Performance across different versions of an artificial intelligence model for screen-reading of mammograms
Question Studies have reported promising results regarding the use of AI in mammography screening, but comparisons of updated versus older versions are less studied. Findings In our study, 87.1% (642/737) of the screen-detected cancers were classified with a high malignancy risk score by the old version, while it was 93.5% (689/737) for the newer version. Clinical relevance Understanding how version updates of AI models might impact screening mammography performance will be important for future quality assurance and validation of AI models.
Authors
- Jan F. Nygård (ORCID: https://orcid.org/0000-0001-9655-7003)
- Heather A. Carlos (ORCID: https://orcid.org/0000-0002-8182-8648)
- Solveig Roth Hoff (ORCID: https://orcid.org/0000-0001-7928-3821)
- Håkon Lund-Hanssen (ORCID: https://orcid.org/0000-0002-1418-7799)
- A M Holen
- Marie Burns Bergan
- Steinar Auensen
- Giske Ursin
- Solveig Hofvind
- Yan Chen
- Marthe Larsen
- Kristina Lång
Institutions
- University of Southern California (US)
- Norwegian Institute of Public Health (NO)
- University of Wisconsin–Madison (US)
- University of Nottingham (GB)
- University of Oslo (NO)
- Lund University (SE)
- Norwegian University of Science and Technology (NO)
- Nottingham City Hospital (GB)
- Ålesund Hospital (NO)
- St Olav's University Hospital (NO)
- Cancer Registry of Norway (NO)
- Skåne University Hospital (SE)
- UiT The Arctic University of Norway (NO)
Publication Details
- Journal
- European Radiology
- Published
- 2026-01-13
- DOI
- https://doi.org/10.1007/s00330-025-12240-6
- Primary Topic
- AI in cancer detection
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- Norwegian Institute of Public Health