Leveraging electrocardiography signals for deep learning-driven cardiovascular disease classification model.

Electrocardiography (ECG) is the most non-invasive diagnostic tool for cardiovascular diseases (CVDs). Automatic analysis of ECG signals assists in accurately and rapidly detecting life-threatening arrhythmias like atrioventricular blockage, atrial fibrillation, ventricular tachycardia, etc. The ECG recognition models need to utilize algorithms to detect various kinds of waveforms in the ECG and identify complicated relationships over time. However, the high variability of wave morphology among patients and noise are challenging issues. Physicians frequently utilize automated ECG abnormality recognition models to classify long-term ECG signals. Recently, deep learning (DL) models can be used to achieve enhanced ECG recognition accuracy in the healthcare decision making system. In this aspect, this study introduces an automated DL enabled ECG signal recognition (ADL-ECGSR) technique for CVD detection and classification. The ADL-ECGSR technique employs three most important subprocesses: pre-processed, feature extraction, parameter tuning, and classification. Besides, the ADL-ECGSR technique involves the design of a bidirectional long short-term memory (BiLSTM) based feature extractor, and the Adamax optimizer is utilized to optimize the trained method of the BiLSTM model. Finally, the dragonfly algorithm (DFA) with a stacked sparse autoencoder (SSAE) module is applied to recognize and classify EEG signals. An extensive range of simulations occur on benchmark PTB-XL datasets to validate the enhanced ECG recognition efficiency. The comparative analysis of the ADL-ECGSR methodology showed a remarkable performance of 91.24 % on the existing methods.

Multimodal insights into granger causality connectivity: Integrating physiological signals and gated eye-tracking data for emotion recognition using convolutional neural network.

This study introduces a groundbreaking method to enhance the accuracy and reliability of emotion recognition systems by combining electrocardiogram (ECG) with electroencephalogram (EEG) data, using an eye-tracking gated strategy. Initially, we propose a technique to filter out irrelevant portions of emotional data by employing pupil diameter metrics from eye-tracking data. Subsequently, we introduce an innovative approach for estimating effective connectivity to capture the dynamic interaction between the brain and the heart during emotional states of happiness and sadness. Granger causality (GC) is estimated and utilized to optimize input for a highly effective pre-trained convolutional neural network (CNN), specifically ResNet-18. To assess this methodology, we employed EEG and ECG data from the publicly available MAHNOB-HCI database, using a 5-fold cross-validation approach. Our method achieved an impressive average accuracy and area under the curve (AUC) of 91.00 % and 0.97, respectively, for GC-EEG-ECG images processed with ResNet-18. Comparative analysis with state-of-the-art studies clearly shows that augmenting ECG with EEG and refining data with an eye-tracking strategy significantly enhances emotion recognition performance across various emotions.

A computational study on the influence of antegrade accessory pathway location on the 12-lead electrocardiogram in Wolff-Parkinson-White syndrome.

Wolff-Parkinson-White syndrome is a cardiovascular disease characterized by abnormal atrio-ventricular conduction facilitated by accessory pathways (APs). Invasive catheter ablation of the AP represents the primary treatment modality. Accurate localization of APs is crucial for successful ablation outcomes, but current diagnostic algorithms based on the 12 lead electrocardiogram (ECG) often struggle with precise determination of AP locations. In order to gain insight into the mechanisms underlying localization failures observed in current diagnostic algorithms, we employ a virtual cardiac model to elucidate the relationship between AP location and ECG morphology. We first introduce a cardiac model of electrophysiology that was specifically tailored to represent antegrade APs in the form of a short atrio-ventricular bypass tract. Locations of antegrade APs were then automatically swept across both ventricles in the virtual model to generate a synthetic ECG database consisting of 9271 signals. Regional grouping of antegrade APs revealed overarching morphological patterns originating from diverse cardiac regions. We then applied variance-based sensitivity analysis relying on polynomial chaos expansion on the ECG database to mathematically quantify how variation in AP location and timing relates to morphological variation in the 12 lead ECG. We utilized our mechanistic virtual model to showcase limitations of AP localization using standard ECG-based algorithms and provide mechanistic explanations through exemplary simulations. Our findings highlight the potential of virtual models of cardiac electrophysiology not only to deepen our understanding of the underlying mechanisms of Wolff-Parkinson-White syndrome but also to potentially enhance the diagnostic accuracy of ECG-based algorithms and facilitate personalized treatment planning.

Unlocking Hidden Risks: Harnessing Artificial Intelligence (AI) to Detect Subclinical Conditions from an Electrocardiogram (ECG).

Recent artificial intelligence (AI) advancements in cardiovascular medicine offer potential enhancements in diagnosis, prediction, treatment, and outcomes. This article aims to provide a basic understanding of AI enabled ECG technology. Specific conditions and findings will be discussed, followed by reviewing associated terminology and methodology. In the appendix, definitions of AUC versus accuracy are explained. The application of deep learning models enables detecting diseases from normal electrocardiograms at accuracy not previously achieved by technology or human experts. Results with AI enabled ECG are encouraging as they considerably exceeded current screening models for specific conditions (i.e., atrial fibrillation, left ventricular dysfunction, aortic stenosis, and hypertrophic cardiomyopathy). This could potentially lead to a revitalization of the utilization of the ECG in the insurance domain. While we are embracing the findings with this rapidly evolving technology, but cautious optimism is still necessary at this point.

Ecg signal watermarking using QR decomposition.

This study introduces a novel watermarking technique for electrocardiogram (ECG) signals. Watermarking embeds critical information within the ECG signal, enabling data origin authentication, ownership verification, and ensuring the integrity of research data in domains like telemedicine, medical databases, insurance, and legal proceedings. Drawing inspiration from image watermarking, the proposed method transforms the ECG signal into a two-dimensional format for QR decomposition. The watermark is then embedded within the first row of the resulting R matrix. Three implementation scenarios are proposed: one in the spatial domain and two in the transform domain utilizing discrete wavelet transform (DWT) for improved watermark imperceptibility. Evaluation on real ECG signals from MIT-BIH Arrhythmia database and comparison to existing methods demonstrate that the proposed method achieves: (1) higher Peak Signal-to-Noise Ratio (PSNR) indicating minimal alterations to the watermarked signal, (2) lower bit error rates (BER) in robustness tests against external modifications such as AWGN noise (additive white Gaussian noise), line noise and down-sampling, and (3) lower computational complexity. These findings emphasize the effectiveness of the proposed QR decomposition-based watermarking method, achieving a balance between robustness and imperceptibility. The proposed approach has the potential to improve the security and authenticity of ECG data in healthcare and legal contexts, while its lower computational complexity enhances its practical applicability.

A coordinated adaptive multiscale enhanced spatio-temporal fusion network for multi-lead electrocardiogram arrhythmia detection.

The multi-lead electrocardiogram (ECG) is widely utilized in clinical diagnosis and monitoring of cardiac conditions. The advancement of deep learning has led to the emergence of automated multi-lead ECG diagnostic networks, which have become essential in the fields of biomedical engineering and clinical cardiac disease diagnosis. Intelligent ECG diagnosis techniques encompass Recurrent Neural Networks (RNN), Transformers, and Convolutional Neural Networks (CNN). While CNN is capable of extracting local spatial information from images, it lacks the ability to learn global spatial features and temporal memory features. Conversely, RNN relies on time and can retain significant sequential features. However, they are not proficient in extracting lengthy dependencies of sequence data in practical scenarios. The self-attention mechanism in the Transformer model has the capability of global feature extraction, but it does not adequately prioritize local features and cannot extract spatial and channel features. This paper proposes STFAC-ECGNet, a model that incorporates CAMV-RNN block, CBMV-CNN block, and TSEF block to enhance the performance of the model by integrating the strengths of CNN, RNN, and Transformer. The CAMV-RNN block incorporates a coordinated adaptive simplified self-attention module that adaptively carries out global sequence feature retention and enhances spatial-temporal information. The CBMV-CNN block integrates spatial and channel attentional mechanism modules in a skip connection, enabling the fusion of spatial and channel information. The TSEF block implements enhanced multi-scale fusion of image spatial and sequence temporal features. In this study, comprehensive experiments were conducted using the PTB-XL large publicly available ECG dataset and the China Physiological Signal Challenge 2018 (CPSC2018) database. The results indicate that STFAC-ECGNet surpasses other cutting-edge techniques in multiple tasks, showcasing robustness and generalization.

ECG Abnormalities, Diastolic Blood Pressure, and Adverse Events After Systemic to Pulmonary Artery Shunt in Infants with Congenital Heart Disease.

All patients with Systemic to Pulmonary Artery (SPA) shunt as the index surgical procedure at a single center were studied to determine the association between post-operative ECG repolarization abnormalities, diastolic blood pressure (DBP), and adverse outcomes. Postoperative ECGs were categorized into three grades, Grade 2 defined as ST elevation/depression ≥ 2 mm in ≥ 2 precordial or ≥ 1 mm in ≥ 2 limb leads; Grade 1-T-wave inversion or flattening in ≥ 3 leads; and Grade 0-no criteria for grades 1 or 2. For each patient, time with invasive DBP below 25, 25-29, 30-34, or above 34 mmHg in the first 24 h was calculated. The primary outcome was a pre-discharge composite of death, cardiac arrest, ECMO, unplanned shunt reintervention, and necrotizing enterocolitis after 24 h of surgery. Of the 109 patients included in final analysis, 17 (15.6%) had the composite outcome. Grade 2 ECG abnormality occurred in 12%, and Grade 1 in 37%. There was no association between ECG abnormalities and adverse events. Increasing time with DBP < 30 was not associated with adverse outcomes, while increasing time with DBP 30-34 was associated with decreased odds, and increasing time with DBP > 34 mmHg was associated with increased odds of adverse outcomes on multivariable analysis accounting for indexed shunt size and chromosomal abnormalities. In conclusion, after SPA shunt placement, ECG repolarization abnormalities and low DBP within 24 h were common and not associated with adverse outcomes. Sustained elevation of DBP above 34 mmHg was not protective, especially in patients with high indexed shunt size and chromosomal abnormalities.

Extended Ambulatory ECG Monitoring Enhances Identification of Higher-Risk Ventricular Tachyarrhythmias in Patients with Hypertrophic Cardiomyopathy.

BACKGROUND: In hypertrophic cardiomyopathy (HCM), 48-hour ambulatory monitoring has been standard practice to detect nonsustained ventricular tachycardia (NSVT), a sudden death risk marker. Extended wear ambulatory ECG devices have more recently utilized for monitoring HCM patients. OBJECTIVE: Evaluate NSVT burden identified with continuous ambulatory monitoring for up to 2 weeks compared to initial 48 hours. METHODS: 236 consecutive HCM patients (49 ± 12 years) underwent 14-day continuous ambulatory monitoring (Zio XT, iRhythm Technologies); diagnostic yield of NSVT compared for initial 48 hours vs. extended for 14 days. RESULTS: Of 236 patients, 114 (48%) had ≥1 runs of NSVT (median 2) over 14 days. Median length of NSVT was 7 beats (range: 3 to 67) at rates of 120 to 240 bpm (median, 167 bpm). In 42 of the 114 patients (37%), initial NSVT occurred ≤ 48 hours and in 72 (63%) only during the extended monitoring period (3 to 14 days). Diagnostic yield for detecting NSVT over 14 days was 2.7-fold greater than ≤ 48 hours (p<0.001). NSVT judged at higher risk (≥8 beats, >200 bpm, ≥2 runs in consecutive 2-day period) was identified more frequently during extended monitoring, diagnostic yield 3.0-fold greater than ≤ 48 hours (p<0.001). CONCLUSION: In HCM, NSVT episodes are frequent, however, in most patients, both NSVT and higher risk NSVT were not detected during initial 48-hours and confined solely to extended monitoring period. These data support additional clinical studies to evaluate the significance of NSVT on extended monitoring on sudden death risk in HCM.

Improved ECG-gated cardiac cine imaging with variable initial value based tiny golden-angle radial trajectory.

BACKGROUND: Breath-held electrocardiogram-gated cardiac cine imaging (ECG-CINE), as the gold standard for assessing cardiac function in magnetic resonance imaging (MRI), is prone to motion artifacts. Conventional golden-angle (CGA) sampling has emerged as a promising technique for mitigating motion effects in real-time cardiac cine imaging. However, in ECG-CINE, the irregular re-binning of radial k-space profiles based on CGA can exacerbate k-space non-uniformity, resulting in severe streaking artifacts. The recently introduced segmented golden-angle ratio (SGA) scheme aims to solve this problem; nevertheless, it sacrifices the desired motion insensitivity. PURPOSE: The study aims to develop a more efficient k-space sampling scheme for ECG-CINE that guarantees both improved motion insensitivity and optimized k-space coverage. METHOD: Theoretically, to enhance motion insensitivity, it is essential that the single-frame radial k-space profiles acquired within each heartbeat (HB) span as close to a full 360-degree range as possible. Meanwhile, to ensure uniform data coverage, the sequentially acquired k-space profiles need to be evenly distributed both within each HB and across multiple HBs. In this study, we propose a Variable Initial value-based tiny Golden-Angle radial trajectory (VIGA) to achieve these two goals. Specifically, VIGA is a two-step approach: First, the tiny golden-angle ratio is applied to the k-space profiles within each HB to maintain motion insensitivity and k-space uniformity as in CGA. Second, a golden ratio of the golden angle used within each HB is applied to the initial k-space profiles across adjacent HBs to optimize coverage further. We validated the proposed VIGA method through numerical simulations, phantom experiments, and prospective and retrospective in vivo cardiac cine experiments. RESULTS: Numerical simulations revealed that the k-space uniformity of CGA is highly dependent on the number of spokes per HB, whereas VIGA and SGA maintained nearly optimal k-space coverage regardless of this parameter. Both phantom and prospective studies demonstrated that VIGA outperforms CGA when the number of spokes per HB is suboptimal, and surpasses SGA in conditions with residual respiratory motion. The standard deviation of gradient scores indicates statistical significance between CGA and VIGA under free-breathing conditions (p = 0.039) and between SGA and VIGA under all conditions tested (Free-breathing, 200 spokes/HB: p = 0.028; Breath-holding, 200 spokes/HB: p = 0.008; Free-breathing, 200 spokes/HB: p = 0.013; Breath-holding, 200 spokes/HB: p = 0.011). Retrospective results demonstrated that doctor ratings for SGA were lower than those for VIGA, and the ratings for systole images using VIGA were significantly higher than those using CGA (2.55 ± 0.45 vs. 3.29 ± 0.52; p = 0.04). CONCLUSION: A novel and efficient k-space sampling scheme, named VIGA, was proposed to improve k-space uniformity and motion insensitivity. VIGA facilitates robust image quality in both prospective and retrospective cardiac cine imaging, demonstrating its potential as a clinically viable alternative to CGA and SGA.

CardioGuard: AI-driven ECG authentication hybrid neural network for predictive health monitoring in telehealth systems.

The increasing integration of telehealth systems underscores the importance of robust and secure methods for patient data management. Traditional authentication methods, such as passwords and PINs, are prone to breaches, underscoring the need for more secure alternatives. Therefore, there is a need for alternative approaches that provide enhanced security and user convenience. Biometric-based authentication systems uses individuals unique physical or behavioral characteristics for identification, have emerged as a promising solution. Specifically, Electrocardiogram (ECG) signals have gained attention among various biometric modalities due to their uniqueness, stability, and non-invasiveness. This paper presents CardioGaurd, a deep learning-based authentication system that leverages ECG signals-unique, stable, and non-invasive biometric markers. The proposed system uses a hybrid Convolution and Long short-term memory based model to obtain rich characteristics from the ECG signal and classify it as authentic or fake. CardioGaurd not only ensures secure access but also serves as a predictive tool by analyzing ECG patterns that could indicate early signs of cardiovascular abnormalities. This dual functionality enhances patient security and contributes to AI-driven disease prevention and early detection. Our results demonstrate that CardioGaurd offers superior performance in both security and potential predictive health insights compared to traditional models, thus supporting a shift towards more proactive and personalized telehealth solutions.

Ten quick tips for electrocardiogram (ECG) signal processing.

The electrocardiogram (ECG) is a powerful tool to measure the electrical activity of the heart, and the analysis of its data can be useful to assess the patient's health. In particular, the computational analysis of electrocardiogram data, also called ECG signal processing, can reveal specific patterns or heart cycle trends which otherwise would be unnoticeable by medical experts. When performing ECG signal processing, however, it is easy to make mistakes and generate inflated, overoptimistic, or misleading results, which can lead to wrong diagnoses or prognoses and, in turn, could even contribute to bad medical decisions, damaging the health of the patient. Therefore, to avoid common mistakes and bad practices, we present here ten easy guidelines to follow when analyzing electrocardiogram data computationally. Our ten recommendations, written in a simple way, can be useful to anyone performing a computational study based on ECG data and eventually lead to better, more robust medical results.

Feasibility of estimating tidal volume from electrocardiograph-derived respiration signal and respiration waveform.

PURPOSE: Estimating tidal volume (VT) from electrocardiography (ECG) can be quite useful during deep sedation or spinal anesthesia since it eliminates the need for additional monitoring of ventilation. This study aims to validate and compare VT estimation methodologies based on ECG-derived respiration (EDR) using real-world clinical data. MATERIALS AND METHODS: We analyzed data from 90 critically ill patients for general analysis and two critically ill patients for constrained analysis. EDR signals were generated from ECG data, and VT was estimated using impedance-based respiration waveforms. Linear regression and deep learning models, both subject-independent and subject-specific, were evaluated using mean absolute error and Pearson correlation. RESULTS: There was a strong short-term correlation between VT and the respiration waveform (r = 0.78 and 0.96), which weakened over longer periods (r = 0.23 and - 0.16). VT prediction models performed poorly in the general population (R2 = 0.17) but showed satisfactory performance in two constrained patient records using measured respiration waveforms (R2 = 0.84 to 0.94). CONCLUSION: Although EDR-based VT estimation is promising, current methodologies are limited by noisy ICU ECG signals, but controlled environment data showed significant short-term correlations with measured respiration waveforms. Future studies should develop reliable EDR extraction procedures and improve predictive models to broaden clinical applications.

Manifested U-Waves Prior to Seizure Attacks in a Patient Who Had Remote Subarachnoid Hemorrhage: A Case Report.

Sudden unexpected death in epilepsy (SUDEP) refers to unpredictable demise of a person following a seizure. Electroencephalograms can directly measure electrical activity in the brain; however, it cannot predict when seizures will occur. The use of electrocardiograms (ECGs) to monitor changes in brain electrical activity has gained attention, recently. In this case report, we retrospectively reviewed ECGs taken before and after seizure activity in a 75-year-old male who had a remote subarachnoid hemorrhage. Interestingly, U-waves appeared prior to his seizures and disappeared afterward, which suggests ECGs can be used to predict epilepsy in a certain population.

Enhancing P-wave localization for accurate detection of second-degree and third-degree atrioventricular conduction blocks.

Objective.This paper tackles the challenge of accurately detecting second-degree and third-degree atrioventricular block (AVB) in electrocardiogram (ECG) signals through automated algorithms. The inaccurate detection of P-waves poses a difficulty in this process. To address this limitation, we propose a reliable method that significantly improves the performances of AVB detection by precisely localizing P-waves.Approach.Our proposed P-WaveNet utilized an attention mechanism to extract spatial and temporal features, and employs a bidirectional long short-term memory module to capture inter-temporal dependencies within the ECG signal. To overcome the scarcity of data for second-degree and third-degree AVB (2AVB,3AVB), a mathematical approach was employed to synthesize pseudo-data. By combining P-wave positions identified by the P-WaveNet with key medical features such as RR interval rhythm and PR intervals, we established a classification rule enabling automatic AVB detection.Main results. The P-WaveNet achieved an F1 score of 93.62% and 91.42% for P-wave localization on the QT Dataset and Lobachevsky University dataset datasets, respectively. In the BUTPDB dataset, the F1 scores for P-wave localization in ECG signals with 2AVB and 3AVB were 98.29% and 62.65%, respectively. Across two independent datasets, the AVB detection algorithm achieved F1 scores of 83.33% and 84.15% for 2AVB and 3AVB, respectively.Significance.Our proposed P-WaveNet demonstrates accurate identification of P-waves in complex ECGs, significantly enhancing AVB detection efficacy. This paper's contributions stem from the fusion of medical expertise with data augmentation techniques and ECG classification. The proposed P-WaveNet demonstrates potential clinical applicability.

Validity and Reliability of Movesense HR+ ECG Measurements for High-Intensity Running and Cycling.

Low-cost, portable devices capable of accurate physiological measurements are attractive tools for coaches, athletes, and practitioners. The purpose of this study was primarily to establish the validity and reliability of Movesense HR+ ECG measurements compared to the criterion three-lead ECG, and secondarily, to test the industry leader Garmin HRM. Twenty-one healthy adults participated in running and cycling incremental test protocols to exhaustion, both with rest before and after. Movesense HR+ demonstrated consistent and accurate R-peak detection, with an overall sensitivity of 99.7% and precision of 99.6% compared to the criterion; Garmin HRM sensitivity and precision were 84.7% and 87.7%, respectively. Bland-Altman analysis compared to the criterion indicated mean differences (SD) in RR' intervals of 0.23 (22.3) ms for Movesense HR+ at rest and 0.38 (18.7) ms during the incremental test. The mean difference for Garmin HRM-Pro at rest was -8.5 (111.5) ms and 27.7 (128.7) ms for the incremental test. The incremental test correlation was very strong (r = 0.98) between Movesense HR+ and criterion, and moderate (r = 0.66) for Garmin HRM-Pro. This study developed a robust peak detection algorithm and data collection protocol for Movesense HR+ and established its validity and reliability for ECG measurement.

Acute myocardial infarction in a patient with permanent left bundle branch pacing.

BACKGROUND: Electrocardiographic diagnosis of acute myocardial infarction in the setting of cardiac pacing represents diagnostic challenge. There are no focusing data, neither reporting about diagnostic sensitivity of 12­lead ECG with left bundle branch area pacing (LBBAP) during acute myocardial infarction (AMI). CASE SUMMARY: We present 12­lead ECG morphology in a patient with permanent LBBAP during AMI. DISCUSSION: Abnormal repolarization changes induced by ventricular pacing can lead to delay in diagnosis in patients with AMI. LBBAP and overall conduction system pacing may facilitate a timely diagnosis providing additional, still underestimated, advantages of physiological pacing of the heart.

Practice and confidence in electrocardiogram interpretation among ICU nurses: A cross-sectional study.

OBJECTIVES: This study aimed to determine practice and confidence in electrocardiogram (ECG) interpretation among intensive care unit (ICU) nurses in Fujian Province, China, and identify predictors of ECG interpretation practice. RESEARCH METHODOLOGY/DESIGN: A quantitative cross-sectional study was conducted between October 2021 and December 2021 among 357 respondents. SETTING: Conducted online at twenty-one hospitals in all nine cities of Fujian Province. MAIN OUTCOME MEASURES: Purposive and convenient sampling techniques were employed in selecting hospitals and respondents, respectively. A validated and pre-tested Chinese version of the questionnaire was used in data collection. We conducted binary logistic regression to identify the predictors of ICU nurses' ECG interpretation practice, and linear regression to analyze the relationship between ECG interpretation practice and confidence. We considered statistically significant a p-value < 0.05. RESULTS: The practice mean score of the respondents was 5.54 (SD = 2.26) out of 10 points, and only 2.2 % of nurses correctly interpreted all the patient ECG strips. Few ICU nurses (25.5 %) had good ECG interpretation practice, with a confidence mean score of 2.02 (SD = 0.99) out of 4 points in their overall ability to interpret patient ECG strips. Currently working unit in comparison to cardiac ICU (emergency ICU: AOR = 5.71, 95 % CI: 1.84-17.75); previous ECG training (AOR = 2.02, 95 % CI: 1.10-3.70); source of ECG training (university/school) (AOR = 2.02, 95 % CI: 1.12-3.65); and ECG knowledge (AOR = 16.18, 95 % CI: 7.43-35.25) were significantly associated with the ECG interpretation practice. CONCLUSIONS: ICU nurses' ECG interpretation practice in the current study was relatively poor. An ECG education program is recommended to impart ICU nurses with basic ECG knowledge for enhancing good ECG interpretation practice and confidence in nursing care provision. IMPLICATIONS FOR CLINICAL PRACTICE: Good ECG interpretation skills are paramount among ICU nurses for better patient outcomes. ECG knowledge among ICU nurses is an important predictor of effective ECG monitoring for cardiac arrhythmias. Therefore, frequent, continuouszgood practice and boost confidence in the provision of quality nursing care.

ECG Findings Are Poor Predictors for Adverse Events and Cardiac Death in Barth Syndrome.

Background: Patients with Barth syndrome (BTHS) can present with cardiomyopathy. BTHS subjects are at risk for cardiac adverse outcomes throughout life, including malignant arrhythmias and death. Electrocardiogram (ECG) parameters have never been assessed as a tool to predict adverse outcomes in individuals with BTHS. Objectives: The purpose of this study was to identify any ECG parameters including QRS fragmentation, presence of arrhythmia, or abnormal intervals that could predict adverse outcomes and cardiac death among the BTHS population. Methods: We performed a retrospective case referent study on subjects with BTHS (n=43), and compared them with our reference group, subjects with idiopathic dilated cardiomyopathy (DCM) from a single institution (n=53) from 2007-2021. BTHS data was obtained from subjects attending the biennial Barth Syndrome Foundation International Scientific, Medical, and Family Conferences (BSFISMFC) from 2002-2018. ECG data from first and last available ECG's prior to an adverse event or cardiac death was analyzed, and then multivariable regression was performed to determine odd ratios between ECG characteristics and adverse events/cardiac death. Results: No ECG variables were statistically significant predictors of adverse events or cardiac death in the BTHS group. Last ECG QRS fragmentation trended to statistically significance (OR 13.3, p=0.12) in predicting adverse events in the DCM group. Conclusion: No ECG parameters, including QRS fragmentation, presence of arrhythmia, or abnormal interval values predict adverse events or cardiac death among BTHS patients. QRS fragmentation may be a predictor of adverse events in the DCM population.

Machine Learning Algorithm to Predict Atrial Fibrillation Using Serial 12-Lead ECGs Based on Left Atrial Remodeling.

BACKGROUND: We hypothesized that analysis of serial ECGs could predict new-onset atrial fibrillation (AF) more accurately than analysis of a single ECG by detecting the subtle cardiac remodeling that occurs immediately before AF occurrence. Our aim in this study was to compare the performance of 2 types of machine learning (ML) algorithms. METHODS AND RESULTS: Standard 12-lead ECGs of patients selected by cardiologists between January 2010 and May 2021 were used for ML model development. Two ML models (single ECG and serial ECG) were developed using a light gradient boosting machine-learning algorithm. Model performance was evaluated based on the area under the receiver operating characteristic curve, sensitivity, specificity, accuracy, and F1 score. We trained the ML models on 415 964 ECGs from 176 090 patients. When testing the 2 ML models using external validation data sets, the performance of the serial-ML model was significantly better than that of the single-ML model for predicting new-onset AF (single- versus serial-ML model: sensitivity 0.744 versus 0.810; specificity 0.742 versus 0.822; accuracy 0.743 versus 0.816; F1 score 0.743 versus 0.815; area under the receiver operating characteristic curve 0.812 versus 0.880; P<0.001). The Shapley Additive Explanations analysis ranked P-wave duration and amplitude among the top 10 ECG parameters. CONCLUSIONS: An ML model based on serial ECGs from an individual had greater ability to predict new-onset AF than the ML model based on a single ECG. P-wave morphologies were associated with future AF prediction.

Associations between epilepsy, respiratory impairment, and minor ECG abnormalities in children.

OBJECTIVE: We sought to examine the effects of acute seizures and respiratory derangement on the cardiac electrical properties reflected on the electrocardiogram (ECG); and to analyze their potential interactions with a diagnosis of epilepsy in children. METHODS: Emergency center (EC) visits with seizure or epilepsy diagnostic codes from 1/2011-12/2013 were included if they had ECG within 24 h of EC visit. Patients were excluded if they had pre-existing cardiac conditions, ion channelopathy, or were taking specific cardiac medications. Control subjects were 1:1 age and gender matched. Abnormal ECG was defined as changes in rhythm, PR, QRS, or corrected QT intervals; QRS axis or morphology; ST segment; or T wave morphology from normal standards. We identified independent associations between clinical factors and abnormal ECG findings using multivariable logistic regression modeling. RESULTS: Ninety-five children with epilepsy presented to the EC with seizures, respiratory distress, and other concerns. Three hundred children without epilepsy presented with seizures. There was an increased prevalence of minor ECG abnormalities in children with epilepsy (49 %) compared to the control subjects (29 %) and those without epilepsy (36 %). Epilepsy (OR: 1.61, 95 %CI: 1.01-2.6), need for supplemental oxygen (OR 3.06, 95 % CI: 1.45-6.44) or mechanical ventilation (OR: 2.5, 95 % CI: 1.03-6.05) were independently associated with minor ECG abnormalities. Secondary analyses further demonstrated an independent association between level of respiratory support and ECG abnormalities only in the epilepsy group. SIGNIFICANCE: Independent association of increased respiratory support with minor ECG abnormalities suggests a potential respiratory influence on the hearts of children with epilepsy.