2965

Volumetric mapping of the intracellular magnesium ion concentration in human calf muscles using ³¹P MRSI at 7 Tesla

Vanessa L. Franke¹, Mark E. Ladd¹, Peter Bachert¹, and Andreas Korzowski¹
¹Division of Medical Physics in Radiology, German Cancer Research Center (DKFZ), Heidelberg, Germany

Synopsis

Intracellular free magnesium ion concentration [Mg²⁺_free] is an important parameter in the regulation of the energy metabolism and can be assessed non-invasively by the use of ³¹P MRSI. The purpose of this study is to show that volumetric mapping of [Mg²⁺_free] in the human calf muscle is possible at B₀=7T with a voxel size of 1 ml. The [Mg²⁺_free] maps of healthy volunteers showed local differences, presumably resulting from different muscle fiber compositions. Two different approaches employed for the calculation of [Mg²⁺_free] yielded comparable maps but different absolute values, indicating the need for further investigation.

Introduction

Phosphorus magnetic resonance spectroscopic imaging (³¹P MRSI) enables the localized non-invasive investigation of the energy metabolism in vivo. By using the chemical shift differences of the observed phosphorus resonances, one can calculate intracellular pH values as well as free magnesium ion concentrations ([Mg²⁺_free])^1-3, which may provide deeper insight into the metabolic processes of the investigated tissue. Recently, we observed local pH differences for different muscle groups when performing volumetric ³¹P MRSI at B₀=7T with high spatial resolution⁴. To determine [Mg²⁺_free] by means of ³¹P MRSI, up to now different approaches have been applied at field strengths of 1.5 and 3T in the human brain and calf muscle^5-7. The mentioned studies suggest local differences in [Mg²⁺_free], but were performed with only low spatial resolution.
The purpose of this study was to demonstrate feasibility of volumetric mapping of the intracellular magnesium ion concentration in human calf muscles with high spatial resolution using ³¹P MRSI at B₀=7T.

Methods

Four healthy volunteers (3 male / 1 female, age: 25-32) were examined on a 7-T whole-body MR system (Siemens Healthineers) using a double-resonant ³¹P-¹H volume resonator (RAPID Biomedical). ³¹P MRSI data of calf muscles were acquired using an acquisition-weighted 3D CSI sequence (parameters: matrix size = 24×24×16, spatial resolution = 8×8×16 mm³, T_R = 240ms, α = 20°, Δf = 5000 Hz, 1024 time points, Hamming-weighted k-space averaging with 16 averages in k-space center, T_tot = 56 min). All MRSI datasets were processed by one-fold spatial zero-filling and application of a 10-Hz Gaussian filter in the time domain. Localized ³¹P spectra were evaluated for seven detectable resonances (Phosphocreatine (PCr), three times Adenosine-5’-Triphosphate (ATP), Nicotinamide-Adenine-Dinucleotide (NAD), inorganic phosphate (P_i), Glycerophosphocholine (GPC)) using a customized Matlab (The MathWorks, Inc.) implementation of the AMARES algorithm⁸.
[Mg²⁺_free] was calculated using the local chemical shift separation of the α- and β-resonance of ATP, employing two different approaches. First, we calculated [Mg²⁺_free] according to the approach after Widmaier et al.⁶ Second, we applied the approach by Barker et al.⁵, taking also into account influences of locally varying pH values for the calculation of [Mg²⁺_free]. For this, intracellular pH values were calculated using the chemical shift difference between P_i and PCr according to the modified Henderson-Hasselbalch equation¹. Voxelwise calculation yielded volumetric pH and [Mg²⁺_free] maps for each volunteer.
Regions of interest (ROIs) covering the muscles tibilias anterior (TA) and gastrocnemius medialis (GM) were drawn on the central transversal slice of the morphological ¹H images using MITK⁹. The average number of voxels covered by the ROIs was (41±7) in the TA and (44±7) in the GM. Mean values and standard deviations of [Mg²⁺_free] values were calculated for each volunteer individually, as well as across all volunteers.

Results

In all volunteers, localized ³¹P MR spectra of high quality were obtained, enabling the robust quantification of all detectable metabolites (Figure 1). These spectra revealed local differences in the chemical shifts of β-ATP and P_i, which persist after correction of local B₀shifts (Figure 2). The calculated [Mg²⁺_free] maps reflect these local differences (Figure 3a and b) and show comparable structures in all volunteers. Local differences were also observable in the pH maps (Figure 3c).
The [Mg²⁺_free] maps generated according to the two mentioned approaches resemble each other in structure, but deviate in the calculated absolute values (Figure 3a and b). Absolute [Mg²⁺_free] values calculated according to Widmaier et al. deviated on average by 0.33 mM from the values calculated after Barker et al.
The ROI analysis showed varying [Mg²⁺_free] values for different muscle groups (Table 1).

Discussion

The observed local differences in β-ATP chemical shifts most likely reflect physiological differences (e.g. due to different muscle fiber compositions), as B₀ field inhomogeneity (shifts persist after correction) and fitting uncertainties (Cramer-Rao Lower Bounds are smaller than the observed chemical shift differences) can be excluded as reasons. This is further supported by similar observed structures in pH and [Mg²⁺_free] maps. A mutual effect on pH and [Mg²⁺_free] calculation is unlikely, as [Mg²⁺_free] concentrations are too low to explain the P_i shifts and β-ATP shifts are too strong to be induced by pH only.
Local differences in [Mg²⁺_free] have been reported before⁶, and the herein calculated values using the approach by Widmaier et al. agree with the reported ones for the muscle gastrocnemius. However, the calculated [Mg²⁺_free] values using the approach by Barker et al. are significantly higher than those reported previously^6,7. A possible reason could be the inclusion of the local pH value into the calculations of [Mg²⁺_free] in this approach.
The total measurement time of the acquired datasets at 56 min is relatively long, but could be potentially reduced by factor of 2, as the signal-to-noise ratio of the spectra is sufficiently high.

Conclusion

³¹P MRSI at B₀=7T enables the volumetric mapping of free, intracellular magnesium ion concentration in the human calf muscle, with voxels sizes of about 1 ml. The local differences in [Mg²⁺_free] values observed in healthy volunteers most likely result from different muscle fiber compositions of various muscle groups. Depending on the evaluation method, different absolute [Mg²⁺_free] values were obtained, showing the need for future validation of these approaches.

Acknowledgements

No acknowledgement found.

References

1. de Graaf RA. In Vivo NMR Spectroscopy: Principles and Techniques: 2nd Edition.; 2007.

2. Gupta RK, Benovic JL, Rose ZB. The determination of the free magnesium level in the human red blood cell by 31 P NMR. J Biol Chem. 1978;253(17):6172-6176.

3. Golding EM, Golding RM. Interpretation of 31PMRS spectra. Neurology.

4. Korzowski A, Bachert P. High-resolution 31 P echo-planar spectroscopic imaging in vivo at 7T. Magn Reson Med. 2018;79(3):1251-1259.

5. Barker PB, Butterworth EJ, Boska MD, Nelson J, Welch KMA. Magnesium and pH imaging of the human brain at 3.0 Tesla. Magn Reson Med. 1999;41(2):400-406.

6. Widmaier S, Hoess T, Jung WI, Staubert A, Dietze GF, Lutz O. 31P NMR Studies of human soleus and gastrocnemius show differences in the Jγβ coupling constant of ATP and in intracellular free magnesium. Magn Reson Mater Physics, Biol Med. 1996;4(1):47-53.

7. Iotti S, Malucelli E. In vivo assessment of Mg2+ in human brain and skeletal muscle by 31P-MRS. Magnes Res. 2008;21(3):157-162.

8. Vanhamme L, Van Huffel S. AMARES: Advanced Method for Accurate, Robust and Efficient Spectral fitting of MRS data with use of prior knowledge. J Magn Reson. 1997;43(129):1-2.

9. Nolden M, Zelzer S, Seitel A, et al. The medical imaging interaction toolkit: Challenges and advances: 10 years of open-source development. Int J Comput Assist Radiol Surg. 2013;8(4):607-620.

Figures

Figure 1: Representative ³¹P spectra localized in the muscles tibialis anterior (top) and gastrocnemius medialis (bottom) of one healthy volunteer. Locations of the displayed voxels are depicted in Figure 2. Spectra (black) are corrected for local B₀ shifts and overlaid with the corresponding fitted spectra (red). The chemical shifts of P_i and β-ATP deviate between the two muscles, whereas the chemical shifts of all other metabolites do not show any visible differences.

Figure 2: Representative transversal map of the chemical shift difference (δ) between α-ATP and β-ATP in one volunteer overlaid on the morphological ¹H image. The map shows local differences in the chemical shifts of up to 0.15 ppm. The voxels displaying the representative ³¹P spectra of the tibilias anterior and gastrocnemius medialis in Figure 1 are marked in red.

Figure 3: Calculated [Mg²⁺_free] maps following the approach of a) Widmaier et al. and b) Barker et al., as well as the c) pH map generated according to the modified Henderson-Hasselbalch equation for one representative volunteer (note the different scales in pannels a and b). The maps are overlaid onto morphological ¹H images. The ROIs covering the tibilias anterior (blue) and gastrocnemius medialis (red) were drawn for each volunteer separately in MITK⁹.

Table 1: Results of the ROI analysis for the calculated [Mg²⁺_free] values following the method of Widmaier et al. (Wid.) and Barker et al. (Bar.). Mean values and standard deviations of [Mg²⁺_free] over all voxels of the drawn ROI in the tibilias anterior (TA) and gastrocnemius medialis (GM) are given for each individual volunteer, and also averaged over all volunteers.

Proc. Intl. Soc. Mag. Reson. Med. 28 (2020)

2965