Serologic Diagnosis of Human Neuropsychiatric Lupus Using the Immunarray ICHIP (R)

Chaim Putterman, D. Scott Batty, I. R. Cohen, Nicole Jordan, Debbie Rybak, Tamar Rubinstein, Keren Jakobi, Rachel Sorek, Anat Reiner-Benaim, Yakov Blumenstein, Pennina Safer

Research output: Contribution to journalMeeting Abstractpeer-review


Peripheral neurologic syndromes and central nervous system (CNS) manifestations are recognized as primary disease manifestations in systemic lupus erythematosus (SLE). Neuropsychiatric SLE (NPSLE) involves a wide range of nervous system disorders, and can affect 50% or more of SLE patients(1). There is no single diagnostic test specific for NPSLE; rather, the diagnosis is currently based on the combined use of serological testing, functional and/or structural neuroimaging, and standardized neurological and neuropsychological assessments(1). A sensitive and specific serological biomarker for NPSLE would be very helpful in the management of lupus patients. Using a novel antigen microarray platform, unique multivariate classification models were developed to distinguish between patients with NPSLE and SLE patients without active neuropsychiatric manifestations (non-NPSLE).
Thirty-eight SLE serum samples were obtained from the Einstein Lupus Cohort at the Albert Einstein College of Medicine (Bronx, NY) and tested using the ImmunArray iCHIP®(2), printed with a set of 225 antigens associated with SLE and/or brain injury. All SLE samples satisfied the ACR classification criteria. Twelve of the 38 patients were diagnosed as positive for NPSLE based on the use of a validated questionnaire (3). Two independent classification methods were developed. Classifier training and testing were performed based on 5-fold cross validation on all samples.


Both classification methods differentiated between the lupus patients with and without neuropsychiatric symptoms. The support vector machine (SVM) classification model performed with sensitivity greater than 99% and specificity of 88%. The logistic regression method separated the populations with 83% sensitivity and 96% specificity. Several auto-antigens were shared between the two models, increasing the confidence in the selected antigen lists.
In a proof of concept study, classification methods based on autoantibody profiles from the ImmunArray iCHIP® were able to successfully distinguish between lupus patients with and without neuropsychiatric symptoms. Our preliminary results based on 38 patients are very promising and warrant additional validation in a larger cohort of NPSLE patients.
Original languageEnglish
JournalArthritis and Rheumatology
Issue number10
StatePublished - Oct 2015
Externally publishedYes


Dive into the research topics of 'Serologic Diagnosis of Human Neuropsychiatric Lupus Using the Immunarray ICHIP (R)'. Together they form a unique fingerprint.

Cite this