Artificial Intelligence Not Yet Ready for Clinical Breast Cancer Screening

FRIDAY, Sept. 17, 2021 (HealthDay News) — Artificial intelligence (AI) systems for the early detection of breast cancer from digital mammography are not yet ready for the clinic, according to a review published online Sept. 2 in The BMJ.

Karoline Freeman, from the University of Warwick in the United Kingdom, and colleagues conducted a systematic literature review to identify studies assessing test accuracy of AI algorithms, alone or in combination with radiologists, to detect cancer in women’s digital mammograms.

The researchers identified 12 studies (131,822 screened women) and found that in a subset of three retrospective studies (79,910 women, 1,878 cancers), 34 of 36 AI systems evaluated were less accurate than a single radiologist, and all were less accurate than the consensus of two or more radiologists. In contrast, five smaller studies (1,086 women, 520 cancers) showed that all five evaluated AI systems (as standalone to replace a radiologist or as a reader aid) were more accurate than a single radiologist reading a test set in the laboratory. Three studies using AI for triage screened out 53, 45, and 50 percent of women at low risk but also 10, 4, and 0 percent of radiologist-detected cancers.

“Current evidence for AI does not yet allow judgment of its accuracy in breast cancer screening programs, and it is unclear where on the clinical pathway AI might be of most benefit,” the authors write.

Abstract/Full Text