Speaker: Artificial intelligence may be future of skin cancer detection
Click Here to Manage Email Alerts
MIAMI BEACH, Fla. — Artificial intelligence should be thoughtfully evaluated for the use of skin cancer detection, according to a presenter at South Beach Symposium.
“There is an explosion of new technology out there and I think what we’re learning now is there are lots of ways to actually have dermatologist-like sensitivity and specificity in AI,” Laura K. Ferris, MD, PhD, professor of dermatology at the University of Pittsburgh, told Healio.
Artificial intelligence should be thoughtfully evaluated for the use of skin cancer detection.
During her presentation, Ferris discussed a number of past and current devices using artificial intelligence (AI) in skin cancer detection, including the now discontinued MelaFind (MELA Sciences), which showed that AI devices often compromise specificity for sensitivity. In contrast, Ferris also highlighted a device currently being used in the United Kingdom in primary care settings that has demonstrated strong detection of not just melanoma but all skin cancers by utilizing Deep Ensemble for the Recognition of Malignancy (Skin Analytics), an AI algorithm that determines the likelihood of skin cancer from evaluating dermatoscopic images of pigmented skin lesions. Due to the device’s ability to detect multiple skin cancers, Ferris describes it as “a next generation tool.”
As these technologies advance, Ferris said dermatologists should question whether these devices can be validated in all circumstances.
“We ought to think about how we’re going to use these devices, how they’re going to best fit in our practice and how we make sure that they're inclusive of all of our patients,” Ferris said during her presentation.
Using AI devices in skin cancer detection is an “evolving conversation with the FDA,” Ferris told Healio, and a balance of maximizing safety with maximizing availability.
“We want to make sure that these devices are safe before we put them in the hands of patients or doctors,” Ferris said, “but we also want to make sure as they improve that we can update devices and not have to go through an entire new clinical trial, validation process.”