Experts Find That New AI Is Officially Better Than Humans In Diagnosing Breast Cancer
The fight against breast cancer just got a ray of hope especially for those who need to go for regular screenings and those who may be healthy but receive false alarms by their bodies. Digital mammographies have evolved one step further where artificial intelligence is now being used to detect and diagnose the disease. Yes, scientists have invented the procedure of AI diagnosing breast cancer and found that it is more accurate than human diagnosis.
1Scientists invent artificial intelligence to detect breast cancer with more accuracy than humans
A Google-based model has beaten radiologists in the ability to spot breast cancer simply by scanning X-ray results. Even for experts, it can be difficult to read mammograms and it can be easy to make false positive or negative results. Not only do these inaccuracies lead to delay of treatment, they can also create unnecessary depression in patients without risk. Often such problems lead to delays, stress, and a stressful working environment for doctors and radiologists who are in short supply everywhere.
Image Source: www.telegraph.co.uk
2The AI system created by Google Health
The massive breakthrough was revealed in a Google blogpost where the new AI system to detect breast cancer was developed by Google Health. The programme was in collaboration with DeepMind Cancer research UK Imperial Center, Northwestern University and Royal Surrey County Hospital. According to researchers, the AI model was programmed and trained to tune in to annonymised mammograms done on 76,000 women from the UK and 15,000 from the US. The scientists wanted to test it to see if it could spot various forms of breast cancer or wherever there were none. In effect, it was to test the accuracy of the AI in spotting breast cancer.
Image Source: health.google.com
3Tested on thousands of women with positive results
After the initial phase, the AI was also tested on a separate selection of data involving a further 25,000 women from the United Kingdom and again 3000 women in the US. The results were better than expected. The researchers found that where false positives were concerned, the AI showed a 1.2% reduction. A false positive is when a mammogram is reported or diagnosed as abnormal even when there is no cancer. The reduction of false negatives showed a 2.7% reduction. False negatives are when a report states normal even though there is a presence of breast cancer. These were for the figured in the UK.
Image Source: cdc.gov
