Age and Sex Estimation Using Artificial Intelligence From Standard 12-Lead ECGs

Zachi I. Attia, Paul Andrew Friedman, Peter Noseworthy, Francisco Lopez-Jimenez, Dorothy J. Ladewig, Gaurav Satam, Patricia Pellikka, Thomas M. Munger, Samuel J Asirvatham, Christopher G. Scott, Rickey E. Carter, Suraj Kapa

Research output: Contribution to journalArticle

Abstract

BACKGROUND: 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. METHODS: 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. RESULTS: 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). CONCLUSIONS: 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.

Original languageEnglish (US)
Pages (from-to)e007284
JournalCirculation. Arrhythmia and electrophysiology
Volume12
Issue number9
DOIs
StatePublished - Sep 1 2019

Fingerprint

Artificial Intelligence
Electrocardiography
Aptitude
Health
Lead
Area Under Curve
Coronary Disease
Learning
Hypertension

Keywords

  • artificial intelligence
  • coronary disease
  • electrocardiography
  • hypertension
  • neural network

ASJC Scopus subject areas

  • Cardiology and Cardiovascular Medicine
  • Physiology (medical)

Cite this

Age and Sex Estimation Using Artificial Intelligence From Standard 12-Lead ECGs. / Attia, Zachi I.; Friedman, Paul Andrew; Noseworthy, Peter; Lopez-Jimenez, Francisco; Ladewig, Dorothy J.; Satam, Gaurav; Pellikka, Patricia; Munger, Thomas M.; Asirvatham, Samuel J; Scott, Christopher G.; Carter, Rickey E.; Kapa, Suraj.

In: Circulation. Arrhythmia and electrophysiology, Vol. 12, No. 9, 01.09.2019, p. e007284.

Research output: Contribution to journalArticle

@article{4abc605167584fd88a6c4dd4499129da,
title = "Age and Sex Estimation Using Artificial Intelligence From Standard 12-Lead ECGs",
abstract = "BACKGROUND: 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. METHODS: 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. RESULTS: 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). CONCLUSIONS: 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.",
keywords = "artificial intelligence, coronary disease, electrocardiography, hypertension, neural network",
author = "Attia, {Zachi I.} and Friedman, {Paul Andrew} and Peter Noseworthy and Francisco Lopez-Jimenez and Ladewig, {Dorothy J.} and Gaurav Satam and Patricia Pellikka and Munger, {Thomas M.} and Asirvatham, {Samuel J} and Scott, {Christopher G.} and Carter, {Rickey E.} and Suraj Kapa",
year = "2019",
month = "9",
day = "1",
doi = "10.1161/CIRCEP.119.007284",
language = "English (US)",
volume = "12",
pages = "e007284",
journal = "Circulation: Arrhythmia and Electrophysiology",
issn = "1941-3149",
publisher = "Lippincott Williams and Wilkins",
number = "9",

}

TY - JOUR

T1 - Age and Sex Estimation Using Artificial Intelligence From Standard 12-Lead ECGs

AU - Attia, Zachi I.

AU - Friedman, Paul Andrew

AU - Noseworthy, Peter

AU - Lopez-Jimenez, Francisco

AU - Ladewig, Dorothy J.

AU - Satam, Gaurav

AU - Pellikka, Patricia

AU - Munger, Thomas M.

AU - Asirvatham, Samuel J

AU - Scott, Christopher G.

AU - Carter, Rickey E.

AU - Kapa, Suraj

PY - 2019/9/1

Y1 - 2019/9/1

N2 - BACKGROUND: 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. METHODS: 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. RESULTS: 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). CONCLUSIONS: 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.

AB - BACKGROUND: 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. METHODS: 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. RESULTS: 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). CONCLUSIONS: 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.

KW - artificial intelligence

KW - coronary disease

KW - electrocardiography

KW - hypertension

KW - neural network

UR - http://www.scopus.com/inward/record.url?scp=85071631170&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85071631170&partnerID=8YFLogxK

U2 - 10.1161/CIRCEP.119.007284

DO - 10.1161/CIRCEP.119.007284

M3 - Article

C2 - 31450977

AN - SCOPUS:85071631170

VL - 12

SP - e007284

JO - Circulation: Arrhythmia and Electrophysiology

JF - Circulation: Arrhythmia and Electrophysiology

SN - 1941-3149

IS - 9

ER -