RESEARCH · RESEARCH · #1619
Google study: AI interprets blind-sweep ultrasounds to estimate gestational age and fetal presentation
Google researchers, working with Northwestern Medicine and Jacaranda Health, trained healthcare workers to perform predefined “blind sweep” scans on handheld devices and used a machine learning model to interpret the images. In a study of roughly 2,000 pregnancies in Nairobi and Chicago, the model detected gestational age and fetal presentation with accuracy comparable to trained sonographers.
KEY POINTS
- Google researchers, working with Northwestern Medicine and Jacaranda Health, trained healthcare workers to perform predefined “blind sweep” scans on handheld devices and used a machine learning model to interpret the images.
- In a study of roughly 2,000 pregnancies in Nairobi and Chicago, the model detected gestational age and fetal presentation with accuracy comparable to trained sonographers.
- If replicable and deployable, this approach could expand access to basic prenatal imaging in low-resource settings by allowing minimally trained operators and handheld devices to produce clinically useful results.
WHY IT MATTERS
If replicable and deployable, this approach could expand access to basic prenatal imaging in low-resource settings by allowing minimally trained operators and handheld devices to produce clinically useful results.