Experimental estimation of local SAR in a multi-transmit system

Introduction: SAR management is a widely discussed topic at main fields of 3T or above, particularly in the framework of multi-transmit systems [1,2]. Typically, local SAR is estimated via patient model simulations, i.e., extensive numerical calculations not suitable for clinical routine. An alternative way to estimate local SAR might be given by “Electric Properties Tomography” (EPT), i.e., the analysis of the RF coil's spatial transmit and receive sensitivity distribution [3]. However, this approach assumes a dominant positive circular magnetic field component H >> H of the RF coil used, which is usually only fulfilled for birdcage-type coils in the quadrature mode. For multi-transmit systems, this assumption can be violated, particularly while performing RF shimming. RF shimming, i.e., the adjustment of amplitudes and phases of the different transmit channels to obtain a homogeneous total B1 field, makes use of all available birdcage modes, and thus, yields H in the same order of magnitude as H. This study investigates the possibility to circumvent this problem by a suitable handling of the acquired sensitivities, potentially opening a short and easy way of local SAR determination in multi-transmit systems. Theory: Electric conductivity σ and permittivity ε can be estimated via Eq. (1), a combination of Ampere's and Faraday's law, with ω the Larmor frequency, Axz a coronal integration area and ∂Axz the boundary of this area [3]. Equation (1) holds inside compartments with constant ε and σ and assumes Hz << Hx,Hy (usually fulfilled for birdcage-type RF coils). After solving Eq. (1), the electric fields can be derived from Ampere’s law (2). Finally, results from Eqs. (1,2) are used to calculate local SAR (3). Cartesian Hx, Hy and circular field components H, H are related via (4). The positive circular components + n H of the n = 1...N different