Summary: Integrating AI with radiologists’ assessments could improve breast cancer screening, introducing an algorithm designed to accurately identify normal mammograms, potentially reducing false positives and unnecessary follow-ups while maintaining cancer detection rates.

Key Takeaways:

  1. Combining AI with radiologists’ evaluations can enhance breast cancer screening by accurately identifying normal mammograms, potentially reducing false positives and unnecessary follow-ups while maintaining cancer detection rates.
  2. The partnership between Washington University School of Medicine and led to WRDensity, an algorithm cleared by the FDA in 2020, facilitating breast density assessment on mammograms and identifying individuals for further screening.
  3. Incorporating AI into mammogram assessments could notably decrease callbacks for diagnostic exams and biopsies while preserving high accuracy in cancer detection, potentially reducing false positives.


A recent study from Washington University School of Medicine and suggests that combining AI with radiologists’ evaluations could improve breast cancer screening. Published in Radiology: Artificial Intelligence, the study introduces an algorithm specifically tailored to identify normal mammograms accurately. Simulations indicate that by allowing AI to handle low-risk scans, unnecessary follow-ups could decrease while preserving cancer detection rates.

“False positives are when you call a patient back for additional testing, and it turns out to be benign,” says senior author Richard L. Wahl, MD, a professor of radiology and radiation oncology at St. Louis-based Washington University’s Mallinckrodt Institute of Radiology (MIR). “That causes a lot of unnecessary anxiety for patients and consumes medical resources. This simulation study showed that very low-risk mammograms can be reliably identified by AI to reduce false positives and improve workflows.”

Wahl collaborated with on an algorithm for assessing breast density on mammograms, aiding in identifying individuals who may need additional screening. This algorithm, named WRDensity, obtained FDA clearance in 2020 and is currently marketed by

AI Enhances Mammogram Evaluation

In this study, Wahl and colleagues at worked together to develop a way to rule out cancer using AI to evaluate mammograms. They trained the AI model on 123,248 2D digital mammograms (containing 6,161 showing cancer) that were largely collected and read by Washington University radiologists. Then, they validated and tested the AI model on three independent sets of mammograms, two from institutions in the U.S. and one in the United Kingdom.

First, the researchers figured out what the doctors did: how many patients were called back for secondary screening and biopsies; the results of those tests; and the final determination in each case. Then, they applied AI to the datasets to see what would have been different if AI had been used to remove negative mammograms in the initial assessments and physicians had followed standard diagnostic procedures to evaluate the rest.

AI and Mammogram Accuracy

Consider the largest dataset of 11,592 mammograms. When scaled to 10,000 for simplicity in the simulation, AI identified 34.9% as negative. Removing these 3,485 negative mammograms would have resulted in 23.7% fewer callbacks for diagnostic exams and 6.9% fewer biopsies compared to reality.

Both the AI and standard-of-care approaches detected the same 55 cancers among the 10,000 mammograms. Essentially, this AI study suggests that out of 10,000 people, 262 could have avoided diagnostic exams and 10 could have avoided biopsies without missing any cancer cases.

“At the end of the day, we believe in a world where the doctor is the superhero who finds cancer and helps patients navigate their journey ahead,” says co-author Jason Su, co-founder and chief technology officer at 

“The way AI systems can help is by being in a supporting role. By accurately assessing the negatives, it can help remove the hay from the haystack so doctors can find the needle more easily,” Su adds. “This study demonstrates that AI can potentially be highly accurate in identifying negative exams. More importantly, the results showed that automating the detection of negatives may also lead to a tremendous benefit in the reduction of false positives without changing the cancer detection rate.”