Multifactorial analysis of opsoclonus-myoclonus syndrome etiology ("Tumor" vs. "No tumor") in a cohort of 356 US children.

Abstract

Pediatric opsoclonus-myoclonus syndrome (OMS) presents a paradox of etiopathogenesis: A neuroblastic tumor (NB) is found in only one half of the cases, the others are ascribed to infections or designated as idiopathic. From an IRB-approved observational study of 356 US children with OMS, secondary analysis of "etiology" and related factors was performed on a well-characterized cohort. The "Tumor" (n=173) and "No Tumor" groups (n=183), as defined radiologically, were compared according to multiple factors considered potentially differentiating. Data were analyzed retrospectively using parametric and nonparametric tests as indicated. Patients with NB were not distinguishable by prodromal symptoms, OMS onset age, gender, race/ethnicity, OMS severity, rank order of neurological sign appearance, or geographic distribution. Various CSF immunologic biomarker abnormalities of OMS did not vary in the presence or absence of a detectable tumor: frequency of six lymphocyte subsets, or concentrations of 18 cytokines/chemokines, cytokine antagonists, chemokine receptors, cell adhesion molecules, or neuronal/glial markers. Prior responsiveness to conventional immunotherapy was not contingent on tumor/no tumor designation. Multiple convergent factors provide compelling empirical evidence and rationalize the concept that OMS is one neurological disorder, regardless of apparent etiology. Limitations to the current clinical etiologic classifications as paraneoplastic, parainfectious/post-infectious, and idiopathic etiology require antigen-based biological solutions to tease out the molecular pathophysiology of viral/tumoral mechanisms. Systematic studies, regardless of presumed etiology, will be necessary to find the highest-yield combination of imaging approaches, screening for infectious agents, and new biomarkers. Two testable hypotheses for future research are presented.