Magnetic resonance imaging (MRI) examinations of patients with implanted medical devices comprises health risks. In particular, radiofrequency (RF) induced heating due to coupling of RF fields with the metallic parts of the implant could occur in the tissues surrounding the electrodes. To overcome this hazard, only patients implanted with devices labeled by regulatory bodies as “MRI safe” should be scanned. The safety of implants in an MRI environment is assessed by a combination of experimental and numerical techniques¹. The latter include the evaluation of the fields induced in patients exposed to RF fields during MRI examinations. The absorption of RF fields by the human body during an MRI examination depends on many factors, e.g., the frequency, coil dimensions, coil polarization, coil topology, and positioning landmarks of the body, among others. A dosimetric study of the in vivo RF absorption due to MRI exposure must cover the parameters of existing clinical scanners.The objective of this study was to develop a minimal set of numerical models of birdcages that comprehensively represent all the commercial closed-bore systems with respect to RF exposure. Simulations were performed on a subset of five birdcage resonators with high pass topology, bore diameters of 600, 700, and 800 mm and rung lengths ranging from 350 to 700 mm. All birdcages were surrounded by 850 mm long cylindrical shields. The human exposure envelope was defined for a set of anatomical human models that represent the patient population. The models were positioned at a series of clinically relevant landmarks that correspond to scanning positions for imaging targets ranging from the head to the extremities². The variability and uncertainty were comprehensively assessed. All birdcages were driven in quadrature at 64 MHz. Finite-difference time-domain simulations and post-processing were performed with Sim4Life Version 1.2 (ZMT, Zurich, Switzerland) and SEMCAD X Version 14.8 (SPEAG, Zurich, Switzerland).

All quantities reported in this section are for a B₁-field of 1 μT at the isocenter of the coils. Only a subset of the results are shown here. Fig 1 shows the E-field distributions in coronal (top) and axial (bottom) slices of the Ella model positioned for thoracic imaging. Exposures in birdcages with a fixed diameter (D) of 750 mm and lengths of 350, 500, and 700 mm (Fig 1-top, marked with white rectangles for the coronal views) or with a fixed length (L) of 500 mm and diameters of 650, 750, and 850 mm (Fig 1-bottom, marked with white circles for the axial views). The coronal E-field distributions in Fig 1-top show how the E-field level increases as the birdcage length grows for a fixed coil diameter. In Fig 2, the coronal E-field distributions in Fig 1 plotted against the average E-field exposure for the five coils show the correlation of coil length and field strength. The shorter coils produce E-field below the average, whereas the E-field for the longer coil are in general above the average.

In the context of implant safety, the incident fields at possible locations of a pacemaker were obtained to determine whether a reduced set of birdcages is valid for demonstration of the MRI safety of implants^{3, 4, 5}. Sets of realistic clinical trajectories representing the possible positions of the devices were generated for all models; Fig 3-left shows examples of cardiac trajectories in Ella. For the assessment of exposures in the patient population, the tangential E-fields for the representative trajectories were extracted for each human model/landmark/birdcage combination. The average E-field tangential to 300 trajectories was plotted for each landmark group and for each of the five birdcages (Fig 3-right shows the tangential E-fields in Ella). For cardiac exposure, the head and thorax landmarks yield the highest E-field values. For a fixed landmark, however, the range of the obtained values is large depending on the birdcage the body is exposed to.

The envelope for in vivo MRI exposure to a set of RF birdcages that cover the variation existing in clinical scanners has been investigated. For this purpose, a set of simulations were performed with permutations of human model, birdcage model, and landmark. The preliminary assessment shows large variations in the tangential E-field values for the different coils. From the large variations observed, we conclude that further investigations are needed to derive a minimal set of birdcages that can be used for safety assessment of implants; and to determine the impact of different implant types (e.g. cardiac devices, neurostimulation devices, etc.). Future investigations need also to include variation of B₁ polarization as a parameter in the simulation matrix.