Identifying breast cancer from routine scans can become more comfortable with the new computer program that beats human experts’ accuracy. Researchers hope their study and results are the breakthroughs we all need when fighting against an unforgiving killer. One of the most common cancers in women is breast cancer, so it is essential to perform regular screening and detect the early signs of patients that show no symptoms.
It is advised for women over 50 to get a mammogram once every three years and have the results analyzed by two different experts. Nonetheless, the interpretation of scans can lead to false positives or false negatives. Here is where researchers at Google Health are trying to improve cancer detection. They have trained an AI model to detect breast cancer after looking at thousands of scans from patients in the UK and the US.
The images the AI model received were also reviewed by doctors. Researchers found that the AI model was as good as expert radiographers and what bested the doctors was that it reduced the cases of incorrectly identified cases – 5.7% in the US and 1.2% in the UK. The AI also reduced the missed diagnoses by 9.4% in the US and 2.7% in the UK.
Think of this AI as a Second Opinion
Dominic King, UK lead at Google Health, stated that it is best to identify breast cancer the earliest possible and that their tech “supports and enables an expert, or a patient ultimately, to get the best outcome from whatever diagnostics they’ve had.”
The Google Health team compared their AI with an expert’s opinion. When the two diagnoses were the same, the case was resolved. When the two results were different, the AI was asked to compare its result with the second expert’s opinion. Using the AI to verify experts’ reviews and diagnosis could save a lot of work – 88% of the workload, to be more exact. King added:
“Find me a country where you can find a nurse or doctor that isn’t busy. There’s the opportunity for this technology to support the existing excellent service of the (human) reviewers.” The research team admits they need to conduct more research before this tech is able to act as the “second opinion” in cancer diagnosis.