yablonskiy^1
1Mallinckrodt Institute of Radiology, Washington University in St. Louis, St. Louis, MO, United States, ²Neurology, Washington University in St. Louis, St. Louis, MO, United States

Physiological fluctuations in biological tissues adversely affect MR images. In this abstract, a navigator- and a keyhole-based methods are used to reduce physiologically-induced artifacts. We study brains in normal subjects and subjects with multiple sclerosis and demonstrate that employed strategies substantially reduce the width of the R2*=1/T2* distribution within a human brain and provide significant improvement in quantifying tissue damage in multiple sclerosis. We also show improvement in the quality of GEPCI, SWI and QSM images. These correcting strategies greatly improve the reliability of quantitative gradient echo MRI techniques.

Yi Sui^1,2, Frederick C. Damen^1,3, Karen Xie³, and Xiaohong Joe Zhou^1,4
1Center for MR Research, University of Illinois at Chicago, Chicago, Illinois, United States, ²Bioengineering, University of Illinois at Chicago, Chicago, Illinois, United States, ³Department of Radiology, University of Illinois Hospital & Health Sciences System, Chicago, Illinois, United States, ⁴Departments of Radiology, Neurosurgery and Bioengineering, University of Illinois Hospital & Health Sciences System, Chicago, Illinois, United States

An image domain segmented acquisition technique has been developed to reduce distortion in diffusion imaging. The full FOV along the phase-encoding direction was divided into N parallel segments and acquired sequentially using a 2D RF excitation pulse. These image segments were then combined using a weighting function determined by the spatial excitation profile of the 2D RF pulse. Using this technique, substantial reduction in geometric distortion and signal pile up was demonstrated in human brain diffusion images at 3T.

Rüdiger Stirnberg¹, Daniel Brenner¹, and Tony Stöcker^1
1German Center for Neurodegenerative Diseases (DZNE), Bonn, Germany

Two novel sequence modifications for segmented 3D-EPI are proposed, which increase signal sensitivity by boosting signal-to-noise ratio on the one hand (variable flip angles) and by reducing the effective volume repetition time on the other hand (variable echo train lengths). Simulations and experiments are performed at 3 and 7 Tesla confirming the effectiveness of these modifications. On the example of 7 Tesla whole brain acquisitions at 0.75mm isotropic resolution it is shown that, with respect to temporal SNR, the proposed 3D-EPI based method is superior to a popular simultaneous multi-slice method using equivalent acceleration factors.

Enhao Gong¹, Feng Huang², and John M Pauly^1
1Electrical Engineering, Stanford University, Stanford, CA, United States, ²Philips Healthcare, Gainesville, FL, United States

Random undersampling is an important component used with Parallel Imaging (PI) and Compressed Sensing (CS) and their combination (PI-CS) for fast acquisition. Optimized pseudo-random trajectory results in better reconstruction yet the optimization is computational costly. Lately, we proposed an efficient scheme for 1D random undersampling optimization using stochastic method and reference k-space. Here we extended and improved the scheme to optimize the 2D Cartesian undersampling for both PI and CS using Nonlinear Grappa Operator and Coherence based objective function. In-vivo experiments demonstrated greater performance improvement for reconstruction using PI-CS. The scheme is also applicable for non-Cartesian undersampling.

Anthony G. Christodoulou¹, Yijen L. Wu², T. Kevin Hitchens², Chien Ho², and Zhi-Pei Liang^1
1Department of Electrical and Computer Engineering, University of Illinois at Urbana-Champaign, Urbana, IL, United States, ²Pittsburgh NMR Center for Biomedical Research, Department of Biological Sciences, Carnegie Mellon University, Pittsburgh, PA, United States

We investigate 2D and 3D non-Cartesian k-space navigator trajectories for low-rank (subspace) myocardial perfusion imaging, replacing 1D Cartesian navigators (which are highly sensitive to readout direction). A rodent ischemic re-perfusion injury animal model was used for both whole-heart 3D first-pass and delayed myocardial perfusion imaging in rats. The resulting 3D images have high spatiotemporal resolution (128 x 128 x 24 matrix size, 0.31mm x 0.31mm x 1.0 mm spatial resolution, 74 frames per second) and were used for analysis of both the first-pass of contrast and of late enhancement.

Diego Hernando¹, Samir D. Sharma¹, and Scott B. Reeder^1,2
1Radiology, University of Wisconsin-Madison, Madison, WI, United States, ²Medicine, University of Wisconsin-Madison, Madison, WI, United States

Chemical shift-encoded fat quantification requires the acquisition of images at multiple echo times to enable fat-water separation and R2* estimation. Excessive tissue iron deposition can severely increase R2*, complicating fat quantification. In this work, we assess the feasibility of chemical shift encoded liver fat quantification in the presence of liver iron using theoretical analysis and in vivo patient data. Our results demonstrate that fat quantification fails at high iron levels when using standard protocols. The use of shorter initial echo time and echo spacing significantly extends the range of iron levels over which accurate fat quantification is possible.

James Korte¹, Kelvin J. Layton¹, Bahman Tahayori¹, Peter M. Farrell¹, Stephen M. Moore¹, and Leigh A. Johnston^1
1The University of Melbourne, Melbourne, Victoria, Australia

The response of a spin system to Rabi modulated continuous wave excitation has recently been shown to achieve a substantial steady-state magnetization that exhibits periodic orbits. We exploit this steady-state behavior and explore off resonance effects, in two proof-of-concept experiments and corresponding simulations of the Bloch equation. Our experiments confirm that harmonics of the periodic steady-state magnetization orbits are affected by off-resonance effects and therefore contain chemical shift information, in agreement with predictions. We furthermore demonstrate that chemical shift information can be encoded in a series of CW excitations and used to reconstruct a simple spectrum.

Samir D. Sharma¹, Nathan S. Artz¹, Diego Hernando¹, Debra E. Horng^1,2, and Scott B. Reeder^1,2
1Radiology, University of Wisconsin - Madison, Madison, WI, United States, ²Medical Physics, University of Wisconsin - Madison, Madison, WI, United States

The primary challenge in water-fat separation lies in estimating the B0 field map, which is composed of the background field and susceptibility-induced field. The susceptibility-induced field can be estimated if the susceptibility distribution is known or can be approximated. In this work, the susceptibility distribution is approximated from the source images using the known susceptibility values of tissue and air. This object-based information is then used to improve the robustness to swaps for existing water-fat separation methods. Cases are shown in which water-fat swaps were avoided by using the object-based information of the B0 field map.

Brian A Hargreaves¹, Valentina Taviani¹, and Daehyun Yoon^2
1Radiology, Stanford University, Stanford, CA, United States, ²Radiology, Stanford University, Stanford, California, United States

Multi-spectral imaging (MSI) methods such as SEMAC, MAVRIC, and MAVRIC-SL, offer excellent metal artifact reduction. Variants of these methods all use 3D imaging, limiting both minimum scan time and flexibility. Here we present an approach using 2D imaging of limited volumes, produced by flipping the selection gradient sign between excitation and refocusing pulses. The volumes are combined directly, avoiding the need for z phase encoding and view-angle tilting. Artifact correction is comparable to MSI methods, though at an SNR penalty. Advantages include fast imaging of a small number of slices, and easy localization of off-resonance signal to improve scan efficiency.

Jonathan I. Tamir¹, Peng Lai², Martin Uecker¹, and Michael Lustig^1
1Electrical Engineering and Computer Sciences, University of California, Berkeley, Berkeley, CA, United States, ²Global Applied Science Laboratory, GE Healthcare, Menlo Park, CA, United States

Volumetric Fast Spin Echo (FSE) is an attractive alternative to 2D FSE as it provides isotropic resolution. Because 3D FSE employs long echo trains to reduce scan time, the resulting image suffers from blurring due to T2 decay. In this work, we model the temporal behavior of the acquisition to reconstruct a full time series of images. In addition to exploiting parallel imaging and spatial sparsity, we constrain the temporal decay to a low-dimensional subspace. We show that randomizing the echo train ordering in tandem with this temporal model can produce a multi-contrast time series of images with reduced blurring.