AI Model Reduces False Positives in Screening Mammograms

THURSDAY, April 25, 2024 (HealthDay News) — A semiautonomous breast cancer screening system reduces false positives with screening mammograms, according to a study published online April 10 in Radiology: Artificial Intelligence.

Stefano Pedemonte, Ph.D., from Whiterabbit.ai in Santa Clara, California, and colleagues evaluated the ability of a semiautonomous artificial intelligence (AI) model to identify screening mammograms not suspicious for breast cancer. The analysis compared performance of humans and AI using three nonoverlapping datasets of 14,831 screening mammography examinations (1,026 cancers).

The researchers found minimal changes to the cancer detection rate with use of the AI device (noninferiority margin of 0.25 cancers per 1,000 examinations). For U.S. dataset 1 (11,592 mammograms; 101 cancers), the AI model reduced screening examinations requiring radiologist interpretation by 41.6 percent, diagnostic examination callbacks by 31.1 percent, and benign needle biopsies by 7.4 percent. For U.S. dataset 2 (1,362 mammograms; 330 cancers), reductions were 19.5, 11.9, and 6.5 percent, respectively. For the U.K. dataset (1,877 mammograms; 595 cancers), reductions were 36.8, 17.1, and 5.9 percent, respectively.

“This work demonstrates the potential of a semiautonomous breast cancer screening system to reduce false positives, unnecessary procedures, patient anxiety, and medical expenses,” the authors write.

Several authors disclosed financial ties to Whiterabbit.ai.

Abstract/Full Text (subscription or payment may be required)