Absolute metabolite concentrations were assessed in the muscle, the liver, and the kidney of healthy human volunteers by 31P MRS. Fully relaxed in vivo spectra were acquired with a surface coil and were localized with an adiabatic ISIS pulse sequence. The spectra were quantified with a subsequent measurement of a calibration phantom and were processed iteratively in the time domain. The following mean metabolite concentrations (mmol/liter) were measured in the resting male calf muscle (n = 9), in the fasting liver (n = 12), and in the orthotopic kidney (n = 5): [PME] = 2.0 +/- 0.6, 3.8 +/- 0.7, and 2.6 +/- 0.9, [Pi] = 2.9 +/- 0.3, 1.8 +/- 0.3, and 1.6 +/- 0.4, [PDE] = 3.8 +/- 0.8, 9.7 +/- 1.5, and 4.9 +/- 1.1, [PCr] = 22.0 +/- 1.2, 0, and 0, [NTP] = 5.7 +/- 0.4, 2.9 +/- 0.4, and 2.0 +/- 0.3, respectively. Several interesting findings are to be emphasized: The concentrations of Pi, PCr, and NTP were 20% lower in the muscle of women than of men. In addition, the pHi was significantly lower in female muscle (6.99 +/- 0.03) than in male muscle (7.05 +/- 0.03). The pHi in the liver (7.12 +/- 0.09) and in the kidney (7.09 +/- 0.08) were higher than in the muscle of both genders. The free magnesium concentration (mmol/liter) was higher in the liver (1.40 +/- 0.64) than in the kidney (0.79 +/- 0.39) and in the muscle (0.52 +/- 0.10).

In order to determine metabolite concentrations in human skeletal muscles by in vivo 31P MRS, different quantification methods were analyzed with regard to the accuracy and reproducibility of results and the simplicity of handling. Each quantification method comprised a calibration strategy and a localization technique. Extensive in vivo and in vitro tests showed that homonuclear phantom-based calibration strategies yielded significantly more accurate (lower systematic errors) and more reproducible (lower statistical errors) concentration estimates than heteronuclear strategies using internal water as a concentration standard. Additionally, the former strategies are easier to handle than the latter. Localization with the volume-selective sequence ISIS yielded slightly more reproducible results than localization by surface coil. We conclude that phosphorus metabolite concentrations are determined most accurately with phantom-based calibration strategies in combination with ISIS localization (measurement errors approximately 5-7%).

Magnetic radiofrequency fields applied in magnetic resonance imaging examinations induce electrical currents in metallic implants. These eddy currents may heat up the implants and thus may be capable of causing localized tissue heating. The rf power deposition and the joule heating of the implant can be calculated by solving Maxwell's equations for the specific problem. First, extreme in vitro worst-case experiments were performed with a large and very thin aluminum sheet, which was placed in a 1.5-T MRI device in a position parallel to the magnetic rf field. In agreement with the theoretical results the temperature rise of a thermally insulated sheet amounted to only 0.08 degrees C after a 15-min MRI examination at 64 MHz. No temperature rise in the aluminum sheet could be measured for a sheet immersed in a saline solution. Second, in vitro experiments with a hip joint prosthesis and an osteosynthetic plate were performed to confirm the theoretical results, which predict nearly no temperature rise in the metallic implants. No temperature rise in the implants could be measured.