Duration-dependent hippocampal structural changes in focal epilepsy: multicenter neuroimaging evidence

BACKGROUND: Hippocampal sclerosis (HS) represents a higher proportion of surgical pathology in adult epilepsy patients compared to pediatric surgical cohorts. We hypothesized that while HS may serve as an initial pathological substrate, hippocampal abnormalities in epilepsy may also represent a duration-related structural change process that evolves with disease progression. To test this duration-dependent hypothesis, we investigated associations between disease duration and hippocampal morphological features using advanced neuroimaging in a large multicenter cohort. METHODS: This study included 705 patients with unilateral focal epilepsy (age 25.4 ± 14.9 years, disease duration 14.8 ± 13.6 years) and 424 healthy controls from three centers in China. Global hippocampal features (volume, thickness, gyrification, mean curvature, intrinsic curvature) and subfield volumes were extracted using HippUnfold and AID-HS. Z-scores were calculated using normative models controlling for age, sex, total intracranial volume, and site. Associations between disease duration and hippocampal features were assessed using Spearman correlations and duration-stratified group comparisons. Longitudinal changes were evaluated in 80 patients with serial MRI scans using linear mixed-effects models. RESULTS: Patients with epilepsy showed significant bilateral hippocampal abnormalities compared to controls, with more severe changes ipsilateral to the seizure focus. Disease duration significantly correlated with all ipsilateral hippocampal metrics: negatively with volume (ρ = -0.181, p < 0.001), thickness (ρ = -0.135, p < 0.001), and gyrification (ρ = -0.141, p < 0.001), and positively with mean curvature (ρ = 0.121, p = 0.003) and intrinsic curvature (ρ = 0.171, p < 0.001). Contralateral associations were observed for volume (ρ = -0.085, p = 0.044), thickness (ρ = -0.082, p = 0.049), and intrinsic curvature (ρ = 0.080, p = 0.049). Longitudinal analysis revealed significant annual decline in ipsilateral hippocampal volume (β = -0.21 Z-score units/year, p = 0.042). CONCLUSIONS: We provide novel neuroimaging evidence supporting the duration-related structural alteration hypothesis of hippocampal structural changes in focal epilepsy. These findings may help explain the higher proportion of HS observed in adult surgical cohorts compared to pediatric surgical cohorts and enhance understanding of epilepsy pathophysiology, suggesting that disease duration is significantly associated with hippocampal structural abnormalities, indicating that it may serve as an important reference indicator for clinical monitoring and therapeutic decision-making.