Diagnostic Performance of Automated versus Human ECG Interpretation for Prior Myocardial Infarction Using CMR as the Reference Standard

Purpose To compare automated 12-lead ECG analysis using the ECG Treatment Module (ETM) with human ECG interpretation for detecting prior myocardial infarction (pMI), using cardiovascular magnetic resonance (CMR) as the reference standard. Methods Consecutive patients referred for CMR evaluation of coronary artery disease or cardiomyopathy were included. pMI was centrally adjudicated using Late Gadolinium Enhancement (LGE). LGE was assessed using the 17-segment model for ischaemic pattern, degree of transmurality, and presence of non-ischaemic LGE. ECGs were automatically analysed using the ETM (ETM-2.8.0, Schiller AG, Switzerland). At least two independent cardiologists, blinded to all clinical and imaging data, interpreted the ECGs for pMI according to the fourth universal definition of myocardial infarction. Results Among 1038 patients, automated ECG analysis achieved an overall diagnostic accuracy of 80.3% (77.7% - 82.6%) for pMI detection with a specificity of 89.4% (87.1% - 91.3%) and a sensitivity of 43.8% (37.2% - 50.5%). Compared with human readers, ETM demonstrated similar overall diagnostic accuracy. Without ECG confounders such as bundle branch blocks, the ETM algorithm outperformed humans; with ECG confounders, humans outperformed the algorithm. Transmural infarctions were detected more frequently than subendocardial infarctions (53.9% vs 34.0%, p=0.0058), and detection was higher in infarctions with >50% transmurality (p=0.0464). Of the 179 patients with automatically presumed pMI, 50.8% had an ischaemic LGE pattern and 27.9% had non-ischaemic LGE. Overall, 56.2% of CMR-confirmed pMIs were missed, most commonly in the lateral segments (61.1%). Conclusions Automated ECG analysis using ETM demonstrated diagnostic performance comparable to human readers for detection of pMI, but limited sensitivity precludes reliable exclusion of myocardial scar.

Authors

Institutions

Publication Details

Journal
European Heart Journal - Digital Health
Published
2026-11-18
DOI
https://doi.org/10.1093/ehjdh/ztag180
Primary Topic
ECG Monitoring and Analysis
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Diagnostic Performance of Automated versus Human ECG Interpretation for Prior Myocardial Infarction Using CMR as the Reference Standard

Jędrzej Sarnecki, Ramun Schmid, Ivo Strebel, Maurice Pradella et al.
European Heart Journal - Digital Health
ECG Monitoring and Analysis
article

Diagnostic Performance of Automated versus Human ECG Interpretation for Prior Myocardial Infarction Using CMR as the Reference Standard

Jędrzej Sarnecki, Ramun Schmid, Ivo Strebel, Maurice Pradella, Felix Mahfoud, Jan Gerrit van der Stouwe, Philip Haaf, Simon Frey, Jens Bremerich, Patrick Badertscher, Michael Johannes Zellweger, Martin Segeroth, Stefan Gherca, Christian Eugen Mueller, Yves Schaulin
article en

Abstract

Purpose To compare automated 12-lead ECG analysis using the ECG Treatment Module (ETM) with human ECG interpretation for detecting prior myocardial infarction (pMI), using cardiovascular magnetic resonance (CMR) as the reference standard. Methods Consecutive patients referred for CMR evaluation of coronary artery disease or cardiomyopathy were included. pMI was centrally adjudicated using Late Gadolinium Enhancement (LGE). LGE was assessed using the 17-segment model for ischaemic pattern, degree of transmurality, and presence of non-ischaemic LGE. ECGs were automatically analysed using the ETM (ETM-2.8.0, Schiller AG, Switzerland). At least two independent cardiologists, blinded to all clinical and imaging data, interpreted the ECGs for pMI according to the fourth universal definition of myocardial infarction. Results Among 1038 patients, automated ECG analysis achieved an overall diagnostic accuracy of 80.3% (77.7% - 82.6%) for pMI detection with a specificity of 89.4% (87.1% - 91.3%) and a sensitivity of 43.8% (37.2% - 50.5%). Compared with human readers, ETM demonstrated similar overall diagnostic accuracy. Without ECG confounders such as bundle branch blocks, the ETM algorithm outperformed humans; with ECG confounders, humans outperformed the algorithm. Transmural infarctions were detected more frequently than subendocardial infarctions (53.9% vs 34.0%, p=0.0058), and detection was higher in infarctions with >50% transmurality (p=0.0464). Of the 179 patients with automatically presumed pMI, 50.8% had an ischaemic LGE pattern and 27.9% had non-ischaemic LGE. Overall, 56.2% of CMR-confirmed pMIs were missed, most commonly in the lateral segments (61.1%). Conclusions Automated ECG analysis using ETM demonstrated diagnostic performance comparable to human readers for detection of pMI, but limited sensitivity precludes reliable exclusion of myocardial scar.

European Heart Journal - Digital Health
University of Basel (CH), University Hospital of Basel (CH), Kantonsspital Baden (CH), University Children’s Hospital Basel (CH)
Good health and well-being
ECG Monitoring and Analysis
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon