Mammography Study: Multi-Stage Use of AI for DBT Exams Yields Over 21 Percent Increase in Breast Cancer Detection

Could a multi-stage AI-driven workflow lead to significant improvement in breast cancer detection with digital breast tomosynthesis (DBT)?

In the new multicenter ASSURE study, recently published in Nature Health, researchers assessed the combination of artificial intelligence detection with Saige-Dx (DeepHealth/RadNet) with an AI-supported safeguard review evaluation (ASSURE) to facilitate a second review of the top 10 percent of malignancy scores identified with Saige-Dx.

In comparison to the standard of care, the multi-stage AI-driven workflow led to a 21.6 percent increase in breast cancer detection. In a recent interview with Diagnostic Imaging, study co-author Christoph Lee, M.D., noted that the multi-stage approach with AI demonstrated a consistent improvement in the cancer detection rate (CDR) across women of different races, ethnicities and breast densities.

“ … The AI-driven workflow actually detected just as many additional cancers in women with dense breasts versus (women with non-dense breasts. So it gives a level of hope for improved cancer detection among women with dense breasts,”pointed out Dr. Lee, a professor and vice chair in the Department of Radiology at the University of Wisconsin-Madison School of Medicine and Public Health.

While the multi-stage AI-enabled workflow had a 5.7 percent increased recall rate, Dr. Lee also noted a 15 percent increase in the positive predictive value (PPV).

“ … Yes, we are calling more women back with this program, but we're detecting more cancers because of it, and not leading to an inordinate amount of false positive results. So, the balance of risk versus benefits really tilts toward potentially more benefits with this multi-stage AI-driven workflow,” emphasized Dr. Lee.

(Editor’s note: For related content, see “Predicting Interval Breast Cancer Risk: Can a Mammography Deep Learning Model Have an Impact?,” “Mammography Study Shows Elevated Future Breast Cancer Risk with Initial Concordance of Radiologist and AI Interpretation” and “Mammography Research: Can Adjunctive AI for DBT Provide Insights into Tumor Biology with Breast Cancer?”)

For more insights from Dr. Lee, watch the video below.