ScholarWorks@성균관대학교

초록

Purpose: This study aimed to compare the localizing accuracy, interobserver reliability, and prognostic value of ictal SPECT analyzed by Statistical Parametric Mapping (ISAS) versus Subtraction Ictal SPECT Co-registered to MRI (SISCOM) in patients with drug-resistant temporal lobe epilepsy (TLE). Methods: Sixty consecutive TLE patients who underwent resective epilepsy surgery were analyzed. All ictal SPECT studies were processed by ISAS and SISCOM. Two blinded reviewers independently identified the most prominent hyperperfusion regions across 40 predefined brain subregions. Localization concordance with the actual resection site was scored at both lobar and sublobar levels. Interobserver agreement and the correlation between imaging localization and postsurgical seizure outcomes were analyzed. Results: ISAS achieved significantly higher interobserver agreement than SISCOM for the localization (kappa = 0.34-0.88 vs. 0.14-0.69). ISAS concordance with surgical resection was 88.7% at the lobar level and 79.7% at the sublobar level, compared to 71.7% and 60.9% for SISCOM, respectively. Logistic regression revealed that only ISAS localization at both scales significantly predicted seizure-free outcome, whereas SISCOM did not. Conclusions: ISAS outperforms SISCOM in both localizing accuracy and observer reliability, and its concordance with surgical resection robustly predicts postoperative seizure freedom. These findings support the adoption of ISAS as the first-line SPECT postprocessing technique for presurgical evaluation of TLE.

키워드