Meng Gu1, Adam Kerr2, Ralph Hurd3, and Daniel Spielman1
1Radiology, Stanford University, Stanford, CA, United States, 2Electrical Engineering, Stanford University, Stanford, CA, United States, 3GE Healthcare, Menlo Park, CA, United States
Synopsis
MEGA PRESS has been used to edit the GABA resonance
at 3ppm. Due to the wide transition bandwidth of the editing pulse,
macromolecule resonances are coedited. To suppress macromolecule signals, a
symmetric suppression method has been proposed resulting in reduced GABA
signal. We present a new editing method by incorporating spatial and spectral
selectivity into the SPECIAL refocusing RF pulses to achieve both GABA editing
and macromolecule suppression. Phantom studies showed higher edited GABA signal
compared with MEGA PRESS and 90% macromolecule suppression. In-vivo studies
demonstrated significantly higher edited GABA signal compared with MEGA PRESS.
Purpose
Gamma-aminobutyric-acid (GABA) is the major
inhibitory neurotransmitter in the human brain and in-vivo measurement of GABA
using magnetic resonance spectroscopy offers valuable information in
understanding brain functions. Based on J-difference editing, MEGA PRESS [1],
has been used to detect the GABA resonance at 3ppm. In order to achieve the
maximum signal strength, the editing pulse is generally applied at 1.9ppm and
7.5ppm with an echo time of 68ms. However, due to the wide transition bandwidth
of the editing pulse, macromolecule resonances are coedited, with the signal at
3ppm mostly contributed from lysine residues in macromolecules, whose 3ppm
resonance is coupled with its 1.7ppm resonance. To suppress macromolecule
signals, a symmetric suppression method has been proposed where the editing
pulse is applied at 1.9ppm and 1.5ppm. The assumption of this method is that the
macromolecule 1.7ppm resonances are equally affected from the editing pulses so
that the partially edited resonances at 3ppm are cancelled out in the edited
spectrum. Unfortunately, this method results in reduced GABA signal because the
editing pulse applied at 1.5ppm partially inverts the 1.9ppm resonance even
when longer and more selective editing pulses are used at a longer TE of 80ms
[2].
Here, we present a new editing method by
incorporating spatial and spectral selectivity into the
refocusing RF pulses of the SPECIAL sequence with 1D ISIS based localization to
achieve both GABA editing and macromolecule suppression. Because a single spin
echo is used for localizaiton, longer pulses with narrower transition band can
be used than the double spin-echo used in MEGA-PRESS . As shown in Figure 1, by
using spectral-spatial refocusing RF pulses with narrow transition bandwidths,
GABA resonance at 3ppm can be edited by refocusing ("ON" case)/not
refocusing ("OFF" case) the 1.9ppm resonance. As the lysine 1.7ppm
resonance is not refocused in either case, its resonance at 3ppm is suppressed
in the edited spectrum.
Methods
A spectral-spatial 180° pulse with a linear-phase
spectral filter was designed with a 2D inverse SLR approach to achieve
refocusing with narrow transition bandwidth [3-5]. The variable-rate selective
excitation (VERSE) algorithm was applied to reduce the peak B1 for each
subpulse [6]. The final pulse used had a 66ms duration, 260Hz spectral
bandwidth, 22Hz transition bandwidth and a 27µT peak B1. Pulse and gradient
waveforms are shown in Figure 2.
RESULTS
A 50mM GABA phantom
and a 50mM lysine phantom were built to test the GABA editing and macromolecule
suppression. The spectral-spatial based editing sequence was compared with the
MEGA PRESS with symmetrical suppression at TE/TR=80ms/2s, 16 averaging with a
voxel size of 2.5x2.5x2.5cm. The MEGA PRESS editing pulses were applied at
1.9ppm and 1.5ppm while the spectral-spatial pulse frequencies were shifted by
30Hz. Spectra of the "ON" and "OFF" cases and the edited
spectrum for the MEGA PRESS and spectral-spatial SPECIAL are shown in Figure 3
and Figure 4. Measured from the phantom experiments, the edited GABA signal
using spectral-spatial PRESS is 24% higher than that using MEGA PRESS with
symmetrical suppression. The MEGA PRESS achieved 88% lysine suppression while
the spectral-spatial PRESS achieved 90% lysine suppression. The comparison of
MEGA PRESS and spectral-spatial SPECIAL was performed on several human subjects
with the same prescription as the phantom studies except for with 64 averaging
at an acquisition time of 4:30 minutes. Shown in Figure 5 are the results from
a representative 42-year-old healthy male subject. The measured GABA signal
with spectral-spatial PRESS is 28% higher than with MEGA PRESS.
CONCLUSIONS
A new editing method
based on spectral-spatial RF pulses was developed for GABA editing and
macromolecule suppression. Phantom studies showed higher edited GABA signal
compared with MEGA PRESS with symmetrical suppression and 90% lysine
suppression. In-vivo studies demonstrated significantly higher edited GABA
signal compared with MEGA PRESS with symmetrical suppression. In addition to more
efficient editing of the GABA signal, the use of spectral-spatial RF pulses
offers other valuable benefits of lipid suppression and improved spatial
selectivity compared with MEGA PRESS.
Acknowledgements
Lucas foundation, GE Health Care, NIH P41 EB 015891. References
[1] Mescher, M.M.,
et al. NMR Biomed, 11:266, 1998.
[2]
Henry, PG., et al. MRM 45:517, 2001.
[3]
Pauly, J., et al. MRM 29:776, 1993
[4]
Kerr, A.B., et al. Proc. ISMRM p226, 2008
[5] Larson, P.E., et al. JMR 194(1):121
2008
[6]
Larson, P.E., et al. Proc. ISMRM p3149, 2008