Transforming ANA Testing: Leveraging Epidemiology and AI to Detect Autoimmune Disease

Antinuclear antibody (ANA) testing is critical for the diagnosis of organ-specific and autoimmune diseases, such as systemic lupus erythematosus (SLE), which affect millions of patients. Although ANA production is not unique to a single disease, specific autoantibodies (e.g., anti-DNA, anti-centromere) can serve as highly characteristic biomarkers. However, ANA testing presents several challenges in clinical practice. In particular, a growing portion of the US population is ANA-positive (estimated at 16%), only a subset of whom have autoimmune disease, creating difficulties distinguishing pathological ANA from incidental findings. The long-term risk of autoimmune disease in ANA-positive patients without an immediate diagnosis is also unknown. In addition, the gold standard ANA test—the indirect immunofluorescence (IIF) assay using human HEp-2 cells—suffers from key technical limitations. This test identifies fluorescence staining patterns that guide further testing for specific autoantibodies; however, reliance on expert visual interpretation introduces variability and inconsistency and limits clinical utility. These issues lead to further diagnostic ambiguity, resulting in patient anxiety, unnecessary referrals, over- treatment, and strains on healthcare systems suffering from a shortage of rheumatology specialists. Here, we aim to address these issues by filling two key knowledge gaps: 1) the long-term risk of developing autoimmune diseases in ANA-positive individuals is poorly defined, and 2) the ANA IIF test lacks standardization and predictive integration. To this end, we will leverage epidemiological data, biobank resources, and advances in artificial intelligence (AI) technologies to improve ANA testing and advance understanding of ANA positivity and progression to autoimmune disease. In Aim 1, we will assess the link between ANA positivity and long-term progression to autoimmune disease, testing our hypothesis that ANA-positive patients without autoimmune disease at time of testing are at higher long-term risk of developing systemic and organ-specific autoimmune diseases than ANA-negative controls. These studies will utilize two complementary resources, the Rochester Epidemiology Project and Mayo Clinic Biobank, both with decades of longitudinal data. In Aim 2, based on our preliminary data showing that AI-based computer vision can detect specific autoantibodies with high accuracy from IIF images, we will develop and validate AI models for identifying specific autoantibodies and autoimmune diseases directly from IIF images. This novel tool will use computer vision to analyze IIF images and identify imperceptible diagnostic patterns linked to specific autoantibodies and diseases, thereby standardizing interpretation, improving precision, and facilitating the early detection of autoimmune diseases, in alignment with the NIAMS Strategic Plan (2025–2029) priority of advancing knowledge through AI. Through these studies, which integrate unique epidemiological resources, biobank data, and AI-driven advances to address longstanding challenges in ANA testing, we aim to improve diagnostic accuracy, optimize specialty referrals, and enhance care for individuals with or at risk of autoimmune diseases. Project Number: 1R01AI196126-01 | Fiscal Year: 2026 | NIH Institute/Center: National Institute of Allergy and Infectious Diseases (NIAID) | Principal Investigator: Ali Duarte Garcia | Institution: MAYO CLINIC ROCHESTER, ROCHESTER, MN | Award Amount: $800,097 | Activity Code: R01 | Study Section: Aging, Injury, Musculoskeletal, and Rheumatologic Disorders Study Section[AIMR] View on NIH RePORTER: https://reporter.nih.gov/project-details/1R01AI19612601