Activation Time Detection Algorithms Used in Computerized Intraoperative Cardiac Mapping: A Comparison with Manually Determined Activation Times

Activation Time Detection for Intraoperative Mapping. Introduction: Local activation determined from monopolar electrograms usually is marked at the time of the minimum slope. In bipolar recordings, the peak, the maximum absolute slope, the baseline crossing of the segment containing the maximum absolute slope, and morphological criteria have been used. The purpose of this work is: (1) to compare activation times determined using these algorithms with activation times corrected by electrophysiologists; and (2) to determine which computerized algorithm requires fewest operator correction of outliers. Methods and Results: Monopolar and bipolar electrograms were collected from the epicardium of subjects undergoing surgery for ventricular tachycardia and Wolff‐Parkinson‐White syndrome, and the above bipolar and monopolar algorithms were evaluated. The electrophysiologists viewed adjacent channels and evolving isochronal maps but not the computer markings. The morphological algorithm compared most favorably with times marked up by electrophysiologists having an overall average difference of 2.2 ± 3.1 msec. For monopolar data, a five point, least square fit, second‐order slope approximation algorithm most closely agreed with the electrophysiologist corrected algorithm (3.2 ± 4.0 msec). The morphological and least square algorithms also produced the fewest outliers (3.1% and 5.7%, respectively). When outliers were removed, bipolar and monopolar discrepancies were reduced to 1.8 ± 1.9 and 2.5 ± 2.3 msec, respectively. Conclusions: Thus, among those algorithms tested, an algorithm based on morphological criteria is best suited for computer determination of bipolar activation times and compares favorably to variations between electrophysiologist corrected and computer marked activation times determined traditionally by the largest negative slope in monopolar data. Identification of outliers using contextual and improved morphological algorithms may improve computerized mapping system performance. Copyright © 1991, Wiley Blackwell. All rights reserved

Duke Scholars

Author Carl F. Pieper Biostatistics & Bioinformatics, Division of Biostatistics