0486
Full coverage 31P MRSI of the liver with a body coil at 7T1Radiology, University Medical Center Utrecht, Utrecht, Netherlands
Synopsis
We demonstrate increased 31P metabolite sensitivity by acquiring 31P signals over a large volume in the liver of a volunteer using a 31P whole body birdcage coil at 7T and minimize spatial reach of muscle signal leakage plus investigate frequency alignment due to B0 homogeneities. Sufficient SNR could be obtained in weighted average spectrum of all liver voxels and of four local liver voxels. Correcting B0 inhomogeneities by aligning resulted in a two-fold and 30% increase compared to non-aligned for the liver voxel average and the local voxel average respectively. The full setup allowed for full liver coverage and minimized muscle signal leakage.
Introduction
Conventional therapy in liver metastasis is often palliative care, as therapy effectiveness can only be monitored after several months. Metabolic imaging with 31P MRSI can potentially provide insight in tumor staging and therapy effectiveness, thus aiding in therapy response monitoring before any morphological changes are visible. However, current 31P applications in the body are still limited as conventional surface transceivers used with 31P MRSI, are restricted to small field of views. Surface coils require high energetic adiabatic pulses to provide a uniform flip angle distribution, at the cost of increasing SAR and thus limiting SNR per unit of time. 31P body coils are proposed to overcome the need for high SAR adiabatic RF pulses as these provide uniform transmit RF fields over larger field of views1. Previous studies have shown muscle signal leakage into liver spectra, therefore discarding numerous voxels2-4. Pohmann et al. reported on the feasibility of 31P MRSI in the heart, using conventional surface coils and weighted k-space sampling acquisition using a Hamming filter5. In combination with small nominal voxel sizes, the spatial reach of signal leakage from muscles can be minimized. However, small voxels result in low SNR, which in combination with the low concentrations of the observed 31P metabolites, make this approach challenging. Here, we propose to increase metabolite sensitivity by acquiring 31P signals from a large volume, and investigate the need for frequency alignment due to B0 inhomogeneities. Using acquisition weighted 3D spectroscopic imaging in combination with a 31P whole body birdcage coil, we obtained full liver coverage at 7T, with the ability to exclude contaminated spectra and align spectra prior to averaging to counteract B0 offsets over the large field of view in the liver.Methods
31P MRSI was performed using an in-house designed 31P whole body birdcage coil integrated in a 7T MR system (Philips Healthcare, Best, Netherlands) in combination with two 31P receive coils in quadrature mode. The body coil, tuned at 120Mhz is powered by a 25kW amplifier. Two fractionated dipole antennas were driven in quadrature transceiver mode to allow simultaneous proton MRI for anatomy localization and image based B0 shimming6. A volunteer was positioned in a right decubitus position on the 31P receiver coil with the liver on top, to reduce motion induced B0 inhomogeneity, and the arms stretched upwards. Phosphorous (31P) spectra were acquired using a 3D 31P chemical shift imaging protocol with Hamming weighted acquisition at a 15mm isotropic nominal resolution using a flip angle of 8° and a TR of 60ms for optimal SNR. Other CSI parameters were TE, 0.44ms; Bandwidth, 4800Hz; matrix size, 12x8x8; NSA, 80 and 256 acquisition points for a total scan duration of 21:48min. All data were processed in Matlab 2017b. 31P MRSI data were averaged using hamming weighting and filtering. FIDs were apodized using a 24Hz Gaussian filter and zero filled. Zeroth-order phase corrections were applied by changing the phase to match the real signal to the magnitude signal of α-ATP of the non-phased spectrum. Liver voxels with an SNR of PCr above four plus voxels with an SNR of α-ATP below 2.5 were excluded from processing. Spectra were aligned to the α-ATP peaks prior to generating the weighted average of the remaining liver volume and of four local liver voxels, using the SNR of α-ATP as the weighting factor. First order phase corrections were applied on both the aligned and nonaligned averages. Spectral SNR was calculated by dividing the real part of the signal from a metabolite peak by the absolute standard deviation of the last 50 noise samples of the spectrum.Results
Spectra from a large field of view in the liver were acquired as shown in figure 1. Alignment of the spectra increased the SNR for all labeled metabolites by 100% on average for all included liver voxels (figure 2) and 30% on average for the four local liver voxels (figure 3). Practically no PCr is visible in the spectra from the full liver nor in the four voxel average relatively close to the muscles.Discussion and Conclusion
The body coil setup,
positioning of the volunteer and chosen protocol parameters allowed for full
liver coverage and minimized signal leakage from the muscles. Averaging of the
liver voxels regained metabolite signal without introducing PCr signal.
Alignment of the spectra prior to averaging to minimize B0
inhomogeneity by respiratory motion increases spectral SNR two-fold. As the
current setup only used two receiver coils, further SNR improvements of the
setup by using a full receiver array may still be expected, possibly increasing
metabolite discrimination7.Acknowledgements
No acknowledgement found.References
1. J. Löring et al., “Whole-body radiofrequency coil for (31) P MRSI at 7 T,” NMR Biomed, vol. 29, no. 6, pp. 709–720, Jun. 2016.
2. M. Chmelik et al., “In vivo 31P magnetic resonance spectroscopy of the human liver at 7 T: an initial experience,” NMR in Biomedicine, vol. 27, no. 4, pp. 478–485, Apr. 2014.
3. Q. van Houtum et al., “Large FOV phosphor MR Spectroscopic imaging with multi-transmit proton MR imaging in the liver at 7 tesla”, “ISMRM 26 (2018)”, 0623.
4. J. H. Runge et al., “2D AMESING multi-echo 31P-MRSI of the liver at 7T allows transverse relaxation assessment and T2-weighted averaging for improved SNR,” Magnetic Resonance Imaging, vol. 34, no. 2, pp. 219–226, Feb. 2016.
5. R. Pohmann et al., “Accurate phosphorus metabolite images of the human heart by 3D acquisition-weighted CSI,” Magnetic Resonance in Medicine, vol. 45, no. 5, pp. 817–826, May 2001.
6. A. J. E. Raaijmakers et al., “The fractionated dipole antenna: A new antenna for body imaging at 7 Tesla,” Magn Reson Med, vol. 75, no. 3, pp. 1366–1374, Mar. 2016.
7. L. Valkovič et al., “Using a whole-body 31P birdcage transmit coil and 16-element receive array for human cardiac metabolic imaging at 7T,” PLoS ONE, vol. 12, no. 10, p. e0187153, 2017.
Figures
