Supportive Data

An international collaborative group (Mayo Clinic, Oxford University, and McGill University) compared sensitivity and specificity of AQP4 FACS assays to other assay types including fixed, permeabilized cell-based assays (CBA, observer-scored immunofluorescence, Euroimmun), tissue-based immunofluorescence (IF), ELISA, and immunoprecipitation assay (IPA) in a blinded fashion among 60 neuromyelitis optica spectrum disorder (NMOSD) cases and 86 control subjects. Clinical sensitivity of AQP4 FACS was superior to the other assay types. Sensitivities were: AQP4 FACS (M23 isoform), 77%, AQP4-CBA (M1 isoform), 73%; M1-AQP4-ELISA, 60%; IPA, 53%; tissue-based IF, 48%. Specificities were 100% for all assay types, except the Mayo Clinic IPA (97%).(6)

In 2014, a systematic comparison of AQP4-IgG assays, in a clinical service setting, confirmed superiority of FACS assays over ELISA. Higher-order arrays of M23-AQP4 (M23-AQP4-FACS) and M1-AQP4-ELISA were associated with false-positive results. Overall, M1-AQP4-FACS was 83% sensitive for NMO compared with 75% for M23-AQP4-FACS, 75% for M1-AQP4-CBA and 58% for M1-AQP4-ELISA. Assays specificities for NMO were: M1-AQP4-FACS, 100%, M1-AQP4-CBA, 100%, M1-AQP4- ELISA, 99%; and M23-AQP4- FACS, 95%.(7)

AQP4 FACS analysis was done for serum samples from 36 random patients with a diagnosis of NMO. All samples were tested with our validated AQP4 CBA assay. Thirty samples (83.33%) were positive by FACS and 29 samples (80.55%) were positive by CBA. All 6 samples that were negative by FACS also tested negative by CBA.

To measure the specificity, AQP4 FACS analysis was done for CSF of 338 non-NMO(SD) patients. None of the samples were positive by this assay (specificity=100%).