top of page

Google AI to help detect breast cancer more accurately than medical professionals!

Updated: Apr 13, 2020


Google AI for Breast Cancer

Artificial Intelligence or AI technology does not need any introduction. However, in the last year or so, AI has shown impressive advancements particularly in the field of the health industry. AIs have proven their capability in detecting heart conditions early or even ocular diseases just like medical professionals. Now, Google is using an AI system to detect breast cancer.


As a matter of fact, Google is trying to improve breast cancer detection rates with the help of AI. According to a Google Blog post, over 55,000 people in the U.K are diagnosed with breast cancer annually and one in eight American women are likely to develop it at some point in their lifetime.


We all know that breast cancer is a well-known disease that is difficult to treat. The best chance of treating cancer lies in early detection. If the cancer symptoms set in, it may be in its advanced stage and difficult to treat.


Currently, mammograms that are used to detect early cancer detection miss about 20 percent of cases. An 80% success rate is good enough but it is far from ideal. However, Google’s artificial intelligence technology has already proven its capability of improving detection rates.


To prove, Google fed its AI 76,000 “de-identified” mammograms from UK women and 15,000 from US women. It was put to test in a separate data set that contained mammograms from over 25,000 UK residents and 3000 US residents.


During the test, the AI managed to reduce false positives by 5.7 % and false negatives by 9.4% (compared to average detection). According to Google, these results are promising and if used in actual clinical settings could help doctors “reduce stress and wait times” for potential cancer patients in the near future.


Whether Google’s AI will be successful in the real-world settings remains to be seen as there are several hurdles (legal and scientific) for the AI to clear.

bottom of page