Introduction

Magnetic resonance spectroscopy and spectroscopic imaging is used to obtain information on ³¹P metabolites related to, e.g., energy or cell membrane metabolism. Despite moving to high field strengths, such as 7T or higher, the resolution of CSI acquisition remains fairly course with long scan times due to the low natural abundance of ³¹P in the body.
Recent developments incorporated a transmit body coil tuned for ³¹P, 121MHZ, and a 16 channel ³¹P receiver array into the system^1,2,3, to facilitate homogeneous B₁⁺ fields combined with sensitive receiver arrays. These developments allow to increase spatial coverage and resolution with high SNR in reasonable acquisition times for whole body CSI. As importantly, this opens up the possibility to perform ³¹P MRI at 7T and the use of all MRI methods available for ¹H considering distinct metabolic profiles with substantial variances in spatial domain (figure 1).
In this study, we investigated the feasibility to obtain ³¹P gradient echo MRI images to measure the ³¹P distribution in muscle, and compared this to conventional ³¹P CSI.

Methods

All experiments were performed at a whole body 7T MR system (Philips Healthcare, Best, NL) equipped with a home-built integrated quadrature transmit body coil and a 16-channel local receiver array, both tuned for ³¹P (121 MHz). Proton information was obtained using an 8-channel Tx/Rx antenna array embedded in the receiver array.
A set of three ³¹P experiments was performed on a phantom consisting of 2 tubes (2cm diameter) containing a 0.1M phosphate solution embedded in 1H containing tube, as well as on the thigh muscles of a healthy volunteer. The experiments comprised a 3D CSI scan, a 2D multi-slice gradient echo scan, and a 2D multi-slice multi-echo gradient scan, detailed parameters are shown in the table in figure 2. For both cases, a ¹H gradient echo scan was performed for localization purposes.
Data reconstruction and processing was performed offline using the open source package QMRI Tools⁴. For both the imaging and the CSI data the raw 16 channel data was reconstructed using SENSE reconstruction where the coil sensitivity maps were estimated by dividing the individual coil data by the sum of squares addition of all coils.

Results

Figure 3 shows the gradient echo image of the phantom experiment, where both tubes were clearly visible.
Figure 4 shows the transversal multi-echo gradient echo ³¹P images of the in-vivo experiment where both legs are clearly visible. In the top row displays the results of the multi-slice gradient echo image. The bottom row displays the magnitude images of the multi-echo acquisition, as well as the fitted T₂^* map. Fitted T₂^* values were in the order of 20ms. Signal voids in the leg are the bones, where no ³¹P signal is expected.
Figure 5 shows the CSI spectra overlaid on the ¹H as well as the ³¹P MRI acquisition, both for the phantom and in-vivo experiments. A good correlation was found between the ¹H image, the ³¹P image and the CSI data, where the ³¹P has a better spatial localization compared to the CSI data.

Discussion and conclusion

This study shows that ³¹P MR imaging is feasible, using a transmit body coil and a receiver array at 7T. Currently, ³¹P MRI does not quantify the metabolites to the extent as MRS does but is can allow for better spatial localization. The feasibility of ³¹P MRI will allow the use of all quantitative ¹H MRI methods. Therefore, using methods such as Dixon can potentially allow for metabolite quantification.

Acknowledgements

No acknowledgement found.