Abstracts

Rationale: Given the risk of developing Sturge-Weber Syndrome (SWS), the presence of a facial port-wine birthmark (PWB) bears considerable concern for pediatricians and families of the newborns who bear them. There is no consensus on optimal timing for magnetic resonance imaging (MRI) with gadolinium, the gold standard for SWS diagnosis. Early identification of disease may allow for prophylactic therapies. MRI, however, has low sensitivity in the newborn period. If completed early, it may need to be repeated to determine presence and extent of brain involvement, exposing the infant again to the risks of sedation and contrast exposure. With low cost and favorable safety profile, quantitative EEG (qEEG) holds considerable promise in risk stratification for infants born with PWBs. A qEEG threshold for asymmetry has previously been determined to identify SWS in a small group of infants with facial PWB (Ewen et al, Clin Neurophysiol 2009; 120; 1433-1440). We aim to validate qEEG’s use in providing prognostic information for SWS in a larger cohort of infants. Methods: Participants were part of a larger, longitudinal natural history study of SWS that recruited infants with facial PWBs. Participants were retrospectively identified and included if they had an MRI with contrast, were followed clinically through 1 year of age, and had an EEG prior to 1 year. Infants were classified in one of two SWS classification groups: SWS- (normal brain MRI after 1 year of age) or SWS+ (appearance of contrast-enhancing leptomeningeal lesions on brain MRI). A pediatric electroencephalographer blinded to clinical history selected the first 30 artifact-free epochs of 2 seconds duration. Fast Fourier transform was applied to all bipolar channels in each epoch to generate a power value, which was summed into frequency bands. A mean laterality score was generated to quantify the power asymmetry in each band by the formula (ipsilatera-contralateral)/(ipsilateral+contralateral) for pairs of symmetrical bipolar channel (Hatfield et al., Epilepsia 2007; 48(1); 191-195). Using the previously determined and validated threshold of 0.2, the qEEG was deemed abnormal if this threshold was exceeded in whole hemisphere or occipital bands, indicating lower power on the side of the facial PWB. Results: 48 infants met inclusion criteria; 32 eventually were diagnosed with SWS and the other 16 had a negative MRI and no symptom emergence after 1 year of age. The median age at EEG was 3 months. When comparing qEEG results (normal or abnormal) to eventual SWS diagnosis, the sensitivity and specificity of qEEG in predicting accurate diagnosis were 53% and 81%, respectively (p 0.02). Pre-test probability was estimated at 25% to approximate the risk of SWS in facial PWBs. An abnormal qEEG increased the probability of eventual SWS diagnosis 25% to 48%. A normal qEEG increased the probability of a negative MRI after age 1 year from 75 to 84%. Conclusions: QEEG contains information relevant to the risk of development of SWS in infants with facial port-wine birthmarks. QEEG may serve as a prognostic biomarker for development of SWS in infants with facial PWBs, the current reference standard of which is diagnosis with MRI at a later age. Furthermore, qEEG may provide adequate reassurance that an asymptomatic infant with a normal qEEG will not develop SWS. This may allow families and clinicians to feel comfortable deferring an MRI until after at least 1 year of age. Funding: Brain Vascular Malformation Consortium, 5U54NS065705-08, NIH/ NINDS/ORD