Surgical Pathologic Findings of Extratemporal-Based Intractable Epilepsy

The surgical treatment of epilepsy is now an established mode of therapy in various clinical settings.1 Engel estimated that as of the early 1990s more than 8000 patients had undergone a surgical procedure as the definitive treatment for pharmacoresistant seizure disorder.2 As Engel emphasized, only a small fraction of surgical candidates are at present referred to epilepsy surgery centers, meaning that a huge treatment opportunity is being missed.Until the 1990s, anterior temporal lobectomy, or amygdalohippocampectomy, accounted for more than two thirds of all surgical procedures used to treat epilepsy.2 Within the past 10 years, focal neocortical resections, ‘lesionectomies,’ and hemispherectomies have been used increasingly more often to treat intractable seizures in some of the patients most refractory to drug therapy, that is, infants and children with infantile spasms and intractable seizures that begin in the first years—often the first weeks—of life.3,4 This radical form of neurosurgical intervention has provided pathologists with unique and challenging tissue specimens, whose range was well described in an article by Frater et al,5 which appeared in the April issue of the Archives. Understanding the pathogenesis and prognosis of lesions within such specimens may tell us a great deal about the neurobiology of epileptogenesis.Frater et al nicely summarize the variety of lesions encountered among many epilepsy surgery centers. The relative frequency of the different specimen types varied tremendously among centers adept at this complex form of surgical intervention, often depending on local expertise and interests. As with all studies of archival material, the pathologic specimens were obtained over a long period of time from patients who were operated on over a broad age range, from 3 months to 57 years. As one would expect, the material available for study from different resections was tremendously variable in amount and was probably examined using a variety of special and immunohistochemical stains.Cortical dysplasia6 (CD) was noted to be the major neuropathologic abnormality in more than a third of cases. Neoplasms (usually of an indolent nature) were found in more than a quarter of patients, and evidence of ‘remote’ encephalomalacia was identified in less than a fifth. The etiology of these destructive lesions almost always remains obscure, even after they have been studied in great detail, but they are presumed (perhaps incorrectly) to represent sequelae of intrauterine or perinatal/neonatal ischemia in the brain. The developing central nervous system is vulnerable to a variety of vascular and hypoxic-ischemic insults, including periventricular white matter necrosis, germinal matrix hemorrhage, and venous thrombosis. The precise etiology of a given region of encephalomalacia may be impossible to discern when it is examined weeks, months, or years after the precipitating event.7–9 Furthermore, toxic and infectious processes may cause tissue destruction, which closely mimics that caused by vascular lesions.10 Small numbers of cases of vascular malformation, Sturge-Weber disease, and Rasmussen encephalitis were also found.5Of interest, just less than 20% of specimens showed no discernible pathologic abnormality. As Frater et al appropriately surmise, the origin of seizures in these cases may be at a cellular or molecular level that has no obvious morphologic correlate. However, it is also possible that evaluation of the tissue using rigorous morphometric techniques might reveal structural abnormalities not apparent on routine light microscopic examination. Such methodologies are currently beyond the range of activities in most diagnostic pathology laboratories, although they are likely to play a major role in the future.11,12The high incidence of malformative and destructive lesions studied by Frater et al5 reinforces the impression from pediatric epilepsy surgery centers that these structural changes (sometimes described as neuronal migration disorder [NMD]) most commonly account for intractable epilepsy (including infantile spasms/West syndrome) in infants and children.13–17 The authors did not, unfortunately, substratify patients by the nature of their seizure disorder or by their age (at either presentation or surgery). It would also have been of interest to know how effective epilepsy surgery was in the context of each type of neuropathologic abnormality encountered (ie, in currently fashionable terminology, what was the outcome of surgical intervention in a malformative vs a destructive or neoplastic lesion).While identifying morphologic abnormalities in extratemporal corticectomies from epileptic patients is both valuable and of importance for assessing prognosis, a note of caution is in order. An epilepsy-associated lesion, bizarre though it may appear (as in many cases of CD), might have little to do with the cause(s) of a given seizure disorder. Rather, it may simply represent a marker for underlying brain disease that is the proximate cause of epilepsy. Patients with tuberous sclerosis (TSC) may, for example, have several cortical tubers of similar morphologic appearance within the brain, yet only one of these may be epileptogenic.18 In CD/NMD, abnormalities of expression of neurotransmitters/neurotransmitter receptors on aberrant neurons may represent a ‘molecular lesion,’ which (rather than the structural lesion itself) predisposes to seizure genesis.19–21Cortical dysplasia is perhaps the most neurobiologically intriguing of the lesions described in the study by Frater et al, and in many other articles on extratemporal resections for seizure disorder as well. As the authors indicated, both the gross and microscopic morphologic appearances of CD can be remarkably variable, possibly reflecting the timing of the (presumed) intrauterine insult that resulted in anomalous neuronal migration to, or maturation within, the neocortex.6 In its most extreme form, CD may be associated with hemi-megalencephaly/hemi-lissencephaly.22 By light microscopy, severe CD is often associated with profound neuronal cytoskeletal abnormalities.23 Frater et al appropriately pointed out the possible linkage or continuum between CD/NMD and tumors that appear to have a glioneuronal malformative component, for example, gangliogliomas and dysembryoplastic neuroepithelial tumors.Perhaps an even more intriguing etiologic connection exists between CD/NMD and TSC. Cortical tubers of TSC may give rise to epilepsy and morphologically show a striking resemblance to foci of severe CD, including ‘balloon’ cell change, neuronal cytomegaly and dysmorphism, and profound cytoskeletal abnormalities.6,18,24 Tuberous sclerosis-associated genes (TSC1, TSC2), mutations in which determine the TSC phenotype (including brain lesions), have recently been cloned.25,26 Both TSC1 and TSC2 transcripts and their encoded proteins, hamartin and tuberin, respectively, are widely and abundantly expressed in viscera and the developing and mature healthy central nervous system, as well as in TSC brain lesions.27–30 Precisely how they determine the formation of multifocal dysplastic lesions in the central nervous system is not, however, understood.Compiling a description of extratemporal neocortical lesions encountered among surgical resection specimens obtained in the course of treating epilepsy represents an important starting point for understanding the potential neurobiologic significance and impact of these lesions on brain function. Making use of this material for mechanistic studies aimed at understanding seizures represents the next logical step in this analysis.