Quantitative SISCOM assessment for epileptogenic zone localization: insights from a multicenter study comparing two software platforms in temporal lobe epilepsy

Abstract

Introduction: Pharmacoresistant epilepsy affects around one-third of individuals with epilepsy, requiring precise diagnosis, particularly in cases where surgical resection of the epileptogenic zone (EZ) is an option. Functional imaging techniques, such as ictal-interictal subtraction SPECT coregistered to MRI (SISCOM), have proven useful in pre-surgical evaluation by improving EZ localization accuracy. However, the widespread use of SISCOM is limited by the high costs and technical complexity of commercial software. Statistical Parametric Mapping (SPM) has been demonstrated to be a viable alternative for SISCOM analysis, displaying the potential for cost-effective EZ localization.

Materials and methods: In this retrospective study, we evaluated patients with pharmacoresistant temporal lobe epilepsy from two reference centers of epilepsy in Brazil, who underwent ictal and interictal SPECT imaging as part of their pre-surgical evaluation, achieving favorable outcomes (Engel I or II) after surgical resection. The EZ reference standard was determined according to anatomopathological findings and good clinical outcomes. SISCOM was performed using a semi-automated approach with Statistical Parametric Mapping (SPM) and a proprietary software - Analyze. Data from each method were compared to the EZ reference standard and classified as concordant, partially concordant, or discordant.

Results: We included 20 patients, 14 (70%) with left temporal lobe epilepsy and six (30%) with right temporal lobe epilepsy. Hippocampal sclerosis was the most common pathology (80%). Both SPM and Analyze were concordant with the EZ reference standard in 14 cases (70%), showing no difference in sensitivity between the methods. However, SPM generated smaller, more localized clusters, while Analyze produced larger clusters with broader spatial coverage. Concordance between the two methods was poor (Kappa = 0.0179), reflecting methodological differences.

Conclusion: This study evidences technical differences between SISCOM performed with SPM and Analyze, but with similar sensitivity (70%) for EZ localization. Further studies with larger sample sizes are required to confirm these findings. The data presented here suggest that SISCOM-SPM, due to its rapid and semi-automated workflow, may offer a practical and accessible alternative to proprietary software for epilepsy surgical planning.

Keywords: SISCOM; SPECT; epileptogenic zone; pharmacoresistant epilepsy; quantitative analysis; temporal lobe epilepsy.

Figure 1

Patient 4: 50-year-old female with left hippocampal sclerosis. Both methods (SPM and Analyze) demonstrated concordance with the gold standard, identifying the EZ in the left temporal lobe. (A) Axial slices from the SISCOM analysis using SPM, showing the largest cluster and the voxel with the highest intensity located in the left temporal lobe, concordant with the EZ. (B) Axial slices from the Analyze software, also identifying the largest cluster and the voxel with the highest intensity in the left temporal lobe, both concordant with the EZ.

Figure 2

Patient 17: 45-year-old female with focal cortical dysplasia (FCD) IB in the right temporal lobe. Neither method demonstrated concordance with the gold standard. (A) Axial slices from the SISCOM analysis using SPM, showing multiple clusters discordant with the EZ, which was located in the right temporal lobe. (B) Axial slices from the Analyze software, also displaying multiple clusters inconsistent with the EZ in the right temporal lobe.