Detection of skin cancer with artificial intelligence

European and American researchers have developed a system of artificial intelligence, which seems to be even better than dermatologists in diagnosing the most aggressive and fatal skin cancer, melanoma. It is a deep learning convolutional neural network, trained to detect cancer, previously fed with more than 100,000 images of malignant melanomas, as well as – for comparison purposes – images of harmless moles.

Scientists from Germany, France and the United States, headed by Professor Holger Hensle of the Department of Dermatology at the University of Heidelberg, published the publication in the Annals of Oncology journal of the European Society of Oncology.

The diagnostic credibility of the artificial intelligence system was compared to that of 58 dermatologists from 17 countries, including Greece. It was found that the system diagnosed more melanomas (95%) than doctors (89%), and had fewer false positive diagnoses (innocuous naïves that were misdiagnosed as cancers).

The system, which mimics the functioning of the human brain, learns to get better as many images see.

Melanoma incidents are constantly increasing, with about 232,000 new diagnoses and 55,500 deaths each year worldwide. Melanoma can be cured if diagnosed early, but unfortunately often the diagnosis is delayed when the cancer progresses.

The researchers said they did not foresee artificial intelligence to replace dermatologists, but would help them make better diagnoses in the future.
However, before such an artificial intelligence system finds broad clinical application, some technical problems, such as the difficulty of correctly depicting certain melanomas in areas such as the fingers and toes or the skull, system to “read” the images.