Each year, there are more than 300,000 new cases of Lyme disease in the United States. Treatments for Lyme disease can be effective, but it depends on how early the disease is detected.

To develop a quick, easy way to detect Lyme disease, scientists have developed deep learning models that can analyze images of rashes to identify the erythema migrans (EM) skin redness associated with acute Lyme disease. They “trained” the deep learning models to discern the appearance of EM using images of non-EM rashes and normal skin available in the public domain and clinical photos of patients with EM. That way, the computer models can accurately pick out EM from other dermatological conditions and normal skin. 

This is a big improvement from the current methods being used to detect Lyme disease. Blood tests detecting the presence of antibodies to Borrelia burgdorferi, the cause of Lyme disease, are often unreliable, and more direct ways of detecting the disease through skin biopsies aren’t readily available to clinicians.

Now that the researchers have shown the potential of their EM rash digital analysis as a prescreening diagnostic tool for Lyme disease, they plan to further test and refine the technology in upcoming studies.