A single setup approach for the MRI-based measurement and validation of the transfer function of elongated medical implants

Abstract

Purpose: To propose a single setup using the MRI to both measure and validate the transfer function (TF) of linear implants. Conventionally, the TF of an implant is measured in one bench setup and validated using another.

Methods: It has been shown that the TF can be measured using MRI. To validate this measurement, the implant is exposed to different incident electric fields, while the temperature increase at the tip is monitored. For a good validation, the incident electric fields that the implant is exposed to should be orthogonal. We perform a simulation study on six different methods that change the incident electric field. Afterward, a TF measurement and validation study using the best method from the simulations is performed. This is done with fiberoptic temperature probes at 1.5 T for four linear implant structures using the proposed single setup.

Results: The simulation study showed that positioning local transmit coils at different locations along the lead trajectory has a similar validation quality compared with changing the implant trajectory (ie, the conventional validation method). For the validation study that was performed, an R² ≥ 0.91 was found for the four investigated leads.

Conclusion: A single setup to both measure and validate the transfer function using local transmit coils has been shown to work. The benefits of using the proposed validation method are that there is only one setup required instead of two and the implant trajectory is not varied; therefore, the relative distance between the leap tip and the temperature probe is constant.

Keywords: measurement; simulation; transfer function; validation.

FIGURE 1

All of the simulation setups that were used to investigate the methodologies to change the incident tangential electric field. A, The phantom was shifted out of the birdcage coil. B1,B2, The side and top view of the dielectric pads. The electric field is extracted at the location of the light green trajectory; this location within the phantom is used for all methodologies except changing the trajectory. C1,C2, Top and side view of the passive RF coil positions. D, The local transmit coil positions. E, The 100 random implant trajectories that were extracted. F, Example of how the phantom was wrapped in aluminum foil

FIGURE 2

The ends of the implants for which the transfer functions (TFs) were measured and validated. A,B, Both ends of the bare copper wire. C,D, Both ends of the insulated copper wire where the ends of the insulation are removed for about 1 cm. E,F, The ends of the coaxial lead and a proper lead tip. G,H, The Implantable Pulse Generator and the electrode patch of the spinal cord stimulator, respectively

FIGURE 3

The measurement setup. A, The positioning of the temperature probe with respect to the lead tip. B, The temperature measurement device. C, How the probes are entered through the RF waveguide into the MRI room. D, Using the ruler, we measured the distance from the end of the ASTM phantom toward the place of the local transmit coil when it is placed underneath the phantom

FIGURE 4

Overview of the uncertainty analysis. A, The curve fit with uncertainty of a single temperature measurement; this fit is used to find the measured specific absorption rate (SAR). B, Transverse slice of the fast field‐echo (FFE) sequence, where the artifact created by the lead is visible. C, The steps in the Monte Carlo simulation approach find the uncertainty of the measured TF, and thereby the uncertainty in the predicted SAR. Step 1 is used to sample the noise distribution and that noise to the original FFE series to create a new FFE series. Step 2 is used to fit the $\left|{\mathrm{B}}_1^{+}\right|$. Step 3 is used to fit the current from the $\left|{\mathrm{B}}_1^{+}\right|$. In step 4 we fit the TF using the current and the simulated E_tan . The blue‐shaded area shows the SD of the measured TF

FIGURE 5

The incident tangential electric fields for all of the described methods, to alter the exposure conditions. Each column represents the electric field along the implant trajectory. The columns are normalized to have the same vector length. Afterward, these matrices are used to compare the normalized singular values between the methods

FIGURE 6

The normalized singular values of the different excitation methods. The legend indicates the sum for the different methods, where a higher number indicates more equivalent orthogonal exposure conditions; thus, more information is obtained using that method

FIGURE 7

The two local transmit coils that were constructed to create different incident tangential electric field exposures along the lead trajectory. A,B, The length and width of the larger loop coil. C,D, The length and width of the smaller loop coil. E, The generated signal of the larger loop coil in a sagittal slice through the phantom. F, The corresponding z‐component of the electric field

FIGURE 8

A comparison between the transfer functions obtained with finite‐difference time‐domain simulations and the transfer function obtained with MRI. The top row shows the magnitude and phase of the transfer function for the bare copper wire; the second row shows the same for the insulated copper wire; the third row shows the TF for the coaxial lead; and the bottom row is the TF for the spinal cord stimulator. For all of the TFs, the blue‐shaded area displays the SD of the measured TF

FIGURE 9

The SAR that is calculated using the transfer function and the known exposure condition correlated with the SAR that is calculated from the measured temperature curves. Six measurements were done with the small transmit coil, and four were done with the large transmit coil for a total of 10 measurements per lead