|Year : 2008 | Volume
| Issue : 1 | Page : 48-54
Imaging in epilepsy
Wellspring Jankharia Imaging, Bhaveshwar Vihar, 383 Sardar V. P. Road, Mumbai - 400 004, India
Wellspring Jankharia Imaging, Bhaveshwar Vihar, 383 Sardar V. P. Road, Mumbai - 400 004
Source of Support: None, Conflict of Interest: None
| Abstract|| |
The various imaging techniques used in the diagnostic workup of the epileptic child are presented in this review, with special emphasis on magnetic resonance imaging (MRI) which has become the mainstay for the anatomical imaging workup. The MRI findings in various childhood epilepsies are discussed.
Keywords: Imaging, magnetic resonance imaging, epilepsy, hippocampus
|How to cite this article:|
Ursekar M. Imaging in epilepsy. J Pediatr Neurosci 2008;3:48-54
In the diagnostic workup of the epileptic child, imaging plays an important role. There are situations however, dependant on the clinical evaluation findings, in which imaging may not be needed at all. Prior assessment of the clinical and EEG characteristics, recorded ideally during the seizure, or as soon as possible after it, and at rest, will, in most instances define the type of epilepsy, and the decision as to whether imaging is required or not. From the radiologist's perspective, there is a great advantage in this electroclinical approach, in that there is prior information concerning the type of epilepsy, which, in turn guides the specific protocol selected, the correct choice of the reference plane for imaging and the decision whether to use I. V. contrast or not.
During the diagnostic stage, imaging (particularly Magnetic resonance imaging, MRI), is used to detect lesions which could be causal, potentially causal or facilitating. The effects of the epilepsy itself, such as progressive atrophy or shrinkage, post status, are also shown. In situations where the electroclinical data do not fit into a specific, defined framework, the imaging study may be of help. In surgical candidates, imaging is used to locate the anatomic lesion, and in such situations, it becomes even more imperative that the studies are conducted in close association with the clinicians, to avoid unnecessary repeated investigations requiring sedation or general anesthesia, and to enhance the diagnostic yield of the test. In presurgical evaluation, functional imaging techniques may be used to locate eloquent cortex, for estimating surgical risk and for defining the surgical approach.
| Technique|| |
MRI has become the mainstay and standard of anatomic epilepsy imaging, replacing CT studies in almost all situations owing to its greater sensitivity in detecting lesions. There is usually a need for sedation or anesthesia in the uncooperative or small child. The specific imaging protocol used is tailored according to the clinical assessment findings. The standard sequences include T1 weighted (preferably by inversion recovery, IR), T2 weighted fast spin-echo, and fluid-attenuated inversion recovery (FLAIR) sequences. The resolution and quality have to be high in order to detect small lesions and subtle cortical dysplasias. Spatial resolution can be enhanced by using three-dimensional (3D) Fourier transform (FT), gradient-echo T1-weighted acquisitions such as MPRAGE or SPGR, which allow thinner sections (1-2 mm) to be obtained, while maintaining a good signal-to-noise ratio despite the small voxel size.
The reference plane chosen for the examination depends on the electroclinical data. The bicommissural plane of Talairach,  which is the basic reference plane of the Talairach stereotactic system, is widely accepted, reproducible, and almost perpendicular to the central sulcus. It is the standard axial reference plane for anatomical imaging and is the plane most used in functional MRI studies, as well . Coronal planes that are acquired for the purpose of detecting lesions in the cerebral cortex outside the temporal lobes (in extra temporal partial epilepsies), are obtained perpendicular to this reference plane. The hippocampal plane passes through the planes of both hippocampal formations and is the reference plane used in patients with temporal lobe epilepsy (TLE). It is perpendicular to the tangent line drawn along the anterior surface of the pons on midsagittal images, or parallel to the plane of the left sylvian fissure.  This plane depicts the length of the hippocampi and the occipital lobes, and images perpendicular to it, section the hippocampal bodies exactly perpendicularly. In summary, the bicommisural reference plane is used in all cases of extra temporal partial epilepsy to study pathologies arising in the frontal lobes, paracentral cortex, and parietal lobes, while the hippocampal plane is used in temporal and occipital lobe epilepsies. The complete evaluation requires that both axial and coronal sections are performed in every case, and complemented by sagittal slices.
MRI pulse sequences
At least three sequences are routinely performed, irrespective of the reference plane chosen. The author prefers 4 mm axial T2-weighted FSE slices, 3 mm coronal FLAIR, and 3 mm T1-weighted IR, as essential sequences, supplemented by sagittally acquired 3D FT T1-weighted gradient echo MPRAGE sequences providing near-isotropic submillimeter voxel images which can be reconstructed in any oblique plane, depending on the location of the suspected pathology. Intravenously injected contrast material is used only in specific instances such as tumors, arteriovenous malformations, infections, suspected Sturge- Weber syndrome More Details More Details, etc. Postprocessing of 3D FT GE sequences is occasionally required for morphometry, stereotactic electrode placement, neuronavigation, and for radiosurgery.
Usually, in young children and in developmental delay, anesthesia, administered by an anesthesiologist or deep sedation is necessary.
Several studies have advocated the quantification of hippocampal atrophy to evaluate its severity. , Volumetry is performed on the 3D FT GE sequences reconstructed perpendicular to the hippocampal reference plane. With sufficiently high resolution and good interpolation algorithms, there should be no blurring of the boundaries of the various anatomic structures to be assessed. Anatomic boundaries that are chosen and the structures to be included for calculating the volumes have been defined previously.  The boundary limits of the hippocampus are the CSF of the temporal horn above (alveus included), the ambient cistern medially and the uncal recess between the hippocampal head and the amygdala. The most anterior plane through the head defines the anterior limit while the plane containing the posterior crus fornix defines the posterior limit of the tail. A manual tracing on adjacent coronal thin sections is performed while the workstation automatically calculates the sum of all the pixels chosen. Problems with the accurate selection of anatomic boundaries arise usually when defining the limits of the hippocampal head anteriorly from the basal ganglia. Similarly, the boundaries of the entorhinal cortex may be quite arbitrary.  In epileptic children, growth in size and the maturation of the temporal lobes have to be taken into account while interpreting volumetric data.  The data from volumetric studies of the temporal lobes in children with TLE, have shown good correlation with the clinical features and have shown good sensitivity and specificity  and correlation with postoperative outcome. It is generally believed however, that in most situations, visual evaluation of the hippocampus by an experienced observer, is sufficient.  Quantitative measurements have also been applied to brain structures apart from the hippocampal formations, such as the anterior temporal cortex, and the fornix and mammilary bodies, in the context of clinical research.
Other imaging studies
Sometimes, in spite of every attempt to optimize the anatomic imaging technique with tailored protocols, the MRI study fails to show a lesion, or there is discordance between the findings on MRI and the results of the electroclinical evaluation. In refractory cases being considered for surgery, imaging may have to be repeated after further investigations, ,, such as video recorded EEG, single photon emission computed tomography (SPECT) imaging, positron emission tomography (PET), and invasive EEG. Other sophisticated MRI techniques like MR spectroscopy have also been useful in localizing epileptogenic foci by showing a reduced magnitude of the N -acetlyaspartate (NAA) peak and the NAA-to-choline (Cho) ratio in the vicinity of the epileptogenic focus. Functional MR imaging (fMRI) can be used clinically to in presurgical localization of eloquent cortex.  Using the interictal spikes recorded by EEG, to start the MRI acquisition, the fMRI may also help in demonstrating the cortical area from where the epileptic activity is generated more precisely than by the use of scalp electrodes alone.
MRI findings in various childhood epilepsies
Benign idiopathic epilepsies
There are many syndromes of benign idiopathic epilepsy, further classified into generalized and partial, and recognized by their clinical features, EEG pattern and familial background. The prognosis is usually good. There is no need for imaging, as the brain is morphologically normal. It is possible however, to have typical generalized epilepsy with a fortuitously detectable cerebral lesion, unrelated to the epilepsy.  Although rare, such a situation emphasizes the need to recognize that the clinical and EEG findings, are of utmost importance, and over that of the morphology, shown by imaging. Further, although it is believed that the brain should be morphologically normal in cases with idiopathic epilepsies, the "normalcy" has been questioned, based on evidence gathered through some autopsy studies. Morphological abnormalities like subpial bands of abnormal neurons, subtle increases in the density of neurons in the molecular layer of the cortex, heterotopic neurons in the white matter, abnormality in the cellular organization of the hippocampii, besides other subtle morphological brain abnormalities, have been observed in a series of eight autopsy cases by Meencke and Janz.  In future, as the efficacy of imaging improves even further, it is likely that a larger number of "cryptogenic" epilepsies will be found to have subtle abnormalities at imaging.
| Symptomatic Generalized Epilepsy|| |
Generalized epilepsies may present as any type of seizure and can be generalized from the onset, or develop as a spreading of partial seizures. Any generalized disorder of the brain, such as cerebral malformations, can cause symptomatic generalized epilepsy. Other specific morphological abnormalities include tuberous sclerosis, metabolic diseases, and nonspecific abnormalities such as diffuse atrophy, multifocal encephalomalacia, or hydrocephalus.  Neuroimaging by MRI is essential to recognize the underlying anatomic lesion that may be the causing or facilitating the seizures. Imaging is also used to formulate the prognosis. Within the group of symptomatic (or cryptogenic) generalized epilepsies, certain characteristic syndromes with specific features, treatments and prognoses, are recognized. Their imaging features are described below.
Infantile spasms (West syndrome)
Infantile spasms usually present between 3 and 7 months of age and are thought to represent a single-specific response of the infantile brain (characteristically hypsarrythmia on EEG) to a variety of insults. In approximately 15%, there is no detectable, underlying pathology (cryptogenic West syndrome). In the more typical symptomatic forms, the underlying insults include developmental abnormalities such as commissural agenesis, hemimegalencephaly, pachygyria, holoprosencephaly, and schizencephaly as well as other prenatal, perinatal, and postnatal insults. The prenatal causes include TORCH infections, and fetal anoxic-ischemic injury such as intrauterine growth delay and porencephaly. Malformations of the brain account for 30% in autopsy series. In Aicardi syndrome infantile spasms occur with agenesis of the corpus callosum, chorioretinal lacunae, and sometimes cortical malformation and heterotopias. Tuberous sclerosis is seen in 7-13% cases with infantile spasms. Downs syndrome is also a facilitating condition. Amongst the perinatal causes, hypoglycemic brain injury seems to be a common cause in the Indian context.
The so-called cryptogenic infantile spasms are characterized by normal brain imaging, and a developmentally and neurologically normal presentation prior to the seizure onset. Some of these patients develop severe encephalopathies, while others, and normalize. In patients with worse outcomes, imaging studies show the development of significant atrophy on follow-up.
Early myoclonic encephalopathy
The onset is between 2 weeks and 3 months of age. Imaging findings include subcortical cystic encephalomalacia due to HIE or nonketotic hyperglycenemia. 
Early infantile epileptic encephalopathy
The onset is between 4 weeks and 2 years of age. The imaging findings usually include brain malformations or metabolic disorders.
It may rarely be idiopathic or develop in patients with preexisting neurologic problems such as delayed development, infantile spasm, and other types of epilepsy. In one pathologic study in 15 cases, findings included selective postsynaptic neuronal necrosis in the neocortex, hippocampus, thalamus, and cerebellum. MRI shows the preexisting abnormality or nonspecific findings such as atrophy.
Progressive myoclonic epilepsy (PME)
Although rare as a group (1% of the epilepsies in a referral center),  the PMEs present with relatively characteristic features. MR imaging, supplemented by MR spectroscopy may be helpful in identifying specific metabolic diseases underlying PME, more often, a neuronal storage disorder. In some situations, the imaging features are quite suggestive of the specific metabolic disorder, such as in the case of GM1 and GM2 gangliosidosis and mitochondrial encephalopathies such as MELAS (stroke-like lesions) and MERFF (predominant alteration of the basal ganglia with atrophy, high T2-signal, and mineralization). Other underlying disorders like sialidosis, neuronal ceroid lipofuscinosis, Niemann-Pick disease, Gaucher disease, biopterin deficiency, Lafora disease, and the degenerative PMEs (Unverrich-Lundborg disease) have no specific imaging features. The final diagnosis is based on biologic and biochemical data, and on biopsies.
Symptomatic partial-onset epilepsy
The seizures in partial-onset epilepsies arise from a specific area of the cortex, have a tendency to become drug resistant, and are frequently associated with an underlying causal lesion, which, when removed, often leads to a cure. Neuroimaging plays a key role in the detection and characterization of specific underlying pathological substrates of partial-onset epilepsy. Undoubtedly, high-quality MR imaging is the study of choice and is easily applicable to children, with the added advantage of greater sensitivity over other anatomic imaging techniques, thereby making the diagnosis possible much earlier, in the first months or years of the disease. A reasonable early surgery is now recommended in children with refractory partial-onset seizures, because over time, the epilepsy tends to become more severe and to develop secondary foci.  A conventional MRI may sometimes fail to demonstrate the underlying lesion, although the clinical and EEG findings point to a specific functional zone. Such situations (especially patients being worked up for epilepsy surgery) demand further refinements in the MRI technique through the use of higher-field strength (3 T vs. 1.5 T) and specialized surface-array head coils, both, potentially having the potential to increase the spatial resolution of the anatomic images. Ideally, the MRI study should be performed after an intensive clinical search through simultaneous EEG and video recording (video telemetry) has been made to focus the MR imaging to the specific cortical area of interest. It is most useful from the radiological perspective that the MR analysis is made with prior knowledge of the clinical and EEG data. In some cases, despite every effort made, the MRI fails to demonstrate a lesion. The imaging investigations would then have to be supplemented by functional imaging techniques such as ictal, postictal and interictal PET and SPECT, and perhaps, a cortical or subcortical EEG. It is also important to remember that, in some instances, the epileptogenic focus is different from the lesion shown by MR imaging and the radiologic findings have to be interpreted together with the clinical and EEG data.
Malformative and neoplastic glioneuronal lesions
This group consists of developmental abnormalities of the cortex with characteristic histopathological features of poorly differentiated ganglionic (neuronal) cells with abnormal connections, making the malformations intrinsically epileptogenic.
Transmantle dysplasias include:  focal cortical dysplasia of Taylor (FCD)  ; isolated tuber of tuberous sclerosis complex (TSC) or forme fruste of TSC in which the other stigmata of TSC are lacking  ; hemimegalencephaly, which is histologically indistinguishable from the FCD. 
The transmantle dysplasias are postulated to arise from poor premigratory differentiation of the stem cells of the cerebral mantle, resulting in giant, bizarre neurons, and the "giant astrocytes" of TSC.  These poorly differentiated neuroglial cells are scattered across the cerebral mantle between the surface and the ventricle, because of abnormal migration, resulting in blurring of the junction between the cortex and white matter.
The term FCD is used to describe a focal area of dysplasia in the cortex. Typically, the involved cortical area demonstrates a focally expanded gyrus with high signal intensity, elongated triangular-shaped region seen in the underlying white matter on FLAIR and T2-W image. The apex of the hyperintense signal extends into the ventricular margin, and the base lies at the cortex. The cortical-subcortical interface is blurred.  The fan-shaped transmantle dysplasia appears darker than the normal white matter surrounding it on IR sequences. Thin sections must be performed in every case to delineate the fan-shaped extension, which is often small, but constitutes the most characteristic radiological feature of this entity. In some cases, the lesion is limited to the cortex and appears hyperintense on FLAIR images. They can usually be distinguished from cortical gliomas. 
Isolated tuber : These have radiological features similar to the tubers of TSC, but unlike TSC, do not show the other typical radiological findings of subependymal hamartomas, subependymal giant cell astrocytoma, or multiplicity of the cortical tubers. The isolated tuber has appearances of an expanded gyrus with a cortex that appears thicker than usual, and a sucortical area of extremely low signal intensity on T1 and high signal on T2 and FLAIR images. The pathology of these lesions is identical to the tubers of TSC.
This entity is characterized on MR images by an increase in size of one cerebral hemisphere, with a thickened cortical ribbon, enlargement of the ipsilateral ventricle with an abnormally shaped frontal horn, broad featureless gyri with shallow sulci, abnormal primitive veins overlying shallow sulci, and abnormal white matter signal. In neonates and infants, the dysplastic cortex may appear hypointense on T2WI . On FLAIR images, the white matter shows increased signal intensity reflecting altered myelination. Diffusion tractography can show increased number or density of fiber tracts. An anomalous venous pattern is seen on MR venograms. In 50% of the patients, the PET studies show glucose hypometabolism in the affected hemisphere. In the presurgical evaluation of cases being considered for hemispherectomy, it is important for the radiologist to look carefully at the contralateral hemisphere, and report on whether it is normal or abnormal.
Hemimegalencephaly may be associated with neurocutaneous and overgrowth syndromes such as NF type 1, tuberous sclerosis, Klippel-Trenaunay-Weber, Proteus syndrome, unilateral hypomelanosis of Ito, epidermal nevus syndrome, and incontinentia pigmenti.
Dysplastic glioneuronal tumors
This group includes dysembryoplastic neuroepithelial tumor (DNT), ganglioglioma, and dysplastic gangliocytoma. Most patients with DNT present with a longstanding, early-onset partial epilepsy and with no deficit or only a stable congenital neurologic deficit. The tumor remains static in morphology over repeated follow-up MRI studies and is most often located in the temporal lobes. The tumor is cortical and hypodense or pseudo cystic on CT, tends to appear very hyperintense on T2W images attributable to its rich interstitial matrix, devoid of edema, and may cause calvarial remodelling due to its indolent nature. Approximately one-third of these lesions enhances, usually with a ring-like pattern. These tumors have to be distinguished radiologically from astrocytomas which appear more infiltrating, while the margin of a DNT is always sharply defined. Gangliogliomas are composed of abnormal ganglion cells admixed with neoplastic astrocytes. The ganglionic component is thought to be purely dysplastic and stable in its behavior, whereas the glial component, neoplastic,  most often low grade, but in some cases it can transform to an anaplastic tumor. Radiologically, the tumor is often a well-circumscribed cystic lesion of variable size, calcified in a third of cases, and contains a mural nodule that usually enhances. 
| Heterotopia|| |
These are masses of normal neurons in abnormal locations, the resultant mass of gray matter (GM) being functional and connected with the rest of the brain thereby contributing to their epileptogenesis.  It is postulated that heterotopias not only arise because of disorder during the period of neuronal migration, leading to either local or diffuse, nodular or laminar, and periventricular or subcortical, but also transcerebral or intracortical masses . Apart from heterotopias, the agyria-pachygyria complex is also postulated to arise out of disordered neuronal migration. At MR imaging, heterotopias show signal characteristic of normal GM and are easy to recognize. Various types of heterotopias are encountered including  : isolated subependymal nodular heterotopias. They are small none calcified, noncystic, masses lying adjacent to the ventricular walls creating a mild undulation of the ventricular margin. When bilateral they are often symmetrical. They may be isolated or associated with commissural agenesis, polymicrogyria or schizencephaly.  Diffuse subependymal nodular heterotopias are an X-linked disorder, almost exclusively occurring in girls, and usually lethal in boys. They appear as numerous, conglomerating masses of GM along the ventricular margins, associated with ventricular enlargement and atrophic appearance of the cortex. They occur with mutations in filamin-1 gene on Xq28.  Subcortical nodular heterotopias are located between the cortex and ventricular margin, are often large, and may be associated with dysplasia in the overlying cortex and subependymal heterotopias.  Transcerebral heterotopias extend across the mantle from the cortex to the ventricle, and may be bilateral. They could be classified as extreme forms of closed-lip schizencephaly but differ, in that they are solid masses of GM without any pia-mater, vessels or CSF, rather than areas of infolded cortex with pia-mater and CSF (closed lip schizencephaly). Their morphogenesis however, still remains unclear.  Laminar or Band heterotopias are layers of GM extending parallel with the cortical contour, but located away from the cortex, under it. The appearance is of a "double cortex" when diffuse, and almost always in girls. Partial bands may be seen in the frontal lobes or the posterior aspect of the cerebral hemispheres.  Lissencephaly : in classical lissencephaly (LIS1), there is arrested neuronal migration resulting in a four-layer cortex and a smooth brain surface. This phenotype correlates with gene alterations, most commonly LIS1 and DCX (double cortex also called XLIS). An hour-glass configuration of the brain often somewhat incomplete (pachygyria-agyria) is seen. The cortical changes are predominantly parietal-occipital when incomplete. A typical layering is seen in the cortex on T2W images, with a thick inner band of GM, cell sparse white matter zone, and thin outer layer of GM. In DCX, bands of thinner, symmetric subcortical ribbons of GM paralleling cortex and embedded in white matter are noted in the hemispheres. The thickness of the band predicts the thickness of the overlying cortex, for instance, a thick band is seen with a thin, abnormally convoluted overlying cortex. The severity is more in the temporal and subfrontal lobes as opposed to LIS1. The genetic lissencephalies are differentiated from CMV -related lissencephaly which shows no cortical thickening, and presence of typical calcifications. Lissencephaly with cerebellar hypoplasia (LCH) can be seen with LIS1 and XLIS mutations. Although lissencephaly can be diagnosed in utero by fetal MRI studies, it has to be remembered that the normal fetal brain displays a smooth cortex up to 26 weeks. The presence or absence of the parieto-occipital fissure and a shallow Sylvian fissure, are more specific signs.
| Polymicrogyria|| |
Polymicrogyria arises due to an abnormality arising in late neuronal migration and cortical organization, in which the neurons reach the cortex, but distribute abnormally forming multiple small undulating gyri. Schizencephaly is also thought to arise similarly in late neuronal migration. On MRI, the cortical-white matter junction is always irregular. The polymicrogyria has a predilection for perisylvian regions, and when bilateral, is often syndromic. The cortex is of variable thickness, being thickened or normal, bumpy or smooth, with shallow or absent sulci. The appearances on T2W images are influenced by the stage of maturation of the brain. In babies >12 months, the cortex appears thin, with fine undulations. In brains >18 months, the cortex is thick, bumpy with or without underlying hypomyelination. Dysplastic leptomeningeal veins on the surface are seen on postcontrast T1W images and at MR venography. On PET studies, foci of polymicrogyria show increased metabolism during ictus and hypometabolism interictally. CT aids in the detection of periventricular calcifications when polymicrogyria is associated with CMV. Polymicrogyric mimicks can be seen in malformations due to some inborn errors of metabolism such as Zellweger syndrome and the mitochondrial and pyruvate metabolism disorders.
| Schizencephaly|| |
The term schizencephaly describes a cleft in the brain parenchyma that extends from the cortical surface to the ventricle and is lined by dysplastic GM. The cleft may be wide open or narrow and even closed, when the only sign of its presence is the appearance of a dimple in the wall of the ventricle. Clefts are usually located near the central sulcus and may be either unilateral, or bilateral. The GM lining of the cleft may be indistinct on T1-weighted MR images, and is better delineated on T2-weighted images, prior to myelination. After myelination is complete, the lining appears "cobblestone"-like on T1-weighted MRI. A canal of T2-hyperintense CSF can be seen in the cleft when it is open. Clefts associated with CMV may be calcified. The pia overlying the clefts contains anomalous venous channels, which are seen on postcontrast T1-weighted images. The relationship of clefts to the surrounding cortex is best appreciated on 3D surface rendered MRI. Functional reorganization of the normal cortical areas around the cleft has been noted on BOLD fMRI studies. Schizencephaly should be differentiated radiologically from encephaloclastic porencephaly, which is lined by gliotic white matter and not by dysplastic GM and transmantle heterotopias, which infact, may represent a form of closed-lip schizencephaly.
| Hamartomas|| |
Hamartomas that occur in the hypothalamus are masses of neural tissue attached to the tuber cinereum at the anterior aspect of the floor of the third ventricle. They may be sessile or pedunculated. Their vascularity is usually connected with that of the adjacent parenchyma or the circle of Willis. They may project into the third ventricle or into the suprasellar-interpeduncular cisterns. Some may be asymptomatic, while others present in the first year of life with epilepsy or hormonal disturbances.  MR imaging is diagnostic in this condition. A mass is seen attached to the tuber cinereum, identical to GM signal intensity on T1-weighted images and similar to, or slightly higher signal intensity than that of the surrounding brain on FLAIR and T2-weighted images. The signal intensity of the mass in young children may change in young infants as the brain matures.
| Cavernous Angioma|| |
Partial-onset seizures may be caused by cavernomas in children. Histologically, they show large vascular spaces without enlarged feeding arteries or draining veins. In some familial cases, cavernomas are known to appear, develop, and multiply with growth of the child. They appear lobular in outline and contain foci of high signal on T1, T2, and FLAIR images. A rim of signal void related to the paramagnetic effects of hemosiderin/ferritin is seen at the margin on T2 weighted and gradient echo images.
| Cerebral Scars|| |
Acquired lesions such as cerebral scars are reported to occur in more than 10% cases of partial-onset epilepsy.  In the pediatric population, these are usually the result of perinatal or infantile insults. The imaging findings vary from subtle (areas of cortical thinning, gyral atrophy either focal or diffuse, subtle T2 hyperintense changes in the parenchyma) to obvious (porencephalic cavities, larger areas of cortical or subcortical involvement as in old infarctions). Hippocampal necrosis resulting from excitotoxic injury via neurotransmitters, like glutamate, which is released in excess by anoxia and prolonged seizures, may lead to scarring within the hippocampus (hippocampal sclerosis, HS).
| Mesial temporal Sclerosis|| |
Medically intractable TLE may have many pathological substrates, including tumors, vascular malformations, and most importantly, HS. Also referred to as mesial temporal sclerosis, HS is a condition in which neuronal loss and gliosis within the hippocampus leads to reorganization of neuronal circuitry within the hippocampus, which consequently results in an epileptogenic focus. Treatment of HS by surgical resection has a good outcome, and therefore it is extremely important to diagnose it. High resolution MRI performed in the coronal plane, is the best technique available for evaluating the morphologic and metabolic changes (with MR spectra) that establish the diagnosis. An asymmetrically small or shrunken hippocampus ipsilateral to the seizure focus is the most significant MR finding. Other findings include hyperintense hippocampal signal on T2-weighted images and ipsilateral atrophy of the temporal lobe. 93% sensitivity and 86% specificity for MRI have been reported, for the diagnosis of HS. The accuracy is probably even higher, when the study is optimized and performed with newer scan protocols.
Quantitative MR volume measurements may increase detection of HS, especially when there is bilateral atrophy.
MRI can easily distinguish HS from tumors (in which the hippocampus is enlarged and not shrunken) and hamartomas.
| Rasmussen Encephalitis|| |
Patients with epilepsia partialis continua are characterized at imaging studies by progressive brain atrophy. In early stages of the disease, MRI shows an edematous swelling of the cortex, often diffusely involving an area of the cerebral hemisphere, low signal on T1-weighted and high signal on T2-weighted images, with no contrast enhancement. In time, the area of inflammatory gliosis expands the cortex and the underlying white matter and atrophy progressively develops with enlargement of the ipsilateral ventricle. MRI studies are used to map out the extent of anatomic involvement prior to surgery and fMRI studies have shown brain plasticity in young children after surgery.
| Status Epilepticus|| |
In children, the main causes of status are a systemic non-central nervous system (CNS) infection, and remote causes such as brain malformations, inadequate levels of antiepileptic drugs, stroke, metabolic causes, hypoxia, and CNS infections. Imaging is usually not required in the acute phase of status, but may be performed after the acute phase, to detect an underlying brain disorder, especially in patients with refractory status. Prolonged status epilepticus causes brain edema, which may be diffuse, unilateral, or focal, associated with a lasting vasodilatation.  These lesions may further develop into permanent brain atrophy. It has been believed that status in early childhood or prolonged episodes of febrile seizure may lead to mesial temporal sclerosis. This assumption has now been contested, and epidemiologic studies suggest that, instead, a preexisting HS would have facilitated the episodes of prolonged convulsions or status. 
Implications of imaging for epilepsy surgery
Surgical resection for epilepsy used to be based on EEG and clinical findings, before the advent of MRI. False localization due to wide propagation of the surface EEG abnormalities was frequent, occurring in 10-15%. Invasive EEG techniques had to be used for more accurate localization. This was especially true for the lateralization of HS. With the availability of MRI, presurgical evaluation of the epileptic patient has dramatically changed from being invasive with the attendant risk of complications, to noninvasive. Neuroimaging is used along with surface EEG to lateralize the seizure focus. PET and SPECT have also been used for the same purpose but remain ancillary techniques, which are used when results of MRI, EEG, and clinical evaluation are discordant. If all preoperative data such as clinical and neuropsychological testing, scalp EEG, 24-h continuous audio/visual monitoring, and MRI are concordant, then the patient may be considered for surgical resection.
In summary, CT and MRI are useful to identify the origin and cause of seizure disorders. CT scan is cheaper and more widely available. It can be useful for the detection of acquired epileptogenic lesions like tumors and infectious granulomas. MRI is clearly superior, especially for the detection of neuronal migration disorders and HS. MRI plays an important role in the surgical management of patients with medically intractable seizures. Its use in the work up of these patients has led to a decrease in the number of preoperative invasive tests and increase in the precision of surgical resection.
| References|| |
|1.||Talairach J, Tournoux P. Referentially oriented cerebral MRI anatomy. Atlas of stereotaxic anatomical correlations for gray and white matter. Stuttgart: Georg Thieme Verlag; 1993. |
|2.||Jackson GD. Visual analysis in mesial temporal |
|3.||Jack CR Jr, Twomey CK, Zinsmeister AR, Sharbrough FW, Petersen RC, Cascino GD. Anterior temporal lobes and hippocampal formations: Normative volumetric measurements from MR images in young adults. Radiology 1989;172:549-54. [PUBMED] [FULLTEXT]|
|4.||Watson C, Jack CR, Cendes F. Volumetric magnetic resonance imaging: Clinical applications and contributions to the understanding of temporal lobe epilepsy. Arch Neurol 1997;54:1521-31. |
|5.||Insausti R, Juotten K, Soininen H, Insausti AM, Partanen K, Vainio P, et al . MR Volumetric analysis of the human entorhinal, perirhinal and temporopolar cortices. AJNR Am J Neuroradiol 1998;19:659-71. |
|6.||Utsunomiya H, Takano K, Okazaki M, Mitsudome A. Development of the temporal lobes in infants and children: Analysis by MR-based volumetry. AJNR Am J Neuroradiol 1999;20:717-23. [PUBMED] [FULLTEXT]|
|7.||Lawson JA, Nguyen W, Bleasel AF, Pereira JK, Vogrin S, Cook MJ, et al . ILAE-defined epilepsy syndrome in children: Correlation with quantitative MRI. Epilepsia 1998;39:1345-9. [PUBMED] |
|8.||Jackson GD. Visual analysis in mesial temporal sclerosis. In: Cascino GD, Jack CR, editors. Neuroimaging in Epilepsy. Principles and Practice. Boston: Butterworth-Heinemann; 1996. p. 73-110. |
|9.||International league against epilepsy commission report. Guidelines for neuroimaging evaluation of patients with uncontrolled epilepsy considered for surgery. Epilepsia 1998;39:1375-6. |
|10.||Kuzniecky RI. Neuroimaging in paediatric epilepsy. Epilepsia 1996;37:S10-21. [PUBMED] |
|11.||Ng TC, Comair YG, Xue M, So N, Majors A, Kolem H, et al . Temporal lobe epilepsy: Pre-surgical localization with chemical shift imaging. Radiology 1994;193:465-72. [PUBMED] [FULLTEXT]|
|12.||King D, Baltuch GH. Magnetic resonance imaging and temporal lobe epilepsy. Acta Neurol Scand 1998;98:217-23. [PUBMED] |
|13.||Santanelli P, Bureau M, Maggauda A, Gobbi G, Roger J. Benign partial epilepsy with centrotemporal (or Rolandic) spikes and brain lesion. Epilepsia 1989;30:182-8. |
|14.||Meencke HJ, Janz D. Neuropathological findings in primary generalized epilepsy: A study of 8 cases. Epilepsia 1989;25:8-21. |
|15.||Battaglia A. Seizures and dysplasias of cerebral cortex in dysmorphic syndromes. In: Guerrini R, Andermann F, Canapicchi R, et al. , editors. Dysplasia of Cerebral cortex and Epilepsy. Philadelphia: Lippincott-Raven; 1996. p. 199-209. |
|16.||Dreifuss FE. Malignant syndromes of childhood epilepsy. In: Porter RJ, Chadwick D, editors. The Epilepsies 2. Blue Books of Practical Neurology. Boston: Butterworth-Heinemann; 1997. p. 157-65. |
|17.||Roger J, Gambarelli-Dubois D. Neuropathological studies of the Lennox-Gastaut syndrome. In Niedermeyer E, Degen R, editors. The Lennox-Gastaut syndrome. New York: Alan R. Riss; 1988. p. 73-93. |
|18.||Roger J, Genton P, et al . Les epilepsies myocloniques progressives de l'enfant et de l'adolescent. In: Roger J, Bureau M, Dravet Ch, editors. Epileptic Syndromes in Infancy, Childhood and Adolescence, 2 nd ed. London: John Libbey; 1992. p. 381-400. |
|19.||Morrell F, de Toledo-Morrel L. From mirror focus to secondary epileptogenesis in man: An historical review. In: Stefan H, Andermann F, Chauvel P, et al. , editors. Plasticity in Epilepsy: Dynamic Aspects of Brain Function. Advances in Neurology, vol 81. Philadelphia: Lippincott Williams and Wilkins; 1999. p. 11-23. |
|20.||Barkovich AJ, Kuzniecky RI, Bollen AW, Grant PE. Focal transmantle dysplasia: A specific malformation of cortical development. Neurology 1997;49:1148-52. [PUBMED] |
|21.||Robain O. Introduction to the pathology of cerebral cortical dysplasia. In: Guerrini R, Andermann F, Canapicchi R, et al. , editors. Dysplasias of Cerebral Cortex and Epilepsy. Philadelphia: Lippincott-Raven; 1996. p. 1-9 |
|22.||Yagishita A, Arai N, Maehara T, Shimizu H, Tokumaru AM, Oda M. Focal cortical dysplasia: Appearance on MR images. Radiology 1997;203:553-9. [PUBMED] [FULLTEXT]|
|23.||Bronen RA, Vives KP, Kim JH, Fulbright RK, Spencer SS, Spencer DD. Focal cortical dysplasia of Taylor, balloon cell subtype: MR differentiation from low grade tumours. AJNR Am J Neuroradiol 1997;18:1141-51. [PUBMED] [FULLTEXT]|
|24.||Wolf HK, Wiestler OD. Malformative and neoplastic glioneuronal lesions in patients with chronic pharmacoresistant epilepsies. In: Stefan H, Andermann F, Chauvel P, et al. , editors. Plasticity in epilepsy: Dynamic aspects of brain function. Advances in neurology, vol 81. Philadelphia: Lippincott Williams and Wilkins; 1999. p. 69-79. |
|25.||Castillo M, Davis PC, Takei Y, Hoffman JC Jr. Intracranial ganglioglioma: MR, CT and clinical findings in 18 patients. AJNR Am J Neuroradiol 1990;11:109-14. [PUBMED] |
|26.||Chevassus-au-Louis N, Represa A. The right neuron at the wrong place: Biology of heterotopic neurons in cortical neuronal migration disorders, with special reference to associated pathologies. Cell Mol Like Sci 1999;55:1206-15. |
|27.||Breningstall GN. Gelastic seizures, precocious puberty, and hypothalamic hamartomas. Neurology 1983;35:1180-3. |
|28.||Peretti P, Raybaud C, Dravet C, Mancini J, Pinsard N. Magnetic resonance imaging in partial epilepsy of childhood. J Neuroradiol 1989;16:308-16. [PUBMED] |
|29.||Sanmaritano M, Andermann F, Melanson D, Pappius HM, Camfield P, Aicardi J, et al . Prolonged focal cerebral edema associated with partial status epilepticus. Epilepsia 1985;26:334-9. |
|30.||Maytal J, Shinnar S. Status epilepticus in children. In: Shiinar LJ, Amir N, Branski AD, editors. Childhood Seizures. Pediatric and Adolescent Medicine, vol 6. Basel: Karger; 1995. p. 111-22. |
|This article has been cited by|
||Relation between electroencephalogram and neuroimaging present in children with epilepsy of difficult control | [Relación entre electroencefalograma y neuroimagen en niños con epilepsia focal de difícil control]
| ||Valdivia Álvarez, I., Aguilar Fabré, L., Francisco Pérez, A. |
| ||Revista Cubana de Pediatria. 2009; 81(3): 2 |