Abstracts

Ictal video-EEG telemetry (VET) remains the gold standard to determine the localization and resectability of seizure foci. However, there is a notion that in well-selected patients, VET may only prove the obvious and it may be redundant. Few studies have formally investigated how well outpatient investigations (MRI, outpatient EEG, neuropsychological tests) predict the findings of VET in patients with temporal lobe epilepsy (TLE). We explored the predictive value of outpatient investigations and clinicians[apos] judgement with regards to VET findings and surgical decision in TLE patients.
We reviewed charts of 63 patients with TLE and mesial temporal sclerosis (MTS), who underwent epilepsy surgery. Based on one outpatient EEG, neuropsychological findings and seizure semiology, 3 blinded adjudicators (epileptologists) determined the laterality of the epileptogenic focus and rated on a five-point Likert scale (5=[quot]definitely[quot] to 0=[quot]definitely not[quot]) whether each patient should undergo a temporal lobectomy. In addition, an MRI report for each patient was adjudicated by a blinded neuroradiologist, who determined the side of the MTS and how confident he was in this assessment. We calculated the sensitivity and specificity of the epileptologists[apos] decision using the results of VET as the gold standard.
In 45 patients the neuroradiologist predicted the side of MTS with a high degree of certainty (either 5 or 4 on a five-point Likert scale) coupled with a definite or probable recommendation for temporal lobectomy from an epileptologist. The epileptologists[apos] sensitivity for TLE seizure lateralization was 97% (95% CI: 94-99%), and specificity was 33% (95% CI: 13-53%). The correlation between the adjudicators[apos] predicted side (both the epileptologists[apos] and neuroradiologist[apos]s ratings) and the actual side as determined by VET was 0.92 (95% CI: 88-96%). In no instance did an adjudicator predict a focus contralateral to the side found on VET.
Based solely on outpatient investigations (MRI, EEG and neuropsychological tests) experienced clinicians can predict laterality of TLE and surgical decision with a high level of accuracy (no errors). With regards to surgical decision, epileptologists rarely made an incorrect recommendation for temporal lobectomy (sensitivity 94-99%), but erred on the side of caution by not identifying patients who might be treated surgically (specificity 13-53%). Arguably, the latter could be improved by further outpatient testing.
[Supported by: Student Research Training Program - The University of Western Ontario]