- Workshop presentation
- Open access
- Published:
Automated removal of gradient-induced voltages from 12-lead ECG traces during high-gradient duty-cycle MRI sequences
Journal of Cardiovascular Magnetic Resonance volume 18, Article number: W4 (2016)
Background
During cardiac MR imaging (CMR), ECG traces (<10 mV) are highly distorted due to strong gradient-induced voltages (GIV). GIV reaches 1000-3000 mV in high gradient-duty-cycle [GDC = (Total Gradient Activity Time per cardiac cycle)/(Cardiac cycle time)] sequences, for ECG electrodes located farther from magnet's iso-center. An approach which predicted and removed GIV overlaid on ECG traces acquired during CMR using a 19-parameter analytical equation based on Maxwell's equations [Bowtell, MRM. ‘00] and the concomitant fields equation [Bernstein, MRM. '89], was previously verified in 10 subjects at rest and following stress [Zhang et Al, DOI: 10.1002/mrm.25810]. Automating this method would enable its use for efficiently synchronizing CMR (not missing QRS peaks), and for detecting cardiac events (VF, VT) during higher-risk CMR and MR-guided interventions.
To this end we created a self-triggered software pipeline, which requires minimal human intervention, in order to compute and display 12-lead ECG traces without GIV. We verify its efficiency in human subjects during high GDC sequences; SSFP and Short-TR GRE.
Methods
CMR was performed at 3T (Skyra, Siemens). All data used in the analysis were recoded using a National Instruments Data Acquisition card controlled by MATLAB [Natick, MA]. A commercial CardioLab system (GE, Waukesha, WI) was used to amplify and filter the signal from the ECG leads.
The process is outlined in Figure 1b. Prior to the beginning of imaging, ECG traces without GIV are recorded, after which the software awaits a triggering event to start acquiring data corrupted by GIV. A pulse, marking the start of gradient activity, triggers this acquisition at the start of a training sequence. Imaging gradient-waveforms Gx, Gy, Gz) were simultaneously recorded directly from Siemens hardware. Once all the traces have been acquired for at least three QRS cycles, GIVs are calculated by subtracting the clean ECG template from the corrupted traces, and then fitted to the equation, yielding the coefficients for each sequence. These coefficients are automatically transferred to the real-time program-component, which calculates GIV for each ECG electrode at each time-point during CMR, using the simultaneously-acquired gradient-waveforms, and then subtracts them from the corrupted signals, thereby cleaning the traces (Figure 1).
Results
In two volunteers, we obtained >95% noise removal in 12-lead ECG traces for SSFP and GRE using this automated approach which functioned equivalently to the manual method. The approach succeeded in removing GIV even during changes in heart-rate (Figure 2). Since QRS peaks are clearly discerned in the cleaned ECGs, a VCG-approach [Gregory et al, DOI: 10.1002/mrm.25078] provides >90% efficiency in CMR synchronization.
Conclusions
An efficient automated procedure to remove gradient-induced voltages from ECG traces permits real-time 12-lead ECG-monitoring during high gradient-duty-cycle sequences.
Author information
Authors and Affiliations
Rights and permissions
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
About this article
Cite this article
Dabaghyan, M., Zhang, S.H., Ward, J. et al. Automated removal of gradient-induced voltages from 12-lead ECG traces during high-gradient duty-cycle MRI sequences. J Cardiovasc Magn Reson 18 (Suppl 1), W4 (2016). https://doi.org/10.1186/1532-429X-18-S1-W4
Published:
DOI: https://doi.org/10.1186/1532-429X-18-S1-W4