Development and Validation of a Deep-Learning Model to Screen for Hyperkalemia from the Electrocardiogram

Conner D. Galloway, Alexander V. Valys, Jacqueline B. Shreibati, Daniel L. Treiman, Frank L. Petterson, Vivek P. Gundotra, David E. Albert, Zachi I. Attia, Rickey E. Carter, Samuel J Asirvatham, Michael John Ackerman, Peter Noseworthy, John J. Dillon, Paul Andrew Friedman

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

Importance: For patients with chronic kidney disease (CKD), hyperkalemia is common, associated with fatal arrhythmias, and often asymptomatic, while guideline-directed monitoring of serum potassium is underused. A deep-learning model that enables noninvasive hyperkalemia screening from the electrocardiogram (ECG) may improve detection of this life-threatening condition. Objective: To evaluate the performance of a deep-learning model in detection of hyperkalemia from the ECG in patients with CKD. Design, Setting, and Participants: A deep convolutional neural network (DNN) was trained using 1576581 ECGs from 449380 patients seen at Mayo Clinic, Rochester, Minnesota, from 1994 to 2017. The DNN was trained using 2 (leads I and II) or 4 (leads I, II, V3, and V5) ECG leads to detect serum potassium levels of 5.5 mEq/L or less (to convert to millimoles per liter, multiply by 1) and was validated using retrospective data from the Mayo Clinic in Minnesota, Florida, and Arizona. The validation included 61965 patients with stage 3 or greater CKD. Each patient had a serum potassium count drawn within 4 hours after their ECG was recorded. Data were analyzed between April 12, 2018, and June 25, 2018. Exposures: Use of a deep-learning model. Main Outcomes and Measures: Area under the receiver operating characteristic curve (AUC) and sensitivity and specificity, with serum potassium level as the reference standard. The model was evaluated at 2 operating points, 1 for equal specificity and sensitivity and another for high (90%) sensitivity. Results: Of the total 1638546 ECGs, 908000 (55%) were from men. The prevalence of hyperkalemia in the 3 validation data sets ranged from 2.6% (n = 1282 of 50099; Minnesota) to 4.8% (n = 287 of 6011; Florida). Using ECG leads I and II, the AUC of the deep-learning model was 0.883 (95% CI, 0.873-0.893) for Minnesota, 0.860 (95% CI, 0.837-0.883) for Florida, and 0.853 (95% CI, 0.830-0.877) for Arizona. Using a 90% sensitivity operating point, the sensitivity was 90.2% (95% CI, 88.4%-91.7%) and specificity was 63.2% (95% CI, 62.7%-63.6%) for Minnesota; the sensitivity was 91.3% (95% CI, 87.4%-94.3%) and specificity was 54.7% (95% CI, 53.4%-56.0%) for Florida; and the sensitivity was 88.9% (95% CI, 84.5%-92.4%) and specificity was 55.0% (95% CI, 53.7%-56.3%) for Arizona. Conclusions and Relevance: In this study, using only 2 ECG leads, a deep-learning model detected hyperkalemia in patients with renal disease with an AUC of 0.853 to 0.883. The application of artificial intelligence to the ECG may enable screening for hyperkalemia. Prospective studies are warranted.

Original languageEnglish (US)
JournalJAMA cardiology
DOIs
StatePublished - Jan 1 2019

Fingerprint

Hyperkalemia
Electrocardiography
Learning
Potassium
Chronic Renal Insufficiency
Area Under Curve
Serum
Sensitivity and Specificity
Artificial Intelligence
ROC Curve
Cardiac Arrhythmias
Outcome Assessment (Health Care)
Prospective Studies
Guidelines
Kidney

ASJC Scopus subject areas

  • Cardiology and Cardiovascular Medicine

Cite this

Development and Validation of a Deep-Learning Model to Screen for Hyperkalemia from the Electrocardiogram. / Galloway, Conner D.; Valys, Alexander V.; Shreibati, Jacqueline B.; Treiman, Daniel L.; Petterson, Frank L.; Gundotra, Vivek P.; Albert, David E.; Attia, Zachi I.; Carter, Rickey E.; Asirvatham, Samuel J; Ackerman, Michael John; Noseworthy, Peter; Dillon, John J.; Friedman, Paul Andrew.

In: JAMA cardiology, 01.01.2019.

Research output: Contribution to journalArticle

Galloway, Conner D. ; Valys, Alexander V. ; Shreibati, Jacqueline B. ; Treiman, Daniel L. ; Petterson, Frank L. ; Gundotra, Vivek P. ; Albert, David E. ; Attia, Zachi I. ; Carter, Rickey E. ; Asirvatham, Samuel J ; Ackerman, Michael John ; Noseworthy, Peter ; Dillon, John J. ; Friedman, Paul Andrew. / Development and Validation of a Deep-Learning Model to Screen for Hyperkalemia from the Electrocardiogram. In: JAMA cardiology. 2019.
@article{c7c8f73006924cc1b6db7678656f8253,
title = "Development and Validation of a Deep-Learning Model to Screen for Hyperkalemia from the Electrocardiogram",
abstract = "Importance: For patients with chronic kidney disease (CKD), hyperkalemia is common, associated with fatal arrhythmias, and often asymptomatic, while guideline-directed monitoring of serum potassium is underused. A deep-learning model that enables noninvasive hyperkalemia screening from the electrocardiogram (ECG) may improve detection of this life-threatening condition. Objective: To evaluate the performance of a deep-learning model in detection of hyperkalemia from the ECG in patients with CKD. Design, Setting, and Participants: A deep convolutional neural network (DNN) was trained using 1576581 ECGs from 449380 patients seen at Mayo Clinic, Rochester, Minnesota, from 1994 to 2017. The DNN was trained using 2 (leads I and II) or 4 (leads I, II, V3, and V5) ECG leads to detect serum potassium levels of 5.5 mEq/L or less (to convert to millimoles per liter, multiply by 1) and was validated using retrospective data from the Mayo Clinic in Minnesota, Florida, and Arizona. The validation included 61965 patients with stage 3 or greater CKD. Each patient had a serum potassium count drawn within 4 hours after their ECG was recorded. Data were analyzed between April 12, 2018, and June 25, 2018. Exposures: Use of a deep-learning model. Main Outcomes and Measures: Area under the receiver operating characteristic curve (AUC) and sensitivity and specificity, with serum potassium level as the reference standard. The model was evaluated at 2 operating points, 1 for equal specificity and sensitivity and another for high (90{\%}) sensitivity. Results: Of the total 1638546 ECGs, 908000 (55{\%}) were from men. The prevalence of hyperkalemia in the 3 validation data sets ranged from 2.6{\%} (n = 1282 of 50099; Minnesota) to 4.8{\%} (n = 287 of 6011; Florida). Using ECG leads I and II, the AUC of the deep-learning model was 0.883 (95{\%} CI, 0.873-0.893) for Minnesota, 0.860 (95{\%} CI, 0.837-0.883) for Florida, and 0.853 (95{\%} CI, 0.830-0.877) for Arizona. Using a 90{\%} sensitivity operating point, the sensitivity was 90.2{\%} (95{\%} CI, 88.4{\%}-91.7{\%}) and specificity was 63.2{\%} (95{\%} CI, 62.7{\%}-63.6{\%}) for Minnesota; the sensitivity was 91.3{\%} (95{\%} CI, 87.4{\%}-94.3{\%}) and specificity was 54.7{\%} (95{\%} CI, 53.4{\%}-56.0{\%}) for Florida; and the sensitivity was 88.9{\%} (95{\%} CI, 84.5{\%}-92.4{\%}) and specificity was 55.0{\%} (95{\%} CI, 53.7{\%}-56.3{\%}) for Arizona. Conclusions and Relevance: In this study, using only 2 ECG leads, a deep-learning model detected hyperkalemia in patients with renal disease with an AUC of 0.853 to 0.883. The application of artificial intelligence to the ECG may enable screening for hyperkalemia. Prospective studies are warranted.",
author = "Galloway, {Conner D.} and Valys, {Alexander V.} and Shreibati, {Jacqueline B.} and Treiman, {Daniel L.} and Petterson, {Frank L.} and Gundotra, {Vivek P.} and Albert, {David E.} and Attia, {Zachi I.} and Carter, {Rickey E.} and Asirvatham, {Samuel J} and Ackerman, {Michael John} and Peter Noseworthy and Dillon, {John J.} and Friedman, {Paul Andrew}",
year = "2019",
month = "1",
day = "1",
doi = "10.1001/jamacardio.2019.0640",
language = "English (US)",
journal = "JAMA Cardiology",
issn = "2380-6583",
publisher = "American Medical Association",

}

TY - JOUR

T1 - Development and Validation of a Deep-Learning Model to Screen for Hyperkalemia from the Electrocardiogram

AU - Galloway, Conner D.

AU - Valys, Alexander V.

AU - Shreibati, Jacqueline B.

AU - Treiman, Daniel L.

AU - Petterson, Frank L.

AU - Gundotra, Vivek P.

AU - Albert, David E.

AU - Attia, Zachi I.

AU - Carter, Rickey E.

AU - Asirvatham, Samuel J

AU - Ackerman, Michael John

AU - Noseworthy, Peter

AU - Dillon, John J.

AU - Friedman, Paul Andrew

PY - 2019/1/1

Y1 - 2019/1/1

N2 - Importance: For patients with chronic kidney disease (CKD), hyperkalemia is common, associated with fatal arrhythmias, and often asymptomatic, while guideline-directed monitoring of serum potassium is underused. A deep-learning model that enables noninvasive hyperkalemia screening from the electrocardiogram (ECG) may improve detection of this life-threatening condition. Objective: To evaluate the performance of a deep-learning model in detection of hyperkalemia from the ECG in patients with CKD. Design, Setting, and Participants: A deep convolutional neural network (DNN) was trained using 1576581 ECGs from 449380 patients seen at Mayo Clinic, Rochester, Minnesota, from 1994 to 2017. The DNN was trained using 2 (leads I and II) or 4 (leads I, II, V3, and V5) ECG leads to detect serum potassium levels of 5.5 mEq/L or less (to convert to millimoles per liter, multiply by 1) and was validated using retrospective data from the Mayo Clinic in Minnesota, Florida, and Arizona. The validation included 61965 patients with stage 3 or greater CKD. Each patient had a serum potassium count drawn within 4 hours after their ECG was recorded. Data were analyzed between April 12, 2018, and June 25, 2018. Exposures: Use of a deep-learning model. Main Outcomes and Measures: Area under the receiver operating characteristic curve (AUC) and sensitivity and specificity, with serum potassium level as the reference standard. The model was evaluated at 2 operating points, 1 for equal specificity and sensitivity and another for high (90%) sensitivity. Results: Of the total 1638546 ECGs, 908000 (55%) were from men. The prevalence of hyperkalemia in the 3 validation data sets ranged from 2.6% (n = 1282 of 50099; Minnesota) to 4.8% (n = 287 of 6011; Florida). Using ECG leads I and II, the AUC of the deep-learning model was 0.883 (95% CI, 0.873-0.893) for Minnesota, 0.860 (95% CI, 0.837-0.883) for Florida, and 0.853 (95% CI, 0.830-0.877) for Arizona. Using a 90% sensitivity operating point, the sensitivity was 90.2% (95% CI, 88.4%-91.7%) and specificity was 63.2% (95% CI, 62.7%-63.6%) for Minnesota; the sensitivity was 91.3% (95% CI, 87.4%-94.3%) and specificity was 54.7% (95% CI, 53.4%-56.0%) for Florida; and the sensitivity was 88.9% (95% CI, 84.5%-92.4%) and specificity was 55.0% (95% CI, 53.7%-56.3%) for Arizona. Conclusions and Relevance: In this study, using only 2 ECG leads, a deep-learning model detected hyperkalemia in patients with renal disease with an AUC of 0.853 to 0.883. The application of artificial intelligence to the ECG may enable screening for hyperkalemia. Prospective studies are warranted.

AB - Importance: For patients with chronic kidney disease (CKD), hyperkalemia is common, associated with fatal arrhythmias, and often asymptomatic, while guideline-directed monitoring of serum potassium is underused. A deep-learning model that enables noninvasive hyperkalemia screening from the electrocardiogram (ECG) may improve detection of this life-threatening condition. Objective: To evaluate the performance of a deep-learning model in detection of hyperkalemia from the ECG in patients with CKD. Design, Setting, and Participants: A deep convolutional neural network (DNN) was trained using 1576581 ECGs from 449380 patients seen at Mayo Clinic, Rochester, Minnesota, from 1994 to 2017. The DNN was trained using 2 (leads I and II) or 4 (leads I, II, V3, and V5) ECG leads to detect serum potassium levels of 5.5 mEq/L or less (to convert to millimoles per liter, multiply by 1) and was validated using retrospective data from the Mayo Clinic in Minnesota, Florida, and Arizona. The validation included 61965 patients with stage 3 or greater CKD. Each patient had a serum potassium count drawn within 4 hours after their ECG was recorded. Data were analyzed between April 12, 2018, and June 25, 2018. Exposures: Use of a deep-learning model. Main Outcomes and Measures: Area under the receiver operating characteristic curve (AUC) and sensitivity and specificity, with serum potassium level as the reference standard. The model was evaluated at 2 operating points, 1 for equal specificity and sensitivity and another for high (90%) sensitivity. Results: Of the total 1638546 ECGs, 908000 (55%) were from men. The prevalence of hyperkalemia in the 3 validation data sets ranged from 2.6% (n = 1282 of 50099; Minnesota) to 4.8% (n = 287 of 6011; Florida). Using ECG leads I and II, the AUC of the deep-learning model was 0.883 (95% CI, 0.873-0.893) for Minnesota, 0.860 (95% CI, 0.837-0.883) for Florida, and 0.853 (95% CI, 0.830-0.877) for Arizona. Using a 90% sensitivity operating point, the sensitivity was 90.2% (95% CI, 88.4%-91.7%) and specificity was 63.2% (95% CI, 62.7%-63.6%) for Minnesota; the sensitivity was 91.3% (95% CI, 87.4%-94.3%) and specificity was 54.7% (95% CI, 53.4%-56.0%) for Florida; and the sensitivity was 88.9% (95% CI, 84.5%-92.4%) and specificity was 55.0% (95% CI, 53.7%-56.3%) for Arizona. Conclusions and Relevance: In this study, using only 2 ECG leads, a deep-learning model detected hyperkalemia in patients with renal disease with an AUC of 0.853 to 0.883. The application of artificial intelligence to the ECG may enable screening for hyperkalemia. Prospective studies are warranted.

UR - http://www.scopus.com/inward/record.url?scp=85064052220&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85064052220&partnerID=8YFLogxK

U2 - 10.1001/jamacardio.2019.0640

DO - 10.1001/jamacardio.2019.0640

M3 - Article

JO - JAMA Cardiology

JF - JAMA Cardiology

SN - 2380-6583

ER -