Simulation of B1 field distribution and intrinsic signal-to-noise in cardiac MRI as a function of static magnetic field.

Two issues that pertain to the optimal static magnetic field for cardiac MRI were addressed: intrinsic signal-to-noise ratio (ISNR) and radiofrequency power deposition. From 1.5 to 9.5 T, proton Larmor frequencies of 63 to 400 MHz, numerical simulations were performed of the RF fields from a surface coil and a body coil loaded by a heterogeneous, three-dimensional, symmetric model of the human chest. The RF field distribution, the power required to produce the RF field, and the ISNR at the center of the heart were computed. The model was validated by comparison with experimental data up to 4 T. The RF field distortion was quantified and found to increase linearly up to 6 T due mostly to dielectric resonance modes. Body coil simulations beyond 6 T showed the onset of higher-order modes at the center of the heart. A range of expected RF power requirements was constructed as a function of field up to 9.5 T for surface coils and up to 6.8 T for body coils. Over this range of static field, ISNR for a constant coil geometry was bracketed by an upper limit that was slightly greater than linear with field and a lower limit that was slightly less than linear with field. The RF power and ISNR showed a strong dependence on chest thickness at 1.5 and 4.0 T. Additionally, independent of chest thickness, the model predicts a lower limit of a factor of 5 increase in RF power as the static field is increased from 1.5 to 4 T. Implications for imaging with other nuclei are discussed. Methods for checking the self-consistency of electrodynamic simulations are presented.