Introduction

Intraoperative neuromonitoring (IONM) is critical to reduce neural injury when ablating lesions near nerves and has been introduced into MRI-guided ablations [1, 2]. As there are no MRI-conditional IONM systems currently available, a conventional neuromonitoring system with metallic needle electrodes has been tested for use under specific conditions [2]. Generalizing the use of IONM to wider use conditions is needed; however, the risk of radiofrequency (RF)-heating of electrodes in certain configurations can be significant. This work further investigates the factors contributing to RF-heating risk during IONM and presents several strategies to enable IONM during MRI-guided ablation even for challenging configurations.

Methods

Experiments with ex vivo porcine tissue were performed on a 1.5 T scanner (Ingenia, Philips Healthcare, Best, Netherlands) using a conventional IONM system (Cascade Pro, Cadwell, Kennewick, WA). One IONM use scenario with high RF-heating risk was monitoring MRI-guided ablation of a vascular malformation in the right shoulder near the brachial plexus, and was simulated with porcine tissue fashioned to mimic a head, shoulders, and arms (Figure 1a). One pair of stimulator electrodes was inserted in the “head,” and four pairs of recording electrodes were inserted into the left and right “shoulders,” right “arm,” and “wrist.” A temperature sensor was attached to each recording electrode to track temperature at the tissue surface. The recording electrode wires were routed either along the two sides of the bore, or along the middle line of the patient table at different heights. By default, the image volume was centered at the “shoulder” level.

With the wires in the “Large” position (Figure 1b), the image volume was shifted in 5 cm increments in the superior/inferior direction to study the impact of wire lengths within the MRI bore on RF-heating. At the worst heating configuration, the impact of local tissue property was investigated by moving electrode #2 in Figure 1a from a location undergone multiple tests to a “fresh” location ~1 cm away without moving anything else. To investigate the impact of connection to the IONM system, experiments in the “Sides” configuration with image volume at “shoulder” level were repeated with the electrode wires connected to/disconnected from the extender pod connected to the IONM system. The channels the electrodes were plugged into were also swapped in experiments to study whether that had any impact on RF-heating.

For imaging, a fast-spin-echo (FSE) sequence (specific absorption rate (SAR): 1.9 W/kg) was used unless otherwise specified. To examine the dependence of heating on SAR, a lower-SAR (0.9 W/kg) FSE sequence as well as a gradient-echo (GRE) sequence (SAR <0.1 W/kg) were also used. All sequences had a scan time of ~2 minutes.

Results

Figure 2 shows temperature elevations during imaging for all recording electrodes in one of the worst-case scenarios investigated. Temperature elevations up to 63.9 °C were observed and varied significantly among different electrodes. Relocating electrode #2 to a slightly different position ~1 cm away resulted in a dramatic decrease in heating, with other unmoved electrodes exhibiting different changes in temperature elevation as well. Figure 3 shows that maximum temperature elevations of electrode #2 increased as the wires moved away from the center of the scanner bore. Temperature elevations for the “Large” and “Sides” configurations were comparable. Figure 4 shows that maximum temperature elevations increased as wire length within the bore increased. In Figure 5, the two FSE series showed temperature elevations that were approximately proportional to SAR. The GRE sequence resulted in negligible heating.

Disconnecting electrode wires from the IONM system resulted in up to a 30% change in electrode temperatures (~6° C), with temperature elevations either increasing or decreasing for each individual electrode. Altering electrode connections (swapping or disconnecting a subset) at the extender pod also caused changes in the recorded temperature on each electrode and relative to other electrodes.

Discussion

Significant RF-heating at the IONM electrode insertion locations was demonstrated in the ex vivo experiments. RF-heating was generally more severe with wires closer to the scanner bore, longer wire lengths within the scanner bore, and higher SAR sequences. Additionally, local tissue properties and inter-electrode interactions in both tissue and the IONM system had substantial impacts on RF-heating. Using a GRE sequence with less than 0.1 W/kg SAR resulted in negligible temperature elevations even with the relatively high-risk “Large” wire configuration.

Conclusion

The temperature elevation on IONM electrodes was demonstrated to be dependent on the imaging sequence SAR and wire configurations within the scanner bore. Using optimal wire routing along with an extremely low-SAR sequence can significantly reduce the RF-heating risk during MRI-guided ablations with IONM monitoring.

Acknowledgements

No acknowledgement found.

References

1. Kurup AN, Morris JM, Boon AJ, Strommen JA, Schmit GD, Atwell TD, Carter RE, Brown MJ, Wass CT, Rose PS, Callstrom MR. Motor Evoked Potential Monitoring during Cryoablation of Musculoskeletal Tumors. J Vasc Interv Radiol, 2014. 25(11): 1657-1664.

2. Thompson SM, Gorny KR, Koepsel EMK, Welch BT, Mynderse L, Lu A, Favazza CP, Felmlee JP, Woodrum DA. Body interventional MRI for diagnostic and interventional radiologists: current practice and future prospects. Radiographics, 2021. 41(6): 1785-1801.