Epilepsy surgery, by definition is the removal of an area of brain with the aim of alleviating seizures. In order to proceed, evaluation needs to be undertaken to determine whether seizure onset can be localized to one area, and whether that area is functionally silent. In the majority, the aim is to alleviate seizures with no deterioration in function.
There are no current randomized studies in children demonstrating the superiority of surgery over medical treatment for this group of patients. However, many nonrandomized studies strongly suggest that epilepsy surgery may produce seizure freedom in a substantial number of nonidiopathic focal epilepsy (NIFE) cases, regardless of age or cause. The evidence base now suggests that 40%–80% will become seizure free. This figure is dependent on underlying pathology (less likely in developmental malformations and cryptogenic cases) and extent of resection. Therefore, surgery is now established in the routine management of children with presumed lesional focal epilepsy, and criteria for referral and evaluation for pediatric epilepsy surgery have been recently proposed.1
SURGERY IN CHILDREN NOT ONLY AIMS AT SEIZURE ALLEVIATION
Children should not be referred for surgery as a “last resort.” The rate of cognitive and behavioral comorbidity associated with early onset of epilepsy is high, and early cessation of seizures likely leads to improved neurobehavioral outcome. The concept of “epileptic encephalopathy” implies ongoing cognitive impairment as a result of underlying epileptic activity and may be potentially reversible. Longitudinal data on cognitive outcome are lacking due to the absence of standardized assessments over the lifespan and lack of a control group.
There is some evidence of stable postoperative intelligence quotient (IQ) suggesting at least an unchanged developmental trajectory that may be related to duration of epilepsy rather than ultimate seizure outcome. Brain plasticity and relocalization of function are additional factors that are influenced by age and type of surgery.
SURGERY IN CHILDREN IS NOT THE VERY LAST OPTION
There is no minimal age for surgical referral. All children under the age of 5 years with a definitive magnetic resonance imaging (MRI) lesion are potential surgical candidates. Should seizure control be achieved with medication a conservative wait period is indicated. Children older than the age of 5 years with a definitive lesion may be left to determine response to anticonvulsants but failure of two medications in children with apparent focal seizures should be prompt referral for assessment. Delayed surgical referral depends on adequacy of neurodevelopmental progress. As a rule, all children suffering from pharmacoresistant focal epilepsy should be evaluated at a specialized epilepsy center if they exhibit behavioral and cognitive dysfunction.
The number of antiepileptic drugs (AEDs) utilized and duration of the epilepsy are of strict importance in the decision to perform epilepsy surgery in adults for epilepsy. In contrast, the pediatric evaluation is influenced by a variety of additional factors such as type of epilepsy and prognosis for seizure freedom. For instance, the likelihood of favorable surgical outcome is lower in nonlesional cases but the absence of a brain lesion does not exclude definitive surgical possibilities.
SURGERY IN CHILDREN DIFFERS FROM SURGERY IN ADULTS
Once the history of disabling medically refractory focal seizures is clearly established—which requires at least a good description of seizure semiology, repeated interictal electroencephalographic (EEG) data, and an optimal MRI scan—the principal goal for resective epilepsy surgery is the identification and accurate localization of the epileptogenic zone. To answer this question, many investigations are available, but their use must be decided in a stepwise fashion based on the individual anatomo-electro-clinical features of each patient (Fig. 55–1). Not surprisingly, topographic diagnosis in children may prove challenging for several age-related issues.
Interictal EEG abnormalities may be diffuse and the EEG may need to be repeated over time to better define focal abnormalities that only appear or become more evident during evolution. Also, the coexistence of age-related “benign spikes” with other focal abnormalities may erroneously conduct to the diagnosis of multifocal epilepsy.
The collection of precise information on seizure semiology is often difficult, especially in young children. Notably, description of the initial symptom, when found, is much less detailed than in adult patients, and the level of awareness is rarely possible to evaluate.
Focal seizures in children can manifest in very unusual clinical presentations, when compared to classical adult patterns. Ictal symptomatology may pass almost unnoticed or be particularly “explosive.” Very young children often exhibit global motor signs that may mask the focal character of the seizures, or which precedes the more clearly focal semiology, as it is typically the case with epileptic spasms.
The clinical expression of focal seizures at a given age may evolve with time under the influence of many factors such as progression of brain maturation, increasing ability of children to describe subjective feelings, and treatment.
Although a significant number of children undergo surgery to remove a localized brain area, more commonly children often undergo lobar or multilobar resection; approximately one third will have a hemidisconnection, whether hemispherectomy or hemispherotomy.2
Children considered for hemidisconnection usually have a preexistent congenital hemiplegia due to structural abnormality of the contralateral hemisphere. Presurgical assessment is aimed at lateralisation of seizure onset, and any risk for functionality. Early seizure onset is often associated with complex patterns of language representation, especially if language is represented in the normal dominant hemisphere. Affected children may not evidence significant verbal/nonverbal discrepancy on neuropsychological evaluation. Remarkably, hemispherectomy is often associated with little change in cognitive or functional assessment although a visual field defect is inevitable if not already present.
Children with acquired disease such as Sturge–Weber syndrome and Rasmussen’s encephalitis typically suffer greater functional postoperative deficits. Hemispheric procedures in children with dominant hemisphere involvement in Rasmussen’s Encephalitis require careful consideration regarding the potential for of relocalization of language function. The final decision rests on the risk-benefit assessment of preexisting cognitive compromise versus surgically induced deficits and the prevention of future deterioration.
Children who are candidates for focal/multilobar resection (or disconnection) should have seizures arising from one area of the brain, which can be safely resected without functional consequence. Functionality is in turn determined by the region of brain involvement, the structural abnormality on neuroimaging and the pattern of representation of cognitive function.
In the presence of a discrete structural lesion in a silent cortical region, surgical planning is relatively straightforward. Removing the lesion is possible without further investigation, if there is concordance between the lesion, the clinical semiology, and interictal scalp EEG data. Video-EEG recording of seizures recording provides further confirmation of the epileptogenic zone.
In the presence of poorly defined lesions, lesions that encroach on motor or language cortex, or discordant findings, invasive EEG monitoring helps determine the extent of surgery. Intracranial EEG recording is almost always necessary in MRI-negative cases.
In all pediatric surgical candidates, the aims of the family must be carefully explored to ascertain that they have realistic expectations.
Children presenting with focal epilepsy often have a catastrophic course. Demonstration of focal seizure onset a may be challenging. Young infants may present with focal seizures and evolve through infantile spasms that may or may not be lateralized in their presentation. In view of the devastating effect that early onset epilepsy may have on ongoing neurodevelopment (Jonas et al, 2005), potential surgical candidates should be considered early; onset under the age of 2 years is particularly correlated to low long-term IQ.3,4
Data on the long-term outcome of children who have undergone surgery reveal maintenance of the learning trajectory postoperatively. The greatest gains are observed when seizure freedom is achieved, and after a longer postsurgery interval.5 In children under the age of 3 years, gains were more likely if surgery was performed under the age of 1 year, and with a history of infantile spasms.6
The International League Against Epilepsy (ILAE) survey of pediatric procedures for epilepsy (2004) revealed that although two-thirds of children had onset of epilepsy less than the age of 2 years, only one-third had surgery within two years of seizure onset.2 However, although major gains can occur with early seizure cessation, the primary goal of pediatric epilepsy surgery remains seizure freedom, as likely neurodevelopmental gains are unpredictable.
A high index of suspicion should be entertained when evaluating a child with medically resistant focal seizures. Assessment should be performed by a team with expertise in epilepsy Ictal semiology and the EEG may be difficult to interpret in young children;8,9 focal seizures may occur in specific epilepsy syndromes may be apparent. Certainly all children with a focal or lateralized structural brain abnormality should be referred for surgical consideration at diagnosis, even though response to medication may be unclear at that point. It is also has to be considered that incomplete myelination may make MRI difficult to interpret; just as abnormalities may not become apparent until myelination is complete, lesions may also “disappear” despite being apparent on early scans as myelination progresses.10
An 8-month-old boy presents with a history of asymmetric spasms. His first seizure occurred on day 1; his mother reported both his legs would draw up flexed, and he would look uncomfortable. This was originally diagnosed as infantile colic, but by 4 weeks of age the movements became more prominent and increased in frequency. They evolved into recognizable spasms; asymmetric posturing with head turning to the left then right, with fisting of the right hand. Clusters would occur on waking, for up to 20 minutes five to six times per day. Some reduction was noted with vigabatrin with subsequent resolution with steroids. At 8 months, he demonstrates a left-hand preference with evidence of a right hemiparesis, and visual inattention. MRI revealed left hemimegalancephaly (see Fig. 55–2).
Despite apparent seizure control, developmental assessment at 7 months suggested right visual field loss, right hemiplegia with nonverbal skills at 2–3 months age equivalent across all scales. His EEG showed evolution from asymmetric burst to continuing spike–wave discharges over the left hemisphere.
Although clinical seizures remain controlled, his EEG was markedly abnormal and he remained developmentally delayed. Neurological examination revealed a right hemiplegia and right visual field defect, with inattention of vision to the right. He is at high risk for return of seizures, and for continued developmental compromise. Hemidisconnection was offered to reduce the risk of recurrent seizures and optimize neurodevelopmental outcome. Twelve months following surgery he remains seizure free on medication and is at a 6-month developmental level.
A 22-month-old boy presented at age 15 months with a first seizure. He was found motionless and staring in his buggy; he was hypotonic when lifted. This lasted a few minutes. An episode of abnormal jerking movements was noted the following day. Further episodes consisted of “head nods,” singly or in clusters. Although he retained awareness, in some episodes he had a puzzled expression on his face, and in protracted clusters would tremble. At time of assessment he was experiencing 2–3 clusters per day, each involving 5–10 “spasms.”
His neurodevelopment was normal until the time of seizure onset. Three months after seizure onset he underwent a marked regression of cognitive skills to a 6-month developmental level with loss of social interaction. There was a clinical suggestion of left visual inattention. He was treated with vigabatrin, levetiracetam, and oxcarbazepine without success.
MRI revealed an area of focal cortical dysplasia in the right mid temporal region. Interictal EEG showed multifocal sharp waves on a slow background, with variable asymmetry. A burst suppression pattern was evident in sleep (Fig. 55–3) with no consistent lateralization of seizure onset on ictal EEG recording. Semiology was also variable, but consistently included behavior arrest consistent with a temporal lobe onset prior to clusters of “spasms.” Subsequent MRI evaluation demonstrated abnormality of the entire right temporal lobe extending posteriorly into the occipital region.
Interictal electroencephalography (EEG) of case 2 (A) demonstrating burst suppression pattern seen in sleep, and spasm with lateralizaiton to the left on ictal recording (B). A dysplastic lesion of whole temporal lobe on the right with extension into the occipital lobe is demonstrated on magnetic resonance imaging (C).
This case illustrates early onset catastrophic epileptic encephalopathy with focal onset. There was little likelihood of response to medication. Seizure onset was compatible with the MRI lesion, felt to be multilobar. Therefore, he proceeded to right tempo-occipital resection, maximizing chance of seizure freedom with counseling for a definite visual field defect, and unpredictable prognosis for cognitive outcome. He underwent a right temporo-occipital disconnection and remains seizure free postoperatively at 12 months, off medication, with increased interest and developmental level.