Fetal ultrasound is a screening tool that has an established role in the diagnosis of cerebral malformations, congenital anomalies, and complex syndromes. In the postnatal period, ultrasound is often the initial study in a diagnostic workup for seizure activity. It has the benefits of being noninvasive and avoiding radiation exposure.
Neonatal head ultrasound's primary role has traditionally been in the evaluation of the preterm newborn for diagnosis of parenchymal hemorrhage, germinal matrix hemorrhage, and hydrocephalus. Ultrasound may detect changes of hypoxic ischemic injury, vascular anomalies, or brain malformations.
Computed tomography (CT) is most helpful in the identification of intracranial hemorrhage, major vascular malformations, and ventriculomegaly. The sensitivity of CT is approximately 30% in the detection of many of the causes of epilepsy (1-5). Given these limitations, as well as the risks of radiation exposure in infants and young children, CT has been replaced by MRI in the elective workup of childhood epilepsy.
In cases of recording in order to localize epileptic foci with electrode grids, CT may also be used to obtain images of electrodes in situ that can later be fused onto preoperative MR images with standard software. This technique allows the use of neuronavigation to guide operative resection of epileptogenic lesions (6).
Technical Considerations. The continued development of MRI has improved our ability to visualize the most common causes of epilepsy. Routine sequences should include high-resolution T2-weighted, fluid attenuation inversion recovery (FLAIR), T-2*- weighted gradient echo, 3D high resolution T1W gradient echo, and diffusion-weighted imaging. Imaging after administration of gadolinium contrast agents helps differentiate neoplasia from dysplasia and should be done whenever a structural abnormality is present.
The advent of high-resolution, multichannel coils and higher field strength (^ 3 Tesla) has enhanced signal-to-noise ratio and reduced scan times. Phasedarray coil imaging has improved cortical lesion detection by 64% compared with standard 1.5-T coils (7). As phased-array technology is adapted for 3-T systems, signal-to-noise ratio and resolution should improve even more (8). Research with 7-T multichannel array technology promises visualization of cortical abnormalities of less than 100 pm in the near future, prompting some authorities to predict an era of "in vivo histology" (9).
Any interpretation of magnetic resonance imaging of the developing brain must take into account the dynamic process of myelination and changes in water content, with corresponding signal intensity changes on T1- and T2-weighted imaging. In most basic terms, T1 and T2 relaxation times become shorter as maturation progresses, as a result of increased myelin and decreased water content in imaged tissue. Myelination begins in the dorsal brain-stem and cerebellum as early as the fifth fetal month, then involves the pyramidal and somatosensory radiations, and progresses later to the subcortical white matter of the prefrontal and anterior temporal regions over the first two years of life (10). Thus, white-matter maturation is evaluated best with T1-weighted images until 6 to 8 months of age, after which T2-weighted images become more sensitive to white-matter changes.
Detailed manual segmentation of specific brain regions and structures such as hippocampus and amygdala has permitted a more quantitative formulation of standardized growth trajectories and better understanding of both normal and abnormal development (8). For example, mesial temporal sclerosis (MTS) has been studied using quantitative techniques that have excellent sensitivity and specificity for detecting hippocampal asymmetry (11). Quantitative studies of regional brain volume also have found decreases in cerebral and cerebellar size in children with epilepsy (12).
Surface-based and volume-based representations of cortex provide population-averaged "architectonic" maps to permit truer comparisons of functional or structural data between individuals or research cohorts (13). For example, construction of a surface-based atlas has permitted the identification of cortical folding abnormalities in Williams syndrome (14).
A wide spectrum of cortical malformations, due to genetic or environmental insults, can cause pediatric epilepsy. These can be classified in terms of disruption of neuronal production, differentiation, and migration. The most common malformations are well visualized on MRI, especially with the use of high resolution 3-T and modern phased-array technology. The examples that follow illustrate the types of abnormalities that can be identified. For a more extensive review of this topic, see the references by Barkovich and by Raybaud et al (15, 16).
Transmantle dyplasias are disorders of neuronal and glial differentiation with or without abnormal migration. Histologically, these abnormalities contain dysplastic glioneuronal elements that increase signal on T2/FLAIR imaging depending on the child's age and state of myelina-tion of the brain. In many ways these dysplasias represent a one pole of a spectrum that blends into low-grade neoplasia. As a result, it can be difficult to differentiate the two based on imaging alone.
Was this article helpful?