This AI system outperforms humans in detecting cancer tumours from CT scans

Last Updated: Sat, Aug 25, 2018 18:51 hrs
Scan Machine in Hospital (Image Courtesy: Wikipedia)

An image of a large CT scan system. Image Courtesy: Wikimedia-Commons

Scientists have developed an artificial intelligence system that can accurately detect tiny specks of lung cancer in CT scans, which radiologists often have a difficult time identifying.

The AI system is about 95% accurate, compared to 65% when done by human eyes, researchers were quoted as saying.

"We used the brain as a model to create our system," said Rodney LaLonde, a doctoral candidate at University of Central Florida in the US.

The approach is similar to the algorithms that facial-recognition software uses. It scans thousands of faces looking for a particular pattern to find its match.

The group fed more than 1,000 CT scans into the software they developed to help the computer learn to look for the tumours.

They had to teach the computer different things to help it learn properly. The system was taught to ignore other tissue, nerves and other masses it encountered in the CT scans and analyse lung tissues.

Researchers are fine-tuning the AI's ability to identify cancerous versus benign tumours.