Non-algorithmic approach to accessory pathway localization based on a mobile app

Abstract

Abstract Funding Acknowledgements Type of funding sources: Public Institution(s). Main funding source(s): Polish Cardiac Society - Section of Non-invasive Electrocardiology and Telemonitoring (PTK SENIT) Background. Electrocardiographical localization of accessory pathways (AP) in WPW syndrome remains an elusive goal. Despite several, often very complex algorithms that were developed over the last three decades, none of them is very accurate - as demonstrated by several independent validation studies. Moreover, algorithms are difficult to remember, which further limits their usefulness. The accessory pathway regions are arbitrary and transitions between them are smooth, not binary. The same applies to ECG features such as QRS polarity – they are not binary. Moreover, even the most careful ECG lead placement cannot compensate for the variability in heart position within the chest. Therefore, it is unrealistic to expect that any ECG feature or feature set will precisely identify a particular location in a 0 vs. 1 fashion. Aim We intend to create a mobile app / web based localizing application, that would support a novel method for AP localization: a colour density map generator - based on actual distribution of a sizable number of accessory pathway with a particular QRS features. Methods A total of 881 consecutive AP ablation procedures were reviewed and patients with successful ablation of an overt accessory pathway were included. ECG with baseline and full preexcitation obtained during incremental atrial pacing were obtained. AP localization was based on stored fluoroscopic images in several projections and procedure description. Each AP localization, as well as QRS polarity in all 12 leads, separately for full and baseline preexcitation patterns, were coded and introduced into the application database. An application web-based, optimized for mobile devices was developed. This application enabled to graphically show changing localization of AP with regard to the introduced, at each step, QRS polarity in a particular ECG lead. User can choose his/her own steps or criteria; there is no need to adhere to the proposed steps. Results A total of 604 patients/ ECGs were available for density map generation. With each additional lead QRS information introduced into the application the user can appreciate more and more precise localization of the AP. This quickly enables to pinpoint the area of maximum probability, i.e. the most likely/frequent location, followed by the immediate adjacent locations. Representative screen shots of to typical inital steps for left- and right-sided APs are provided below (Figure 1 and Figure 2). Limited prospective assessment was very favourable. Conclusions A novel method, easy to use, not requiring to remember any algorithmic steps, and based on an analysis of by far the largest cohort of over accessory pathways (in comparison with published algorithms) was developed. Prospective validation of the localization application seems warranted. Abstract Figure.