Patients’ (pre-operative) handedness was coded as one of five possible groups: Left, Right, Ambidextrous, or No clear hand preference.
Patients’ (pre-operative) educational status was coded as one of six possible groups: Mainstream school, Mainstream school with additional support, Special educational needs school, Home schooled, Residential home, or Not of school age.
Age at first seizure was a continuous variable used to capture the age at which the patient experienced their first seizure. This could have been an afebrile or febrile seizure.
Age of epilepsy onset was a continuous variable used to capture the age at which the patient was diagnosed with epilepsy. The variables “Age at first seizure” and ”Age of epilepsy onset” were kept distinct to account for early, isolated occurrences of febrile seizure(s), which did not result in a diagnosis of epilepsy.
Patients were classified as either having or not having a history of febrile seizures.
Patients were classified as either having or not having a family history of epilepsy. No criteria were applied in terms of how closely related the affected family member(s) needed to be.
Patients’ antiseizure medication was characterised in terms of:
- the number of antiseizure medications the patient was receiving at time of pre-surgical evaluation, and
- the total number of different antiseizure medications the patient had trialled from epilepsy onset to pre-surgical evaluation.
Age at surgery was a continuous variable used to capture the age at which the patient underwent their first resective or disconnective surgery for epilepsy.
Duration of epilepsy was a continuous variable used to capture the time elapsed between the patient’s epilepsy onset and surgery (“Age at surgery” - “Age of epilepsy onset” = “Duration of epilepsy”).
Patients’ seizure history as well as their seizures, seizure semiology, and EEG recordings at the time of pre-surgical evaluation were retrospectively characterised by four experienced paediatric neurophysiology consultants: FM, KD, CE and GC. The consultants had access only to the patients’ pre-operative information (i.e. they did not have access to the patients’ seizure outcomes).
The consultants coded whether patients had a history of: Infantile seizures, Electrical status epilepticus during slow-wave sleep (ESES), Status epilepticus, and Generalised tonic-clonic seizures.
The consultants coded the patients’ seizures at the time of pre-surgical evaluation: Spasms, Number of seizure types, Focal versus generalised onset (all focal/all generalised/mixture of focal and generalised), Motor versus non-motor (all motor/all non-motor/mixture of motor and non-motor), and Aware versus impaired awareness (all aware/all impaired awareness/mixture of aware and impaired awareness).
The consultants coded the patients’ seizure semiology at the time of pre-surgical evaluation: Localising + lesion concordant, Lateralising + lesion concordant, and Non-concordant.
The consultants coded the patients’ ictal and interictal EEG recordings at the time of pre-surgical evaluation: Localising + lesion concordant, Lateralising + lesion concordant, and Non-concordant.
Patients’ pre-operative MRI scan was classified as either abnormal (MRI positive) or normal (MRI negative). MRI scans classified as abnormal were further classified as comprising a unilateral (right- or left-sided) abnormality or a bilateral abnormality.
In addition to MRI bilaterality, MRI findings were characterised as one of fifteen possible diagnoses.
- FCD-II (focal cortical dysplasia type II)
- FCD-NOS (focal cortical dysplasia not otherwise specified)
- LEAT (low-grade epilepsy-associated tumour, including ganglioglioma, dysembryoplastic neuroepithelial tumour, pilocytic astrocytoma, and papillary glioneuronal tumour)
- LGL (low-grade lesion not otherwise specified, comprising possible diagnoses focal cortical dysplasia, mesial temporal sclerosis, and low-grade epilepsy-associated tumour)
- MCD-Other (including hypothalamic hamartoma, hemimegalencephaly, polymicrogyria, and focal cortical dysplasia type III)
- MTS (mesial temporal sclerosis, including hippocampal sclerosis, neuronal loss in the hippocampus/temporal lobe, gliosis in hippocampus/temporal lobe, granule cell dispersion, and Chaslin’s gliosis)
- Negative
- N-LEAT (non-low-grade epilepsy-associated tumour, including astrocytoma, neurocytoma, and pleomorphic xanthoastrocytoma)
- NSC (non-specific epilepsy-associated changes)
- Not determined lesion/abnormality
- Rasmussen encephalitis
- Scarring (as result of an acute vascular event, i.e. hemorrhagic infarction, trauma, or infection in the past),
- TS (tuberous sclerosis)
- Unspecified tumour
- Vascular diagnoses (including cavernoma, meningioangiomatosis, as seen in Sturge-Weber syndrome, and arterio-venous malformations)
All genetic tests performed in children who underwent surgery were retrospectively retrieved. The reported genetic variants were reviewed by APC, an experienced clinical scientist with a specialization in molecular genetics. Variant classification was carried out using Alamut Visual version 2.15 (SOPHiA GENETICS, Lausanne, Switzerland). Population data (https://gnomad.broadinstitute.org/), in silico tools predictions, functional analysis, segregation studies, and previous reports were taken into account. Variants were classified using the American College of Medical Genetics (ACMG)/Association for Molecular Pathology (AMP) guidelines (Richards et al., 2015) and Association for Clinical Genomic Science (ACGS) Best Practice Guidelines for Variant Classification in Rare Disease 2020 (https://www.acgs.uk.com/quality/best-practice-guidelines/).
Findings were discussed with LM, a clinical geneticist, and AM, a consultant paediatric neurologist. Variants were classified into five categories as per standard ACMG guidelines, namely: 1) benign, 2) likely benign, 3) uncertain significance, 4) likely pathogenic, and 5) pathogenic. We considered patients with variants classified as pathogenic, likely pathogenic and variants of uncertain significance with a suspicion of being pathogenic to have a genetic cause of epilepsy, in accordance with ACGS Best Practice Guidelines for Variant Classification in Rare Disease 2020 (https://www.acgs.uk.com/quality/best-practice-guidelines/). The clinicians did not have access to the patients’ seizure outcomes during classification of genetic information.
Surgeries were classified by visual inspection of pre- and post-operative MRI scans and retrieval of medical records describing the pre-operative intent, surgical procedure, and post-operative discharge. Surgeries were characterised according to: Type of procedure performed, Side operated on, and Lobe operated on.
Type of procedure performed included four categories: Lesionectomy, Lobectomy, Disconnection and Hemispherotomy. Procedures comprising more than one type of procedure, so-called “Combined procedures”, were reported as such, but re-classified for modelling purposes. For re-classification, the more extensive procedure was used. For example, a surgery comprising both a Lobectomy and Lesionectomy would be re-classified as a Lobectomy.
Side operated on included two categories: Right and Left. When classifying Side operated on, “Not applicable” was assigned to focal resections that involved the removal of a hypothalamic hamartoma.
Lobe operated on included six categories: Temporal, Frontal, Parietal, Occipital, Insular and Multilobar. When classifying Lobe operated on, “Not applicable” was assigned to hemispherotomy procedures as well as focal resections that involved the removal of a hypothalamic hamartoma.
Histopathology diagnosis was grouped into one of thirteen categories.
- FCD-II (focal cortical dysplasia type II)
- FCD-NOS (focal cortical dysplasia not otherwise specified)
- LEAT (low-grade epilepsy-associated tumour, including ganglioglioma, dysembryoplastic neuroepithelial tumour, pilocytic astrocytoma, and papillary glioneuronal tumour)
- M-MCD (mild malformation of cortical development, including focal cortical dysplasia type I)
- MCD-Other (including hypothalamic hamartoma, hemimegalencephaly, polymicrogyria, and focal cortical dysplasia type III)
- MTS (mesial temporal sclerosis, including hippocampal sclerosis, neuronal loss in the hippocampus/temporal lobe, gliosis in hippocampus/temporal lobe, granule cell dispersion, and Chaslin’s gliosis)
- N-LEAT (non-low-grade epilepsy-associated tumour, including astrocytoma, neurocytoma, and pleomorphic xanthoastrocytoma)
- Normal result
- NSC (non-specific epilepsy-associated changes)
- Rasmussen encephalitis
- Scarring (as result of an acute vascular event, i.e. haemorrhagic infarction, trauma, or infection in the past),
- TS (tuberous sclerosis)
- Vascular diagnoses (including cavernoma, meningioangiomatosis, as seen in Sturge-Weber syndrome, and arterio-venous malformations)