Session S34.1

Non-Invasive Cardiac Imaging Based on Just the Standard 12-Lead Signals?

PM van Dam*, TF Oostendorp, A van Oosterom

Radboud University Medical Center
Arnhem, Netherlands

Currently, methods exist to estimate non-invasively the electrical activity of the heart from up to 265 ECG signals. The number of ECG leads hampers the clinical application of such methods. In our research we have been using 64 leads, a number tuned to the number of independent components observed in large data bases. We report on progress in using the non-invasive imaging method based on a double layer at the myocardial surface as the equivalent source of cardiac activity. The local source strength is the transmembrane potential, its waveform described analytically, specified by timing parameters of activation and recovery. The timing sequences were estimated by minimizing the difference between measured and simulated ECGs. The parameter estimation procedure involved is non-linear, and consequently requires the specification of an initial estimate. For the depolarization timing, this was based on the fastest route algorithm, while taking into account the anisotropic nature of propagation; for the recovery times it was based on electrotonic effects. Body surface potentials and individual geometry were recorded on: a healthy subject, a WPW patient and a Brugada patient during an Ajmaline provocation test. The resulting activation and repolarization sequences were in agreement with the (sparse) knowledge available from invasive data. We investigated how much the quality of the inverse results would be affected by using fewer leads. After first testing the 32 leads of the Lux lead system, we went as far as also testing the performance based on using just the signals of the standard 12-lead system. Both lead-sets were tested, for all three cases, by comparing the inversely estimated timing with the results based on 64 signals. When using 32 leads, the activation and recovery sequences closely matched the ones estimated from 64 signals (average correlation 90% for the depolarization times). The observed differences were larger, as expected, when using the signals of the standard 12-lead system (8 independent components). However, the nature of the resulting timing sequences (patterns, locations of extremes) were similar. The robustness of the method is attributed to our use of initial estimates based on the general electrophysiology of propagation. The results hold a clear promise for estimating activation and recovery times in dedicated clinical applications.

(Abstract Control Number: 104)