Artificial Intelligence-Enabled Electrocardiography Predicts Left Ventricular Dysfunction and Future Cardiovascular Outcomes: A Retrospective Analysis

Comprehensive clinical application analysis of artificial intelligence-enabled electrocardiograms for screening multiple valvular heart diseases

BACKGROUND

Valvular heart disease (VHD) is becoming increasingly important to manage the risk of future complications. Electrocardiographic (ECG) changes may be related to multiple VHDs, and (AI)-enabled ECG has been able to detect some VHDs. We aimed to develop five deep learning models (DLMs) to identify aortic stenosis, aortic regurgitation, pulmonary regurgitation, tricuspid regurgitation, and mitral regurgitation.

METHODS

Between 2010 and 2021, 77,047 patients with echocardiography and 12-lead ECG performed within 7 days were identified from an academic medical center to provide DLM development (122,728 ECGs), and internal validation (7,637 ECGs). Additional 11,800 patients from a community hospital were identified to external validation. The ECGs were classified as with or without moderate-to-severe VHDs according to transthoracic echocardiography (TTE) records, and we also collected the other echocardiographic data and follow-up TTE records to identify new-onset valvular heart diseases.

RESULTS

AI-ECG adjusted for age and sex achieved areas under the curves (AUCs) of >0.84, >0.80, >0.77, >0.83, and >0.81 for detecting aortic stenosis, aortic regurgitation, pulmonary regurgitation, tricuspid regurgitation, and mitral regurgitation, respectively. Since predictions of each DLM shared similar components of ECG rhythms, the positive findings of each DLM were highly correlated with other valvular heart diseases. Of note, a total of 37.5-51.7% of false-positive predictions had at least one significant echocardiographic finding, which may lead to a significantly higher risk of future moderate-to-severe VHDs in patients with initially minimal-to-mild VHDs.

CONCLUSION

AI-ECG may be used as a large-scale screening tool for detecting VHDs and a basis to undergo an echocardiography.

Estimating Ejection Fraction from the 12 Lead ECG among Patients with Acute Heart Failure

Background

Identifying patients with low left ventricular ejection fraction (LVEF) in the emergency department using an electrocardiogram (ECG) may optimize acute heart failure (AHF) management. We aimed to assess the efficacy of 527 automated 12-lead ECG features for estimating LVEF among patients with AHF.

Method

Medical records of patients >18 years old and AHF-related ICD codes, demographics, LVEF %, comorbidities, and medication were analyzed. Least Absolute Shrinkage and Selection Operator (LASSO) identified important ECG features and evaluated performance.

Results

Among 851 patients, the mean age was 74 years (IQR:11), male 56% (n=478), and the median body mass index was 29 kg/m2 (IQR:1.8). A total of 914 echocardiograms and ECGs were matched; the time between ECG-Echocardiogram was 9 hours (IQR of 9 hours); ≤30% LVEF (16.45%, n=140). Lasso demonstrated 42 ECG features important for estimating LVEF ≤30%. The predictive model of LVEF ≤30% demonstrated an area under the curve (AUC) of 0.86, a 95% confidence interval (CI) of 0.83 to 0.89, a specificity of 54% (50% to 57%), and a sensitivity of 91 (95% CI: 88% to 96%), accuracy 60% (95% CI:60 % to 63%) and, negative predictive value of 95%.

Conclusions

An explainable machine learning model with physiologically feasible predictors may be useful in screening patients with low LVEF in AHF.

Artificial intelligence-enabled electrocardiography contributes to hyperthyroidism detection and outcome prediction

BACKGROUND

Hyperthyroidism is frequently under-recognized and leads to heart failure and mortality. Timely identification of high-risk patients is a prerequisite to effective antithyroid therapy. Since the heart is very sensitive to hyperthyroidism and its electrical signature can be demonstrated by electrocardiography, we developed an artificial intelligence model to detect hyperthyroidism by electrocardiography and examined its potential for outcome prediction.

METHODS

The deep learning model was trained using a large dataset of 47,245 electrocardiograms from 33,246 patients at an academic medical center. Patients were included if electrocardiograms and measurements of serum thyroid-stimulating hormone were available that had been obtained within a three day period. Serum thyroid-stimulating hormone and free thyroxine were used to define overt and subclinical hyperthyroidism. We tested the model internally using 14,420 patients and externally using two additional test sets comprising 11,498 and 596 patients, respectively.

RESULTS

The performance of the deep learning model achieves areas under the receiver operating characteristic curves (AUCs) of 0.725-0.761 for hyperthyroidism detection, AUCs of 0.867-0.876 for overt hyperthyroidism, and AUC of 0.631-0.701 for subclinical hyperthyroidism, superior to a traditional features-based machine learning model. Patients identified as hyperthyroidism-positive by the deep learning model have a significantly higher risk (1.97-2.94 fold) of all-cause mortality and new-onset heart failure compared to hyperthyroidism-negative patients. This cardiovascular disease stratification is particularly pronounced in subclinical hyperthyroidism, surpassing that observed in overt hyperthyroidism.

CONCLUSIONS

An innovative algorithm effectively identifies overt and subclinical hyperthyroidism and contributes to cardiovascular risk assessment.

Artificial intelligence-based identification of left ventricular systolic dysfunction from 12-lead electrocardiograms: external validation and advanced application of an existing model

Aims

The diagnostic application of artificial intelligence (AI)-based models to detect cardiovascular diseases from electrocardiograms (ECGs) evolves, and promising results were reported. However, external validation is not available for all published algorithms. The aim of this study was to validate an existing algorithm for the detection of left ventricular systolic dysfunction (LVSD) from 12-lead ECGs.

Methods and results

Patients with digitalized data pairs of 12-lead ECGs and echocardiography (at intervals of ≤7 days) were retrospectively selected from the Heart Center Leipzig ECG and electronic medical records databases. A previously developed AI-based model was applied to ECGs and calculated probabilities for LVSD. The area under the receiver operating characteristic curve (AUROC) was computed overall and in cohorts stratified for baseline and ECG characteristics. Repeated echocardiography studies recorded ≥3 months after index diagnostics were used for follow-up (FU) analysis. At baseline, 42 291 ECG-echocardiography pairs were analysed, and AUROC for LVSD detection was 0.88. Sensitivity and specificity were 82% and 77% for the optimal LVSD probability cut-off based on Youden's J. AUROCs were lower in ECG subgroups with tachycardia, atrial fibrillation, and wide QRS complexes. In patients without LVSD at baseline and available FU, model-generated high probability for LVSD was associated with a four-fold increased risk of developing LVSD during FU.

Conclusion

We provide the external validation of an existing AI-based ECG-analysing model for the detection of LVSD with robust performance metrics. The association of false positive LVSD screenings at baseline with a deterioration of ventricular function during FU deserves a further evaluation in prospective trials.

Artificial Intelligence-Enabled Electrocardiography Detects B-Type Natriuretic Peptide and N-Terminal Pro-Brain Natriuretic Peptide

BACKGROUND: The B-type natriuretic peptide (BNP) and N-terminal pro-brain natriuretic peptide (pBNP) are predictors of cardiovascular morbidity and mortality. Since the artificial intelligence (AI)-enabled electrocardiogram (ECG) system is widely used in the management of many cardiovascular diseases (CVDs), patients requiring intensive monitoring may benefit from an AI-ECG with BNP/pBNP predictions. This study aimed to develop an AI-ECG to predict BNP/pBNP and compare their values for future mortality. METHODS: The development, tuning, internal validation, and external validation sets included 47,709, 16,249, 4001, and 6042 ECGs, respectively. Deep learning models (DLMs) were trained using a development set for estimating ECG-based BNP/pBNP (ECG-BNP/ECG-pBNP), and the tuning set was used to guide the training process. The ECGs in internal and external validation sets belonging to nonrepeating patients were used to validate the DLMs. We also followed-up all-cause mortality to explore the prognostic value. RESULTS: The DLMs accurately distinguished mild (≥500 pg/mL) and severe (≥1000 pg/mL) an abnormal BNP/pBNP with AUCs of ≥0.85 in the internal and external validation sets, which provided sensitivities of 68.0-85.0% and specificities of 77.9-86.2%. In continuous predictions, the Pearson correlation coefficient between ECG-BNP and ECG-pBNP was 0.93, and they were both associated with similar ECG features, such as the T wave axis and correct QT interval. ECG-pBNP provided a higher all-cause mortality predictive value than ECG-BNP. CONCLUSIONS: The AI-ECG can accurately estimate BNP/pBNP and may be useful for monitoring the risk of CVDs. Moreover, ECG-pBNP may be a better indicator to manage the risk of future mortality.

Multi-Modal Stacking Ensemble for the Diagnosis of Cardiovascular Diseases

BACKGROUND

Cardiovascular diseases (CVDs) are a leading cause of death worldwide. Deep learning methods have been widely used in the field of medical image analysis and have shown promising results in the diagnosis of CVDs.

METHODS

Experiments were performed on 12-lead electrocardiogram (ECG) databases collected by Chapman University and Shaoxing People's Hospital. The ECG signal of each lead was converted into a scalogram image and an ECG grayscale image and used to fine-tune the pretrained ResNet-50 model of each lead. The ResNet-50 model was used as a base learner for the stacking ensemble method. Logistic regression, support vector machine, random forest, and XGBoost were used as a meta learner by combining the predictions of the base learner. The study introduced a method called multi-modal stacking ensemble, which involves training a meta learner through a stacking ensemble that combines predictions from two modalities: scalogram images and ECG grayscale images.

RESULTS

The multi-modal stacking ensemble with a combination of ResNet-50 and logistic regression achieved an AUC of 0.995, an accuracy of 93.97%, a sensitivity of 0.940, a precision of 0.937, and an F1-score of 0.936, which are higher than those of LSTM, BiLSTM, individual base learners, simple averaging ensemble, and single-modal stacking ensemble methods.

CONCLUSION

The proposed multi-modal stacking ensemble approach showed effectiveness for diagnosing CVDs.

Artificial intelligence enabled ECG screening for left ventricular systolic dysfunction: a systematic review

Screening for left ventricular systolic dysfunction (LVSD), defined as reduced left ventricular ejection fraction (LVEF), deserves renewed interest as the medical treatment for the prevention and progression of heart failure improves. We aimed to review the updated literature to outline the potential and caveats of using artificial intelligence-enabled electrocardiography (AIeECG) as an opportunistic screening tool for LVSD.We searched PubMed and Cochrane for variations of the terms "ECG," "Heart Failure," "systolic dysfunction," and "Artificial Intelligence" from January 2010 to April 2022 and selected studies that reported the diagnostic accuracy and confounders of using AIeECG to detect LVSD.Out of 40 articles, we identified 15 relevant studies; eleven retrospective cohorts, three prospective cohorts, and one case series. Although various LVEF thresholds were used, AIeECG detected LVSD with a median AUC of 0.90 (IQR from 0.85 to 0.95), a sensitivity of 83.3% (IQR from 73 to 86.9%) and a specificity of 87% (IQR from 84.5 to 90.9%). AIeECG algorithms succeeded across a wide range of sex, age, and comorbidity and seemed especially useful in non-cardiology settings and when combined with natriuretic peptide testing. Furthermore, a false-positive AIeECG indicated a future development of LVSD. No studies investigated the effect on treatment or patient outcomes.This systematic review corroborates the arrival of a new generic biomarker, AIeECG, to improve the detection of LVSD. AIeECG, in addition to natriuretic peptides and echocardiograms, will improve screening for LVSD, but prospective randomized implementation trials with added therapy are needed to show cost-effectiveness and clinical significance.

A Deep Learning Algorithm for Detecting Acute Pericarditis by Electrocardiogram

(1) Background: Acute pericarditis is often confused with ST-segment elevation myocardial infarction (STEMI) among patients presenting with acute chest pain in the emergency department (ED). Since a deep learning model (DLM) has been validated to accurately identify STEMI cases via 12-lead electrocardiogram (ECG), this study aimed to develop another DLM for the detection of acute pericarditis in the ED. (2) Methods: This study included 128 ECGs from patients with acute pericarditis and 66,633 ECGs from patients visiting the ED between 1 January 2010 and 31 December 2020. The ECGs were randomly allocated based on patients to the training, tuning, and validation sets, at a 3:1:1 ratio. We used raw ECG signals to train a pericarditis-DLM and used traditional ECG features to train a machine learning model. A human–machine competition was conducted using a subset of the validation set, and the performance of the Philips automatic algorithm was also compared. STEMI cases in the validation set were extracted to analyze the DLM ability of differential diagnosis between acute pericarditis and STEMI using ECG. We also followed the hospitalization events in non-pericarditis cases to explore the meaning of false-positive predictions. (3) Results: The pericarditis-DLM exceeded the performance of all participating human experts and algorithms based on traditional ECG features in the human–machine competition. In the validation set, the pericarditis-DLM could detect acute pericarditis with an area under the receiver operating characteristic curve (AUC) of 0.954, a sensitivity of 78.9%, and a specificity of 97.7%. However, our pericarditis-DLM also misinterpreted 10.2% of STEMI ECGs as pericarditis cases. Therefore, we generated an integrating strategy combining pericarditis-DLM and a previously developed STEMI-DLM, which provided a sensitivity of 73.7% and specificity of 99.4%, to identify acute pericarditis in patients with chest pains. Compared to the true-negative cases, patients with false-positive results using this strategy were associated with higher risk of hospitalization within 3 days due to cardiac disorders (hazard ratio (HR): 8.09; 95% confidence interval (CI): 3.99 to 16.39). (4) Conclusions: The AI-enhanced algorithm may be a powerful tool to assist clinicians in the early detection of acute pericarditis and differentiate it from STEMI using 12-lead ECGs.

Artificial intelligence-enabled electrocardiogram screens low left ventricular ejection fraction with a degree of confidence

Background

Artificial intelligence-enabled electrocardiogram has become a substitute tool for echocardiography in left ventricular ejection fraction estimation. However, the direct use of artificial intelligence-enabled electrocardiogram may be not trustable due to the uncertainty of the prediction.

Objective

The study aimed to establish an artificial intelligence-enabled electrocardiogram with a degree of confidence to identify left ventricular dysfunction.

Methods

The study collected 76,081 and 11,771 electrocardiograms from an academic medical center and a community hospital to establish and validate the deep learning model, respectively. The proposed deep learning model provided the point estimation of the actual ejection fraction and its standard deviation derived from the maximum probability density function of a normal distribution. The primary analysis focused on the accuracy of identifying patients with left ventricular dysfunction (ejection fraction ≤ 40%). Since the standard deviation was an uncertainty indicator in a normal distribution, we used it as a degree of confidence in the artificial intelligence-enabled electrocardiogram. We further explored the clinical application of estimated standard deviation and followed up on the new-onset left ventricular dysfunction in patients with initially normal ejection fraction.

Results

The area under receiver operating characteristic curves (AUC) of detecting left ventricular dysfunction were 0.9549 and 0.9365 in internal and external validation sets. After excluding the cases with a lower degree of confidence, the artificial intelligence-enabled electrocardiogram performed better in the remaining cases in internal (AUC = 0.9759) and external (AUC = 0.9653) validation sets. For the application of future left ventricular dysfunction risk stratification in patients with initially normal ejection fraction, a 4.57-fold risk of future left ventricular dysfunction when the artificial intelligence-enabled electrocardiogram is positive in the internal validation set. The hazard ratio was increased to 8.67 after excluding the cases with a lower degree of confidence. This trend was also validated in the external validation set.

Conclusion

The deep learning model with a degree of confidence can provide advanced improvements in identifying left ventricular dysfunction and serve as a decision support and management-guided screening tool for prognosis.