JLE

Epileptic Disorders

MENU

MRI in rolandic epilepsy Volume 2, supplément 4, Supplément 1, Décembre 2000

Magnetic resonance imaging (MRI) studies in a group of children with typical rolandic epilepsy (RE) has never been performed. However, there are occasional case reports of apparent RE in which clear structural lesions on MRI have been found. Ambrosetto [1] described a case with unilateral opercular macrogyria, Sheth et al. [2] a case with cortical dysplasia, and Stephani and Doose [3] a case with posttraumatic bilateral multiple brain lesions. In addition de Saint-Martin et al. [4] reported a case with normal MRI but with bilateral hypermetabolism in opercular regions on positron emission tomography. Shevell et al. [5] documented a mass lesion in five children with RE with atypical features. They were designated "pseudo-BECRS".

Stimulated by the finding of a distinct hippocampal (HC) asymmetry on MRI in a child with genuine RE and recurrent seizures in spite of antiepileptic drug treatment, an investigation to look for the occurrence of MRI abnormalities in other children with the syndrome was undertaken [6].

Materials and methods

All children with typical RE, based on both characteristic seizure manifestations and typical EEG (rolandic sharp waves), seen in our epilepsy clinic between September 1995 and May 1998 were included in this study. The series comprised 18 children, 9 girls and 9 boys, aged 6 to 12 years. Age at seizure onset varied between 3 and 10 years (median 6 years). All children were mentally and neurologically normal. Sixteen of the children were right-handed and two were left-handed. All but one child were treated with antiepileptic drugs (carbamazepine or valproate). The seizures were strictly nocturnal in 15 cases, diurnal in one case and both nocturnal and diurnal in two cases. Ten children had partial seizures with secondary generalization, five children were observed to have nocturnal generalized tonic-clonic seizures, and in three children there were strictly partial seizures. The series included two pairs of siblings. One child had a history of simple febrile seizures and two, a history of migraine. A family history of epilepsy was found in six cases (33%), febrile seizures in two cases (11%), and migraine in 12 cases (66%).

MRI was performed with an imager operating at 1.5 T. Sagittal T1-weighted spin echo (SE) images were obtained in 5 mm sections with 0.5 mm gaps between the slices. Oblique axial images were obtained in the plane of the long axis of the HC. Proton density- and T2-weighted SE images were 4 mm thick with 0.4 mm interslice gaps. Turbo inversion recovery (IR) images were 2.0-3.5 mm thick. There were no gaps between the IR images. Oblique coronal images were obtained in a plane perpendicular to the long axis of the HC and there were no gaps between the slices. T2-weighted turbo SE images were 2.5 mm thick, and turbo IR images were 2.0-3.5 mm thick. The images were reviewed by an experienced neuroradiologist (RR), who had no knowledge of the clinical histories or the EEG patterns.

The interval between the first seizure event and the MRI examination varied from 4 months to 6 years 9 months (median 2 years 2 months). All children had been seizure-free at least one month preceding the examination. Repeated MRI after two years have so far been performed in six children.

Results

Out of the 18 children hippocampal asymmetry was obvious in five (28%). The left HC was smaller than the right in three and the right smaller than the left in two cases. The smaller HC was ipsilateral to the spike lateralization and contralateral to the clinical manifestations in all of them. In two of these children, there was a high signal intensity in the affected HC, diffuse in one, and small focal in three adjacent slices in the other one. The first also had a heterotopic nodule in close connection, and rostral to the contralateral frontal horn, and a subtle, hyperintense linear extension from the nodule towards the cortex. In addition, one child with HC asymmetry showed a Chiari 1 malformation. One child had a high signal focus seen in two adjoining slices of the right hippocampus without obvious HC asymmetry. Thus, altogether six children (33%) showed some HC abnormality, and they all presented ipsilateral EEG findings.

Five children showed local high signal intensities (T2-weighted images) just under the cortex of the frontal and temporal lobes. One of them also had an HC asymmetry. Thus, localized high signal intensities were found in eight children (44%). In eight children (44%), the MRI examination was normal. In the six children with MRI follow-up, the findings remained unchanged.

Discussion

The present study group comprised children with the genuine RE syndrome; they presented with typical seizure manifestations and a typical EEG. They were also neurologically normal. The case reports by Ambrosetto [1], Stephani and Doose [3] and de Saint-Martin et al. [4] must be regarded as atypical cases in conformity with the report by Shevell et al. [5]. That atypical clinical and electrographic findings is the rule rather than the exception in RE was discussed by Wirrell et al. [7], who presented a series of 42 children. MRI was not performed, but computed tomographic scans were performed in 15 of them with normal findings. These results may confuse the whole definition of RE, but at the same time it strengthens the need for a comprehensive discussion of the syndrome.

Concerning the HC abnormalities found in our study, they were all ipsilateral to the main EEG finding. Thus, it may be of interest to consider the function of an abnormal HC in epileptogenesis, and especially its relation to epilepsy arising from the sylvian cortex. Apparently HC has a central role in regulating different processes such as motor and sensory functions as well as memory and cognition, due to its afferent and efferent pathways, the so-called HC circuit system. This complicated system has been thoroughly discussed by Braak et al. [8].

The presence of subcortical high signal intensities on T2-weighted images in both temporal and frontal gyri of five children is difficult to explain. Focal high signal intensities in the subcortical white matter are common in focal cortical dysplasia. In this condition blurring of the gray-white matter junction, thickening and signal intensity abnormalities of the covering cortex, as well as gyral enlargement are characteristic findings in the histologically confirmed cases [9-12]. Similar findings can also be seen in tuberous sclerosis [13]. One of the causes of the MRI characteristics may be a dysmyelination [10]. In our series the subcortical, high signal areas were covered by a regular cortex with normal thickness and signal intensity and there was no gyral enlargement or other deformity. Thus, these hyperintensities were not interpreted as representing dysplastic cortical changes, but they could probably represent local defective or delayed myelination. This hypothesis is adopted from studies on children with neurofibromatosis type 1 (NF1), where high signal intensities on T2-weighted images are common but with other localization [14]. The lesions remained static for several years and decreased with time during puberty and adolescence. DiPaolo et al. [15] suggested that the high signal intensity lesions on MRI in NF 1 represented increased fluid within the myelin associated with hyperplastic or dysplastic glial proliferation. It was postulated that the abnormal MRI signals may disappear with time, with resolution of the intramyelinic edema and replacement of abnormal myelin. It is possible that the high signal intensity lesions in RE may represent a transitory defective myelination, which could explain the mostly "benign" course of the syndrome. Besides, this phenomenon, in addition to hippocampal abnormalities, could be related to affection of cognitive ability in RE. That neuropsychological deficits are present in many children with RE has also been shown in this study group [16].

The relation between brain pathology and RE is far from clear. No pathology studies have so far been reported similar to those in idiopathic generalized epilepsy [17], where marked cerebral microdysgenesis was found.

The structural abnormalities found in our series are probably maturation disturbances with a possible genetic background. It can be suggested that they are related to the well-known outcome of seizures and EEG manifestations in RE. The finding of major brain lesions in single case reports of RE may be accidental but seem to worsen the seizure situation.

In order to get a better understanding of RE it is of importance to report goups of children with typical seizures, with and without rolandic sharp waves, as well as groups of children with typical rolandic sharp waves and atypical seizures. In addition, appropriate MRI studies should be performed.