Synopsis

Fat quantification by MRI plays an important role in assessing hepatic steatosis and bone marrow changes. We propose a new method for fat quantification in which multiband pulses are used to separately but simultaneously excite fat and water. CAIPIRINHA encoding allows overlapping images of the two species to be separated. Fat fraction maps can be calculated from fat and water images with appropriate consideration of relaxation times. The proposed method is validated on phantoms with different fat concentrations, where it yields similar fat fraction values to the Dixon approach and single-voxel spectroscopy.

Introduction

Fat quantification plays an important role in many contexts, including assessing hepatic steatosis and bone marrow changes. Currently, fat quantification is performed either using single-voxel spectroscopy, which provides only averaged information over an ROI, or using the Dixon method; a multi-echo approach which prolongs acquisition times and requires postprocessing which is prone to errors. We propose a method for separate fat and water imaging which allows fat quantification without fat-water ambiguity and uses only single-echo acquisition and well-established reconstruction techniques. Multiband pulses¹ simultaneously excite fat and water at their characteristic resonance frequencies, and the use of CAIPIRINHA² in combination with parallel imaging^3,4 allows separate fat and water images to be reconstructed. Due to the similarities with Simultaneous Multi-Slice imaging^2,5, we call this method Simultaneous Multi-Metabolite (SMM) imaging.

Once the water and fat are separated, signal fat fraction maps (FF_S) are calculated as $$FF_{s}=\frac{S_{f}}{S_{w}+S_{f}},$$ where S_w and S_f represent complex water and fat signals respectively, given by $$S=\frac{PD(1-e^{-\frac{TR}{T_1}})e^{-\frac{TE}{T_2^*}}\sin\alpha}{1-e^{-\frac{TR}{T_1}}\cos\alpha},$$ where PD is the proton density and α the flip angle. To minimize the bias caused by different T₁ values, small flip angles are used⁶. Alternatively, a correction based on measured⁷ or reference T₁ values could be applied. Since T₂^* values of fat and water in homogeneous mixture are expected to be similar⁸, no T₂^* correction is required in SMM.

Methods

The SMM approach was implemented in a gradient-echo sequence. For non-selective 3D imaging, the multiband pulse comprised two 17ms Shinar-Le-Roux pulses^9,10 of BW=350Hz, 0.5% out-of-passband and 1.5% over-passband ripples^11,12. For 2D imaging and slab-selective 3D imaging, spatial-spectral pulses¹³ can be used to avoid cross-excitation of metabolites between slices.

The accuracy of fat quantification by SMM was assessed using 3 water-fat phantoms - emulsions of peanut oil, 1% lecithin, 1% agar and 0.9% NaCl with fat fractions of 5%, 10% and 20%. Each phantom was measured separately in a water-filled container using a 3T Siemens PRISMA scanner and a 20-channel head coil.

Non-selective 3D gradient-echo SMM scans were acquired with TE=11.6ms, TR=25ms, resolution=1.15x1.15x3.0mm, FOV=140x140x132, FA(water)=3/5/9/9° and FA(fat)=3/5/9/24°, to assess the effects of SNR and T₁ bias on fat quantification. Fat fraction maps were calculated from complex, water-only and fat-only images, unaliased using slice-GRAPPA⁵ and combined over coils¹⁴. Mean fat fractions were calculated over an ROI localized in the centre of the image.

For comparison, two monopolar 3D gradient-echo Dixon scans were acquired using a vibe sequence¹⁵ with FAs of 3° and 5°, TE={1.23,3.09,4.95,6.81,8.67,10.53}ms, TR=25ms and the same FOV and resolution as for SMM. Water-fat separation was performed using the hierarchical IDEAL approach with T₂^* correction¹⁶ from the water-fat toolbox¹⁷, using the peanut oil spectral model¹⁸ and noise bias correction. Further, T₂-corrected single-voxel multi-echo 1H MRS (“HISTO”) data were acquired with a 20x20x20mm voxel, TE={14,28,42,56,70}ms and TR=3000ms.

Results

The SMM approach yielded well-separated fat and water images (Figure 1), with minimal residual aliasing and cross-excitation (due to e.g. poor shimming or out-of-passband ripples of RF pulses).

SMM fat fractions were similar to those calculated with DIXON, with a slight underestimation in water-fat phantoms and overestimation in water (the container itself; Figure 2).

SNR increased with flip angle (towards Ernst angles). Fat fractions were increasingly overestimated, however, due to T₁ bias (Figure 3 and Figure 4).

Mean fat fractions were underestimated by approximately 10% by SMM compared to Dixon and single-voxel spectroscopy (Figure 4).

Discussion

Fat quantification was shown to be possible using the proposed Simultaneous Multi-Metabolite imaging method. Fat fractions were close to those obtained using the Dixon method and single-voxel spectroscopy, despite being a very new method requiring further optimisation. The source of systematic underestimation by all methods (compared to the expected fat fractions) is under investigation.

In this study, low flip acquisitions had to be used to avoid T₁ bias, limiting the achievable SNR. In future work, T₁ correction⁷ could be used, allowing both metabolites to be excited using their respective Ernst angles, providing high SNRs. Further SNR increase could be achieved by reducing TE if shorter RF pulses with the same spectral selectivity could be achieved using optimal control¹⁹, for instance. Although complex data were used for fat fraction calculations, residual noise bias was observed, requiring further corrections for reliable assessment of low concentrations.

Underestimation of fat fractions with the proposed ‘naïve’ SMM approach was expected in the light of the spectral complexity of fat, as some fat peaks overlap with water. This could be corrected, since the relative amplitudes of peaks are generally known.

Conclusion

We have presented a new method for separate water-fat imaging and shown its potential for fat quantification.

Acknowledgements

References

1. Muller S. Multifrequency Selective RF Pulses for Multislice MR Imaging. Magn Reson Med 1988;6:364–371.
2. 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.
3. Pruessmann KP, Weiger M, Scheidegger MB, Boesiger P. SENSE: Sensitivity encoding for fast MRI. Magn Reson Med 1999;42:952–962.
4. Griswold MA, Jakob PM, Heidemann RM, Nittka M, Jellus V, Wang J, Kiefer B, Haase A. Generalized autocalibrating partially parallel acquisitions (GRAPPA). Magn Reson Med 2002;47:1202–1210.
5. 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. Magn Reson Med 2012;67(5):1210–1224.
6. Liu CY, McKenzie CA, Yu H, Brittain JH, Reeder SB. Fat quantification with IDEAL gradient echo imaging: correction of bias from T1 and noise. Magn Reson Med 2007;58(2): 354–364.
7. Yang IY, Cui Y, Wiens CN, Wade TP, Friesen-Waldner LJ, McKenzie CA. Fat fraction bias correction using T1 estimates and flip angle mapping. J Magn Reson Imaging. 2014;39:217–223.
8. Yu H, McKenzie CA, Shimakawa A, Vu AT, Brau AC, Beatty PJ, Pineda AR, Brittain JH, Reeder SB. Multiecho reconstruction for simultaneous water‐fat decomposition and T2* estimation. J Magn Reson Imaging 2007;26:1153–1161.
9. Shinnar M, Eleff S, Subramanian H, Leigh JS. The synthesis of pulse sequences yielding arbitrary magnetization vectors. Magn. Reson Med 1989;12:74–80.
10. Le Roux P. Exact synthesis of radio frequency waveforms. Proc. 7th SMRM, 1988;1049.
11. 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.
12. http://scion.duhs.duke.edu/vespa/project
13. Meyer CH, Pauly JM, Macovski A, Nishimura DG. Simultaneous spatial and spectral selective excitation. Magn Reson Med 1990;15:287–304.
14. Parker DL, Payne A, Todd N, Hadley JR. Phase reconstruction from multiple coil data using a virtual reference coil. Magn Reson Med 2014;72(2):563–569.
15. Rofsky NM, Lee VS, Laub G, Pollack MA, Krinsky GA, Thomasson D, Ambrosino MM, Weinreb JC. Abdominal MR imaging with a volumetric interpolated breath-hold examination. Radiology 1999;212:876–884.
16. Tsao J, Jiang Y. I.D.E.A.L. Hierarchical Fast, robust, and multiresolution separation of multiple chemical species from multiple echo times. Magn Reson Med 2013;70:155–159.
17. Hu HH, Bornert P, Hernando D, Kellman P, Ma J, Reeder S, Sirlin C. ISMRM workshop on fat-water separation: insights, applications and progress in MRI. Magn Reson Med 2012;68(2):378–388.
18. Guillen MD, Ruiz A. Rapid simultaneous determination by proton NMR of unsaturation and composition of acyl groups in vegetable oils. European Journal of Lipid Science and Technology 2003;105(11):688–696.
19. 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.