AI and automation have been touted to help increase efficiency and productivity for a while now, and Google is proving that true with their AI models for lung and breast cancer detection.
Radiologists typically look through hundreds of 2D images from CT scans to determine whether a patient has cancer or not, and malignant nodules can be difficult to spot. Using a single CT scan, an AI model trained to spot lung cancer actually detected five percent more cancer cases and reduced false-positive flags by 11 percent compared to radiologists.
An additional test was done with digital mammographs to see if AI could assist doctors in spotting breast cancer more accurately and earlier. The model was trained and resulted in a 5.7 percent reduction of false positives in the U.S, and a 1.2 percent reduction in the UK. It produced a 9.4 percent reduction in false negatives in the U.S., and a 2.7 percent reduction in the UK.
It’s also important to note that the AI model only had access to the most recent anonymised mammograms while doctors had access to patient histories and previous mammograms. This means that the AI model actually performed better despite having less information to work with.
That’s not the only case of AI being able to detect health issues before human experts. It’s been reported on Wired that a Toronto start-up, BlueDot, actually alerted their clients to a potential outbreak (which eventually turned out to be the 2019-nCoV) on 31 December 2019 by using natural-language processing, machine-learning techniques and big data analytics to sift through news reports and more to predict potential situations.
As the technology evolves even further, AI systems are all but guaranteed to improve their prediction accuracy, resulting in better diagnostics for patients. More information can be found on Google’s website for the tests on lung cancer and breast cancer. More information about BlueDot can be found here.