Abstract
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.
Graphical Abstract
The application of artificial intelligence to the standard electrocardiogram enables it to diagnose conditions not previously identifiable by an electrocardiogram, or to do so with a greater performance than previously possible. This includes identification of the current rhythm, identification of episodic atrial fibrillation from an ECG acquired during sinus rhythm, the presence of ventricular dysfunction (low ejection fraction), the presence of valvular heart disease, channelopathies (even when electrocardiographically ‘concealed’), and the presence of hypertrophic cardiomyopathy.
Asymptomatic left ventricular dysfunction (ALVD) is present in 3-6% of the general population, is associated with reduced quality of life and longevity, and is treatable when found
. An inexpensive, ...noninvasive screening tool for ALVD in the doctor's office is not available. We tested the hypothesis that application of artificial intelligence (AI) to the electrocardiogram (ECG), a routine method of measuring the heart's electrical activity, could identify ALVD. Using paired 12-lead ECG and echocardiogram data, including the left ventricular ejection fraction (a measure of contractile function), from 44,959 patients at the Mayo Clinic, we trained a convolutional neural network to identify patients with ventricular dysfunction, defined as ejection fraction ≤35%, using the ECG data alone. When tested on an independent set of 52,870 patients, the network model yielded values for the area under the curve, sensitivity, specificity, and accuracy of 0.93, 86.3%, 85.7%, and 85.7%, respectively. In patients without ventricular dysfunction, those with a positive AI screen were at 4 times the risk (hazard ratio, 4.1; 95% confidence interval, 3.3 to 5.0) of developing future ventricular dysfunction compared with those with a negative screen. Application of AI to the ECG-a ubiquitous, low-cost test-permits the ECG to serve as a powerful screening tool in asymptomatic individuals to identify ALVD.
Atrial fibrillation is frequently asymptomatic and thus underdetected but is associated with stroke, heart failure, and death. Existing screening methods require prolonged monitoring and are limited ...by cost and low yield. We aimed to develop a rapid, inexpensive, point-of-care means of identifying patients with atrial fibrillation using machine learning.
We developed an artificial intelligence (AI)-enabled electrocardiograph (ECG) using a convolutional neural network to detect the electrocardiographic signature of atrial fibrillation present during normal sinus rhythm using standard 10-second, 12-lead ECGs. We included all patients aged 18 years or older with at least one digital, normal sinus rhythm, standard 10-second, 12-lead ECG acquired in the supine position at the Mayo Clinic ECG laboratory between Dec 31, 1993, and July 21, 2017, with rhythm labels validated by trained personnel under cardiologist supervision. We classified patients with at least one ECG with a rhythm of atrial fibrillation or atrial flutter as positive for atrial fibrillation. We allocated ECGs to the training, internal validation, and testing datasets in a 7:1:2 ratio. We calculated the area under the curve (AUC) of the receiver operatoring characteristic curve for the internal validation dataset to select a probability threshold, which we applied to the testing dataset. We evaluated model performance on the testing dataset by calculating the AUC and the accuracy, sensitivity, specificity, and F1 score with two-sided 95% CIs.
We included 180 922 patients with 649 931 normal sinus rhythm ECGs for analysis: 454 789 ECGs recorded from 126 526 patients in the training dataset, 64 340 ECGs from 18 116 patients in the internal validation dataset, and 130 802 ECGs from 36 280 patients in the testing dataset. 3051 (8·4%) patients in the testing dataset had verified atrial fibrillation before the normal sinus rhythm ECG tested by the model. A single AI-enabled ECG identified atrial fibrillation with an AUC of 0·87 (95% CI 0·86–0·88), sensitivity of 79·0% (77·5–80·4), specificity of 79·5% (79·0–79·9), F1 score of 39·2% (38·1–40·3), and overall accuracy of 79·4% (79·0–79·9). Including all ECGs acquired during the first month of each patient's window of interest (ie, the study start date or 31 days before the first recorded atrial fibrillation ECG) increased the AUC to 0·90 (0·90–0·91), sensitivity to 82·3% (80·9–83·6), specificity to 83·4% (83·0–83·8), F1 score to 45·4% (44·2–46·5), and overall accuracy to 83·3% (83·0–83·7).
An AI-enabled ECG acquired during normal sinus rhythm permits identification at point of care of individuals with atrial fibrillation.
None.
The application of artificial intelligence (AI) to the electrocardiogram (ECG), a ubiquitous and standardized test, is an example of the ongoing transformative effect of AI on cardiovascular ...medicine. Although the ECG has long offered valuable insights into cardiac and non-cardiac health and disease, its interpretation requires considerable human expertise. Advanced AI methods, such as deep-learning convolutional neural networks, have enabled rapid, human-like interpretation of the ECG, while signals and patterns largely unrecognizable to human interpreters can be detected by multilayer AI networks with precision, making the ECG a powerful, non-invasive biomarker. Large sets of digital ECGs linked to rich clinical data have been used to develop AI models for the detection of left ventricular dysfunction, silent (previously undocumented and asymptomatic) atrial fibrillation and hypertrophic cardiomyopathy, as well as the determination of a person's age, sex and race, among other phenotypes. The clinical and population-level implications of AI-based ECG phenotyping continue to emerge, particularly with the rapid rise in the availability of mobile and wearable ECG technologies. In this Review, we summarize the current and future state of the AI-enhanced ECG in the detection of cardiovascular disease in at-risk populations, discuss its implications for clinical decision-making in patients with cardiovascular disease and critically appraise potential limitations and unknowns.
Hypertrophic cardiomyopathy (HCM) is an uncommon but important cause of sudden cardiac death.
This study sought to develop an artificial intelligence approach for the detection of HCM based on ...12-lead electrocardiography (ECG).
A convolutional neural network (CNN) was trained and validated using digital 12-lead ECG from 2,448 patients with a verified HCM diagnosis and 51,153 non-HCM age- and sex-matched control subjects. The ability of the CNN to detect HCM was then tested on a different dataset of 612 HCM and 12,788 control subjects.
In the combined datasets, mean age was 54.8 ± 15.9 years for the HCM group and 57.5 ± 15.5 years for the control group. After training and validation, the area under the curve (AUC) of the CNN in the validation dataset was 0.95 (95% confidence interval CI: 0.94 to 0.97) at the optimal probability threshold of 11% for having HCM. When applying this probability threshold to the testing dataset, the CNN’s AUC was 0.96 (95% CI: 0.95 to 0.96) with sensitivity 87% and specificity 90%. In subgroup analyses, the AUC was 0.95 (95% CI: 0.94 to 0.97) among patients with left ventricular hypertrophy by ECG criteria and 0.95 (95% CI: 0.90 to 1.00) among patients with a normal ECG. The model performed particularly well in younger patients (sensitivity 95%, specificity 92%). In patients with HCM with and without sarcomeric mutations, the model-derived median probabilities for having HCM were 97% and 96%, respectively.
ECG-based detection of HCM by an artificial intelligence algorithm can be achieved with high diagnostic performance, particularly in younger patients. This model requires further refinement and external validation, but it may hold promise for HCM screening.
Display omitted
Sex and age have long been known to affect the ECG. Several biologic variables and anatomic factors may contribute to sex and age-related differences on the ECG. We hypothesized that a convolutional ...neural network (CNN) could be trained through a process called deep learning to predict a person's age and self-reported sex using only 12-lead ECG signals. We further hypothesized that discrepancies between CNN-predicted age and chronological age may serve as a physiological measure of health.
We trained CNNs using 10-second samples of 12-lead ECG signals from 499 727 patients to predict sex and age. The networks were tested on a separate cohort of 275 056 patients. Subsequently, 100 randomly selected patients with multiple ECGs over the course of decades were identified to assess within-individual accuracy of CNN age estimation.
Of 275 056 patients tested, 52% were males and mean age was 58.6±16.2 years. For sex classification, the model obtained 90.4% classification accuracy with an area under the curve of 0.97 in the independent test data. Age was estimated as a continuous variable with an average error of 6.9±5.6 years (R-squared =0.7). Among 100 patients with multiple ECGs over the course of at least 2 decades of life, most patients (51%) had an average error between real age and CNN-predicted age of <7 years. Major factors seen among patients with a CNN-predicted age that exceeded chronologic age by >7 years included: low ejection fraction, hypertension, and coronary disease (P<0.01). In the 27% of patients where correlation was >0.8 between CNN-predicted and chronologic age, no incident events occurred over follow-up (33±12 years).
Applying artificial intelligence to the ECG allows prediction of patient sex and estimation of age. The ability of an artificial intelligence algorithm to determine physiological age, with further validation, may serve as a measure of overall health.
Abstract
Aims
Early detection of aortic stenosis (AS) is becoming increasingly important with a better outcome after aortic valve replacement in asymptomatic severe AS patients and a poor outcome in ...moderate AS. We aimed to develop artificial intelligence-enabled electrocardiogram (AI-ECG) using a convolutional neural network to identify patients with moderate to severe AS.
Methods and results
Between 1989 and 2019, 258 607 adults mean age 63 ± 16.3 years; women 122 790 (48%) with an echocardiography and an ECG performed within 180 days were identified from the Mayo Clinic database. Moderate to severe AS by echocardiography was present in 9723 (3.7%) patients. Artificial intelligence training was performed in 129 788 (50%), validation in 25 893 (10%), and testing in 102 926 (40%) randomly selected subjects. In the test group, the AI-ECG labelled 3833 (3.7%) patients as positive with the area under the curve (AUC) of 0.85. The sensitivity, specificity, and accuracy were 78%, 74%, and 74%, respectively. The sensitivity increased and the specificity decreased as age increased. Women had lower sensitivity but higher specificity compared with men at any age groups. The model performance increased when age and sex were added to the model (AUC 0.87), which further increased to 0.90 in patients without hypertension. Patients with false-positive AI-ECGs had twice the risk for developing moderate or severe AS in 15 years compared with true negative AI-ECGs (hazard ratio 2.18, 95% confidence interval 1.90–2.50).
Conclusion
An AI-ECG can identify patients with moderate or severe AS and may serve as a powerful screening tool for AS in the community.
Graphical Abstract
AI-ECG for Aortic Stenosis screening using convolutional neural network (CNN).
Heart rate-corrected QT interval (QTc) prolongation, whether secondary to drugs, genetics including congenital long QT syndrome, and/or systemic diseases including SARS-CoV-2-mediated coronavirus ...disease 2019 (COVID-19), can predispose to ventricular arrhythmias and sudden cardiac death. Currently, QTc assessment and monitoring relies largely on 12-lead electrocardiography. As such, we sought to train and validate an artificial intelligence (AI)-enabled 12-lead ECG algorithm to determine the QTc, and then prospectively test this algorithm on tracings acquired from a mobile ECG (mECG) device in a population enriched for repolarization abnormalities.
Using >1.6 million 12-lead ECGs from 538 200 patients, a deep neural network (DNN) was derived (patients for training, n = 250 767; patients for testing, n = 107 920) and validated (n = 179 513 patients) to predict the QTc using cardiologist-overread QTc values as the "gold standard". The ability of this DNN to detect clinically-relevant QTc prolongation (eg, QTc ≥500 ms) was then tested prospectively on 686 patients with genetic heart disease (50% with long QT syndrome) with QTc values obtained from both a 12-lead ECG and a prototype mECG device equivalent to the commercially-available AliveCor KardiaMobile 6L.
In the validation sample, strong agreement was observed between human over-read and DNN-predicted QTc values (-1.76±23.14 ms). Similarly, within the prospective, genetic heart disease-enriched dataset, the difference between DNN-predicted QTc values derived from mECG tracings and those annotated from 12-lead ECGs by a QT expert (-0.45±24.73 ms) and a commercial core ECG laboratory 10.52±25.64 ms was nominal. When applied to mECG tracings, the DNN's ability to detect a QTc value ≥500 ms yielded an area under the curve, sensitivity, and specificity of 0.97, 80.0%, and 94.4%, respectively.
Using smartphone-enabled electrodes, an AI DNN can predict accurately the QTc of a standard 12-lead ECG. QTc estimation from an AI-enabled mECG device may provide a cost-effective means of screening for both acquired and congenital long QT syndrome in a variety of clinical settings where standard 12-lead electrocardiography is not accessible or cost-effective.
There is no established screening approach for hypertrophic cardiomyopathy (HCM). We recently developed an artificial intelligence (AI) model for the detection of HCM based on the 12‑lead ...electrocardiogram (AI-ECG) in adults. Here, we aimed to validate this approach of ECG-based HCM detection in pediatric patients (age ≤ 18 years).
We identified a cohort of 300 children and adolescents with HCM (mean age 12.5 ± 4.6 years, male 68%) who had an ECG and echocardiogram at our institution. Patients were age- and sex-matched to 18,439 non-HCM controls. Diagnostic performance of the AI-ECG model for the detection of HCM was estimated using the previously identified optimal diagnostic threshold of 11% (the probability output derived by the model above which an ECG is considered to belong to an HCM patient).
Mean AI-ECG probabilities of HCM were 92% and 5% in the case and control groups, respectively. The area under the receiver operating characteristic curve (AUC) of the AI-ECG model for HCM detection was 0.98 (95% CI 0.98–0.99) with corresponding sensitivity 92% and specificity 95%. The positive and negative predictive values were 22% and 99%, respectively. The model performed similarly in males and females and in genotype-positive and genotype-negative HCM patients. Performance tended to be superior with increasing age. In the age subgroup <5 years, the test's AUC was 0.93. In comparison, the AUC was 0.99 in the age subgroup 15–18 years.
A deep-learning, AI model can detect pediatric HCM with high accuracy from the standard 12‑lead ECG.
•No established approach for screening for HCM exists currently.•A deep-learning AI algorithm applied on the ECG can detect HCM with high accuracy in children and adolescents.•The AI-ECG algorithm may have optimal performance in the oldest subgroup of pediatric patients (15–18 years).
PDF
