Dr Joseph Zabinski Discusses Challenges in Identifying Patients With GPP


Joseph Zabinski, PhD, MEM, vice president, head of commercial strategy and AI, OM1, addresses challenges in identifying patients with generalized pustular psoriasis (GPP) and how artificial intelligence (AI) can help alleviate these concerns.

Artificial intelligence (AI) has the capability to close gaps in care by reducing delays in diagnoses among patients with generalized pustular psoriasis (GPP), according to Joseph Zabinski, PhD, MEM, vice president, head of commercial strategy and AI, OM1.


Are there any specific challenges when identifying patients with GPP, considering the diverse presentation and severity of symptoms often seen in this patient population?

Yeah, the challenges with GPP; there are several. One is the similarity of the disease in some of its manifestations to other, often more common dermatologic conditions. As you mentioned, there is heterogeneity in the presentation and manifestation. GPP can manifest in these flare eruptions, but what triggers those can be different among different patients. Sometimes there's not an obvious trigger at all. And all that heterogeneity makes the picture cloudier when you look at it from a human eye or a human lens.

The nice thing with AI is it's able to sort through some of the noise in those data and say, "Under the surface, there are still common patterns that are shared amongst these patients." And they look different than they do in hidradenitis suppurativa (HS) or in other dermatologic conditions, where we might also be asking questions about diagnosis. And I have found in practice, that the AI systems are not super confused when it comes to GPP. They're able to isolate and elevate those at much greater likelihood in GPP quite well. So, it's quite a good application, I think.

Can you discuss the potential long-term implications and benefits of integrating AI technology into clinical practice for identifying patients with GPP?

As I was describing, right now, [in] this state of the world, there's often a significant, years-long delay between patients' first manifestation of symptoms and when they get a correct diagnosis. It's only that diagnosis that will let them have access to effective treatment. If the doctor doesn't know that they have this condition, they're not going to be able to connect the patient to the appropriate treatment.

What I would love to see, I don't think [in] the immediate future, but perhaps in the medium-term future, is exactly as you said: the integration of tools like our patient finder, our version of this with AI, to read patient data and highlight those at risk, so that they can get diagnosed and treated much more quickly. The reason integrating this is important is because we don't want AI tools to blow up workflows and blow up how clinicians interact with patients. We want them to be seamlessly running behind the scenes, in a regulated and careful way, that just helps us to close gaps around who is experiencing what and who is at elevated risk of what. The ideal state there is to say as soon as we have some good evidence that a patient is sort of bending in an arc towards GPP. They can get the care that they need and get diagnosed quickly [but] doesn't mean that they're magically going to get diagnosed immediately. But if we can reduce that years-long delay in diagnosis substantially, if we can reduce the odysseys patients go through, misdiagnosis into other conditions, that can make quite a dramatic difference. Having this kind of technology integrated into EMRs [emergency medical records] [that’s] familiar to providers is the way to make that happen.

Related Videos
Monica Kraft, MD, ATSF.
Michael Arzt, MD.
Ravin Ratan, MD, MEd, MD Anderson
Dr Ajay Goel
Julie Linton, MD, FAAP.
Jan Hedner, MD, PhD.
Dr Migvis Monduy
Kimberly Westrich, MA, chief strategy officer of the National Pharmaceutical Council
Amy Shapiro, MD
Adam Benjafield, PhD.
Related Content
CH LogoCenter for Biosimilars Logo