The Pregnancy Ultrasound ResourcE study (PURe study)

Research summary

Ultrasound is a mandatory tool for safely monitoring the growth and development of fetuses during pregnancy. Although pregnancy ultrasound is routinely used, it has a major disadvantage, namely a high level of operator dependence: non-experts (or infrequent) users of the technology are often not able to acquire and interpret diagnostic ultrasound images consistently and reliably. Consequently, there is a significant requirement for training, and lack of adequately trained staff is the most commonly reported barrier to the use of ultrasound in developing countries. Currently, there are several factors that limit automated ultrasound systems, such as the use of non-standard imaging planes and artefacts.Therefore, automated obstetric ultrasound systems are used in research settings only. One way to help with automated image analysis systems may be to understand how sonographers examine the image in real time using hand- and eye-tracking. We are working to create automatic ultrasound systems that can work by themselves and only require little control by a health professional so that the scan machine will automatically identify and take measurements of babies. In order to automate the process, very large numbers of videos in different formats from ultrasound scans are necessary to train computers through a process called machine learning. To do this, we will ask the pregnant woman to have her scan video recorded. In parallel to recording the scans of babies, we will ask the sonographer to record her/his hand and eye-tracking while the scan is being performed. In the future, an automated image analysis during screening will make ultrasound scans more accessible to those living in areas where today ultrasound is not part of routine clinical care.

Principal Investigator

Prof Aris Papageorghiou

Contact us

Email: osprea@wrh.ox.ac.uk

IRAS number

225829