Canadian Journal of Cardiology

Usefulness of Machine Learning-Based Detection and Classification of Cardiac Arrhythmias With 12-Lead Electrocardiograms

Published:March 05, 2020DOI:



      Deep-learning algorithms to annotate electrocardiograms (ECGs) and classify different types of cardiac arrhythmias with the use of a single-lead ECG input data set have been developed. It remains to be determined whether these algorithms can be generalized to 12-lead ECG-based rhythm classification.


      We used a long short-term memory (LSTM) model to detect 12 heart rhythm classes with the use of 65,932 digital 12-lead ECG signals from 38,899 patients, using annotations obtained by consensus of 3 board-certified electrophysiologists as the criterion standard.


      The accuracy of the LSTM model for the classification of each of the 12 heart rhythms was ≥ 0.982 (range 0.982-1.0), with an area under the receiver operating characteristic curve of ≥ 0.987 (range 0.987-1.0). The precision and recall ranged from 0.692 to 1 and from 0.625 to 1, respectively, with an F1 score of ≥ 0.777 (range 0.777-1.0). The accuracy of the model (0.90) was superior to the mean accuracies of internists (0.55), emergency physicians (0.73), and cardiologists (0.83).


      We demonstrated the feasibility and effectiveness of the deep-learning LSTM model for interpreting 12 common heart rhythms according to 12-lead ECG signals. The findings may have clinical relevance for the early diagnosis of cardiac rhythm disorders.



      Des algorithmes d’apprentissage profond conçus pour annoter les électrocardiogrammes (ECG) et classifier différents types d’arythmie cardiaque à partir des données d’un ECG à une seule dérivation ont été mis au point. Nous avons tenté de déterminer si ces algorithmes peuvent être généralisés pour obtenir une classification à partir des données d’un ECG à 12 dérivations.


      Nous avons utilisé un modèle LSTM (Long Short-Term Memory) pour reconnaître 12 catégories de rythmes cardiaques à partir de 65 932 signaux numériques d’ECG à 12 dérivations obtenus auprès de 38 899 patients; nous avons utilisé les annotations consensuelles établies par trois électrophysiologistes spécialisés comme critères de référence.


      L’exactitude du modèle LSTM utilisé pour classifier les rythmes cardiaques de chacune des 12 catégories s’établissait à ≥ 0,982 (plage : de 0,982 à 1,0), l’aire sous la courbe caractéristique de la performance du test étant de ≥ 0,987 (plage : de 0,987 à 1,0). La précision et le rappel allaient de 0,692 à 1 et de 0,625 à 1, respectivement, le score F1 s’établissant à ≥ 0,777 (plage : de 0,777 à 1,0). L’exactitude du modèle (0,90) était supérieure à l’exactitude moyenne des internistes (0,55), des urgentologues (0,73) et des cardiologues (0,83).


      Nous avons démontré la faisabilité et l’efficacité de l’emploi du modèle d’apprentissage profond LSTM pour l’interprétation de 12 rythmes cardiaques courants à partir des signaux d’un ECG à 12 dérivations. Ces résultats pourraient être utiles sur le plan clinique aux fins du diagnostic précoce des arythmies.
      To read this article in full you will need to make a payment

      Purchase one-time access:

      Academic & Personal: 24 hour online accessCorporate R&D Professionals: 24 hour online access
      One-time access price info
      • For academic or personal research use, select 'Academic and Personal'
      • For corporate R&D use, select 'Corporate R&D Professionals'


      Subscribe to Canadian Journal of Cardiology
      Already a print subscriber? Claim online access
      Already an online subscriber? Sign in
      Institutional Access: Sign in to ScienceDirect


        • LeCun Y.
        • Bengio Y.
        • Hinton G.
        Deep learning.
        Nature. 2015; 521: 436-444
        • Lee J.G.
        • Jun S.
        • Cho Y.W.
        • et al.
        Deep learning in medical imaging: general overview.
        Korean J Radiol. 2017; 18: 570-584
        • Becker A.S.
        • Marcon M.
        • Ghafoor S.
        • et al.
        Deep learning in mammography: diagnostic accuracy of a multipurpose image analysis software in the detection of breast cancer.
        Invest Radiol. 2017; 52: 434-440
        • Becker A.S.
        • Bluthgen C.
        • Phi Van V.D.
        • et al.
        Detection of tuberculosis patterns in digital photographs of chest X-ray images using deep learning: feasibility study.
        Int J Tuberc Lung Dis. 2018; 22: 328-335
        • Baltruschat I.M.
        • Nickisch H.
        • Grass M.
        • Knopp T.
        • Saalbach A.
        Comparison of deep learning approaches for multi-label chest X-ray classification.
        Sci Rep. 2019; 9: 6381
        • Pehrson L.M.
        • Lauridsen C.
        • Nielsen M.B.
        Machine learning and deep learning applied in ultrasound.
        Ultraschall Med. 2018; 39: 379-381
        • Akkus Z.
        • Galimzianova A.
        • Hoogi A.
        • Rubin D.L.
        • Erickson B.J.
        Deep learning for brain MRI segmentation: state of the art and future directions.
        J Digit Imaging. 2017; 30: 449-459
        • Al’Aref S.J.
        • Anchouche K.
        • Singh G.
        • et al.
        Clinical applications of machine learning in cardiovascular disease and its relevance to cardiac imaging.
        Eur Heart J. 2019; 40: 1975-1986
        • Schlant R.C.
        • Adolph R.J.
        • DiMarco J.P.
        • et al.
        Guidelines for electrocardiography. A report of the American College of Cardiology/American Heart Association Task Force on Assessment of Diagnostic and Therapeutic Cardiovascular Procedures (Committee on Electrocardiography).
        J Am Coll Cardiol. 1992; 19: 473-481
        • Schlapfer J.
        • Wellens H.J.
        Computer-interpreted electrocardiograms: benefits and limitations.
        J Am Coll Cardiol. 2017; 70: 1183-1192
        • Mjahad A.
        • Rosado-Munoz A.
        • Bataller-Mompean M.
        • Frances-Villora J.V.
        • Guerrero-Martinez J.F.
        Ventricular fibrillation and tachycardia detection from surface ECG using time-frequency representation images as input dataset for machine learning.
        Comput Methods Programs Biomed. 2017; 141: 119-127
        • Mathews S.M.
        • Kambhamettu C.
        • Barner K.E.
        A novel application of deep learning for single-lead ECG classification.
        Comput Biol Med. 2018; 99: 53-62
        • Guglin M.E.
        • Thatai D.
        Common errors in computer electrocardiogram interpretation.
        Int J Cardiol. 2006; 106: 232-237
        • Hannun A.Y.
        • Rajpurkar P.
        • Haghpanahi M.
        • et al.
        Cardiologist-level arrhythmia detection and classification in ambulatory electrocardiograms using a deep neural network.
        Nat Med. 2019; 25: 65-69
        • Karpathy A.
        • Fei-Fei L.
        Deep visual-semantic alignments for generating image descriptions.
        IEEE Trans Pattern Anal Mach Intell. 2017; 39: 664-676
        • Hochreiter S.
        • Schmidhuber J.
        Long short-term memory.
        Neural Comput. 1997; 9: 1735-1780
        • Kligfield P.
        • Gettes L.S.
        • Bailey J.J.
        • et al.
        Recommendations for the standardization and interpretation of the electrocardiogram: part I: the electrocardiogram and its technology: a scientific statement from the American Heart Association Electrocardiography and Arrhythmias Committee, Council on Clinical Cardiology; the American College of Cardiology Foundation; and the Heart Rhythm Society: endorsed by the International Society for Computerized Electrocardiology.
        Circulation. 2007; 115: 1306-1324
      1. Mostayed A, Luo J, Shu XL, Wee W. Classification of 12-lead ECG signals with bi-directional LSTM network 2018 [e-print]. arXiv:1811.02090.

        • Graves A.
        • Liwicki M.
        • Fernandez S.
        • et al.
        A novel connectionist system for unconstrained handwriting recognition.
        IEEE Trans Pattern Anal Mach Intell. 2009; 31: 855-868
        • Graves A.
        • Schmidhuber J.
        Framewise phoneme classification with bidirectional LSTM and other neural network architectures.
        Neural Netw. 2005; 18: 602-610
        • Rahhal M.M.A.
        • Bazi Y.
        • AlHichri H.
        • Alajlan N.
        • Melgani F.
        • Yager R.R.
        Deep learning approach for active classification of electrocardiogram signals.
        Inf Sci. 2016; 345: 340-354
        • Moody G.B.
        • Mark R.G.
        The impact of the MIT-BIH arrhythmia database.
        IEEE Eng Med Biol Mag. 2001; 20: 45-50
        • Yildirim O.
        A novel wavelet sequence based on deep bidirectional LSTM network model for ECG signal classification.
        Comput Biol Med. 2018; 96: 189-202
        • Afsar F.A.
        • Arif M.
        • Yang J.
        Detection of ST segment deviation episodes in ECG using KLT with an ensemble neural classifier.
        Physiol Meas. 2008; 29: 747-760