Google's AI can detect lung cancer better than radiologists

Artificial intelligence is officially joining the fight against cancer.

Google on Monday announced new research in training an AI tool to recognize the signs of lung cancer from a CT scan of a patient's chest. The research, which began in late 2017, has culminated in an AI capable of diagnosing lung cancer with better accuracy than certified radiologists.

In order to test the AI, Google showed it 45,856 chest CT scans, comparing the AI's diagnoses with those of six board-certified radiologists, Engadget reported. Google's AI was able to detect cancer in 5 percent more of these screenings than the radiologists; it also reduced false diagnoses by "more than 11 percent."

The Week

Escape your echo chamber. Get the facts behind the news, plus analysis from multiple perspectives.

SUBSCRIBE & SAVE

Sign up for The Week's Free Newsletters

From our morning news briefing to a weekly Good News Newsletter, get the best of The Week delivered directly to your inbox.

From our morning news briefing to a weekly Good News Newsletter, get the best of The Week delivered directly to your inbox.

Sign up

Radiologists typically have to view up to hundreds of images from a single CT scan in order to make a successful diagnosis for lung cancer; Google's AI is able to generate a three-dimensional image instead of 2-D ones, as well as detecting specific areas of malignant tissue in the lungs, which radiologists are often unable to do from images alone.

This AI modeling technique represents a breakthrough in the ability to diagnose cancers early on. Lung cancer, which accounts for more than 1.7 million deaths every year around the world, is often not caught until later stages, when treatment has a much lower chance of success. Google's technology will have to undergo further research and testing before becoming available as a diagnosis tool, but the company hopes to "make early detection more accessible."