Constructing a clinical radiographic knee osteoarthritis database using artificial intelligence tools with limited human labor: A proof of principle
Authors' conclusion
This study shows that clinical knee OA databases can be built using AI with limited human reading time for uniform grading and measurements.
Study context
- Population
- Clinical consecutive knee radiographs over 6 years from patients aged 35–79, Bispebjerg-Frederiksberg Hospital, Denmark25,778 knee radiographs
- Participants
- 8,575 participants
- Outcome
- AI-built knee OA radiographic database: completeness and human labour saved
Methodological detailsproof of principle, consecutive, single-centre, Four artificial intelligence (AI) tools (two commercial, two custom-built) provided Kellgren-Lawrence (KL) grades, joint space widths, patella osteophyte, radiographic view, knee implant and radiographic marker detection
- Proof of principle
- Consecutive
- Single-centre
- Four AI tools (two commercial, two custom-built) provided KL grades, joint space widths, patella osteophyte, radiographic view, knee implant and radiographic marker detection
Main finding
As reported in the abstract
- Knee radiographs included
- 25,778
- Patients included
- 8575
- Knees with a complete osteoarthritis (OA) dataset
- 92.5%
- Radiologist reading time saved
- about 800 hours
- Images manually reviewed
- 16.0%
Observational data: associations, not effects.
Original publication
Constructing a clinical radiographic knee osteoarthritis database using artificial intelligence tools with limited human labor: A proof of principle. Osteoarthritis Cartilage. Published 1 December 2023. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →