Ayhan Gursan1, Arjan D. Hendriks1, Dimitri Welting1, Pim A. de Jong1, Dennis W. Klomp1, and Jeanine J. Prompers1
1Department of Radiology, University Medical Center Utrecht, Utrecht, Netherlands
Synopsis
Deuterium Metabolic Imaging (DMI)
is a new technique, which could potentially be used to study liver metabolism
in vivo. The aim of this study was to develop a 4-channel 2H
transmit/receive array for DMI of the liver at 7T, combined with 4 1H
dipole elements for conventional anatomical MRI. Natural abundance DMI water
and lipid maps measured on a phantom showed good correspondence with 1H
Dixon water and lipid images. The multi-channel setup allowed us to pick up the
natural abundance 2H water signal throughout the whole liver in a 3D
DMI acquisition with a nominal resolution of 20x20x20 mm3.
Introduction
Deuterium (2H) metabolic
imaging (DMI) is a new emerging technique to study metabolism in vivo and has
been applied both in rats1,2 and in humans1. The disturbed metabolism of
glucose in the diabetic liver could potentially be investigated non-invasively
with DMI after oral intake of deuterated glucose. As the sensitivity of DMI
scales supralinear with the magnetic field strength3, the application of DMI at ultra-high
field leads to a significant gain in SNR, which can be employed to reduce scan
time or to improve spatial resolution. In
addition, DMI is characterized by
relatively low radiofrequency (RF) power requirements, making the method
extremely suited for application at ultra-high field, while 1H-MRI/MRS
methods are typically challenging at high field due to much higher RF power
demands. The aim of this study is to develop a 4-channel 2H
transmit/receive array for DMI of the liver at 7T, combined with 4 1H
dipole elements for conventional anatomical MRI.Methods
A DMI body coil setup was developed
for a 7T whole-body MRI system (Philips
Healthcare, Best, Netherlands), consisting of 4 2H
transmit/receive loops. Two loops were incorporated in a posterior element and
two other two loops in an anterior element (Figure 1). The loops within one
element were geometrically decoupled. Each 2H transmit/receive loop
was combined with a 1H dipole element for reference anatomical imaging
and B0 shimming (Figure 1). All DMI measurements were performed with a
pulse-acquire sequence using a 500 μs
block pulse, followed by phase encoding gradients for 3D spatial encoding. Sensitivity
profiles of the 2H coils were assessed individually on a body size
phantom by driving each transmit coil consecutively. This was compared to an
acquisition in which all 4 transmit coils were driven simultaneously (TR= 333
ms, FA= 90°, CSI matrix= 6 x 7 x 7, voxel size= 40 x 40 x 40 mm3, NSA=
4). Next, DMI was performed on a multi-compartment phantom, filled with water
(inner compartment) and soybean oil (outer compartment). B1 and second order B0
shimming were applied before Dixon images were acquired, which were also used for
CSI planning. A 3D DMI scan was performed (TR= 333 ms, FA= 90°, CSI matrix= 13
x 13 x 11, voxel size= 15 x 15 x 15 mm3, NSA= 4), from which water
and lipid maps were generated by Lorentzian fitting and integrating the area
under the respective peaks. For the in-vivo scan, coils were placed on the right
lateral side of a healthy subject to maximize acquired signal from liver. The 3D
DMI grid was planned on a Dixon scan and covered the whole liver. Acquisition
was performed with elliptical k-space acquisition without respiratory gating (TR=
333 ms, CSI matrix= 12 x 14 x 13, voxel size= 20 x 20 x 20 mm3, NSA=
4, total acquisition time= 22 min). All 3D DMI data were constructed with Hamming
filtering in the spatial domain, 10-Hz Gaussian apodization in the spectral
domain and WSVD coil combination4 using the CSIgui toolbox5
in MATLAB.Results
Figure 2 shows sensitivity profiles
of the 4 2H coils, as measured from the natural abundance HDO signal
in a body size phantom, using single-channel transmit for each channel (left)
and 4-channel transmit (right). The acquired signal patterns were similar for
both cases, showing that there was no destructive interference between the
effective B1+ fields of the 4 coils. Figure 3 displays 1H Dixon
images (a,b) and DMI metabolite maps of the natural abundance 2H
water and lipid signals (c,d) measured on a multi-compartment phantom filled
with water and soybean oil, showing good correspondence in the localization of
the water and lipid signals. The sensitivity of the developed setup was
sufficient to pick up the natural abundance 2H signal from
water in the liver (Figure 4). Signal could be detected throughout the whole
liver; however, signal intensity dropped at further distances from the
posterior and anterior coil elements.Discussion and Conclusion
This is the first application of a multi-channel
body coil for DMI to our knowledge. Natural abundance DMI water and lipid maps measured
on a phantom showed good correspondence with 1H Dixon water and
lipid images. The developed multi-channel coil setup allowed us to pick up the
natural abundance 2H water signal throughout the whole liver in a 3D
DMI acquisition with a nominal resolution of 20 x 20 x 20 mm3. When
the setup is combined with oral intake of deuterium labeled glucose or water in
the next phase, it could enable us to study metabolic processes in the liver in
more detail.Acknowledgements
This
work was funded by a HTSM grant from NWO TTW (project number 17134).References
1.De
Feyter HM, Behar KL, Corbin ZA, et al. Deuterium metabolic imaging (DMI) for
MRI-based 3D mapping of metabolism in vivo. Sci Adv. 2018;4(8):1-12.
doi:10.1126/sciadv.aat7314
2.Lu M, Zhu XH, Zhang Y, Mateescu G,
Chen W. Quantitative assessment of brain glucose metabolic rates using
in vivo deuterium magnetic resonance spectroscopy. J Cereb Blood Flow
Metab. 2017;37(11):3518-3530. doi:10.1177/0271678X17706444
3.de Graaf RA, Hendriks AD, Klomp DW,
et al. On the magnetic field dependence of deuterium metabolic imaging (DMI). Proc
27th Annu Meet ISMRM. 2019; #0492.
4.Rodgers CT, Robson MD. Receive
array magnetic resonance spectroscopy: Whitened singular value decomposition
(WSVD) gives optimal bayesian solution. Magn Reson Med. 2010.
doi:10.1002/mrm.22230
5.CSIgui
v2, http://ww.csigui.tk,
visited on 25/09/2019.