Predicting Pre-Eclampsia

Pre-eclampsia, a condition affecting pregnant women, remains poorly understood despite affecting up to one in ten pregnancies worldwide. The condition kills 70,000 mothers and half a million babies globally each year.

If diagnosed too late, the only known treatment is to deliver the baby as soon as possible. But this too has issues, as being born early can lead to lifelong health challenges.

Predicting pe-eclampsia before symptoms, so clinicians can strive to stop it from developing, is therefore, the best way to reduce the harm it causes.

Dr Pradeepa Madapura, Reader in Genome Biology, and Dr Manthan Patel, postdoctoral researcher at Queen Mary University of London, invented a new blood test for predicting pre-eclampsia just five to sixteen weeks into a pregnancy.

The test could fit directly into the current standard of care practice, making it easy to implement.

It works by analysing mobile DNA sequences in plasma (known as transposable elements) to check for abnormal activity associated with inflammation in the mother’s placenta. The researchers discovered this using machine learning to identify a transposable element signature (a specific pattern of DNA sequences) in maternal plasma. The presence of this signature provides a robust means of predicting pre-eclampsia before symptoms emerge.

This is a significant improvement on existing diagnostic methods, which occur much later in the pregnancy and usually after symptoms have already appeared.

Knowing that pre-eclampsia is likely means you can mitigate against it. This innovation could, therefore, have a lifesaving impact on mothers and babies, especially in developing countries.

The invention has been proven in principle using an AI model – a notable demonstration of the value of AI in research. Queen Mary is now backing the research team via the Queen Mary Impact Fund to validate the test in a group of patients.

We are looking to partner with a diagnostics company to develop the test for clinical use.