A contribution to MRI safety testing related to gradient-induced heating of medical devices

Abstract

Purpose: To theoretically investigate the feasibility of a novel procedure for testing the MRI gradient-induced heating of medical devices and translating the results into clinical practice.

Methods: The concept of index of stress is introduced by decoupling the time waveform characteristics of the gradient field signals from the field spatial distribution within an MRI scanner. This index is also extended to consider the anisotropy of complex bulky metallic implants. Merits and drawbacks of the proposed index of stress are investigated through virtual experiments. In particular, the values of the index of stress evaluated for realistic orthopedic implants placed within an ASTM phantom are compared with accurate heating simulations performed with 2 anatomic body models (a man and a woman) implanted through a virtual surgery procedure.

Results: The manipulation of the proposed index of stress allows to identify regions within the MRI bore where the implant could affect the safety of the examinations. Furthermore, the conducted analysis shows that the power dissipated into the implant by the induced eddy currents is a dosimetric quantity that estimates well the maximum temperature increase in the tissues surrounding the implant.

Conclusion: The results support the adoption of an anisotropic index of stress to regulate the gradient-induced heating of geometrically complex implants. They also pave the way for a laboratory characterization of the implants based on electrical measurements, rather than on thermal measurements. The next step will be to set up a standardized experimental procedure to evaluate the index of stress associated with an implant.

Keywords: MRI safety; gradient coil heating; medical devices; numerical simulation.

FIGURE 1

Spatial distribution in the plane y = 0 of the magnetic field components produced by tubular gradient coils with a unit gradient (1 T/m) around the isocenter. The first row depicts the x components, the second row the y components and the third row the z components

FIGURE 2

Waveforms of the gradients produced by each gradient coil during 1 repetition time of a 3D FISP or an EPI‐X pulse sequence with the frequency encoding signal associated to the coil X (A,B) and their time derivatives (C,D)

FIGURE 3

Sketch of the procedure for the evaluation of the weighting coefficient associated with a generic direction of the magnetic field starting from a limited number of measured samples. The set of the admissible directions coincides with a hemisphere, where each direction is identified by a segment OP from the center of the hemisphere (O) to a point on the hemisphere (P). The hemisphere is described in spherical coordinates, in which the latitude is the angle between OP and its projection on the plane z = 0, whereas the longitude is the angle between the projection of OP in the plane z = 0 and the x‐axis. The reported weighting coefficient is the one obtained for the hip implant of Yoon‐Sun model with total power as dosimetric quantity

FIGURE 4

Schematic summary of the proposed procedure. The input data are in bold

FIGURE 5

Weighting coefficient distributions in spherical coordinates for each considered implant (hip, knee, and shoulder of Yoon‐Sun model) computed with 65 virtual experiments with reference to different dosimetric quantities: peak power density (p _max), total power (P), temperature increase after 360 s (ΔT _max,360s), and after 900 s (ΔT _max,900s). For the latter, the direction with the maximum coefficient is highlighted by a red circle and is represented by arrows in the 3D representation

FIGURE 6

Scatter plots of ΔT _max,900s versus P/S (ratio between the total power deposited inside the implant and the external surface of the implant itself) for Yoon‐Sun model with each considered implant (first column), compared with the corresponding cases in phantom (second column). The color is representative of the considered sequence: EPI‐X (red), EPI‐Y (blue), EPI‐Z (green), 3D FISP (yellow). The results obtained combining all implants together are reported in the last row. The linear fits (green lines) are depicted together with lower (cyan) and upper (magenta) lines that include 95% of data

FIGURE 7

In the first row, spatial distribution in the plane y = 0 of the index of stress associated to the heating induced by the magnetic field generated by a tubular gradient coil system during the application of a 3D FISP pulse sequence. The isotropic index (Equation [3]) and the anisotropic index (Equation [4]) for 3 implants with weighting coefficient estimated from the peak temperature increase after 900 s are reported. In the second row, for each implant, the correlation between the anisotropic index of stress and the peak temperature increase after 900 s is shown by evaluating them when the implant is located in the position denoted by the white circles in the color maps. The considered implants are those implanted in the Yoon‐Sun model

FIGURE 8

The colored areas represent the regions of exclusion for 3 implants during a 3D FISP (first row) and an EPI‐X (second row) pulse sequence executed in a tubular gradient coil system. These correspond to the positions where the implants should not be located during the exam to avoid a temperature increase, induced by the gradient fields, larger than 1 K after 900 s. For each implant, the regions are computed based on both the isotropic index of stress (Equation [3]) and the anisotropic index (Equation [4]), with weighting coefficients estimated from different dosimetric quantities. The uncertain orientation of the implant is also accounted. The considered implants are those implanted in the Yoon‐Sun model