AI-Powered Ultrasound Tool Receives FDA Clearance for More Accurate Delivery Date Prediction

A new artificial intelligence (AI) tool, Delivery Date AI, has received De Novo clearance from the Food and Drug Administration (FDA) to assist healthcare providers in predicting a pregnant woman’s delivery date using standard ultrasound images. Developed by Ultrasound AI, this cloud-based software aims to improve the accuracy of delivery date estimations, particularly in cases where traditional methods are unreliable.

How Delivery Date AI Works

Delivery Date AI analyzes entire ultrasound images, identifying fetal and maternal characteristics that may be missed by standard biometric methods. The technology utilizes deep-learning neural networks trained on millions of de-identified ultrasound images from diverse pregnancies and clinical settings [1]. It provides a Predicted Delivery Date (PDD) in real-time, aiding clinical decision-making.

Addressing the Challenges of Traditional Dating Methods

Traditional methods for estimating delivery dates, such as last menstrual period (LMP) and early ultrasound gestational-age dating, can be inaccurate. Delivery Date AI is designed as an adjunctive tool to complement clinical judgment when these traditional methods are unreliable. It doesn’t replace the expertise of healthcare professionals but provides an additional data point for more informed care planning [1].

Benefits and Applications

• Improved Accuracy: The AI analyzes a broader range of ultrasound data, potentially leading to more precise delivery date predictions.
• Real-Time Results: The PDD is provided quickly, facilitating timely clinical decisions.
• Compatibility: The software is compatible with most existing ultrasound machines and installs quickly [3].
• Scalability: Its cloud-based nature allows for potential use in various healthcare settings, including hospitals, imaging centers, and clinics.

Clinical Validation

Delivery Date AI has been evaluated in a peer-reviewed study published in The Journal of Maternal-Fetal & Neonatal Medicine, conducted in collaboration with the University of Kentucky [2]. A study of over 5,700 patients demonstrated an R-squared value of 0.92 in predicting days to delivery using only ultrasound images [3].

Availability

With FDA clearance, Delivery Date AI is now available to practices, hospitals, imaging centers, and ultrasound equipment partners across the United States [2]. The technology has likewise received approval in Brazil (ANVISA registration number: 82499100004) and is currently offered in Chile under local regulatory allowances [1].

Looking Ahead

The FDA clearance of Delivery Date AI represents a significant advancement in prenatal care, offering the potential to reduce uncertainty in delivery date predictions and improve outcomes for mothers and babies. As AI continues to evolve, we can expect further innovations in medical imaging and diagnostic tools.