A new method for detecting heart rhythm abnormalities in unborn babies

 

3 Minute read

Although many cases of fetal arrhythmia pose no harm, up to 10 per cent can lead to severe injury or death of the unborn child. Proper treatment depends on fast, reliable detection and diagnosis.


Read the abstract

A method that automatically identifies fetal arrhythmias from non-invasive electrocardiography (ECG) recordings has been developed by University of Melbourne researchers. The research was led by Dr Emerson Keenan, a postdoctoral researcher working with Dr Fiona Brownfoot and Professor Marimuthu Palaniswami, and involved an international group of collaborators.

Fetal arrhythmias affect up to 2 per cent of pregnancies and include cases where the heart rhythm is too fast, too slow or irregular. Detection of fetal arrhythmias can be difficult and time-consuming. The most common method requires a sonographer to assess fetal heart activity using ultrasound and a cardiologist to interpret the results.

To help automate the process of arrhythmia detection, previous algorithms have been designed to analyse the unpredictability, or entropy, in ECG recordings. Entropy algorithms measure similarities in the heart rate pattern over time. However, these methods require researchers to set a threshold for which patterns are classified as similar in each application. This makes it difficult to directly apply existing entropy algorithms to novel uses such as detecting fetal arrhythmias.

The new method developed by the team, called TotalSampEn, overcomes this limitation by measuring the similarity over a range of data-driven thresholds for each ECG recording. In this way, TotalSampEn creates an entropy profile for each recording, which has more useful information about heart rate irregularities than a single entropy value.

Using 318 ECG recordings from fetuses with and without an arrhythmia, the researchers showed that TotalSampEn identified over 80 percent of fetal arrhythmias with a low rate of false positives, outperforming previous entropy algorithms. When low-quality ECG recordings were excluded, the performance of TotalSampEn was even higher.

These findings show that TotalSampEn can be used as a first-line tool to detect fetal arrhythmias from ECG recordings. In cases where an arrhythmia is detected, a cardiologist can examine the ECG recording to diagnose the type of arrhythmia and choose the best pathway for treatment.

Next steps

The researchers now plan to develop a wearable device that can reliably capture high-quality ECG recordings of fetal heart activity. Using this device, they hope to conduct clinical trials to validate the performance of their automated arrhythmia detection method in a clinical setting.

Funding

Department of Obstetrics and Gynaecology Early Career Fellowship to Dr Emerson Keenan

National Health and Medical Research Council Early Career Fellowship (1142636) to Dr Fiona Brownfoot

Norman Beischer Medical Research Foundation Early-Mid Career Clinical Research Fellowship to Dr Fiona Brownfoot

Australian Research Council Discovery Project (DP190101248) to Professor Marimuthu Palaniswami

Publication

Keenan E et al (2022) Detection of fetal arrhythmias in non-invasive fetal ECG recordingsusing data-driven entropy profiling. Physiological Measurement 43: 025008. doi: 10.1088/1361-6579/ac4e6d

Re-use this text

Please use the text of this article for your own purposes. The text is licensed under the Creative Commons Attribution (CC BY) 4.0 International license. This lets you copy, transform and share the text without restriction. We appreciate appropriate credit and links back to this website. Other content on this page (such as images, videos and logos) is not covered by the CC BY license and may not be used without permission from the copyright holder. If you have any questions about using this text, please contact the research web team.

First published on 12 January 2023.


Share this article