Cortical thickness analysis in operculo-insular epilepsy

Progressive remodeling of structural networks following surgery for operculo-insular epilepsy

Introduction

Operculo-insular epilepsy (OIE) is a rare condition amenable to surgery in well-selected cases. Despite the high rate of neurological complications associated with OIE surgery, most postoperative deficits recover fully and rapidly. We provide insights into this peculiar pattern of functional recovery by investigating the longitudinal reorganization of structural networks after surgery for OIE in 10 patients.

Methods

Structural T1 and diffusion-weighted MRIs were performed before surgery (t0) and at 6 months (t1) and 12 months (t2) postoperatively. These images were processed with an original, comprehensive structural connectivity pipeline. Using our method, we performed comparisons between the t0 and t1 timepoints and between the t1 and t2 timepoints to characterize the progressive structural remodeling.

Results

We found a widespread pattern of postoperative changes primarily in the surgical hemisphere, most of which consisted of reductions in connectivity strength (CS) and regional graph theoretic measures (rGTM) that reflect local connectivity. We also observed increases in CS and rGTMs predominantly in regions located near the resection cavity and in the contralateral healthy hemisphere. Finally, most structural changes arose in the first six months following surgery (i.e., between t0 and t1).

Discussion

To our knowledge, this study provides the first description of postoperative structural connectivity changes following surgery for OIE. The ipsilateral reductions in connectivity unveiled by our analysis may result from the reversal of seizure-related structural alterations following postoperative seizure control. Moreover, the strengthening of connections in peri-resection areas and in the contralateral hemisphere may be compatible with compensatory structural plasticity, a process that could contribute to the recovery of functions seen following operculo-insular resections for focal epilepsy.

Establishment of a normal control model of children's brain 18-fluorodeoxyglucose positron emission tomography and analysis of the changing pattern in patients aged 0-14 years

Background

It is difficult to obtain 18-fluorodeoxyglucose positron emission tomography (18FDG-PET) data from normal children, and changes in brain metabolism in children due to growth and development are poorly understood. For the first time, we established a normal control model of brain 18FDG-PET in children and evaluated its feasibility. The association of PET with age in children aged 0-14 years was analyzed. This study aimed to establish a normal control model of brain 18FDG-PET in children for the first time and to verify its feasibility, and to analyze the trend of PET with age in children aged 0-14 years.

Methods

In this retrospective cohort study, the 18FDG-PET imaging data of patients with no epileptiform discharge involvement contralateral to the epileptogenic zone were consecutively collected from January 2015 to June 2022 according to strictly defined screening criteria. For the normal control data, the hemisphere contralateral to the epileptogenic zone was mirrored and spliced to form an intact brain. The cohort of children aged 0-14 years was divided into 14 groups according age by year. Subsequently, patients who underwent lesionectomy with clear hypometabolism that roughly coincided with the extent of surgical resection were examined. The PET scan was compared with the control model, and the ratio of overlapping parts (hypometabolic areas ∩ surgical resection area) to hypometabolic parts (ROH) was calculated. Multiple regression analysis was performed on the normal control model for every 3- to 4-year age interval.

Results

A total of 159 normal control models were established. Five patients were randomly selected to verify the reliability of each yearly model. The average ROH was 0.968. Metabolism increasing with age in the different brain regions was observed at ages 0-2~, 3-5~, and 6-10 years. No age-related metabolic increase or decrease was found in the 10- to 14-year-old group. The metabolism in the 7- to 8-year-old group was higher than that in the 13- to 14-year-old group.

Conclusions

With strict screening criteria, the method of mirroring the contralateral hemisphere of the epileptic zone and splicing it into a complete brain as a means of creating a normal control group is feasible. The method offers convenience to the studies that lack healthy pediatric controls. Children under 10 years of age (especially 0-6 years old) experience considerable metabolic changes year on year. After the age of 10 years, the changes in metabolism gradually decrease, and metabolism also slowly decreases. Our findings provide guidance the clinical interpretation of areas with hypometabolism and emphasize the importance of establishing a normal control model of the child's brain, which should not be replaced by an adult model.

Identifying important factors for successful surgery in patients with lateral temporal lobe epilepsy

OBJECTIVE

Lateral temporal lobe epilepsy (LTLE) has been diagnosed in only a small number of patients; therefore, its surgical outcome is not as well-known as that of mesial temporal lobe epilepsy. We aimed to evaluate the long-term (5 years) and short-term (2 years) surgical outcomes and identify possible prognostic factors in patients with LTLE.

METHODS

This retrospective cohort study was conducted between January 1995 and December 2018 among patients who underwent resective surgery in a university-affiliated hospital. Patients were classified as LTLE if ictal onset zone was in lateral temporal area. Surgical outcomes were evaluated at 2 and 5 years. We subdivided based on outcomes and compared clinical and neuroimaging data including cortical thickness between two groups.

RESULTS

Sixty-four patients were included in the study. The mean follow-up duration after the surgery was 8.4 years. Five years after surgery, 45 of the 63 (71.4%) patients achieved seizure freedom. Clinically and statistically significant prognostic factors for postsurgical outcomes were the duration of epilepsy before surgery and focal cortical dysplasia on postoperative histopathology at the 5-year follow-up. Optimal cut-off point for epilepsy duration was eight years after the seizure onset (odds ratio 4.375, p-value = 0.0214). Furthermore, we propose a model for predicting seizure outcomes 5 years after surgery using the receiver operating characteristic curve and nomogram (area under the curve = 0.733; 95% confidence interval, 0.588-0.879). Cortical thinning was observed in ipsilateral cingulate gyrus and contralateral parietal lobe in poor surgical group compared to good surgical group (p-value < 0.01, uncorrected).

CONCLUSIONS

The identified predictors of unfavorable surgical outcomes may help in selecting optimal candidates and identifying the optimal timing for surgery among patients with LTLE. Additionally, cortical thinning was more extensive in the poor surgical group.

Predictors of outcomes after surgery for medically intractable insular epilepsy: A systematic review and individual participant data meta-analysis

Insular epilepsy (IE) is an increasingly recognized cause of drug-resistant epilepsy amenable to surgery. However, concerns of suboptimal seizure control and permanent neurological morbidity hamper widespread adoption of surgery for IE. We performed a systematic review and individual participant data meta-analysis to determine the efficacy and safety profile of surgery for IE and identify predictors of outcomes. Of 2483 unique citations, 24 retrospective studies reporting on 312 participants were eligible for inclusion. The median follow-up duration was 2.58 years (range, 0-17 years), and 206 (66.7%) patients were seizure-free at last follow-up. Younger age at surgery (≤18 years; HR = 1.70, 95% CI = 1.09-2.66, P = .022) and invasive EEG monitoring (HR = 1.97, 95% CI = 1.04-3.74, P = .039) were significantly associated with shorter time to seizure recurrence. Performing MR-guided laser ablation or radiofrequency ablation instead of open resection (OR = 2.05, 95% CI = 1.08-3.89, P = .028) was independently associated with suboptimal or poor seizure outcome (Engel II-IV) at last follow-up. Postoperative neurological complications occurred in 42.5% of patients, most commonly motor deficits (29.9%). Permanent neurological complications occurred in 7.8% of surgeries, including 5% and 1.4% rate of permanent motor deficits and dysphasia, respectively. Resection of the frontal operculum was independently associated with greater odds of motor deficits (OR = 2.75, 95% CI = 1.46-5.15, P = .002). Dominant-hemisphere resections were independently associated with dysphasia (OR = 13.09, 95% CI = 2.22-77.14, P = .005) albeit none of the observed language deficits were permanent. Surgery for IE is associated with a good efficacy/safety profile. Most patients experience seizure freedom, and neurological deficits are predominantly transient. Pediatric patients and those requiring invasive monitoring or undergoing stereotactic ablation procedures experience lower rates of seizure freedom. Transgression of the frontal operculum should be avoided if it is not deemed part of the epileptogenic zone. Well-selected candidates undergoing dominant-hemisphere resection are more likely to exhibit transient language deficits; however, the risk of permanent deficit is very low.

Safety of an operculoinsulectomy in the language-dominant hemisphere for refractory epilepsy

BACKGROUND

Operculoinsular cortectomy is increasingly recognized as a therapeutic avenue for perisylvian refractory epilepsy. However, most neurosurgeons are reluctant to perform this type of procedure because of feared neurological complications, especially in the language-dominant hemisphere, as the insula is involved in speech and language processes. The goal of this retrospective study is to quantify the incidence and types of speech and language deficits associated with operculoinsulectomies in the dominant hemisphere for language, and to identify factors associated with these complications.

METHODS

Clinical, imaging, and surgical data of all patients who had an operculoinsulectomy for refractory epilepsy at our center between 1998 and 2018 were reviewed. Language lateralization was determined by functional magnetic resonance imaging (fMRI) and/or Wada test. Speech and language assessments were carried out by neurosurgeons, neurologists, neuropsychologists and/or speech language pathologists, before surgery, during the first week after surgery, and at least 6 months after surgery.

RESULTS

Amongst 44 operculoinsulectomies, 13 were performed in the language-dominant hemisphere. 46% of these patients presented with transient aphasia post-surgery. However, a few months later, the patients' performances on language assessments were not statistically different from before surgery, thus suggesting a complete recovery of speech and language functions.

CONCLUSION

Temporary aphasias after operculoinsulectomy for refractory epilepsy in the language-dominant hemisphere are frequent, but eventually subside. Potential mechanisms underlying this recovery are discussed.

Structural Connectivity Alterations in Operculo-Insular Epilepsy

Operculo-insular epilepsy (OIE) is an under-recognized condition that can mimic temporal and extratemporal epilepsies. Previous studies have revealed structural connectivity changes in the epileptic network of focal epilepsy. However, most reports use the debated streamline-count to quantify ‘connectivity strength’ and rely on standard tracking algorithms. We propose a sophisticated cutting-edge method that is robust to crossing fibers, optimizes cortical coverage, and assigns an accurate microstructure-reflecting quantitative conectivity marker, namely the COMMIT (Convex Optimization Modeling for Microstructure Informed Tractography)-weight. Using our pipeline, we report the connectivity alterations in OIE. COMMIT-weighted matrices were created in all participants (nine patients with OIE, eight patients with temporal lobe epilepsy (TLE), and 22 healthy controls (HC)). In the OIE group, widespread increases in ‘connectivity strength’ were observed bilaterally. In OIE patients, ‘hyperconnections’ were observed between the insula and the pregenual cingulate gyrus (OIE group vs. HC group) and between insular subregions (OIE vs. TLE). Graph theoretic analyses revealed higher connectivity within insular subregions of OIE patients (OIE vs. TLE). We reveal, for the first time, the structural connectivity distribution in OIE. The observed pattern of connectivity in OIE likely reflects a diffuse epileptic network incorporating insular-connected regions and may represent a structural signature and diagnostic biomarker.

Widespread cortical dyslamination in epilepsy patients with malformations of cortical development

PURPOSE

Recent research in epilepsy patients confirms our understanding of epilepsy as a network disorder with widespread cortical compromise. Here, we aimed to investigate the neocortical laminar architecture in patients with focal cortical dysplasia (FCD) and periventricular nodular heterotopia (PNH) using clinically feasible 3 T MRI.

METHODS

Eighteen epilepsy patients (FCD and PNH groups; n = 9 each) and age-matched healthy controls (n = 9) underwent T1 relaxation 3 T MRI, from which component probability T1 maps were utilized to extract sub-voxel composition of 6 T1 cortical layers. Seventy-eight cortical areas of the automated anatomical labeling atlas were divided into 1000 equal-volume sub-areas for better detection of cortical abnormalities, and logistic regressions were performed to compare FCD/PNH patients with healthy controls with the T1 layers composing each sub-area as regressors. Statistical significance (p < 0.05) was determined by a likelihood-ratio test with correction for false discovery rate using Benjamini-Hochberg method.

RESULTS

Widespread cortical abnormalities were observed in the patient groups. Out of 1000 sub-areas, 291 and 256 bilateral hemispheric cortical sub-areas were found to predict FCD and PNH, respectively. For each of these sub-areas, we were able to identify the T1 layer, which contributed the most to the prediction.

CONCLUSION

Our results reveal widespread cortical abnormalities in epilepsy patients with FCD and PNH, which may have a role in epileptogenesis, and likely related to recent studies showing widespread structural (e.g., cortical thinning) and diffusion abnormalities in various human epilepsy populations. Our study provides quantitative information of cortical laminar architecture in epilepsy patients that can be further targeted for study in functional and neuropathological studies.