PARADIGM: The diagnostic potential of 3D ultrasound with AI in maternity care

Problem description:

A national study is currently taking place in the Netherlands to determine the value and efficiency of structural ultrasound examination at 13 weeks gestation. Routine prenatal ultrasound scanning is a time and labour-intensive process. The 2D planes of interest are acquired by a trained, skilled sonographer during the scanning process and fetal measurements are made on these images. To acquire all the images and measurements needed requires a visit of at least 20 minutes for the patient and sonographer. This process could be streamlined through the use of 3D ultrasound and Artificial Intelligence.

Project summary

Modern ultrasound equipment is capable of acquiring 3D volumetric images in an instant, using a specialized probe. Since the positioning of the fetus in these images is not fixed it is however, not trivial for a viewer to find the 2D planes of interest within this volume. The PARADIGM project focuses on the development of AI tools for analysis of 3D ultrasound images in collaboration with the RTC Deep Learning. The project will focus on comparison with the current state-of-the-art by automatic extraction of the 2D planes of interest and provision of the images and measurements of interest. In the longer term the advantages of volumetric imaging could be exploited for improved analysis and accuracy.

Funding

This research project is funded by the Nijmegen Foundation for Prenatal Screening.

People

Keelin Murphy

Keelin Murphy

Assistant Professor

Bram van Ginneken

Bram van Ginneken

Professor, Scientific Co-Director

Tessa Meuleman

Tessa Meuleman

Sonographer

Gynaecology, Radboudumc

Marc Spaanderman

Marc Spaanderman

Professor Obstetrics

Gynaecology, Radboudumc

Ajay Patel

Ajay Patel

RTC Deep Learning

RTC Deep Learning

Ward Hendrix

Ward Hendrix

Research Scientist RTC Deep Learning

RTC Deep Learning