Abstracts

Rationale: To date, children with epileptic seizures undergo brain MRI to determine a structural etiology. The indication for brain MRI in these children is largely based on expert opinion rather than objective criteria and algorithms. The aim of this study was to identify clinical factors that predict an abnormal brain MRI in a retrospective cohort of children with epilepsy at our pediatric medical center. Methods: An 18 months retrospective chart review was conducted in all children age 1 month to 18 years of age diagnosed with epilepsy, who underwent a brain MRI from January 1, 2010 through July 30, 2011. Each child's medical chart was reviewed and specific and detailed clinical markers including patient history, physical examination, EEG and MRI findings were identified. This information was entered into a clinical database Multivariable Poisson regression with robust variance estimates were used to calculate adjusted relative risks using the GEE capability of SAS PROC GENMOD. To assess model predictive capability, ROC curves were also estimated. Results: A total of 228 patients were identified, 54 % males. The average age of the children at diagnosis of epilepsy was 5.52 years old. When all components of the physical exam were combined, 39.9% of patients had an abnormality. A total of 225 EEG's were reviewed, of which 92% were abnormal. The majority of abnormalities found were focal epileptiform discharges in 125 patients' EEG's (55.6%), whereas diffuse epileptiform discharges were only found in 55 EEG's (24.4%). EEG slowing was also analyzed, with the majority having diffuse changes in 102 EEG's (45.3%), and focal slowing in 66 EEG's (29.3%). Of the 228 MRI's reviewed, 91 were abnormal (39.9%). In the multivariate analysis, developmental delay was found to have a relative risk of 1.4995 (CI: 1.0648 - 2.1118, p=0.0204) for an abnormal brain MRI; in addition, patients with a diffusely abnormal physical exam had a relative risk of 1.5753 (CI: 1.1279 - 2.2002, p=0.0077). Interestingly, focal findings on physical exam were not found to be predictive (OR: 1.147; CI: 0.494 - 2.666, p=0.7494). Conclusions: : The expected role of unilateral focal findings on clinical examination and/or EEG was not found to be predictors of an abnormal MRI of brain. This study found that diffuse abnormal physical examination findings and developmental delay are independent predictors of an abnormal brain MRI in children with epilepsy. The results from this study are not statistically robust to build a diagnostic algorithm; however, to our knowledge this is the first study of its kind to assist physicians in their clinical decision to obtain MRI imaging in children with epilepsy. Since this is a retrospective study with its known limitations, further prospective research is needed to create a diagnostic algorithm for clinicians to aid in the management of children with epilepsy.