Filtering of pathological ventricular rhythms during MRI scanning

Julien Oster (1) , Matthieu Geist (2) , Olivier Pietquin (2) , Gary D. Clifford (1)

1 IBME - Institute of Biomedical Engineering [Oxford]
2 IMS : Information, Multimodalité & Signal

The detection of pathological ventricular repolarization during Magnetic Resonance Imaging (MRI) examination becomes necessary with the development of MRI guided surgery or electrophysiology and stress-test MRI. The Electrocardiogram (ECG) signal is distorted by the MagnetoHydroDynamic (MHD) effect, arising from blood flow and thus the motion of electrolytes inside a magnetic field. This distortion makes ECG diagnosis impossible, and thus the development of an accurate MHD suppression is mandatory. An extension of the ECG model based nonlinear Bayesian filtering is proposed to deal specifically with the MHD artefact by incorporating a model of the MHD based on Gaussian waves. The assessment of an MHD suppression technique is difficult due to the high level of noise and the lack of information about the ventricular repolarization. The method has thus been applied on an artificial signals, modelling an acquisition of ECG during MRI examination. Two different cases were modelled: one ECG signal with T wave inversion and an ECG, which exhibited a prolonged QT interval. The results show promising denoising of the ECG signal enabling detection of pathological ventricular repolarization, but application of this technique on real acquisitions are needed to assess the filtering behaviour on real signals.

Julien Oster, Matthieu Geist, Olivier Pietquin, Gary D. Clifford. Filtering of pathological ventricular rhythms during MRI scanning. BSI2012, Jul 2012, Como, Italy. pp.97-100. ⟨hal-00749457⟩