AI more accurate than breast density status for predicting cancer risk
Artificial intelligence could significantly improve breast cancer risk assessments, according to new data being presented during this week’s annual meeting of the Radiological Society of North America.
The findings suggest that AI assessments more accurately stratify an individual’s cancer risk compared to the standard of care that integrates information relative to patients’ age, family history, genetics and breast density. Constance D. Lehman, MD, PhD, professor of radiology at Harvard Medical School in Boston, says traditional assessments, including those related to density status, don’t go far enough.
“Over 2 million women are diagnosed with breast cancer annually, and for most, it comes as a complete shock,” Lehman said in an RSNA news release on the findings. “Only 5% to 10% of breast cancer cases are considered hereditary, and breast density alone is a very weak predictor of risk.”
Recently, Lehman and colleagues tested the first FDA-authorized image-only AI breast cancer risk model (Clairity Breast) on a set of over 200,000 screening mammograms from across five sites in the United States, plus an additional 8,810 from one site in Europe to compare its utility to standard prediction methods. They used follow-up data to determine accuracy, paying close attention to how radiologists’ assessments of patients’ breast density affected five-year outcomes. The AI risk model’s predictions were categorized using National Comprehensive Cancer Network thresholds: average (less than 1.7%), intermediate (1.7%–3%) and high (greater than 3%).
Researchers determined that the model, which was trained on longitudinal data from women who did and did not develop breast cancer over a period of five years, made significantly more accurate predictions compared to traditional methods. When accounting for breast density, the cases the model categorized as high-risk had more than a fourfold higher cancer incidence compared to those it categorized as average risk (5.9% vs. 1.3%). Conversely, breast density as the lone risk factor (dense vs. non-dense) did not significantly alter prediction accuracy (3.2% vs 2.7%).
While breast density status is an important factor to consider with respect to cancer risk, Lehman suggested that integrating AI assessments as a complimentary tool could have significant advantages.
“An AI image-based risk score can help us identify high-risk women more accurately than traditional methods and determine who may need screening at an earlier age,” Lehman said. “We already screen some women in their 30s when they are clearly at high risk based on family history or genetics. In the future, a baseline mammogram at 30 could allow women with a high image-based risk score to join that earlier, more effective screening pathway.”
In addition to her background in journalism, Hannah also has patient-facing experience in clinical settings, having spent more than 12 years working as a registered rad tech. She began covering the medical imaging industry for Innovate Healthcare in 2021.