Key result
Explainable deep learning achieves ~89% accuracy discriminating abnormal ventricular potentials in post-ischemic VT.
Why the study?
Detection of abnormal ventricular potentials marking arrhythmogenic sites during post-ischemic VT electroanatomic mapping is strongly operator-dependent and time-consuming.
Does explainable deep learning improve the identification of abnormal ventricular potentials in patients with post-ischemic ventricular tachycardia?
Observational (n=9)
No
Does explainable deep learning improve the identification of abnormal ventricular potentials in patients with post-ischemic ventricular tachycardia?
Deep learning models using time-frequency representations can accurately identify arrhythmogenic sites in post-ischemic ventricular tachycardia, potentially reducing operator dependence during electroanatomic mapping.
May reduce operator dependence in post-ischemic VT mapping; leaves open prospective validation before clinical adoption.
• Operator-independent tool to support cardiac electrophysiology procedures. • Deep learning succeeds in the identification of VT arrhythmogenic sites (89% accuracy). • Customized saliency maps highlight signatures of abnormal ventricular potentials. Abnormal ventricular potentials (AVPs) in intracardiac electrograms (EGMs) are frequently considered as markers of arrhythmogenic sites in post-ischemic ventricular tachycardia (VT) during electroanatomic mapping (EAM) procedures. Their detection is strongly operator-dependent and time-consuming. This work explores the adoption of explainable deep learning to support the discrimination between physiological EGMs and AVPs. Three convolutional neural networks were trained to discriminate the target signals based on their time–frequency representations by synchrosqueezed wavelet transform. The efficacy of the method was assessed on 2561 real bipolar EGMs collected from nine post-ischemic VT patients. The proposed approach achieved high performance, with accuracy levels reaching up to 89%. It also demonstrated coherent localization of the arrhythmogenic sites with respect to conventional voltage and local activation time maps. Moreover, by using saliency maps, AVPs discriminant signatures were highlighted at high frequencies (i.e., in the 103–125 Hz band, which was generally relevant for every network), in line with prior evidence. For the first time, deep learning has been successfully applied and robustly evaluated in the field. The proposed approach paves the way to the development of effective AI-driven systems. These systems will enable a faster, trustworthy and operator-independent identification of AVPs in VT EAM procedures. Furthermore, even without injecting prior knowledge in the adopted models, the analysis of saliency maps revealed that CNNs are prone to autonomously select time–frequency ranges of the EGMs in agreement with the current knowledge.
No takes yet. Share an insight, caveat, or question.
Pitzus et al. (2024) conducted an observational in Post-ischemic ventricular tachycardia (n=9). Explainable deep learning (Convolutional Neural Networks) was evaluated on Accuracy in discriminating abnormal ventricular potentials from physiological electrograms. Explainable deep learning using convolutional neural networks achieved up to 89% accuracy in discriminating abnormal ventricular potentials from physiological electrograms in post-ischemic ventricular tachycardia.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: