Background/Purpose:

Rheumatoid Arthritis (RA) is a chronic inflammatory autoimmune disease primarily targeting the synovium. RA is B-cell driven, and is associated with autoantibody production. The clinical heterogeneity in RA and overlapping symptoms with other diseases can pose diagnostic challenges, especially in early disease. An accurate diagnosic tool or a means to further stratify patients could have a significant impact on patient care.

Immunosignature (IS) technology permits differential diagnosis of related automimmune diseases based on distinct serum autoantibody profiles, as determined by binding to a miroarray containing 126K distinct peptides with an average length of 9 amino acids. Our unique approach to peptide microarray fabrication, combining photolithography with optimized peptide chemistry and MALDI-based quality control, enables low-cost, rapid, and reproducible testing.

Here, the IS technology was applied to the serological differentiation of RA from other rheumatic diseases (inflammatory and non-inflammatory) and healthy controls.

Methods:

379 serum samples were prospectively collected, including RA (n=95), systemic lupus erythematosus (SLE) (n=75), Sjögren’s syndrome (SS) (n=20), osteoarthritis (OA) (n=24), fibromyalgia (n=22), other disease (OD) (n=76) and healthy controls (HC) (n=59). Subjects with rheumatological diseases were diagnosed based on ACR criteria. There were no significant differences in gender, race, or ethnicity across all groups. Antibody(IgG)-peptide binding was quantified and peptides with significant intensity differences between contrasting groups were identified by Bonferroni adjusted t-test. Support vector machine classifiers were trained using the most distinguishing peptides between contrasts. Classifier performance was evaluated by a cross-validation routine that included feature selection, model training, and model testing.

Results:

The number of significant peptides that discriminate RA from other groups and classification cross-validated area under the curve (cvAUC) values are summarized in the table below.

Conclusion:

Using IS technology, RA is best discriminated from patients with SLE and HC. Nevertheless, RA can also be differentiated from closely-related conditions such as SS with modest cvAUCs. The results presented represent a step toward creating a single test using a small serum sample capable of multi-classification across a range of symptomatically related diseases and in patients with conditions referred to rheumatologic evaluation. Whether diagnostic accuracy would be improved by combining results from standard serological tests is being studied. Verification in cohorts from other sites and validation in blinded studies would allow for further model refinement to create a robust diagnostic assay.

« Back to 2017 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/a-single-immunosignature-test-accurately-discriminates-ra-from-related-autoimmune-and-inflammatory-disorders/