Noninvasive Skin Protein Assessment to Guide Targeted Treatment Selection in Inflammatory Skin Diseases

Using DIPS, a non-invasive skin sampling technique, to build cytokine-based biomarker panels that predict biologic treatment response in psoriasis and atopic dermatitis.
Completed Allergic & Atopic Diseases Dermatologic Diseases Psoriasis Biological & Mechanistic Biomarker Discovery Cytokine Signaling Data-Driven & Quantitative Disease Subtyping Experimental Platforms & Models Human Cohorts Immune Profiling Machine Learning & AI Precision Medicine Translational & Clinical

Funding Year: 2023

Project Overview

Psoriasis and atopic dermatitis are common inflammatory skin diseases where biologics work well for some patients but not others, likely due to individual variation in cytokine signaling. While skin biopsies have shown promise for predicting treatment response, they are invasive and impractical at scale. This project applies DIPS (Detergent-based Immune Profiling of Skin) — a non-invasive technique that painlessly collects immune proteins from the skin surface — to identify cytokine biomarker panels that predict biologic responsiveness. Early results include a 5-biomarker panel distinguishing PS from AD with 99.8% ROC-AUC accuracy, and distinct panels predicting responsiveness to apremilast and dupilumab, respectively. The team is working with Yale Ventures to develop new IP and explore commercialization.

Impact & Innovation

Predicting biologic response from a painless skin swab.

 

DIPS enables rapid, non-invasive collection of skin immune proteins at scale — and early results show it can distinguish disease subtypes and predict treatment responsiveness with near-perfect accuracy, pointing toward a point-of-care diagnostic platform.

  • Demonstrates that non-invasive cytokine profiling via DIPS can predict responsiveness to apremilast and dupilumab, addressing the core clinical challenge of biologic treatment selection in PS and AD
  • Generates strong IP and commercial potential through Yale Ventures collaboration, with a company in development to license and scale DIPS for clinical use across dermatology and beyond
  • Advances the Consortium’s From Mechanistic Insight to Translation pillar by converting skin immune profiling discoveries into a scalable, point-of-care diagnostic tool for precision treatment selection in inflammatory skin diseases
Research Approach

A framework designed for discovery

This project applies DIPS-based non-invasive skin immune protein collection with machine learning-driven biomarker analysis to identify cytokine panels that distinguish inflammatory skin disease subtypes and predict biologic treatment response in psoriasis and atopic dermatitis.

Non-invasive skin sampling using DIPS (small probe and surfactants) to painlessly collect immune proteins from patient skin; protein-level cytokine analysis of collected samples; machine learning approaches to develop and validate 5-biomarker panels distinguishing PS from AD and predicting responsiveness to apremilast and dupilumab; and advanced data analysis for biomarker identification and method refinement.

Non-invasive skin protein samples from PS and AD patients, cytokine profiling datasets measuring immune marker abundance at the protein level, machine learning training and validation datasets for biomarker panel development, and prior skin biopsy datasets used to validate DIPS against established methods.

Development and validation of DIPS-based cytokine biomarker panels for disease subclassification and biologic treatment prediction in PS and AD, with a translational goal of commercializing DIPS as a point-of-care diagnostic tool. A company is being developed with Yale Ventures to license and scale the technology, enabling healthcare providers to collect and analyze samples rapidly and deliver personalized treatment guidance.

Investigators & Institutions

Powering the science

Principal Investigators

Jeffrey M. Cohen, MD, MPH, Colton Consortium Member

Associate Professor, Department of Dermatology, Yale School of Medicine, Yale University

William Damsky, MD, PhD, Colton Consortium Member

Associate Professor, Department of Dermatology (Dermatopathology), Yale School of Medicine, Yale University

Colton Centers & Institutions

Colton Center for Autoimmunity at Yale University

Research Outputs

From insight to impact

Murphy MJ, Hwang E, Singh K, Lee T, Cohen JM, Damsky W. Machine learning analysis of pretreatment skin biopsies predicts nonresponse to dupilumab in patients with eczematous dermatitis. Br J Dermatol. 2023 Dec 20;190(1):132–134. doi: 10.1093/bjd/ljad389. PMID: 37818837.

Damsky and Wang et al. Non-invasive epidermal proteome assessment-based diagnosis and molecular subclassification of psoriasis and eczematous dermatitis. Manuscript in preparation.

Yale Ventures IP development in progress.