Precision Medicine for AD

The flooding of the dermatology market with safe, effective systemic therapies for various skin conditions has led many to speculate that the next frontier is not necessarily a new drug class, but rather precision medicine tools that optimize the selection of available systemics. That frontier seems closer than ever after Castle Biosciences announced the launch of AdvanceAD-Tx™, a gene expression profile (GEP) test designed to guide systemic treatment decision making in patients aged 12 and older with moderate-to-severe atopic dermatitis (AD). Practical Dermatology sat down with Derek Maetzold, CEO of Castle Biosciences, at the 2025 Fall Clinical Dermatology Conference to learn more about the test and what it could mean for practitioners and patients (Editor’s note: This conversation has been edited for brevity and clarity).

WHAT IS MOST NOTEWORTHY ABOUT THE DATA THAT CASTLE PRESENTED AT THIS CONFERENCE?

What we have here is the first unveiling of the topline results of our large, multicenter, prospective validation study. We performed this study across 49 clinical sites in the US, and the objective was to address the complex biological diversity that drives symptoms of AD. We know that many patients develop AD through a dysregulation of the main Th2 immune pathway system, but the biology is far more complex than a single pathway. Past attempts to identify biomarkers in AD typically focused on one or two genes or proteins, but none reliably differentiated treatment response. Our objective was to go back to the literature, be thoughtful about it, and instead of trying to narrow the biology and get to a small number of genes, let’s just see what the AD tells us.

What we ended up with is a 487-gene test, where we are measuring the expression levels of 487 genes through a technique called RNA Seq. We map them to 12 immune inflammatory or skin barrier pathways. Some genes occur in multiple pathways and others are unique to one. When we analyzed the data, two reproducible biological profiles emerged. Most patients largely have their symptoms driven by a Th2 molecular profile or phenotype, but approximately three out of 10 in our study have what we are calling a JAK inhibitor responder profile. That is interesting because, in hindsight, the literature indicates approximately 65% of patients who start on Th2¹ targeted therapy maintain that monotherapy over time, so our 70/30 split closely mirrors what real-world data show.

What does that mean for patient care? In this initial study, we demonstrated that regardless of whether a patient with a Th2 molecular phenotype starts on a Th2-targeted therapy drug or on a JAK inhibitor therapy, the patient shows similar outcomes. However, the other 30% is where it is quite an interesting story. If you have a patient whom we would classify as a JAK inhibitor responder phenotype, they have a substantially higher likelihood of achieving Eczema Area and Severity Index (EASI) 90 on a JAK inhibitor than on a Th2-targeted drug. It is a dramatic differentiation: a five- or six-fold difference in attaining EASI 90, and similar differences with investigator global assessment (IGA), patient-reported itch, and lack of flares. No matter which efficacy end point one looks at, there is a strong differentiation.

Practically speaking, if you have a patient with moderate-to-severe disease, and you are ready to take that leap to a systemic therapy, our test can provide a result back saying this patient fits a Th2 molecular phenotype or a JAK inhibitor phenotype. With a Th2 phenotype, the efficacy end points are likely to be similar for either type of drug, so you can base your therapy on the patient’s preferences involving injections, routine blood monitoring, etc. With a JAK inhibitor responder profile, however, if you prescribe a Th2-targeted therapy drug, the patient will likely have a poorer response than if you put them on a JAK inhibitor. Learning that ahead of time is the benefit of bringing personalized care to AD.

WHY DID YOU SET THE TARGET AT EASI 90 INSTEAD OF EASI 75?

The FDA utilizes EASI 75 as a primary regulatory threshold for evaluating drug efficacy in AD. If we set that target for our test, there would not be much clinical use for it. If we were going to invest in developing something that would help you make an improved objective treatment decision in AD, we wanted to set the bar higher, and the clinician feedback indicated that if we get 90% clearance, consistent with the other measurement scales, then that would be a home run to show differentiation. Based on the data from our validation study, if you test three patients, you would likely make a change in your first-line therapy for one of them. That is a fantastic outcome for patient care.

HOW DO YOU ENVISION THIS TEST BEING PAID FOR, AND HOW DID YOU NEED TO APPROACH THIS DIFFERENTLY THAN WITH YOUR SKIN CANCER TESTS?

Our skin cancer tests use biopsies, and we know that process quite well. However, AD does not involve biopsy as part of a routine clinical workup, so we developed a non-invasive tissue collection process that involves light scraping with the blunt side of a curette. Since this is a slight change in patient flow, we will have a limited access launch to understand that and publish the data in a peer-reviewed journal. At that point, our expectation is to go to commercial insurers and Medicare along with Medicaid on a state-by-state basis and tell a great economic story in terms of helping to save costs. Starting doses of systemic therapies involve frontloading cost expectations. Initial failure to control the disease can also lead to other costs as patients are driven to increased healthcare visits. It’s an overall increase in healthcare system utilization. We have a very nice model showing that, at the appropriate reimbursement rate, we have a healthcare extraction. Systemic therapies are expensive because they work and they’re newer. If we can help people avoid going through two or three of these in the course of a year, and instead find that the third one should be their first choice, that should save healthcare systems money.

COULD THIS PROCESS WORK FOR OTHER DISEASE STATES SUCH AS PSORIASIS?

We believe the skin scraping technique will create opportunities in other disease states comfortably. We did study psoriasis, and between the complexity of what we were seeing from a biological signal standpoint, plus the very, very high efficacy of all the newer therapies, we did not feel that we could offer very much differentiation at this point, though that could change based on different therapeutic classes coming out that have different risk-benefit profiles. Still, a whole variety of other indications in dermatology could be natural next steps, especially where JAK inhibitors are being used.

ARE THERE ANY PLANS TO GET MORE GRANULAR IN THE TESTING?

Yes. In the Th2-targeted therapies available today and even within the two currently approved systemic JAK therapies as well as some coming in the pipeline, there are enough differing mechanisms of action to make that worth exploring. Our initial study was not powered for that, but our next set of studies certainly will be powered to look at a broader sample set, focusing on AD only. One hypothesis would be that, by having these 12 immune barrier inflammatory pathways covered, we should be able to get there. Conversely, I’m not sure we need to get there. Determining that would be one of the next steps. We’ll have to see how the data pans out.

COULD AN IN-OFFICE TESTING MECHANISM BE IN THE FUTURE TO SAVE TURNAROUND TIME BY CUTTING OUT THE LAB?

That’s a good question. The instrumentation is likely too expensive for most practices, but if we see future consolidation of private equity groups and bringing dermatopathology in-house completely, some might make that investment.

WHAT ELSE COULD BE NEXT?

When we started out with cutaneous melanoma, we were not even thinking beyond skin cancer, but eventually could not see anything else in skin cancer to pursue. Our customers asked us to look at psoriasis and AD as new areas for precision medicine. More recently, as we saw the data from this test, our question was whether we should be thinking beyond gene expression profiling. We began a collaboration with a company that has an electrical impedance spectroscopy (EIS) device system used to obtain many data points. We were particularly interested in this technology for AD. What if we can use this handheld device to predict inflammatory or immune changes, or skin barrier function? What if EIS technology could predict a flare based on changes in the skin structure to give a patient lead time to apply their rescue medications prophylactically? Could that stop the flare from ever happening, or take it from a clinically significant flare down to just a bothersome nuisance flare? Some early studies conducted in Europe indicate that the signals are there. The question is, can you get there with a reliable tool? If that’s the case, then we would have two solutions for AD: one to help you and your patients make a better choice when they are ready for a systemic, and one for people with mild disease to use at home to better control it. That is exciting.

DEREK MAETZOLD

President and CEO of Castle Biosciences, Inc.