Abstracts

Rationale: A machine learning model for estimating chronological age from structural MRI scans has shown promise as a biomarker of overall brain health. Increased brain-predicted age (relative to actual age) indicates accelerated aging or higher cumulative exposure/sensitivity to pathological brain insults, in contrast to brain resilience. We have recently reported increased Brain-Age-Gap-Estimation (BrainAGE) in patients with refractory focal epilepsy and that BrainAGE is reduced after surgical intervention (De Bezenac et al., 2021), which further highlighted the advantages of resective epilepsy surgery on overall brain health. However, there has been no work to date that has investigated the effects of a first seizure on overall brain health. In this study we investigated the effects of a first unprovoked seizure on BrainAGE as a proxy of overall brain health.

Methods: We studied 68 well-characterised adult patients who had a first unprovoked seizure and were in a specialist epilepsy centre. Patients were followed up for an average of 4.8 years following their index seizure. We also studied 50 age- and sex-matched healthy controls. All patients and controls had 3D T1-weighted MRI performed. Brain-predicted age was computed from T1-weighted MRI scans using the brainAgeR analysis pipeline (De Bezenac et al., 2021). A trained Gaussian processes regression (GPR) model following tissue segmentation, vectorization and principal component analysis (PCA)based dimension reduction was performed for each scan. BrainAGE was calculated as the brain-predicted age minus chronological age at the time of the MRI scan. A multivariate analysis controlling for sex, age, grey matter, white matter, and cerebrospinal fluid (CSF) brain volume was carried out.

Results: There was a significant increase in mean BrainAGE in patients (42.7 [95% CI = 38.9, 46.5]) in comparison to healthy controls (32.1 [95% CI = 29.4, 34.8]; F=9.46, p 0.003). Patient sub-group analysis revealed no statistically significant difference in mean BrainAGE between patients who had seizure recurrence compared to those who did not have any subsequent seizures (F=1.39, p 0.243).

Conclusions: Morphological brain alterations linked to accelerated aging in epilepsy have been previously reported and our findings suggest that at least some changes in overall brain health may even be present as early as the index seizure. Further work is required to understand the biology underpinning these changes and whether early and appropriate treatment can potentially reverse premature brain aging.

Funding: Please list any funding that was received in support of this abstract.: Nil.