“Why Did My Seizures Start Now?”: A Singular Question With Multiple Potential Answers

Abstract

Why Did My Seizures Start Now? Influences of Lesion Connectivity and Genetic Etiology on Age at Seizure Onset in Focal Epilepsy

Macdonald-Laurs E, Warren AEL, Leventer RJ, Harvey, AS. Epilepsia, 2024,65(6):1644–1657. https://doi.org/10.1111/epi.17947

Objective: Patients with focal, lesional epilepsy present with seizures at variable ages. Larger lesion size and overlap with sensorimotor or default mode network (DMN) have been associated with younger age at seizure onset in cohorts with mixed types of focal cortical dysplasia (FCD). Here, we studied determinants of age at seizure onset in patients with bottom-of-sulcus dysplasia (BOSD), a discrete type of FCD with highly localized epileptogenicity. Methods: Eighty-four patients (77% operated) with BOSD were studied. Demographic, histopathologic, and genetic findings were recorded. BOSD volume and anatomical, primary versus association, rostral versus caudal, and functional network locations were determined. Normative functional connectivity analyses were performed using each BOSD as a region of interest in resting-state functional magnetic resonance imaging data of healthy children. Variables were correlated with age at seizure onset. Results: Median age at seizure onset was 5.4 (interquartile range = 2–7.9) years. Of 50 tested patients, 22 had somatic and nine had germline pathogenic mammalian target of rapamycin (mTOR) pathway variants. Younger age at seizure onset was associated with greater BOSD volume (p = .002), presence of a germline pathogenic variant (p = .04), DMN overlap (p = .04), and increased functional connectivity with the DMN (p < .05, false discovery rate corrected). Location within the sensorimotor cortex and networks was not associated with younger age at seizure onset in our relatively small but homogenous cohort. Significance: Greater lesion size, pathogenic mTOR pathway germline variants, and DMN connectivity are associated with younger age at seizure onset in small FCD. Our findings strengthen the suggested role of DMN connectivity in the onset of FCD-related focal epilepsy and reveal novel contributions of genetic etiology.

Commentary

The title of the recent article by Macdonald-Laurs et al ¹ is a question that I frequently hear in my epilepsy clinic, particularly from patients with lesional adult-onset epilepsy: “Why did my seizures start now if I’ve had this all my life?” Emerging research in this area is beginning to shed light on the complex intersections of neurodevelopment and epileptogenesis. Age of seizure onset can be a genetic trait, with members of familial epilepsies presenting at similar ages. ² Epigenetic factors, related to timing of gene expression in different regions of the brain and at different stages of neurodevelopment (or senescence) likely also play a role. Recently, there has been interest in the functional connectivity of regions of focal cortical dysplasia (FCD), suggesting that age-related maturation of specific brain networks may be an important variable in the timing of epilepsy onset.^3,4

This study used a single pathological entity, bottom-of-sulcus dysplasia (BOSD) to focus on factors that may influence the age of seizure onset. ¹ A cohort of pediatric patients with BOSD was identified at a single hospital in Australia between 2005 and 2022. This cohort's clinical, electrographic, radiographic and genetic features were recently published; it is the largest cohort of BOSD described to date. ⁵ The authors used 84 of the 85 patients in the original cohort to examine characteristics associated with age of onset including genetic, neuroanatomic, and functional connectivity variables. The median age of seizure onset was 5.4 years (interquartile range [IQR] = 2.0–7.9). Twelve children (14%) had developmental delays at the time of seizure onset, and this was associated with earlier seizures (z = −2.75, p = 0.006). Sixty-five patients (77%) had surgery and 50 had genetic testing. Pathological subtype (FCD IIA vs. FCD IIB) was not associated with age of onset. Most of the genetic testing was research testing of resected brain tissue, with six individuals having pre-operative testing due to a family history of epilepsy. Nine cases were found to have germline pathogenic variants in mammalian target of rapamycin (mTOR) pathway genes (DEPDC5 −4, NPRL3 −4, TSC2 – 1). One patient with a DEPDC5 germline variant was also found to have a somatic “second-hit” variant in DEPDC5. Twenty-two cases were found to have somatic pathogenic variants only (21 in the gene MTOR and one in TSC2). Among the 31 patients with identified pathogenic variants, those with germline variants had an earlier age of onset than those with somatic variants only (z = 1.98, p = 0.04).

In addition to genetic and histologic variables, the authors examined neuroanatomical features that might be associated with the age of onset. The ratio of BOSD size to total brain volume was used for comparison between individuals. Larger relative BOSD volume was associated with younger age of onset (ρ= −0.32, p = 0.002). Sulcal depth, lesion laterality, and location were not associated with the age of onset.

Another focus of this article was network connectivity. Functional MRI (fMRI) was not available for all children in the study, so connectivity analysis was based on resting-state MRIs of a normative cohort of 109 children. BOSD segmentations were co-registered to each resting state fMRI in the normative set to obtain connectivity maps. These were then averaged together to determine the likely functional connectivity of each BOSD. Dominant network overlap was somewhat equally distributed between dorsal attention, frontoparietal control, ventral attention, somatomotor, and default mode networks (DMNs); involvement of the visual (4%) and limbic (1%) networks was rare. There was a correlation between the dominant network overlap and age of onset (X²(4) = 9.8, p = 0.04). The youngest median age of onset was associated with dominant default mode network (DMN) involvement [2 years (IQR 1.2–3.9)]. When specific networks were compared pairwise with correction for multiple comparisons, the DMN was associated with younger onset when compared to the dorsal attention network, but other pairwise comparisons were not significant. Finally, the authors mapped the regions of connectivity of individual BOSDs associated with early age of onset, demonstrating that they resembled the cortical regions in the DMN (medial prefrontal cortex, posterior cingulate, angular gyrus, temporal pole, and pre-cuneus). They used a Dice coefficient, a measurement of similarity between imaging data sets, to demonstrate that the DMN overlap associated with younger age of onset was much greater than that generated by random permutations.

The neuroimaging findings of this study are similar to other studies in confirming that lesion size, but not sulcal depth, is associated with age of onset.^3,4 The finding that DMN involvement is correlated with an earlier age of onset was also reported in one other pediatric FCD cohort. ³ Both groups of authors suggest that this network undergoes rapid growth during infancy which may be a reason why lesional epilepsies associated with it emerge earlier. However, in the multi-center epilepsy lesion detection (MELD) study, which evaluated pediatric and adult patients with FCD, having the DMN as a dominant network was associated with older, not younger, onset. ⁴ One possible reason for this discrepancy is that the MELD study included adult patients; The majority of patients with dominant lesional DMN connectivity had epilepsy onset before age 10, but there were individuals presenting at age 40–50. If there is a bi-modal peak to DMN-associated epileptogenic lesions, this may not be apparent in pediatric-focused studies. Another consideration for the difference is statistical: unlike the two pediatric studies, the analyses in the MELD study were multivariate and controlled for interactions between other factors that could impact both connectivity and age of onset. ⁴ In their multivariate model, lesion size and dominant network were both significantly associated with age of onset and interacted with each other. ⁴ Future studies will require large datasets to allow for assessments of these interactions.

In dissecting the age of onset question, future studies should also control for the influence of genetic factors on age of presentation. As demonstrated in this study by Macdonald-Laurs et al, ¹ germline pathogenic variants and the presence of developmental delay are both associated with earlier age of epilepsy presentation and it is possible that both may influence brain development and age of onset in ways that supersede network connectivity. As identification of somatic variants becomes easier, separating germline and somatic cohorts for analysis may be important. Furthermore, consideration of the variant allele frequency of somatic variants will be key as it will likely play a role in lesion size as well as age of onset.

In summary, this study of a single lesional pathology, BOSD, takes an important step forward in answering our patient's questions about why seizures present at various ages. It suggests several potential determinants of earlier onset epilepsy in children: the presence of germline mutations, developmental delay, lesion size, and lesion connectivity. The next phase of answering the question will be to figure out how these factors interact with each other to predict the age of onset in a given individual.