Can AI Detect Breast Cancer on DBT Years Earlier than Radiologists?

In new study findings presented at the recent Society of Breast Imaging (SBI) conference, researchers found that the use of AI software may facilitate breast cancer detection on screening digital breast tomosynthesis (DBT) up to three years earlier than unassisted interpretation by radiologists.

For the retrospective study, researchers examined the utility of AI detection software (Genius AI Detection 2.0, Hologic) in a review of data from 341 women (mean age of 66) who had screening-detected cancer and three prior DBT screening exams deemed negative by radiologists.¹

For the index screening DBT in which the breast cancer was detected by the radiologist, the study authors found that the AI software correctly identified breast cancer in 87.7 percent of the cases.¹

However, researchers also determined that AI detected breast cancer on one prior DBT screening in 26.8 percent of patients, on two prior DBT exams in 8.7 percent of patients and on three prior DBT exams in 11 percent of patients.¹

“AI applied to screening DBT demonstrated the capability to detect breast cancers up to three years before radiologist detection,” noted study author Leslie Lamb, MD, MSc, an assistant professor of radiology at Harvard Medical School.

The study authors also pointed out that for those patients in whom AI detected breast cancer on at least one prior DBT exam, axillary lymph node biopsy results found that 8.8 percent had a lymph node-positive diagnosis.¹

“Earlier recognition of subtle findings could enable more timely diagnosis and intervention, supporting AI’s potential role as a complementary tool for advancing early breast cancer detection,” added Dr. Lamb.

Reference

1. Lamb LR. Artificial intelligence for earlier detection of breast cancer on screening digital breast tomosynthesis. Presented at the Society of Breast Imaging Symposium, April 16-19, Seattle. https://2026-sbi-symposium.eventscribe.net/