A deep neural network learning algorithm outperforms a conventional algorithm for emergency department electrocardiogram interpretation

Smith, Stephen W.; Walsh, Barbara C.; Grauer, Ken; Wang, Kyuhyun; Rapin, Jérémy; Li, Jia; Fennell, William H.; Taboulet, P.

doi:10.1016/j.jelectrocard.2018.11.013

Cited by 84 publications

(66 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The second general observation is that the performances of both convolutional as well as recurrent deep learning models is significantly better than the performance of the baseline algorithm operating on wavelet features in line with literature results [41], [44]. However, this statement has to be taken with caution, as the performance of feature-based classifiers is typically rather sensitive to details of feature selection choice of derived and details of the preprocessing procedure.…”

Section: B Ecg Statement Prediction On Ptb-xlsupporting

confidence: 57%

“…Feature-based approaches, where a classifier is trained on precomputed statistical features such as Fourier or Wavelet coefficients have been the predominant approach in the ECG analysis literature until fairly recently, see [42], [43] for reviews. Similar to the image domain, there is increasing evidence that deep learning algorithms trained in an end-toend fashion are able to outperform feature-based approaches [41], [44]. To the best of our knowledge, there is no public implementation of a state-of-the-art feature-based algorithm for multi-channel ECG classification.…”

Section: B Time Series Classification Algorithmsmentioning

confidence: 99%

See 1 more Smart Citation

Deep Learning for ECG Analysis: Benchmarks and Insights from PTB-XL

Strodthoff

Wagner

Schaeffter

et al. 2021

IEEE J. Biomed. Health Inform.

223

137

View full text Add to dashboard Cite

Electrocardiography (ECG) is a very common, non-invasive diagnostic procedure and its interpretation is increasingly supported by algorithms. The progress in the field of automatic ECG analysis has up to now been hampered by a lack of appropriate datasets for training as well as a lack of well-defined evaluation procedures to ensure comparability of different algorithms. To alleviate these issues, we put forward first benchmarking results for the recently published, freely accessible clinical 12-lead ECG dataset PTB-XL, covering a variety of tasks from different ECG statement prediction tasks to age and sex prediction. Among the investigated deep-learning-based timeseries classification algorithms, we find that convolutional neural networks, in particular resnet-and inception-based architectures, show the strongest performance across all tasks. We find consistent results on the ICBEB2018 challenge ECG dataset and discuss prospects of transfer learning using classifiers pretrained on PTB-XL. These benchmarking results are complemented by deeper insights into the classification algorithm in terms of hidden stratification, model uncertainty and an exploratory interpretability analysis, which provide connecting points for future research on the dataset. Our results emphasize the prospects of deep-learning-based algorithms in the field of ECG analysis, not only in terms of quantitative accuracy but also in terms of clinically equally important further quality metrics such as uncertainty quantification and interpretability. With this resource, we aim to establish the PTB-XL dataset as a resource for structured benchmarking of ECG analysis algorithms and encourage other researchers in the field to join these efforts.

show abstract

Section: B Ecg Statement Prediction On Ptb-xlsupporting

confidence: 57%

Section: B Time Series Classification Algorithmsmentioning

confidence: 99%

Deep Learning for ECG Analysis: Benchmarks and Insights from PTB-XL

Strodthoff

Wagner

Schaeffter

et al. 2021

IEEE J. Biomed. Health Inform.

223

137

View full text Add to dashboard Cite

show abstract

“…An obstacle to the widespread application of the ACOMI/non-ACOMI concept is its dependence on better ECG interpreting skills, which may be hard to achieve in the real clinical world, but this is an unavoidable necessary step for improvement. Improved computer interpretation algorithms, especially use of neural networks [40] , may partly overcome this issue. As a universally agreed definition of ACO is not present, we created our own arbitrary definition.…”

Section: Discussionmentioning

confidence: 99%

DIagnostic accuracy oF electrocardiogram for acute coronary OCClUsion resuLTing in myocardial infarction (DIFOCCULT Study)

Aslanger

Yıldırımtürk

Şimşek

et al. 2020

IJC Heart & Vasculature

Self Cite

View full text Add to dashboard Cite

Background Although ST-segment elevation (STE) has been used synonymously with acute coronary occlusion (ACO), current STE criteria miss nearly one-third of ACO and result in a substantial amount of false catheterization laboratory activations. As many other electrocardiographic (ECG) findings can reliably indicate ACO, we sought whether a new ACO/non-ACO myocardial infarction (MI) paradigm would result in better identification of the patients who need acute reperfusion therapy. Methods A total of 3000 patients were enrolled in STEMI, non-STEMI and control groups. All ECGs were reviewed by two cardiologists, blinded to any outcomes, for the current STEMI criteria and other subtle signs. A combined ACO endpoint was composed of peak troponin level, troponin rise within the first 24 h and angiographic appearance. The dead or alive status was checked from hospital records and from the electronic national database. Results In non-STEMI group, 28.2% of the patients were re-classified by the ECG reviewers as having ACO. This subgroup had a higher frequency of ACO, myocardial damage, and both in-hospital and long-term mortality compared to non-STEMI group. A prospective ACOMI/non-ACOMI approach to the ECG had superior diagnostic accuracy compared to the STE/non-STEMI approach in the prediction of ACO and long-term mortality. In Cox-regression analysis early intervention in patients with non-ACO-predicting ECGs was associated with a higher long-term mortality. Conclusions We believe that it is time for a new paradigm shift from the STEMI/non-STEMI to the ACOMI/non-ACOMI in the acute management of MI. (DIFOCCULT study; ClinicalTrials.gov number, NCT04022668.)

show abstract

“…A recent work attempted to automatically classify 12-lead ECG into 17 groups of conditions 27 . Although this work overcame the challenge of identifying both rhythm and morphology conditions and successfully detected more than one disease per ECG, it provided low accuracy and was only suitable for digital signals.…”

Section: Discussionmentioning

confidence: 99%

Automatic classification of healthy and disease conditions from images or digital standard 12-lead electrocardiograms

Gliner

Keidar

Makarov

et al. 2020

Sci Rep

View full text Add to dashboard Cite

Standard 12-lead electrocardiography (ECG) is used as the primary clinical tool to diagnose changes in heart function. The value of automated 12-lead ECG diagnostic approaches lies in their ability to screen the general population and to provide a second opinion for doctors. Yet, the clinical utility of automated ECG interpretations remains limited. We introduce a two-way approach to an automated cardiac disease identification system using standard digital or image 12-lead ECG recordings. Two different network architectures, one trained using digital signals (CNN-dig) and one trained using images (CNN-ima), were generated. An open-source dataset of 41,830 classified standard ECG recordings from patients and volunteers was generated. CNN-ima was trained to identify atrial fibrillation (AF) using 12-lead ECG digital signals and images that were also transformed to mimic mobile device camera-acquired ECG plot snapshots. CNN-dig accurately (92.9–100%) identified every possible combination of the eight most-common cardiac conditions. Both CNN-dig and CNN-ima accurately (98%) detected AF from standard 12-lead ECG digital signals and images, respectively. Similar classification accuracy was achieved with images containing smartphone camera acquisition artifacts. Automated detection of cardiac conditions in standard digital or image 12-lead ECG signals is feasible and may improve current diagnostic methods.

show abstract

A deep neural network learning algorithm outperforms a conventional algorithm for emergency department electrocardiogram interpretation

Cited by 84 publications

References 27 publications

Deep Learning for ECG Analysis: Benchmarks and Insights from PTB-XL

Deep Learning for ECG Analysis: Benchmarks and Insights from PTB-XL

DIagnostic accuracy oF electrocardiogram for acute coronary OCClUsion resuLTing in myocardial infarction (DIFOCCULT Study)

Automatic classification of healthy and disease conditions from images or digital standard 12-lead electrocardiograms

Contact Info

Product

Resources

About