Convolutional neural network (CNN)-enabled electrocardiogram (ECG) analysis: a comparison between standard twelve-lead and single-lead setups
BackgroundArtificial intelligence (AI) has shown promise in the early detection of various cardiac conditions from a standard 12-lead electrocardiogram (ECG). However, the ability of AI to identify abnormalities from single-lead recordings across a range of pathological conditions remains to be syst...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2024-02-01
|
| Series: | Frontiers in Cardiovascular Medicine |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fcvm.2024.1327179/full |
| _version_ | 1827318195004899328 |
|---|---|
| author | Andrea Saglietto Andrea Saglietto Daniele Baccega Daniele Baccega Roberto Esposito Matteo Anselmino Matteo Anselmino Veronica Dusi Veronica Dusi Attilio Fiandrotti Gaetano Maria De Ferrari Gaetano Maria De Ferrari |
| author_facet | Andrea Saglietto Andrea Saglietto Daniele Baccega Daniele Baccega Roberto Esposito Matteo Anselmino Matteo Anselmino Veronica Dusi Veronica Dusi Attilio Fiandrotti Gaetano Maria De Ferrari Gaetano Maria De Ferrari |
| author_sort | Andrea Saglietto |
| collection | DOAJ |
| description | BackgroundArtificial intelligence (AI) has shown promise in the early detection of various cardiac conditions from a standard 12-lead electrocardiogram (ECG). However, the ability of AI to identify abnormalities from single-lead recordings across a range of pathological conditions remains to be systematically investigated. This study aims to assess the performance of a convolutional neural network (CNN) using a single-lead (D1) rather than a standard 12-lead setup for accurate identification of ECG abnormalities.MethodsWe designed and trained a lightweight CNN to identify 20 different cardiac abnormalities on ECGs, using data from the PTB-XL dataset. With a relatively simple architecture, the network was designed to accommodate different combinations of leads as input (<100,000 learnable parameters). We compared various lead setups such as the standard 12-lead, D1 alone, and D1 paired with an additional lead.ResultsThe CNN based on single-lead ECG (D1) achieved satisfactory performance compared to the standard 12-lead framework (average percentage AUC difference: −8.7%). Notably, for certain diagnostic classes, there was no difference in the diagnostic AUC between the single-lead and the standard 12-lead setups. When a second lead was detected in the CNN in addition to D1, the AUC gap was further reduced to an average percentage difference of −2.8% compared with that of the standard 12-lead setup.ConclusionsA relatively lightweight CNN can predict different classes of cardiac abnormalities from D1 alone and the standard 12-lead ECG. Considering the growing availability of wearable devices capable of recording a D1-like single-lead ECG, we discuss how our findings contribute to the foundation of a large-scale screening of cardiac abnormalities. |
| first_indexed | 2024-03-08T00:50:54Z |
| format | Article |
| id | doaj.art-52303fbdb09b4f3fbbe4b2beb20d4a48 |
| institution | Directory Open Access Journal |
| issn | 2297-055X |
| language | English |
| last_indexed | 2024-04-24T23:53:55Z |
| publishDate | 2024-02-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Cardiovascular Medicine |
| spelling | doaj.art-52303fbdb09b4f3fbbe4b2beb20d4a482024-03-14T16:01:20ZengFrontiers Media S.A.Frontiers in Cardiovascular Medicine2297-055X2024-02-011110.3389/fcvm.2024.13271791327179Convolutional neural network (CNN)-enabled electrocardiogram (ECG) analysis: a comparison between standard twelve-lead and single-lead setupsAndrea Saglietto0Andrea Saglietto1Daniele Baccega2Daniele Baccega3Roberto Esposito4Matteo Anselmino5Matteo Anselmino6Veronica Dusi7Veronica Dusi8Attilio Fiandrotti9Gaetano Maria De Ferrari10Gaetano Maria De Ferrari11Division of Cardiology, Cardiovascular and Thoracic Department, “Citta della Salute e della Scienza” Hospital, Turin, ItalyDepartment of Medical Sciences, University of Turin, Turin, ItalyDepartment of Computer Science, University of Turin, Turin, ItalyLaboratorio InfoLife, Consorzio Interuniversitario Nazionale per l'Informatica (CINI), Rome, ItalyDepartment of Computer Science, University of Turin, Turin, ItalyDivision of Cardiology, Cardiovascular and Thoracic Department, “Citta della Salute e della Scienza” Hospital, Turin, ItalyDepartment of Medical Sciences, University of Turin, Turin, ItalyDivision of Cardiology, Cardiovascular and Thoracic Department, “Citta della Salute e della Scienza” Hospital, Turin, ItalyDepartment of Medical Sciences, University of Turin, Turin, ItalyDepartment of Computer Science, University of Turin, Turin, ItalyDivision of Cardiology, Cardiovascular and Thoracic Department, “Citta della Salute e della Scienza” Hospital, Turin, ItalyDepartment of Medical Sciences, University of Turin, Turin, ItalyBackgroundArtificial intelligence (AI) has shown promise in the early detection of various cardiac conditions from a standard 12-lead electrocardiogram (ECG). However, the ability of AI to identify abnormalities from single-lead recordings across a range of pathological conditions remains to be systematically investigated. This study aims to assess the performance of a convolutional neural network (CNN) using a single-lead (D1) rather than a standard 12-lead setup for accurate identification of ECG abnormalities.MethodsWe designed and trained a lightweight CNN to identify 20 different cardiac abnormalities on ECGs, using data from the PTB-XL dataset. With a relatively simple architecture, the network was designed to accommodate different combinations of leads as input (<100,000 learnable parameters). We compared various lead setups such as the standard 12-lead, D1 alone, and D1 paired with an additional lead.ResultsThe CNN based on single-lead ECG (D1) achieved satisfactory performance compared to the standard 12-lead framework (average percentage AUC difference: −8.7%). Notably, for certain diagnostic classes, there was no difference in the diagnostic AUC between the single-lead and the standard 12-lead setups. When a second lead was detected in the CNN in addition to D1, the AUC gap was further reduced to an average percentage difference of −2.8% compared with that of the standard 12-lead setup.ConclusionsA relatively lightweight CNN can predict different classes of cardiac abnormalities from D1 alone and the standard 12-lead ECG. Considering the growing availability of wearable devices capable of recording a D1-like single-lead ECG, we discuss how our findings contribute to the foundation of a large-scale screening of cardiac abnormalities.https://www.frontiersin.org/articles/10.3389/fcvm.2024.1327179/fullartificial intelligencedeep learningelectrocardiogramsingle-leadscreening |
| spellingShingle | Andrea Saglietto Andrea Saglietto Daniele Baccega Daniele Baccega Roberto Esposito Matteo Anselmino Matteo Anselmino Veronica Dusi Veronica Dusi Attilio Fiandrotti Gaetano Maria De Ferrari Gaetano Maria De Ferrari Convolutional neural network (CNN)-enabled electrocardiogram (ECG) analysis: a comparison between standard twelve-lead and single-lead setups Frontiers in Cardiovascular Medicine artificial intelligence deep learning electrocardiogram single-lead screening |
| title | Convolutional neural network (CNN)-enabled electrocardiogram (ECG) analysis: a comparison between standard twelve-lead and single-lead setups |
| title_full | Convolutional neural network (CNN)-enabled electrocardiogram (ECG) analysis: a comparison between standard twelve-lead and single-lead setups |
| title_fullStr | Convolutional neural network (CNN)-enabled electrocardiogram (ECG) analysis: a comparison between standard twelve-lead and single-lead setups |
| title_full_unstemmed | Convolutional neural network (CNN)-enabled electrocardiogram (ECG) analysis: a comparison between standard twelve-lead and single-lead setups |
| title_short | Convolutional neural network (CNN)-enabled electrocardiogram (ECG) analysis: a comparison between standard twelve-lead and single-lead setups |
| title_sort | convolutional neural network cnn enabled electrocardiogram ecg analysis a comparison between standard twelve lead and single lead setups |
| topic | artificial intelligence deep learning electrocardiogram single-lead screening |
| url | https://www.frontiersin.org/articles/10.3389/fcvm.2024.1327179/full |
| work_keys_str_mv | AT andreasaglietto convolutionalneuralnetworkcnnenabledelectrocardiogramecganalysisacomparisonbetweenstandardtwelveleadandsingleleadsetups AT andreasaglietto convolutionalneuralnetworkcnnenabledelectrocardiogramecganalysisacomparisonbetweenstandardtwelveleadandsingleleadsetups AT danielebaccega convolutionalneuralnetworkcnnenabledelectrocardiogramecganalysisacomparisonbetweenstandardtwelveleadandsingleleadsetups AT danielebaccega convolutionalneuralnetworkcnnenabledelectrocardiogramecganalysisacomparisonbetweenstandardtwelveleadandsingleleadsetups AT robertoesposito convolutionalneuralnetworkcnnenabledelectrocardiogramecganalysisacomparisonbetweenstandardtwelveleadandsingleleadsetups AT matteoanselmino convolutionalneuralnetworkcnnenabledelectrocardiogramecganalysisacomparisonbetweenstandardtwelveleadandsingleleadsetups AT matteoanselmino convolutionalneuralnetworkcnnenabledelectrocardiogramecganalysisacomparisonbetweenstandardtwelveleadandsingleleadsetups AT veronicadusi convolutionalneuralnetworkcnnenabledelectrocardiogramecganalysisacomparisonbetweenstandardtwelveleadandsingleleadsetups AT veronicadusi convolutionalneuralnetworkcnnenabledelectrocardiogramecganalysisacomparisonbetweenstandardtwelveleadandsingleleadsetups AT attiliofiandrotti convolutionalneuralnetworkcnnenabledelectrocardiogramecganalysisacomparisonbetweenstandardtwelveleadandsingleleadsetups AT gaetanomariadeferrari convolutionalneuralnetworkcnnenabledelectrocardiogramecganalysisacomparisonbetweenstandardtwelveleadandsingleleadsetups AT gaetanomariadeferrari convolutionalneuralnetworkcnnenabledelectrocardiogramecganalysisacomparisonbetweenstandardtwelveleadandsingleleadsetups |