Search

Atrial fibrillation is a clinically important arrhythmia. There are some reports on machine learning models for AF diagnosis using electrocardiogram data. However, few reports have proposed an eXplainable Artificial Intelligence (XAI) model to enable physicians to easily understand the machine learning model's diagnosis results.We developed and validated an XAI-enabled atrial fibrillation diagnosis model based on a convolutional neural network (CNN) algorithm. We used Holter electrocardiogram monitoring data and the gradient-weighted class activation mapping (Grad-CAM) method.Electrocardiogram data recorded from patients between January 4, 2016, and October 31, 2019, totaling 57,273 electrocardiogram waveform slots of 30 seconds each with diagnostic information annotated by cardiologists, were used for training our proposed model. Performance metrics of our AI model for AF diagnosis are as follows: sensitivity, 97.1% (95% CI: 0.969-0.972); specificity, 94.5% (95% CI: 0.943-0.946); accuracy, 95.3% (95% CI: 0.952-0.955); positive predictive value, 89.3% (95% CI: 0.892-0.897); and F-value, 93.1% (95% CI: 0.929-0.933). The area under the receiver operating characteristic curve for AF detection using our model was 0.988 (95% CI: 0.987-0.988). Furthermore, using the XAI method, 94.5 ± 3.5% of the areas identified as regions of interest using our machine learning model were identified as characteristic sites for AF diagnosis by cardiologists.AF was accurately diagnosed and favorably explained with Holter ECG waveforms using our proposed CNN-based XAI model. Our study presents another step toward realizing a viable XAI-based detection model for AF diagnoses for use by physicians.

Aims: This study aimed to evaluate the electrocardiographic characteristics and predictors of successful ablation for ventricular arrhythmias (VAs) with superior axis originating in the ventricular septum., Methods and Results: This study included 385 consecutive patients with VAs undergoing radiofrequency ablation (RFA). Of these, 14 patients (3.7%) were identified who had VAs with superior axis that were mapped to and ablated at the left ventricular (LV) septum. These patients were classified into two groups, successful (n = 9, Success-RFA) and failed (n = 5, Fail-RFA) ablation. The QRS duration of the VAs was longer in the Success-RFA than the Fail-RFA [median (25%, 75% quartile), 140 (134, 149) vs. 128 (116, 132) ms; P = 0.007]. In the Success-RFA, the QRS morphology in lead V1 exhibited qR or rSR (r < 0.2 mV) pattern. In the Fail-RFA, QRS in lead V1 demonstrated an initial R-wave of ≥0.2 mV except for one patient who demonstrated a qR pattern. The initial R-wave amplitude of <0.2 mV in lead V1 identified successful ablation cases with 100% sensitivity and 80% specificity. The magnitude of the initial R-wave amplitude in lead V1 could be related to the originating site's depth within the septal tissue, which could also explain the RFA results., Conclusion: Four percent of VA patients had superior axis on electrocardiogram and foci that mapped to the LV septum, two-thirds of which were successfully ablated. The initial R-wave amplitude of <0.2 mV in lead V1 identified RFA success with high sensitivity and specificity., (Published on behalf of the European Society of Cardiology. All rights reserved. © The Author 2015. For permissions please email: journals.permissions@oup.com.)

Background: The aim of this study was to identify the ECG features that might differentiate between anterior interventricular vein (AIV) and distal great cardiac vein (d-GCV) outflow tract-ventricular arrhythmias (OT-VAs)., Methods and results: Radiofrequency catheter ablation was performed in 13 of 375 patients (3.5%) for AIV or d-GCV OT-VAs. We grouped the 13 patients by the origin, d-GCV (n=9) or AIV (n=4), and compared their ECGs and electrophysiological data. The OT-VA ECGs had S waves in lead I in all 13 patients. The voltage in the inferior lead III and peak deflection index showed no significant between-group differences (2.3±0.7 vs. 2.5±0.3 mV and 0.65±0.04 vs. 0.68±0.04 mV, respectively) for the d-GCV and AIV groups. There were also no significant between-group differences in the QaVL/QaVR, where Q denotes the amplitude of the Q wave in the suffix lead. However, the R/S ratio in V1 (1.7±1.0 [n=5] vs. 0.2±0.05, P=0.04), and QRS duration (149±16.6 vs. 123±3.8 ms, P=0.012) were greater in the d-GCG group than in the AIV group. There were no significant between-group differences in the activation time or pace mapping score at the optimal ablation sites., Conclusions: A low R/S ratio in V1 and shorter QRS duration may help identify AIV sites of epicardial OT-VA origin.

Introduction: Confined thoracic vein fibrillation (cTVT) is a finding that suggests that thoracic vein acts as a rapid driver to maintain atrial fibrillation (AF). However, little is known about the cTVT., Methods and Results: Among consecutive 655 patients (age 62 ± 0 years, 492 men, 421 paroxysmal) who underwent circumferential pulmonary vein (PV) antrum isolation for AF, cTVT was identified in 28 (4.3%) patients. The prevalence was significantly higher in patients with paroxysmal AF than in those with nonparoxysmal AF (5.9% vs 1.3%, P = .002). The cTVT was observed in left PVs in 15 (53.6%), right PVs in 11 (39.3%), and superior vena cava in 2 (7.1%) patients. The median cycle length of cTVT was 150 (110-170) ms. The cTVT was recognized when sinus rhythm was restored from AF during vein isolation in 14 patients. Dissociated activity was seen after the termination of cTVT in 23 (82.1%) patients, and cTVT reinitiated spontaneously after the dissociated activity. In 2 patients, AF was not terminated by multiple cardioversions before the isolation, even with a maximal energy delivery. At a median follow-up of 12.0 (7.5-20.5) months, 26 patients (92.9%) were free from AF without antiarrhythmic drugs after a mean of 1.4 ± 0.5 procedures per patient. Notably, recurrent arrhythmia was not observed in any patient (n = 6), wherein cTVT was terminated by additional radiofrequency applications inside an isolated area after the achievement of vein isolation., Conclusions: Confined thoracic vein fibrillation is not a rare finding in patients with paroxysmal AF, and its elimination results in an excellent clinical outcome., (Copyright © 2014 Mosby, Inc. All rights reserved.)