SPECT brain imaging in epilepsy: a meta-analysis

A meta-analysis of SPECT brain imaging in epilepsy was performed to derive the sensitivity and specificity of interictal, postictal or ictal rCBF patterns to identify a seizure focus in medically refractory patients.

Papers were obtained by pooling all published articles identified by two independent literature searches: (a) Dialnet (EMBASE) or Radline by CD-ROM and (b) Current Contents searched manually. Literature inclusion criteria were: (a) patients had a localization-related epileptic syndrome; (b) more than six patients were reported; and (c) patients had at least an interictal EEG-documented epileptiform abnormality. Of 46 papers meeting these criteria, 30 contained extractable data. SPECT results were compared to localization by standard diagnostic evaluation and surgical outcome. Meta-analytic sensitivities for SPECT localization in patients with temporal lobe seizures relative to diagnostic evaluation were 0.44 (interictal), 0.75 (postictal) and 0.97 (ictal). Similar results were obtained relative to surgical outcome. False-positive rates were low relative to diagnostic evaluation (7.4% for interictal and 1.5% for postictal studies) and surgical outcome (4.4% for interictal and 0.0% for postictal studies).

The results were not dependent on tracer used (or dose), the presence of CT-identified structural abnormalities, blinding of image interpretation or camera quality (although data were more variable with low-resolution cameras). There were insufficient data for conclusions regarding extratemporal-seizure or pediatric epilepsy populations.

Insights gained from reviewing this literature yielded recommendations for minimal information that should be provided in future reports. Additional recommendations regarding the nature and focus of future studies also are provided. The most important of these is that institutions using SPECT imaging in epilepsy should perform ictal, preferably, or postictal scanning in combination with interictal scanning.