AI Tools Show Potential to Detect Subtle Signs of Interval Breast Cancer

Subtle findings on prior images may be missed, while even normal-appearing studies can precede elevated cancer risk. Against this backdrop, investigators reviewed commercially available AI tools for screening mammography to assess their ability to detect interval cancers and flag early risk signals.

Led by UCLA Health Jonsson Comprehensive Cancer Center, the review examined AI systems that assist radiologists in interpreting screening mammograms. These tools analyze images to highlight subtle abnormalities that were present but unrecognized on earlier exams. Some systems also generate risk scores from ostensibly normal mammograms to identify patients who may develop cancer before the next round of screening. The work was published in the Journal of Breast Imaging on August 18, 2026.

Across retrospective studies, AI identified a substantial share of interval cancers with prior subtle signs, though estimates varied widely from about 5% to 78% depending on study design, algorithm, and methodology. These findings reflect potential detection rather than proven reductions in interval cancer rates. A large 2026 randomized trial of more than 105,000 women found AI-supported screening had an interval cancer rate of 1.55 per 1,000 versus 1.76 with standard double reading, with higher sensitivity (80.5% vs. 73.8%) and identical specificity (98.5%). Workload decreased by 44.3% without lowering the overall detection rate.

Evidence translation to routine practice remains uncertain. Definitions of interval cancer, screening technologies, intervals, and workflows varied across studies, and few algorithms have been evaluated widely. The authors call for prospective research in settings aligned with U.S. programs, long-term outcomes assessment, and post-market surveillance to establish safety, effectiveness, and optimal clinical responses to high AI risk scores when no visible abnormality is present. Authors were from UCLA and Larkin Health System.

Related Links
UCLA Health Jonsson Comprehensive Cancer Center