TY - JOUR
T1 - Using Deep-Learning Algorithms to Simultaneously Identify Right and Left Ventricular Dysfunction From the Electrocardiogram
AU - Vaid, Akhil
AU - Johnson, Kipp W.
AU - Badgeley, Marcus A.
AU - Somani, Sulaiman S.
AU - Bicak, Mesude
AU - Landi, Isotta
AU - Russak, Adam
AU - Zhao, Shan
AU - Levin, Matthew A.
AU - Freeman, Robert S.
AU - Charney, Alexander W.
AU - Kukar, Atul
AU - Kim, Bette
AU - Danilov, Tatyana
AU - Lerakis, Stamatios
AU - Argulian, Edgar
AU - Narula, Jagat
AU - Nadkarni, Girish N.
AU - Glicksberg, Benjamin S.
N1 - Publisher Copyright:
© 2022 The Authors
PY - 2022/3
Y1 - 2022/3
N2 - Objectives: This study sought to develop DL models capable of comprehensively quantifying left and right ventricular dysfunction from ECG data in a large, diverse population. Background: Rapid evaluation of left and right ventricular function using deep learning (DL) on electrocardiograms (ECGs) can assist diagnostic workflow. However, DL tools to estimate right ventricular (RV) function do not exist, whereas those to estimate left ventricular (LV) function are restricted to quantification of very low LV function only. Methods: A multicenter study was conducted with data from 5 New York City hospitals: 4 for internal testing and 1 serving as external validation. We created novel DL models to classify left ventricular ejection fraction (LVEF) into categories derived from the latest universal definition of heart failure, estimate LVEF through regression, and predict a composite outcome of either RV systolic dysfunction or RV dilation. Results: We obtained echocardiogram LVEF estimates for 147,636 patients paired to 715,890 ECGs. We used natural language processing (NLP) to extract RV size and systolic function information from 404,502 echocardiogram reports paired to 761,510 ECGs for 148,227 patients. For LVEF classification in internal testing, area under curve (AUC) at detection of LVEF ≤40%, 40% < LVEF ≤50%, and LVEF >50% was 0.94 (95% CI: 0.94-0.94), 0.82 (95% CI: 0.81-0.83), and 0.89 (95% CI: 0.89-0.89), respectively. For external validation, these results were 0.94 (95% CI: 0.94-0.95), 0.73 (95% CI: 0.72-0.74), and 0.87 (95% CI: 0.87-0.88). For regression, the mean absolute error was 5.84% (95% CI: 5.82%-5.85%) for internal testing and 6.14% (95% CI: 6.13%-6.16%) in external validation. For prediction of the composite RV outcome, AUC was 0.84 (95% CI: 0.84-0.84) in both internal testing and external validation. Conclusions: DL on ECG data can be used to create inexpensive screening, diagnostic, and predictive tools for both LV and RV dysfunction. Such tools may bridge the applicability of ECGs and echocardiography and enable prioritization of patients for further interventions for either sided failure progressing to biventricular disease.
AB - Objectives: This study sought to develop DL models capable of comprehensively quantifying left and right ventricular dysfunction from ECG data in a large, diverse population. Background: Rapid evaluation of left and right ventricular function using deep learning (DL) on electrocardiograms (ECGs) can assist diagnostic workflow. However, DL tools to estimate right ventricular (RV) function do not exist, whereas those to estimate left ventricular (LV) function are restricted to quantification of very low LV function only. Methods: A multicenter study was conducted with data from 5 New York City hospitals: 4 for internal testing and 1 serving as external validation. We created novel DL models to classify left ventricular ejection fraction (LVEF) into categories derived from the latest universal definition of heart failure, estimate LVEF through regression, and predict a composite outcome of either RV systolic dysfunction or RV dilation. Results: We obtained echocardiogram LVEF estimates for 147,636 patients paired to 715,890 ECGs. We used natural language processing (NLP) to extract RV size and systolic function information from 404,502 echocardiogram reports paired to 761,510 ECGs for 148,227 patients. For LVEF classification in internal testing, area under curve (AUC) at detection of LVEF ≤40%, 40% < LVEF ≤50%, and LVEF >50% was 0.94 (95% CI: 0.94-0.94), 0.82 (95% CI: 0.81-0.83), and 0.89 (95% CI: 0.89-0.89), respectively. For external validation, these results were 0.94 (95% CI: 0.94-0.95), 0.73 (95% CI: 0.72-0.74), and 0.87 (95% CI: 0.87-0.88). For regression, the mean absolute error was 5.84% (95% CI: 5.82%-5.85%) for internal testing and 6.14% (95% CI: 6.13%-6.16%) in external validation. For prediction of the composite RV outcome, AUC was 0.84 (95% CI: 0.84-0.84) in both internal testing and external validation. Conclusions: DL on ECG data can be used to create inexpensive screening, diagnostic, and predictive tools for both LV and RV dysfunction. Such tools may bridge the applicability of ECGs and echocardiography and enable prioritization of patients for further interventions for either sided failure progressing to biventricular disease.
KW - artificial intelligence
KW - echocardiography
KW - electrocardiogram
KW - left heart failure
KW - left ventricular ejection fraction
KW - right heart failure
UR - http://www.scopus.com/inward/record.url?scp=85125198210&partnerID=8YFLogxK
U2 - 10.1016/j.jcmg.2021.08.004
DO - 10.1016/j.jcmg.2021.08.004
M3 - Article
C2 - 34656465
AN - SCOPUS:85125198210
SN - 1936-878X
VL - 15
SP - 395
EP - 410
JO - JACC: Cardiovascular Imaging
JF - JACC: Cardiovascular Imaging
IS - 3
ER -