Apps to detect possible melanomas fine

Many applications have emerged in recent months to help detect skin cancers photographing, for example, a mole suspect. But U.S. researchers reviewed these apps, and their results are not encouraging.

Researchers at the University of Pittsburgh Medical Center have loaded 188 images of skin lesions in four different applications, based mainly on algorithms to detect, often within a few seconds, if the lesion was benign or whether it should move in dermatologist.

But three out of four applications have failed to identify at least three cases of melanoma. In addition, they also falsely identified many benign lumps as potentially cancerous. The most effective service based on four college certified dermatologists giving their opinions on the images, and within 24 hours.

Scientists have not given the names of the applications they tested. The results of this study are published in the January 16 edition of the review JAMA Dermatology.

Rather than leave it to their smartphone to diagnose dangerousness of his beauty spots, experts suggest to use only applications to facilitate the process of self-examination, such as sending reminders when it's time return to his dermatologist. UMSkinCheck free app of the University of Michigan is one of those services. It also proposes a model for organizing photos taken over time showing the evolution of his moles to present to their physician.