A comparison of local SAR using individual patient data and a patient template

INTRODUCTION: Higher field strengths and new multi-coil-transmission methods require special attention regarding the patient safety. An often used approach to determine SAR is electromagnetic field simulation based on the Finite-Difference Time-Domain method (FDTD). This method requires accurate modelling of transmitting elements and the object itself. Due to long computation times, field calculations prior to each scan are hardly feasible and instead the storage of field data for different patient constitutions and body regions is required. Recently an alternative method has been proposed that uses measured B1 data to estimate SAR and electric properties of the object [1]. However, if the distribution of the electric properties is not exactly known for the individual patient, errors are introduced to the SAR prediction. Particularly the fat distribution can differ strongly from patient to patient. As has been reported in the literature, SAR peaks can occur at tissue-fat and tissue-bone interfaces due to their high dielectric contrast [2] [3]. The present work therefore compares the effect of incorrectly assumed patient data in two methods: The standard FDTD method and the method of solving Ampère’s law for time harmonic fields (Equ.1) using B1 fields, which will in the following be named Finite-Difference Frequency-Domain method (FDFD). METHODS: Using the FDTD method (SEMCAD X, Schmid & Partner Engineering AG, Switzerland) electromagnetic fields were calculated for a shielded strip line driven at its resonance frequency of 125MHz. Simulations were performed for two adult phantoms of the Virtual Family Project [4] with an isotropic resolution of 2mm. The phantoms were scaled to make them comparably sized and positioned with the liver above the centre of the coil. The first phantom is in the following treated as the individual patient, the second one serves as a patient template (Fig.1). Then the simulated transversal components of the magnetic field in the individual patient were used to calculate E by solving Equ.1. In Equ.1 uv E and μμ = uuv uv