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

Due to dielectric losses in body tissue, radio-frequency (RF) energy is absorbed and converted into heat. Heat can increase tissue temperature, and high temperatures subsequently result in tissue damage. The IEC safety standard1 for medical equipment defines limits for local tissue temperature and specific absorption rate (SAR) for the safe use of an MR scanner. According to the IEC standard, compliance with temperature limits may be achieved by limiting the SAR. In the majority of cases, RF safety assessment of transmit coils is based on SAR, whose determination is well defined in the standard.

SAR depends on the body physique (tissue distribution) and on the Tx coil design and MR environment (geometry). As temperature is directly correlated with tissue damage whereas SAR is not, the evaluation of tissue temperature allows a more precise safety assessment of RF coils. Tissue temperature, however, depends on the thermal tissue properties (conductivity, specific heat), ambient temperature, heat convection on body surface (ventilation and clothing), metabolism, and blood perfusion. The SAR influences the temperate as an RF-induced heat source.

Individual health status affects the thermoregulatory system and can be taken into account by different temperature-dependent blood perfusion models proposed in the literature but which are not specified in safety standards.

In this study, the safety of an RF coil was assessed based on both SAR and temperature limits by numerical simulations. The correlation between SAR and temperature were analyzed with respect to spatial distribution and maximum permissible input power.

Numerical calculations were performed for a coil model of a 4-ch Tx/Rx bilateral breast coil for 7 T² (Fig. 1a). A new body model conforming to the coil housing was generated from 3 Tesla data (1.2 mm tissue resolution, 17 different tissues). The maximum permissible input power was determined for the normal (NM) and first-level (FLM) controlled operating modes (10g-averaged SAR limits of 10 W/kg and 20 W/kg, respectively). Subsequently, temperature simulations were performed with the Pennes bio heat equation³ (Fig. 2). Different health states were considered via temperature-dependent blood perfusion models: A constant blood perfusion was assumed for an impaired thermoregulation system, e.g. due to diabetes, whereas a linear increase⁴ and an exponential increase⁵ in blood perfusion were assumed as possible models for a healthy thermoregulatory system, Fig. 3. An ambient temperature of 25 °C, a blood temperature of 37 °C, and a heat convection coefficient of 2 W/(m²K) were used.

Thermal simulations were performed based on the two power levels, and resulting tissue temperature distributions were analyzed. In addition, the maximum permissible input power was determined for temperature limits in NM (39 °C) and FLM (40 °C). Simulations were performed with CST Studio Suite 2015 (CST AG, Darmstadt, Germany).

Maximum 10g-averaged SAR was located inside the left mamma, where the prevailing tissue is fat, Fig. 1b². Maximum permissible time-averaged input power was 36.5 W (NM) and 73 W (FLM). For the constant perfusion model, the corresponding temperature limits were exceeded by 2.28 °C (NM) and 6.17 °C (FLM), Table 1. To satisfy the SAR and the temperature limit, the input power would have to be decreased by 46% (NM) and 62% (FLM). For the linear perfusion model, the maximum tissue temperatures were lower compared to the constant model but also exceeded the temperature limits in NM and FLM. For the exponential perfusion model, the tissue temperature stays below the limits in both operating modes. Consequently, for a healthy subject, the input power level can be increased by 37% (NM) or 40% (FLM) when applying temperature limits rather than SAR limits. For all simulations, the SAR and temperature distributions correlated well for the configuration considered here, Fig. 4, which is a consequence of the low perfusion in fat tissue. This finding is not expected for tissues with higher blood perfusion⁶.

The proposed perfusion models yield a large variation in tissue temperature. So far, there is no general agreement on which perfusion models yield realistic tissue temperature for impaired or healthy subjects. The results indicate that a validation of the thermal models is needed and that different thermal models for different health status may be required.

The results indicate that it is possible to exceed the IEC local temperature limit while satisfying the IEC limits for local SAR when a patient with impaired thermoregulatory system is examined. As a consequence, the maximum permitted input power needs to be decreased if temperature is considered to be the fundamental limit most closely related to possible tissue damage.