Thermometry by measuring the chemical shift of lanthanide complex

BACKGROUND: In the long-term, non-invasive thermometry is vital for the continued clinical and technological development of regional hyperthermia. In magnetic resonance tomography. T1 relaxation time, diffusion and proton resonance frequency are used to measure temperature distributions. When used clinically in the pelvic region, all of these methods are plagued with errors and artefacts on account of the tissue relationships, tissue changes under hyperthermia, physiological and stochastic movements, inhomogeneities, drift phenomena and instabilities. MATERIAL AND METHOD: We tested the relationship between the temperature and the chemical shift of a methyl group of a lanthanide complex with central atom praseodymium (Pr-MOE-DO3A. Schering AG). To do this we used cylindrical phantoms containing a 5-mmol-solution of this temperature-sensitive substance. High resolution spectra and relaxation times were determined in a Bruker AMX at 11.5 T. A calibration curve was then recorded by a Siemens Magnetom SP63 at 1.5 T. Local temperature distributions were determined using the chemical shift imaging method, with a matrix size of 16 x 8 and a narrow-band excitation pulse. The temperature distribution was created using a Nd:YAG laser applicator. RESULTS: At a distance of -25.7 ppm from the water line, we found a singlet line with a temperature-dependent chemical shift of 0.13 ppm/C. In the phantom experiment we found that the chemical shift had a linear relationship with a gradient independent of the surroundings, and a temperature resolution of +/-0.6 degree C. With a concentration of 1 mmol/l, a matrix size of 8 x 8 and a measurement period of 5 s per acquisition, phantom measurements using the CSI method produced a signal to noise ratio of 3.5 per acquisition, i.e a measurement period of 10 to 20 s per spectrum. CONCLUSIONS: Our in vitro data show that spectroscopic temperature measurement using a temperature-sensitive praseodymium complex with a therapeutically practical concentration of 1 mmol/l already appears to be suitable for clinical use Compared with the methods tested so far (T1, diffusion, proton resonance), this method has the special advantage of not being very susceptible to artefacts. The competing methods of non-invasive thermometry using magnetic resonance tomography/spectroscopy will be investigated next.