Global Longitudinal Active Strain Energy Density (GLASED): A Powerful Prognostic Marker in a Community-Based Cohort.

BACKGROUND: Identifying the imaging method that best predicts all-cause mortality, cardiovascular adverse events and heart failure risk is crucial for tailoring optimal management. Potential prognostic markers include left ventricular myocardial mass, ejection fraction, myocardial strain, stroke work, contraction fraction, pressure-strain product and a new measurement called global longitudinal active strain density (GLASED). OBJECTIVES: This study sought to compare the utility of 23 potential left ventricular prognostic markers of structure and contractile function in a community-based cohort. METHODS: The impact of cardiovascular magnetic resonance image-derived markers extracted by machine learning algorithms was compared to the future risk of adverse events in a group of 44,957 UK Biobank participants. RESULTS: Most markers, including the left ventricular ejection fraction, have limited prognostic value. GLASED was significantly associated with all-cause mortality and major adverse cardiovascular events, with the largest hazard ratio, highest ranking and differentiated risk in all three tertiles (P ≤ 0.0003). CONCLUSIONS: GLASED predicted all-cause mortality and major cardiovascular adverse events better than conventional markers of risk and is recommended for assessing patient prognosis.

Quantifying myocardial active strain energy density: A comparative analysis of analytic and finite element methods for estimating left ventricular wall stress and strain.

AIMS: This study compared commonly used methods for calculating left ventricular wall stress with the finite element analysis and evaluated different approaches to strain estimation. We sought to improve the accuracy of contractance estimation by developing a novel stress equation. BACKGROUND: Multiple methods for calculating LV contractile stress and strain exist. Contractance is derived from stress and strain information and is a measure of myocardial work per unit volume of muscle. Precise stress and strain information are essential for its accurate evaluation. METHODS AND RESULTS: We compared widely used methods for stress and strain calculations across diverse clinical scenarios representing distinct types of left ventricular myocardial disease. Our analysis revealed significant discrepancies in both the stress and strain values obtained with different methods. However, a newly developed modified version of the Mirsky equation demonstrated close agreement with the finite element analysis results for circumferential stress, while the Lamé method produced results close to those of finite element analysis for longitudinal stress and improved contractance accuracy. CONCLUSION: This study highlights significant inconsistencies in stress and strain values calculated using different methods, emphasising the potential impact on contractance calculations and subsequent clinical interpretation. We recommend adopting the Lamé method for longitudinal stress assessment and the modified Mirsky equation for circumferential stress analysis. These methods offer a balance between accuracy and feasibility, making them advantageous for clinical practice. By adopting these recommendations, we can improve the accuracy of LV wall stress and strain estimates, leading to more dependable contractance calculations, better prognostication and improved clinical decisions. CLINICAL AND TRANSLATIONAL IMPACT STATEMENT: Accurately estimating myocardial stress and strain is of paramount significance in clinical practice because the calculation of the contractance, defined and quantified by myocardial active strain energy density, necessitates correct stress and strain data. Contractance, which assesses myocardial work per unit muscle volume, has emerged as a promising indicator of contractile function and a predictor of future risk. The new recommendations for calculating myocardial stress improve the reliability of calculating contractance and enhance the understanding of myocardial diseases.

Proteomics couples electrical remodelling to inflammation in a murine model of heart failure with sinus node dysfunction.

AIMS: In patients with heart failure (HF), concomitant sinus node dysfunction (SND) is an important predictor of mortality, yet its molecular underpinnings are poorly understood. Using proteomics, this study aimed to dissect the protein and phosphorylation remodelling within the sinus node in an animal model of HF with concurrent SND. METHODS AND RESULTS: We acquired deep sinus node proteomes and phosphoproteomes in mice with heart failure and SND and report extensive remodelling. Intersecting the measured (phospho)proteome changes with human genomics pharmacovigilance data, highlighted downregulated proteins involved in electrical activity such as the pacemaker ion channel, Hcn4. We confirmed the importance of ion channel downregulation for sinus node physiology using computer modelling. Guided by the proteomics data, we hypothesized that an inflammatory response may drive the electrophysiological remodeling underlying SND in heart failure. In support of this, experimentally induced inflammation downregulated Hcn4 and slowed pacemaking in the isolated sinus node. From the proteomics data we identified proinflammatory cytokine-like protein galectin-3 as a potential target to mitigate the effect. Indeed, in vivo suppression of galectin-3 in the animal model of heart failure prevented SND. CONCLUSION: Collectively, we outline the protein and phosphorylation remodeling of SND in heart failure, we highlight a role for inflammation in electrophysiological remodelling of the sinus node, and we present galectin-3 signalling as a target to ameliorate SND in heart failure.

Cardiac GR Mediates the Diurnal Rhythm in Ventricular Arrhythmia Susceptibility.

BACKGROUND: Ventricular arrhythmias (VAs) demonstrate a prominent day-night rhythm, commonly presenting in the morning. Transcriptional rhythms in cardiac ion channels accompany this phenomenon, but their role in the morning vulnerability to VAs and the underlying mechanisms are not understood. We investigated the recruitment of transcription factors that underpins transcriptional rhythms in ion channels and assessed whether this mechanism was pertinent to the heart's intrinsic diurnal susceptibility to VA. METHODS AND RESULTS: Assay for transposase-accessible chromatin with sequencing performed in mouse ventricular myocyte nuclei at the beginning of the animals' inactive (ZT0) and active (ZT12) periods revealed differentially accessible chromatin sites annotating to rhythmically transcribed ion channels and distinct transcription factor binding motifs in these regions. Notably, motif enrichment for the glucocorticoid receptor (GR; transcriptional effector of corticosteroid signaling) in open chromatin profiles at ZT12 was observed, in line with the well-recognized ZT12 peak in circulating corticosteroids. Molecular, electrophysiological, and in silico biophysically-detailed modeling approaches demonstrated GR-mediated transcriptional control of ion channels (including Scn5a underlying the cardiac Na+ current, Kcnh2 underlying the rapid delayed rectifier K+ current, and Gja1 responsible for electrical coupling) and their contribution to the day-night rhythm in the vulnerability to VA. Strikingly, both pharmacological block of GR and cardiomyocyte-specific genetic knockout of GR blunted or abolished ion channel expression rhythms and abolished the ZT12 susceptibility to pacing-induced VA in isolated hearts. CONCLUSIONS: Our study registers a day-night rhythm in chromatin accessibility that accompanies diurnal cycles in ventricular myocytes. Our approaches directly implicate the cardiac GR in the myocyte excitability rhythm and mechanistically link the ZT12 surge in glucocorticoids to intrinsic VA propensity at this time.

Evaluating pro-arrhythmogenic effects of the T634S-hERG mutation: insights from a simulation study.

A mutation to serine of a conserved threonine (T634S) in the hERG K+ channel S6 pore region has been identified as a variant of uncertain significance, showing a loss-of-function effect. However, its potential consequences for ventricular excitation and arrhythmogenesis have not been reported. This study evaluated possible functional effects of the T634S-hERG mutation on ventricular excitation and arrhythmogenesis by using multi-scale computer models of the human ventricle. A Markov chain model of the rapid delayed rectifier potassium current (IKr) was reconstructed for wild-type and T634S-hERG mutant conditions and incorporated into the ten Tusscher et al. models of human ventricles at cell and tissue (1D, 2D and 3D) levels. Possible functional impacts of the T634S-hERG mutation were evaluated by its effects on action potential durations (APDs) and their rate-dependence (APDr) at the cell level; and on the QT interval of pseudo-ECGs, tissue vulnerability to unidirectional conduction block (VW), spiral wave dynamics and repolarization dispersion at the tissue level. It was found that the T634S-hERG mutation prolonged cellular APDs, steepened APDr, prolonged the QT interval, increased VW, destablized re-entry and augmented repolarization dispersion across the ventricle. Collectively, these results imply potential pro-arrhythmic effects of the T634S-hERG mutation, consistent with LQT2.

Dbh+ catecholaminergic cardiomyocytes contribute to the structure and function of the cardiac conduction system in murine heart.

The heterogeneity of functional cardiomyocytes arises during heart development, which is essential to the complex and highly coordinated cardiac physiological function. Yet the biological and physiological identities and the origin of the specialized cardiomyocyte populations have not been fully comprehended. Here we report a previously unrecognised population of cardiomyocytes expressing Dbhgene encoding dopamine beta-hydroxylase in murine heart. We determined how these myocytes are distributed across the heart by utilising advanced single-cell and spatial transcriptomic analyses, genetic fate mapping and molecular imaging with computational reconstruction. We demonstrated that they form the key functional components of the cardiac conduction system by using optogenetic electrophysiology and conditional cardiomyocyte Dbh gene deletion models. We revealed their close relationship with sympathetic innervation during cardiac conduction system formation. Our study thus provides new insights into the development and heterogeneity of the mammalian cardiac conduction system by revealing a new cardiomyocyte population with potential catecholaminergic endocrine function.

Non-canonical role of the sympathetic nervous system in the day-night rhythm in heart rate.

Although, for many decades, the day-night rhythm in resting heart rate has been attributed to the parasympathetic branch of the autonomic nervous system (high vagal tone during sleep), recently we have shown that there is a circadian clock in the cardiac pacemaker, the sinus node, and the day-night rhythm in heart rate involves an intrinsic rhythmic transcriptional remodelling of pacemaker ion channels, particularly Hcn4. We have now investigated the role of the sympathetic branch of the autonomic nervous system in this and shown it to have a non-canonical role. In mice, sustained long-term block of cardiac ß-adrenergic receptors by propranolol administered in the drinking water abolished the day-night rhythm in pacemaking in the isolated sinus node. Concomitant with this, there was a loss of the normal day-night rhythm in many pacemaker ion channel transcripts. However, there was little or no change in the local circadian clock, indicating that the well-known day-night rhythm in sympathetic nerve activity is directly involved in pacemaker ion channel transcription. The day-night rhythm in pacemaking helps explain the occurrence of clinically significant bradyarrhythmias during sleep, and this study improves our understanding of this pathology. This article is part of the theme issue 'The heartbeat: its molecular basis and physiological mechanisms'.

Human atrial fibrillation and genetic defects in transient outward currents: mechanistic insights from multi-scale computational models.

Previous studies have linked dysfunctional Ito arising from mutations to KCND3-encoded Kv4.3 and KCND2-encoded Kv4.2 to atrial fibrillation. Using computational models, this study aimed to investigate the mechanisms underlying pro-arrhythmic effects of the gain-of-function Kv4.3 (T361S, A545P) and Kv4.2 (S447R) mutations. Wild-type and mutant Ito formulations were developed from and validated against experimental data and incorporated into the Colman et al. model of human atrial cells. Single-cell models were incorporated into one- (1D) and two-dimensional (2D) models of atrial tissue, and a three-dimensional (3D) realistic model of the human atria. The three gain-of-function mutations had similar, albeit quantitatively different, effects: shortening of the action potential duration; lowering the plateau membrane potential, abbreviating the effective refractory period (ERP) and the wavelength (WL) of atrial excitation at the tissue level. Restitution curves for the WL, the ERP and the conduction velocity were leftward shifted, facilitating the conduction of atrial excitation waves at high excitation rates. The mutations also increased lifespan and stationarity of re-entry in both 2D and 3D simulations, which further highlighted a mutation-induced increase in spatial dispersion of repolarization. Collectively, these changes account for pro-arrhythmic effects of these Kv4.3 and Kv4.2 mutations in facilitating AF. This article is part of the theme issue 'The heartbeat: its molecular basis and physiological mechanisms'.

Remodelling and dysfunction of the sinus node in pulmonary arterial hypertension.

Patients with pulmonary arterial hypertension (PAH) have a high burden of arrhythmias, including arrhythmias arising from sinus node dysfunction, and the aim of this study was to investigate the effects of PAH on the sinus node. In the rat, PAH was induced by an injection of monocrotaline. Three weeks after injection, there was a decrease of the intrinsic heart rate (heart rate in the absence of autonomic tone) as well as the normal heart rate, evidence of sinus node dysfunction. In the sinus node of PAH rats, there was a significant downregulation of many ion channels and Ca2+-handling genes that could explain the dysfunction: HCN1 and HCN4 (responsible for pacemaker current, If), Cav1.2, Cav1.3 and Cav3.1 (responsible for L- and T-type Ca2+ currents, ICa,L and ICa,T), NCX1 (responsible for Na+-Ca2+ exchanger) and SERCA2 and RYR2 (Ca2+-handling molecules). In the sinus node of PAH rats, there was also a significant upregulation of many fibrosis genes that could also help explain the dysfunction: vimentin, collagen type 1, elastin, fibronectin and transforming growth factor ß1. In summary, in PAH, there is a remodelling of ion channel, Ca2+-handling and fibrosis genes in the sinus node that is likely to be responsible for the sinus node dysfunction. This article is part of the theme issue 'The heartbeat: its molecular basis and physiological mechanisms'.

Mechanistic insights into spontaneous transition from cellular alternans to ventricular fibrillation.

T-wave alternans (TWA) has been used for predicting the risk of malignant cardiac arrhythmias and sudden cardiac death (SCD) in multiple clinical settings; however, possible mechanism(s) underlying the spontaneous transition from cellular alternans reflected by TWA to arrhythmias in impaired repolarization remains unclear. The healthy guinea pig ventricular myocytes under E-4031 blocking IKr (0.1 µM, N = 12; 0.3 µM, N = 10; 1 µM, N = 10) were evaluated using whole-cell patch-clamp. The electrophysiological properties of isolated perfused guinea pig hearts under E-4031 (0.1 µM, N = 5; 0.3 µM, N = 5; 1 µM, N = 5) were evaluated using dual- optical mapping. The amplitude/threshold/restitution curves of action potential duration (APD) alternans and potential mechanism(s) underlying the spontaneous transition of cellular alternans to ventricular fibrillation (VF) were examined. There were longer APD80 and increased amplitude and threshold of APD alternans in E-4031 group compared with baseline group, which was reflected by more pronounced arrhythmogenesis at the tissue level, and were associated with steep restitution curves of the APD and the conduction velocity (CV). Conduction of AP alternans augmented tissue's functional spatiotemporal heterogeneity of regional AP/Ca alternans, as well as the AP/Ca dispersion, leading to localized uni-directional conduction block that spontaneous facilitated the formation of reentrant excitation waves without the need for additional premature stimulus. Our results provide a possible mechanism for the spontaneous transition from cardiac electrical alternans in cellular action potentials and intercellular conduction without the involvement of premature excitations, and explain the increased susceptibility to ventricular arrhythmias in impaired repolarization. In this study, we implemented voltage-clamp and dual-optical mapping approaches to investigate the underlying mechanism(s) for the arrhythmogenesis of cardiac alternans in the guinea pig heart at cellular and tissue levels. Our results demonstrated a spontaneous development of reentry from cellular alternans, arising from a combined actions of restitution properties of action potential duration, conduction velocity of excitation wave and interplay between alternants of action potential and the intracellular Ca handling. We believe this study provides new insights into underlying the mechanism, by which cellular cardiac alternans spontaneously evolves into cardiac arrhythmias.

Knowledge graph analysis and visualization of artificial intelligence applied in electrocardiogram.

Background: Electrocardiogram (ECG) provides a straightforward and non-invasive approach for various applications, such as disease classification, biometric identification, emotion recognition, and so on. In recent years, artificial intelligence (AI) shows excellent performance and plays an increasingly important role in electrocardiogram research as well. Objective: This study mainly adopts the literature on the applications of artificial intelligence in electrocardiogram research to focus on the development process through bibliometric and visual knowledge graph methods. Methods: The 2,229 publications collected from the Web of Science Core Collection (WoSCC) database until 2021 are employed as the research objects, and a comprehensive metrology and visualization analysis based on CiteSpace (version 6.1. R3) and VOSviewer (version 1.6.18) platform, which were conducted to explore the co-authorship, co-occurrence and co-citation of countries/regions, institutions, authors, journals, categories, references and keywords regarding artificial intelligence applied in electrocardiogram. Results: In the recent 4 years, both the annual publications and citations of artificial intelligence in electrocardiogram sharply increased. China published the most articles while Singapore had the highest ACP (average citations per article). The most productive institution and authors were Ngee Ann Polytech from Singapore and Acharya U. Rajendra from the University of Technology Sydney. The journal Computers in Biology and Medicine published the most influential publications, and the subject with the most published articles are distributed in Engineering Electrical Electronic. The evolution of research hotspots was analyzed by co-citation references' cluster knowledge visualization domain map. In addition, deep learning, attention mechanism, data augmentation, and so on were the focuses of recent research through the co-occurrence of keywords.

Left ventricular contractance: A new measure of contractile function.

AIMS: Myocardial contractility is poorly defined and difficult to compare between studies. Contractance or myocardial active strain energy density (MASED) measures the mechanical work done per unit volume (with units of kJ/m3) by any cardiac tissue during contraction. Contractance is an ideal candidate for measuring contractile function as it combines information from both stress and strain. METHODS AND RESULTS: Data obtained from three previously published experimental studies using trabecular tissue was used to provide contemporaneous nominal stress and strain data in 18 different scenarios with different loading conditions. Contractance varied in the differing loading conditions with values of 1.16 (low preload), 2.02 (high afterload) and 3.76 kJ/m3 (normal). Contractance varied between 0 with isometric loading and 2.14 kJ/m3 with an isotonic and moderate afterload. Increasing inotropy increased contractance to 4.7 kJ/m3. CONCLUSION: We showed that calculating MASED was feasible and provided a measure of energy production (work done) per unit volume of myocardium during contraction. The new term for contractile function, contractance, can be defined and quantified by MASED. Contractance measures contractile function in differing preload, afterload and inotropic settings. The method of measuring contractance is transferable to the assessment of global and regional systolic function.

A robust multiple heartbeats classification with weight-based loss based on convolutional neural network and bidirectional long short-term memory.

Background: Analysis of electrocardiogram (ECG) provides a straightforward and non-invasive approach for cardiologists to diagnose and classify the nature and severity of variant cardiac diseases including cardiac arrhythmia. However, the interpretation and analysis of ECG are highly working-load demanding, and the subjective may lead to false diagnoses and heartbeats classification. In recent years, many deep learning works showed an excellent role in accurate heartbeats classification. However, the imbalance of heartbeat classes is universal in most of the available ECG databases since abnormal heartbeats are always relatively rare in real life scenarios. In addition, many existing approaches achieved prominent results by removing noise and extracting features in data preprocessing, which relies heavily on powerful computers. It is a pressing need to develop efficient and automatic light weighted algorithms for accurate heartbeats classification that can be used in portable ECG sensors. Objective: This study aims at developing a robust and efficient deep learning method, which can be embedded into wearable or portable ECG monitors for classifying heartbeats. Methods: We proposed a novel and light weighted deep learning architecture with weight-based loss based on a convolutional neural network (CNN) and bidirectional long short-term memory (Bi-LSTM) that can automatically identify five types of ECG heartbeats according to the AAMI EC57 standard. It was also true that the raw ECG signals were simply segmented without noise removal and other feature extraction processing. Moreover, to tackle the challenge of classification bias due to imbalanced ECG datasets for different types of arrhythmias, we introduced a weight-based loss function to reduce the influence of over-weighted categories in the ECG dataset. For avoiding the influence of the division of validation dataset, k-fold method was adopted to improve the reliability of the model. Results: The proposed algorithm is trained and tested on MIT-BIH Arrhythmia Database, and achieves an average of 99.33% accuracy, 93.67% sensitivity, 99.18% specificity, 89.85% positive prediction, and 91.65% F1 score.

Effects of fibroblast on electromechanical dynamics of human atrial tissue-insights from a 2D discrete element model.

Roughly 75% of normal myocardial tissue volume is comprised of myocytes, however, fibroblasts by number are the most predominant cells in cardiac tissue. Previous studies have shown distinctive differences in cellular electrophysiology and excitability between myocytes and fibroblasts. However, it is still unclear how the electrical coupling between the two and the increased population of fibroblasts affects the electromechanical dynamics of cardiac tissue. This paper focuses on investigating effects of fibroblast-myocyte electrical coupling (FMEC) and fibroblast population on atrial electrical conduction and mechanical contractility by using a two-dimensional Discrete Element Method (DEM) model of cardiac tissue that is different to finite element method (FEM). In the model, the electro-mechanics of atrial cells are modelled by a biophysically detailed model for atrial electrical action potentials and myofilament kinetics, and the atrial fibroblasts are modelled by an active model that considers four active membrane ionic channel currents. Our simulation results show that the FMEC impairs myocytes' electrical action potential and mechanical contractibility, manifested by reduced upstroke velocity, amplitude and duration of action potentials, as well as cell length shortening. At the tissue level, the FMEC slows down the conduction of excitation waves, and reduces strain of the tissue produced during a contraction course. These findings provide new insights into understandings of how FMEC impairs cardiac electrical and mechanical dynamics of the heart.

Enhancing Protein Function Prediction Performance by Utilizing AlphaFold-Predicted Protein Structures.

The structure of a protein is of great importance in determining its functionality, and this characteristic can be leveraged to train data-driven prediction models. However, the limited number of available protein structures severely limits the performance of these models. AlphaFold2 and its open-source data set of predicted protein structures have provided a promising solution to this problem, and these predicted structures are expected to benefit the model performance by increasing the number of training samples. In this work, we constructed a new data set that acted as a benchmark and implemented a state-of-the-art structure-based approach for determining whether the performance of the function prediction model can be improved by putting additional AlphaFold-predicted structures into the training set and further compared the performance differences between two models separately trained with real structures only and AlphaFold-predicted structures only. Experimental results indicated that structure-based protein function prediction models could benefit from virtual training data consisting of AlphaFold-predicted structures. First, model performances were improved in all three categories of Gene Ontology terms (GO terms) after adding predicted structures as training samples. Second, the model trained only on AlphaFold-predicted virtual samples achieved comparable performances to the model based on experimentally solved real structures, suggesting that predicted structures were almost equally effective in predicting protein functionality.

Left ventricular active strain energy density is a promising new measure of systolic function.

The left ventricular ejection fraction does not accurately predict exercise capacity or symptom severity and has a limited role in predicting prognosis in heart failure. A better method of assessing ventricular performance is needed to aid understanding of the pathophysiological mechanisms and guide management in conditions such as heart failure. In this study, we propose two novel measures to quantify myocardial performance, the global longitudinal active strain energy (GLASE) and its density (GLASED) and compare them to existing measures in normal and diseased left ventricles. GLASED calculates the work done per unit volume of muscle (energy density) by combining information from myocardial strain and wall stress (contractile force per unit cross sectional area). Magnetic resonance images were obtained from 183 individuals forming four cohorts (normal, hypertension, dilated cardiomyopathy, and cardiac amyloidosis). GLASE and GLASED were compared with the standard ejection fraction, the corrected ejection fraction, myocardial strains, stroke work and myocardial forces. Myocardial shortening was decreased in all disease cohorts. Longitudinal stress was normal in hypertension, increased in dilated cardiomyopathy and severely decreased in amyloid heart disease. GLASE was increased in hypertension. GLASED was mildly reduced in hypertension (1.39 ± 0.65 kJ/m3), moderately reduced in dilated cardiomyopathy (0.86 ± 0.45 kJ/m3) and severely reduced in amyloid heart disease (0.42 ± 0.28 kJ/m3) compared to the control cohort (1.94 ± 0.49 kJ/m3). GLASED progressively decreased in the hypertension, dilated cardiomyopathy and cardiac amyloid cohorts indicating that mechanical work done and systolic performance is severely reduced in cardiac amyloid despite the relatively preserved ejection fraction. GLASED provides a new technique for assessing left ventricular myocardial health and contractile function.

Pro-Arrhythmic Effects of Discontinuous Conduction at the Purkinje Fiber-Ventricle Junction Arising From Heart Failure-Induced Ionic Remodeling - Insights From Computational Modelling.

Heart failure is associated with electrical remodeling of the electrical properties and kinetics of the ion channels and transporters that are responsible for cardiac action potentials. However, it is still unclear whether heart failure-induced ionic remodeling can affect the conduction of excitation waves at the Purkinje fiber-ventricle junction contributing to pro-arrhythmic effects of heart failure, as the complexity of the heart impedes a detailed experimental analysis. The aim of this study was to employ computational models to investigate the pro-arrhythmic effects of heart failure-induced ionic remodeling on the cardiac action potentials and excitation wave conduction at the Purkinje fiber-ventricle junction. Single cell models of canine Purkinje fiber and ventricular myocytes were developed for control and heart failure. These single cell models were then incorporated into one-dimensional strand and three-dimensional wedge models to investigate the effects of heart failure-induced remodeling on propagation of action potentials in Purkinje fiber and ventricular tissue and at the Purkinje fiber-ventricle junction. This revealed that heart failure-induced ionic remodeling of Purkinje fiber and ventricular tissue reduced conduction safety and increased tissue vulnerability to the genesis of the unidirectional conduction block. This was marked at the Purkinje fiber-ventricle junction, forming a potential substrate for the genesis of conduction failure that led to re-entry. This study provides new insights into proarrhythmic consequences of heart failure-induced ionic remodeling.

Generalizable Beat-by-Beat Arrhythmia Detection by Using Weakly Supervised Deep Learning.

Beat-by-beat arrhythmia detection in ambulatory electrocardiogram (ECG) monitoring is critical for the evaluation and prognosis of cardiac arrhythmias, however, it is a highly professional demanding and time-consuming task. Current methods for automatic beat-by-beat arrhythmia detection suffer from poor generalization ability due to the lack of large-sample and finely-annotated (labels are given to each beat) ECG data for model training. In this work, we propose a weakly supervised deep learning framework for arrhythmia detection (WSDL-AD), which permits training a fine-grained (beat-by-beat) arrhythmia detector with the use of large amounts of coarsely annotated ECG data (labels are given to each recording) to improve the generalization ability. In this framework, heartbeat classification and recording classification are integrated into a deep neural network for end-to-end training with only recording labels. Several techniques, including knowledge-based features, masked aggregation, and supervised pre-training, are proposed to improve the accuracy and stability of the heartbeat classification under weak supervision. The developed WSDL-AD model is trained for the detection of ventricular ectopic beats (VEB) and supraventricular ectopic beats (SVEB) on five large-sample and coarsely-annotated datasets and the model performance is evaluated on three independent benchmarks according to the recommendations from the Association for the Advancement of Medical Instrumentation (AAMI). The experimental results show that our method improves the F 1 score of supraventricular ectopic beats detection by 8%-290% and the F1 of ventricular ectopic beats detection by 4%-11% on the benchmarks compared with the state-of-the-art methods of supervised learning. It demonstrates that the WSDL-AD framework can leverage the abundant coarsely-labeled data to achieve a better generalization ability than previous methods while retaining fine detection granularity. Therefore, this framework has a great potential to be used in clinical and telehealth applications. The source code is available at https://github.com/sdnjly/WSDL-AD.

Mechanisms of ventricular arrhythmias elicited by coexistence of multiple electrophysiological remodeling in ischemia: A simulation study.

Myocardial ischemia, injury and infarction (MI) are the three stages of acute coronary syndrome (ACS). In the past two decades, a great number of studies focused on myocardial ischemia and MI individually, and showed that the occurrence of reentrant arrhythmias is often associated with myocardial ischemia or MI. However, arrhythmogenic mechanisms in the tissue with various degrees of remodeling in the ischemic heart have not been fully understood. In this study, biophysical detailed single-cell models of ischemia 1a, 1b, and MI were developed to mimic the electrophysiological remodeling at different stages of ACS. 2D tissue models with different distributions of ischemia and MI areas were constructed to investigate the mechanisms of the initiation of reentrant waves during the progression of ischemia. Simulation results in 2D tissues showed that the vulnerable windows (VWs) in simultaneous presence of multiple ischemic conditions were associated with the dynamics of wave propagation in the tissues with each single pathological condition. In the tissue with multiple pathological conditions, reentrant waves were mainly induced by two different mechanisms: one is the heterogeneity along the excitation wavefront, especially the abrupt variation in conduction velocity (CV) across the border of ischemia 1b and MI, and the other is the decreased safe factor (SF) for conduction at the edge of the tissue in MI region which is attributed to the increased excitation threshold of MI region. Finally, the reentrant wave was observed in a 3D model with a scar reconstructed from MRI images of a MI patient. These comprehensive findings provide novel insights for understanding the arrhythmic risk during the progression of myocardial ischemia and highlight the importance of the multiple pathological stages in designing medical therapies for arrhythmias in ischemia.

Inter-subject registration-based one-shot segmentation with alternating union network for cardiac MRI images.

Medical image segmentation based on deep-learning networks makes great progress in assisting disease diagnosis. However, currently, the training of most networks still requires a large amount of data with labels. In reality, labeling a considerable number of medical images is challenging and time-consuming. In order to tackle this challenge, a new one-shot segmentation framework for cardiac MRI images based on an inter-subject registration model called Alternating Union Network (AUN) is proposed in this study. The label of the source image is warped with deformation fields discovered from AUN to segment target images directly. Initially, the volumes are pre-processed by aligning affinely and adjusting the global intensity to simplify subsequent deformation registration. AUN consists of two kinds of subnetworks trained alternately to optimize segmentation gradually. The first kind of subnetwork takes a pair of volumes as inputs and registers them using global intensity similarity. The second kind of subnetwork, which takes the predicted labels generated from the previous subnetwork and the labels refined using the information of intrinsic anatomical structures of interest as inputs, is intensity-independent and focuses attention on registering structures of interest. Specifically, the input of AUN is a pair of a labeled image with the texture in regions of interest removed and a target image. Additionally, a new similarity measurement more appropriate for registering such image pair is defined as Local Squared Error (LSE). The proposed registration-based one-shot segmentation pays attention to the problem of the lack of labeled medical images. In AUN, only one labeled volume is required and a large number of unlabeled ones can be leveraged to improve segmentation performance, which has great advantages in clinical application. In addition, the intensity-independent subnetwork and LSE proposed in this study empower the framework to segment medical images with complicated intensity distribution.