AI can now assess skin cancer as 'accurately as dermatologists'

The digital doctor can examine the severity of a case of skin melanoma

Artificial intelligence (AI) can now assess the severity of skin cancer as 'accurately as dermatologists', claims a new study from University of Gothenburg.

Researchers based in Sweden have developed an algorithm that determines whether a skin melanoma is invasive and at risk of spreading, or at its earliest stage and confined to the epidermis.

The technology was trialled on more than 900 dermoscopic images of melanoma before being tested on 200 images of skin cancer melanoma that had been diagnosed by a dermatopathologist.

According to the team, the algorithm also scored the same result as seven independent dermatologists when classifying the severity of a melanoma.

"None of the dermatologists significantly outperformed the ML algorithm," said Sam Polesie, a researcher at the University of Gothenburg and specialist doctor at Sahlgrenska University Hospital.

The scientists predict that the algorithm could support surgeons in terms of how extensive an operation to remove the cancer needs to be.

"The results of the study are interesting, and the hope is that the algorithm can be used as clinical decision support in the future.

"But it needs refining further, and prospective studies that monitor patients over time are necessary, too."