Classification Of Atrial Fibrillation Research Articles

Preprocedural atrial fibrillation burden is associated with recurrence in patients undergoing atrial fibrillation ablation: findings from the ISOLATION AF Ablation Cohort Study

Abstract Background Non-invasive quantification of electrophysiological properties of the atria is increasingly used for outcome prediction in atrial fibrillation (AF) patients. Data on the benefit of the current classification of AF compared to AF burden based on single lead ECG measurements to predict AF recurrences are scarce. Objective This study sought to assess the predictive performance of low versus high AF burden compared to the conventional AF classification for predicting recurrences after AF catheter ablation. Methods A cross-sectional analysis was conducted in AF patients scheduled for catheter ablation included in the prospective multicenter ISOLATION cohort study between July 2020 and November 2022. Clinical characteristics, routine tests, blood tests and rhythm monitoring through single lead ECGs were collected prior to ablation. ECGs were recorded three times daily for four weeks, with additional measurements in case of the onset or relief of symptoms. The primary endpoint was "ablation success" defined as freedom from documented arrhythmia after the blanking period. Of 650 included patients, pre-procedural rhythm monitoring was available in 580. 261 patients had at least 14 days of ECG measurements and completed 12 months of follow-up. In these patients, we compared the recurrences after 12 months between paroxysmal and persistent AF and high versus low AF burden through survival analysis. Maximally selected rank statistics were used to determine the best cut-off value for AF burden. Results In 261 patients, the mean age was 64 ± 9 years, and 33.7% were female. Physicians classified 170 patients (65.1%) as paroxysmal AF and 91 (34.9%) as persistent AF based on medical history. The AF burden is shown in Figure 1. The optimal calculated cut-off value for AF burden was 42%. Out of 170 patients with paroxysmal AF, 17 (10%) had a burden of > 42%. Of the 91 patients with persistent AF, 35 (38%) had a burden of < 42%. After 12 months, a significant difference in recurrences was observed between patients with paroxysmal AF versus persistent AF (p= 0.015). Patients with high AF burden, had more recurrences after 12 months, than patients with a low AF burden (p= 0.006) (Figure 2). Conclusion Our preliminary results suggest that the addition of a new classification of AF based on burden might provide additional benefit in predicting AF recurrences after catheter ablation compared to the current existing classification and may improve prognostication. Completion of patient follow-up is necessary to determine the predictive value of AF burden derived from single lead ECGs for the prediction of AF recurrence.Figure 2.

Predictors of quality of life improvement in patients undergoing catheter ablation for atrial fibrillation: a cluster-based analysis

Abstract Introduction Traditional classification of Atrial Fibrillation (AF) based on temporal patterns doesn't fully encompass the condition's complexity. Unsupervised cluster analyses have been used for classifying patients into groups with similar comorbid profiles. However, inconsistent comorbidity patterns have been found across different cohorts. Bayesian profile regression, a semi-supervised machine learning technique, may improve the precision of clustering by integrating outcome data, diminishing the variability observed with previous methods. Since quality of life (QoL) primarily drives the indication for AF ablation, we utilized the AF Effect on QualiTy-of-Life (AFEQT) questionnaire to cluster patients. Purpose Our aim is to use patient-reported outcome measures to identify clinically relevant phenotypes of AF that benefit the most from catheter ablation. Methods We implemented a single center follow-up (FUP) program for patients with AF referred for ablation since august 2020, including scheduled visits and remote monitoring through a digital health platform. AFEQT score reported by patients was analyzed using a non-linear mixed model and change in AFEQT by month 12 was used as outcome. Comorbidities, echocardiographic data, and biomarkers were collected. Profile regression mixture modelling guided by QoL improvement was performed to create clinically relevant patient groupings. Penalized multinomial logistic regression model was used for validation. Results From the 466 ablations performed, 354 patients were enrolled until July 2023. 211 patients had FUP > 12 months (age 59±10 years, 31% female, 80% paroxysmal). The overall AFEQT questionnaire completeness rate was 72%. Patients were classified into 3 clusters: 1) patients with long duration AF, left heart cavities enlargement and diastolic dysfunction (n=57); 2) patients with self-reported anxiety/depression (n=50); and 3) patients with low rates of comorbidities (n=104). For clinical application, 8 variables (dyslipidemia, AF type and duration since diagnosis, mitral E/A ratio, systolic pulmonary arterial pressure, left ventricle ejection fraction, and PROMIS T-score for anxiety and depression) were sufficient for classifying patients (85% accuracy in internal cross-validation). Patients with higher depressive and anxiety burden (cluster 2) had worse baseline AFEQT score (median: 46, IQR: 39-60) and derived less benefit (median improvement: 21, IQR: 12-32), while patients with low comorbid burden (cluster 3) displayed higher baseline (median: 58, IQR: 48-62) and final QoL (median improvement: 28, IQR: 21-36, p<0.01). Conclusion Our analysis identified 3 reproducible clinically relevant phenotypes of AF using widely available clinical data. These clusters have distinct associations with QoL, underscoring the heterogeneity of AF and importance of comorbidities and substrates. This approach may allow for the prediction of which patients are most likely to benefit from treatment.

A three-dimensional left atrial motion estimation from retrospective gated computed tomography: application in heart failure patients with atrial fibrillation.

A reduced left atrial (LA) strain correlates with the presence of atrial fibrillation (AF). Conventional atrial strain analysis uses two-dimensional (2D) imaging, which is, however, limited by atrial foreshortening and an underestimation of through-plane motion. Retrospective gated computed tomography (RGCT) produces high-fidelity three-dimensional (3D) images of the cardiac anatomy throughout the cardiac cycle that can be used for estimating 3D mechanics. Its feasibility for LA strain measurement, however, is understudied. The aim of this study is to develop and apply a novel workflow to estimate 3D LA motion and calculate the strain from RGCT imaging. The utility of global and regional strains to separate heart failure in patients with reduced ejection fraction (HFrEF) with and without AF is investigated. A cohort of 30 HFrEF patients with (n = 9) and without (n = 21) AF underwent RGCT prior to cardiac resynchronisation therapy. The temporal sparse free form deformation image registration method was optimised for LA feature tracking in RGCT images and used to estimate 3D LA endocardial motion. The area and fibre reservoir strains were calculated over the LA body. Universal atrial coordinates and a human atrial fibre atlas enabled the regional strain calculation and the fibre strain calculation along the local myofibre orientation, respectively. It was found that global reservoir strains were significantly reduced in the HFrEF + AF group patients compared with the HFrEF-only group patients (area strain: 11.2 ± 4.8% vs. 25.3 ± 12.6%, P = 0.001; fibre strain: 4.5 ± 2.0% vs. 15.2 ± 8.8%, P = 0.001), with HFrEF + AF patients having a greater regional reservoir strain dyssynchrony. All regional reservoir strains were reduced in the HFrEF + AF patient group, in whom the inferior wall strains exhibited the most significant differences. The global reservoir fibre strain and LA volume + posterior wall reservoir fibre strain exceeded LA volume alone and 2D global longitudinal strain (GLS) for AF classification (area-under-the-curve: global reservoir fibre strain: 0.94 ± 0.02, LA volume + posterior wall reservoir fibre strain: 0.95 ± 0.02, LA volume: 0.89 ± 0.03, 2D GLS: 0.90 ± 0.03). RGCT enables 3D LA motion estimation and strain calculation that outperforms 2D strain metrics and LA enlargement for AF classification. Differences in regional LA strain could reflect regional myocardial properties such as atrial fibrosis burden.

Empowering AI-Diagnosis: Deep Learning Abilities for Accurate Atrial Fibrillation Classification

Artificial intelligence (AI) is a powerful technology that can enhance clinical decision-making and the efficiency of global health systems. An AI-enabled electrocardiogram (ECG) is an essential tool for diagnosing heart abnormalities such as arrhythmias. The most prevalent arrhythmia globally is atrial fibrillation (AF), which is an irregular heart rhythm that originates in the atria and can lead to other heart-related complications. A trusted AI classification of AF is explored in this study. Deep learning (DL) has been used to analyze large amounts of publicly available ECG datasets in order to classify normal sinus rhythm (NSR), AF, and other types of arrhythmias. A convolutional neural network (CNN) has been proposed to extract ECG features and classify ECG signals. Based on a 10-fold cross-validation strategy, we conducted experiments involving three scenarios for AF classification: (i) a balanced set, an imbalanced set, and an extremely imbalanced set; (ii) a comparison of ECG denoising algorithms; and (iii) the classification of AF, NSR, and other arrhythmia types (15 classes). As a result, we have achieved 100% accuracy, sensitivity, specificity, precision, and F1-score for the AF, NSR, and non-AF classifications, both for balanced and imbalanced sets. In addition, for the classification of AF, NSR, and other types of arrhythmia (15 classes), the performance results achieved an accuracy of 99.77%, sensitivity of 96.48%, specificity of 99.87%, precision of 97.03%, and F1-score of 96.68%. The results can empower AI diagnosis and assist clinicians in classifying AF on routine screening ECGs.

Diagnosis of atrial fibrillation based on AI-detected anomalies of ECG segments

Early detection of atrial fibrillation (AF) is crucial for its effective management and prevention. Various methods for detecting AF using deep learning (DL) based on supervised learning with a large labeled dataset have a remarkable performance. However, supervised learning has several problems, as it is time-consuming for labeling and has a data dependency problem. Moreover, most of the DL methods do not provide any clinical evidence to physicians regarding the analysis of electrocardiography (ECG) for classification or detection of AF. To address these limitations, in this study, we proposed a novel AF diagnosis system using unsupervised learning for anomaly detection with three segments, PreQ, QRS, and PostS, based on the normal ECG. Two independent datasets, PTB-XL and China, were used in three experiments. We used a long short-term memory (LSTM)-based autoencoder to train the segments of the normal ECG. Based on the threshold of anomaly scores using mean squared error (MSE), it distinguished between normal and AF segments. In Experiment A, the best score was that of PreQ, which detected AF with an AUROC score of 0.96. In Experiment B and C for cross validation of each dataset, the best scores were also of PreQ, with AUROC scores of 0.9 and 0.95, respectively. To verify the significance of the anomaly score in distinguishing between AF and normal segments, we utilized an XG-Boosted model after generating anomaly scores in the three segments. The XG-Boosted model achieved an AUROC score of 0.98 and an F1 score of 0.94. AF detection using DL has been controversial among many physicians. However, our study differentiates itself from previous studies in that we can demonstrate evidence that distinguishes AF from normal segments based on the anomaly score.

Duration of left and right atrial activation wavefront assessed non-invasively is significantly higher in patients with atrial fibrillation than in controls

Abstract Introduction Ablation strategies in atrial fibrillation (AF) are currently a "one-size- fits- all" type of approach. Pulmonary vein isolation as the standard of care is not perfect, especially in the persistent forms, but none of the alternatives has shown better results. The dichotomic paroxysmal versus persistent AF classification in no longer acceptable. Characterization of the atrial substrate before ablation is needed. Electrocardiographic imaging (ECGI) is capable of non-invasively assess atrial activation in a single beat and in a panoramic fashion. The aim of this study was to evaluate the feasibility and accuracy of ECGI to assess the activation speed and identify possible targets for ablation. Methods We studied 18 consecutive patients, 9 referred for ablation of paroxysmal AF and 9 without AF, referred for ablation of premature ventricular contractions (PVCs), that had an ECGI performed before ablation. No patient had structural heart disease and all patients had normal electrocardiographic evaluation in sinus rhythm (SR). The ECGI was performed in SR with the AMYCARD 01 system, based on body surface electrocardiograms of a maximum of 224 electrodes and computed tomography imaging data. Total duration of atrial activation, left atrial (LA), and right atrial (RA) activation duration, and presence of lines of conduction block (LOB) defined as the presence of a jump in the local activation time of 50 ms or more between contiguous points, were assessed. The results were compared in patients with and without AF. Results The results are depicted in the Table. Patients with AF were older. Total atrial activation duration, LA, and RA activation duration were longer in patients with AF and there was a trend towards a higher prevalence of LOB in the LA in patients with AF (Figure). Discussion & Conclusion Total atrial activation duration as well as LA and RA activation duration assessed with the ECGI were significantly prolonged in patients with AF. Presence of LOB in the LA showed a trend towards a higher prevalence in AF patients. These pilot study has proved that this technology is accurate in identifying a pathological substrate in AF patients. We intend to use ECGI in a subsequent study that will include patients with AF both in paroxysmal and persistent forms, in order to identify markers of disease as well as possible targets to perform a customized AF ablation.Comparison between AF and control groupsECGI showing local activation times

Abstract 11962: Breaking Down the Relationship Between Atrial Fibrillation Class and Risk of Embolic Stroke: Beyond CHADSVASC Score

Introduction: Atrial fibrillation (AF) is recognized as a risk factor for stroke. To evaluate the risk of stroke in AF patients, the CHADSVASC score is commonly utilized. However, the correlation between AF classification and the risk of embolic stroke beyond the CHADSVASC score remains unclear. Methods: We analyzed data from 488,790 admissions with embolic stroke, out of which 133,429 had atrial fibrillation, between 2016-2020 to examine the association between AF class and risk of embolic stroke. The CHADSVASC score was computed for each patient, with risk factors being assigned point values ranging from 0 to 2 based on their association with embolic stroke. Patients were categorized into one of four AF classes, namely paroxysmal, persistent, long-standing persistent, or permanent. Results: Patients with persistent, long-standing persistent, and permanent AF had a higher risk of embolic stroke compared to those with paroxysmal AF, even after adjusting for CHADSVASC score (OR: 3.15, 95% CI: 3.08-3.22, P=0.000, 3.48, 95% CI: 3.26-3.71, P=0.000, 3.56, 95% CI: 2.84-4.47, P=0.000, OR: 3.21, 95% CI: 3.12-3.30, P=0.000). Conclusion: Based on our study findings, it can be concluded that nonparoxysmal atrial fibrillation is associated with a higher risk of embolic stroke compared to paroxysmal atrial fibrillation , regardless of the CHADSVASC score. Therefore, patients diagnosed with persistent, long-standing persistent, and chronic AF should receive enhanced monitoring and more proactive management strategies to effectively mitigate the risk of embolic stroke.

Classification of Atrial Fibrillation using Random Forest Algorithm

The electrocardiogram is indicates the electrical activity of the heart and it can be used to detect cardiac arrhythmias. In the present work, we exhibited a methodology to classify Atrial Fibrillation (AF), Normal rhythm, and Other abnormal ECG rhythms using a machine learning algorithm by analyzing single-lead ECG signals of short duration. First, the events of ECG signals will be detected, after that morphological features and HRV features are extracted. Finally, these features are applied to the Random Forest classifier to perform classification. The Physionet challenge 2017 dataset with more than 8500 ECG recordings is used to train our model. The proposed methodology yields an F1 score of 0.86, 0.97, and 0.83 respectively in classifying AF, normal, other rhythms, and an accuracy of 0.91 after performing a 5-fold cross-validation.

Classification of Atrial Fibrillation In ECG Signal Using Deep Learning

Atrial fibrillation is a type of heart rhythm disorder that most often occurs in the world and can cause death. Atrial fibrillation can be diagnosed by reading an Electrocardiograph (ECG) recording, however, an ECG reading takes a long time and requires specialists to analyze the type of signal pattern. The use of deep learning to classify Atrial Fibrillation in ECG signals was chosen because deep learning has 10% higher performance compared to machine learning methods. In this research, an application for classification of Atrial Fibrillation was developed using the 1-Dimentional Convolutional Neural Network (CNN 1D) method. There are 6 configurations of the 1D CNN model that were developed by varying the configuration on the learning rate and batch size. The best model obtained 100% accuracy, 100% precision, 100% recall, and 100% F1 Score.

Multi-center atrial fibrillation electrocardiogram (ECG) classification using Fourier space convolutional neural networks (FD-CNN) and transfer learning

There has been a proliferation of machine learning (ML) electrocardiogram (ECG) classification algorithms reaching >85% accuracy for various cardiac pathologies. Despite the high accuracy at individual institutions, challenges remain when it comes to multi-center deployment. Transfer learning (TL) is a technique in which a model trained for a specific task is repurposed for another related task, in this case ECG ML model trained at one institution is fine-tuned to be utilized to classify ECGs at another institution. Models trained at one institution, however, might not be generalizable for accurate classification when deployed broadly due to differences in type, time, and sampling rate of traditional ECG acquisition. In this study, we evaluate the performance of time domain (TD) and frequency domain (FD) convolutional neural network (CNN) classification models in an inter-institutional scenario leveraging three different publicly available datasets. The larger PTB-XL ECG dataset was used to initially train TD and FD CNN models for atrial fibrillation (AFIB) classification. The models were then tested on two different data sets, Lobachevsky University Electrocardiography Database (LUDB) and Korea University Medical Center database (KURIAS). The FD model was able to retain most of its performance (>0.81 F1-score), whereas TD was highly affected (<0.53 F1-score) by the dataset variations, even with TL applied. The FD CNN showed superior robustness to cross-institutional variability and has potential for widespread application with no compromise to ECG classification performance.

Filtering of remote monitoring alerts transmitted by cardiac implantable electronic devices and reclassification of atrial fibrillation events by a new algorithm

Cardiac implantable electronic devices (CIEDs) are an important means of atrial fibrillation (AF) detection. However, the AF burden measurements and notifications transmitted by CIEDs are not directly related to the clinical classification of paroxysmal, persistent, or permanent AF. Moreover, AF alerts are the most frequent form of notification, imposing a time-consuming review on caregivers. The purpose of this study was to compare the incidence of standard AF burden-related notifications in remotely monitored (RM) patients with the incidence of events detected after filtering by a new proprietary algorithm implementing the standard European Society of Cardiology classification of AF. Between 2017 and 2022, all RM patients with daily AF burden measurements available for ≥30 days and ≥1 AF burden-related alerts were enrolled at 68 medical centers. The incidence of CIED-transmitted alerts was compared to that of AF episodes detected by a new proprietary algorithm and classified as "first recorded episode of AF", "paroxysmal AF", "increased paroxysmal AF", "persistent AF", or "end of persistent AF back to paroxysmal AF or back to sinus rhythm." Between January 2017 and September 2022, this retrospective study analyzed data from 4162 recipients of an Abbott, Biotronik, Boston Scientific, or Medtronic CIED, RM over mean follow-up of 605 ± 386 days. The algorithm broke down 67,883 AF burden-related alerts into 9728 (14.3%) clinically relevant AF events. A new AF alert algorithm successfully identified clinically significant AF events in RM CIED recipients and would markedly limit the total number of transmitted alerts that require review by caregivers.

Design and Implementation of an Atrial Fibrillation Detection Algorithm on the ARM Cortex-M4 Microcontroller.

At present, a medium-level microcontroller is capable of performing edge computing and can handle the computation of neural network kernel functions. This makes it possible to implement a complete end-to-end solution incorporating signal acquisition, digital signal processing, and machine learning for the classification of cardiac arrhythmias on a small wearable device. In this work, we describe the design and implementation of several classifiers for atrial fibrillation detection on a general-purpose ARM Cortex-M4 microcontroller. We used the CMSIS-DSP library, which supports Naïve Bayes and Support Vector Machine classifiers, with different kernel functions. We also developed Python scripts to automatically transfer the Python model (trained in Scikit-learn) to the C environment. To train and evaluate the models, we used part of the data from the PhysioNet/Computing in Cardiology Challenge 2020 and performed simple classification of atrial fibrillation based on heart-rate irregularity. The performance of the classifiers was tested on a general-purpose ARM Cortex-M4 microcontroller (STM32WB55RG). Our study reveals that among the tested classifiers, the SVM classifier with RBF kernel function achieves the highest accuracy of 96.9%, sensitivity of 98.4%, and specificity of 95.8%. The execution time of this classifier was 720 μs per recording. We also discuss the advantages of moving computing tasks to edge devices, including increased power efficiency of the system, improved patient data privacy and security, and reduced overall system operation costs. In addition, we highlight a problem with false-positive detection and unclear significance of device-detected atrial fibrillation.

A Deep Learning Approach for Atrial Fibrillation Classification Using Multi-Feature Time Series Data from ECG and PPG.

Atrial fibrillation is a prevalent cardiac arrhythmia that poses significant health risks to patients. The use of non-invasive methods for AF detection, such as Electrocardiogram and Photoplethysmogram, has gained attention due to their accessibility and ease of use. However, there are challenges associated with ECG-based AF detection, and the significance of PPG signals in this context has been increasingly recognized. The limitations of ECG and the untapped potential of PPG are taken into account as this work attempts to classify AF and non-AF using PPG time series data and deep learning. In this work, we emploted a hybrid deep neural network comprising of 1D CNN and BiLSTM for the task of AF classification. We addressed the under-researched area of applying deep learning methods to transmissive PPG signals by proposing a novel approach. Our approach involved integrating ECG and PPG signals as multi-featured time series data and training deep learning models for AF classification. Our hybrid 1D CNN and BiLSTM model achieved an accuracy of 95% on test data in identifying atrial fibrillation, showcasing its strong performance and reliable predictive capabilities. Furthermore, we evaluated the performance of our model using additional metrics. The precision of our classification model was measured at 0.88, indicating its ability to accurately identify true positive cases of AF. The recall, or sensitivity, was measured at 0.85, illustrating the model's capacity to detect a high proportion of actual AF cases. Additionally, the F1 score, which combines both precision and recall, was calculated at 0.84, highlighting the overall effectiveness of our model in classifying AF and non-AF cases.

Automated atrial fibrillation classification based on denoising stacked autoencoder and optimized deep network

This study proposed an efficient ECG preprocessing technique, and the preprocessed ECG data was subsequently used to classify atrial fibrillation (AFib) using end-to-end deep neural networks. This is significant since early identification of AFib can help to prevent mortality. With this in mind, the two-fold study is proposed, in which three denoising autoencoders for ECG signal pre-processing were evaluated and compared. Denoising autoencoders (DAEs) utilising convolutional neural networks (CNNs) as backbones outperformed the other two DAE models. As a result, the preprocessed signal is then used to classify atrial fibrillation using the three classification models. The combination of CNN-based DAE and CNN-based classification model yielded the best results, with 99.20 percent accuracy, 99.50 percent specificity, 99.50 percent sensitivity, and 99.00 percent true positive rate. The average accuracy of the algorithms we investigated was 96.26 percent, and our technique was 3.2 percent more accurate than the other algorithms we compared. Furthermore, with a 24-h ECG signal processing time of only 1.3 s, the proposed model is computationally inexpensive for real-time applications. To determine its robustness, the proposed framework is tested on previously unseen datasets with varied proportions of arrhythmias, producing a 99.10 percent recall rate and a 98.50 percent accuracy.

Accuracy and variability of cardiologist interpretation of single lead electrocardiograms for atrial fibrillation: The VITAL-AF trial

Screening for atrial fibrillation (AF) using consumer-based devices capable of producing a single lead electrocardiogram (1L ECG) is increasing. There are limited data on the accuracy of physician interpretation of these tracings. The goal of this study is to assess the sensitivity, specificity, confidence, and variability of cardiologist interpretation of point-of-care 1L ECGs. Fifteen cardiologists reviewed point-of-care handheld 1L ECGs collected from patients aged 65 years or older enrolled in the VITAL-AF clinical trial [NCT035115057] who underwent cardiac rhythm assessments with a 1L ECG using an AliveCor KardiaMobile device. Random sampling of 1L ECGs for cardiologist review was stratified by the AliveCor algorithm interpretation. A 12L ECG performed on the same day for clinical purposes was used as the gold standard. Cardiologists each reviewed a common sample of 200 1L ECG tracings and completed a survey associated with each tracing. Cardiologists were blinded to both the AliveCor algorithm and same day 12L ECG interpretation. For each tracing, study cardiologists were asked to assess the rhythm (sinus rhythm, AF, unclassifiable), report their assessment of the quality of the tracing, and rate their confidence in rhythm interpretation. The outcomes included the sensitivity, specificity, variability, and confidence in physician interpretation. Variables associated with each measure were identified using multivariable regression. The average sensitivity for AF was 77.4% (range 50%-90.6%, standard deviation [SD]=11.4%) and the average specificity was 73.0% (range 41.3%-94.6%, SD=15.4%). The mean variability was 30.8% (range 0%-76.2%, SD=23.2%). The average reviewer confidence of 1L ECG rhythm assessment was 3.6 out of 5 (range 2.5-4.2, SD=0.6). Patient and tracing factors associated with sensitivity, specificity, variability, and confidence were identified and included age, body mass index, and presence of artifact. Cardiologist interpretation of point-of-care handheld 1L ECGs has modest diagnostic sensitivity and specificity with substantial variability for AF classification despite high confidence. Variability in cardiologist interpretation of 1L ECGs highlights the importance of confirmatory testing for diagnosing AF.

An Atrial Fibrillation Classification Method Study Based on BP Neural Network and SVM

Atrial fibrillation is a common arrhythmia, and its diagnosis is interfered by many factors. In order to achieve applicability in diagnosis and improve the level of automatic analysis of atrial fibrillation to the level of experts, the automatic detection of atrial fibrillation is very important. This study proposes an automatic detection algorithm for atrial fibrillation based on BP neural network (back propagation network) and support vector machine (SVM). The electrocardiogram (ECG) segments in the MIT-BIH atrial fibrillation database are divided into 10, 32, 64, and 128 heartbeats, respectively, and the Lorentz value, Shannon entropy, K-S test value and exponential moving average value are calculated. These four characteristic parameters are used as the input of SVM and BP neural network for classification and testing, and the label given by experts in the MIT-BIH atrial fibrillation database is used as the reference output. Among them, the use of atrial fibrillation in the MIT-BIH database, the first 18 cases of data are used as the training set, and the last 7 cases of data are used as the test set. The results show that the accuracy rate of 92% is obtained in the classification of 10 heartbeats, and the accuracy rate of 98% is obtained in the latter three categories. The sensitivity and specificity are both above 97.7%, which has certain applicability. Further validation and improvement in clinical ECG data will be done in next study.

Enhancing Atrial Fibrillation detection accuracy: A wavelet transform filtered single lead ECG signal analysis with artificial neural networks and novel feature extraction

Atrial Fibrillation (AF) is a common cardiac arrhythmia that can lead to fatal outcomes. Detecting AF early is crucial for improving patient outcomes and reducing complications. However, AF is difficult to detect due to asymptomatic cases, false-positive test results, and patient non-adherence. Therefore, it is essential to develop effective methods for early AF detection that can prolong patient life. This study proposes a novel approach for AF detection using a wavelet transform-based filtered ECG signal and a neural network with a novel feature extractor. Firstly, the wavelet transform is applied to filter the ECG signal; secondly, the novel feature extractor is designed to extract important features from the filtered ECG signal. The extracted features are then used as input to a neural network for AF classification. The proposed method is evaluated on two publicly available ECG datasets, and the results show a prominent accuracy of 96% with a ROC is 0.95 and 86% with a ROC is 0.84 for the two datasets, respectively. In addition, to enhance the model’s performance, we have utilized 5-fold cross-validation which achieved an accuracy of 0.99 for Dataset 1 and 0.90 for Dataset 2, respectively. Compared to existing AF detection methods, our approach provides a promising solution that can improve the early diagnosis and treatment of AF. This study presents a new avenue for AF detection and has the potential to advance the field of cardiac arrhythmia detection.

Prognosis of various atrial fibrillation patterns in the general population

Abstract Funding Acknowledgements Type of funding sources: Foundation. Main funding source(s): Senior Scientist Grant from the Dutch Heart Foundation. Aims Due to their dependence on temporality, current atrial fibrillation (AF) classifications are inaccurate and impractical in large cohort studies, resulting in scarce evidence on prognostic differences between AF patterns in the general population. Using our recently developed AF classification, we investigated potential differences in cardiovascular prognosis among participants with various AF patterns and matched non-AF individuals from the general population. Methods 1,111 individuals from the general population were classified as 'single-documented AF episode' (sAF), 198 as 'multiple-documented AF episodes' (mAF), and 52 as 'longstanding persistent AF' (lpAF). Each group was independently matched with non-AF participants through propensity score matching. Participants were followed for the development of coronary heart syndrome (CHD), stroke, heart failure (HF), and cardiovascular mortality. Cox proportional hazards regression analyses were performed to calculate hazard ratios (HRs) and 95% confidence intervals (CIs) for each AF pattern. Results Stroke risk was 1.01; 0.82-1.28 (HR; 95% CI) for sAF, 1.36; 0.87-2.11 for mAF, and 1.44; 0.79-3.54 for lpAF, as compared to their matched participants. HF risk was larger in all AF patterns, with the gradient corresponding to AF severity: 3.48; 1.61-7.55 for lpAF, 2.78; 1.94-3.98 for mAF, and 1.71; 1.44-2.03 for sAF. Compared to non-AF participants, cardiovascular mortality risk was larger in the sAF (1.56; 1.37-1.79) and mAF (1.53; 1.11-2.12) categories, but not for lpAF (0.85; 0.40-1.82). Conclusion More severe AF patterns in the general population carried a larger HF risk. We observed a trend regarding a larger stroke risk in more severe AF patterns, albeit statistically not significant. Our new AF classification method may benefit comorbidity and mortality risk stratification associated with AF in a general population.

A Holistic Overview of Artificial Intelligence in Detection, Classification and Prediction of Atrial Fibrillation Using Electrocardiogram: A Systematic Review and Meta-analysis

A Holistic Overview of Artificial Intelligence in Detection, Classification and Prediction of Atrial Fibrillation Using Electrocardiogram: A Systematic Review and Meta-analysis

PO-01-042 OUTCOMES IN ATRIAL FIBRILLATION USING ADENOSINE TO CHECK FOR DORMANT CONNECTIONS DURING CATHETER ABLATION

Adenosine is used to uncover dormant connections after pulmonary vein isolation (PVI) for atrial fibrillation (AF). There are conflicting reports in the literature on whether this technique reduces AF recurrence. To determine if adenosine provocation (APROV) to check for dormant conduction following PVI reduces AF recurrence. AF ablation procedures performed at a single center from 2013 to 2021 were analyzed. The endpoint was AF recurrence after a 3-month blanking period. AF recurrence was defined as AF lasting for more than 30 seconds. 1:1 propensity score match (PSM) was performed with covariates that have been shown to predict AF recurrence in previous studies (gender, age, BMI, AF classification, obstructive sleep apnea (OSA), hypertension (HTN), and diabetes mellitus (DM). The minimum duration of follow-up was 6 months. Adenosine was used at the procedure physician’s discretion with a minimum dose of 12 mg. Dormant connections were ablated, and persistent PVI with elimination of dormant conduction with adenosine re-challenge was confirmed in all cases. Kaplan-Meier analysis was conducted to assess time to AF recurrence in the entire cohort and the PSM cohort. We analyzed 1544 AF ablations (1111 de novo, 428 repeat ablation) at a single academic medical center. Adenosine was used in 64% of cases. The study population was 65.5% male with a mean age 66 years. Patient co-morbidities included DM (17%), HTN (55%), and OSA (18%). There was no significant difference in time to AF recurrence in the entire cohort with APROV (HR 0.95, CI 0.78-1.16, p=0.6). After PSM, there were 450 patients in each group with AF ablations with 72% paroxysmal AF and 28% persistent AF. There was no difference in time to AF recurrence in the PSM cohort with APROV (HR 0.94, CI 0.75-1.18, p=0.6, Figure 1), with no differences observed between de novo vs repeat ablations. Adenosine provocation is not associated with decreased AF recurrence.