Prediction of atrial fibrillation inducibility using spatiotemporal activation analysis combined with network mapping. (April 2021)

• Record Type: Journal Article
• Authors: He, Kaiyue; Feng, Xujian; Wu, Ziqian; Yang, Cuiwei; Wu, Zhong; Chen, Ying
• Is Part Of: Biomedical signal processing and control. Volume 66(2021)
• Publication Date: 2021-04
• Subjects: Atrial fibrillation (AF) prediction -- Activating rule -- Network mapping -- Sinus rhythm -- Spatiotemporal features; Signal processing -- Periodicals; Biomedical engineering -- Periodicals; Signal Processing, Computer-Assisted -- Periodicals; Image Processing, Computer-Assisted -- Periodicals; Biomedical Engineering -- Periodicals; 610.28
• DOI: 10.1016/j.bspc.2021.102460
• Languages: English
• ISSNs: 1746-8094

Abstract

Highlights: AF vulnerability correlates with atrial waveform variability and depolarization asynchrony. Multi-lead signals analysis benefits the holistic observation of atrial electrophysiology. Combining spatial and temporal features better reflects atrial status than independently. Electrical measuring might be a convenient and general-purpose way to assess the atrial status of different animal AF models. Abstract: Objective: Atrial fibrillation inducibility has recently been shown to be associated with the increased clinical recurrence after ablation. Previous studies have identified unstable sinus rhythm before the occurrence of paroxysmal AF, yet earlier subtle changes, related to AF inducibility, are less be investigated. The purposed of this study is to predict AF inducibility, here we established a controllable canine model and prepared various sinus rhythms with different AF inducibility to study atrial electrophysiological changes. Methods: The data were derived from acetylcholine induced acute canine AF models ( n = 5, data samples = 90) through epicardial mapping system using 64 electrodes attached to the left and right atrial appendages. Signal preprocessing consists of three steps: noise reduction, removal of ventricular artifact and extraction of activation time (AT). Two classifiers based on activating rule and spatiotemporal features, visualized by network mapping, were established by binary logical regression analysis. Results: The sensitivity, specificity and accuracy of activation analysis are 83.30%, 91.70% and 87.80% respectively, while those of spatiotemporal analysis are apparently increased to 97.60%, 93.80% and 95.56%. We observed that the atria (in sinus rhythm) with more disordered electrophysiological activity was more vulnerable to AF. Conclusion: The high accuracies indicate the feasibility of AF inducibility prediction using sinus rhythm electrocardiogram, especially with the aid of superior spatiotemporal analysis. Significance: To our knowledge, this is the first study that demonstrates the possibility of predicting AF inducibility using network mapping. With this study, accurate AF inducibility forecasting may help to evaluate the recurrence of AF after ablation.