Synopsis

mDixon imaging can provide robust anatomical images at 7T. These images have also been used to provide preliminary initial measurements of fat fraction in the liver at 7T.

Introduction

Aim

To explore the potential of mDixon images to provide workhorse anatomical scans with reduced signal drop out at 7T and to obtain preliminary fat quantification maps of the liver.

Methods

Abdominal imaging was carried out on healthy volunteers using a Phillips 7T Achieva 8-channel multi-transmit system with an 8 channel transmit, 32 channel receive fractionated dipole body array [2] (MR Coils, Zaltbommel, Netherlands). B0 shimming was implemented using the Philips volume shimtool over a region encompassing the liver. RF shimming was also performed using a phase nulling method to maximise the B1+ in regions of interest also encompassing the liver.

Preliminary fat quantification was performed on two healthy volunteers. Following the recommendations of Reeder et al 2005 [1], 3D multi gradient echo images were acquired at 6 echo times (TE1=1.4ms, rTE=1.32ms), a field of view (FOV) of 400 x 400 mm² with 4 slices of resolution 2 mm x 2 mm x 5 mm (with in-plane reconstruction 0.93 x 0.93 mm²). Images were acquired during a single breath hold (maximum 30 s). The magnitude signal was fitted to a 6 peak model using non-linear least squares fit in Matlab assuming a single R2* for fat and water.

S = ( W + F ΣC_i exp(jω_it)) exp( j(Ψt+φ) - R2*t)

S is complex signal for n components varying with time t where S_i= component signal, ω_i= component resonant frequency, Ψ= local field offset, φ= constant phase offset, R2*= component relaxation, W = water signal, F = fat signal with m fat peaks c_i= proportional amplitude constant at ω_i resonant frequency where Σc_i=1. The average fat fraction was measured in an ROI where the signal in the base multiecho images was high.

Results

Figure 1 shows in vivo coronal water only calculated images covering the kidneys and part of the liver. Figure 2 shows the fat, water, in-phase and opposed-phase images generated over the liver in the transverse plane. Figure 3 show the liver and part of the bowel surrounded by fat in the transverse plane. Figure 4 shows the magnitude signal from the first echo of the fat quantification acquisition along with the fat fraction maps. The fat fraction calculated for subject 1 was 4.0%, and the fat fraction calculated for subject 2 was 1.7%.

Discussion

High quality, normalized mDIXON images were obtained from several subjects and highlight the potential for mDIXON to be used as a rapid workhorse anatomical reference scan at 7T. Promising results were seen across several areas in the body including the kidneys, liver and bowel.

Accurate fat quantification proves a challenge at 7T. The magnitude fitting used here has the advantage over fitting complex data in that it is less affected by shimming errors. These are likely to be more of a problem at 7T, given that the gradients available do not allow us to shorten the echo time as the frequency shift between water and fat increases. However magnitude imaging miscalculates fat fractions above 50%, so that any fraction above this is misreported as a high water voxel. Despite this the preliminary maps are promising and show that at 7T in vivo fat quantification is feasible.

Conclusions

Multiecho acquisition with MDIXON reconstruction can provide a robust set of work horse images for body rapid imaging at 7T. It has also been used for preliminary measurement of fat fraction at 7T.

Acknowledgements

References

1. 1. Vaughan et al, Whole-Body Imaging at 7T: Preliminary Results, MRM. 2009;61:244-248
2. 1. Raaijmakers et al, The Fractionated Dipole Antenna: A New Antenna for Body Imaging at 7 Tesla, MRM. 2016;75:1366-1374
3. Reeder, S.B., et al., Iterative decomposition of water and fat with echo asymmetry and least-squares estimation (IDEAL): Application with fast spin-echo imaging. Magnetic Resonance in Medicine, 2005. 54(3): p. 636-644.