Electromagnetic and Thermal Aspects of Radiofrequency Field Propagation in Ultra-High Field MRI

Abstract

In MRI, a radiofrequency (RF) pulse is used to generate a signal from the spins that are polarized by a strong magnetic field. For higher magnetic field strengths, a higher frequency of the RF pulse is required in order to match the Larmor frequency. A higher frequency, in turn, leads to a shorter wavelength. This results in undesirable spatial fluctuations of the RF magnetic field. Those fluctuations are caused by interference and lead to reduced image quality. Not only the magnetic field, but also the electric field is inhomogeneous. Part of the electric field is absorbed which results in tissue heating. Due to interference, hotspots are expected in the energy absorption patterns, as well as the tissue temperature distribution. Tissue heating can be harmful in case the reached temperature is too high. In the first part of the thesis, the energy absorption and resulting tissue temperature in the head during 7T MRI was investigated using simulations. It was shown that the current safety guidelines regarding the maximal allowable energy absorption closely match the maximum allowable temperature. Therefore patient-to-patient variations in anatomy and physiology, can lead to a higher temperature than expected, thus lead to a risk for the patients. Therefore, the remainder of the work was focused on mapping of one of the input variables: the dielectric properties. It was shown that the dielectric properties can be mapped with MRI using Electrical Properties Tomography (EPT). Paradoxically, the inhomogeneous RF field is in this case very useful; the local dielectric properties are encoded in the propagation pattern of the RF field, as they affect the local wave behavior. Therefore, it seems beneficial to measure with a high RF frequency, thus at a high static magnetic field strength. This was evaluated using a comparison study at 1.5, 3 and 7T of EPT of the brain. It was found that measuring at a higher static magnetic field strength indeed leads to a higher sensitivity of EPT. However, certain assumptions which are needed for EPT, are less valid at higher field strengths. Therefore it was concluded that for the investigated case, 3T was the best field strength for conductivity mapping. For permittivity mapping, however, it was found that the sensitivity at 3T was much lower, and therefore, 7T was the optimal field strength for this case. Mapping of the dielectric properties is not only useful for RF safety assessment of MRI, but can also be used to image tumors. In the thesis it was shown that in brain tumors the electrical conductivity is elevated compared to the healthy surrounding tissue. This finding was consistent with earlier work on conductivity measurements, however, in that case invasive measurement techniques were used. One remaining challenge for EPT, is to reduce the errors which originate from two presently used assumptions, such that the accuracy of the reconstruction improves and the maps can be used as input for simulations. Clinically, more insight should be gained on the physiological changes that are the basis for the elevated conductivity.