[Advanced Data-Aided Medicine part coronary artery disease]

Developing and validating an AI-driven clinical decision support tool for the optimization of treatment plans for patients with coronary artery disease. On single-lead electrocardiogram, this research team is building an Artificial Intelligence model with risk factors detecting atrial fibrillation during sinus rhythm.

---

Objective

Research designed to improve the detection of atrial fibrillation (AF), a type of irregular heartbeat that is often temporary and intermittent (paroxysmal AF) via ECG screening presenting a normal heath rhythm called sinus rhythm. On single-lead electrocardiogram, this research team is building an Artificial Intelligence model with risk factors detecting atrial fibrillation during sinus rhythm.

Methodology

Guidelines recommend opportunistic screening for atrial fibrillation (AF), using a 30 s single-lead electrocardiogram (ECG) recorded by a wearable device. Since many patients have paroxysmal AF, identification of patients at high risk presenting with sinus rhythm (SR) may increase the yield of subsequent long-term cardiac monitoring. The aim is to evaluate an AI-algorithm trained on 10 s single-lead ECG with or without risk factors to predict AF.

This retrospective study used 13 479 ECGs from AF patients in SR around the time of diagnosis and 53 916 age- and sex-matched control ECGs, augmented with 17 risk factors extracted from electronic health records. AI models were trained and compared using 1- or 12-lead ECGs, with or without risk factors. Model bias was evaluated by age- and sex-stratification of results. Random forest models identified the most relevant risk factors. The single-lead model achieved an area under the curve of 0.74, which increased to 0.76 by adding six risk factors (95% confidence interval: 0.74-0.79). This model matched the performance of a 12-lead model. Results are stable for both sexes, over ages ranging from 40 to 90 years. Out of 17 clinical variables, 6 were sufficient for optimal accuracy of the model: hypertension, heart failure, valvular disease, history of myocardial infarction, age, and sex.

Impact and future directions

A clinical decision support might improve standardization of treatment plans across similar patients and guide decisions for clinically more complex cases. More specifically an AI model using a single-lead SR ECG and six risk factors can identify patients with concurrent AF with similar accuracy as a 12-lead ECG-AI model. An age- and sex-matched data set leads to an unbiased model with consistent predictions across age groups.

General info and contact

Keywords (#): Artificial intelligence, Atrial fibrillation, Screening, Single-lead ECG, Sinus rhythm

RADar project research lead: Ir. Louise Vander Heyde

RADar project researchers: Ir. Louise Vander Heyde, Prof. Dr. Ir. Peter De Jaeger

Principal investigator: MD Karl Dujardin

Timeline: January 2021 – December 2023

Status: First research finished, ready for validation

Publications / presentations:
Single-lead electrocardiogram Artificial Intelligence model with risk factors detects atrial fibrillation during sinus rhythm – PubMed (nih.gov)

Partners: UGent IDLab, Resero

Funding: Funded with resources from the European Recovery Fund via Vlaio O&O

• GitHub