A paper by researchers from our university has been published in npj Digital Medicine.

The research "Diagnosis of Cardiac Conditions from 12-Lead Electrocardiogram through Natural Language Supervision" by Dr. ZHOU Xue in Prof. Wenxi Chen's Lab has been published in npj Digital Medicine (Impact Factor: 15.1), which is part of the Nature Portfolio.
This study was conducted in collaboration with colleagues from Institute of Science Tokyo, Toho University Ohashi Medical Center, and Chaoyang University of Technology.

Paper Title
"Diagnosis of cardiac conditions from 12-lead electrocardiogram through natural language supervision"

Research Highlights
This study introduces ECG-CLIP, an innovative AI model that enables zero-shot diagnosis of multiple cardiac conditions from 12-lead ECGs using natural language supervision. Trained on 800,034 ECG-text pairs, the model successfully evaluated 18 cardiac conditions without requiring condition-specific labeled training data. The approach achieved excellent performance for rhythm abnormalities (AUROC > 0.90) and demonstrated robust generalization across different patient populations, including remarkable zero-shot performance on pediatric patients despite no pediatric cases in the training data.

This breakthrough addresses fundamental scalability limitations in current medical AI by eliminating dependence on condition-specific labeled datasets, potentially expanding global access to expert-level ECG interpretation.
The patent application is currently in progress.

About npj Digital Medicine
npj Digital Medicine is an open access, international, peer-reviewed journal within the Nature Portfolio that is dedicated to publishing the highest quality research relevant to all aspects of digital medicine and health.

Link to the paper : https://www.nature.com/articles/s41746-025-02074-3

Link to the journal : https://www.nature.com/npjdigitalmed/journal-impact

Link to the laboratory : https://bitlab.u-aizu.ac.jp/