Optical Mapping-Validated Machine Learning Improves Atrial Fibrillation Driver Detection by Multi-Electrode Mapping

Abstract

Background - Atrial fibrillation (AF) can be maintained by localized intramural reentrant drivers. However, AF driver detection by clinical surface-only multi-electrode mapping (MEM) has relied on subjective interpretation of activation maps. We hypothesized that application of Machine Learning (ML) to electrogram frequency spectra may accurately automate driver detection by MEM and add some objectivity to the interpretation of MEM findings. Methods - Temporally and spatially stable single AF drivers were mapped simultaneously in explanted human atria (n=11) by subsurface near-infrared optical mapping (NIOM) (0.3mm 2 resolution) and 64-electrode MEM (Higher-Density (HD) or Lower-Density (LD) with 3mm 2 and 9mm 2 resolution, respectively). Unipolar MEM and NIOM recordings were processed by Fourier Transform analysis into 28407 total Fourier spectra. Thirty-five features for ML were extracted from each Fourier spectrum. Results - Targeted driver ablation and NIOM activation maps efficiently defined the center and periphery of AF driver preferential tracks and provided validated classifications for driver vs non-driver electrodes in MEM arrays. Compared to analysis of single electrogram frequency features, averaging the features for each surrounding 8 electrodes neighborhood, significantly improved classification of AF driver electrograms. The classification metrics increased when less strict annotation including driver periphery electrodes were added to driver center annotation. Notably, f1-score for the binary classification of HD catheter dataset were significantly higher than that of LD catheter (0.81 ± 0.02 vs 0.66 ± 0.04, p<0.05). The trained algorithm correctly highlighted 86% of driver regions with HD but only 80% with LD MEM arrays (81% for LD+HD arrays together). Conclusions - The ML model pre-trained on Fourier spectrum features allows efficient classification of electrograms recordings as AF driver or non-driver compared to the NIOM gold-standard. Future application of NIOM-validated ML approach may improve the accuracy of AF driver detection for targeted ablation treatment in patients.

Computational models of atrial fibrillation: achievements, challenges, and perspectives for improving clinical care

“…148 In brief, the ability of ML and ‘big data’ to identify complex associations between numerous variables of interest in a data-driven, hypothesis-free approach make them attractive for identifying occult AF determinants and establishing clinical decision support systems. 146 ML has been employed to predict the future AF incidence from electronic health records 95 , 96 and sinus rhythm ECGs, 97 predict stroke risk from a daily AF burden signature, 149 define AF clinical classifications based on different risks for adverse clinical outcomes, 150 classify intracardiac activation patterns during AF to detect regional rotational activity, 105 , 106 identify patients who may benefit from AF cardioversion, 107 and predict AF recurrence after the first catheter ablation procedure. 102–104 These ML approaches exploited diverse pre-procedural patient characteristics as inputs, including laboratory and clinical parameters, 102 , 103 atrial geometry, 151 and imaging data.…”

Deep Learning Classification of Unipolar Electrograms in Human Atrial Fibrillation: Application in Focal Source Mapping

“…Similar to Hannun et al (2019), we adopt residual CNN because all EGM signals were of the same duration, so that LSTM was not required. Machine learning models have also been developed to detect rotational activation during human AF, but the input training dataset was either color-coded phase maps (Alhusseini et al, 2020) or EGM frequency spectral features (Zolotarev et al, 2020) from a multielectrode array, and not raw EGMs as in our study. In the CNN model by Alhusseini et al (2020), rotational activation was detected with an accuracy of 95%, while more classic machine learning models by Zolotarev et al (2020) achieved an accuracy of 80-90% depending on size of the multi-electrode mapping array input into the model.…”

“…Machine learning models have also been developed to detect rotational activation during human AF, but the input training dataset was either color-coded phase maps (Alhusseini et al, 2020) or EGM frequency spectral features (Zolotarev et al, 2020) from a multielectrode array, and not raw EGMs as in our study. In the CNN model by Alhusseini et al (2020), rotational activation was detected with an accuracy of 95%, while more classic machine learning models by Zolotarev et al (2020) achieved an accuracy of 80-90% depending on size of the multi-electrode mapping array input into the model. In our study, the performance of classic machine learning models, such as logistic regression, SVM and KNN, in classifying FaST sites was inferior to that of DL, which highlights the computational proficiency of DL in EGM classification without the requisite for discrete feature input, such as unipolar EGM onset.…”

Understanding PITX2-Dependent Atrial Fibrillation Mechanisms through Computational Models

“…The 3D virtual human atria not only allows to differentiate the relative contribution of each variable (i.e., gene variants, ion channels, In order to study human AF mechanisms and treatment, a lot of effort has been invested to integrate detailed anatomical, structural and electrophysiological information in the three-dimensional (3D) computer atrial modeling [100,101]. Based on experimental/clinical data on medical imaging and invasively acquired electroanatomic maps, atrial geometry with wall thickness [102], fibrosis distribution [103,104], myofibre orientation, regional electrical heterogeneities and AF driver distribution [105] were used to develop patient-specific 3D models [106]. In details, models with real atrial geometry are reconstructed from medical imaging, specifically from cardiac MRI and/or cardiac CT scans using image segmentation and 3D reconstruction algorithms [107][108][109][110].…”