Non-invasive fetal electrocardiogram and Doppler in early pregnancy

Made with heart. Why and how our dream became reality

Like Comment
Read more

Sardinia, the Italian region where the dataset was collected and the largest part of this research carried out, records an incidence of congenital heart disease about double that of the rest of Italy (about 15‰of live births, 18‰ if we also consider intrauterine deaths). Every year about 800 fetuses are cardiologically evaluated at the Pediatric Cardiology and Congenital Heart Disease Unit of the Brotzu Hospital in Cagliari, and many have arrhythmias with different levels of severity. Early discovery of fetal heart problems could allow in-utero treatment (trans-placental drug delivery) and birth planning, eventually leading to obtaining better prenatal prevention.

When we started the research in the field of non-invasive fetal electrocardiography (fECG), about 15 years ago, there were no certified medical devices able to provide a clinically acceptable, non averaged, fECG. To date, some devices are starting to appear but most of them still present severe limitations for a cardiological assessment of the fetus, so the research focused on new signal processing methods is still a hot topic. Several anomalies observed on the fetuses could be easily detected from an ECG but, for the time being, they can be only deduced with ultrasound techniques that, in turn, require very expensive instrumentation and expert cardiologists. This limits the screening to risky pregnancy.

Dr. R. Tumbarello performing a fetal echocardiographic exam for the development of the NInFEA multimodal dataset at the Pediatric Cardiology and Congenital Heart Disease Unit of the Brotzu Hospital in Cagliari, Italy.

The idea of developing a dataset for non-invasive fetal electrocardiography (fECG) dated about ten years ago, in the context of our research in the field. The lack of publicly available datasets that can be used for the development of signal processing algorithms aimed at the extraction of the small and elusive fECG signal from the instrumental and maternal physiological interferences, with high flexibility in the number and position of the electrodes on the maternal body, is a dramatic problem. It imposes to research groups the independent development of a proprietary dataset of real signals, with consequent considerable, often insurmountable, costs and difficulties. During the years, we struggled to find an adequate measurement setup, with huge costs for the acquisition of biopotential recording devices, which often proved unable to provide good quality signals, and clinical partners with enough experience on antenatal cardiological assessment to collect the signals with additional information on the fetal heart activity in early pregnancy, when no invasive methods can be adopted. Without that, no ground truth is available to understand the effectiveness of the signal processing methods.

We did it, and here is the result of our work, with the data available for free to everyone, and we will enlarge the dataset in the next few years with more signals. Several people worked hard to achieve this goal, and we are sure their dream is now a reality. We are sure this will foster the development of non-invasive fECG processing algorithms, with exciting new studies carried out without the burden of data acquisition. Eventually, we hope this will yield progress in fetal cardiac physiology knowledge and the development of prenatal diagnostics. 

Danilo Pani

Associate Professor, University of Cagliari

I was born in Cagliari in 1978. Since November 2019 I'm Associate Professor of biomedical engineering at the Dept. of Electrical and Electronic Engineering, University of Cagliari, Italy. My research is focused on medical devices and advanced biomedical signal processing. In particular: - Cardiac electrophysiology and non-invasive fetal ECG - Neural signal processing - Wearable electronics and textile electrodes for electrophysiological signals monitoring - Tele-medicine, tele-health and tele-care systems