Analyzing Sleep-Only EEG Improves Positive and Negative Predictive Values of Quantitative EEG Method for Predicting Sturge-Weber Syndrome
Abstract number :
2.144
Submission category :
3. Clinical Neurophysiology
Year :
2011
Submission ID :
14880
Source :
www.aesnet.org
Presentation date :
12/2/2011 12:00:00 AM
Published date :
Oct 4, 2011, 07:57 AM
Authors :
J. B. Ewen, B. Lakshmanan, K. Lanier, E. H. Kossoff, , T. A. Zabel, N. E. Crone, A. M. Comi
Rationale: Children with port-wine birthmark (PWB) in the V1 distribution (forehead and/or upper eyelid) have about a 20% chance of developing brain involvement characteristic of Sturge-Weber syndrome (SWS). Many children with a V1 facial PWB receive an MRI in infancy to rule-out SWS despite risks of contrast and sedation. We have previously published a quantitative EEG (qEEG) method to differentiate infants with PWB plus SWS from those with PWB without SWS; we validated this technique in 9 infants. Here we examined a larger cohort to develop sensitivity, specificity, and positive and negative predictive values. We subsequently examined a subset of EEGs in which we analyzed only sleeping tracings in order to determine whether movement-free signals from sleep EEG would offer improved sensitivity and specificity.Methods: Using a previously published qEEG method of assessing amplitude asymmetries (Ewen et al, Clin Neurophysiol 2009), we examined 44 EEGs from 21 subjects (SWS: 15 EEGs from 10 subjects; PWB only: 29 EEGs from 11 subjects), in which either sleep or waking EEG was analyzed (group 1). We then analyzed the subset of EEGs (group 2) that had sleep tracings (SWS: 8 EEGs in 6 subjects; PWB only: 20 EEGs in 10 subjects). We calculated sensitivity and specificity in both groups. Assuming a prevalence of SWS of 20% in infants with facial PWB, we calculated positive and negative predictive values (PPV/NPV).Results: When all EEGs were examined (group 1), this analysis had a sensitivity of 73% and a specificity of 72% (PPV was 41%; NPV was 92%). When only sleep EEGs were analyzed (group 2), the sensitivity was 88% and the specificity was 80% (PPV was 53%; NPV was 97%).Conclusions: This technique offers a robust NPV that may reassure clinicians and parents of infants with V1 facial PWB that the child does not have SWS and may decrease the number of negative MRIs performed. The examination of sleep tracings may improve the performance of this test, perhaps because the data are less contaminated with muscle and movement artifact.
Neurophysiology