Title:Cardiac and extracardiac discharge diagnosis prediction from emergency department ECGs using deep learning

Abstract:Current deep learning algorithms designed for automatic ECG analysis have exhibited notable accuracy. However, akin to traditional electrocardiography, they tend to be narrowly focused and typically address a singular diagnostic condition. In this study, we specifically demonstrate the capability of a single model to predict a diverse range of both cardiac and non-cardiac discharge diagnoses based on a sole ECG collected in the emergency department. Among the 1,076 hierarchically structured ICD codes considered, our model achieves an AUROC exceeding 0.8 in 439 of them. This underscores the models proficiency in handling a wide array of diagnostic scenarios. We emphasize the potential of utilizing this model as a screening tool, potentially integrated into a holistic clinical decision support system for efficiently triaging patients in the emergency department. This research underscores the remarkable capabilities of comprehensive ECG analysis algorithms and the extensive range of possibilities facilitated by the open MIMIC-IV-ECG dataset. Finally, our data may play a pivotal role in revolutionizing the way ECG analysis is performed, marking a significant advancement in the field.