Interpreting mammography results "remains challenging," cancer researchers at Google Health wrote in a new study published recently in the journal Nature.
“False positives can lead to patient anxiety, unnecessary follow-up and invasive diagnostic procedures,” they wrote, adding, “Cancers that are missed at screening may not be identified until they are more advanced and less amenable to treatment.”
Artificial intelligence might be able to help, the study found.
» NIH links hair dyes, straighteners to increased risk of breast cancer
Breast cancer is the second leading cause of death from cancer in women, but early screening can help spot the cancer early to begin treatment.
So cancer researchers collaborated with Google Health to create an AI model that was trained to read digital breast cancer scans of thousands of women in the United States and United Kingdom.
The test set from the U.S., where women are screened every one to two years, consisted of mammograms collected between 2001 and 2018 from 3,097 women at one academic medical center.
» Molecular patterns could indicate whether breast cancer returns, study finds
The UK test set consisted of screening mammograms collected between 2012 and 2015 from 25,856 women at two screening centers in England.
“We show an absolute reduction of 5.7% and 1.2% (USA and UK) in false positives and 9.4% and 2.7% in false negatives,” the researchers wrote.
The Google AI model outperformed the six radiologists who read the screenings by 11.5%.
Why is this important? Researchers found the “AI system maintained non-inferior performance and reduced the workload of the second reader by 88%,” freeing up radiologists and other doctors to focus on patient care.
You can read the full study here.
» New blood test could help with early detection of breast cancer
» These two common foods could lower breast cancer risk, study suggests
