Researchers at Vanderbilt University have developed a machine learning algorithm that performs as well as human at identifying skin lesions in clinical photographs of people with monkeypox, Healthcare Purchasing News September 27 reports. The report appeared in the Journal of Investigative Dermatology on September 15.
The severity of an individual’s monkeypox can be identified by the number of skin lesions present, the primary measurement the World Health Organization uses. With no current drug treatment for monkeypox, identifying the number of lesions can be used to assess a patients response in drug trials.
This development may be a gamechanger for the efficiency of daily lesion counting, Eric Tkaczyk, MD, PhD, assistant professor of Dermatology, staff physician at the Department of Veterans Affairs and director of the Vanderbilt Dermatology Translational Research Clinic, explained. “While our proof-of-concept study uses limited data, our results clearly demonstrate that an AI solution to speed intensive assessment of monkeypox severity is within reach.”
The project used 66 photographs of 18 patients from an observational study in the Democratic Republic of Congo. In each image, a member marked borders around lesions, providing training data for machine learning. They found that the lesion counts by the machine matched with counts by two human raters.Read More >>