Using AI as Augmented Intelligence to Improve Rare Dermatologic Skin Diseases

This content was produced independently by The American Journal of Managed Care^® (AJMC^®) and is not endorsed by the American Academy of Dermatology.

Steven Daniel Daveluy, MD, FAAD, associate professor, program director, and clinical educator at Wayne State University, discussed the potential of artificial intelligence (AI) in teledermatology, particularly for rare diseases like generalized pustular psoriasis and hidradenitis suppurativa, at the recent American Academy of Dermatology annual conference.

He highlighted the use of large data registries like DataDerm to train AI in identifying patterns and improving early diagnosis by analyzing patient journeys and various medical records. Daveluy emphasized "augmented intelligence," noting that AI guided by expert dermatologists can expedite data analysis beyond human capabilities, ultimately aiming to educate clinicians and facilitate earlier interventions for patients with rare dermatological conditions.

This transcript was lightly edited for clarity; captions were auto-generated.

Transcript

What are some potential benefits of AI playing a role in teledermatology for rare diseases? What are some limitations?

We're seeing some real opportunities for AI to help us in dermatology. There's a lot of big data out there, and if we can access that data and really tease out what it can teach us, we can learn a lot and really help. Two areas that we've really dove into this are generalized pustular psoriasis, and we're starting with hidradenitis suppurativa, where we can use our DataDerm data registry, which has tons of information from over 1 million patients.

We can help AI to look into their records and see what was their journey like before diagnosis. Where could we have made diagnosis earlier? How can we help these patients get an earlier diagnosis and start treatment sooner? AI is really well equipped, because when you have that much data, no human being can look at it as quickly, but we really like to call it augmented intelligence, because it's not just working on its own, we have expert dermatologists, who are helping guide the process and say, does that make sense what the AI is thinking, or do we need to dive into that a little bit more?

Looking ahead 5 to 10 years, how do you envision AI transforming the landscape of rare disease diagnosis and treatment in dermatology?

I'm really optimistic that AI is going to help us to get rare diseases diagnosed earlier. We can look back not only at patients' dermatology encounters, but our partner, OM1 who's helping us with this AI technology, can look back at other notes that patients have had from their primary care, the emergency room, so we can really see what it was like before they even came to our office and hopefully find some opportunities to then have educational efforts to help all of our colleagues and our dermatologists to know how to find these patients sooner, get the right diagnosis, and get treatment started earlier.