AI better at detecting skin cancer than docs: Study

PARIS • A computer was better than human dermatologists at detecting skin cancer in a study that pitted human against machine in the quest for better, faster diagnostics, researchers said yesterday.

A team from Germany, the United States and France taught an artificial intelligence system to distinguish dangerous skin lesions from benign ones, showing it more than 100,000 images.

The machine - a deep learning convolutional neural network or CNN - was then tested against 58 dermatologists from 17 countries, shown photos of malignant melanomas and benign moles.

Just over half the dermatologists were at "expert" level with more than five years of experience, while the rest had less than two years to five years experience under their belt.

"Most dermatologists were outperformed by the CNN," the research team wrote in a paper published in the journal Annals of Oncology.

On average, dermatologists accurately detected 86.6 per cent of skin cancers from the images, compared to 95 per cent for the CNN.

"The CNN missed fewer melanomas, meaning it had a higher sensitivity than the dermatologists," the study's first author Holger Haenssle of the University of Heidelberg said.

It also "misdiagnosed fewer benign moles as malignant melanoma... this would result in less unnecessary surgery."

The dermatologists' performance improved when they were given more information on the patients and their skin lesions.

The team said AI may be a useful tool for faster, easier diagnosis of skin cancer, allowing surgical removal before it spreads. They said there are about 232,000 new cases of melanoma and 55,500 deaths in the world a year. But it is unlikely a machine will take over from human doctors entirely, rather acting as an aid.

AGENCE FRANCE-PRESSE

Join ST's Telegram channel and get the latest breaking news delivered to you.

A version of this article appeared in the print edition of The Straits Times on May 30, 2018, with the headline AI better at detecting skin cancer than docs: Study. Subscribe