Researchers involved in RF safety, high-field and/or X-nuclei MRI

It has already been shown that dielectric pads with high permittivity can be used to improve transmit/receive sensitivities and, consequently, image quality in brain imaging¹. Besides these B1 effects, also effects on the specific absorption rate (SAR) have been discussed², and it has been shown that specific configurations of high-permittivity pads can lead to an increase in local SAR of 92% as well as a displacement of the 10g local maximum³. Thus far, only the RF frequency of ¹H at 7 Tesla has been investigated.

Higher static field strength and improvements in coil and sequence design have increased the clinical interest of X-nuclei in MRI, although the spatial resolution is still low compared to ¹H imaging. Examinations of X-nuclei are typically combined with ¹H sequences. The additional information obtained from high-resolution ¹H images can be used as an anatomical reference, for image post-processing steps, e.g. partial volume correction⁴, or for the iterative reconstruction of ²³Na data using an anatomical prior⁵.

Dual-tuned coils or coil combinations, e.g. for ¹H (297.15 MHz) and ²³Na (78.6 MHz), provide images of both nuclei without the disadvantage of moving the patient or changing the measurement setup. For such configurations, high-permittivity pads used to optimize B₁⁺ for ¹H imaging will also be in place during X-nuclei imaging. In this work, we investigate the effects of high-permittivity pads on RF fields during sodium (²³Na) imaging at 7T.

To account for different head sizes and tissue distributions, simulations were performed with a male (Duke, 34y, height 1.77 m, body mass 72.4 kg) and a female body model (Ella, 26y, 1.63 m, 58.7 kg)⁶. Dielectric pads were modelled with dimensions of 195x110x10 mm³ and dielectric parameters of ε_r’ = 110 as well as 500, and σ = 0.0918 S/m at 78.6 MHz. One pad was placed at the back of the head with its longitudinal axis parallel to the body axis. A volume transmit head coil^3,7, modelled as a shielded bandpass birdcage with 16 rungs, was used for excitation. The coil, originally designed for 297.15 MHz, was tuned to 78.6 MHz. Field results were normalized to an input power of 1 W and simulations were performed with CST Studio Suite 2015⁸. The maximum 10g-averaged specific absorption rates (SAR_10g,max) were determined, and SAR elevations were compared for configurations with and without pads. Further, the effect of the pads on the spin excitation B₁⁺ field was evaluated in the cerebellum. Additionally, configurations with a doubled pad thickness of 20 mm were considered.

Fig. 1 shows the tissue distribution in the head models and the SAR_10g,max distribution in the central sagittal plane. Both body models show an increased local SAR located in muscle/fat tissue of the neck at the inferior, shorter edge of the pad. The configuration with ε_r’ = 500 and thickness of 10 mm turned out to be the most critical case: here, the 10g-averaged SAR was increased by 58% (Ella) and 81% (Duke). For the configurations without pad and with a pad with ε_r’ = 110 and 10 mm thickness, SAR_10g,max was located in the left frontal lobe, whereas in all other configurations SAR_10g,max was displaced and located at the inferior, shorter edge of the pad. Results are summarized in Table 1.

B₁⁺ distribution for the considered pad configurations are shown in Fig. 2. The almost homogeneous field of the birdcage was altered by the pads and focused in the posterior part of the brain close to the pads. Higher permittivity and pad thickness enhanced this effect. The increased B₁⁺ extends through half of the head. However, evaluation in the cerebellum did not show a change in mean or max B₁⁺ for any pad configuration. Also, configurations with 3 pads distributed around the head^1,2 (not shown) did not show an improvement in the cerebellum.

None of the dielectric pad configurations considered in the simulations showed an improvement in B₁⁺ in the cerebellum for ²³Na imaging. It is possible, however, that other configurations might achieve this. Certain combinations of pad position, geometry, and material properties can lead to significant SAR elevations. These results are reproducible for multiple body models. Results for SAR_10g,max are similar to previous findings for ¹H at 7T. Since the polarization of the incident field³ and the geometry of the pads are important factors, detailed SAR analysis for any possible configuration is necessary at all used frequencies. If increased RF exposure of the volunteer is found, the maximum permissible input power of the RF coil must be decreased.