JACKSONVILLE, Fla. — Between the sand and the sun, skin problems can develop fast, especially in the Florida elements.
“Do I have a rash on my neck because I’ve been wearing a nickel necklace? Or am I allergic to a cosmetic?” Dr. Alison Bruce, a dermatologist and department chair with Mayo Clinic Florida, said.
>>> STREAM ACTION NEWS JAX LIVE <<<
Those are questions that Bruce said she commonly hears from her patients.
“About one in five people may suffer from allergic contact dermatitis,” she said.
Traditional patch testing helps pinpoint what’s triggering those reactions -- a red, itchy, eczema-like rash.
But that kind of testing can be time-consuming for both patients and providers, and could lead to multiple, sometimes pricey, visits to the doctor.
“It’s also sometimes very hard for patients to get access to a dermatologist,” Bruce said.
Bruce and her team explored options, asking “Could artificial intelligence help do what doctors normally do in person?” by reading photos of skin reactions.
“Can we take photographs … and then can AI interpret those photographs … determining is it a reaction, yes or no … and … is there a way to do this remotely through photographs?” she said.
To teach AI what a reaction looks like, Mayo researchers fed it data -- a lot of data.
About 28,000 patch test images collected over the last decade through their system, then compared the AI’s readings to dermatologists.
“We found that the AI was pretty much as good as the dermatologist in interpreting the photographic images,” Bruce said.
But building trustworthy AI in medicine means making sure it works for everyone.
That includes training the system to recognize reactions across all skin tones.
“If you’re gonna train AI accurately, you have to have good representation of all skin types. Because you have bear in mind a red rash on a dark black skin is gonna look very different than on a very pale white skin,” she said.
Action News Jax’s Chandler Morgan sat down with some of the trial participants.
They said it was the appeal of the AI technology itself that drew them in.
“Having AI as a potential to read these results … I think that’s really cool and innovative,” Natalie Thackerdin, a pharmacist and trial participant, said.
During their testing phase, patients still came into the clinic, but the goal is to eventually let people do much of this from home, taking photos with their phones for the AI to help interpret.
[DOWNLOAD: Free Action News Jax app for alerts as news breaks]
“So it was pretty convenient, pretty straightforward. It didn’t take more than a few minutes for him to basically read it using the AI algorithm,” Guillermo Pradieu, GI Researcher & Trial Participant, said.
Researchers are now training the AI to handle real-world challenges -- different phones, lighting, angles.
Long-term they want to develop a patient-friendly patch kit, a secure cloud system, and eventually FDA clearance.
[SIGN UP: Action News Jax Daily Headlines Newsletter]
Both the trial participants and Bruce agree that this kind of AI could help unlock easier access to healthcare.
“I think if we frame AI as something that can be used as a supplementary to aid the tools that we already have, it’ll make it a lot easier for patients to accept and be willing to use it. Because it’ll be on their phone, and that AI will be the connection between the provider and the patient,” Pradieu said.
“We hope to reduce the burden on patients so that they don’t have to come in but instead be able to potentially do that remotely through photography and AI helping to interpret,” Bruce said.
Click here to download the free Action News Jax news and weather apps, click here to download the Action News Jax Now app for your smart TV and click here to stream Action News Jax live.
