The tissue–pH value is an important parameter to assess physiological function. A noninvasive technique for determination of intracellular pH value is in vivo ³¹P NMR spectroscopy¹. Three-dimensional (3D) ³¹P spectroscopic imaging at high field strength B₀ enables spatial resolution (voxel size) of 1 ml and less. Total measurement time T_tot is reduced by employing echo–planar spectroscopic imaging (EPSI)².

The purpose of this work was to explore the potential of ³¹P–{¹H} EPSI at B₀ = 7 T for mapping of intracellular pH in the human calf muscle.

The calf muscle of three healthy volunteers (2 female / 1 male, age: 22–31 y, weight: 60–75 kg, volunteer #1 and #2 with comparable daily physical activity) was measured with a 3D ³¹P–{¹H} EPSI sequence utilizing the ³¹P–{¹H} nuclear Overhauser effect (NOE) and acquisition weighting of the central k–space lines for signal enhancement (sequence parameters: voxel size 7 x 7 x 20 mm³, matrix 32 x 32 x 16, 4 interleaves, frequency–selective excitation with α = 30°, Δf = 2000 Hz, 512 echoes, 24 weighted averages, NOE preparation during 100 ms, T_R = 390 ms, T_tot = 66 min). The measurements were performed on a MAGNETOM 7 T (Siemens Healthcare, Erlangen, Germany) using a double–resonant ³¹P/¹H volume coil (Rapid Biomedical, Rimpar, Germany). EPSI datasets were reconstructed and corrected with an own MATLAB routine (The MathWorks, Natick, MA, USA). Localized ³¹P spectra (postprocessed with Hamming–windowed k–space interpolation to a 64 x 64 x 16 matrix and application of Gaussian line broadening by 11 Hz) were evaluated with the AMARES algorithm³ implemented in jMRUI⁴. Intracellular pH value was calculated for each voxel using the modified Henderson–Hasselbalch equation $$$pH = pK_A + log\frac{\delta - \delta_{HA}}{\delta_A - \delta}$$$ with pK_A = 6.77, δ_HA = 3.23 ppm, δ_A = 5.70 ppm (values from Ref. 5), and δ the frequency difference between the phosphocreatine (PCr) and inorganic phosphate (P_i) resonance.Figure 1 shows a representative in vivo ³¹P EPSI spectrum with resolved PCr, P_i, and ATP (adenosine 5’–triphosphate) peaks from a 0.98–ml voxel in the M. gastrocnemius of volunteer #2. In all volunteers, the ³¹P spectra of voxels localized in the most sensitive volume of the coil yielded sufficient signal–to–noise-ratio (SNR) for accurate quantification of the P_i–PCr frequency difference and for calculation of 3D pH maps. Figure 2 shows two pH maps from volunteer #2. The pH maps obtained from all the other volunteers were of the same quality. The transversal slices showed distinct pH variations between different muscle groups (Tab. 1). In all volunteers the mean pH value in the tibialis anterior muscle was lower than in the soleus and medial gastrocnemius muscle. The series of transversal pH maps revealed that the distribution of pH differences follows structures arranged along individual muscle groups (coronal view in Fig. 2b).

The data demonstrate that the proposed 3D EPSI method allows the robust quantification of intracellular pH with high spatial resolution. The observed pH variations may result from differences in energy metabolism determined by the muscle fiber type composition (oxidative vs. glycolytic).

High–resolution ³¹P spectroscopic imaging at high field strength allows to map pH heterogeneity of different tissues. The total measurement time for a 3D ³¹P EPSI dataset can be reduced to less than 30 min by choosing slightly lower in–plane resolution and thicker slices.