The 487-gene expression profile test guides systemic therapy selection to improve outcomes for patients with atopic dermatitis: Results from a prospective, multi-center trial

Abstract

Introduction: Atopic dermatitis (AD) is an inflammatory skin disease driven by complex biological interplay between type 1, 2, and 17 inflammatory processes. Systemic therapies, including biologics that target T helper type 2 inflammatory pathway (Th2-targeted therapies) and small molecules targeting multiple inflammatory pathways (JAK inhibitors, JAKi), have improved care for AD. However, current therapy selection is trial and error without consideration of each patient’s underlying biology, resulting in inadequate disease control for approximately 45% of patients taking systemic therapies for AD. An objective test to measure each patient’s underlying disease biology to guide therapy selection with the goal of reducing the trial-and-error phase of therapy selection should lead to earlier response to treatment and an improved proportion of patients who achieve adequate disease control. The objective of this study was to develop and validate a non-invasive gene expression profile (GEP) test to guide systemic therapy selection for AD to improve health outcomes.

Methods: A prospective observational study enrolled patients with AD or psoriasis. Samples were acquired via non-invasive skin scraping from affected skin lesions and analyzed by RNA sequencing. Therapy use, eczema area and severity index (EASI) outcomes, peak pruritus in the last week on a 0-10 numeric rating scale (PP-NRS) and flare frequency data were collected. A neural network ensemble algorithm (487-GEP test) was developed using 12 inflammatory and cutaneous biology pathways containing a total of 487 genes. A training set of n=192 patients was used to create the 487-GEP, which identified patients with a Th2 Molecular Profile or a JAKi Responder Profile. The test was then validated in an independent validation cohort of patients age 12 and older receiving Th2-targeted therapies or JAKi for AD (n=113 patients).

Results: The 487-GEP test scored 29.4% of AD samples as having a JAKi Responder Profile. This subset of patients achieved significantly higher rates of EASI90 (50% vs. 8%, p=0.012), achieved EASI90 2.3 times faster (p=0.008), were more likely to have PP-NRS 0 or 1 (little to no itch) (40% vs. 8%, p=0.043) and were more likely to be flare-free (60% vs. 16%, p=0.016,) within 3 months when treated with a JAKi at low dose versus a Th2-targeted therapy. In contrast, patients with a Th2 Molecular Profile had similar responses to Th2-targeted therapies and JAKi within 3 months.

Limitations: The study included patients treated with JAKi as first line therapy as well as patients who switched to JAKi as a second line treatment. However, this is representative of clinical practice in which patients frequently switch medications with the goal of disease control. The 487-GEP test is predictive of response to JAKi as a class rather than predictive of response to specific therapies; however, this provides clinicians and patients with therapy guidance that also allows choice of specific therapy within the JAKi class.

Conclusions: In patients with AD, the 487-GEP test is predictive of rapid and improved response to systemic JAKi, thereby identifying a clinically meaningful group of patients with AD who are significantly more likely to benefit from JAKi than Th2-targeted therapies.