AI collaborating in spotting Breast Cancer

A new Nature Cancer study explores whether AI can safely step into one of healthcare’s most critical workflows: breast cancer screening.

Using data from ~50,000 women in NHS (UK Health service) screening programmes, researchers tested replacing the traditional second human reader of mammograms with an AI system.

AI can match human performance as a second reader in screening workflows and proves to be 'non-inferior' (no worse) than a second radiologist in accuracy, while significantly reducing workload.

AI worked well, but not independently. Human oversight remained essential, especially in edge cases and disagreements.

With the growing shortage of radiologists, integrating AI into the process leads to result being processed faster and allows radiologists to focus their time and care on patients who needed it the most.

SOURCE

https://www.nature.com/articles/s43018-026-01128-z

https://www.youtube.com/watch?v=4BzoMc3acOY

Previous
Previous

Will the US midterms be driven by AI?

Next
Next

Does AI make things up?