0480

First assessment of cardiac energy metabolism in the lateral and inferior segments of the left ventricle in vivo at 7T
Ladislav Valkovic1,2, Jane Ellis1, Lucian AB Purvis1, Albrecht Ingo Schmid1,3, Stefan Neubauer1, and Christopher T Rodgers1,4

1OCMR, RDM Cardiovascular Medicine, University of Oxford, Oxford, United Kingdom, 2Department of Imaging Methods, Institute of Measurement Science, Slovak Academy of Sciences, Bratislava, Slovakia, 3High-Field MR Centre, Center for Medical Physics and Biomedical Engineering, Medical University of Vienna, Vienna, Austria, 4The Wolfson Brain Imaging Centre, University of Cambridge, Cambridge, United Kingdom

Synopsis

Using a novel whole-body transmit coil with an integrated 30-element receive array, powered by a 35kW RF power amplifier, we show the feasibility of whole-heart cardiac 31P-MRS at 7T. PSF simulations were performed to directly compare the new coil to a published protocol using a 16-element surface coil. The whole-body coil delivers analysable spectra from all left ventricular segments in the hearts of 3 volunteers, whereas the 16-element surface coil allowed probing energetics only in the proximal half of the left ventricle. This coil will enable new studies of focal disease and improve the SNR for regionally homogenous disease investigations.

Introduction:

Phosphorus MR spectroscopy (31P-MRS) plays an important role in the assessment of cardiac energy metabolism in vivo, giving unique insight into high-energy metabolites1,2. Yet the low intrinsic 31P-SNR and low peak 31P-B1+ from clinical scanners and surface coils typically limits cardiac 31P-MRS to measurements in the mid-ventricular septum. SNR can be considerably improved by using ultra-high field, e.g. 7T. But B1+ inhomogeneity of surface transmit coils at 7T still makes the acquisition of spatially-resolved 31P-MRSI across the heart challenging. Recently, a whole-body-sized RF birdcage coil for 31P at 7T was shown to generate homogeneous RF excitation for 31P-MRS across the human body3,4. Furthermore, in combination with a 35kW RF power amplifier (RFPA), strong and uniform B1+ was demonstrated in phantoms5.

This study demonstrates the potential of a whole-body coil transmit, 30-channel receive coil setup driven by a 35kW RFPA. We report the first regionally-resolved energy metabolism assessment across the left ventricle (LV) of healthy volunteers at 7T.

Materials and Methods:

The 31P whole-body transmit coil (MR Coils, Zaltbommel, Netherlands) is incorporated into an extension of the normal patient table, which rests on custom-built support rails at the service-end of the magnet4,5. The 30ch 31P receive array (16 anterior + 14 posterior circular loops, 88mm diameter) and 8-element 1H transceiver array (fractionated dipole antennas) is integrated inside the body coil, i.e. the posterior array is built into the new patient table5. The coil was powered by an MKS-S41 RFPA which was able to excite at 120.3MHz, i.e. the 31P resonance frequency of our 7T MR system (Siemens Healthcare, Erlangen, Germany).

Three healthy men (mean age 33±9, BMI 22.9±1.6 kg/m2) were recruited for this study and scanned lying supine in the whole-body coil with the anterior array positioned over their hearts. An acquisition weighted 3D UTE-CSI6 sequence was used for data acquisition. For a fair comparison of the coil performance against previous 7T 31P-MRS investigations in the heart7,8, the parameters of the sequence were selected based on point-spread-function (PSF) simulations to match the real voxel size and acquisition time of previously published studies. The original field-of-view (FOV) of 240×240 ×200 mm3 and matrix size of 16×16×8, had to be extended due to larger B1+/- profile of the new coil setup. Based on the PSF simulations, the new calculated FOV and matrix size were 480×360×280 mm3 and 28×26×10, respectively. To match the total acquisition time of ~28 minutes, a single average was acquired and TR was reduced from 1000 ms to 800 ms.

All volunteers were also scanned using the established protocol using a 16ch receive array with a local transmit loop (Rapid Biomedical, Rimpar, Germany)8. Spectra were combined with WSVD9 and the results from the two coils were compared for all segments of the LV.

Results and Discussion:

Figure 1 shows the results of the PSF simulations. Although the nominal voxel size was larger for the new acquisition parameters, i.e. 6.6 mL vs. 5.6 mL, due to more favourable PSF the final real voxel size is smaller for the body coil acquisition, i.e. 21.5 mL vs. 21.8 mL. Figure 2 compares spectral quality in several segments of the LV between the surface coil and the whole-body coil. Table 1 quantifies the increase in SNR over the LV segments. Good quality, i.e. high SNR, data is demonstrated in the mid-ventricular septum using the surface coil, in accordance with previous reports7,8. However, moving towards inferior and lateral segments of the heart, i.e. away from the coil, SNR decreases rapidly. However, using the whole-body coil, high SNR data are obtained from all segments of the LV; even in the voxels of the basal inferior wall of the LV where only noise was recorded by the surface coil, see Figure 3.

For studies of regionally homogenous cardiac disease such as dilated cardiomyopathy, it is reasonable to average all good quality spectra around the LV to give a global assessment of LV energetics. This is illustrated in Figure 4, where all voxels with SNR>1.5 were averaged leading to an increase in SNR from 13.4 to 17.6 for PCr for global LV energetics assessment.

Conclusion:

A whole-body transmit 31P coil with integrated 30-element receive array, driven by a 35kW RFPA has been successfully used to assess energy metabolism in the anterior and posterior myocardium. Superior data quality in comparison to a 16ch receive array with a 26x28cm2 surface transmit loop was demonstrated across all left-ventricular segments. These results show the first regionally-resolved, whole-heart cardiac 31P-MRSI at 7T.

Acknowledgements

Funded by a Sir Henry Dale Fellowship from the Wellcome Trust and the Royal Society (098436/Z/12/B); a Science Enhancement from the Wellcome Trust (Grant No. 098436/Z/12/A); the EPA Cephalosporin Fund (Grant No. CF 284); the Oxford BHF Centre of Research Excellence (Grant No.RE/13/1/30181); the Austrian Science Fund FWF (Grant No. J 4043); and Slovak Grant Agencies VEGA (Grant No. 2/0001/17) and APVV (Grant No. 15-0029).

References

1. Bottomley P, et al. Radiology 1987;165(3):703-7
2. Neubauer S, et al. Circulation 1997;96(7):2190-6
3. Loring J, et al. NMR Biomed 2016;29(6):709-20
4. Valkovič L, et al. PLoS ONE 2017;12(10):e0187153
5. Valkovič L, et al. Proc. ISMRM 2018;26:143
6. Robson MD, et al. Magn Reson Med 2005;53(2):267-74
7. Rodgers CT, et al. Magn Reson Med 2014;72(2):304-15
8. Stoll VM, et al. Radiology 2016;281(2):409-17
9. Rodgers CT & Robson MD. Magn Reson Med 2016;75(2):473-87

Figures

Figure 1: Comparison of 1D and 2D cross-sections of simulated point spread functions (PSF) for weighted acquisition MRSI. The previously used protocol with the field of view 240×240×200 mm3 and matrix size 16×16×8 is depicted in red and the new protocol for the whole body coil excitation using field of view 480×360×280 mm3 and matrix size 28×26×10 is depicted in blue. Note that higher number of phase encoding steps translates into a smaller real voxel size, i.e. 21.5 mL vs. 21.8 mL, for the whole-body coil despite a larger nominal voxel size, i.e. 6.6 mL vs. 5.6 mL.

Figure 2: Comparison of the spectral quality across several segments of the left ventricle (midseptal – blue; anterior – red; mid lateral – green; inferior - cyan) in the same volunteer. A) shows the data acquired using the surface coil and B) shows the data acquired using the birdcage coil. Spectra are depicted after apodization using a 25Hz exponential filter (no line broadening) and scaled to noise level of each coil separately and the SNR of each spectrum is given. Note that while lateral voxel in A) is only noise, in B) a good quality cardiac spectrum can be observed.

Table 1: Increase in SNR (in %) when using the new whole-body coil setup in comparison to the 16-element surface coil over the segments of left ventricle in all volunteers

Figure 3: Spectra acquired using the whole-body coil system from voxels in inferior segment of a more basal slice through the heart from all three volunteers. 25Hz exponential filter was applied for better visualization.

Figure 4: Averaged spectra from all the voxels of left ventricle with SNR > 1.5, contributing voxels are depicted in respective localizers. Spectra are normalized to the same noise level. Note that while only the voxels close to the surface coil contributed to the averaged spectrum, in case of the body coil all LV voxels were included leading to increased SNR.

Proc. Intl. Soc. Mag. Reson. Med. 27 (2019)
0480