HRX 2024 AbstracX Library-Artificial intelligence enabled electrocardiogram algorithm accurately predicts atrial fibrosis

Artificial intelligence enabled electrocardiogram algorithm accurately predicts atrial fibrosis

Pdf Summary

The document describes a study led by Henri Gruwez, MD, PhD, from Ziekenhuis Oost-Limburg, Belgium, focusing on an artificial intelligence (AI)-enabled electrocardiogram (ECG) algorithm. This innovative algorithm is designed to accurately predict atrial fibrosis, a common cardiac condition, which often accompanies atrial fibrillation but remains challenging to diagnose non-invasively. <br /><br />The algorithm's development and validation involved a clinical study published in the "JACC Clinical Electrophysiology" journal in 2023, co-authored by Gruwez and Barthels, among others. The study involved 60 patients, with an average age of 67 years and 32% female participation. They used an ECG-based deep learning method to assess and predict the presence of atrial fibrosis. <br /><br />Key findings indicated that the algorithm delivers a high level of accuracy, with a sensitivity of 86.7% and specificity of 80.0%. The algorithm's effectiveness is further demonstrated by an area under the curve (AUC) of 0.87, showing a significant ability to distinguish between patients with and without atrial fibrosis. The results are statistically significant with a p-value of less than 0.001.<br /><br />The study thus highlights that ECG-based deep learning can act as a novel digital biomarker, potentially transforming the way atrial fibrosis is diagnosed, offering a non-invasive, reliable diagnostic alternative. This could lead to more timely and accurate identification of atrial fibrosis in clinical settings, contributing to improved patient management and outcomes.

Keywords

artificial intelligence

electrocardiogram

atrial fibrosis

atrial fibrillation

deep learning

Henri Gruwez

JACC Clinical Electrophysiology

non-invasive diagnosis

digital biomarker

patient management