Soon you could be choosing a computer over a doctor when it comes to a cancer diagnosis.
According to a new study, an artificial intelligent (AI) system outperformed dermatologists when it came to diagnosing skin cancer.
The computer, a deep learning convolutional neural network (CNN), was trained by a team from Germany, France, and the US, by looking at over 100,000 images of cancerous moles and benign spots.
After its training, the scientists put the computer to work by pitting it against 58 dermatologists from 17 countries around the world.
After being shown images of different types of moles, the CNN was able to accurately detect skin cancer in 95 per cent of the images.
The human dermatologists, on the other hand, accurately detected cancer 86.6 per cent of the time.
“The CNN missed fewer melanomas, meaning it had a higher sensitivity than the dermatologists,” said the study’s first author, Holger Haenssle from Germany’s University of Heidelberg, in a statement.
This meant the computer also misdiagnosed fewer benign moles as malignant. This could result in less unnecessary surgery, according to Haenssle.
In the future, the AI could be used as a tool for faster and easier diagnosis of skin cancer.
The use of AI in health care, whether it's cancer diagnoses or in eye disease research, is still a relatively new practice. Companies like IBM and Microsoft have been using the technology to improve cancer diagnoses and treatment plans for the past few years, as well as charities such as Cancer Research.
However, there is more of an emphasis on using this technology in the UK. Earlier this month, UK prime minister Theresa May pledged millions in government funding for research into using AI to fight cancer.
May said AI could be used to help stop 22,000 cancer deaths a year by 2033.
We are already using AI in our daily lives, whether it’s to help us save money, improve our holidays, and make our jobs easier, so it makes sense then that this transformative technology is put to work to prevent cancer.
