Why Anomaly Detection Matters in Prenatal Ultrasound
Prenatal ultrasound is a high-stakes environment where clinicians must interpret complex, fast-moving information in real time. Within that context, anomaly detection plays a critical role: it supports the identification of findings that may require closer evaluation, additional views, or escalation to expert review. Research shows that up to 51% of fetal anomalies* are missed during standard prenatal ultrasound screenings, with 31% of these missed cases due to misinterpretation of high-quality images*.
From a workflow standpoint, anomaly detection embedded in US machine is also a logical entry point for broader clinical adoption of AI. When designed appropriately, AI is designed to help reduce variability, support consistency across care settings, and reinforce best practices, while keeping the clinician firmly in control of interpretation and decision-making.
A Regulatory Milestone that Enables Broader Deployment
This clearance also builds on Sonio Suspect web-based application’s earlier FDA-clearance for integration in Sonio's cloud-native platform, now extending those capabilities into an integrated, on‑machine configuration and showing that automatic detection of abnormal fetal ultrasound findings can preserve performance when integrated in settings with reduced computational power.
With this clearance, Sonio Suspect now supports two configurations:
- Web-Based Application
- A new Integration into Third-Party Ultrasound Systems
This expanded configuration matters because it broadens access to Sonio Suspect: available via the web-based application or embedded directly into third‑party ultrasound systems, so AI support can fit naturally into existing ultrasound workflows.
The First AI Module in a US Machine for Anomaly Detection in Fetal Medicine
Sonio Suspect Integration Configuration is the first AI cleared for integration module in an US ultrasound machine for anomaly detection.
As AI becomes more common in medical imaging, it appears key to include AI in tools embedded directly into the ultrasound workflow, whether running in the cloud or on the machine, so AI support is available at each workflow step without adding friction.
Continued Momentum: Fifth Clearance in Under Three Years
This announcement reinforces a broader trajectory for Sonio: consistent delivery against regulatory pathways, and sustained investment in products intended to support clinical teams. Achieving a fifth FDA clearance in less than three years is a signal of both execution and long-term commitment to building AI that meets regulatory standards and clinical expectations.
What’s Next: Real-World Integration Testing
With this clearance secured, the next phase is focused on in‑real‑life testing to validate performance in clinical environments and ensure a smooth integration into existing ultrasound workflows moving forward.
*van Nisselrooij et al. (2020), Why are congenital heart defects being missed?. Ultrasound Obstet Gynecol, 55: 747-757.
“Achieving this FDA clearance for Sonio Suspect Integration Configuration is a major milestone for our team, and a strong validation of what’s possible when clinicians and the Sonio team build together. We’re deeply grateful to clinical partners who keep challenging our assumptions and grounding us in what matters most at the point of care, and to the Sonio teams across R&D, clinical, and regulatory, whose rigor turned that feedback into a cleared, deployable solution. This marks the first integration of an anomaly‑detection algorithm for fetal ultrasound directly into ultrasound machines, helping bring reliable, embedded AI seamlessly into everyday workflows and expanding access where it matters most.”
Rémi Besson
Chief Medical & Scientific Officer
