Related Experiment Video
Updated: Sep 28, 2025
Detection of Architectural Distortion in Prior Mammograms via Analysis of Oriented Patterns
Published on: August 30, 2013
Artificial Intelligence Evaluation of 122 969 Mammography Examinations from a Population-based Screening Program
Marthe Larsen1, Camilla F Aglen1, Christoph I Lee1
1From the Section for Breast Cancer Screening (M.L., C.F.A., S.H.) and Department of Register Informatics (J.F.N.), Cancer Registry of Norway (G.U.), P.O. Box 5313, 0304 Oslo, Norway; Department of Health and Care Sciences, Faculty of Health Sciences, The Arctic University of Norway, Tromsø, Norway (S.H.); Department of Radiology, University of Washington School of Medicine, Seattle, Wash (C.I.L.); Department of Health Systems and Population Health, University of Washington School of Public Health, Seattle, Wash (C.I.L.); Department of Radiology, Ålesund Hospital, Møre og Romsdal Hospital Trust, Ålesund, Norway (S.R.H.); Department of Circulation and Medical Imaging, Faculty of Medicine and Health Sciences, National University for Science and Technology, Trondheim, Norway (S.R.H.); Department of Radiology and Nuclear Medicine, St Olavs University Hospital, Trondheim, Norway (H.L.H.); Department of Translational Medicine, Lund University, Lund, Sweden (K.L.); and Unilabs Mammography Unit, Skåne University Hospital, Malmö, Sweden (K.L.).
Artificial intelligence (AI) shows promise in mammographic screening for cancer detection. In a real-world setting, the AI system demonstrated strong performance, with less than 20% of screen-detected cancers being unselected across evaluated thresholds.
Area of Science:
- Radiology
- Medical Imaging
- Artificial Intelligence
Background:
- Artificial intelligence (AI) demonstrates potential in mammographic cancer detection.
- Real-world evidence on AI in screening settings is limited.
- This study evaluates AI performance in a population-based screening program.
Purpose of the Study:
- To compare a commercial AI system's performance against double reading with consensus in mammography.
- To explore histopathologic tumor characteristics associated with different AI scores.
Main Methods:
- Retrospective analysis of 122,969 mammographic examinations from BreastScreen Norway (2009-2018).
- Inclusion of 752 screen-detected and 205 interval cancers.
- AI scoring (1-10) and evaluation at three thresholds to assess performance as a binary tool.
Main Results:
- At threshold 1 (AI score 10), 86.8% of screen-detected cancers were identified.
- Using threshold 3 (similar to individual radiologist rate), 80.1% of screen-detected cancers were selected.
- Unselected screen-detected cancers had favorable histopathology; interval cancers showed opposite results.
Conclusions:
- The AI system did not select less than 20% of screen-detected cancers across the evaluated thresholds.
- The artificial intelligence system exhibited promising overall performance for cancer detection in mammographic screening.
