by
Gus Iversen, Editor in Chief | October 29, 2025
A recent study published in Radiology highlights the potential of an AI tool to identify women at elevated risk for interval breast cancers; those diagnosed between routine screening mammograms and often associated with poorer outcomes.
Researchers from the University of Cambridge examined more than 134,000 mammograms performed between 2014 and 2016 across two centers in the U.K.’s triennial Breast Screening Program. The cohort included women between the ages of 50 and 70, with 524 cases of interval breast cancer.
The team applied Mirai, a deep learning-based algorithm, to negative screening mammograms to produce individualized three-year risk scores. These predictions were based on imaging data alone, such as tumor characteristics and breast density, rather than clinical or demographic information.
Ad Statistics
Times Displayed: 5266
Times Visited: 18 Brand-New FDA-cleared Advanced Ultrasound Medical Device available for sale or lease to Wound Care Centers or any other Medical Facilities.The Arobella 1000D is designed for non-contact or debridement ultrasound wound healing therapy, or any other wounds
The model's risk scores retrospectively flagged a substantial proportion of interval cancers. Specifically, 42.4% of interval cancers occurred in women within the top 20% of predicted risk scores. At lower thresholds, the tool identified 3.6% of cancers in the highest 1% of risk scores, and 26.1% in the top 10%.
“Our results suggest that further workup of mammograms within the top 20% of scores could yield 42.4% of interval cancers, meaning that Mirai could be used to identify women for supplemental imaging or a shortened screening interval, instead of or in addition to breast density,” said lead author Joshua W. D. Rothwell, an M.B.B.S./Ph.D. student at Cambridge.
While performance dropped in women with extremely dense breast tissue, the model outperformed traditional risk stratification tools overall and was most effective at predicting cancers that developed within a year of the original screening.
The findings support potential use of AI to refine screening intervals or target supplemental imaging like MR or contrast-enhanced mammography. However, logistical considerations remain. “If we called back 20% of women for supplemental imaging, we’d have to find the capacity to offer contrast-enhanced mammography or MR to 440,000 women,” said co-author Dr. Fiona Gilbert, professor of radiology at Cambridge.
The team plans to evaluate other commercial AI models, conduct cost-effectiveness analyses, and design trials to assess AI-guided screening strategies.
