01Fracture detection

Evaluating fracture-detection AI: the questions to ask

Independent studies, results by region, patient age, clearance, integration, testing on your own cases: a checklist to compare fracture-detection software.

Medically reviewed by Dr Alexandre Parpaleix, MD-PhD, CEO and co-founder of Milvue

Which studies should you look at first?

Independent studies, published in peer-reviewed journals, that compare several software products on the same images (head-to-head studies). Prefer consecutive series of real patients, for example every emergency visit over a period, to image sets enriched in fractures. Check who funded the study and who set the reference standard.

Sensitivity, specificity, accuracy, AUC: what does each figure measure?

Sensitivity is the share of real fractures the AI flags; specificity, the share of fracture-free exams it leaves without an alert; accuracy, the share of correct answers overall. The AUC sums up performance across all possible thresholds. Tell the standalone evaluation of the AI apart from the reader study, which measures what radiologists gain when they read with the AI: the second figure is the one that describes real use.

Why ask for results by anatomical region?

A global figure hides large gaps from one region to another. Ask for sensitivity and specificity on the regions you read most: wrist, ankle, elbow, hip, ribs.

What about children?

Growing bone, cartilage and normal variants mislead models trained on adults. Check the population covered by the clearance (minimum age, in each territory) and whether independent pediatric studies exist, with results by region. The Milvue Suite's fracture detection is CE marked with no age limit in Europe; in the United States, TechCare Trauma is FDA 510(k) cleared for fractures from birth, depending on the body region; in Canada, SmartTrauma is licensed by Health Canada with no age limit.

Which regulatory clearance should you check?

In Europe, CE marking under the MDR 2017/745, the device class and the notified body. In the United States, the 510(k) number, searchable in the FDA public database. In Canada, the Health Canada licence. In the United Kingdom, the NICE recommendation: in January 2025, the HTG739 assessment listed four fracture-detection technologies that can be used in the NHS during the evidence generation period, among them Milvue's TechCare Alert.

How does the tool fit into the reading workflow?

Results should reach the existing PACS, on the image, with no workstation change: DICOM in and out, IHE profiles. Look at the delay from acquisition to result, how suspicious exams move up the worklist, and how the tool shows uncertainty.

How do you test before deciding?

Run a retrospective series of your own consecutive exams through the software, with a reference set by your radiologists. Count missed fractures, and also false alerts per reading day: that is the load your teams will live with.

Where does Milvue stand on these criteria?

In an independent emergency department study comparing three commercial algorithms on 1,210 patients (Bousson et al., Academic Radiology, 2023), Milvue reached the highest fracture-detection accuracy in published independent head-to-head studies. With the AI, +22% reader sensitivity in adults and +20% in children under 22. In Europe, TechCare Alert is part of the Milvue Suite, a CE marked, Class IIa medical device; in the United States, TechCare Trauma is FDA 510(k) cleared (K242171); in Canada, SmartTrauma is licensed by Health Canada. Detection integrates natively with the RIS and PACS.

Book a demo

The Milvue solution TechCare Alert
Explore

See Milvue AI in real conditions.