Google has revealed that it has developed an AI system with the ability to detect the presence of breast cancer more accurately than doctors.
The study, which was conducted in collaboration with colleagues at DeepMind, Cancer Research UK Imperial Centre, Northwestern University and Royal Surrey County Hospital, was recently published in the scientific journal Nature.
By using tens of thousands of mammograms from women in the UK and the US, the program was trained to detect cancer and so far early research has shown that it can produce more accurate detection than human radiologists.
- DeepMind's new AI learns like a human
- NHS trials AI software to beat breast cancer
- Google makes it easier to move data to its cloud
According to the study (opens in new tab), Google's AI technology led to fewer false positives, where test results say someone has cancer when they don't, as well as false negatives, where existing cancer goes undetected. The program was able to reduce false positives by 5.7 percent for US subjects and 1.2 percent for UK subjects while also reducing false negatives by 9.4 percent for US subjects and 2.7 percent for UK subjects.
AI and healthcare
Despite having less information such as patient histories and prior mammograms to work with, Google's AI system was more accurate than human experts.
Director of the Cancer Research UK Imperial Centre and one of the authors of the study, Professor Ara Darzi was quite impressed with the system's results and explained to CNN Business how it could improve doctor's productivity, saying:
"This is one of those transformational discoveries you have in your hand, which could disrupt the way we deliver screening in terms of improving accuracy and productivity."
In the UK, two radiologists are needed to interpret each mammogram but the new AI system could eventually be used to replace the second reader according to Darzi, though the system is not yet at the point where it could replace humans entirely.
- Also check out the best Electronic Health Record software
Via CNN (opens in new tab)