Accuracy of Antibody Testing in Myasthenia Gravis: The ADAM Study

Antibody testing has a pivotal role in the diagnosis of myasthenia gravis (MG). However, there are limited data on the accuracy of the different methods available for acetylcholine receptor (AChR) and muscle-specific kinase (MuSK) autoantibody (Ab) detection in a real-world setting. Our aim was to compare the diagnostic accuracy of in-house live cell-based assay (L-CBA), a commercially available fixed CBA (F-CBA), and an indirect ELISA in detecting AChR and MuSK antibodies in patients with suspected MG.

Between July 2023 and December 2025, we prospectively recruited consecutive adult and pediatric patients with suspected MG across 4 Italian centers. Patients undergoing immunotherapy and those with incomplete diagnostic work-up or insufficient samples were excluded. Serum samples were tested blindly by L-CBA, F-CBA, and ELISA for AChR and MuSK antibodies. MG diagnosis was established independently based on neurologic examination, electrophysiologic studies, and/or sustained clinical response to pyridostigmine or corticosteroids as well as exclusion of other diagnoses. Diagnostic sensitivity, specificity, and receiver operating characteristic (ROC) curves were calculated using final diagnosis as the reference standard.

Of the 327 included patients (median age 63 years, interquartile range 26, 52.9% female), 152 were diagnosed with MG and 175 with other diagnoses. For AChR antibodies, sensitivity was 71.7% (95% CI 63.8-78.7) for L-CBA, 69% (61.0-76.3) for F-CBA, and 63.8% (55.6-71.4) for ELISA, while specificity was 97.7% (94.2-99.3) for both CBAs vs 69.1% (61.7-75.9) for ELISA. ELISA showed lower concordance with CBAs and produced substantially more false positives (30.9%). For MuSK antibodies, sensitivity was similar (6.6 vs 7.2%) and specificity 100% across assays. ROC curve analysis confirmed higher diagnostic accuracy for L-CBA (AUC 0.85) compared with F-CBA (0.83) (p = 0.043) and ELISA (0.72) (p < 0.0001). The AUCs of MuSK-Ab were not significantly different between different tests.

CBAs, especially live, demonstrated superior diagnostic performance for AChR antibody detection compared with ELISA, particularly in ocular MG. L-CBA showed a modest but significant diagnostic advantage over F-CBA, although fixed assays can be considered as a practical, nearly equivalent alternative, suitable for routine laboratories. By contrast, ELISA demonstrated lower specificity and should be interpreted cautiously.

This study shows Class I data that L-CBA and F-CBA compare favorably with ELISA testing of AChR-Ab and MuSK-Ab, in the diagnosis of MG.