Joint Estimation of Water/Fat Images and Field Inhomogeneity Map

Department of Electrical and Computer Engineering, University of Illinois at Urbana-Champaign, Urbana, Illinois, USA.
Magnetic Resonance in Medicine (Impact Factor: 3.57). 03/2008; 59(3):571-80. DOI: 10.1002/mrm.21522
Source: PubMed


Water/fat separation in the presence of B 0 field inhomogeneity is a problem of considerable practical importance in MRI. This article describes two complementary methods for estimating the water/fat images and the field inhomogeneity map from Dixon-type acquisitions. One is based on variable projection (VARPRO) and the other on linear prediction (LP). The VARPRO method is very robust and can be used in low signal-to-noise ratio conditions because of its ability to achieve the maximum-likelihood solution. The LP method is computationally more efficient, and is shown to perform well under moderate levels of noise and field inhomogeneity. These methods have been extended to handle multicoil acquisitions by jointly solving the estimation problem for all the coils. Both methods are analyzed and compared and results from several experiments are included to demonstrate their performance.

Download full-text


Available from: Brad P Sutton, Jan 07, 2014
1 Follower
11 Reads
  • Source
    • "The water separation algorithm uses the variable projection (VARPRO) method to find an iterative solution for the entire image [22]. The VARPRO algorithm can use four or more echoes with non-uniformly spaced echo times (phase shifts between water and fat) to solve the nonlinear problem with a globally optimized solution. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Previous data using T1-weighted MRI demonstrated neck muscle fat infiltration (MFI) in patients with poor functional recovery following whiplash. Such findings do not occur in those with milder symptoms of whiplash, chronic non-traumatic neck pain or healthy controls, suggesting traumatic factors play a role. Muscle degeneration could potentially represent a quantifiable marker of poor recovery, but the temporal constraints of running a T1-weighted sequence and performing the subsequent analysis for muscle fat may be a barrier for clinical translation. The purpose of this preliminary study was to evaluate, quantify and compare MFI for the cervical multifidus muscles with T1-weighted imaging and a more rapid quantitative 3D multi-echo gradient echo (GRE) Dixon based method in healthy subjects. 5 asymptomatic participants with no history of neck pain underwent cervical spine MRI with a Siemens 3 Tesla system. The muscle and fat signal intensities on axial spin-echo T1-weighted images were quantitatively classified for the cervical multifidii from C3-C7, bilaterally. Additional axial GRE Dixon based data for fat and water quantification were used for comparison via paired t-tests. Inter-tester reliability for fat and water measures with GRE images were examined using 1) Pearson's Intra-class correlation coefficient 2) Bland-Altman Plots and 3) Lin's-Concordance Coefficient. P < 0.05 was used to indicate significance. Total mean (SD) MFI (C3-C7) for the multifidii obtained with T1-weighted imaging and GRE were 18.4% (3.3) (range 14-22%) and 18.8% (2.9) (range 15-22%), respectively. The Pearson correlation coefficients for inter-tester reliability on the GRE sequences for the C3-C7 multifidii ranged from .83 - .99, indicating high levels of agreement with segmental MFI measures. Bland-Altman Plots revealed all data points were within 2 SDs and concordance was established between 2-blinded raters, suggesting good agreement between two raters measuring fat and water with GRE imaging. Results of this preliminary study demonstrate reliability between 2 raters of varying experience for MRI analysis of MFI with 3D GRE MRI. The quantification of MFI for healthy cervical musculature is comparable to T1-weighted images. Inclusion of larger samples of symptomatic data and histological comparison with the reference standard biopsy is warranted.
    BMC Medical Imaging 09/2013; 13(1):30. DOI:10.1186/1471-2342-13-30 · 1.31 Impact Factor
  • Source
    • "There are two parts in the objective function formulated in VARPRO: One is the data cost that measures the goodness of fit at an individual pixel; the other is the smoothness cost that measures the similarity of field map values at neighboring pixels. The optimization problem is tackled by iterated conditional modes (ICM) algorithm in [9] and graph cut algorithm in [10], respectively. The regularization parameter, which is often empirically chosen to combine the two parts in the objective function, determines the degree of smoothing in the resultant field maps. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Water-fat separation in magnetic resonance imaging (MRI) is of great clinical importance, and the key to uniform water-fat separation lies in field map estimation. This work deals with three-point field map estimation, in which water and fat are modelled as two single-peak spectral lines, and field inhomogeneities shift the spectrum by an unknown amount. Due to the simplified spectrum modelling, there exists inherent ambiguity in forming field maps from multiple locally feasible field map values at each pixel. To resolve such ambiguity, spatial smoothness of field maps has been incorporated as a constraint of an optimization problem. However, there are two issues: the optimization problem is computationally intractable and even when it is solved exactly, it does not always separate water and fat images. Hence, robust field map estimation remains challenging in many clinically important imaging scenarios. This paper proposes a novel field map estimation technique called JIGSAW. It extends a loopy belief propagation (BP) algorithm to obtain an approximate solution to the optimization problem. The solution produces locally smooth segments and avoids error propagation associated with greedy methods. The locally smooth segments are then assembled into a globally consistent field map by exploiting the periodicity of the feasible field map values. In vivo results demonstrate that JIGSAW outperforms existing techniques and produces correct water-fat separation in challenging imaging scenarios.
    IEEE Transactions on Medical Imaging 08/2011; 30(7-30):1417 - 1426. DOI:10.1109/TMI.2011.2122342 · 3.39 Impact Factor
  • Source
    • "We initialize this scheme by setting [ ] 0 α = n . This approach is similar to the VARPRO formulation, used by [15]. The main differences are 1) the inclusion of the decay term in α , which enables us to consider larger delays and 2) the use of more delay terms to improve the estimates. "
    [Show abstract] [Hide abstract]
    ABSTRACT: We introduce a novel algorithm to address the challenges in magnetic resonance (MR) spectroscopic imaging. In contrast to classical sequential data processing schemes, the proposed method combines the reconstruction and postprocessing steps into a unified algorithm. This integrated approach enables us to inject a range of prior information into the data processing scheme, thus constraining the reconstructions. We use high resolution, 3-D estimate of the magnetic field inhomogeneity map to generate an accurate forward model, while a high resolution estimate of the fat/water boundary is used to minimize spectral leakage artifacts. We parameterize the spectrum at each voxel as a sparse linear combination of spikes and polynomials to capture the metabolite and baseline components, respectively. The constrained model makes the problem better conditioned in regions with significant field inhomogeneity, thus enabling the recovery even in regions with high field map variations. To exploit the high resolution MR information, we formulate the problem as an anatomically constrained total variation optimization scheme on a grid with the same spacing as the magnetic resonance imaging data. We analyze the performance of the proposed scheme using phantom and human subjects. Quantitative and qualitative comparisons indicate a significant improvement in spectral quality and lower leakage artifacts.
    04/2010; 29(6):1297-309. DOI:10.1109/TMI.2010.2046673
Show more