Automated anomaly screening in the first trimester of pregnancy using Artificial Intelligence

Project summary

There is critical need for automated analysis of ultrasonography to get a more detailed picture of the developing fetus. Together with GE Healthcare, our aim is to develop innovative Artificial Intelligence (AI) algorithms to support the screening for congenital anomalies in the first trimester of pregnancy.

Each year, approximately 130 million children are born worldwide, of which 6% have congenital anomalies. About half of these congenital anomalies are only discovered after birth. Having a congenital anomaly has direct consequences for the short-term and long-term health of the newborn. Ultrasound imaging techniques are used as the ‘gold standard’ for detecting congenital anomalies during pregnancy. However, significant limitations include: 1) low detection rates, 2) time-consuming and dependent on the healthcare provider, 3) low success rates of measurements, and 4) the lack of a comprehensive view of the developed fetus. It is time to address these limitations with an innovative method.

In this project, we aim to develop fully automated AI algorithms to support the detection of congenital anomalies in the first trimester of pregnancy. We will investigate the clinical relevance of the developed algorithms in already existing imaging datasets including >4000 pregnancies. Finally, we will implement and test the final prototype in daily clinical practice. Together with patients and healthcare providers, we will evaluate this process.

The final deliverable of the project is a tested prototype of the AI algorithms embedded on the ultrasound machine. The implementation of this prototype in clinical practice will result in better detection of congenital anomalies, with direct consequences for the lifelong health of the unborn child and the parents. Additionally, this will reduce the workload by shortening scan times and requiring fewer experienced healthcare providers. Ultimately, this will result in a decrease in overall healthcare costs.