Arrhythmias & electrophysiology
Identifying the Presence and Characteristics of Mid-Myocardial and Epicardial Fibrosis From Intracardiac Electrograms in Patients Undergoing Ventricular Arrhythmia Ablation Using a Transformer-Based Self-Supervised Classifier
Authors' conclusion
Our study demonstrates that routine intracardiac electrograms enable the identification of endocardial, mid-myocardial, and epicardial scar using a transformer-based self-supervised deep learning model. Ultimately, our model has the potential to provide magnetic resonance imaging-like fibrosis maps during EP procedures without the need for external imaging and to increase the data set size for patient-specific models in ventricular arrhythmia research.
Study context
- Population
- Patients undergoing ventricular arrhythmia ablationpatients, 131,584 cardiac contact electrogram (EGM) signals · training/validation set 37 · testing set 9
- Participants
- 46 patients
- Exposure
- EGM2Scar-AI vs basic convolutional neural network models, with scar labels from coregistered cardiac magnetic resonance imaging and electroanatomic mapping surface meshes
- Outcome
- Average task-specific area under the receiver operating characteristic curve across all 3 scar types
Methodological detailsstratified patient-wise split
- Stratified patient-wise split
Main finding
As reported in the abstract
Basic convolutional neural network
0.729
0.725–0.732
average area under the receiver operating characteristic curve
Short-time Fourier transform-based convolutional neural network
0.729
0.726–0.733
average area under the receiver operating characteristic curve
EGM2Scar-AI
0.822
0.819–0.825
average area under the receiver operating characteristic curve
Observational data: associations, not effects.
Further findings
Performance by scar type
- All models performed better on endocardial and mid-myocardial fibrosis identification, with a modest reduction in performance for epicardial fibrosis
Sensitivity and specificity
- Sensitivity improved significantly with the transformer-based architecture without appreciable changes in specificity
Original publication
Summarised from the published abstract; every number is checked against it