Indications

Hemidecortication has been performed in a variety of clinical settings, ranging from congenital to acquired conditions. The overall goal in such a surgery is to remove cortex with pathology while allowing underlying white and deep gray matter to remain. This may be particularly useful when abnormal tissue is primarily located in the cortical gray matter. Given that seizures are generated by abnormal neuronal activity, hemidecortication may be useful for removing only the most affected tissue while preserving structures that still may be functioning normally.

In one large series of hemidecortications, the most common underlying pathology was Rasmussen syndrome.4 Rasmussen syndrome is a progressive inflammatory condition of the brain characterized by intractable epilepsy and neurological impairments.5 With a typical picture of early childhood onset, unilateral brain involvement, and a progressive nature, this syndrome appears to be quite amenable to hemispherectomy if medical management fails. The etiology of Rasmussen syndrome is not known; suggestions include a viral illness (including Epstein-Barr virus, cytomegalovirus, and herpes simplex-1 virus), antibodies (case reports and series include patients with antibodies against one sub-unit of the ionotropic glutamate receptor, GluR3, and the presynaptic protein munc-8), or T-cell dysregulation.5,6 No unifying hypothesis has emerged however, despite years of effort.7 Clinically, the median age of presentation is 5 years, although adults present with Rasmussen syndrome, as well.8 Three stages have been described: a prodrome characterized by seizures and a mild hemiparesis, followed by an acute stage characterized by more frequent seizures and progressive neurological deficits referent to one hemisphere, then a residual stage with fewer seizures but persistent functional deficits.5 Any type of seizure may be seen, but focal motor seizures (including epilepsia partialis continua), sometimes appearing to represent a noncontiguous patchy distribution (e.g., face and leg), are most frequently noted.7 Electroen-cephalography (EEG) can be confusing (i.e., abnormalities are sometimes seen bilaterally), but imaging typically shows unihemispheric atrophy over time.7 The patchy nature of the seizures probably reflects the distribution of findings when pathological tissue is examined: perivascular cuffing, microglial nodules, and neuronal loss interspersed with areas of normal-appearing tissue.9 This patchy involvement is one reason we avoid biopsies to diagnose Rasmussen syndrome in our practice. Diagnostic criteria based on a variety of clinical and paraclinical findings have been proposed.5 Medical therapy includes anticonvulsants for controlling seizures (although this practice frequently is unsuccessful), immuno-therapy (including intravenous immune globulin, cortico-steroids, or plasmapheresis), and immune modulators (e.g., tacrolimus).5 Some patients are helped to a limited extent by medical management, but the most effective therapy for control of the seizures (or relief from seizures in the context of unacceptable medication side effects), particularly if there is evidence of unilateral hemispheric dysfunction (e.g., hemiparesis, language involvement), is hemispherectomy.

Another common indication for hemidecortication is malformations of cortical development.4 Cortical development can be affected adversely at any stage of development: cell proliferation, neuronal migration, or cortical organization.10 Anomalies can range in size from heterotopias involving small areas of cortex to hemimegalencephaly, which involves the entire hemisphere. In addition to intractable epilepsy, these children may have developmental disabilities, contralateral motor deficits, associated neurocutaneous signs (e.g., epidermal nevi), and craniofacial anomalies (e.g., hemicrania). In those malformations not typically limited to one hemisphere, one of the greatest challenges in evaluation is ensuring that the contralateral hemisphere functions reasonably well. Some patients with Rasmussen syndrome have evidence of malformations of cortical development.9 This might partly explain why EEG findings can be seen bilaterally; it also makes the preoperative assessment of risks and benefits critical in determining the likelihood of postoperative success.

Vascular injury (i.e., perinatal stroke) is another common pathology leading to hemidecortication.4 Ischemic (arterial and venous) and hemorrhagic strokes may be seen in the neonatal period and can lead to seizures.11 A variety of prenatal and perinatal events, coagulopathies, congenital cardiac lesions, infections, and metabolic abnormalities can lead to cerebrovascular cortical lesions that may eventually be amenable to surgical resection for control of seizures.

Much of the early literature on hemidecortication highlighted patients with Sturge-Weber syndrome.12,13 Sturge-Weber syndrome is characterized by facial port-wine stains associated with leptomeningeal venous malformations and intracranial calcification in a gyriform pattern.14 Many patients also suffer from intractable epilepsy, developmental delays, glaucoma, and cerebrovascular complications. Indications for and timing of hemispherectomy are somewhat controversial: patients with seizures starting younger than 1 year of age appear to have more intractable seizures, yet it is difficult to make early predictions of how intractable a given patient’s seizures will become over time.14 There is no solid evidence to indicate the best timing or the optimal choice of surgery (i.e., hemispherectomy versus focal resection), although seizure freedom is higher after hemispherectomy than after focal resections (although nonrandom choice of patients for each procedure may be a major confounder of this finding).15,16

Other pathologies in which hemidecortication may be useful include large tumors and tuberous sclerosis, both of which involve mass lesions associated with epileptogenic cortex, sometimes in the context of associated malformations of cortical development.

Presurgical Planning

The most important first step in the evaluation of a candidate for hemidecortication is diagnosing the correct epilepsy syndrome. When considering the removal of one hemisphere, several questions arise concerning the expected benefits of the surgery (i.e., seizure control and improved developmental outcomes after relief from seizures and medication side effects), potential risks (i.e., loss of full function in the contralateral hand and possibly partial visual field loss), freedom of the remaining hemisphere from seizure activity, the ability of the remaining hemisphere to assume new functions postoperatively, and finally, the timing of the surgery (i.e., early in life versus later).17 Several modalities exist to make some assessment of these parameters, although none have been completely predictive, based on clinical experience.18 In the ideal situation, both structural and functional studies point to the remaining hemisphere as structurally normal, free from seizures, and fully capable of accommodating new function. Unfortunately, this is not always the case.

Imaging studies such as magnetic resonance imaging and computed tomography aid in the assessment of the structural characteristics such as the extent of pathology (i.e., involvement of gray or white matter, single versus multiple lobar involvement, and possibly involvement of the more normal side).18 Magnetic resonance spectroscopy may contribute to the localization of involved cortex but rarely is the only abnormal study in a patient’s evaluation. Functional magnetic resonance imaging may be useful for delineating the location of areas involved in critical language and motor function, particularly in patients with lesions thought to originate early in development, because their localization may be anomalous.19 Positron emission tomography (PET) using radiolabeled fluorodeoxyglucose may identify areas of relative hypometabolism, indicating the extent of metabolic dysfunction in various regions.20 We routinely obtain a concurrent EEG during PET studies to avoid false-negative tests (i.e., abnormal areas with epileptic foci that are active during the study-these areas may be more metabolically active than at baseline during an extended run of epileptiform activity, which would make them appear to be more normal than they actually are). EEG provides a functional picture of the source of epileptiform activity and can be useful for defining both the electrical source of seizures as well as one indication of the general electrical health of the presumably normal hemisphere before operation. Extended video-EEG monitoring in an epilepsy monitoring unit may be particularly useful if multiple different seizure types need to be characterized or if the nature of different types of events needs to be studied in the context of a question about whether a focal resection or hemispherectomy should be performed. Developmental screening or neuropsycho-logical testing are invaluable tools in the assessment of a patient’s current level of cognitive and language function. They may play a role in preparation of the patient for postoperative rehabilitation and serve as a baseline comparison for the postoperative evaluation.20 We do not routinely use other studies such as the intracarotid amobarbital (Wada) test and magnetoencephalography in patients being considered for hemidecortication, given the high prevalence of developmental disabilities, lack of cooperation, and the fact that these studies rarely make a meaningful contribution to surgical decisions in this population. The importance of the assessment of the more normal side using various modalities cannot be overemphasized in their importance for pre-surgical counseling and prognostication ( Table 26.1 ).