Hybrid machine learning to localize atrial flutter substrates using the surface 12-lead electrocardiogram

ECG Modeling for Simulation of Arrhythmias in Time-Varying Conditions

The present article proposes an ECG simulator that advances modeling of arrhythmias and noise by introducing time-varying signal characteristics. The simulator is built around a discrete-time Markov chain model for simulating atrial and ventricular arrhythmias of particular relevance when analyzing atrial fibrillation (AF). Each state is associated with statistical information on episode duration and heartbeat characteristics. Statistical, time-varying modeling of muscle noise, motion artifacts, and the influence of respiration is introduced to increase the complexity of simulated ECGs, making the simulator well suited for data augmentation in machine learning. Modeling of how the PQ and QT intervals depend on heart rate is also introduced. The realism of simulated ECGs is assessed by three experienced doctors, showing that simulated ECGs are difficult to distinguish from real ECGs. Simulator usefulness is illustrated in terms of AF detection performance when either simulated or real ECGs are used to train a neural network for signal quality control. The results show that both types of training lead to similar performance.

Non-invasive localization of the ventricular excitation origin without patient-specific geometries using deep learning

Cardiovascular diseases account for 17 million deaths per year worldwide. Of these, 25% are categorized as sudden cardiac death, which can be related to ventricular tachycardia (VT). This type of arrhythmia can be caused by focal activation sources outside the sinus node. Catheter ablation of these foci is a curative treatment in order to inactivate the abnormal triggering activity. However, the localization procedure is usually time-consuming and requires an invasive procedure in the catheter lab. To facilitate and expedite the treatment, we present two novel localization support techniques based on convolutional neural networks (CNNs) that address these clinical needs. In contrast to existing methods, our approaches were designed to be independent of the patient-specific geometry and directly applicable to surface ECG signals, while also delivering a binary transmural position. Moreover, one of the method's outputs can be interpreted as several ranked solutions. The CNNs were trained on a dataset containing only simulated data and evaluated both on simulated test data and clinical data. On a novel large and open simulated dataset, the median test error was below 3 mm. The median localization error on the unseen clinical data ranged from 32 mm to 41 mm without optimizing the pre-processing and CNN to the clinical data. Interpreting the output of one of the approaches as ranked solutions, the best median error of the top-3 solutions decreased to 20 mm on the clinical data. The transmural position was correctly detected in up to 82% of all clinical cases. These results demonstrate a proof of principle to utilize CNNs to localize the activation source without the intrinsic need for patient-specific geometrical information. Furthermore, providing multiple solutions can assist physicians in identifying the true activation source amongst more than one possible location. With further optimization to clinical data, these methods have high potential to accelerate clinical interventions, replace certain steps within these procedures and consequently reduce procedural risk and improve VT patient outcomes.

The digital journey: 25 years of digital development in electrophysiology from an Europace perspective

AIMS

Over the past 25 years there has been a substantial development in the field of digital electrophysiology (EP) and in parallel a substantial increase in publications on digital cardiology.In this celebratory paper, we provide an overview of the digital field by highlighting publications from the field focusing on the EP Europace journal.

RESULTS

In this journey across the past quarter of a century we follow the development of digital tools commonly used in the clinic spanning from the initiation of digital clinics through the early days of telemonitoring, to wearables, mobile applications, and the use of fully virtual clinics. We then provide a chronicle of the field of artificial intelligence, a regulatory perspective, and at the end of our journey provide a future outlook for digital EP.

CONCLUSION

Over the past 25 years Europace has published a substantial number of papers on digital EP, with a marked expansion in digital publications in recent years.

MedalCare-XL: 16,900 healthy and pathological synthetic 12 lead ECGs from electrophysiological simulations

Mechanistic cardiac electrophysiology models allow for personalized simulations of the electrical activity in the heart and the ensuing electrocardiogram (ECG) on the body surface. As such, synthetic signals possess known ground truth labels of the underlying disease and can be employed for validation of machine learning ECG analysis tools in addition to clinical signals. Recently, synthetic ECGs were used to enrich sparse clinical data or even replace them completely during training leading to improved performance on real-world clinical test data. We thus generated a novel synthetic database comprising a total of 16,900 12 lead ECGs based on electrophysiological simulations equally distributed into healthy control and 7 pathology classes. The pathological case of myocardial infraction had 6 sub-classes. A comparison of extracted features between the virtual cohort and a publicly available clinical ECG database demonstrated that the synthetic signals represent clinical ECGs for healthy and pathological subpopulations with high fidelity. The ECG database is split into training, validation, and test folds for development and objective assessment of novel machine learning algorithms.

Feasibility of three-dimensional artificial intelligence algorithm integration with intracardiac echocardiography for left atrial imaging during atrial fibrillation catheter ablation

AIMS

Intracardiac echocardiography (ICE) is a useful but operator-dependent tool for left atrial (LA) anatomical rendering during atrial fibrillation (AF) ablation. The CARTOSOUND FAM Module, a new deep learning (DL) imaging algorithm, has the potential to overcome this limitation. This study aims to evaluate feasibility of the algorithm compared to cardiac computed tomography (CT) in patients undergoing AF ablation.

METHODS AND RESULTS

In 28 patients undergoing AF ablation, baseline patient information was recorded, and three-dimensional (3D) shells of LA body and anatomical structures [LA appendage/left superior pulmonary vein/left inferior pulmonary vein/right superior pulmonary vein/right inferior pulmonary vein (RIPV)] were reconstructed using the DL algorithm. The selected ultrasound frames were gated to end-expiration and max LA volume. Ostial diameters of these structures and carina-to-carina distance between left and right pulmonary veins were measured and compared with CT measurements. Anatomical accuracy of the DL algorithm was evaluated by three independent electrophysiologists using a three-anchor scale for LA anatomical structures and a five-anchor scale for LA body. Ablation-related characteristics were summarized. The algorithm generated 3D reconstruction of LA anatomies, and two-dimensional contours overlaid on ultrasound input frames. Average calculation time for LA reconstruction was 65 s. Mean ostial diameters and carina-to-carina distance were all comparable to CT without statistical significance. Ostial diameters and carina-to-carina distance also showed moderate to high correlation (r = 0.52-0.75) except for RIPV (r = 0.20). Qualitative ratings showed good agreement without between-rater differences. Average procedure time was 143.7 ± 43.7 min, with average radiofrequency time 31.6 ± 10.2 min. All patients achieved ablation success, and no immediate complications were observed.

CONCLUSION

DL algorithm integration with ICE demonstrated considerable accuracy compared to CT and qualitative physician assessment. The feasibility of ICE with this algorithm can potentially further streamline AF ablation workflow.

Detailed analysis of tachycardia cycle length aids diagnosis of the mechanism and location of atrial tachycardias

AIMS

Although the mechanism of an atrial tachycardia (AT) can usually be elucidated using modern high-resolution mapping systems, it would be helpful if the AT mechanism and circuit could be predicted before initiating mapping.

OBJECTIVE

We examined if the information gathered from the cycle length (CL) of the tachycardia can help predict the AT-mechanism and its localization.

METHODS

One hundred and thirty-eight activation maps of ATs including eight focal-ATs, 94 macroreentrant-ATs, and 36 localized-ATs in 95 patients were retrospectively reviewed. Maximal CL (MCL) and minimal CL (mCL) over a minute period were measured via a decapolar catheter in the coronary sinus. CL-variation and beat-by-beat CL-alternation were examined. Additionally, the CL-respiration correlation was analysed by the RhythmiaTM system. : Both MCL and mCL were significantly shorter in macroreentrant-ATs [MCL = 288 (253-348) ms, P = 0.0001; mCL = 283 (243-341) ms, P = 0.0012], and also shorter in localized-ATs [MCL = 314 (261-349) ms, P = 0.0016; mCL = 295 (248-340) ms, P = 0.0047] compared to focal-ATs [MCL = 506 (421-555) ms, mCL = 427 (347-508) ms]. An absolute CL-variation (MCL-mCL) < 24 ms significantly differentiated re-entrant ATs from focal-ATs with a sensitivity = 96.9%, specificity = 100%, positive predictive value (PPV) = 100%, and negative predictive value (NPV) = 66.7%. The beat-by-beat CL-alternation was observed in 10/138 (7.2%), all of which showed the re-entrant mechanism, meaning that beat-by-beat CL-alternation was the strong sign of re-entrant mechanism (PPV = 100%). Although the CL-respiration correlation was observed in 28/138 (20.3%) of ATs, this was predominantly in right-atrium (RA)-ATs (24/41, 85.7%), rather than left atrium (LA)-ATs (4/97, 4.1%). A positive CL-respiration correlation highly predicted RA-ATs (PPV = 85.7%), and negative CL-respiration correlation probably suggested LA-ATs (NPV = 84.5%).

CONCLUSION

Detailed analysis of the tachycardia CL helps predict the AT-mechanism and the active AT chamber before an initial mapping.

Identification of supraventricular tachycardia mechanisms with surface electrocardiograms using a convolutional neural network

Background

It remains difficult to definitively distinguish supraventricular tachycardia (SVT) mechanisms using a 12-lead electrocardiogram (ECG) alone. Machine learning may identify visually imperceptible changes on 12-lead ECGs and may improve ability to determine SVT mechanisms.

Objective

We sought to develop a convolutional neural network (CNN) that identifies the SVT mechanism according to the gold standard of SVT ablation and to compare CNN performance against experienced electrophysiologists among patients with atrioventricular nodal re-entrant tachycardia (AVNRT), atrioventricular reciprocating tachycardia (AVRT), and atrial tachycardia (AT).

Methods

All patients with 12-lead surface ECG during sinus rhythm and SVT and had successful SVT ablation from 2013 to 2020 were included. A CNN was trained using data from 1505 surface ECGs that were split into 1287 training and 218 test ECG datasets. We compared the CNN performance against independent adjudication by 2 experienced cardiac electrophysiologists on the test dataset.

Results

Our dataset comprised 1505 ECGs (368 AVNRT, 304 AVRT, 95 AT, and 738 sinus rhythm) from 725 patients. The CNN areas under the receiver-operating characteristic curve for AVNRT, AVRT, and AT were 0.909, 0.867, and 0.817, respectively. When fixing the specificity of the CNN to the electrophysiologist adjudicators' specificity, the CNN identified all SVT classes with higher sensitivity: (1) AVNRT (91.7% vs 65.9%), (2) AVRT (78.4% vs 63.6%), and (3) AT (61.5% vs 50.0%).

Conclusion

A CNN can be trained to differentiate SVT mechanisms from surface 12-lead ECGs with high overall performance, achieving similar performance to experienced electrophysiologists at fixed specificities.

The Use of Artificial Intelligence for Detecting and Predicting Atrial Arrhythmias Post Catheter Ablation

Catheter ablation (CA) is considered as one of the most effective methods technique for eradicating persistent and abnormal cardiac arrhythmias. Nevertheless, in some cases, these arrhythmias are not treated properly, resulting in their recurrences. If left untreated, they may result in complications such as strokes, heart failure, or death. Until recently, the primary techniques for diagnosing recurrent arrhythmias following CA were the findings predisposing to the changes caused by the arrhythmias on cardiac imaging and electrocardiograms during follow-up visits, or if patients reported having palpitations or chest discomfort after the ablation. However, these follow-ups may be time-consuming and costly, and they may not always determine the root cause of the recurrences. With the introduction of artificial intelligence (AI), these follow-up visits can be effectively shortened, and improved methods for predicting the likelihood of recurring arrhythmias after their ablation procedures can be developed. AI can be divided into two categories: machine learning (ML) and deep learning (DL), the latter of which is a subset of ML. ML and DL models have been used in several studies to demonstrate their ability to predict and identify cardiac arrhythmias using clinical variables, electrophysiological characteristics, and trends extracted from imaging data. AI has proven to be a valuable aid for cardiologists due to its ability to compute massive amounts of data and detect subtle changes in electric signals and cardiac images, which may potentially increase the risk of recurrent arrhythmias after CA. Despite the fact that these studies involving AI have generated promising outcomes comparable to or superior to human intervention, they have primarily focused on atrial fibrillation while atrial flutter (AFL) and atrial tachycardia (AT) were the subjects of relatively few AI studies. Therefore, the aim of this review is to investigate the interaction of AI algorithms, electrophysiological characteristics, imaging data, risk score calculators, and clinical variables in predicting cardiac arrhythmias following an ablation procedure. This review will also discuss the implementation of these algorithms to enable the detection and prediction of AFL and AT recurrences following CA.

Artificial intelligence-enabled electrocardiogram to distinguish cavotricuspid isthmus dependence from other atrial tachycardia mechanisms. EUROPEAN HEART JOURNAL. DIGITAL HEALTH 2022;3:405-414. [PMID: 36712163 PMCID: PMC9708023 DOI: 10.1093/ehjdh/ztac042]

Aims

Accurately determining atrial arrhythmia mechanisms from a 12-lead electrocardiogram (ECG) can be challenging. Given the high success rate of cavotricuspid isthmus (CTI) ablation, identification of CTI-dependent typical atrial flutter (AFL) is important for treatment decisions and procedure planning. We sought to train a convolutional neural network (CNN) to classify CTI-dependent AFL vs. non-CTI dependent atrial tachycardia (AT), using data from the invasive electrophysiology (EP) study as the gold standard.

Methods and results

We trained a CNN on data from 231 patients undergoing EP studies for atrial tachyarrhythmia. A total of 13 500 five-second 12-lead ECG segments were used for training. Each case was labelled CTI-dependent AFL or non-CTI-dependent AT based on the findings of the EP study. The model performance was evaluated against a test set of 57 patients. A survey of electrophysiologists in Europe was undertaken on the same 57 ECGs. The model had an accuracy of 86% (95% CI 0.77-0.95) compared to median expert electrophysiologist accuracy of 79% (range 70-84%). In the two thirds of test set cases (38/57) where both the model and electrophysiologist consensus were in agreement, the prediction accuracy was 100%. Saliency mapping demonstrated atrial activation was the most important segment of the ECG for determining model output.

Conclusion

We describe the first CNN trained to differentiate CTI-dependent AFL from other AT using the ECG. Our model matched and complemented expert electrophysiologist performance. Automated artificial intelligence-enhanced ECG analysis could help guide treatment decisions and plan ablation procedures for patients with organized atrial arrhythmias.