A newly approved artificial intelligence tool could soon make prenatal ultrasounds far more reliable, helping doctors detect problems that are often missed during routine pregnancy scans.
Real-Time Support During Ultrasound Exams
The US Food and Drug Administration has cleared an AI-powered software developed by American start-up BioticsAI for clinical use. Designed to work with existing ultrasound machines, the system analyses images in real time while a scan is taking place, alerting clinicians to possible malformations or areas of concern as they examine a developing fetus.
Ultrasound scans are a standard part of prenatal care and play a vital role in monitoring fetal development. However, many abnormalities affecting organs or limbs can go undetected. This new tool aims to act as an extra set of eyes for doctors, improving accuracy without replacing human judgement.
How the AI Tool Works
As soon as an ultrasound image is captured, the software checks its quality and suggests adjustments if the image is unclear. It then verifies whether all essential parts of the baby are visible, flagging anything missing from the scan.
By comparing the images to patterns drawn from global data, the system can identify signs of potential issues, including heart and limb defects. After the scan, it automatically compiles its findings into a report that doctors can review, a process developers say can save clinicians around eight minutes per patient.
The FDA’s clearance confirms the software meets required safety and performance standards and can be integrated into clinical workflows as a regulated medical device.
Why Better Detection Is Needed
Congenital abnormalities remain a significant challenge in prenatal care. According to European Commission figures, major birth defects occur in nearly 24 out of every 10,000 births across Europe.
A major review by Cochrane, which analysed data from more than seven million pregnancies, found that a single early ultrasound scan between 11 and 14 weeks detects just 38 percent of birth defects. A mid-pregnancy scan between 18 and 24 weeks identifies around 51 percent. When both scans are combined, detection improves to 84 percent.
Similar AI-driven solutions are also emerging in Europe. French companies Diagnoly and Sonio Detect have received approval for tools that automatically identify fetal structures and heart conditions, signalling a broader shift toward technology-assisted prenatal screening.
