Yuchi Liu1, Yun Jiang1, Charlie Yi Wang1, Mark Alan Griswold1,2, and Xin Yu1,2,3,4
1Biomedical Engineering, Case Western Reserve University, Cleveland, OH, United States, 2Radiology, Case Western Reserve University, Cleveland, OH, United States, 3Physiology and Biophysics, Case Western Reserve University, Cleveland, OH, United States, 4Case Center for Imaging Research, Case Western Reserve University, Cleveland, OH, United States
Synopsis
In this study, a fast phosphocreatine (PCr) imaging
method was developed by combining spectrally selective bSSFP with single-shot
spiral-in/out encoding for measuring PCr distribution in vivo. 2 min acquisition yielded an SNR of ~9 in rat calf muscles
with high spatial resolution (1.25 mm×1.25 mm×5 mm)
at 9.4T. The SNR was increased to 25.5 with 30 min acquisition. This method also has the potential for imaging other
metabolites such as ATP with a different carrier frequency, dynamic PCr imaging
in exercise-recovery or ischemia-reperfusion studies, and quantification of the
absolute metabolite concentration with appropriate T1, T2,
and flip angle calibration/correction.Target
Audience
Researchers interested in hetero-nuclei imaging and
non-Cartesian acquisition scheme.
Introduction
31P MR spectroscopy has
been used to quantify non-localized high-energy phosphate in vivo. Spectroscopic
imaging offers both spatial and spectral encoding. However, due to the low
metabolite concentration,
31P spectroscopic imaging with high
spatial resolution requires long acquisition time that renders the technique impractical.
Alternatively, imaging using spectrally selective excitation allows only one
metabolite to be imaged using conventional imaging sequence. Further
acceleration can be achieved by using a balanced steady-state free precession
(bSSFP) approach. In addition, the relatively long T
2 relaxation
time of phosphocreatine (PCr) provides more favorable T
2/T
1 contrast in
bSSFP
1. In this study, a fast PCr imaging method
was developed by combining spectrally selective bSSFP with single-shot
spiral-in/out encoding for measuring PCr distribution in vivo.
Methods
A user-defined Gaussian pulse with 4 ms duration was
designed to spectrally select PCr. For image acquisition, a single-shot, zero moment compensated spiral trajectory was designed using the minimum-time gradient method2.
The
spiral waveform was executed twice within one repetition time (TR) in a
spiral-in/out fashion and the center k-space was acquired at the middle of each
TR. The spiral-in/out trajectory was measured manually3 and images
were reconstructed with NUFFT4 using the measured trajectory. 2D
imaging was performed by combining the spectrally selective pulse and outer
volume saturation without slice selection gradient.
All experiments were performed on a Bruker 9.4T horizontal
scanner (Bruker Biospin Co., Billerica, MA) using an in-house built 31P
transmit/receive saddle coil. The selectivity of the Gaussian pulse was evaluated
on a 500 mM inorganic phosphate (Pi) phantom by sweeping the carrier frequency
from -5 to 5 ppm with a step size of 0.5 ppm. The Gaussian pulse combined with
the spiral-in/out bSSFP sequence was validated on two phantoms with 40 mM
phosphocreatine (PCr) and 20 mM adenosine triphosphate (ATP), respectively.
In vivo PCr imaging
of the calf muscle was performed on a 10 month-old Fischer rat. An M2M volume
coil was used for proton imaging. The B1 fields of the 31P
coil and the 1H coil were placed orthogonal to minimize the coupling
between the two coils. Localized shimming using PRESS was performed before the
imaging experiments. Reference 1H image was acquired using FLASH
sequence. Acquisition parameters were: TR, 500 ms; TE, 7 ms; flip angle, 40°; FOV
4×4 cm2;
matrix size 256×256; slice thickness,
5 mm. PCr image was acquired using the proposed spiral-in/out bSSFP sequence
with outer volume saturation. Acquisition parameters were: TR, 20 ms; TE, 10
ms; flip angle, 25°; FOV 4×4 cm2; matrix size 32×32; slice thickness, 5 mm. The saturation module
was implemented at every 32 TRs, followed by 10 dummy scans. A delay of 350 ms
was also inserted between every 32 TRs to prevent gradient from overheating. A
total of 48000 repetitions were acquired in 30 min. Image reconstruction used
an average of 1600, 3200, 8000, 16000, 24000, 32000, 40000, and 48000 repetitions,
corresponding to an acquisition time of 1, 2, 5, 10, 15, 20, 25, and 30 min,
respectively.
Results
Designed and measured trajectory is shown in Fig. 1.
Fig. 2 shows the selectivity of the 4-ms Gaussian pulse. Residual signal at 2.5
ppm, where γATP resonance occurs,
was 5.8%. This profile allowed the excitation of PCr only. Fig. 3 shows the
reference anatomy image (Fig. 3a) and PCr selective image (Fig. 3b) using the
proposed method on PCr and ATP phantoms.
In vivo PCr images were reconstructed using different
number of signal averages, corresponding to an acquisition time of 1, 2, 5, 10,
15, 20, 25, and 30 min, respectively, and the corresponding signal to noise
ratio (SNR) was calculated (Fig. 4). 2 min acquisition yielded an SNR of ~9. The
SNR was increased to 25.5 with 30 min acquisition. Representative
images with 5, 10, 20, and 30 min acquisition time are shown in Fig. 5. PCr
images on phantom and in vivo were zero padded to matrix size 64×64, and FOV of
2cm×2cm is shown.
Discussion
& Conclusion
This study shows a fast PCr imaging
method with spiral encoding. PCr map of rat calf muscles with high spatial
resolution (1.25 mm×1.25 mm×5 mm) was achieved in acceptable acquisition time. With
a different carrier frequency, this method can also be used for imaging other
metabolites such as ATP. Quantification of the absolute metabolite
concentration can be achieved with appropriate T
1, T
2, and
flip angle calibration/correction. This method also has the potential for
dynamic PCr imaging in exercise-recovery or ischemia-reperfusion studies.
Acknowledgements
This work was supported by National Heart, Lung, and Blood
Institute Grants R21 HL126215.References
1. Speck O et al. MR. 2002.
2.
Lee et al. MRM. 2003.
3.
Duyn et al. JMRI. 1998.
4. Fessler and Sutton. IEEE Trans. Signal Process. 2003.