AI shows value interpreting breast-cancer scans

March 2, 2020

AI shows value interpreting breast-cancer scans

Can a computer algorithm interpret a mammogram as well as an expert radiologist? Not yet – but in a study published today, the most accurate algorithm in a coding contest, when combined with a radiologist's reading of the mammogram, actually did improve the interpretations' agreement with the patients' clinical outcomes. 

Dr. Christoph Lee, a professor of radiology at the University of Washington School of Medicine, was co-senior author of the study. The findings, he said, show that artificial intelligence could help by "clearing the obviously negative mammograms" and allow expert radiologists to spend more time on patient scans that are more complex. Read our news release.


Terms of appropriate usage of file downloads

  • News reporters and news organizations may freely republish and distribute videos, still images and audio files produced by UW Medicine and the University of Washington School of Medicine.
  • Works must be attributed/credited appropriately (for example, “UW Medicine” – as denoted in the file) and must not be used for commercial purposes.
  • These visual and audio files may not be used to exploit or misrepresent UW Medicine or the University of Washington.
  • UW Medicine often licenses still images from Thinkstock but cannot grant republishing rights. You may not republish single image files credited to Thinkstock.
  • Logos of UW Medicine and University of Washington Health Sciences schools may not be republished without explicit permission. Contact us by phone or email: 206.543.3620 or