Population-wide evaluation of artificial intelligence and radiologist assessment of screening mammograms

• Published in: European radiology 2024-06, Vol.34 (6), p.3935-3946
• Main Authors: Kühl, Johanne, Elhakim, Mohammad Talal, Stougaard, Sarah Wordenskjold, Rasmussen, Benjamin Schnack Brandt, Nielsen, Mads, Gerke, Oke, Larsen, Lisbet Brønsro, Graumann, Ole
• Format: Article
• Language: English
• Subjects: Accuracy; Aged; Arbitration; Artificial Intelligence; Breast; Breast cancer; Breast Neoplasms - diagnostic imaging; Denmark; Diagnostic Radiology; Early Detection of Cancer - methods; Female; Humans; Imaging; Internal Medicine; Interventional Radiology; Mammography; Mammography - methods; Mass Screening - methods; Medicine; Medicine & Public Health; Middle Aged; Neuroradiology; Population (statistical); Radiologists - statistics & numerical data; Radiology; Sensitivity; Sensitivity and Specificity; Statistical analysis; Subgroups; Ultrasound
• ISSN: 1432-1084; 0938-7994; 1432-1084
• DOI: 10.1007/s00330-023-10423-7

Objectives To validate an AI system for standalone breast cancer detection on an entire screening population in comparison to first-reading breast radiologists. Materials and methods All mammography screenings performed between August 4, 2014, and August 15, 2018, in the Region of Southern Denmark with follow-up within 24 months were eligible. Screenings were assessed as normal or abnormal by breast radiologists through double reading with arbitration. For an AI decision of normal or abnormal, two AI-score cut-off points were applied by matching at mean sensitivity (AI sens ) and specificity (AI spec ) of first readers. Accuracy measures were sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and recall rate (RR). Results The sample included 249,402 screenings (149,495 women) and 2033 breast cancers (72.6% screen-detected cancers, 27.4% interval cancers). AI sens had lower specificity (97.5% vs 97.7%; p