SNR Improvement in Single-Echo Dixon Imaging
Publication Date: 2015-Mar-06
The IP.com Prior Art Database
Title: SNR Improvement in Single-Echo Dixon Imaging
The approach relates to a magnetic resonance imaging method which involves a single-echo Dixon water-fat separation approach.
According to the approach voxels are identified with a high SNR (SNR exceeding a preset threshold). From these high SNR voxels, voxels are selected with a high or a low fat fraction (also based on preset thresholds, and on a standard water-fat separation). Then, areas in the reconstructed magnetic resonance image can be identified that are strongly dominant in water or in fat, respectively. Finally, in the identified strongly dominant water (or fat) areas, all signal is attributed to water (or fat).
Thus an effective fat suppression is achieved with a good SNR that is less sensitive to the signal echo time.
It appears that the approach of the approach is clinical notably useful in angiography.
The suppression of fat signal is a basic requirement in many applications of magnetic resonance imaging. So-called Dixon imaging is increasingly employed for this purpose, mainly because of its robustness with respect to spatial variation of the main magnetic field. It involves an encoding of the chemical shifts of the hydrogen atoms in water and fat by repeated measurements at different echo times.
Relying on a single echo only in Dixon imaging seems very attractive, since it promises a substantial reduction in scan time compared to using two or more echoes . However, the choice of the echo time of this single echo proves to be more critical for preserving sufficient SNR, which compromises some of the expected gains.
 Ma J. A single-point Dixon technique for fat-suppressed 3D gradient-echo imaging with a flexible echo time. J Magn Reson Imaging 2008; 27:881-890.
The present approach aims at improving SNR in potential applications of single-echo Dixon imaging to fat suppression. It suggests allocating all signal to the water or fat signal in a voxel, instead of separating water and fat signal, if the latter results in a fat fraction below or above a certain threshold in this voxel and optionally its immediate vicinity.
Let the complex signal S from a voxel be modeled by
where W and F are the unknown real water and fat signal. c and P are a complex weighting factor assumed to be known and an unknown phasor. To determine W and F given S, further assumptions have to be made, due to the presence of the nuisance parameter P. By restricting the fat fraction F/(W+F) to either 0 (water only) or to a value close to 1 (mostly fat), two candidates for P can be derived. One of them can then be selected by requiring the spatial variation of P to be smooth. Finally, W and F can be calculated by
Even if P were known exactly, th...