Fetal Heart

Currently more than 50% of severe heart disease in unborn babies is missed during screening. This project aims to investigate whether Artificial Intelligence (AI) can be used to help identify heart conditions in babies before they are born. 

When a baby has a heart problem before birth, this is known as congenital heart disease (CHD). CHD is the most common health issue affecting unborn babies and is present in around 1 in 100 pregnancies. All pregnant women in the UK are offered an ultrasound scan at roughly half-way through their pregnancy, to try and identify certain health issues including CHD. These scans are performed because we know that babies are more likely to survive after birth if we identify CHD before birth, compared to after birth. They are also less likely to have damage to other organs, such as the brain. This is because they can be treated properly as soon as they are born, rather than after a delay. 

Unfortunately, in the UK only around half of babies with CHD are diagnosed before birth. This is because the ultrasound scans are difficult to perform and interpret, and there is a shortage of people able to do these scans. Evidence suggests that due to this lack of staff, and specialist staff, diagnosis is better in some parts of the UK than others.  

Our research group has already developed AI tools to help recognise different parts of the baby, and to automatically measure the baby’s size. the project team aims to develop these to also show which babies have heart conditions. To do this, the team will develop AI tools by using past patient scans to teach the computer program how to recognise differences between hearts with CHD and healthy hearts. They will test how well this work by using past medical pictures where we already know the answer. If this is successful, the AI would then be tested by scanning pregnant women who volunteer to be part of the study.  

By developing new AI enabled technologies, ultrasound scans will identify babies with CHD which will lead to a higher chance of diagnosis. The success of this application will also boost the effectiveness of antenatal and postnatal healthcare services across the UK.