RF safety assessment of a bilateral four‐channel transmit/receive 7 Tesla breast coil: SAR versus tissue temperature limits

Purpose: The purpose of this work was to perform an RF safety evaluation for a bilateral four‐channel transmit/receive breast coil and to determine the maximum permissible input power for which RF exposure of the subject stays within recommended limits. The safety evaluation was done based on SAR as well as on temperature simulations. In comparison to SAR, temperature is more directly correlated with tissue damage, which allows a more precise safety assessment. The temperature simulations were performed by applying three different blood perfusion models as well as two different ambient temperatures. The goal was to evaluate whether the SAR and temperature distributions correlate inside the human body and whether SAR or temperature is more conservative with respect to the limits specified by the IEC. Methods: A simulation model was constructed including coil housing and MR environment. Lumped elements and feed networks were modeled by a network co‐simulation. The model was validated by comparison of S‐parameters and B1+ maps obtained in an anatomical phantom. Three numerical body models were generated based on 3 Tesla MRI images to conform to the coil housing. SAR calculations were performed and the maximal permissible input power was calculated based on IEC guidelines. Temperature simulations were performed based on the Pennes bioheat equation with the power absorption from the RF simulations as heat source. The blood perfusion was modeled as constant to reflect impaired patients as well as with a linear and exponential temperature‐dependent increase to reflect two possible models for healthy subjects. Two ambient temperatures were considered to account for cooling effects from the environment. Results: The simulation model was validated with a mean deviation of 3% between measurement and simulation results. The highest 10 g‐averaged SAR was found in lung and muscle tissue on the right side of the upper torso. The maximum permissible input power was calculated to be 17 W. The temperature simulations showed that temperature maximums do not correlate well with the position of the SAR maximums in all considered cases. The body models with an exponential blood perfusion increase did not exceed the temperature limit when an RF power according to the SAR limit was applied; in this case, a higher input power level by up to 73% would be allowed. The models with a constant or linear perfusion exceeded the limit for the local temperature when the local SAR limit was adhered to and would require a decrease in the input power level by up to 62%. Conclusion: The maximum permissible input power was determined based on SAR simulations with three newly generated body models and compared with results from temperature simulations. While SAR calculations are state‐of‐the‐art and well defined as they are based on more or less well‐known material parameters, temperature simulations depend strongly on additional material, environmental and physiological parameters. The simulations demonstrated that more consideration needs be made by the MR community in defining the parameters for temperature simulations in order to apply temperature limits instead of SAR limits in the context of MR RF safety evaluations.