Makoto Hirohata

Morphology-matching-based R-wave detection for noise-robust ECG gating

Background Accurate ECG R-wave detection is crucial for cardiac gating in MRI. However, in high-field MRI systems, it is hard to detect R-waves in ECG signals accurately, because the amplitude of the ECG signal may be smaller than that of the noise induced by the MRI system. To overcome this issue, existing studies have focused on (a) acquiring additional ECG signals or on (b) improving the R-wave detector to be robust against noise. In the first approach, ECG gating with a 12-lead ECG has been reported to have high accuracy [1]. However, due to ECG monitor limitations, this study utilizes a common dual-lead ECG. We propose a new morphology-matching-based R-wave detector for noise-robust ECG gating. The morphology is analyzed in filtered ECG signals, and the R-wave is detected by matching the input ECG signals to R-wave templates. The templates are updated when the MRI system is not scanning, and they contribute to robustness against patient variation and noise [1].