Project description
Machine-learning based analysis of the foetal cardiovascular system
Pregnancy complications and stillbirth are associated with abnormalities in the development of the cardiovascular system (CVS). Funded by the Marie Skłodowska-Curie Actions programme, the MEASURE project aims to investigate the CVS in pregnancy, focusing on the coupling between foetal and maternal subsystems, mediated by the uterus and placenta. Synchronised Doppler and electrophysiological acquisitions will produce a dataset for the creation of a CVS mathematical model. This model will be translated into a coupling map with a graphic presentation of the relationships between the CVS subsystems. Finally, the coupling maps of healthy and complicated pregnancies will enable the identification of anomalies in the interactions of the CVS subsystems and the creation of a machine-learning based automatic classifier.
Objective
Complications occur in 5-10% of pregnancies, resulting every year in 2.6 million stillbirths and 303,000 maternal deaths. There is established evidence that complications and stillbirth are associated with abnormal development/adaptation of the cardiovascular system (CVS) during pregnancy. However, the literature lacks a holistic approach and only a few limited aspects have been investigated.
Therefore, MEASURE aims at pursuing a holistic investigation of the CVS and the coupling between its subsystems (fetal and maternal CVS), mediated by the uterus and placenta. Synchronized Doppler and electrophysiological acquisitions will be performed to produce a unique dataset for the identification of a mathematical model describing the CVS. This will be achieved through robust and accurate extraction of time series representing the characteristics of the CVS in healthy pregnancies. Through an innovative aggregating strategy proposed to improve clinical reliability, the obtained model will be translated into a coupling map, i.e. a graphical representation of the relationships between the CVS subsystems and their variables. Being more intuitive and easier to interpret, this representation is particularly suitable for its clinical use and uptake. Finally, the comparison between the coupling maps of healthy and complicated pregnancies will allow identifying anomalies in the interactions between the CVS subsystems and designing/training a machine learning classifier to automatically discern between them.
The outcomes of MEASURE will improve our understanding of the physiological mechanisms underlying pregnancy and original causes of complications, paving the way for the development of novel, powerful, diagnostic tools for timely diagnosis and, possibly, prediction of pregnancy complications. The support of academic, industrial, and clinical partners will provide all the instruments to achieve the envisioned scientific breakthroughs and favour their clinical translation.
Fields of science
Programme(s)
- HORIZON.1.2 - Marie Skłodowska-Curie Actions (MSCA) Main Programme
Funding Scheme
HORIZON-AG-UN - HORIZON Unit GrantCoordinator
5612 AE Eindhoven
Netherlands