Melanoma is a very severe cancer that is often diagnosed too late to save patients’ lives, and most people do not regularly visit a dermatologist for skin exams. Early-stage identification of suspicious pigmented lesions (SPLs), ideally by primary care providers, could lead to improved melanoma prognosis. Researchers at the Wyss Institute and MIT have developed a deep learning algorithm that can differentiate SPLs from non-cancerous skin marks in wide-field images of patients’ skin, such as those taken with a tablet or cellphone. This method could enable improved patient triaging, reduced medical costs, and earlier treatment of melanoma.
Credit: Wyss Institute at Harvard University