Sphenoidal electrodes in localising temporal epileptic focus, in association with CT, MRI and SPECT.

Abstract

We read with interest the article by Duncan et al.'MRI is certainly superior to CT, and SPECT is superior to MRI in detecting lateralising lesions in temporal lobe epilepsy.2Interictal studies in temporal lobe epilepsy using both PET and SPECT show usually focal temporal hypofusion as the most common abnormality.' The danger in localising epileptic focus by PET in the fact that these investigations are carried out interictally. In this, the authors (we infer) regard interictal SPECT as giving information analogous to that given by interictal EEG spikes. However, as they themselves point out, the correlation between the two is often not good. Our experience suggests that simple unilateral temporal hypoperfusion is indeed a reliable localising finding. We would point out, however, that we find this in only around 30% of our overall series of patients with complex partial seizures. A further 30% have other findings (such as more extensive hypoperfusion, focal hyperperfusion or combinations of hypoperfusion and hyperperfusion), the reliability of which we are less sure of. The remaining patients have normal regional cerebral blood flow (rCBF). Series which report localising abnormalities in higher proportions of patients''3 do tend to have some patients who localise falsely using SPECT. Hence we feel strongly that localising reliability depends crucially on conservative reporting of images.SPECT can of course be performed immediately postictally, or even during a seizure, as we are increasingly succeeding in doing. Initial data (our own and from elsewhere4) suggest that this provides localising information in a higher proportion of patients, and may in particular help make the important discrimination between frontal and temporal foci.'We hope very much that the development of SPECT imaging of rCBF (and more recently of benzodiazepine receptor density) will reduce the need for the longterm and invasive EEG monitoring at present necessary in so many patients, rather than simply adding yet another test to an already extensive assessment. To what extent this turns out to be possible will depend on the results of longterm assessment of its ability to predict surgical success.