A Google artificial intelligence system was better than 6 radiologists at identifying whether patients had lung cancer, according to a new Nature Medicine study reported on by Stat News.
How it worked: When looking at a single CT scan, the algorithm detected 5% more cancers than the humans and reduced false positives by 11%. When prior images of patients were included in the evaluation, the algorithm performed the same as the radiologists.
Why it matters: Screening high-risk patients for lung cancer reduces the risk of death, but false positives can lead to unnecessary and potentially harmful surgeries.
- About half of lung cancers are diagnosed after they have already spread.
- "These people have a technology that will improve the precision of screening tremendously," Otis Brawley, an oncology professor at Johns Hopkins University, told Stat.
What next? The system will now be subject to further testing. Google acknowledged this, and said that it's already working with clinical partners and has had pre-submission discussions with the FDA.
Go deeper: Scientists call for rules on evaluating predictive AI in medicine