Probabilistic Cortical Mapping of Paediatric Epileptic Symptomatology

Probabilistic Cortical Mapping of Paediatric Epileptic Symptomatology: The Seizure Semiology-to-Brain Visualisation Tool (SVT) in Paediatric Presurgical Evaluation

Background:
Around 30% of children with epilepsy do not respond to antiepileptic medication and may be candidates for epilepsy surgery. If a seizure focus is identified, surgery can cure epilepsy, resulting in sustained improvements in quality of life (Dwivedi et al. 2017). The main aim of presurgical evaluation is to identify the brain region from which seizures arise in individuals with focal drug-resistant epilepsy.

Seizure semiology is the sequence of symptoms and signs that occur during a seizure and can help determine the cortical region of seizure onset (Tufenkjian and Lüders, 2012). Studies suggest that good detailed semiology is probably as good scalp EEG and MRI for localisation (Elwan et al., 2018), nevertheless the expert interpretation of semiology is largely subjective.

Recently developed open-source software, semiology-to-brain visualisation tool (SVT) - based on 309 journal articles and 4643 patients (Semio2Brain v.1.2.3, 2021, doi:10.5281/zenodo.4606589) – probabilistically maps descriptive seizure phenotypes to cortical epileptogenicity and displays results as 3-dimensional heatmaps (https://github.com/thenineteen/Semiology-Visualisation-Tool). SVT brings adult semiology that had otherwise been left to the clinician, into the same space as other investigative modalities such as structural imaging, and may ultimately improve epilepsy surgery outcomes.

Epileptic symptoms and signs are however different in children than in adults due to a developing brain and the inability of children to accurately report early symptoms. However, SVT contains few and unvalidated paediatric data. It is therefore important to clinically validate SVT for predicting seizure foci in children to give them the best chance at good long-term outcomes.

Aims/Objectives:
This project aims to map paediatric epileptic symptomatology to cortical epileptogenicity as SVT does for adults, by expanding its paediatric data using a large cohort of patients from London, then testing its accuracy in a cohort of children from Toronto.

Methods and Timeline:
1. Evaluation of current SVT (months 1-6)
First, we will retrospectively test the current SVT in 20 random paediatric patients from GOSH who underwent focal resections and subsequently became seizure free. Different SVT settings including all ages versus only paediatric datasets will be explored and results will be compared with the predictions of expert clinicians.

2. Expansion of SVT database (months 7-26)
225 patients from GOSH who underwent focal resections and subsequently became seizure free will be integrated into the database. Paediatric articles published after Jan 2019 will also be added to the database.

3. Evaluation of expanded paediatric SVT (27-32)
Expanded SVT will be tested in an independent dataset of 60 children with focal resections and subsequent seizure freedom from Toronto. To assess the effect of age on seizure semiology we will study three different age groups (20 patients 5-9yres, 20 patients 10-14 years and 20 patients 15-18yrs).

4. Publications and PhD thesis (33-36)
During the last 4 months publications and PhD thesis will be finalised.

Collaboration with University of Toronto:
During the research period (6 months) in Toronto 60 patients (20 patients 5-9yres, 20 patients 10-14 years and 20 patients 15-18yrs) with focal resections and subsequent seizure freedom will be identified. The expanded paediatric SVT will be tested on these patients, comparing the SVT results with the classification of expert clinicians. The PhD student will be locally supervised by Dr Elizabeth Donner and Dr Ayako Ochi.

References:
1.  Dwivedi et al.Surgery for Drug-Resistant Epilepsy in Children. N Engl J Med. 2017 Oct 26;377(17):1639-1647.
2.  Tufenkjian, K., Lüders, 2012. Seizure semiology: its value and limitations in localizing the epileptogenic zone. 8, 243-250.
3.  Elwan, S., et al., 2018. Lateralizing and localizing value of seizure semiology: Comparison with scalp EEG, MRI and PET in patients successfully treated with resective epilepsy surgery. Seizure. 61, 203-208.
4.  Alim-Marvasti, A., et al., 2020. Epileptogenic Zone Prediction from Seizure Semiology: A Data-Driven 3D-Visualisation Tool. American Epilepsy Society. Virtual (Seattle, USA). DOI:10.13140/RG.2.2.28373.65769.