Assessing metabolic progression in temporal lobe epilepsy: biological and methodological considerations

Magnetic resonance (MR) and [¹⁸F]fluorodeoxyglucose (FDG) positron emission tomography (PET) studies suggested that temporal lobe epilepsy (TLE) might be a progressive disease with mesial temporal regions being the most vulnerable to epileptic seizures. However, the biasing effect of brain atrophy and mesial temporal sclerosis (MTS) has not been assessed. We assessed the relationship between epilepsy duration and cerebral glucose metabolism (CMRglu) before and after partial volume correction (PVC).

We used PET with FDG in 22 TLE patients to quantify absolute and normalized CMRglu values. An automated, MR-segmentation-based PVC algorithm was applied.

PVC increased CMRglu variability. Normalization of regional absolute CMRglu values to global gray matter reduced variability. There was no significant relationship between absolute CMRglu values and epilepsy duration. A significant negative relationship was found between normalized CMRglu values in parahippocampus and hippocampus that persisted after PVC only in parahippocampus. A significant positive relationship between CMRglu asymmetry indices (AIs) and epilepsy duration was found in hippocampus before, but not after PVC. In patients with MTS, before PVC a statistical trend between epilepsy duration and AI was obtained for hippocampus, superior temporal cortex, and amygdala. After PVC, statistical significance was attained in amygdala.

Ipsilateral mesial temporal regions are one of the areas most sensitive to epileptic seizures. Assessment of metabolic progression is affected by methodological factors, such as brain atrophy (and the correction thereof) and data variability. Normalized CMRglu values and AIs have greater statistical power than absolute CMRglu values implying no need for quantitative PET studies.