Prior electrocardiograms not useful for machine learning predictions of major adverse cardiac events in emergency department chest pain patients.

At the emergency department (ED), it is important to quickly and accurately determine which patients are likely to have a major adverse cardiac event (MACE). Machine learning (ML) models can be used to aid physicians in detecting MACE, and improving the performance of such models is an active area of research. In this study, we sought to determine if ML models can be improved by including a prior electrocardiogram (ECG) from each patient. To that end, we trained several models to predict MACE within 30 days, both with and without prior ECGs, using data collected from 19,499 consecutive patients with chest pain, from five EDs in southern Sweden, between the years 2017 and 2018. Our results indicate no improvement in AUC from prior ECGs. This was consistent across models, both with and without additional clinical input variables, for different patient subgroups, and for different subsets of the outcome. While contradicting current best practices for manual ECG analysis, the results are positive in the sense that ML models with fewer inputs are more easily and widely applicable in practice.

Serial electrocardiograms at follow-up for early detection of transplanted heart rejection: A viewpoint.

This viewpoint proposed that serial electrocardiograms (ECG) could be used to monitor for heart transplantation (HT) rejection, based on the expected attenuation of the amplitude of ECG QRS complexes (attQRS) engendered by the rejection-induced decrease in electrical resistance due to the underlying myocardial edema (ME). Previous work in humans has shown attQRS in the setting of a diverse array of edematous states, affecting the myocardium (i.e, ME) and the body volume conductor "enveloping" the heart. Also, animal and human experience has revealed low electrical resistance during mild/moderate HT rejection. Studies with serial correlations of endomyocardial biopsy (EMB), echocardiography, cardiac magnetic resonance imaging, and ECG are recommended, which will merely require recording of an ECG, when EMB and imaging studies are carried out for monitoring of post-HT rejection.

Adult congenital heart disease care in a municipal public health system.

Specialty care is associated with improved outcomes for adults with adult CHD and must be extended to the underserved. A retrospective cohort study was performed to describe the provision of care to adult CHD patients in America's largest municipal public health system including patient demographics, diagnostic and therapeutic procedures, and adherence to guideline-recommended surveillance. We identified 229 adult CHD patients aged >18 years through electronic medical records. The most common diagnoses were atrial septal defect, ventricular septal defect, patent ductus arteriosus, and valvular pulmonary stenosis. In total, 65% had moderate or greater anatomic complexity. A large number of patients were uninsured (45%), non-white (96%), and non-English speaking (44%). One hundred forty-six patients (64%) presented with unrepaired primary defects. Fifty eight patients underwent primary repair during the study period; 48 of those repairs were surgical and 10 were transcatheter. Collaboration with an affiliated Comprehensive Care Center was utilised for 28% of patients. A high proportion of patients received adult CHD speciality visits (78%), echocardiograms (66%), and electrocardiograms (56%) at the guideline-recommended frequency throughout the study period. There was no significant difference in the rate of adherence to guideline-recommended surveillance based on insurance status, race/ethnicity, or primary language status. The proportion of patients who had guideline-recommended adult CHD visits, echocardiograms, and electrocardiograms was significantly lower for those with more advanced physiological stages. These results can inform the provision of adult CHD care in other public health system settings.

Deep learning-mediated prediction of concealed accessory pathway based on sinus rhythmic electrocardiograms.

BACKGROUND: Concealed accessory pathway (AP) may cause atrial ventricular reentrant tachycardia impacting the health of patients. However, it is asymptomatic and undetectable during sinus rhythm. METHODS: To detect concealed AP with electrocardiography (ECG) images, we collected normal sinus rhythmic ECG images of concealed AP patients and healthy subjects. All ECG images were randomly allocated to the training and testing datasets, and were used to train and test six popular convolutional neural networks from ImageNet pre-training and random initialization, respectively. RESULTS: We screened 152 ECG recordings in concealed AP group and 600 ECG recordings in control group. There were no statistically significant differences in ECG characteristics between control group and concealed AP group in terms of PR interval and QRS interval. However, the QT interval and QTc were slightly higher in control group than in concealed AP group. In the testing set, ResNet26, SE-ResNet50, MobileNetV3_large_100, and DenseNet169 achieved a sensitivity rate more than 87.0% with a specificity rate above 98.0%. And models trained from random initialization showed similar performance and convergence with models trained from ImageNet pre-training. CONCLUSION: Our study suggests that deep learning could be an effective way to predict concealed AP with normal sinus rhythmic ECG images. And our results might encourage people to rethink the possibility of training from random initialization on ECG image tasks.

Usefulness of the four-variable formula on serial electrocardiograms in detecting subtle anterior myocardial infarction.

OBJECTIVE: To perform serial electrocardiogram (ECG) analyses in patients with subtle ECG changes in the anterior leads and evaluate the performance of the four-variable formula in detecting left anterior descending (LAD) coronary artery occlusion. METHODS: This prospective study included patients admitted to the emergency department with acute chest pain between April 2021 and January 2023, whose initial ECG was not diagnostic but indicated suspicion of myocardial infarction (MI) and who underwent percutaneous coronary intervention in their follow-up. The control group consisted of patients who were diagnosed with benign variant ST-segment elevation (BV-STE) due to ST-segment elevation (STE) of at least 1 mm in the anterior lead, who had normal cardiac troponin levels, and who presented with non-cardiac chest pain. Following admission, six ECGs were taken at 10-min intervals. The scores of all patients were calculated with the four-variable formula on serial ECGs and compared between the groups. RESULTS: A total of 232 patients, including 116 with anterior MI and 116 with BV-STE, were included in the study. When the cut-off value for the four-variable formula was taken as ≥18.2, the sensitivity, specificity, and diagnostic accuracy of the first ECG were determined to be 82.7%, 85.3%, and 83.6%, respectively. We found that the four-variable formula had the highest sensitivity, specificity, and diagnostic accuracy in detecting LAD occlusion for the ECG taken at the 20th minute (83.6%, 89.6%, and 86.2%, respectively). CONCLUSION: The four-variable formula was found to be a valid method for the differentiation of STEMI and BV-STE in patients with subtle ECG changes. While managing this patient group, using serial ECGs rather than a single ECG to evaluate the clinical status of patients can help clinicians make more accurate decisions.

Artificial intelligence cloud platform improves arrhythmia detection from insertable cardiac monitors to 25 cardiac rhythm patterns through multi-label classification.

BACKGROUND: Electrocardiogram (ECG) is the gold standard for the diagnosis of cardiac arrhythmias and other heart diseases. Insertable cardiac monitors (ICMs) have been developed to continuously monitor cardiac activity over long periods of time and to detect 4 cardiac patterns (atrial tachyarrhythmias, ventricular tachycardia, bradycardia, and pause). However, interpretation of ECG or ICM subcutaneous ECG (sECG) is time-consuming for clinicians. Artificial intelligence (AI) classifies ECG and sECG with high accuracy in short times. OBJECTIVE: To demonstrate whether an AI algorithm can expand ICM arrhythmia recognition from 4 to many cardiac patterns. METHODS: We performed an exploratory retrospective study with sECG raw data coming from 20 patients wearing a Confirm Rx™ (Abbott, Sylmar, USA) ICM. The sECG data were recorded in standard conditions and then analyzed by AI (Willem™, IDOVEN, Madrid, Spain) and cardiologists, in parallel. RESULTS: In nineteen patients, ICMs recorded 2261 sECGs in an average follow-up of 23 months. Within these 2261 sECG episodes, AI identified 7882 events and classified them according to 25 different cardiac rhythm patterns with a pondered global accuracy of 88%. Global positive predictive value, sensitivity, and F1-score were 86.77%, 83.89%, and 85.52% respectively. AI was especially sensitive for bradycardias, pauses, rS complexes, premature atrial contractions, and inverted T waves, reducing the median time spent to classify each sECG compared to cardiologists. CONCLUSION: AI can process sECG raw data coming from ICMs without previous training, extending the performance of these devices and saving cardiologists' time in reviewing cardiac rhythm patterns detection.

Clinical valuation of ST changes in a group of patients with ventricular arrhythmias: The inSighT Study.

BACKGROUND: The inSighT study was designed to determine the prevalence of ischemic changes as recorded by implantable cardioverter-defibrillator (ICD) ST deviations in intracardiac electrocardiograms (EGM) over the 24 h preceding malignant ventricular arrhythmias (VT/VF). METHODS: The study enrolled patients with known coronary artery disease (CAD) or high risk of future development of CAD implanted with an ICD equipped with an ST monitoring feature (Ellipse™/Fortify Assura™, St. Jude Medical). Device session records were collected at each in-clinic follow-up. EGM ST levels of the beats over the 15 minutes prior to VT/VF events were compared using a t test with those from a baseline period of 23-24 h prior to the VT/VF event. All events with p < .05 were visually inspected to confirm they were evaluable; additional criteria for exclusion from further analysis included inappropriate therapy, aberrant conduction, and occurrence of VT/VF within 24h prior to the current event. RESULTS: The study enrolled 481 ICD patients (64 ± 11 years, 83% male) in 14 countries and followed them for 15±5 months. A total of 165 confirmed VT/VF episodes were observed, of which 71 events (in 56 patients, 34% of all patients with VT/VF) were preceded by significant (p < .05) ST-segment changes unrelated to known non-ischemic causes. None of the analyzed demographic and clinical factors proved to be associated with greater odds of presenting with ST-segment changes prior to VT/VF episode. CONCLUSION: In this exploratory study, characteristic ST-segment changes, likely representative of ischemic events, were observed in 34% of all patients with VT/VF episodes.

Application of artificial intelligence to the electrocardiogram.

Artificial intelligence (AI) has given the electrocardiogram (ECG) and clinicians reading them super-human diagnostic abilities. Trained without hard-coded rules by finding often subclinical patterns in huge datasets, AI transforms the ECG, a ubiquitous, non-invasive cardiac test that is integrated into practice workflows, into a screening tool and predictor of cardiac and non-cardiac diseases, often in asymptomatic individuals. This review describes the mathematical background behind supervised AI algorithms, and discusses selected AI ECG cardiac screening algorithms including those for the detection of left ventricular dysfunction, episodic atrial fibrillation from a tracing recorded during normal sinus rhythm, and other structural and valvular diseases. The ability to learn from big data sets, without the need to understand the biological mechanism, has created opportunities for detecting non-cardiac diseases as COVID-19 and introduced challenges with regards to data privacy. Like all medical tests, the AI ECG must be carefully vetted and validated in real-world clinical environments. Finally, with mobile form factors that allow acquisition of medical-grade ECGs from smartphones and wearables, the use of AI may enable massive scalability to democratize healthcare.

Multi-level longitudinal learning curve regression models integrated with item difficulty metrics for deliberate practice of visual diagnosis: groundwork for adaptive learning.

Visual diagnosis of radiographs, histology and electrocardiograms lends itself to deliberate practice, facilitated by large online banks of cases. Which cases to supply to which learners in which order is still to be worked out, with there being considerable potential for adapting the learning. Advances in statistical modeling, based on an accumulating learning curve, offer methods for more effectively pairing learners with cases of known calibrations. Using demonstration radiograph and electrocardiogram datasets, the advantages of moving from traditional regression to multilevel methods for modeling growth in ability or performance are demonstrated, with a final step of integrating case-level item-response information based on diagnostic grouping. This produces more precise individual-level estimates that can eventually support learner adaptive case selection. The progressive increase in model sophistication is not simply statistical but rather brings the models into alignment with core learning principles including the importance of taking into account individual differences in baseline skill and learning rate as well as the differential interaction with cases of varying diagnosis and difficulty. The developed approach can thus give researchers and educators a better basis on which to anticipate learners' pathways and individually adapt their future learning.

Reproducibility and repeatability of identifying the latest electrical activation during mapping of coronary sinus branches in CRT recipients.

INTRODUCTION: Studies have shown an association between the outcome in cardiac resynchronization therapy (CRT) and longer interventricular delay at the site of the left ventricular (LV) lead. Targeted LV lead placement at the latest electrically activated segment increases LV function further as compared with standard treatment. We aimed to determine reproducibility and repeatability of identifying the latest electrically activated segment during mapping of all available coronary sinus (CS) branches in patients receiving CRT. METHODS: We included 35 patients who underwent CRT implantation with protocolled mapping guided LV lead implantation aiming for the site of the latest electrical activation. Three different doctors experienced in electrophysiology and implantation of CRT devices independently measured time interval from the local bipolar right ventricular (RV) electrogram (EGM) to the local unipolar LV EGM at all mapped sites (RV-LV). The segment with the latest electrical activation was defined as the target segment (TS) and the CS tributary containing TS was defined as the target vein (TV). Weighted κ statistics with 95% confidence intervals were computed to assess intra- and interobserver agreement for TS and TV. RESULTS: We mapped 258 segments within 131 veins. Weighted κ values for repeatability were 0.85 (0.81-0.89) for TS and 0.92 (0.89-0.93) for TV, and weighted κ values of interobserver agreement ranged from 0.70 (0.61-0.73) to 0.80 (0.76-0.83) for TS and 0.73 (0.64-0.78) to 0.86 (0.83-0.89) for TV among all three observers. CONCLUSION: The reproducibility and repeatability of identifying the latest electrically activated segment during mapping of all available CS branches in patients receiving CRT range from good to very good.

Insulin resistance and reduced cardiac autonomic function in older adults: the Atherosclerosis Risk in Communities study.

BACKGROUND: Prior studies have shown insulin resistance is associated with reduced cardiac autonomic function measured at rest, but few studies have determined whether insulin resistance is associated with reduced cardiac autonomic function measured during daily activities. METHODS: We examined older adults without diabetes with 48-h ambulatory electrocardiography (n = 759) in an ancillary study of the Atherosclerosis Risk in Communities Study. Insulin resistance, the exposure, was defined by quartiles for three indexes: 1) the homeostatic model assessment of insulin resistance (HOMA-IR), 2) the triglyceride and glucose index (TyG), and 3) the triglyceride to high-density lipoprotein cholesterol ratio (TG/HDL-C). Low heart rate variability, the outcome, was defined by <25th percentile for four measures: 1) standard deviation of normal-to-normal R-R intervals (SDNN), a measure of total variability; 2) root mean square of successive differences in normal-to-normal R-R intervals (RMSSD), a measure of vagal activity; 3) low frequency spectral component (LF), a measure of sympathetic and vagal activity; and 4) high frequency spectral component (HF), a measure of vagal activity. Logistic regression was used to estimate odds ratios (OR) and 95% confidence intervals weighted for sampling/non-response, adjusted for age at ancillary visit, sex, and race/study-site. Insulin resistance quartiles 4, 3, and 2 were compared to quartile 1; high indexes refer to quartile 4 versus quartile 1. RESULTS: The average age was 78 years, 66% (n = 497) were women, and 58% (n = 438) were African American. Estimates of association were not robust at all levels of HOMA-IR, TyG, and TG/HDL-C, but suggest that high indexes were associated consistently with indicators of vagal activity. High HOMA-IR, high TyG, and high TG/HDL-C were consistently associated with low RMSSD (OR: 1.68 (1.00, 2.81), OR: 2.03 (1.21, 3.39), and OR: 1.73 (1.01, 2.91), respectively). High HOMA-IR, high TyG, and high TG/HDL-C were consistently associated with low HF (OR: 1.90 (1.14, 3.18), OR: 1.98 (1.21, 3.25), and OR: 1.76 (1.07, 2.90), respectively). CONCLUSIONS: In older adults without diabetes, insulin resistance was associated with reduced cardiac autonomic function - specifically and consistently for indicators of vagal activity - measured during daily activities. Primary prevention of insulin resistance may reduce the related risk of cardiac autonomic dysfunction.

Simultaneous monitoring of mouse respiratory and cardiac rates through a single precordial electrode.

Normal respiratory and circulatory functions are crucial for survival. However, conventional methods of monitoring respiration, some of which use sensors inserted into the nasal cavity, may interfere with naïve respiratory rates. In this study, we conducted a single-point measurement of electrocardiograms (ECGs) from the pectoral muscles of anesthetized and waking mice and found low-frequency oscillations in the ECG baseline. Using the fast Fourier transform of simultaneously recorded respiratory signals, we demonstrated that the low-frequency oscillations corresponded to respiratory rhythms. Moreover, the baseline oscillations changed in parallel with the respiratory rhythm when the latter was altered by pharmacological manipulation. We also demonstrated that this method could be combined with in vivo whole-cell patch-clamp recordings from the hippocampus. Thus, we developed a non-invasive form of respirometry in mice. Our recording method using a simple derivation algorithm is applicable to a variety of physiological and pharmacological experiments, providing an experimental platform in studying the mechanisms underlying the interaction of the central nervous system and the peripheral functions.

Electrocardiograms: a guide to rhythm recognition for emergency nurses.

Electrocardiogram (ECG) is one of the most commonly performed investigations in emergency departments (EDs), and is an extremely useful adjunct that guides diagnosis, prognosis and treatment. In most cases nurses are the first healthcare professional to assess patients and record an ECG, yet anecdotal evidence suggests that few emergency nurses review, interpret and act on ECG findings. Research suggests this may be due to lack of confidence in, or knowledge about, interpretation of results, often because of inadequate training. This article aims to help emergency nurses understand and interpret the cardiac rhythms commonly encountered on ECGs in EDs, to enable them to support earlier diagnosis and treatment. It describes a simple, five-step method for evaluating the main components of cardiac rhythm.

Is an Abnormal ECG Just the Tip of the ICE-berg? Examining the Utility of Electrocardiography in Detecting Methamphetamine-Induced Cardiac Pathology.

BACKGROUND: Methamphetamine use is escalating in Australia and New Zealand, with increasing emergency department attendance and mortality. Cardiac complications play a large role in methamphetamine-related mortality, and it would be informative to assess the frequency of abnormal electrocardiograms (ECGs) amongst methamphetamine users. OBJECTIVE: To determine the frequency and severity of ECG abnormalities amongst methamphetamine users compared to a control group. METHODS: We conducted a retrospective cohort analysis on 212 patients admitted to a tertiary hospital (106 patients with methamphetamine use, 106 age and gender-matched control patients). Electrocardiograms were analysed according to American College of Cardiology guidelines. RESULTS: Mean age was 33.4 years, with 73.6% male gender, with no significant differences between groups in smoking status, ECG indication, or coronary angiography rates. Methamphetamine users were more likely to have psychiatric admissions (22.6% vs 1.9%, p<0.0001). Overall, ECG abnormalities were significantly more common (71.7% vs 32.1%, p<0.0001) in methamphetamine users, particularly tachyarrhythmias (38.7% vs 26.4%, p<0.0001), right axis deviation (7.5% vs 0.0%, p=0.004), left ventricular hypertrophy (26.4% vs 4.7%, p<0.0001), P pulmonale pattern (7.5% vs 0.9%, p=0.017), inferior Q waves (10.4% vs 0.0%, p=0.001), lateral T wave inversion (3.8% vs 0.0%, p=0.043), and longer QTc interval (436.41±31.61ms vs 407.28±24.38ms, p<0.0001). Transthoracic echocardiogram (n=24) demonstrated left ventricular dysfunction (38%), thrombus (8%), valvular lesions (17%), infective endocarditis (17%), and pulmonary hypertension (13%). Electrocardiograms were only moderately sensitive at predicting abnormal TTE. CONCLUSION: Electrocardiographic abnormalities are more common in methamphetamine users than age and gender-matched controls. Due to the high frequency of abnormalities, ECGs should be performed in all methamphetamine users who present to hospital. Methamphetamine users with abnormal ECGs should undergo further cardiac investigations.

Novice interpretation of screening electrocardiograms and impact of online training.

INTRODUCTION: It is not known whether there is a specific training method that improves the accuracy of physician interpretations of pre-participation electrocardiograms (ECGs). METHODS: Participants took an online test and interpreted a series of normal, normal variant and abnormal ECGs. They then reviewed the BMJ's ECG interpretation online learning module and completed a post-test and a follow-up examination three months later. RESULTS: 28 fellows enrolled. The average correct for the pre-test was 63.57%, which increased to 81.19% for the post-test (p≤0.0001). When evaluating for retention, the average fell to 73.33% (p=0.0116) but was still significantly improved from baseline (p=0.0253). CONCLUSIONS: This study demonstrated that the accuracy of fellows' interpretation of ECGs significantly improved after completion of BMJ modules. Results of this study will likely impact the training of future sports medicine fellows and should encourage fellowship directors to incorporate the BMJ's ECG interpretation module as part of their curriculum.

AttentionCovidNet: Efficient ECG-based diagnosis of COVID-19.

The novel coronavirus caused a worldwide pandemic. Rapid detection of COVID-19 can help reduce the spread of the novel coronavirus as well as the burden on healthcare systems worldwide. The current method of detecting COVID-19 suffers from low sensitivity, with estimates of 50%-70% in clinical settings. Therefore, in this study, we propose AttentionCovidNet, an efficient model for the detection of COVID-19 based on a channel attention convolutional neural network for electrocardiograms. The electrocardiogram is a non-invasive test, and so can be more easily obtained from a patient. We show that the proposed model achieves state-of-the-art results compared to recent models in the field, achieving metrics of 0.993, 0.997, 0.993, and 0.995 for accuracy, precision, recall, and F1 score, respectively. These results indicate both the promise of the proposed model as an alternative test for COVID-19, as well as the potential of ECG data as a diagnostic tool for COVID-19.

Simple models vs. deep learning in detecting low ejection fraction from the electrocardiogram.

Aims: Deep learning methods have recently gained success in detecting left ventricular systolic dysfunction (LVSD) from electrocardiogram (ECG) waveforms. Despite their high level of accuracy, they are difficult to interpret and deploy broadly in the clinical setting. In this study, we set out to determine whether simpler models based on standard ECG measurements could detect LVSD with similar accuracy to that of deep learning models. Methods and results: Using an observational data set of 40 994 matched 12-lead ECGs and transthoracic echocardiograms, we trained a range of models with increasing complexity to detect LVSD based on ECG waveforms and derived measurements. The training data were acquired from the Stanford University Medical Center. External validation data were acquired from the Columbia Medical Center and the UK Biobank. The Stanford data set consisted of 40 994 matched ECGs and echocardiograms, of which 9.72% had LVSD. A random forest model using 555 discrete, automated measurements achieved an area under the receiver operator characteristic curve (AUC) of 0.92 (0.91-0.93), similar to a deep learning waveform model with an AUC of 0.94 (0.93-0.94). A logistic regression model based on five measurements achieved high performance [AUC of 0.86 (0.85-0.87)], close to a deep learning model and better than N-terminal prohormone brain natriuretic peptide (NT-proBNP). Finally, we found that simpler models were more portable across sites, with experiments at two independent, external sites. Conclusion: Our study demonstrates the value of simple electrocardiographic models that perform nearly as well as deep learning models, while being much easier to implement and interpret.

Three-Dimensional Electrodes of Liquid Metals for Long-Term, Wireless Cardiac Analysis and Modulation.

Cardiovascular disease is a major public health issue, and smart diagnostic approaches play an important role in the analysis of electrocardiograms. Here, we present three-dimensional, soft electrodes of liquid metals that can be conformably attached to the surfaces of the heart and skin for long-term cardiac analysis. The fine micropillar structures of biocompatible liquid metals enable precise targeting to small tissue areas, allowing for spatiotemporal mapping and modulation of cardiac electrical activity with high resolution. The low mechanical modulus of these liquid-metal electrodes not only helps avoid inflammatory responses triggered by modulus mismatch between the tissue and electrodes, but also minimizes pain when embedded in biological tissues such as the skin and heart. Furthermore, in vivo experiments with animal models and human trials demonstrate long-term and accurate monitoring of electrocardiograms over a period of 30 days.

Towards federated transfer learning in electrocardiogram signal analysis.

Modern methods in artificial intelligence perform very well on many healthcare datasets, at times outperforming trained doctors. However, many assumptions made in model training are not justifiable in clinical settings. In this work, we propose a method to train classifiers for electrocardiograms, able to deal with data of disparate input dimensions, distributed across different institutions, and able to protect patient privacy. In addition, we propose a simple method for creating federated datasets from any centralized dataset. We use autoencoders in conjunction with federated learning to model a highly heterogeneous modeling problem using the Massachusetts Institute of Technology Beth Israel Hospital Arrhythmia dataset, the Computing in Cardiology 2017 challenge dataset, and the PTB-XL dataset. For an encoding dimension of 1000, our federated classifier achieves an accuracy, precision, recall, and F1 score of 73.0%, 66.6%, 73.0%, and 69.7%, respectively. Our results suggest that dropping commonly made assumptions significantly complicate training and that as a result, estimates of performance of many machine learning models may overestimate performance when adopted for clinical settings.