Fernando Soares Schlindwein
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Fernando Soares Schlindwein has a total of 320 co-authors. Here are the top co-authors by number of shared publications:
- Ghulam AndrÉ Ng
Department of Cardiovascular Sciences, Clinical Science Wing, Glenfield Hospital, University of Leicester, Leicester LE3 9QP, UK
41 shared publications - Xin Li
Department of Engineering, University of Leicester, Leicester LE1 7RH, UK
35 shared publications - Frederique Jos Vanheusden
School of Science & Technology, Nottingham Trent University, United Kingdom
11 shared publications - João Salinet
HeartLab, Federal University of ABC, São Bernardo do Campo, São Paulo, Brazil
9 shared publications Peter J Stafford
8 shared publicationsPeter J. Stafford
University Hospitals of Leicester NHS Trust, U.K.
8 shared publications
Latest Publications
- Open Access
Compressed Deep Learning Models for Wearable Atrial Fibrillation Detection through Attention
Deep learning (DL) models have shown promise for the accurate detection of atrial fibrillation (AF) from electrocardiogram/photoplethysmography (ECG/PPG) data, yet deploying these on resource-constrained wearable devices remains challenging. This study proposes integrating a customized channel attention mechanism to compress DL neural networks for AF detection, allowing the model to focus only on the most salient time-series features. The results demonstrate that applying compression through channel attention significantly reduces the total number of model parameters and file size while minimizing loss in detection accuracy. Notably, after compression, performance increases for certain model variants in key AF databases (ADB and C2017DB). Moreover, analyzing the learned channel attention distributions after training enhances the explainability of the AF detection models by highlighting the salient temporal ECG/PPG features most important for its diagnosis. Overall, this research establishes that integrating attention mechanisms is an effective strategy for compressing large DL models, making them deployable on low-power wearable devices. We show that this approach yields compressed, accurate, and explainable AF detectors ideal for wearables. Incorporating channel attention enables simpler yet more accurate algorithms that have the potential to provide clinicians with valuable insights into the salient temporal biomarkers of AF. Our findings highlight that the use of attention is an important direction for the future development of efficient, high-performing, and interpretable AF screening tools for wearable technology.
Open AccessCompressed Deep Learning Models for Wearable Atrial Fibrillation Detection through Attention
Deep learning (DL) models have shown promise for the accurate detection of atrial fibrillation (AF) from electrocardiogram/photoplethysmography (ECG/PPG) data, yet deploying these on resource-constrained wearable devices remains challenging. This study proposes integrating a customized channel attention mechanism to compress DL neural networks for AF detection, allowing the model to focus only on the most salient time-series features. The results demonstrate that applying compression through channel attention significantly reduces the total number of model parameters and file size while minimizing loss in detection accuracy. Notably, after compression, performance increases for certain model variants in key AF databases (ADB and C2017DB). Moreover, analyzing the learned channel attention distributions after training enhances the explainability of the AF detection models by highlighting the salient temporal ECG/PPG features most important for its diagnosis. Overall, this research establishes that integrating attention mechanisms is an effective strategy for compressing large DL models, making them deployable on low-power wearable devices. We show that this approach yields compressed, accurate, and explainable AF detectors ideal for wearables. Incorporating channel attention enables simpler yet more accurate algorithms that have the potential to provide clinicians with valuable insights into the salient temporal biomarkers of AF. Our findings highlight that the use of attention is an important direction for the future development of efficient, high-performing, and interpretable AF screening tools for wearable technology.
- Open Access
- 12 March 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Biomedical Engineering
- Vol. 71 (7) , 2143-2153
ECGVEDNET: A Variational Encoder-Decoder Network for ECG Delineation in Morphology Variant ECGs
Electrocardiogram (ECG) delineation to identify the fiducial points of ECG segments, plays an important role in cardiovascular diagnosis and care. Whilst deep delineation frameworks have been deployed within the literature, several factors still hinder their development: (a) data availability: the capacity of deep learning models to generalise is limited by the amount of available data; (b) morphology variations: ECG complexes vary, even within the same person, which degrades the performance of conventional deep learning models. To address these concerns, we present a large-scale 12-leads ECG dataset, ICDIRS, to train and evaluate a novel deep delineation model-ECGVEDNET. ICDIRS is a large-scale ECG dataset with 156,145 QRS onset annotations and 156,145 T peak annotations. ECGVEDNET is a novel variational encoder-decoder network designed to address morphology variations. In ECGVEDNET, we construct a well-regularized latent space, in which the latent features of ECG follow a regular distribution and present smaller morphology variations than in the raw data space. Finally, a transfer learning framework is proposed to transfer the knowledge learned on ICDIRS to smaller datasets. On ICDIRS, ECGVEDNET achieves accuracy of 86.28%/88.31% within 5/10 ms tolerance for QRS onset and accuracy of 89.94%/91.16% within 5/10 ms tolerance for T peak. On QTDB, the average time errors computed for QRS onset and T peak are −1.86 $\pm$ 8.02 ms and −0.50 $\pm$ 12.96 ms, respectively, achieving state-of-the-art performances on both large and small-scale datasets. We will release the source code and the pre-trained model on ICDIRS once accepted.
- 12 March 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Biomedical Engineering
- Vol. 71 (7) , 2143-2153
Open AccessECGVEDNET: A Variational Encoder-Decoder Network for ECG Delineation in Morphology Variant ECGs
Electrocardiogram (ECG) delineation to identify the fiducial points of ECG segments, plays an important role in cardiovascular diagnosis and care. Whilst deep delineation frameworks have been deployed within the literature, several factors still hinder their development: (a) data availability: the capacity of deep learning models to generalise is limited by the amount of available data; (b) morphology variations: ECG complexes vary, even within the same person, which degrades the performance of conventional deep learning models. To address these concerns, we present a large-scale 12-leads ECG dataset, ICDIRS, to train and evaluate a novel deep delineation model-ECGVEDNET. ICDIRS is a large-scale ECG dataset with 156,145 QRS onset annotations and 156,145 T peak annotations. ECGVEDNET is a novel variational encoder-decoder network designed to address morphology variations. In ECGVEDNET, we construct a well-regularized latent space, in which the latent features of ECG follow a regular distribution and present smaller morphology variations than in the raw data space. Finally, a transfer learning framework is proposed to transfer the knowledge learned on ICDIRS to smaller datasets. On ICDIRS, ECGVEDNET achieves accuracy of 86.28%/88.31% within 5/10 ms tolerance for QRS onset and accuracy of 89.94%/91.16% within 5/10 ms tolerance for T peak. On QTDB, the average time errors computed for QRS onset and T peak are −1.86 $\pm$ 8.02 ms and −0.50 $\pm$ 12.96 ms, respectively, achieving state-of-the-art performances on both large and small-scale datasets. We will release the source code and the pre-trained model on ICDIRS once accepted.
- Open Access
- 30 January 2024
- Published by Oxford University Press (OUP) in European Heart Journal - Digital Health
- Vol. 5 (3) , 384-388
Artificial intelligence for ventricular arrhythmia capability using ambulatory electrocardiograms
European and American clinical guidelines for implantable cardioverter defibrillators are insufficiently accurate for ventricular arrhythmia (VA) risk stratification, leading to significant morbidity and mortality. Artificial intelligence offers a novel risk stratification lens through which VA capability can be determined from the electrocardiogram (ECG) in normal cardiac rhythm. The aim of this study was to develop and test a deep neural network for VA risk stratification using routinely collected ambulatory ECGs. A multicentre case–control study was undertaken to assess VA-ResNet-50, our open source ResNet-50-based deep neural network. VA-ResNet-50 was designed to read pyramid samples of three-lead 24 h ambulatory ECGs to decide whether a heart is capable of VA based on the ECG alone. Consecutive adults with VA from East Midlands, UK, who had ambulatory ECGs as part of their NHS care between 2014 and 2022 were recruited and compared with all comer ambulatory electrograms without VA. Of 270 patients, 159 heterogeneous patients had a composite VA outcome. The mean time difference between the ECG and VA was 1.6 years (⅓ ambulatory ECG before VA). The deep neural network was able to classify ECGs for VA capability with an accuracy of 0.76 (95% confidence interval 0.66–0.87), F1 score of 0.79 (0.67–0.90), area under the receiver operator curve of 0.8 (0.67–0.91), and relative risk of 2.87 (1.41–5.81). Ambulatory ECGs confer risk signals for VA risk stratification when analysed using VA-ResNet-50. Pyramid sampling from the ambulatory ECGs is hypothesized to capture autonomic activity. We encourage groups to build on this open-source model. Can artificial intelligence (AI) be used to predict whether a person is at risk of a lethal heart rhythm, based solely on an electrocardiogram (an electrical heart tracing)? In a study of 270 adults (of which 159 had lethal arrhythmias), the AI was correct in 4 out of every 5 cases. If the AI said a person was at risk, the risk of lethal event was three times higher than normal adults. In this study, the AI performed better than current medical guidelines. The AI was able to accurately determine the risk of lethal arrhythmia from standard heart tracings for 80% of cases over a year away—a conceptual shift in what an AI model can see and predict. This method shows promise in better allocating implantable shock box pacemakers (implantable cardioverter defibrillators) that save lives.
- 30 January 2024
- Published by Oxford University Press (OUP) in European Heart Journal - Digital Health
- Vol. 5 (3) , 384-388
Open AccessArtificial intelligence for ventricular arrhythmia capability using ambulatory electrocardiograms
European and American clinical guidelines for implantable cardioverter defibrillators are insufficiently accurate for ventricular arrhythmia (VA) risk stratification, leading to significant morbidity and mortality. Artificial intelligence offers a novel risk stratification lens through which VA capability can be determined from the electrocardiogram (ECG) in normal cardiac rhythm. The aim of this study was to develop and test a deep neural network for VA risk stratification using routinely collected ambulatory ECGs. A multicentre case–control study was undertaken to assess VA-ResNet-50, our open source ResNet-50-based deep neural network. VA-ResNet-50 was designed to read pyramid samples of three-lead 24 h ambulatory ECGs to decide whether a heart is capable of VA based on the ECG alone. Consecutive adults with VA from East Midlands, UK, who had ambulatory ECGs as part of their NHS care between 2014 and 2022 were recruited and compared with all comer ambulatory electrograms without VA. Of 270 patients, 159 heterogeneous patients had a composite VA outcome. The mean time difference between the ECG and VA was 1.6 years (⅓ ambulatory ECG before VA). The deep neural network was able to classify ECGs for VA capability with an accuracy of 0.76 (95% confidence interval 0.66–0.87), F1 score of 0.79 (0.67–0.90), area under the receiver operator curve of 0.8 (0.67–0.91), and relative risk of 2.87 (1.41–5.81). Ambulatory ECGs confer risk signals for VA risk stratification when analysed using VA-ResNet-50. Pyramid sampling from the ambulatory ECGs is hypothesized to capture autonomic activity. We encourage groups to build on this open-source model. Can artificial intelligence (AI) be used to predict whether a person is at risk of a lethal heart rhythm, based solely on an electrocardiogram (an electrical heart tracing)? In a study of 270 adults (of which 159 had lethal arrhythmias), the AI was correct in 4 out of every 5 cases. If the AI said a person was at risk, the risk of lethal event was three times higher than normal adults. In this study, the AI performed better than current medical guidelines. The AI was able to accurately determine the risk of lethal arrhythmia from standard heart tracings for 80% of cases over a year away—a conceptual shift in what an AI model can see and predict. This method shows promise in better allocating implantable shock box pacemakers (implantable cardioverter defibrillators) that save lives.
- Open Access
Correction: Mäkynen et al. Wearable Devices Combined with Artificial Intelligence—A Future Technology for Atrial Fibrillation Detection? Sensors 2022, 22, 8588
The authors wish to add two authors to the original paper
Open AccessCorrection: Mäkynen et al. Wearable Devices Combined with Artificial Intelligence—A Future Technology for Atrial Fibrillation Detection? Sensors 2022, 22, 8588
The authors wish to add two authors to the original paper
- Open Access
Wearable Devices Combined with Artificial Intelligence—A Future Technology for Atrial Fibrillation Detection?
Atrial fibrillation (AF) is the most common cardiac arrhythmia in the world. The arrhythmia and methods developed to cure it have been studied for several decades. However, professionals worldwide are still working to improve treatment quality. One novel technology that can be useful is a wearable device. The two most used recordings from these devices are photoplethysmogram (PPG) and electrocardiogram (ECG) signals. As the price lowers, these devices will become significant technology to increase sensitivity, for monitoring and for treatment quality support. This is important as AF can be challenging to detect in advance, especially during home monitoring. Modern artificial intelligence (AI) has the potential to respond to this challenge. AI has already achieved state of the art results in many applications, including bioengineering. In this perspective, we discuss wearable devices combined with AI for AF detection, an approach that enables a new era of possibilities for the future.
Open AccessWearable Devices Combined with Artificial Intelligence—A Future Technology for Atrial Fibrillation Detection?
Atrial fibrillation (AF) is the most common cardiac arrhythmia in the world. The arrhythmia and methods developed to cure it have been studied for several decades. However, professionals worldwide are still working to improve treatment quality. One novel technology that can be useful is a wearable device. The two most used recordings from these devices are photoplethysmogram (PPG) and electrocardiogram (ECG) signals. As the price lowers, these devices will become significant technology to increase sensitivity, for monitoring and for treatment quality support. This is important as AF can be challenging to detect in advance, especially during home monitoring. Modern artificial intelligence (AI) has the potential to respond to this challenge. AI has already achieved state of the art results in many applications, including bioengineering. In this perspective, we discuss wearable devices combined with AI for AF detection, an approach that enables a new era of possibilities for the future.
- 25 August 2022
- Published by Oxford University Press (OUP) in EP Europace
- Vol. 24 (11) , 1777-1787
Machine learning in sudden cardiac death risk prediction: a systematic review
Most patients who receive implantable cardioverter defibrillators (ICDs) for primary prevention do not receive therapy during the lifespan of the ICD, whilst up to 50% of sudden cardiac death (SCD) occur in individuals who are considered low risk by conventional criteria. Machine learning offers a novel approach to risk stratification for ICD assignment. Systematic search was performed in MEDLINE, Embase, Emcare, CINAHL, Cochrane Library, OpenGrey, MedrXiv, arXiv, Scopus, and Web of Science. Studies modelling SCD risk prediction within days to years using machine learning were eligible for inclusion. Transparency and quality of reporting (TRIPOD) and risk of bias (PROBAST) were assessed. A total of 4356 studies were screened with 11 meeting the inclusion criteria with heterogeneous populations, methods, and outcome measures preventing meta-analysis. The study size ranged from 122 to 124 097 participants. Input data sources included demographic, clinical, electrocardiogram, electrophysiological, imaging, and genetic data ranging from 4 to 72 variables per model. The most common outcome metric reported was the area under the receiver operator characteristic (n = 7) ranging between 0.71 and 0.96. In six studies comparing machine learning models and regression, machine learning improved performance in five. No studies adhered to a reporting standard. Five of the papers were at high risk of bias. Machine learning for SCD prediction has been under-applied and incorrectly implemented but is ripe for future investigation. It may have some incremental utility in predicting SCD over traditional models. The development of reporting standards for machine learning is required to improve the quality of evidence reporting in the field.
- 25 August 2022
- Published by Oxford University Press (OUP) in EP Europace
- Vol. 24 (11) , 1777-1787
Machine learning in sudden cardiac death risk prediction: a systematic review
Most patients who receive implantable cardioverter defibrillators (ICDs) for primary prevention do not receive therapy during the lifespan of the ICD, whilst up to 50% of sudden cardiac death (SCD) occur in individuals who are considered low risk by conventional criteria. Machine learning offers a novel approach to risk stratification for ICD assignment. Systematic search was performed in MEDLINE, Embase, Emcare, CINAHL, Cochrane Library, OpenGrey, MedrXiv, arXiv, Scopus, and Web of Science. Studies modelling SCD risk prediction within days to years using machine learning were eligible for inclusion. Transparency and quality of reporting (TRIPOD) and risk of bias (PROBAST) were assessed. A total of 4356 studies were screened with 11 meeting the inclusion criteria with heterogeneous populations, methods, and outcome measures preventing meta-analysis. The study size ranged from 122 to 124 097 participants. Input data sources included demographic, clinical, electrocardiogram, electrophysiological, imaging, and genetic data ranging from 4 to 72 variables per model. The most common outcome metric reported was the area under the receiver operator characteristic (n = 7) ranging between 0.71 and 0.96. In six studies comparing machine learning models and regression, machine learning improved performance in five. No studies adhered to a reporting standard. Five of the papers were at high risk of bias. Machine learning for SCD prediction has been under-applied and incorrectly implemented but is ripe for future investigation. It may have some incremental utility in predicting SCD over traditional models. The development of reporting standards for machine learning is required to improve the quality of evidence reporting in the field.
- Open Access
Rationale and study design of the MINERVA study: Multicentre Investigation of Novel Electrocardiogram Risk markers in Ventricular Arrhythmia prediction—UK multicentre collaboration
Introduction The purpose of this study is to assess the ability of two new ECG markers (Regional Repolarisation Instability Index (R2I2) and Peak Electrical Restitution Slope) to predict sudden cardiac death (SCD) or ventricular arrhythmia (VA) events in patients with ischaemic cardiomyopathy undergoing implantation of an implantable cardioverter defibrillator for primary prevention indication. Methods and analysis Multicentre Investigation of Novel Electrocardiogram Risk markers in Ventricular Arrhythmia prediction is a prospective, open label, single blinded, multicentre observational study to establish the efficacy of two ECG biomarkers in predicting VA risk. 440 participants with ischaemic cardiomyopathy undergoing routine first time implantable cardioverter-defibrillator (ICD) implantation for primary prevention indication are currently being recruited. An electrophysiological (EP) study is performed using a non-invasive programmed electrical stimulation protocol via the implanted device. All participants will undergo the EP study hence no randomisation is required. Participants will be followed up over a minimum of 18 months and up to 3 years. The first patient was recruited in August 2016 and the study will be completed at the final participant follow-up visit. The primary endpoint is ventricular fibrillation or sustained ventricular tachycardia >200 beats/min as recorded by the ICD. The secondary endpoint is SCD. Analysis of the ECG data obtained during the EP study will be performed by the core lab where blinding of patient health status and endpoints will be maintained. Ethics and dissemination Ethical approval has been granted by Research Ethics Committees Northern Ireland (reference no. 16/NI/0069). The results will inform the design of a definitive Randomised Controlled Trial (RCT). Dissemination will include peer reviewed journal articles reporting the qualitative and quantitative results, as well as presentations at conferences and lay summaries. Trial registration number NCT03022487.
Open AccessRationale and study design of the MINERVA study: Multicentre Investigation of Novel Electrocardiogram Risk markers in Ventricular Arrhythmia prediction—UK multicentre collaboration
Introduction The purpose of this study is to assess the ability of two new ECG markers (Regional Repolarisation Instability Index (R2I2) and Peak Electrical Restitution Slope) to predict sudden cardiac death (SCD) or ventricular arrhythmia (VA) events in patients with ischaemic cardiomyopathy undergoing implantation of an implantable cardioverter defibrillator for primary prevention indication. Methods and analysis Multicentre Investigation of Novel Electrocardiogram Risk markers in Ventricular Arrhythmia prediction is a prospective, open label, single blinded, multicentre observational study to establish the efficacy of two ECG biomarkers in predicting VA risk. 440 participants with ischaemic cardiomyopathy undergoing routine first time implantable cardioverter-defibrillator (ICD) implantation for primary prevention indication are currently being recruited. An electrophysiological (EP) study is performed using a non-invasive programmed electrical stimulation protocol via the implanted device. All participants will undergo the EP study hence no randomisation is required. Participants will be followed up over a minimum of 18 months and up to 3 years. The first patient was recruited in August 2016 and the study will be completed at the final participant follow-up visit. The primary endpoint is ventricular fibrillation or sustained ventricular tachycardia >200 beats/min as recorded by the ICD. The secondary endpoint is SCD. Analysis of the ECG data obtained during the EP study will be performed by the core lab where blinding of patient health status and endpoints will be maintained. Ethics and dissemination Ethical approval has been granted by Research Ethics Committees Northern Ireland (reference no. 16/NI/0069). The results will inform the design of a definitive Randomised Controlled Trial (RCT). Dissemination will include peer reviewed journal articles reporting the qualitative and quantitative results, as well as presentations at conferences and lay summaries. Trial registration number NCT03022487.
- 13 September 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 40, 1-4
Spatio-Temporal ECG Network for Detecting Cardiac Disorders from Multi-Lead ECGs
Automatic detection and classification of cardiac disorders play a critical role in the analysis of clinical electrocardiogram (ECG). Deep learning methods are effective for automated feature extraction and have shown promising results in ECG classification. In this work, we proposed a deep spatio-temporal ECG network (ST-ECGNet) to extract robust spatio-temporal features for detecting multiple cardiac disorders from the multi-lead ECG data. The proposed ST-ECGNet combines a Convolutional Neural Network (CNN) module for extracting local spatial features, an attention module for capturing global spatial features, and a Bi-directional Gated Recurrent Unit (Bi-GRU) module for extracting temporal features from ECG data. Specifically, the attention mechanism enables our deep learning architecture to focus on the most important and useful parts of the input to make more accurate predictions. In PhysioNet/Computing in Cardiology Challenge 2021, our entry was not officially ranked and scored on the test data of the Challenge, because our code was not successfully processed during the official phase and failed to run with errors.
- 13 September 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 40, 1-4
Spatio-Temporal ECG Network for Detecting Cardiac Disorders from Multi-Lead ECGs
Automatic detection and classification of cardiac disorders play a critical role in the analysis of clinical electrocardiogram (ECG). Deep learning methods are effective for automated feature extraction and have shown promising results in ECG classification. In this work, we proposed a deep spatio-temporal ECG network (ST-ECGNet) to extract robust spatio-temporal features for detecting multiple cardiac disorders from the multi-lead ECG data. The proposed ST-ECGNet combines a Convolutional Neural Network (CNN) module for extracting local spatial features, an attention module for capturing global spatial features, and a Bi-directional Gated Recurrent Unit (Bi-GRU) module for extracting temporal features from ECG data. Specifically, the attention mechanism enables our deep learning architecture to focus on the most important and useful parts of the input to make more accurate predictions. In PhysioNet/Computing in Cardiology Challenge 2021, our entry was not officially ranked and scored on the test data of the Challenge, because our code was not successfully processed during the official phase and failed to run with errors.
- Open Access
- 30 December 2020
- Published by Computing in Cardiology
Optimizing Atrial Electrogram Classification Based on Local Ablation Outcome in Human Atrial Fibrillation
- 30 December 2020
- Published by Computing in Cardiology
Open AccessOptimizing Atrial Electrogram Classification Based on Local Ablation Outcome in Human Atrial Fibrillation
- Open Access
- 3 September 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Biomedical Engineering
- Vol. 68 (4) , 1131-1141
Unsupervised Classification of Atrial Electrograms for Electroanatomic Mapping of Human Persistent Atrial Fibrillation
Objective: Ablation treatment for persistent atrial fibrillation (persAF) remains challenging due to the absence of a ‘ground truth’ for atrial substrate characterization and the presence of multiple mechanisms driving the arrhythmia. We implemented an unsupervised classification to identify clusters of atrial electrograms (AEGs) with similar patterns, which were then validated by AEG-derived markers. Methods : 956 bipolar AEGs were collected from 11 persAF patients. CARTO variables (Biosense Webster; ICL, ACI and SCI) were used to create a 3D space, and subsequently used to perform an unsupervised classification with k-means. The characteristics of the identified groups were investigated using nine AEG-derived markers: sample entropy (SampEn), dominant frequency, organization index (OI), determinism, laminarity, recurrence rate (RR), peak-to-peak (PP) amplitude, cycle length (CL), and wave similarity (WS). Results : Five AEG classes with distinct characteristics were identified (F = 582, P<0.0001). The presence of fractionation increased from class 1 to 5, as reflected by the nine markers. Class 1 (25%) included organized AEGs with high WS, determinism, laminarity, and RR, and low SampEn. Class 5 (20%) comprised fractionated AEGs with in low WS, OI, determinism, laminarity, and RR, and in high SampEn. Classes 2 (12%), 3 (13%) and 4 (30%) suggested different degrees of AEG organization. Conclusions : Our results expand and reinterpret the criteria used for automated AEG classification. The nine markers highlighted electrophysiological differences among the five classes found by the k-means, which could provide a more complete characterization of persAF substrate during ablation target identification in future clinical studies.
- 3 September 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Biomedical Engineering
- Vol. 68 (4) , 1131-1141
Open AccessUnsupervised Classification of Atrial Electrograms for Electroanatomic Mapping of Human Persistent Atrial Fibrillation
Objective: Ablation treatment for persistent atrial fibrillation (persAF) remains challenging due to the absence of a ‘ground truth’ for atrial substrate characterization and the presence of multiple mechanisms driving the arrhythmia. We implemented an unsupervised classification to identify clusters of atrial electrograms (AEGs) with similar patterns, which were then validated by AEG-derived markers. Methods : 956 bipolar AEGs were collected from 11 persAF patients. CARTO variables (Biosense Webster; ICL, ACI and SCI) were used to create a 3D space, and subsequently used to perform an unsupervised classification with k-means. The characteristics of the identified groups were investigated using nine AEG-derived markers: sample entropy (SampEn), dominant frequency, organization index (OI), determinism, laminarity, recurrence rate (RR), peak-to-peak (PP) amplitude, cycle length (CL), and wave similarity (WS). Results : Five AEG classes with distinct characteristics were identified (F = 582, P<0.0001). The presence of fractionation increased from class 1 to 5, as reflected by the nine markers. Class 1 (25%) included organized AEGs with high WS, determinism, laminarity, and RR, and low SampEn. Class 5 (20%) comprised fractionated AEGs with in low WS, OI, determinism, laminarity, and RR, and in high SampEn. Classes 2 (12%), 3 (13%) and 4 (30%) suggested different degrees of AEG organization. Conclusions : Our results expand and reinterpret the criteria used for automated AEG classification. The nine markers highlighted electrophysiological differences among the five classes found by the k-means, which could provide a more complete characterization of persAF substrate during ablation target identification in future clinical studies.