Background: Current atopic dermatitis (AD) therapy selection does not account for molecular drivers of disease, contributing to inadequate disease control for half of patients.
Objective: Develop and validate a gene expression profile (GEP) test to guide AD systemic therapy selection.
Methods: In a prospective study, AD or psoriasis lesion skin scrapings were analyzed by RNA-Seq, and treatments and outcomes including Eczema Area and Severity Index, itch, and flare frequency were reported. An ensemble algorithm (12 neural networks with 487 genes, 487-GEP) was developed to classify the probability of AD or psoriasis treatment response (n = 192) and independently validated in AD patients (n = 110).
Results: AD patients classified by the 487-GEP as JAK inhibitor (JAKi) Responder Profile (30.4%) and treated with a JAKi achieved significantly higher rates of Eczema Area and Severity Index-90 (45.5% vs 8.3%, P = .021) 3.8 times faster (P = .049), reported "no itch" (45.5% vs 8.3%, P = .021), and increased flare-free rate (54.5% vs 16.7%, P = .041) versus Th2-targeted therapy. For those with a Th2 Molecular Profile, there were no statistical differences in Eczema Area and Severity Index-90 whether they used Th2-targeted (26.5%) or JAKi (33.3%, P = .728).
Limitations: Validation was limited to patients ≥12 years.
Conclusions: The 487-GEP identifies AD patients more likely to benefit from JAKi than Th2-targeted therapies.
Keywords: 487-gene expression profile; JAK inhibitor; JAKi Responder Profile; Th2 Molecular Profile; Th2-targeted therapy; atopic dermatitis; inflammation; systemic therapy.
Copyright © 2026 The Author(s). Published by Elsevier Inc. All rights reserved.