Reconstruction of phase images for GRAPPA accelerated Magnetic Resonance Imaging

DOI: 10.1007/978-3-540-89208-3_192

ABSTRACT In this work we present a method to combine complex-valued phased array MR data based on a uniform sensitivity approach, which
incorporates sensitivity profiles calculated from afore acquired data. The algorithm was implemented on a clinical 3T whole-body
MR-Scanner and embedded into the vendor-specific standard reconstruction chain.

Additionally, it was linked with the GeneRalized Autocalibrating Partial Parallel Acquisition (GRAPPA) procedure to take advantage
of partial parallel acquisition techniques as well. Thus, under-sampled data can be reconstructed with GRAPPA and subsequently
combined with the presented method. Phase images reconstructed with the proposed method were in excellent agreement with reference
data and did not suffer from errors due to incorrectly combined data.

  • [Show abstract] [Hide abstract]
    ABSTRACT: Quantitative susceptibility mapping (QSM) is a novel magnetic resonance-based technique that determines tissue magnetic susceptibility from measurements of the magnetic field perturbation. Due to the ill-posed nature of this problem, regularization strategies are generally required to reduce streaking artifacts on the computed maps. The present study introduces a new algorithm for calculating the susceptibility distribution utilizing a priori information on its regional homogeneity derived from gradient echo phase images and analyzes the impact of erroneous a priori information on susceptibility map fidelity. The algorithm, Homogeneity Enabled Incremental Dipole Inversion (HEIDI), was investigated with a special focus on the reconstruction of subtle susceptibility variations in a numerical model and in volunteer data and was compared with two recently published approaches, Thresholded K-space Division (TKD) and Morphology Enabled Dipole Inversion (MEDI). HEIDI resulted in susceptibility maps without streaking artifacts and excellent depiction of subtle susceptibility variations in most regions. By investigating HEIDI susceptibility maps acquired with the volunteers' heads in different orientations, it was demonstrated that the apparent magnetic susceptibility distribution of human brain tissue considerably depends on the direction of the main magnetic field.
    NeuroImage 05/2012; 62(3):2083-100. DOI:10.1016/j.neuroimage.2012.05.067 · 6.13 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Magnetic susceptibility is an intrinsic tissue property that recently became measureable in vivo by a magnetic-resonance based technique called quantitative susceptibility mapping (QSM). Although QSM may be performed without additional acquisition time, for example, in the course of the well-established susceptibility weighted imaging, the applicability of QSM is currently hampered by the numerical complexity and computational cost associated with the reconstruction procedure. This work introduces a novel QSM framework called superfast dipole inversion which allows rapid online reconstruction of susceptibility maps from wrapped raw gradient-echo phase data. The algorithm relies on the extension and combination of several recent algorithms involving the precalculation of convolution kernels and the correction of inversion artifacts. Reconstruction of three-dimensional high resolution susceptibility maps of the human brain was achieved with superfast dipole inversion in less than 20 s on a conventional workstation computer. Thus, superfast dipole inversion opens the door to an implementation of QSM on MR scanner hardware as well as to the routine reconstruction of large cohorts of datasets. Magn Reson Med, 2012. © 2012 Wiley Periodicals, Inc.
    Magnetic Resonance in Medicine 06/2013; 69(6). DOI:10.1002/mrm.24405 · 3.40 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: A method is presented for the combination of phase images from multi-channel RF coils in the absence of a volume reference coil. It is based on the subtraction of 3D phase offset maps from the phase data from each coil. Phase offset maps are weighted combinations of phase measurements at two echo times. Multi-Channel Phase Combination using measured 3D phase offsets (MCPC-3D) offers a conceptually and computationally simple solution to the calculation of combined phase images. The dual-echo data required for the phase maps can be intrinsic to the high-resolution gradient-echo scan to be reconstructed (MCPC-3D-I). Alternatively, a separate, fast, low-resolution dual-echo scan can be used (MCPC-3D-II). Both variants are shown to give near perfect phase matching, yielding images with high SNR throughout and high GM-WM contrast. MCPC-3D is compared with other reference-free phase image combination methods; high-pass phase filtering, phase difference imaging, and matching using constant offsets (MCPC-C). Multi-Channel Phase Combination using measured 3D phase offsets method does not need an overlap between the signals from individual coils and can be used with parallel imaging, making it ideally suited to multi-channel coils with a large number of elements, and to high and ultra-high field systems.
    Magnetic Resonance in Medicine 06/2011; 65(6):1638-48. DOI:10.1002/mrm.22753 · 3.40 Impact Factor