Synopsis

Imaging of many body regions is adversely affected by the chemical shift between fat and water. We propose a new method for simultaneous, separate imaging of fat and water in which multiband pulses are used to simultaneously excite fat and water at their characteristic resonance frequencies, and CAIPIRINHA in combination with parallel imaging to reconstruct separate images of the two species. The fat image is corrected for chemical shift displacement and either recombined with the water image or evaluated separately. The proposed method achieves reliable water-fat separation and complete elimination of the chemical shift artefact in musculo-skeletal and breast imaging.

Introduction

The chemical shift between fat and water leads to difficulties assessing images of tissues containing predominately water-based and fat-based tissues. We propose a method for simultaneous, separate water and fat imaging which we call Simultaneous Multi-Metabolite (SMM) imaging. Analogous to Simultaneous Multi-Slice imaging^1,2, in which multiband pulses are used to excite slices with different resonance frequencies due to the slice gradient, SMM simultaneously excites metabolites with different resonance frequencies due to chemical shift. This requires that inhomogeneity in B₀ < ½ chemical shift between water and fat, which is 220 Hz at 3T. CAIPIRINHA³ is used to shift the image of one of the metabolites along PE direction(s) and coil sensitivities are used to separate images of the two metabolites. The chemical shift between the two species can be corrected and the images either evaluated separately or recombined, leading to a chemical shift artefact free image, improving the visibility of clinically interesting features such as lesions in breast imaging or cartilage thickness and layering in musculo-skeletal imaging.

Methods

The SMM approach was implemented in a gradient-echo sequence. For non-selective 3D imaging, the multiband pulse comprised two Shinar-Le-Roux^4,5 pulses of BW=350Hz, 0.5% out-of-passband ripples and duration of t_pulse=16.24ms, achieving high spectral selectivity within a reasonable time^6,7. For 2D imaging, the SMM approach requires spatial-spectral pulses⁸ to avoid cross-excitation of metabolites between slices. The spectrally-selective envelope was identical to the multiband pulse used for 3D imaging, with spatially-selective sinc subpulses of t_subpulse=0.56ms executed during both positive and negative phases of the oscillating slice gradient. The Ernst angles were used for water and fat based on estimates of the respective T₁ values.

Two healthy subjects were measured with a 3T Siemens PRISMA scanner using 2D SMM imaging. A knee of one volunteer was measured with: TE/TR=12.3/125ms, resolution=1x1x2.5mm, rBW/pixel =150Hz, FA_water=19°, FA_fat=36°, FOV=150x150mm, 55 slices in 5 slice groups (15-channel coil) and a breast of the second volunteer with: TE/TR=12.3/75ms, resolution=1.2x1.2x2.5mm, rBW/pixel=150 Hz, FA_water=23°, FA_fat=43° (single breath-hold, 18-channel coil).

Dual-echo GE scan was also acquired, both for B₀ field mapping⁹ and for the water-fat unaliasing with slice-GRAPPA². The fat image was shifted by N_cs voxels in the reverse readout direction to correct chemical shift, where $$$N_{cs} = \frac{\omega_{water}-\omega_{fat}}{rBW/pixel}$$$ voxels, in which $$$\omega_{water}-\omega_{fat}$$$ is the difference between the resonance frequencies of water and fat, i.e. the water-fat chemical shift, and rBW/pixel is the receiver bandwidth per pixel.

Results

The field inhomogeneity was less than 220 Hz in both the knee and the breast, meeting the condition for separate excitation of water and CAIPIRINHA-shifted fat using the proposed approach. Cleanly separated water and fat were generated (see Figure 1 and Figure 3, respectively). Artefacts from the use of slice-GRAPPA were at the level of noise, evidenced by the near absence of water signal in the fatty areas of the images (e.g. bone marrow, adipose tissue) and vice versa (e.g . muscles, breasts lobules) – see the same figures.

The chemical shift correction of the fat images, by 3 voxels in the reverse readout direction (F>>H), allowed fat and water image recombination with complete removal of the chemical shift artefact (Figure 2, Figure 4).

Discussion

The proposed Simultaneous Multi-Metabolite imaging method, using selective excitation combined with CAIPIRINHA was shown to allow simultaneous, separate water and fat imaging. SMM allows complete correction of chemical shift, facilitating the assessment of overlapping water-based and fat-based structures, as demonstrated in the breast and the knee.

Similar to other techniques which rely on the chemical shift difference, the SMM method theoretically requires that ΔB₀ < ½ of the water-fat chemical shift, i.e. 220 Hz at 3 T, in order to selectively and separately excite water and fat. In reality, the required field inhomogeneity is slightly lower because the RF pulses are not rectangular. To make the spectral differentiation as reliable as possible, we have used 16.24ms Shinar-Le-Roux pulses, limiting the minimal TE. The optimal control approach¹⁰, however, could allow for shorter pulses with similar selectivity. Moreover, to allow separation even for larger ΔB₀, the B₀ variation over the FOV could be assessed (e.g. from a prescan) and the excitation frequencies of spatial-spectral pulses could be varied accordingly¹¹, allowing for ΔB₀ compensation. The SMM approach was implemented in a gradient-echo sequence, but could be used for other sequences, such as turbo-spin-echo, which is the most commonly used sequence for MSK imaging.

Conclusion

Acknowledgements

References

1. Muller S. Multifrequency Selective RF Pulses for Multislice MR Imaging. Magn Reson Med 1988;6:364–371.
2. Setsompop K, Gagoski BA, Polimeni JR, Witzel T, Wedeen VJ, Wald LL. Blipped-Controlled Aliasing in Parallel Imaging (blipped-CAIPI) for simultaneous multi-slice EPI with reduced g-factor penalty. MagnReson Med 2012;67(5):1210–1224.
3. Breuer FA, Blaimer M, Mueller MF, Seiberlich N, Heidemann RM, Griswold MA, Jakob PM . Controlled aliasing in volumetric parallel imaging (2D CAIPIRINHA). Magn Reson Med 2006;55:549–556.
4. Shinnar M, Eleff S, Subramanian H, Leigh JS. The synthesis of pulse sequences yielding arbitrary magnetization vectors. Magn. Reson Med 1989;12:74–80.
5. Le Roux P. Exact synthesis of radio frequency waveforms. Proc. 7th SMRM, 1988;1049.
6. Soher B, Semanchuk P, Todd D, Steinberg J, Young K. VeSPA: integrated applications for RF pulse design, spectral simulation and MRS data analysis. Proc. 19th ISMRM, 2011;1410.
7. http://scion.duhs.duke.edu/vespa/project
8. Meyer CH, Pauly JM, Macovski A, Nishimura DG. Simultaneous spatial and spectral selective excitation. Magn Reson Med 1990;15:287–304.
9. Jezzard P, Balaban RS. Correction for geometric distortion in echo‐planar images from B0 field variations. Magn Reson Med 1995;34:65–73.
10. Conolly S, Nishimura D, Macovski A. Optimal control solutions to the magnetic resonance selective excitation problem. IEEE transactions on medical imaging 1986; 5(2):106–115.
11. Yang C, Deng W, Stenger VA. Simple Analytical Dual-Band Spectral-Spatial RF Pulses for B1+ and Susceptibility Artifact Reduction in Gradient Echo MRI. Magn Reson Med 2011;65(2):370–376.