Navigation

Master's Thesis

Deep Learning-Based ECG Classification With Focus on Pregnant Women

Advisors
Stefan Gradl (M.Sc.)Prof. Dr. Björn Eskofier 

Duration
06/2020 – 12/2020

Abstract

Electrocardiogram (ECG) analysis has been established at the core of cardiovascular pathology
diagnosis since the early 20th century, due to its non-invasive, painless, and quickly repeatable
nature [1]. The ECG signals reflect the electrical activity of the heart. Furthermore, different
physiological and pathological processes can be derived like heart rhythm, different types of arrhythmia,
gender and many more [2, 3].
There are well known sex differences in the electrophysiological parameters and cardiac arrhythmias
of the ECG. Women have a higher resting heart rate, longer QT interval, shorter PR and
QRS intervals. Many of these differences are related to varying hormonal parameters and anatomy.
This in particular concerns the periods of pregnancy and the menopause. These adaptations may
result in specific arrhythmia conditions that cannot be found in men and are a cause of concern
for the well-being of the woman after the menopause or for the mother and fetus during and after
pregnancy [4, 5].
Monitoring the health during pregnancy may allow early detection of abnormalities induced in
physiological parameters, which would remain unnoticed otherwise and might lead to critical
complications [6]. While existing computer interpreted ECG signals have a limited diagnostic
accuracy and are in need of specialist review, fully automated ECG analysis pipelines using Deep
Learning have the potential to improve the accuracy of these systems [7]. This strategy should be
examined in this project.

 

References:

[1] M. Gertsch, G. Steinbeck, and B. Fässler, „Das EKG: auf einen Blick und im Detail mit 54 Tabellen“,
2. Aufl. Heidelberg: Springer Medizin, 2008.
[2] H. Bidoggia et al., “Sex-dependent electrocardio- graphic pattern of cardiac repolarization”, American
Heart Journal, vol. 140, no. 3, pp. 430–436, Sep. 2000.
[3] Zachi I. Attia et al., “Age and Sex Estimation Using Artificial Intelligence From Standard 12-Lead
ECGs”, AHAJournals: 2019.
[4] B. M. Pampana Gowd, MD and Paul D. Thompson, MD, “Effect of Female Sex on Cardiac Arrhythmias”,
Cardiology in Review vol. 20 no. 6, Dec. 2012
[5] Muti Goloba, Scott Nelson, Peter Macfarlane, „The Electrocardiogram in Pregnancy“, Computing
in Cardiology, vol. 37, pp. 693?696, 2010.
[6] Lakshmi.B.N et al., “An Hybrid Approach for Prediction Based Health Monitoring in Pregnant
Women”, Elsevier 2016.
[7] K.S. Rajput et al., “On Arrhythmia Detection by Deep Learning and Multidimensional Representation”
2019