While Googles latest phones caught the headlines at this week’s Googles i/o keynote – the internet giant also revealed its latest – potentially life-saving – advancements in AI tech.
Google described progress on its artificial intelligence as “promising” and revealed that its AI can now spot lung cancer almost year before a human doctor can. Early detection is key to increasing survival chances for patients.
Google AI is introducing Deep Learning model that can anlyze CT scans and detect lung malignancies. Can help detect early scan cancers that doctors might not be able to detect. Increased survival rate of 40%. #GoogleIO2019 #io19 pic.twitter.com/llqGDppeXJ
— Lance Ulanoff (@LanceUlanoff) May 7, 2019
Google’s health researcher, Lily Peng, explained at it’s i/o developer conference in Mountain View, California that its current ‘deep learning’ model can spot very subtle lung lesions on a computed tomography (CT) scans.
These tiny lesions are so small they are often missed by radiologist. However, the AI can detect them up to a year before a traditional diagnosis.
Peng explained that that year lead time might “translate into an increased survival rate of 40 per cent”.
The research is a result of Google’s AI team working with Google parent Alphabet’s life science company Verily.
Earlier this year Verily was disappointed when it was forced to abandon work on smart contact lens designed to monitor the glucose levels of diabetics (while Verily was able to create a prototype contact lens design it was unable to work correctly when it came into contact with human tears).
Google now hopes to work with hopistals to allow more patients to access the tool.
Verily collaborated with the NHS last year on a project to analyse and predict chronic conditions from anonymised patient data. Hopefully this existing relationship could see the NHS adopting this type of cuttin g edge AI tech to improve treatment for cancer patients.