Obstetrics_gynecology · Fetal ultrasound image-quality assurance and view detection
Sonio Detect
Sonio (Samsung Medison)
Sonio Detect solves an unglamorous but real screening problem: whether the sonographer actually captured the right view, complete and in focus, before the study is called adequate. It carries a solid regulatory footprint — an original 2023 FDA clearance, an extended-scope 2024 clearance, and a 2024 EU MDR Class IIa CE mark — and validation figures of more than 92 percent sensitivity for view and label detection across 17,000-plus images, with 93.4 percent sensitivity for fetal brain structures. Two honest limits hold it at one mark: the device-specific evidence is largely manufacturer-authored with limited independent prospective outcomes, and a single 2025 FDA malfunction report is on file (no injury). Samsung Medison acquired Sonio in 2024.
Performance Metrics
Clinical Evidence
The load-bearing performance figures come from the FDA submissions and manufacturer validation: Sonio Detect was validated on more than 17,000 ultrasound images and reported over 92 percent [[sensitivity]] for detecting image labels and view types, with the extended-scope 2024 clearance adding 93.4 percent sensitivity and 94.9 percent [[specificity]] for detecting fetal brain structures in the second and third trimesters. The company emphasises consistency of these figures across ultrasound manufacturers and across patient BMI, age, ethnicity and gestational age. The published foundational study (Stirnemann et al., Ultrasound in Obstetrics & Gynecology 2023) validated Sonio's real-time AI diagnostic companion on a retrospective set of several hundred phenotypes across two centres — important context for the platform, but it is manufacturer-authored and concerns the diagnostic-companion capability rather than Detect's image-QA function specifically. Independent prospective evidence on Detect's real-time quality feedback and its downstream effect on screening outcomes is limited; most device-specific validation is authored or affiliated with the manufacturer. As a quality-assurance aid the relevant question is whether better view capture improves anomaly detection — an outcome not yet established in independent trials.
| Study | Design | n | Sensitivity | Specificity | AUC | Published |
|---|---|---|---|---|---|---|
| Stirnemann J, et al. (manufacturer-authored; 2 centres) | RetrospectiveRetrospective | 549 | not reported as single figure | not reported as single figure | — | Ultrasound Obstet Gynecol 2023; real-time AI diagnostic companion validation (platform-level, not Detect-QA-specific) |
Clinical Pulse
Verified clinician sentiment
Verified clinician reviews launching soon · Apply to contribute →
Reviews are signed by clinicians verified against their regulator's public register.
Inside the algorithm
Editorial featureHow Sonio Detect checks a fetal scan in real time.
Five stages — from raw input to verdict — drawn from manufacturer documentation and the public regulatory record.
- INGEST
- NORMALISE
- DETECT
- LOCALISE
- VERDICT
Stage 01 · INGEST
Ultrasound images and clips stream in during the exam.
Sonio Detect receives fetal ultrasound images and cine clips as the sonographer acquires them, from ultrasound systems across manufacturers — the software is vendor-agnostic and validated on GE, Samsung and Canon systems. It operates across the first, second and third trimesters.
The input is the live acquisition; the check happens alongside the exam, not after it.
Input
Fetal ultrasound
Inference
Real-time / per-view
Inside the Auris+ Listing
Five more sections complete this device’s Auris+ Listing.
Decision Ledger
ProPro unlocks a private, cross-vendor log of every case you read on this device — what the AI called, what you concluded, and a one-line reason if you overrode it. Ready for the EU AI Act's deployer logging obligations when they land in 2028.
Clinical Evidence Deep Dive
ProPro unlocks the structured clinical-evidence summary — study count, target patient population, and a tabular accuracy-metrics view drawn from peer-reviewed sources.
Peer-Reviewed Publications
ProPro unlocks the curated peer-reviewed publication list with PubMed cross-links — the citation backbone of every editorial verdict.
Post-Market & Regulatory Conditions
ProPro unlocks the post-market surveillance summary, recall record, and the conditions of approval that bound real-world use.
AI Algorithm Version History
ProPro unlocks the chronological record of algorithm version changes — what changed when, drawn from manufacturer changelogs and regulatory filings.
Regulatory Approvals
Safety Record
One MAUDE malfunction report on file for Sonio Detect; no injury or death. Full narrative not retrieved — recorded as a single uninvestigated complaint, not a confirmed failure or recall.
One adverse-event report is on file: a single MAUDE entry categorised as a malfunction (brand "Sonio Detect", received 25 June 2025), with no injury or death reported; the full narrative could not be retrieved from this environment, so it is recorded as a single uninvestigated malfunction complaint, not a confirmed device failure or recall. No FDA recall was identified in public reporting as of July 2026 (MAUDE and the enforcement database could not be queried directly here). By design the software is a concurrent reading aid — it prompts recapture of inadequate views and does not make a diagnosis, so its harm surface is indirect: a missed quality flag could let an inadequate image pass, but the clinician remains responsible for interpretation and for the anomaly screen itself.
Intended Use & Indications
Sonio Detect analyses fetal ultrasound images and cine clips as they are acquired and interpreted, using machine-learning models to identify the standard screening view, detect the anatomical structures expected within it (for example heart and brain structures), and confirm the image satisfies the quality criteria that make a screening study adequate. It is manufacturer-agnostic — validated across ultrasound systems from vendors including GE, Samsung and Canon — and is indicated across the first, second and third trimesters (roughly gestational weeks 11 to 37). The product is a [[concurrent-reading]] quality aid: it flags whether a view is complete and in-spec, prompting the operator to recapture when it is not, and it does not diagnose fetal anomalies or replace the clinician's read. A separate module, Sonio Suspect, addresses anomaly flagging and is a distinct FDA submission. Sonio also markets a broader reporting and decision-support platform; Detect is specifically the image-QA and view-detection component.