Introduction

3T whole-body birdcage coils are designed to generate a homogeneous B₁₊ field while simultaneously maintaining the RF safety limits inside the patient. The electric (E) field inside the patient is difficult to assess but is not a limiting factor for volume transmit coils. SAR averaged over the mass of the body or specific body parts and over a specified time is the primary RF safety limit for these scanners¹. MR examinations are increasingly being performed on patients with medical devices. Full 3D-EM simulations are used for the evaluation of RF safety limits. Usual domain decomposition consists of separating the human EM exposure, due to the MR RF transmit coil, from the assessment of the medical device RF response. The former could be accomplished by collecting EM field exposure data obtained in multiple anatomically correct patient models at different landmark positions and in a variety of MR whole-body birdcage coils². Based on the limited availability of MR coil design data, generic coil models are typically developed to evaluate EM field exposure in human models. Attempts to validate E-field exposure of commercial whole-body birdcage coils using a generic coil model had limited success^3,4. The commercial library MRIxViP3.0T⁵ (IT’IS Foundation, Zurich, Switzerland) provides EM exposure distributions inside human body models. However, MRIxViP3.0T provides limited validation information that the library data covers EM exposure distribution cases occurring in commercial MR scanners, including scanners with two-channel parallel transmission.

Methods

“Generic” and “vendor-specific” coil models were constructed using the same coil geometry but with different azimuthal placement of the excitation ports: (i) 270/0 degrees, (ii) 135/225 degrees, and (iii) as in Siemens commercial coils (exact location cannot be disclosed). Two widely used whole-body birdcage coils from the MAGNETOM Verio (70cm bore) and MAGNETOM (60cm bore) systems (Siemens Healthcare GmbH, Erlangen, Germany) were selected for this investigation. The full information of simulation workflow for “vendor-specific” and “generic” coil models are presented in⁶. Four “generic” coil models were built using the geometry of MRIxViP3.0T coils B75_L40, B75_L50, B65_L40, B65_L50 described in⁷. Quadrature excitation was applied. “Generic” coils were also excited at all 32 ports (so-called G32 design) as in⁸. As a proof of concept, the results for an oval phantom were obtained based on a co-simulation approach⁹.

Results and Discussion

The qualitative comparison of different MR scanner and RF safety related quantities revealed that: (i) B₁₊ profiles (Figs. 2 and 4) were very similar for most investigated cases except in the case of the Siemens “realistically tuned” coil, i.e., the coil with minor deviation from the ideal S parameters. Unexpectedly, this case results in the best B₁₊ homogeneity over the center axis slice. (ii) SAR and E_rms profiles (Figs. 2-5) were very similar for most cases except Siemens “vendor-specific” coil models. The largest discrepancy relative to the “average” SAR profile was observed for the Siemens “realistically tuned” coil. (iii) E_rms profile in air and in the oval phantom for XZ plane, 75mm below the iso-center, which is substantially different between the “vendor-specific” and the “generic” coil models. Changing the “generic” coil length did not make this E_rms profile to become more similar to the “vendor-specific” profile in air. A mismatch up to 20dB was observed at some positions in close proximity to phantom external face.
For the coils with 70cm bores, the negligible difference in SAR profiles can be explained by very similar current distribution in the rings and rungs. For coils with 60cm bores, the coil load without cylindrical symmetry is the reason for the variation of the coil current distributions. For the G32 design, the distribution of the current amplitudes in the rungs was still close to the “ideal” case, whereas other coil designs exhibited larger deviations. Even more substantial variation, i.e. a mismatch up to 5dB at some location inside the phantom, was observed for “vendor-specific” port locations, especially for the 60cm bore coil. In all cases, the variation increased if Siemens Trueform mode was applied, i.e., the excitation with the same amplitude, but the phase other than 90 degrees¹⁰.
A major goal of MR scanners is to generate the most suitable B₁₊ profile in a patient while keeping acceptable levels of SAR (head, whole body, partial body; depending on the patient’s land mark position). The spatial distribution of the E-field inside or outside the patient’s body for volume transmit coils is not regulated by IEC 60601-2-33 or any other RF safety related standard. This E-field distribution, however, is of great importance for the safety assessment of medical devices.
A quantitative comparison of coil design consequences for given AIMDs, passive implants, etc. requires simulation of a set of “vendor-specific” coil models with human models, preferable as in MRIxViP3.0T. This is a large task which may require collaborative projects, especially when also including RF shimming, i.e., excitation of the RF coil ports with different amplitude and phases

Conclusion

Our investigation provided evidence that (i) quadrature excitation in “vendor-specific” coil models generate EM field profiles substantially different to “generic” coil models, especially in case of “realistically tuned” coils, (ii) the location of the excitation ports impacts the prediction of EM field profiles, especially for non-quadrature excitation mode.

Acknowledgements

DISCLAIMER: The mention of commercial products, their sources, or their use in connection with material reported herein is not to be construed as either an actual or suggested endorsement of such products by the Department of Health and Human Services.