pubmed.ncbi.nlm.nih.gov

Investigating cardiac stimulation limits of MRI gradient coils using electromagnetic and electrophysiological simulations in human and canine body models - PubMed

. 2021 Feb;85(2):1047-1061.

doi: 10.1002/mrm.28472. Epub 2020 Aug 19.

Affiliations

Investigating cardiac stimulation limits of MRI gradient coils using electromagnetic and electrophysiological simulations in human and canine body models

Valerie Klein et al. Magn Reson Med. 2021 Feb.

Erratum in

Abstract

Purpose: Cardiac stimulation (CS) limits to gradient coil switching speed are difficult to measure in humans; instead, current regulatory guidelines (IEC 60601-2-33) are based on animal experiments and electric field-to-dB/dt conversion factors computed for a simple, homogeneous body model. We propose improvement to this methodology by using more detailed CS modeling based on realistic body models and electrophysiological models of excitable cardiac fibers.

Methods: We compute electric fields induced by a solenoid, coplanar loops, and a commercial gradient coil in two human body models and a canine model. The canine simulations mimic previously published experiments. We generate realistic fiber topologies for the cardiac Purkinje and ventricular muscle fiber networks using rule-based algorithms, and evaluate CS thresholds using validated electrodynamic models of these fibers.

Results: We were able to reproduce the average measured canine CS thresholds within 5%. In all simulations, the Purkinje fibers were stimulated before the ventricular fibers, and therefore set the effective CS threshold. For the investigated gradient coil, simulated CS thresholds for the x-, y-, and z-axis were at least one order of magnitude greater than the International Electrotechnical Commission limit.

Conclusion: We demonstrate an approach to simulate gradient-induced CS using a combination of electromagnetic and electrophysiological modeling. Pending additional validation, these simulations could guide the assessment of CS limits to MRI gradient coil switching speed. Such an approach may lead to less conservative, but still safe, operation limits, enabling the use of the maximum gradient amplitude versus slew rate parameter space of recent, powerful gradient systems.

Keywords: MRI gradient field switching; cardiac stimulation; electromagnetic exposure safety; electromagnetic field simulation; heart model; magnetostimulation thresholds.

© 2020 The Authors. Magnetic Resonance in Medicine published by Wiley Periodicals LLC on behalf of International Society for Magnetic Resonance in Medicine.

PubMed Disclaimer

Figures

FIGURE 1
FIGURE 1

Overview of the simulation pipeline for the prediction of cardiac stimulation. A, Detailed body model (shown here are the bones, skin, and heart of the female human model) with added realistic networks of cardiac Purkinje and ventricular muscle fibers. B, Simulated electric fields (E-fields) induced by a time-varying current in a coil. C, The E-field is projected onto the cardiac fiber paths. D, The cardiac response to the extracellular electric potential is predicted using electrical-circuit models of Purkinje and ventricular muscle fibers. Abbreviation: EM, electromagnetic

FIGURE 2
FIGURE 2

Surface model of the myocardium, the vena cava, aorta, and pulmonary arteries of the female body model. Superimposed in red are the Purkinje fibers (A) and the ventricular muscle fibers (B) that have been added to the model using rule-based modeling algorithms,

FIGURE 3
FIGURE 3

A, Canine body model (only skin and myocardium are shown) with a pair of coplanar coils (COP) and a solenoid coil (SOL). B, Simulated damped sinusoidal dB/dt waveform as generated in the experiments, by discharging a capacitor into the coils. The time from onset to first zero crossing is 571 μs for COP, and 540 μs for SOL, respectively

FIGURE 4
FIGURE 4

A, The cardiac fiber model consists of single cells coupled longitudinally by gap junctions, modeled as a resistive T-network. We assigned the axial gap junction resistance Rgap with 1 Ωcm2, the leakage resistance to extracellular space Rsh with 10 kΩ, and the resistivity of the myoplasm Rmyo with 162 Ωcm. B, Simplified depiction of the membrane models of Purkinje cells (Stewart model) and ventricular muscle cells (O’Hara model) showing the different ion channels and ion pumps

FIGURE 5
FIGURE 5

Coronal slices of the male and female voxel models (left column). The center column shows an enlarged section (dashed box in the left column) of the E-field induced in this slice. The right column shows E-field maps in the whole torso as maximum intensity projections (MIPs) of E-field values along the y-direction onto the xz-plane. All E-fields were induced by the Sonata gradient’s z-axis at a slew rate of 100 T/m/s. The E-field in the bones is set to zero for better visibility

FIGURE 6
FIGURE 6

E-field (MIP) induced in the myocardium, the vena cava, aorta, and pulmonary arteries of the male and female body models (rows) by each gradient axis (columns) at a slew rate of 100 T/m/s. The E-field outside the heart is set to zero for better visibility

FIGURE 7
FIGURE 7

Stimulation thresholds of the Sonata gradient coil in terms of gradient amplitude ΔG as a function of the rise time t for a gradient waveform with sinusoidal ramps. Thresholds are plotted for all gradient axes (rows) and both human body models (columns). The cardiac stimulation (CS) thresholds are plotted in red (Purkinje fibers as solid lines, ventricular muscle fibers as dashed lines), simulated peripheral nerve stimulation (PNS) thresholds are plotted in blue, and International Electrotechnical Commission (IEC) 60601–2-33 cardiac safety limits are shown in black

FIGURE 8
FIGURE 8

Electric-field rheobase (A) and time constants (B) obtained from fitting exponential and hyperbolic strength-duration models to the “E-field threshold” versus “pulse duration” curves for the two human models and the three gradient axes. The fit results are shown in terms of mean ± SD of all six simulations (see individual fit curves in Supporting Information Figure S6)

FIGURE 9
FIGURE 9

Simulated and experimental CS thresholds of the dogs for the coplanar coils (COP) and the solenoid coil (SOL) in terms of peak dB/dt amplitude. All thresholds are given for an equivalent rectangular waveform of 571 μs (COP) and 540 μs (SOL). The simulated thresholds correspond to the Purkinje fiber thresholds (thresholds of the ventricular muscle fibers were approximately 6-fold higher)

Similar articles

Cited by

References

    1. Listerud J, Einstein S, Outwater E, Kressel HY. First principles of fast spin echo. Magn Reson Q. 1992;8:199–244. - PubMed
    1. Mansfield P, Maudsley AA. Planar spin imaging by NMR. J Magn Reson. 1977;27:101–119.
    1. Bieri O, Scheffler K. Fundamentals of balanced steady state free precession MRI. J Magn Reson Imaging. 2013;38:2–11. - PubMed
    1. Martin J, Endt S, Wetscherek A, et al. Contrast-to-noise ratio analysis of microscopic diffusion anisotropy indices in q-space trajectory imaging. Z Med Phys. 2020;30:4–16. - PubMed
    1. Reilly J. Peripheral nerve stimulation by induced electric currents: Exposure to time-varying magnetic fields. Med Biol Eng Comput. 1989;27:101–110. - PubMed

Publication types

MeSH terms