Machine Learning-Enhanced Autoantibody Discovery and Diagnostics in Systemic Autoimmune Rheumatic Diseases

Clinical Snapshot

Key Findings

• P Value: Not reported

• Effect Size: Not reported — narrative review without pooled effect size calculation

• Primary Outcome: Conceptual overview of ML applications in autoantibody discovery and diagnostics for SARDs — no primary quantitative outcome is reported or synthesised

• Nnt Or Sensitivity: Not reported — individual study diagnostic accuracy metrics (AUC, sensitivity, specificity) may be described narratively within the full text but are not synthesised into summary estimates; no NNT, pooled sensitivity, or pooled specificity is calculable from this review

• Confidence Interval: Not reported

Clinical Application

Clinical implementation of ML-enhanced autoantibody diagnostics remains in early stages. Barriers include: lack of regulatory-approved ML diagnostic tools for SARDs in most jurisdictions; requirement for large, well-annotated training datasets; absence of standardised ML model validation frameworks; need for clinician education in interpreting probabilistic ML outputs; and infrastructure requirements for high-throughput autoantibody platforms (e.g., protein microarrays, PhIP-Seq). Near-term feasibility is limited to research and academic settings. In Australia, autoantibody testing for SARDs is partially supported under the Medicare Benefits Schedule (MBS), with items covering ANA, anti-dsDNA, ENA panels, and specific antibodies (e.g., anti-CCP, ANCA). ML-enhanced autoantibody panels are not currently TGA-approved or MBS-listed. The RACGP and Australian Rheumatology Association (ARA) guidelines for SARD diagnosis rely on conventional serology. Australian rheumatology centres, particularly those affiliated with academic institutions, may be positioned to participate in ML autoantibody research consortia. The diverse ethnic composition of the Australian population — including significant Aboriginal and Torres Strait Islander, Asian, and South Asian communities with differing SARD prevalence and autoantibody profiles — underscores the importance of ethnically representative training datasets for any future ML diagnostic tools deployed in Australia. PBS listing of any novel ML-guided diagnostic or therapeutic pathway would require TGA approval and MSAC health technology assessment. Rheumatologists, clinical immunologists, and general physicians managing patients with suspected or established SARDs, particularly in settings with access to advanced autoantibody testing platforms. Most directly relevant to specialist and academic centre practice rather than primary care, given the complexity of ML-derived biomarker interpretation.

Abstract

The growing implementation of machine learning (ML) has extended into autoantibody research for the study of systemic autoimmune rheumatic diseases (SARDs). ML methods offer a promising approach for efficiently handling and identifying important signals within the big data generated by modern autoantibody technologies. The novel biomarkers identified through advanced ML techniques show promise in outperforming current clinical tools, bringing us closer to the goal of precision medicine. In this article, we will provide an overview of ML approaches and how they have been applied in autoantibody research to improve the diagnosis and characterization of SARDs.

References

1. 1.Mocanu, V., Moghaddam, F., Aminghafari, M., & Choi, M. Y. (2026). Machine learning-enhanced autoantibody discovery and diagnostics in systemic autoimmune rheumatic diseases. Rheumatic Disease Clinics of North America. Advance online publication. https://doi.org/10.1016/j.rdc.2026.03.003