To develop an artificial intelligence (AI)–based tool to detect cardiac amyloidosis (CA) from a standard 12-lead electrocardiogram (ECG).Read more

To develop an artificial intelligence (AI)–based tool to detect cardiac amyloidosis (CA) from a standard 12-lead electrocardiogram (ECG). Less

Chagas cardiomyopathy is a frequent and severe manifestation of Chagas disease (CD) and it is a leading cause of morbidity and death in South America. The dilated cardiomyopathy in CD is often discovered only when patients present with symptomatic heart failure. We developed an artificial Read more

Chagas cardiomyopathy is a frequent and severe manifestation of Chagas disease (CD) and it is a leading cause of morbidity and death in South America. The dilated cardiomyopathy in CD is often discovered only when patients present with symptomatic heart failure. We developed an artificial intelligence electrocardiogram (AI-ECG) based algorithm for the detection of left ventricular systolic dysfunction (LVSD) using a deep convolutional neural network that may be an important tool to screen for LVSD in patients with CD, especially in limited-resource settings. In this study, we validated the AI-ECG algorithm for the first time in a sample of subjects with CD in Brazil. We included 1,437 subjects who had an ECG and echocardiogram, 1330 with CD. 958 were female (66.7%%), the mean age was 60.6 years, 839 (58.4%) had Chagas cardiomyopathy, and 99 patients (6.9%) had EF<=40%. For the detection of EF<=40%, the AI-ECG had an AUC of 0.813, with sensitivity 83.8%, specificity 54.3%, and a negative predictive value of 97.8% using the originally calculated threshold. The sensitivity was 72.7%, specificity 78.9%, negative predictive value 97.5%, positive predictive value 13%, and accuracy 77.3 using an optimal threshold based on the ROC. The AUC was 0.818 for detection of ejection fraction <=35% and 0.805 for the detection of EF<50%. The AI-augmented ECG can facilitate the screening and monitoring of patients with Chagas disease for early detection of LVSD to enable early treatment and to optimize the use of other resources like echocardiography. Less

There are an increasing number of diagnostic tests derived from artificial intelligence (AI) and machine learning algorithms. Spectrum bias can arise when a diagnostic test is derived from study populations with different disease spectra than the target population. This bias is well described Read more

There are an increasing number of diagnostic tests derived from artificial intelligence (AI) and machine learning algorithms. Spectrum bias can arise when a diagnostic test is derived from study populations with different disease spectra than the target population. This bias is well described studies of test performance, has not been previously evaluated in AI-derived algorithms. We used a real-world AI-derived electrocardiogram (AI-ECG) algorithm to detect severe aortic stenosis (AS) to demonstrate spectrum bias across the range of aortic value disease severity. Overall, 258,607 patients had valid ECG and echocardiograms pairs. Using optimal decision threshold, the area under the receiver operator curve was 0.87 and 0.91 for the general and limited models respectively. Sensitivity and specificity for the general model was 80% and 81% respectively, while for the limited model it was 84% and 84% respectively. When applying the AI-ECG derived from the limited cohort to patients in the general cohort, the sensitivity, specificity and AUC were 83%, 73% and 0.86 respectively. In general, models should be applied to classification tasks that are similar to their initial training and exposure. While AI-ECG in both general and limited models performed robustly in identifying severe AS, there is evidence that spectrum bias may exist based on disease-severity selection. While the effect of the bias may be modest in this example, clinicians should be aware of the existence of such a bias in AI-derived algorithms and methods to ensure proper interpretation of test performance and generalizability in clinical practice. Less

Rapid identification of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection is critical to management of the pandemic. We sought to investigate the use of artificial intelligence applied to the ECG to rule out acute COVID-19. A global, volunteer consortium from 4 continents Read more

Rapid identification of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection is critical to management of the pandemic. We sought to investigate the use of artificial intelligence applied to the ECG to rule out acute COVID-19. A global, volunteer consortium from 4 continents identified patients with ECGs obtained around the time of PCR confirmed COVID 19 diagnosis. Clinical characteristics and raw ECG data were collected. A convolutional neural network was trained using 26,153 ECGs (33.2% COVID positive), validated with 3,826 ECGS (33.3% positive) and tested on 7,870 ECGs not included in other sets (32.7% positive). Performance under different prevalence values was tested by adding control ECGs from a single high-volume site. The area under the curve (AUC) for detection of acute COVID 19 infection in the test group was 0.767 (95% CI: 0.756 to 0.778) (sensitivity 98%, specificity 10%, positive predictive value 37%, negative predictive value 91%). When 50,905 normal controls were added to adjust the COVID prevalence to approximately 5% (2,657/58,555), resulting in an AUC of 0.780 (95% CI: 0.771 to 0.790) with a specificity of 12.1% and a negative predictive value to 99.2%. Infection with SARS-CoV-2 results in electrocardiographic changes that permit the AI-ECG to be utilized as a rapid screening test with a high negative predictive value (99.2%). This may permit the development of ECG-based tools to rapidly screen individuals for pandemic control. Less

The objective of this study was to validate a novel artificial-intelligence electrocardiogram algorithm (AI-ECG) to detect left ventricular systolic dysfunction (LVSD) in an external population. LVSD, even when asymptomatic, confers increased morbidity and mortality. We recently derived AI-ECG to Read more

The objective of this study was to validate a novel artificial-intelligence electrocardiogram algorithm (AI-ECG) to detect left ventricular systolic dysfunction (LVSD) in an external population. LVSD, even when asymptomatic, confers increased morbidity and mortality. We recently derived AI-ECG to detect LVSD using ECGs based on a large sample of patients treated at the Mayo Clinic. We performed an external validation study with subjects from the Know Your Heart Study, a cross-sectional study of adults aged 35-69 years residing in two cities in Russia, who had undergone both ECG and transthoracic echocardiography. LVSD was defined as left ventricular ejection fraction ≤ 35%. We assessed the performance of the AI-ECG to identify LVSD in this distinct patient population. Among 4277 subjects in this external population-based validation study, 0.6% had LVSD (compared to 7.8% of the original clinical derivation study). The overall performance of the AI-ECG to detect LVSD was robust with an area under the receiver operating curve of 0.82. When using the LVSD probability cut-off of 0.256 from the original derivation study, the sensitivity, specificity, and accuracy in this population were 26.9%, 97.4%, 97.0%, respectively. Other probability cut-offs were analysed for different sensitivity values. The AI-ECG detected LVSD with robust test performance in a population that was very different from that used to develop the algorithm. Population-specific cut-offs may be necessary for clinical implementation. Differences in population characteristics, ECG and echocardiographic data quality may affect test performance. Less

Atrial fibrillation (AF) is an established risk factor for ischemic stroke, but it can be paroxysmal and may go undiagnosed. An artificial intelligence (AI)-enabled ECG acquired during normal sinus rhythm was recently shown to detect silent AF. The objective of this study was to determine if AI-ECG Read more

Atrial fibrillation (AF) is an established risk factor for ischemic stroke, but it can be paroxysmal and may go undiagnosed. An artificial intelligence (AI)-enabled ECG acquired during normal sinus rhythm was recently shown to detect silent AF. The objective of this study was to determine if AI-ECG AF score is associated with presence of cerebral infarcts. This study included 1,373 individuals. Average age was 69.6 years and 53% of participants were male. There were 136 (10%) individuals with ECG-confirmed AF; 1237 (90%) participants had no AF history. Of participants with AF, 23% (n=31) were on anticoagulation, 47% (n=64) were on antiplatelet and 18% (n=24) were on dual therapy. Only 1.3% (n=16) of patients without AF were on anticoagulation and 47% (n=578) were on antiplatelet therapy. Ischemic infarcts were detected in 214 (15.6%) patients. As a continuous measure AI-ECG was associated with infarcts but not after adjusting for age and sex (p=0.46). AI-ECG AF score > 0.5 was associated with infarcts (p < 0.001); even after adjusting for age and sex (p= 0.03). History of AF was also associated with infarcts after adjusting for age and sex (p = 0.018). AI-ECG AF score and history of AF were associated with presence of cerebral infarcts after adjusting for age and sex. This tool could be useful in select patients with cryptogenic stroke but further investigation would be required. Less

An artificial intelligence-augmented electrocardiogram (AI-ECG) can identify left ventricular systolic dysfunction (LVSD). We examined the accuracy of AI ECG for identification of LVSD (defined as LVEF ≤40% by transthoracic echocardiogram [TTE]) in cardiac intensive care unit (CICU) patients. We Read more

An artificial intelligence-augmented electrocardiogram (AI-ECG) can identify left ventricular systolic dysfunction (LVSD). We examined the accuracy of AI ECG for identification of LVSD (defined as LVEF ≤40% by transthoracic echocardiogram [TTE]) in cardiac intensive care unit (CICU) patients. We included unique Mayo Clinic CICU patients admitted from 2007 to 2018 who underwent AI-ECG and TTE within 7 days, at least one of which was during hospitalization. Discrimination of the AI-ECG for LVSD was determined using receiver-operator characteristic curve (AUC) values. We included 5680 patients with a mean age of 68 ± 15 years (37% females). Acute coronary syndrome (ACS) was present in 55%. LVSD was present in 34% of patients (mean LVEF 48 ± 16%). The AI-ECG had an AUC of 0.83 (95% confidence interval 0.82–0.84) for discrimination of LVSD. Using the optimal cut-off, the AI-ECG had 73%, specificity 78%, negative predictive value 85% and overall accuracy 76% for LVSD. AUC values were higher for patients aged <70 years (0.85 versus 0.80), males (0.84 versus 0.79), patients without ACS (0.86 versus 0.80), and patients who did not undergo revascularization (0.84 versus 0.80). The AI-ECG algorithm had very good discrimination for LVSD in this critically-ill CICU cohort with a high prevalence of LVSD. Performance was better in younger male patients and those without ACS, highlighting those CICU patients in whom screening for LVSD using AI ECG may be more effective. Less

"Pulmonary hypertension (PH) is a life-threatening disease that is typically detected after significant pulmonary vascular remodeling has occurred. Longer diagnostic delays are associated with higher mortality and there is a need for a simple, fast, non-invasive PH screening tool. Currently, Read more

"Pulmonary hypertension (PH) is a life-threatening disease that is typically detected after significant pulmonary vascular remodeling has occurred. Longer diagnostic delays are associated with higher mortality and there is a need for a simple, fast, non-invasive PH screening tool. Currently, electrocardiograms often only identify abnormalities in severe PH. However, deep learning-based algorithms may enable detection of early, subtle, disease-specific changes, and could allow this inexpensive and ubiquitous test to serve as a powerful screening tool for PH. We used convolutional neural networks (CNN) to develop an algorithm for PH using retrospective electrocardiogram voltage-time data from Mayo Clinic. Each standard 12-lead electrocardiogram was paired with right heart catheterization to define patients as PH or non-PH, and the non- PH group was supplemented with patients in whom PH was excluded by echocardiogram. PH was defined as mean pulmonary arterial pressure (mPAP) ≥25 mmHg (at rest or during drug or exercise challenge), and non-PH was defined as mPAP <21 mmHg or tricuspid regurgitation velocity ≤2.8 m/s, if mPAP was not available. All patients were then randomly partitioned into training (48%), validation (12%) and test sets (40%) for building, optimizing and testing the models, respectively. Models were trained using electrocardiograms performed within 1 month of PH diagnosis (diagnostic dataset) and performance was tested on the diagnostic dataset and on electrocardiograms from 6-18 months (pre-emptive dataset) and 36-60 months before diagnosis. Model performance was evaluated by calculating the area under the curve (AUC) of the receiver operating characteristic curve, sensitivity, specificity, and diagnostic odds ratios. In total, 56,612 unique patients were identified: 11,138 PH and 45,474 non-PH patients. Several model structures were tested, and the best performing were CNNs with residual connections incorporating the 12-lead voltage-time electrocardiogram data. The final model yielded an AUC, sensitivity and specificity, respectively, of 0.91, 83.5%, and 83.6% in the diagnostic test set and 0.86, 77.8% and 78.3% in the pre-emptive dataset (Table). AUC remained above 0.81 for detection of PH using electrocardiograms from 6-monthly intervals up to 5 years before diagnosis. Among the PH patients, 2,134 patients had pre-capillary PH, which includes some progressive but potentially treatable forms of PH, and AUC was 0.95 for detection of PH in this diagnostic dataset. The electrocardiogram algorithm was able to detect PH up to 5 years prior to diagnosis. This type of algorithm has the potential to accelerate diagnosis and management of PH." Less

Identification of systolic heart failure among patients presenting to the emergency department (ED) with acute dyspnea is challenging. The reasons for dyspnea are often multifactorial. A focused physical evaluation and diagnostic testing can lack sensitivity and specificity. The objective of this Read more

Identification of systolic heart failure among patients presenting to the emergency department (ED) with acute dyspnea is challenging. The reasons for dyspnea are often multifactorial. A focused physical evaluation and diagnostic testing can lack sensitivity and specificity. The objective of this study was to assess the accuracy of an artificial intelligence-enabled ECG to identify patients presenting with dyspnea who have left ventricular systolic dysfunction (LVSD). A total of 1606 patients were included. Median time from ECG to echocardiogram was 1 day (Q1: 1, Q3: 2). The artificial intelligence-enabled ECG algorithm identified LVSD with an area under the receiver operating characteristic curve of 0.89 (95% CI, 0.86-0.91) and accuracy of 85.9%. Sensitivity, specificity, negative predictive value, and positive predictive value were 74%, 87%, 97%, and 40%, respectively. To identify an ejection fraction <50%, the area under the receiver operating characteristic curve, accuracy, sensitivity, and specificity were 0.85 (95% CI, 0.83-0.88), 86%, 63%, and 91%, respectively. NT-proBNP (N-terminal pro-B-type natriuretic peptide) alone at a cutoff of >800 identified LVSD with an area under the receiver operating characteristic curve of 0.80 (95% CI, 0.76-0.84). The ECG is an inexpensive, ubiquitous, painless test which can be quickly obtained in the ED. It effectively identifies LVSD in selected patients presenting to the ED with dyspnea when analyzed with artificial intelligence and outperforms NT-proBNP. Less

Patients with moderate to severe aortic stenosis (AS) have increased mortality even when asymptomatic. We hypothesized, that artificial intelligence - (AI) enabled electrocardiogram (ECG) - an inexpensive, ubiquitous, 10 second test - could detect patients with moderate/severe AS. 263,570 Patients Read more

Patients with moderate to severe aortic stenosis (AS) have increased mortality even when asymptomatic. We hypothesized, that artificial intelligence - (AI) enabled electrocardiogram (ECG) - an inexpensive, ubiquitous, 10 second test - could detect patients with moderate/severe AS. 263,570 Patients with an echocardiogram (Echo) and ECG performed within 180 days of each other were included. Patients with past cardiac surgery, or pacemaker implantation prior to the echo were excluded. We trained a convolutional neural network (CNN) to detect AS of at least moderate severity (aortic valve velocity ≥ 3 m/sec and/or area ≤ 1.5 cm2), using a 12 lead ECG. The model that achieved the highest AUC on an internal validation set was selected, and the final performance was assessed on third independent dataset. Of the total cohort, 50% were used for training, 10% for internal validation and 40% for testing the network. Of 105,461 testing patients, 5,088 (4.8%) patients had at least moderate AS and were labeled as “positive”. The area under the receiver operating characteristic curve of the classifier was 0.85 (Fig. 1A). The overall sensitivity, specificity and accuracy were 78%, 75% and 75%, respectively. The predicted probabilities for moderate to severe AS by AI track well with AS progression determined by Echo (Fig. 1B). Adding AI to the 12 lead ECG can permit early detection of AS. The ECG may serve as a powerful tool to screen for asymptomatic aortic stenosis in the community. Less

Deep learning algorithms derived in homogeneous populations may be poorly generalizable and have the potential to reflect, perpetuate, and even exacerbate racial/ethnic disparities in health and health care. In this study, we aimed to (1) assess whether the performance of a deep learning algorithm Read more

Deep learning algorithms derived in homogeneous populations may be poorly generalizable and have the potential to reflect, perpetuate, and even exacerbate racial/ethnic disparities in health and health care. In this study, we aimed to (1) assess whether the performance of a deep learning algorithm designed to detect low left ventricular ejection fraction using the 12-lead ECG varies by race/ethnicity and to (2) determine whether its performance is determined by the derivation population or by racial variation in the ECG. Our study demonstrates that while ECG characteristics vary by race, this did not impact the ability of a convolutional neural network to predict low left ventricular ejection fraction from the ECG. We recommend reporting of performance among diverse ethnic, racial, age, and sex groups for all new artificial intelligence tools to ensure responsible use of artificial intelligence in medicine. Less

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). In the combined datasets, mean age was 54.8 ± 15.9 years for the HCM Read more

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). 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. Less

A deep learning algorithm to detect low ejection fraction (EF) using routine 12-lead electrocardiogram (ECG) has recently been developed and validated. The algorithm was incorporated into the electronic health record (EHR) to automatically screen for low EF, encouraging clinicians to obtain a Read more

A deep learning algorithm to detect low ejection fraction (EF) using routine 12-lead electrocardiogram (ECG) has recently been developed and validated. The algorithm was incorporated into the electronic health record (EHR) to automatically screen for low EF, encouraging clinicians to obtain a confirmatory transthoracic echocardiogram (TTE) for previously undiagnosed patients, thereby facilitating early diagnosis and treatment. This trial examines the effectiveness of the AI-enabled ECG for detection of asymptomatic low EF in routine primary care practices and will be among the first to prospectively evaluate the value of AI in real-world practice. Its findings inform the future implementation strategies for the translation of other AI-enabled algorithms. Less

Machine learning applications in cardiology have rapidly evolved in the past decade. With the availability of machine learning tools coupled with vast data sources, the management of atrial fibrillation (AF), a common chronic disease with significant associated morbidity and socioeconomic impact, Read more

Machine learning applications in cardiology have rapidly evolved in the past decade. With the availability of machine learning tools coupled with vast data sources, the management of atrial fibrillation (AF), a common chronic disease with significant associated morbidity and socioeconomic impact, is undergoing a knowledge and practice transformation in the increasingly complex healthcare environment. Among other advances, deep-learning machine learning methods, including convolutional neural networks, have enabled the development of AF screening pathways using the ubiquitous 12-lead ECG to detect asymptomatic paroxysmal AF in at-risk populations (such as those with cryptogenic stroke), the refinement of AF and stroke prediction schemes through comprehensive digital phenotyping using structured and unstructured data abstraction from the electronic health record or wearable monitoring technologies, and the optimization of treatment strategies, ranging from stroke prophylaxis to monitoring of antiarrhythmic drug (AAD) therapy. Although the clinical and population-wide impact of these tools continues to be elucidated, such transformative progress does not come without challenges, such as the concerns about adopting black box technologies, assessing input data quality for training such models, and the risk of perpetuating rather than alleviating health disparities. This review critically appraises the advances of machine learning related to the care of AF thus far, their potential future directions, and its potential limitations and challenges. Less

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 Read more

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. 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. Less