Exploring artificial intelligence methods for cardiac syncope diagnosis combined with electrocardiogram parameters and clinical characteristics

Background:  Cardiac syncope can be life-threatening, but there is no clinical tool for initial screening. The study explored and developed optimal artificial intelligence methods for automatic diagnosis of cardiac syncope based on combinations of electrocardiogram parameters and clinical characteristics.

Methods:  The patients presenting with syncope and hospitalized between June 21, 2018 and August 23, 2022 at the Second Hospital of Tianjin Medical University. The patients enrolled were divided into development cohort who were then randomly split into a training set and an internal validation set (4: 1) and temporal validation cohort. Fifteen features of syncope patients were ranked and valuable features were selected. Six supervised machine learning models were developed to explore a potential prediction model for cardiac syncope. The area under the curve (AUC) was the primary metric used to evaluate classification performance.

Results:  A total of 380 patients (340 in the development cohort and 40 in the temporal validation cohort) were included in the final analysis. The random forest showed the best performance using the top twelve features ranked by importance, demonstrating an AUC of 0.85 (sensitivity: 0.72, specificity: 0.85, F1 score: 0.74) in the development cohort, and an AUC of 0.75 (sensitivity: 0.70, specificity: 0.65, F1 score: 0.68) in the validation cohort. The novel approach for automatic diagnosis of cardiac syncope has been proposed as web service for further application.

Conclusions:  Artificial intelligence methods may assist in syncope classification, and which have the potential to serve as a cost-effective and efficient screening tool for cardiac syncope.

© 2025 Elsevier Inc. All rights are reserved, including those for text and data mining, AI training, and similar technologies.