Central Nervous System Findings on Magnetic Resonance Imaging in Children with Epilepsy

Epilepsy is a relatively frequent pathology in child populations, with an annual incidence rate ranging from 41 to 67 per 100 000 cases1-7. In the past decade, many important advances related to epilepsy have arisen. On one hand, new diagnostic methods (video electroencephalography monitoring, structural and/or functional neuroimaging, metabolic and genetic analysis, etc) facilitate the diagnosis and the syndromic classification; on the other hand, the availability of the new antiepileptic drugs is conditioning new therapeutic possibilities8-11. The International League Against Epilepsy classification for epilepsy and epileptic syndromes suggests prognosis consequences, because it allows estimating outcome and therapeutic responses12-15. In this way, structural neuroimaging plays an important role in the evaluation, management, and treatment of the child with epilepsy. The role of neuroimaging is to detect an underlying cerebral lesion that may be causally related to the child’s seizure disorder. Magnetic resonance imaging (MRI) is the elective of all structural imaging tool. MRI is considerably superior to X-ray CT in terms of its sensitivity and specificity for identifying subtle abnormalities. The principal role of MRI is in the definition of the structural abnormalities that underlie seizure disorders (tumours, malformations of cortical development, hippocampal sclerosis, neurocutaneous diseases, vascular malformations, traumatic lesions, strokes, residual lesions, etc.) and to contribute to the aetiologic diagnosis and classification of the different epilepsies and epileptic syndromes, and thereby provide an accurate prognosis for patients11,16-18. In Spain, there is limited information about its current use in the initial evaluation of pediatric epilepsy and about its performance during the initial diagnosis of epilepsy, which suggests that a descriptive study of MRI findings in epilepsy in children would be necessary. The aim of this study is to analyze the epidemiological characteristics and proportional distribution of the epilepsy and epileptic syndromes in children, and to describe the central nervous system findings on MRI in these patients.


Magnetic resonance imaging
MRI was performed in all patients at the onset of diagnosis according to a standardized paediatric seizure protocol.The standardized pediatric seizure protocol consisted of the following scanning sequences: -T2-weighted fast spin-echo (FSE) images and proton density-weighted in axial plane with 5 mm thick slices.-T2-weighted Fluid attenuated inversion recovery (FLAIR) images in the coronal plane, with 5 mm thick slices.-T1-weighted FLAIR images in the sagittal plane, with 5 mm thick slices.-T1-weighted FLAIR images in the axial plane, with 3 mm thick slices.-Diffusion-weighted (DW) images in the axial plane.Complementary sequences: -T2-weighted gradient-echo images in the axial plane, with 5 mm thick slices, for the detection of calcifications or hemosiderine deposits.-T2-weighted FSE images in the coronal plane, with 3 mm thick slices, for the study of temporal lobe seizure foci.
The scoring system used for MRI examinations in this study classifies various abnormal features by location 20 .Abnormal features include volume loss, leukomalacia/gliosis, encephalomalacia, other white-matter lesions, heterotopia, cortical dysplasia, other graymatter lesions, mass lesions, vascular lesions, ventricular enlargement, prominence of the extra-axial spaces, hippocampal atrophy or signal abnormality, and others structural abnormalities.
Abnormalities on MRI were classified as either significant or nonsignificant based on a system of classification previously developed [21][22][23] .A significant abnormality was defined as a MRI finding reasonably likely to be related to seizure disorder.These were defined as at least one of the following: leukomalacia/gliosis, encephalomalacia, any gray-matter lesion, mass lesion, hemorrhage, vascular lesion, hippocampal abnormality, ventricular enlargement >1.5 cm, or prominence of extra-axial fluid spaces >1.0 cm.Some potential overlap in scoring encephalomalacia and leukomalacia/gliosis was recognized.Encephalomalacia was generally thought to signify ''cystic encephalomalacia'' and to involve cortical regions.If there was some signal abnormality consistent with gliosis adjacent to an area of encephalomalacia, this was scored as encephalomalacia.If there were other discrete areas of abnormal signal consistent with gliosis, these were scored as (additional) areas of leukomalacia/gliosis.All of these studies were read and coded by two neuroradiologists with experience involved in this study.Rater was blind to other rater scoring and to other clinical data (e.g., electroencephalogram findings), except for knowing that the child had presented with a seizure.
The computer program SPSS 17.0 for Windows (Chicago, Illinois, USA) was used to perform the statistical analysis (descriptive study).

Results
Age distribution was 78 infants (from 1 to 12 months of age), 157 in early childhood (from 1 to 6 years), 134 school-aged children (from 6 to 10 years), and 88 adolescents (from 10 to 15 years).

Infants
The sample consisted of 78 patients (37 males and 41 females).In infants, most of the seizures were symptomatic (53.8%), whereas others were idiopathic (25.6%) or cryptogenic (20.5%).Table 1 shows the distribution of the epilepsies and epileptic syndromes in infants.
Epilepsies attributed to and organized by structural or metabolic conditions (34.6%) and West syndrome (30.8%) were the most prevalent syndromes.
Epilepsies attributed to and organized by structural or metabolic conditions (n=27) were secondary to perinatal asphyxia (figure 1), purulent meningitis, congenital malformations of the brain (figure 2), Aicardy syndrome (figure 3), inherited metabolic disorders, phakomatoses (figure 4) and venous thrombosis.Among patients diagnosed with West syndrome, etiology was symptomatic in 62.5% of the cases (n=15); this means, 6 cases related to perinatal ischaemic or anoxic lesions, 3 cases with Down syndrome, 2 cases with tuberous sclerosis, and 1 case related with pathologies as subcortical band heterotopia (figure 5), semilobar holoprosencephaly, neonatal meningoencephalitis (figure 6) and perinatal intracranial hemorrhage.Of 78 infants in this study, 42 (53.8%)demonstrated at least one MRI abnormality.Thirtythree (42.3%) manifested MRI abnormalities that were classified as significant because they were considered to be potentially related to the seizure condition.Boldface rows represented "significant abnormalities" Table 2. MRI abnormalities found in infants diagnosed with epilepsy

Early childhood
The sample was made up of 157 patients (85 males and 72 females).In early childhood, distribution was idiopathic (44,6%), cryptogenic (29.9%) or symptomatic (25.5%).Table 3 shows the distribution of the epilepsies and epileptic syndromes in this period.Epilepsies in which nature of the underlying causes is yet unknown (29.9%) and epilepsies attributed to and organized by structural or metabolic conditions (24.8%) were the most prevalent syndromes.

Central Nervous System Findings on Magnetic Resonance Imaging in Children with Epilepsy 225
Epilepsies associated to structural or metabolic conditions (n=39) were secondary to perinatal asphyxia (figure 7), cerebral infections (figure 8), congenital malformations of the brain, inherited metabolic disorders, phakomatoses (figure 9), chromosomal abnormalities, arterial infarction (figure 10), arteriovenous malformations and supratentorial tumors.Of 157 early infants in this study, 45 (28.7%) had at least one MRI abnormality.Thirty-two (20.4%) showed MRI abnormalities that were classified as significant because they were considered to be potentially related to the seizure condition.Boldface rows represented "significant abnormalities" Table 4. MRI abnormalities found in early childhood diagnosed with epilepsy

School-aged children
The sample was made up of 134 patients (72 males and 62 females).In school-age children, seizures were idiopathic in 55.2%, cryptogenic in 25.4% and symptomatic in 19.4%.Table 5 shows the distribution of the epilepsies and epileptic syndromes in school-aged children.
Epilepsies in which the nature of the underlying causes is yet unknown (29.1%), benign epilepsy with centrotemporal spikes (25.4%) and childhood absence epilepsy (18.7%) were the most prevalent syndromes.
Epilepsies attributed to structural or metabolic conditions (n=39) were secondary to perinatal asphyxia, cerebral infections, congenital malformations of the brain (figure 11), phakomatoses, arterial infarction, arteriovenous malformations (figure 12), suprasellar arachnoid cyst (figure 13), drug toxicity (figure 14) and desmoplastic neuroepithelial tumors.After going through the records of the 134 school-aged children, we found that 28 (20.9%) had at least one MRI abnormality recorded.Twenty-one (15.7%) presented MRI abnormalities that again were considered as significant because they were potentially related to the seizure condition.Boldface rows represented "significant abnormalities" Table 6.MRI abnormalities found in school-aged childhood diagnosed with epilepsy

Adolescents
The sample was consisted of 88 patients (39 males and 49 females).In this group of patients, seizures were idiopathic in 69.3%, cryptogenic in 19.3% and symptomatic in 11.4% of them.Table 7 shows the distribution of the epilepsies and epileptic syndromes in adolescents.Idiopathic generalized epilepsies with variable phenotypes (46.6%) and cryptogenic focal epilepsies and/or unknown cause (19.3%) were the most prevalent syndromes.
Epilepsies associated to structural or metabolic conditions (n=9) were secondary to perinatal asphyxia, malformations of the brain, strokes (figure 15), vascular malformations, chromosomal abnormalities and ependimal cyst (figure 16).Boldface rows represented "significant abnormalities" Table 8.MRI abnormalities found in adolescents diagnosed with epilepsy

Whole sample
The whole group was 457 patients (233 males and 224 females).Etiology was considered idiopathic in 225 (49.2%), cryptogenic in 114 (24.9%), and symptomatic in 118 (25.8%).Table 9 shows the distribution of the epilepsies and epileptic syndromes in the overall sample.
There was at least one MRI abnormality in 134 patients (29.3%).One hundred (21.9%) had significant MRI abnormalities (potentially related to the seizure condition).Table 10 presents the number of children with MRI imaging abnormalities.As we previously said, the sum of abnormalities may exceed the total number of children, since different abnormalities can affect the same locations in one given patient.Of 134 children with at least one MRI abnormality, the most common abnormalities included white-matter lesions (27.6%), volume loss (19.6%), gray-matter lesions (19.6%) and ventricular enlargement (12%).  .Although there are no data about the prevalence of infantile epilepsy in Spain, the extrapolation of published data 7,[24][25][26] allows estimating that the selected population would account for a high percentage of epilepsy diagnosed children in Navarra.That means, it is a wide enough population to consider as representative for this territory, and, also, the results would contribute to determining the epidemiological characteristics of epilepsy in children.A routine RMI has been requested in all patients with epilepsy as it has been our practice for several years in this Pediatric Neurology Unit in order to aid the formulation of syndromic and etiological classification and, therefore, there is no reason to suspect any statistical bias in the results and conclusions.
According to the data, we confirm that age seems to represent a main factor in the clinical expression of the different childhood epilepsy phenotypes.In fact, relative distribution of the different seizure types and epileptic syndromes was different for each age group, as seen in our series.For example, some epileptic syndromes were predominant in some age groups, such as epilepsies associated to structural or metabolic conditions and West syndrome in infants, epilepsies in which the nature of the underlying causes is yet unknown and epilepsies associated to structural or metabolic conditions in early childhood, or absence and focal benign epilepsies with centrotemporal spikes in school-aged children and idiopathic generalised epilepsies with variable phenotypes in adolescents.Thus, when referring to childhood epilepsy, it would be useful to make reference to the age group of the patients, because characteristics for the epilepsies and epileptic syndromes during childhood seem to be close to the structural and functional changes in the brain since birth to adolescence [27][28][29][30] .
Our major finding of this study was that the use of MRI at diagnosis of epilepsy in children demonstrated a high rate of abnormalities findings.There was an inverse relationship between the prevalence of significant MRI abnormalities and age at diagnosis.Thus, the prevalence in the first year of life was 42.3%, in early childhood was 20.4%, in schoolchildren was 15.7%, and in adolescents was 15.9%.White-matter lesions were more commonly associated with epilepsies and epileptic syndromes in children at any age, whereas many other children displayed epileptogenic lesions involving the cortex or gray matter.In addition, volume loss, considered a priori to be a non significant abnormality, was also very frequent in this population.Ventricular enlargement, especially mild, was also fairly frequent (12% of total sample).Structural neuroimaging is recommended for all children with recently diagnosed localization-related or generalized epilepsy who do not have the clinical and electrographic features characteristic of classical idiopathic focal or generalized epilepsy (these include benign epilepsy with centrotemporal spikes, childhood absence epilepsy, juvenile absence epilepsy, and juvenile myoclonic epilepsy) 16,[31][32][33] .However, this series shows how structural abnormalities judged to be etiologically related to the seizure conditions have been found in idiopathic focal or generalized epilepsy, which are consistent with other published data 23,34,35 .This possibility suggests that clinicians should consider obtaining structural neuroimaging with an MRI when feasible for all patients with epilepsy, because MRI abnormalities may be present even in children with apparently benign syndromes or cryptogenic seizures, and the discovery of an epileptogenic lesion has potential implications for diagnosis, prognosis and treatment.Our data cannot be easily compared with those in other studies because the populations are somewhat different 23,[31][32][33]36 . Ourstudy attempted to acquire MRI scans on all children, whereas previous studies excluded some syndromes or children with seizure onset in infancy and early childhood.Therefore, this series extends earlier work by other investigators in several ways.First, imaging was limited to MRI, the current anatomic ''gold standard''.Second, imaging was performed at diagnosis.Third, we applied a standardized classification system to MRI findings which permits to present a comprehensive and systematic description of imaging findings.Finally, data were coded directly from magnetic resonance images by two neuroradiologists who read the images firsthand.
In conclusion, the use of MRI and a reliable standardized scoring system in a large sample of children after their diagnosis of epilepsy identified a high rate of abnormalities (134 of 457, 29.3%).This may have important implications for practice guidelines in this population.First, some findings that might have been regarded as incidental in the past (e.g., volume loss/ventricular enlargement, and others white-matter abnormalities) appear to be present at the onset of seizures and may therefore be clinically significant.Second, the detection of abnormalities in epileptic syndromes other than localization-related symptomatic/cryptogenic syndromes supports an argument for routine MRI at the onset of any seizure condition.

Fig. 5 .
Fig. 5. Band heterotopia: FLAIR T1 images show a thick band of heterotopic grey matter, which is separated from the cortex by a thin layer of myelinated white matter.

Fig. 6 .
Fig. 6.Frontal bilateral porencephalic lesions as sequelae of meningitis.SE T2 image shows two large frontal porencephalic lesions, being the left one comunicated with the frontal horn.

Fig. 7 .Fig. 9 .
Fig. 7. Porencephalic cyst: T2-weighted SE image displays a wide cyst in the left frontal lobe, which is connected to the lateral ventricle and causes a slight lateral midline displacement of the brain.

Fig. 10 .
Fig. 10.Post-stroke sequelae in the right middle cerebral artery: T2-weighted SE image showing residual encephalomalacia with volume loss and ex vacuo enlargement of the right ventricle.An area of periventricular leukomalacia in the left occipital lobe can be appreciated.

Fig. 11 .
Fig. 11.Subependymal heterotopia: T2-weighted SE image shows a gray matter nodule protruding over the occipital horn of the right ventricle.

Fig. 15 .
Fig. 15.Right cerebral hemisphere atrophy after ischemic lesion: T2-weighted SE image presents a volume loss in the right hemisphere, atrophy of cortical gyri, leukomalacia and ex vacuo enlargement of lateral ventricle.A hyperintense nodule, compatible with spongiotic intramyelinic change (neurofibromatosis), can be appreciated in the left globus pallidus.

Fig. 16 .
Fig. 16.Ependymal cyst: T2-weighted SE image shows a cystic lesion that expands the left occipital horn and a small brain herniation inside the interpeduncular cistern.

Table 4
presents the number of early infants with MRI abnormalities (the sum of abnormalities may exceed the total number of children because different abnormalities can affect the same locations for a given patient).In this group of 45 children with at least one MRI abnormality, the most common abnormalities included white-matter lesions (28.7%), volume loss (19.1%), gray-matter lesions (21.3%) and ventricular enlargement (14.9%).

Table 3 .
Distribution of the different epilepsies and epileptic syndromes in early childhood (n=157)

Table 6
presents the number of school-age children with MRI abnormalities.From the 28 children with at least one MRI abnormality, the findings included white-matter lesions (23.4%), gray-matter lesions (23.4%), volume loss (17%) and prominence of extra-axial fluid space (12.8%)

Table 5 .
Distribution of the different epilepsies and epileptic syndromes in school-aged children (n=134 )

Table 7 .
Distribution of the different epilepsies and epileptic syndromes in adolescents

Table 9 .
Distribution of the different epilepsies and epileptic syndromes arranged by age at onset (n=157)

Table 10 .
MRI abnormalities found in children diagnosed with epilepsy

Table 11 .
Distribution of MRI abnormalities found according to age