The ultimate signal‐to‐noise ratio in realistic body models

Abstract

Purpose We compute the ultimate signal-to-noise ratio (uSNR) and G-factor (uGF) in a realistic head model from 0.5 to 21 Tesla. Methods We excite the head model and a uniform sphere with a large number of electric and magnetic dipoles placed at 3 cm from the object. The resulting electromagnetic fields are computed using an ultrafast volume integral solver, which are used as basis functions for the uSNR and uGF computations. Results Our generalized uSNR calculation shows good convergence in the sphere and the head and is in close agreement with the dyadic Green’s function approach in the uniform sphere. In both models, the uSNR versus B0 trend was linear at shallow depths and supralinear at deeper locations. At equivalent positions, the rate of increase of the uSNR with B0 was greater in the sphere than in the head model. The uGFs were lower in the realistic head than in the sphere for acceleration in the anterior-posterior direction, but similar for the left-right direction. Conclusion The uSNR and uGFs are computable in nonuniform body models and provide fundamental performance limits for human imaging with close-fitting MRI array coils.

Computation of ultimate SAR amplification factors for radiofrequency hyperthermia in non-uniform body models: impact of frequency and tumour location

“…Computation of basis-sets with K = 2500 fields for all 10 frequencies in Duke took a combined ~22 d. The longest computation time was for 894.1 MHz (~5 d), since the convergence of the MARIE solver slows at higher frequencies [32,33]. …”

“…For the generation of the EM basis-set at a given frequency, we followed the strategy described by Guérin et al [32]. The approach is in three steps (Figure 1).…”

“…As explained by Guérin et al [32], this is done in order to quickly generate a basis-set of EM fields that almost spans the space of solutions to Maxwell equations in the body model (the term “almost” is because that space in infinite, therefore, some truncation is required. However, our basis-set is convergent, i.e.…”

“…Since storing the spatial field patterns of all the dipoles would require a large amount of memory, a SVD compression of all the dipole fields would be computationally laborious. Instead, we excite all the dipoles simultaneously using random amplitudes and phases, which allow quickly constructing a basis-set of EM fields in the body model without having to compute all the dipole fields or hold them in memory [32]. …”

“…In this work, we used K = 2500 since this number was sufficient to reach convergence in our previous calculation of the ultimate signal-to-noise ratio in MRI [32]. We test that K = 2500 is a sufficient number of basis fields to achieve convergence in the estimation of the uSAF for all tumours and all frequencies in Figure 2 and Supplementary Figures 4 and 5.…”