Early Detection and Diagnosis of Autoimmune Diseases Using Foundation AI Models

Applying self-supervised AI to multi-modal electronic health records — integrating clinical notes, labs, and imaging — this project builds scalable diagnostic models to detect autoimmune diseases earlier and more precisely.
Ongoing Other Bioinformatics Data-Driven & Quantitative Disease Subtyping Early Disease Detection Experimental Platforms & Models Human Cohorts Machine Learning & AI Multi-omics Integration Precision Medicine Real-world Evidence Translational & Clinical

Funding Year: 2024

Project Overview

Early and accurate diagnosis of autoimmune diseases remains a significant clinical challenge, in part because relevant patient data is often disorganized and spread across multiple record types. This project addresses that gap by applying self-supervised AI approaches to multi-modal Electronic Health Record data, integrating large language models with advanced image analysis to build a unified foundation model. By combining patient history, laboratory results, clinical notes, and medical imaging into a single analytical framework, the team aims to develop scalable AI tools applicable across a range of autoimmune conditions — improving diagnostic precision and enabling earlier intervention.

Impact & Innovation

AI that sees what clinicians can’t — yet.

 

By integrating patient histories, labs, clinical notes, and imaging into a single foundation model, this project is building AI that can detect autoimmune disease earlier and more precisely across conditions.

  • Addresses the fragmented, multi-modal nature of EHR data that currently limits early and accurate autoimmune diagnosis
  • Designed for scalable, cross-disease application — not a single-condition tool
  • Advances the Consortium’s Integrated Data and Discovery Platforms pillar, building scalable AI tools designed for real-world EHR integration across autoimmune conditions
Research Approach

A framework designed for discovery

This project applies self-supervised machine learning to large-scale, multi-modal electronic health record (EHR) data to build foundation AI models capable of early autoimmune disease detection and diagnosis. The work spans model development, validation, and preparation for clinical integrations.

Self-supervised learning approaches applied to multi-modal EHR data, integrating large language models for clinical text analysis with advanced image analysis techniques to construct a unified diagnostic foundation model.

Multi-modal EHR datasets encompassing patient history, laboratory results, clinical notes, and medical imaging across multiple autoimmune disease types, used to train and validate scalable AI models.

Development of AI models that improve diagnostic precision and enable earlier detection across a range of autoimmune conditions, with a design optimized for direct integration into existing EHR systems to support clinical decision-making and personalized patient care.

Investigators & Institutions

Powering the science

Principal Investigators

Aristotelis Tsirigos, PhD, Colton Consortium Member

Professor, Department of Medicine, NYU Grossman School of Medicine, NYU Langone Health

Eric K. Oermann, MD, Colton Consortium Member

Associate Professor, Department of Neurosurgery, NYU Grossman School of Medicine / NYU Langone Health

Colton Centers & Institutions

Judith & Stewart Colton Center for Autoimmunity at NYU Langone Health

NYU Grossman School of Medicine

Research Outputs

From insight to impact

Publications

Metabolic coordination between skin epithelium and type 17 immunity sustains chronic skin inflammation

Immunity
Subudhi, I; Konieczny, P; Prystupa, A; Castillo, RL; Sze-Tu, E; Xing, Y; Rosenblum, D; Reznikov, I; Sidhu, I; Loomis, C; Lu, CP; Anandasabapathy, N; Suárez-Fariñas, M; Gudjonsson, JE; Tsirigos, A; Scher, JU; Naik, S July 2024
Animal ModelsBiological & MechanisticCytokine SignalingExperimental Platforms & ModelsImmune ProfilingImmunometabolismIn Vitro ModelsSingle Cell TechnologiesSpatial BiologyT Cell BiologyTherapeutic DevelopmentTranslational & ClinicalAllergic & Atopic DiseasesDermatologic DiseasesPsoriasisNew York University