Google trains AI to identify lung cancer
An immunotherapy drug has been tested on lung cancer patients. © Neustockimages/Istock.com
Published Tuesday, May 21, 2019 6:38AM EDT
On Monday, Google announced their progress in training artificial intelligence to detect lung cancer; since the project began in 2017, researchers have developed their AI to both generate lung cancer malignancy predictions as well as identify lung nodules as or more accurately and consistently than U.S. board-certified radiologists.
Resulting in more deaths than breast, prostate, and colorectal cancers combined, lung cancer has become the deadliest cancer across the world according to the World Health Organization. This disease has one of the worst survival rates due to difficulties in detecting malignant tissue until it's evolved into late stage cancer.
Since 2017, Google has been training AI to not only detect cancerous lung nodules, but also to "generate an overall lung cancer prediction." On Monday, the company outlined that their technology can diagnose lung cancer either on par with or better than U.S. board-certified radiologists while also reducing the number of false-positives.
Additionally, Google's AI model can use information from previous scans to further aid in the detection of any tissues growing at suspicious rates -- a factor that's a common indicator that a lung nodule is malignant.
According to the company, "despite the value of lung cancer screenings, only 2-4 percent of eligible patients in the U.S. are screening today." Using AI to more accurately and consistently predict the disease could potentially make screening more accessible across not only in the United States, but across the world.