Predictive models in the diagnosis and treatment of autoimmune epilepsy

Divyanshu Dubey, Jaysingh Singh, Jeffrey W. Britton, Sean J Pittock, Eoin Flanagan, Vanda A Lennon, Jan-Mendelt Tillema, Elaine C Wirrell, Cheolsu Shin, Elson So, Gregory D Cascino, Dean Marko Wingerchuk, Matthew T. Hoerth, Jerry J. Shih, Katherine C Nickels, Andrew McKeon

Research output: Contribution to journalArticle

35 Scopus citations


Objective: To validate predictive models for neural antibody positivity and immunotherapy response in epilepsy. Methods: We conducted a retrospective study of epilepsy cases at Mayo Clinic (Rochester-MN; Scottsdale-AZ, and Jacksonville-FL) in whom autoimmune encephalopathy/epilepsy/dementia autoantibody testing profiles were requested (06/30/2014-06/30/2016). An Antibody Prevalence in Epilepsy (APE) score, based on clinical characteristics, was assigned to each patient. Among patients who received immunotherapy, a Response to Immunotherapy in Epilepsy (RITE) score was assigned. Favorable seizure outcome was defined as >50% reduction of seizure frequency at the first follow-up. Results: Serum and cerebrospinal fluid (CSF) from 1,736 patients were sent to the Mayo Clinic Neuroimmunology Laboratory for neural autoantibody evaluation. Three hundred eighty-seven of these patients met the diagnostic criteria for epilepsy. Central nervous system (CNS)-specific antibodies were detected in 44 patients. Certain clinical features such as new-onset epilepsy, autonomic dysfunction, viral prodrome, faciobrachial dystonic seizures/oral dyskinesia, inflammatory CSF profile, and mesial temporal magnetic resonance imaging (MRI) abnormalities had a significant association with positive antibody results. A significantly higher proportion of antibody-positive patients had an APE score ≥4 (97.7% vs. 21.6%, p < 0.01). Sensitivity and specificity of an APE score ≥4 to predict presence of specific neural auto-antibody were 97.7% and 77.9%, respectively. In the subset of patients who received immunotherapy (77), autonomic dysfunction, faciobrachial dystonic seizures/oral dyskinesia, early initiation of immunotherapy, and presence of antibodies targeting plasma membrane proteins (cell-surface antigens) were associated with favorable seizure outcome. Sensitivity and specificity of a RITE score ≥7 to predict favorable seizure outcome were 87.5% and 83.8%, respectively. Significance: APE and RITE scores can aid diagnosis, treatment, and prognostication of autoimmune epilepsy. A PowerPoint slide summarizing this article is available for download in the Supporting Information section here.

Original languageEnglish (US)
StateAccepted/In press - 2017


  • Autoimmune limbic encephalitis
  • Diagnosis
  • Epilepsy
  • Immunotherapy
  • Paraneoplastic limbic encephalitis
  • Predictive model

ASJC Scopus subject areas

  • Neurology
  • Clinical Neurology

Fingerprint Dive into the research topics of 'Predictive models in the diagnosis and treatment of autoimmune epilepsy'. Together they form a unique fingerprint.

  • Cite this