Radio-frequency (RF) safety risks associated with thermal damage in patients are a significant problem for ultra high field(UHF) MR applications. In order to predict the temperature increase in a patient during an MR scan, we need validated electromagnetic and thermal simulation tools and realistic anatomical models. In order to reach our goal, we propose to conduct RF heating experiments and measure temperature in anesthetized pigs. We simulated and measured the heating due to RF excitation with a single dipole antenna in an anesthetized pig using temperature probes. By initially validating our simulations with a single dipole antenna we feel more confident to use multiple coil combinations, which would finally pave the way for accurate temperature prediction in transmit array applications.

We used a fractionated dipole antenna [1] for RF excitation in our experiments. Dipoles have simple structures that are easy to model and build. In addition, they have complementary field patterns in comparison to loops (2,3) therefore they can be used to generate a wide variety of excitation profiles to validate EM & thermal simulations. Simulations as well as RF heating experiments were conducted by using an anesthetized pig (Male, 70 kg). Pig was first injected with Telazol (5-10 mg/kg)+Xylazine(1-3 mg/kg). Anesthesia was maintained by providing isoflurane (~1.5-2.5%) in 50% air and 50% oxygen mixture. After preparation, the pig was transferred to the RF safety lab in CMRR where the RF heating experiment was conducted. Anesthesia & vital monitoring continued through the RF heating experiment until euthanasia. Communications Power Corporation (Hauppauge, NY) 2 kW A-B Linear broadband RF amplifiers were used to generate 10% duty cycle hard pulses. The output of the amplifiers (50 dB attenuated) was observed on an oscilloscope screen which enabled precise calibration of the amplitude and phase levels. A fractionated dipole antenna (1) was used to deliver RF energy to the pig’s neck region. 4 fiber optic thermal probes were used for temperature monitoring. CT visible markers were placed on the pig to mark the location of the coil. After the RF heating experiment, euthanasia was performed by delivering KCL intravenously. CT scan of ex-vivo pig (with markers and the sensors in place) was then performed to obtain detailed anatomical images. The acquired images were used to construct a digital pig model as shown in Figure 1. Finally skin, inner air, bone, muscle and fat were segmented. Appropriate EM/thermal properties were assigned to the tissues and simulations were performed. A commercial FDTD solver, SEMCAD (Zurich, Switzerland) was used to solve for both the EM field and the temperature distribution. Temperature obtained as a result of a 15 min RF exposure was calculated. In addition,time variation of the temperature sampled at the 4 probe locations in the model was compared to the experimentally measured temperature data.

Figure 1 shows the segmented pig model that are used in EM simulations. The model includes muscle, fat, bone and skin tissues as well as inner air. In addition, it includes the CT visible markers associated with the dipole's position and the thermal probes that are used for temperature measurements. Figure 2 shows the SAR distribution obtained by a dipole antenna for a power level of 20 W(power delivered to the dipole). The SAR is distribution is clearly modified by the conductivity of the tissue distribution as it can be seen from the figure. Figure 3 shows the temperature distribution obtained by the same antenna. The temperature distribution strictly depends on thermal properties of the tissues therefore it does not fully correlate with the SAR distribution. Finally the variation of the temperature(in time) at 4 probe locations are shown in Figure 4. The comparison between the experimental and simulated values can be made from the same figure.

The simulated and measured temperatures were in quantitative agreement in this work. Our future plan is to experiment with more number of channels and temperature probes for validation. In addition, performing simulations with anatomical models obtained with MR segmentation methods (prior to the RF exposure) is also planned. The goal of this work was not to make definitive comments about the RF safety of dipole antennas. We excited the fractionated dipole with 20 W of RMS power which resulted in SAR levels that are higher than conventional limits. By doing so we were able to acquire faster and cleaner temperature readings. For future animal experiments we plan to validate our simulations with multiple coil elements where faster heating and cooling of the pig tissue will enable us to use multiple RF shimming patterns consecutively.