How Google Is Using AI To Improve Breast Cancer Screening

By Amit Chowdhry ● January 1, 2020
  • Google is using artificial intelligence to improve breast cancer detection by scanning mammograms. Here’s how it works.

It has now been proven that Google’s artificial intelligence technology system does a superb job detecting breast cancer by scanning mammograms. A study about the technology was put together by Alphabet’s DeepMind AI unit — which combined with Google Health several months ago.

This technology could be used to help reduce errors, which is important because one in eight women is affected by breast cancer globally.

The team also included researchers from Cancer Research UK Imperial Centre, Northwestern University, and Royal Surrey County Hospital. The researchers trained the artificial intelligence system to identify breast cancers on tens of thousands of mammograms. From there, the researchers then compared the predictions to the actual results from a set of 25,856 mammograms in the United Kingdom and 3,097 from the United States.

ADVERTISEMENT

And the study was published in the Nature journal today. And Northwestern University Research Assistant Professor Mozziyar Etemadi pointed out that this represents an advancement for potential technology to be used for early detection of breast cancer.

It is estimated that radiologists miss about 20% of breast cancers in mammograms, according to the American Cancer Society. And half of all women who get screenings over a 10-year period will have a false-positive result.

“Reading these X-ray images is a difficult task, even for experts, and can often result in both false positives and false negatives. In turn, these inaccuracies can lead to delays in detection and treatment, unnecessary stress for patients and a higher workload for radiologists who are already in short supply,” wrote Google Health Technical Lead Shravya Shetty and Google Health product manager Daniel Tse in a blog post. “Over the last two years, we’ve been working with leading clinical research partners in the U.K. and U.S. to see if artificial intelligence could improve the detection of breast cancer. Today, we’re sharing our initial findings, which have been published in Nature. These findings show that our AI model spotted breast cancer in de-identified screening mammograms (where identifiable information has been removed) with greater accuracy, fewer false positives, and fewer false negatives than experts. This sets the stage for future applications where the model could potentially support radiologists performing breast cancer screenings.”

The study proved that the AI system would be able to identify cancer with a similar level of accuracy as expert radiologists. And it reduced the number of false-positive results by 5.7% for the US-based group and 1.2% for the British-based group.

As part of a separate test, the group compared the AI system against six radiologists and found that the AI outperformed the radiologists at accurately predicting breast cancer.

Harvard Medical School professor of radiology Connie Lehman said that these results are comparable to findings from other groups using AI for improving cancer detection in mammograms reported Reuters.