Skip to content

Deep Learning Models for Fetal Monitoring and Decision Support in labour

Funding: R: ResearchAward

"Continuous electronic fetal monitoring (EFM) is used worldwide to visually assess whether a fetus is exhibiting signs of distress during labor and if there is a need of an emergency operative intervention. The gold standard to assess whether a baby is at risk of oxygen starvation during birth, is to monitor continuously the fetal heart rate and the uterine contractions with cardiotocography (CTG). However, visual interpretation of the complex EFM graphs, remains unreliable and poorly understood (in the UK alone, every year, about 421 healthy babies die [1], and about 1000 babies sustain brain injury during labor at term [2]). <br/>There are a few classic EFM patterns that have been empirically identified, but for certain type of them, the disagreement in visual interpretation between experts reaches 100% [3]. Computerized detection of such classic patterns is commercially available, but has not shown benefit in randomized clinical trials [4].<br/>In this project, deep learning approaches will be investigated and developed to achieve data-driven automated CTG evaluation. Multimodal Convolutional Neural Network (MCNN) and Stacked MCNN models in addition to Long Short Term Memory (LSTM) neural networks will be used to extract knowledge from the largest available database of routinely collected CTG and linked clinical data (the Oxford dataset, which span many years of varying clinical practice comprising records of thousands of labors) for prediction of fetal compromise. These will then be further used for developing a hybrid model for CTG interpretation and inferring in labor, incorporating other diagnostic models (such as OxSys 1.5 and Clinical practice).<br/>Finally, the work will culminate in developing software application on a hand held portable device (tablet/mobile) as an automated real time tool that will be used to help practitioners take informed decision when predicting clinically relevant outcome of acidemia and/or severe fetal compromise." <br/>

Engineering and Physical Sciences Research Council: £347,101.00

StatusNot started
Effective start/end date1/04/2130/03/24
Award date27/11/20
Relations

ID: 24835916