COVID VaccinesReturn to WorkMental Health

AI Is Better at Diagnosing Skin Cancer Than Your Doctor, Study Finds

May 30, 2018, 6:20 PM UTC

A new study has found that artificial intelligence is better than humans at detecting skin cancer.

The study, published in the journal Annals of Oncology, put dermatologists against a computer that had been trained to tell the difference between cancerous skin lesions and benign ones. In the end, the dermatologists were only 86.6% accurate at diagnosing skin cancer, while the computer was able to diagnose issues with a 95% accuracy.

The 58 dermatologists involved in the study came from a variety of backgrounds and 17 countries. A little more than half were considered “expert” level with more than five years experience, while 19% had between two and five years of experience, and 29% had less than two years of work experience.

Male patient being examined in clinic
Male patient being examined in clinic, image of man on monitor in foreground. Total body photography is used to track potential changes in moles and lesions for early detection of skin cancer.Science Photo Library Getty Images/Science Photo Library RF
Science Photo Library/Getty Images

The computer was also less likely to diagnose a benign mole as cancer, something that would result in a patient undergoing unnecessary stress and surgery.

The hope is that the technology could eventually be implemented as a way to diagnose skin cancer in its early days before it spreads.