Abstracts

RATIONALE: The development of computer-aided subtraction SPECT technique has enhanced the usefulness of SPECT as a presurgical evaluation tool. False localizing ictal SPECT is often attributed to the post-ictal injection or rapid spreading of the epileptic discharges from the epileptogenic zone to secondary symptomatogenic zone. Even though many clinical studies on the value of ictal SPECT have been published, a systematic evaluation of the clinical and neurophysiological characteristics of non-concordant or false localizing ictal SPECT has not been performed.
METHODS: Two of the authors (YDP and DWK) reviewed the results of the SISCOM (Subtraction Ictal SPECT Co-registered to MRI) in 107 patients who underwent presurgical evaluation for intractable epilepsy at the Medical College of Georgia epilepsy program from 1998 to 2001. We defined non-concordance or false localizing SISCOM when: 1) the injection of radioisotope (99mTc-ECD) was clearly done during the patients[ssquote] habitual seizure confirmed by the review of ictal video/EEG; 2) the primary focus of ictal hyperperfusion on SISCOM was not concordant with other localizing methods such as MRI and video/EEG findings; and 3) other technical variants such as movement artifacts, scan time were excluded. The primary focus on SISCOM was based on the intensity and the volume of the hyperperfusion. Patient[ssquote]s clinical, radiological and neurophysiological data were reviewed by a separate author who was blinded to the results of SISCOM.
RESULTS: Among 107 patients, 35 patients (M 21, F 14 age 1-55, mean 23) were classified as having non-concordant SISCOM. The time of radioisotope injection and the duration of seizures from the non-concordant group were not different from the concordant group. Developmental malformation was the most common etiology for the false localizing group (10/35, 29%), followed by head trauma (9/35, 26%). More than half of the non-concordant group had true epileptogenic foci in either frontal (13/35) or temporal lobe (12/35). Common location of false localization was in wrong temporal lobe (8/35, 23%) (contralateral to true epileptogenic zone) followed by wrong occipital lobe (5/35, 15%) (contralateral to true epileptogenic zone) and wrong orbitofrontal lobe (3/35, 9%). Scalp EEG monitoring demonstrated the rapid spread of the epileptic discharges to the contralateral hemisphere before the radioisotope injection in 54% (19/35). Subclinical epilepsia partialis continua during the interictal SPECT was implicated in three patients with Rasmussen Encephalitis and false localzing SISCOM (9%, 3/35). Seven of the 8 false localizing temporal lobe SISCOM patient had bilateral independent temporal lobe onset seizures during the video/EEG monitoring. More than half (18/35) of the patients underwent subsequent intracranial EEG monitoring and resective surgery was performed in 14 patients: extratemporal resection = 9, temporal lobectomy = 3, and hemispherectomy = 2. Their surgical outcome (Class I = 9, II = 2, III = 2, IV = 1 by Engel[ssquote]s classification at 1 year post-op ) was comparable to the concordant SISCOM group (Class I = 22, II = 7, III = 4, IV = 1, n = 34).
CONCLUSIONS: Our study suggests that the subraction SPECT could be falsely localizing in some partial seizures. Careful clinical and neurophysiological correlation is warranted in these patients.