Fracture
Conditions
Keywords
Aged, Radiography, Paediatric imaging, Artificial intelligence, child abuse
Brief summary
The artificial intelligence (AI) software BoneView (GLEAMER Company, Paris, France) has been designed, tested and validated to detect and locate recent or semi-recent fractures on standard radiographs. The objective will be to assess the AI performance for the detection of bone fractures in children aged less than 2 years old in suspected child abuse setting. These patients benefit from a whole body radiography with a double blind reading by a generalist radiologist and a radiologist with expertise in child abuse. This readings will be compared with the AI results. Hypothesis is that AI is effective for child fractures detection and could be of help especially for radiologists who are not experts in child abuse.
Interventions
None listed
Sponsors
University Hospital, Montpellier
Study design
Observational model
COHORT
Time perspective
RETROSPECTIVE
Eligibility
Sex/Gender
ALL
Age
0 Years to 2 Years
Healthy volunteers
No
Inclusion criteria
* aged less than 2 years old * whole body radiography performed for suspected child abuse setting * report available with a double blind reading (generalist radiologist and radiologist with expertise in child abuse)
Exclusion criteria
* Radiograph not interpretable ( poor quality) * AI not applicable
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Percentage of fracture detected by AI on radiographs | 1 day | Percentage of fracture detected by AI on radiographs |
Countries
France
Contacts
Primary ContactIngrid Millet, PUPH
Outcome results
None listed