Associate Professor
Janni Jensen
UNIFY, Department of Radiology
| Projekt styring | ||
| Projekt status | Open | |
| Data indsamlingsdatoer | ||
| Start | 01.02.2025 | |
| Slut | 01.02.2026 | |
External testing of a Deep Learning Algorithm for Fracture Detection: A retrospective & diagnostic accuracy study
Short summary
Testing of a Deep Learning Algorithm for fracture detection on radiographs, compared to expert opinion.
Rationale
Fracture detection using radiography is a common task in emergency departments, as traumatic extremity fractures are a leading cause of visits worldwide. Missed fractures on radiographs are a common cause of diagnostic discrepancies between clinicians and the Department of Radiology. To adress these challenge, artificial intelligence algorithms have emerged as an encouraging solution for fracture detection. Our study aim to externally test an AI fracture detection algorithm, with the specific objective to estimate fracture detection sensitivity and specificity.
Description of the cohort
Retrospectively, patients referred from the Emergency Department at OUH to the Department of Radiology for radiographic assessment of suspected fractures are included in the study population.
Data and biological material
Radiographs. Data from radiology report.