Autoantibody-Based Models for the Diagnosis and Activity Assessment of Takayasu Arteritis

The precise diagnosis and activity assessment of Takayasu arteritis (TA) remain challenging. Using a two-phase approach, we screened serum from 9 TA patients and 10 healthy controls via HuProt microarrays, identifying three novel autoantibodies (anti-Table 2, anti-TRAC, anti-LILRA5). These were validated by ELISA in 175 samples (61 active TA, 42 inactive TA, 72 controls). Logistic regression models were developed from training (70%) and test (30%) sets. For diagnosis, Model 1 achieved AUCs of 0.88 and 0.86, outperforming CRP (0.58) and ESR (0.61). For activity assessment, Model 2 achieved AUCs of 0.78 and 0.79, superior to CRP (0.62) and ESR (0.58). These novel autoantibody-based models enable accurate, non-invasive diagnosis and activity assessment in TA.