Intractable nonlesional epilepsy of temporal lobe origin: lateralization by interictal SPECT versus MRI.

We performed a retrospective study of 53 consecutive "nonlesional" temporal lobectomy patients to assess the relative utility of MRI versus interictal single-photon emission computed tomography (SPECT) in this patient population. We compared the seizure lateralizing properties of MRI and SPECT using multiple blinded expert reviewers for both SPECT and MRI with a test-retest reviewer paradigm and measurements of hippocampal volume from MRI. The criterion standard for seizure lateralization was satisfactory postoperative seizure control (n = 43). The rate of correct seizure lateralization was significantly greater for MRI than for SPECT (p < or = 0.01), and the rate of incorrect lateralization was significantly less for MRI than for SPECT. The most accurate MRI measure was hippocampal volume measurements, which correctly lateralized the seizures in 86.0% of cases. The correct lateralization rate for SPECT was 45.4%. The MRI and SPECT studies tended to be noncomplementary with respect to seizure lateralization, and SPECT was likely to give an incorrect or indeterminate result in patients who were not lateralized by MRI. Concordant MRI-EEG lateralization was a strong predictor of satisfactory postoperative seizure control, while no relationship between postoperative seizure control and SPECT findings was present.