Oct 25, 2019

AI reshapes how doctors treat breast care patients

An algorithm is helping researchers map the medical history of breast cancer patients so they can better predict, treat and maybe even prevent it, The New York Times reports.

The big picture: This database covers more than 100,000 patients over 30 years, from Massachusetts General hospital. That's a lot more data than oncologists can get from clinical trials, the Times notes.

Why it matters: AI could help us understand how tumors responded to different treatments. Machines could also make it easier for clinicians to identify patients with specific disease characteristics and to enroll them in clinical trials.

The other side: Science Magazine published an article Thursday explaining how black patients were susceptible to racial bias in treatment when a popular algorithm predicted who would benefit from extra medical care.

  • Risk scores underestimated the needs of black patients compared to white patients.
Go deeper