An artificial intelligence system has beaten dermatologists for the first time at identifying malignant melanomas, it was announced last night.
A deep learning convolutional neural network was trained by being more than 100,000 images of melanomas and naevi in a project involving researchers in Germany, France and the USA.
According to a report in the Annals of Oncology, the CNN machine then performed better than 58 dermatologists from 78 countries.
Out of 100 cases, dermatologists identified 86.6% of melanomas and 71.3% of non-malignant lesions.
The researchers said they “tuned” the CNN to identify 71.3% of non-malignant lesions – and at this level it identified 95% of melanomas.
In a second round, dermatologists were given more information and identified 88.9% of melanomas – still failing to out-perform the machine.
Researcher Professor Holger Haenssle, senior managing physician at the Department of Dermatology, University of Heidelberg, Germany, said: "These findings show that deep learning convolutional neural networks are capable of out-performing dermatologists, including extensively trained experts, in the task of detecting melanomas.”
He added: "This CNN may serve physicians involved in skin cancer screening as an aid in their decision whether to biopsy a lesion or not. Most dermatologists already use digital dermoscopy systems to image and store lesions for documentation and follow-up.
“The CNN can then easily and rapidly evaluate the stored image for an ‘expert opinion’ on the probability of melanoma.”
Writing in the journal, two Australian dermatologists Dr Victoria Mar and Professor Peter Soyer state: "Currently, there is no substitute for a thorough clinical examination. However, 2D and 3D total body photography is able to capture about 90 to 95% of the skin surface and given exponential development of imaging technology we envisage that sooner than later, automated diagnosis will change the diagnostic paradigm in dermatology.
“Still, there is much more work to be done to implement this exciting technology safely into routine clinical care."
H.A. Haenssle et al. Man against machine: diagnostic performance of a deep learning convolutional neural network for dermoscopic melanoma recognition in comparison to 58 dermatologists. Annals of Oncology 28 May 2018; doi:10.1093/annonc/mdy166
Leave a Reply