B-180 Performance of Anti-14-3-3eta Multiplex Autoantibody Assay in Differentiating Axial Spondyloarthritis from Diseases with Overlapping Symptoms

Abstract

Background Axial Spondyloarthritis (axSpA) is a chronic autoimmune disease often resulting in spinal fusion and significant disability if left untreated. A major challenge in early intervention is the prolonged diagnostic delay, typically 7-10 years, driven by symptom overlap with chronic lower backpain and frequent misdiagnosis in primary care as presentation can mimic other inflammatory diseases. Anti 14-3-3eta autoantibodies have been identified as a biomarker candidate capable of differentiating axSpA from inflammatory back pain. A validated serologic marker capable of discriminating axSpA from clinically overlapping autoimmune and inflammatory diseases could streamline differential diagnosis and accelerate referral. This study evaluates whether anti-14-3-3eta autoantibodies can distinguish axSpA from the broad range of inflammatory and arthritic diseases that commonly delay and complicate diagnosis. Methods Anti-14-3-3eta autoantibody levels incorporating five non-overlapping peptide regions of the 14-3-3eta protein were measured in a total of 151 serum samples using a Luminex multiplex immunoassay. Cohorts included patients diagnosed with axSpA (n=30) according to modified New York criteria, healthy controls (n=60), and disease controls representing conditions with overlapping symptomatology: rheumatoid arthritis (n=19), psoriatic arthritis (n=19), psoriasis (n=4), Crohn’s disease (n=4), ulcerative colitis (n=2), osteoarthritis (n=12), and gout (n=1), for a total of 151 samples. Anti-14-3-3eta autoantibody values were log10-transformed and analyzed using nominal logistic regression (JMP v18). For each peptide, a univariate model was fitted to assess its ability to classify axSpA versus controls. Model performance was summarized using likelihood-ratio, AICc/BIC, and ROC AUC. Optimal classification thresholds were determined using Youden’s index, and diagnostic odds ratios (DxOR) were calculated at these cut points. Statistical significance was set at p<0.05. Results As demonstrated in Figures 1A-E, anti-14-3-3eta autoantibody levels were significantly higher in axSpA when compared to healthy and disease controls across all peptides. ROC analysis (Figure 1F) demonstrated that all five peptides produced statistically significant discrimination. Peptide 2 demonstrated the strongest performance (AUC=0.657, p=0.0080), followed by Peptide 1 (AUC=0.637, p=0.0203), Peptide 3 (AUC=0.634, p=0.0307), Peptide 4 (AUC=0.628, p=0.0229), and Peptide 5 (AUC=0.598, p=0.0969). Univariate logistic regression confirmed significant associations for all peptides (4.49–6.30, p = 0.034–0.012; RSquare(U) = 0.030–0.042). At Youden’s optimal thresholds, diagnostic odds ratios ranged from 3.09 to 47.1, with Peptide 2 demonstrating the strongest discrimination (DxOR=47.1), followed by Peptide 4 (DxOR=20.4), Peptide 1 (DxOR=19.9), Peptide 3 (DxOR=15.7), and Peptide 5 (DxOR=3.09). Conclusion This study demonstrates that anti-14-3-3eta autoantibody responses provide statistically significant and disease specific differentiation of axSpA from a broad set of clinically overlapping conditions. Peptide level performance supports the assay’s biological and diagnostic relevance. These results position anti-14-3-3eta autoantibodies as a promising serologic marker that could help shorten the diagnostic delay that currently impedes timely intervention. Future work will focus on validating these findings in larger cohorts and assessing performance in early disease.