Computer-assisted diagnosis for skin cancer: have we been outsmarted?

For every melanoma diagnosed, there are 3-40 benign lesions biopsied and many of these are biopsied in primary care rather than in specialist settings. A recent report inĀ Nature outlined a new technology using deep convolution neural networks that was found to have similar diagnostic sensitivity and specificity to dermatologists. This current commentary fromĀ The Lancet discusses the interest in this new technology because of its potential to improve accuracy in skin cancer diagnosis. This technology has still to be validated clinically but has the potential to also be utilised on mobile devices through smart phone applications; potentially improving timely diagnosis. Read the full article here.
Scroll to Top