- Email: [email protected]
Histopathologic findings in 37 cases of functional hemispherectomy
Histopathologic Findings in 37 Cases of Functional Hemispherectomy Richard A. Prayson, MD, William Bingaman, MD, John L. Frater, MD, and Elaine Wyllie, MD Hemispherectomy procedures are performed in patients for whom focal cortical resection would be predicted to produce a significant reduction in seizures. The functional hemispherectomy procedure consists of disconnecting the hemispheres while attempting, in some cases, to preserve parenchyma. This study retrospectively reviews the histopathologic findings in 37 cases of functional hemispherectomy performed between 1990 and 1998 at a major epilepsy center. Procedures were performed in 20 males and 17 females who ranged in age from 3 months to 37 years (mean age, 9.6 years). In all but two cases, more than half or all the material submitted for pathologic testing was examined histologically. Cortical dysplasias or hemimegalencephaly were identified in 14 patients. The most common patterns of dysplasia observed included architectural disorganization (n = 13), increased molecular layer neurons (n = 11), and neuronal cytomegaly (n = 11). One patient was known to have epidermal nevus syndrome. Six patients had Sturge-Weber syndrome. Remote infarct/ischemic damage was identified as the etiology of seizures in six patients; four of these patients had mild associated secondary cortical architectural abnormalities. Three patients demonstrated pathology consistent with Rasmussen's encephalitis; one additional patient had chronic encephalitis changes, not otherwise specified. In two cases, changes consistent with hippocampal sclerosis were identified; additionally, hippocampal neuronal loss and gliosis was focally identified in three patients. Most of these patients had coexistent cortical dysplasia or radiographic evidence of remote infarct. One specimen demonstrated areas of infarct following resection of an arteriovenous malformation. In two specimens, significant histopathologic findings were not identified; both of these patients had radiographic evidence of remote infarct. The spectrum of pathologic conditions that may be encountered in the setting of a functional hemispherectomy is varied and in this study most frequently included cortical dysplasia, Sturge-Weber syndrome remote infarct, and Rasmussen's encephalitis. Ann Diagn Pathol 3: 205-212, 1999, Copyright © 1999 by W.B. Saunders Company
Index Words: Cortical dysplasia, epilepsy, functional hemispherectomy, hemispherectomy, Rasmussen's encephalitis, Sturge-Weber syndrome "EMISPHERECTOMY procedures have been per,formed for a variety of conditions since the earliest part of the 20th centmy.I,2 More recently, hemispherectomy procedures have been used predominantly for patients with chronic epilepsy due to conditions involving all or most of one hemisphere in which a smaller
H
FromtheDepartmentsof'AnatomicPathology,Neurosurge~,andEpilepsyand SleepDisorders,ClevelandClinicFoundation, Cleveland,OH. Presentedat the UnitedStates and CanadianAcademyofPathologymeeting, San Frandsco, CA, March 1998. Addressreprintrequeststo RichardA. Prayson,MD, DepartmentofAnatomic Pathology(L25),ClevelandClinicFoundation,9500EuelidAve, Cleveland,OH 44195. Copy@t © 1999by ~d~B.SaundersCompany 1092-9134/99/0304-0001510.00/0
resection will not remove all the epileptogenic tissue. The procedure has proven effective in reducing seizure frequency. Unfortunately, a significant subset of patients who had complete anatomic hemispherectomy eventually developed complications related to superficial cerebral hemosiderosis) ,4 To minimize complications associated with superficial cerebral hemosiderosis syndrome, the hemispherectomy procedure was modified to involve a subtotal resection of the hemisphere associated with a disconnection of the remaining areas from the rest of the brain (so-called "functional hemispherectomy"). Relatively few series have specifically focused on the histopathologic findings in functional hemispherectomy specimens. In most previous reports, functional hemi-
Annals of Diagnostic Pathology, Vol 3, No 4 (August), 1999: pp 205-212
205
206
Prayson et al
spherectomy-associated resections are grouped together with other hemispherectomy-associated resections and extratemporal corticectomies. This report reviews one institution's experience with 37 consecutive functional hemispherectomies in a predominantly pediatric-aged population. Details of the histopathologic findings in these cases are reviewed. Materials and Methods
A search of the neurosurgicai case flies from 1990 through April 1998 was performed to identify patients who underwent functional hemispherectomy. The functional hemispherectomy procedure was performed as described by Rasmussen and Viilemure. 5-7 The surgical pathology files were then searched for cases in which there was tissue excised at the time of surgery. Thirty-seven patients who underwent functional hemispherectomy during this period had excised tissue examined histologically; these patients comprised this study group. All available histologic material was examined in each case. The number of histologic slides reviewed ranged from 1 to 48 (mean, 9.8 slides). Half or more of the tissue submitted for pathologic evaluation was examined microscopically in 35 of 37 cases; in the remaining two cases, less than half of the tissue submitted was examined histologically. Histologic sections were routinely generated from formalin-fixed, paraffinembedded tissues sectioned 4 to 5 ~rn in thickness and stained with hematoxylin-eosin. The histologic features observed in each case were documented. Cortical dyspiasia was defined as a maldevelopmental abnormality of the cortex relative to the neuronal component.8 Histologic classification of the cortical dysplasia patterns observed were partially based on those previously outlined by Mischel et al in 1995.9 Heterotopic white matter neurons situated in deep white matter were noted when identified. Cases in which glial-neuronal tissue was situated in the pia arachnoid region were also noted when identified. Other lesions that were identified and noted included the presence of remote ischemic damage or infarct. Rasmussen's encephalitis was identified according to the previous histologic descriptions of this lesion. 1°,11Evidence of Sturge-Weber disease, as previously described, was also documented when noted.12,1s Hippocampal sclerosis, when present, was identified and characterized by segmental loss of neurons and gliosis in regions of the Ammon's horn, particularly involving CA1/ prosubiculum, dentate gyrus, and CA4 regions. 14,15 Other findings that were documented when noted included the presence of acute and/or chronic leptomeningeal inflammation, meningeal fibrosis, and evidence of recent infarct and/or contusion. Results
Clinical Information The study included 20 males and 17 females. The patients ranged in age from 3 months to 37 years (mean,
9.6 years) at the time of surgery. Five patients were over the age of 21 years at the time of initial surgery. One patient was known to have epidermal nevus syndrome.
Histopathologic Findings The histopathologic findings in the 37 patients are presented in Table 1, Cortical dysplasia was observed in 14 resections. The single most common pattern of dysplasia observed was that of architectural disorganization consisting of either abnormalities of cortical lamination or a haphazard arrangement of neuronal cells within the cortex (Fig 1). This finding was seen in 13 of 14 cases. In 11 cases, increased numbers of molecular layer neurons were identified. Cytomegalic neuronal cells were focally noted within the cortex in 11 cases (Fig 2). Evidence of gyral fusion was seen in four cases. Discrete areas of gray matter situated within deep white matter was also noted in four resections. Pachygyric cortex was identified in three resections. Balloon cells characterized by abundant eosinophilic cytoplasm and round to slightly oval nuclei with prominent nucleoli were identified focally in three cases (Fig 3). Occasional balloon cells were multinucleated. Discrete areas of heterotopic white matter were observed in the cortex in two cases. Discrete clustering of atypical neuronal and glial cells was observed situated in the cortex in one case. Foci of glial-neuronal tissue situated within the pia arachnoid region were seen in four cases (Fig 4). Heterotopic neurons situated in deep white matter were identified in 17 cases. In all but one patient, more than one of the above-mentioned patterns of dysplasia were noted to coexist. Histologic evidence of remote ischemic injury or infarct was observed in six cases involving the middle cerebral artery distribution. These areas were characterized by focal cystic degeneration of the parenchyma Table 1, Summary of Histopathologic Findings in 37 Functional Hemispherectomies N (%) Cortical dysplasia Infarct/remote ischemia Sturge-Weber disease Rasmussen's encephalitis Hippocampal neuronal Ioss/gliosis* Hippocampal sclerosis* No significant pathologic findings1Chronic encephalitis, not otherwise specified
14 (37.8) 6 (16.2) 6 (16.2) 3 (8.1) 3 (8.1) 2 (5.4) 2 (5,4) 1 (2.7)
*Most of these patients had coexistent cortical dysplasia or radiographic evidence of remote infarct. tBoth of these patients had radiographic evidence of remote infarct, which was the presumed cause of seizures.
Histopathology of Functional Hemispherectomy
Figure 1. Loss of the normal laminar architecture of the cortex with gliosis involving cortical layers 1 to 3 and prominent subpial gliosis. (Hematoxylin-eosin stain; medium power.)
Figure 2. Cytomegalic cortical neurons in a cortical dysplasia involving layer 5. (Hematoxylin-eosin stain; high power.)
207
associated with pericystic gliosis (Fig 5). In four of these cases, adjacent cortical dysplasia-like changes were also observed; these were presumed in some cases to be secondary to the ischemic/infarctive event. In three of these four cases, focal architectural disruption with loss of the normal lamination pattern and a haphazard arrangement of neuronal cells was observed. In two cases, a slight increase in the number of molecular layer neurons was noted. In at least one of these cases, the infarct was known to be associated with previous surgery that involved excision of an arteriovenous malformation. In six resections, changes consistent with those described in Sturge-Weber disease were noted. These changes involved a proliferation of predominantly venous-type vessels in the leptomeninges (Fig 6). Increased numbers of vessels were also noted within the cortex, associated with extensive perivascular and parenchymal mineralization of both cortex and less notably of underlying white matter. Increased gliosis was also noted within the cortex. In three resections, clinicopathologic features consistent with Rasmussen's encephalitis were identified. The histopathologic findings in these cases were somewhat similar and involved the presence ofperivascular chronic inflammation in both the leptomeninges and focally surrounding parenchymal vessels associated with a diffuse microglial cell proliferation, focal formation of more discrete microglial nodules, and a mild nonspecific reactive astrocytosis (Fig 7). Specific viral inclusions were not identified in any of these cases. One additional resection demonstrated similar histopathologic findings but lacked the typical clinical and radiographic features
208
Prayson et al
Figure 3. Scattered balloon cells with abundant eosinophilic cytoplasm in a dysplasia. (Hematoxylin-eosin stain; high power.)
of Rasmussen's encephalitis. Viral inclusions were also not identified in this case by light microscopy. Histopathologic findings ofhippocampal sclerosis were definitely identified in two cases. In one of these cases, a concomitant increased number of molecular layer neurons (cortical dysplasia) was observed. In the other case, a coexistent remote infarct was identified. In three additional hippocampal resections, focal areas of neuronal loss and gliosis were identified. Due to the fragmentation and orientation of these particular specimens, a definite diagnosis of hippocampal sclerosis was precluded, although the possibility that these lesions represented that process could not be excluded. In one of these three cases, coexistent cortical dysplasia character-
Figure 4. Glial-neuronal tissue located in the pia-arachnoid region. (Hematoxylin-eosin stain; medium power.)
ized by gray matter heterotopia and a focally disrupted cortical architecture pattern were identified. The other two patients had evidence of remote infarct on imaging studies. No significant histopathologic findings were identified in two limited resection specimens. Both of these patients had evidence of remote left middle cerebral artery infarcts radiographically. Nonspecific subpial gliosis was observed in 31 of 37 resections. Focal areas of perivascular atrophy, most notably invoMng white matter vessels, was seen in 13 cases. Excluding the four cases of encephalitis, focal areas of chronic leptomeningeal inflammation were seen in eight cases and meningeal fibrosis in 11. Most of these patients had a prior history
Histopathology of Functional Hemispherectomy
209
Figure 5. An area of cystic degeneration associated with a remote infarct. (Hematoxylin-eosin stain; medium power.) of recent invasive seizure monitoring. Focal leptomeningal acute inflammation was observed in one resection. Discussion
Figure 6. Leptomeningeal vascular proliferation in SturgeWeber disease. (Hematoxylin-eosin stain; low power.)
The single most common histopathologic finding in this series is that of cortical dysplasia, which was observed in slightly more than one third of the resections. This was not surprising given the nature of the lesion, which can at times be quite extensive to the extreme of hemimegalencephaly. Although comparable series specifically documenting the histopathologic features in functional hemispherectomy cases are uncommon, a fairly comparable incidence of dysplasia was observed in a series of children who underwent hemispherectomy or multilobar cortical resection as treatment for chronic epilepsy at UCLA Medical Center. 16 Cortical dysplasia also was one of the most common etiologies among children and adolescents who had epilepsy surgery at the Cleveland Clinic Foundation.': Farrell et a116 reported evidence of malformations or hamartomatous lesions in 15 of 38 resections (39.5%). In a series of 58 hemispherectomy cases] a 26 of which were functional hemispherectomies, cortical dysplasia, likewise, was the single most common pathologic finding in 33 of 58 patients (57%). The most common patterns of dysplasia observed in this series included architectural laminar or neuronal orientation abnormalities, increased numbers of molecular neurons, and evidence of neuronal cytomegaly. In Mischel et al's 9 evaluation of 77 cortical dysplasia cases derived from a variety of surgeries, the single most common pattern of dysplasia was cortical laminar disorganization, which was observed in
210
Prayson et al
Figure 7. Rasmussen's encephalitis marked by perivascular chronic inflammation and reactive astrocytosis. (Hematoxylin-eosin stain; medium power.)
all cases. According to their proposed grading system, Mischel et aP suggested that neuronal cytomegaly most probably represented early developmental disruptions (occurring during the first trimester of development), whereas cortical laminar disorganization and increased numbers of molecular layer neurons can arise at any time during development, but the most severe forms probably arise earlier during development. It therefore appears that the more severe and extensive cases of cortical dysplasia, which may require a functional hemispherectomy procedure, are more likely to arise from migration defects that occur earlier during development. It appears that the extent of dysplasia and its specific location may be at least, if not more, significant than the histologic type observed with regard to seizure severity. In the vast majority of cases, multiple patterns of coexistent cortical dysplasia were encountered in the same resection. The etiologies of cortical dysplasia still remain uncertain and they likely arise from a variety of conditions that disrupt the process of neuronal migration and differentiation,a,~9 Defects involving any one of a number of factors that modulate neurogenesis, including involvement of the radial glia, programmed cell death, nerve growth factors, mechanisms controlling cell proliferation, neuronal developmental genes, hormonal factors, and cell adhesion molecules, may potentially give rise to abnormalities that affect neuronal migration and result in cortical dysplasia. 19 Superimposed on these factors are environmental conditions, such as ischemic or infectious states, that may potentially alter neurogenesis. Of particular interest in this regard are the cases in which there was clear histologic evidence of a remote,
likely perinatal, ischemic event associated with adjacent architectural disorganization. Whether one chooses to use the term cortical dysplasia for such abnormalities may be a matter of semantics. Cortical dysplasia is generally thought of as a maldevelopmental-associated lesion. It is quite likely that at least a subset of lesions we currently recognize as dysplasia are the result of ischemic events that occurred in utero during the development of the brain but were not severe enough to cause a cavitary lesion. The presence of pathologic patterns virtually identical to those currently recognized as cortical dysplasia adjacent to areas of infarct raise interesting questions regarding the nature of dysplasia and its precise definition. Patients with Sturge-Weber disease likewise have been previously reported in most hemispherectomy series. It was a relatively common etiology in the Cleveland Clinic pediatric epilepsy series. 17 Farrell et a111 noted that three patients had pathologic findings associated with Sturge-Weber disease, u Two of 58 hemispherectomy specimens reported by Peacock et aP a showed changes consistent with Sturge-Weber syndrome. Because of the generally more extensive distribution of pathology in Sturge-Weber syndrome, many of these patients tend to benefit from a hemispherectomy procedure and have generally excellent seizure control following surge~. 2° Another commonly encountered pathologic finding in hemispherectomy series is that of Rasmussen's-type encephalitis. The extensive involvement of the process (usually confined to one hemisphere) and the progressive nature of the disorder both lend themselves to this surgical approach. Numerous studies have examined a
Histopathology of Functional Hemispherectomy
variety of viral agents in an attempt to pinpoint a culprit in terms of the etiology of Rasmussen's encephalitis. Although occasional reports have demonstrated evidence of cytomegalovirus, Epstein-Barr virus, or herpes simplex virus in patients with Rasmussen's encephalitis, this has been a variable finding, n,~1-23 A fairly recent study has found evidence ofautoantibodies to glutamate receptor (GluR3) in these patients, suggesting that an autoimmune process may be in part responsible for the pathologic finding. 24 Certainly, the possibility that a viral infection may trigger this autoimmune mechanism cannot be excluded. Alternately, the somewhat nonspedfic histologic findings in patients with Rasmussen's encephalitis may be due to a variety of etiologies that happen to share a similar pathology. A recent report of a group of patients who developed seizures during adolescence or adulthood with identical histopathologic findings to those found in patients with Rasmussen's encephalitis demonstrated more variable clinical characteristics. 25 The precise relationship of these types of patients with the classic Rasmussen's syndrome is not well understood and indicates we still have much to learn. Hippocampal sclerosis is well known as the most common pathologic finding found in the majority of patients with temporal lobe epilepsy. 14,26 Generally, however, one would not expect the patient to require a more extensive functional hemispherectomy procedure for hippocampal sclerosis. Of particular interest is the fact that in both patients with hippocampal sclerosis, a significant coexistent lesion was present. In one patient, coexistent cortical dysplasia was identified and in the other patient, evidence of remote infarct was noted on imaging studies. The exact relationship between dysplasia and hippocampal sclerosis is unknown. 15,27,28 Whether these lesions merely coexist or represent two different phenotypic manifestations of a common event that occurred in utero or whether one lesion may beget the other is not certain. In two cases, no signifcant pathologic findings were noted by light microscopy; however, both patients had radiographic evidence of infarct. The etiology of seizures in these cases is presumably related to these lesions, which were not excised at the time of surgery.
Acknowledgment The authors thank Denise Egleton for her help in the preparation of this manuscript.
211
References 1. Tinuper P, Andermann F, Villemure J-G, et ah Functional hemispherectomy for treatment of epilepsy associated with hemiplegia: Rationale, indications, results, and comparison with callosotomy. Ann Neurol I988;24:27-34 2. Dandy W: Removal of right cerebral hemisphere for certain tmnors with hemiplegia: Preliminary report.JAMA 1928;90:823-825 3. Oppenheimer DR, Griffith HB: Persistent intracranial bleeding as a complication of hemispherectomy.J Neurol Neurosurg Psychiatry 1966;29:229-240 4. Rasmussen T: Postoperative superficial hemosiderosis of the brain, its diagnosis, treatment and prevention. Trans Am Neurol Assoc 1973;98:133-137 5. Rasmussen T: Hemispherectomy for seizures revisited.J Can Sci Neurol 1983;10:71-78 6. Rasmussen T, Villemure JG: Cerebral hemispherectomy for seizures and hemiplegia. Cleve ClinJ Med 1989;56:562-568 7. Villemure JG: Hemispherectomy: Techniques and complications, in Wyllie E (ed). The Treatment of Epilepsy-: Principles and Practice. Philadelphia, PA, Lea & Febiger, 1993, pp 1116-1119 8. Prayson RA, Estes ML: Cortical dysplasia: A histopathologic study of 52 cases of partial lobectomy in patients with epiiepsy. Hum Pathol 1995;26:493-500 9. Mischel PS, Nguyen LP, Vinters HV: Cerebral cortical dysplasia associated with pediatric epilepsy. Review of neuropathologic features and proposal for a grading system. J Neuropathol Exp Neurol 1995;54:137-153 10. Rasmussen T, OlszewskiJ, Lloyd-Smith D: Focal seizures due to chronic localized encephalitis. Neurology 1958;8:435-445 11. Farrell MA, Droogan O, Secor DL, et ah Chronic encephalitis associated with epilepsy: Immunohistochemical and ultrastructural studies. Acta Neuropathol 1995;89:313-321 12. DiTrapani G, DiRocco C~ Abbamondi AL, et ah Light microscopic and uhrastructural studies of Sturge-Weber disease. Childs Brain 1982;9:23-36 13. Prayson RA, GrewaI 1D, MdvlahonJT, et ai: Leukocyte adhesion molecules and x-ray dispersive spectroscopy in Sturge-Weber disease.J Pediatr Neurol 1996;15:332-336 14. Babb T, Brown W: Pathological findings in epilepsy, in EngelJJ (ed). Surgical Treatment of the Epilepsies. New York, NY, Raven, 1987, pp 511-540 15. Prayson RA, ReithJD, Najm IM: Mesial temporal sclerosis: A clinicopathologic study of 27 patients, including 5 with coexistent cortical dysplasia. Arch Pathol Lab Med 1996;120:532-536 I6. Farrell MA, DeRosa MJ, Curran JG, et ah Neuropathologic findings in cortical resections (including hemispherectomies) performed for the treatment of intractable childhood epilepsy. Acta Neuropathol 1992;83:246-259 17. Wyllie E, Comair YG, Kotagal P, et al: Seizure outcome after epilepsy surgery in children and adolescents. Ann Neurol 1998;44:740748 18. Peacock WJ, Wehby-Grant MC, Shields WD, et ah Hemispherectomy for intractable seizures in children: A report of 58 cases. Childs Nerv Syst 1996;12:376-384 19. Rorke LB: The roIe of disordered genetic control of neurogenesis in the pathogenesis of migration disorders. J Nenropathol Exp Neurol 1994;33:105-i 17 20. Hoffman HJ, Hendrick EB, Dennis M, et al: Hemispherectomy for Sturge-Weber syndrome. Childs Brain 1979;5:233-248
212
Prayson et al
21. Jay V, Becker LE, Otsubo H, et ah Chronic encephalitis and epilepsy (Rasmussen's encephalitis): Detection of cytomegalovirus and herpes simplex virus I by the polymerase chain reaction and in situ hybridization. Neurology 1995;45:108-I 17 22, Power C, Poland SD, Blume WT, et ah Cytomegalovirus and Rasmussen's encephalitis. Lancet 1990;336:1282-1284 23. Atkins MR, Terrell W, Hulette CM: Rasmussen's syndrome: A study of pot ential viral etiology. Clin Neuropathol 1995;14:7-12 24. Rogers SW, Andrews PI, Gahring LG, et al: Autoantibodies to glutamate receptor GluR3 in Rasmussen's encephalitis. Science 1994; 265:648-651 25. Hart YM, Andermann F, Fish DR, et al: Chronic encephalitis
and epilepsy in adults and adolescents: A variant of Rasmussen's syndrome? Neurology 1997;48:418-424 26. Wiiliamson PD, FrenchJA, Thadani VM, et al: Characteristics of medial temporal lobe epilepsy:II. Interictal and ictal scalp electroencephalography, neuropsychological testing, neuroimaging, surgical results, and pathology. Ann Neurol 1993;34:781-787 27. Rush E, Morrell MJ: Cortical dysplasia with mesiotemporal sclerosis: Evidence of kindling in humans. Epilepsia 1993;34:15A (suppl 6) (abstr) 28. Raymond AA, Fish DR, Stevens JM, et al: Association of hippocampal sclerosis with cortical dysgenesis in patients with epilepsy. Neurology 1994;44:1841-1845
Index Words: Cortical dysplasia, epilepsy, functional hemispherectomy, hemispherectomy, Rasmussen's encephalitis, Sturge-Weber syndrome "EMISPHERECTOMY procedures have been per,formed for a variety of conditions since the earliest part of the 20th centmy.I,2 More recently, hemispherectomy procedures have been used predominantly for patients with chronic epilepsy due to conditions involving all or most of one hemisphere in which a smaller
H
FromtheDepartmentsof'AnatomicPathology,Neurosurge~,andEpilepsyand SleepDisorders,ClevelandClinicFoundation, Cleveland,OH. Presentedat the UnitedStates and CanadianAcademyofPathologymeeting, San Frandsco, CA, March 1998. Addressreprintrequeststo RichardA. Prayson,MD, DepartmentofAnatomic Pathology(L25),ClevelandClinicFoundation,9500EuelidAve, Cleveland,OH 44195. Copy@t © 1999by ~d~B.SaundersCompany 1092-9134/99/0304-0001510.00/0
resection will not remove all the epileptogenic tissue. The procedure has proven effective in reducing seizure frequency. Unfortunately, a significant subset of patients who had complete anatomic hemispherectomy eventually developed complications related to superficial cerebral hemosiderosis) ,4 To minimize complications associated with superficial cerebral hemosiderosis syndrome, the hemispherectomy procedure was modified to involve a subtotal resection of the hemisphere associated with a disconnection of the remaining areas from the rest of the brain (so-called "functional hemispherectomy"). Relatively few series have specifically focused on the histopathologic findings in functional hemispherectomy specimens. In most previous reports, functional hemi-
Annals of Diagnostic Pathology, Vol 3, No 4 (August), 1999: pp 205-212
205
206
Prayson et al
spherectomy-associated resections are grouped together with other hemispherectomy-associated resections and extratemporal corticectomies. This report reviews one institution's experience with 37 consecutive functional hemispherectomies in a predominantly pediatric-aged population. Details of the histopathologic findings in these cases are reviewed. Materials and Methods
A search of the neurosurgicai case flies from 1990 through April 1998 was performed to identify patients who underwent functional hemispherectomy. The functional hemispherectomy procedure was performed as described by Rasmussen and Viilemure. 5-7 The surgical pathology files were then searched for cases in which there was tissue excised at the time of surgery. Thirty-seven patients who underwent functional hemispherectomy during this period had excised tissue examined histologically; these patients comprised this study group. All available histologic material was examined in each case. The number of histologic slides reviewed ranged from 1 to 48 (mean, 9.8 slides). Half or more of the tissue submitted for pathologic evaluation was examined microscopically in 35 of 37 cases; in the remaining two cases, less than half of the tissue submitted was examined histologically. Histologic sections were routinely generated from formalin-fixed, paraffinembedded tissues sectioned 4 to 5 ~rn in thickness and stained with hematoxylin-eosin. The histologic features observed in each case were documented. Cortical dyspiasia was defined as a maldevelopmental abnormality of the cortex relative to the neuronal component.8 Histologic classification of the cortical dysplasia patterns observed were partially based on those previously outlined by Mischel et al in 1995.9 Heterotopic white matter neurons situated in deep white matter were noted when identified. Cases in which glial-neuronal tissue was situated in the pia arachnoid region were also noted when identified. Other lesions that were identified and noted included the presence of remote ischemic damage or infarct. Rasmussen's encephalitis was identified according to the previous histologic descriptions of this lesion. 1°,11Evidence of Sturge-Weber disease, as previously described, was also documented when noted.12,1s Hippocampal sclerosis, when present, was identified and characterized by segmental loss of neurons and gliosis in regions of the Ammon's horn, particularly involving CA1/ prosubiculum, dentate gyrus, and CA4 regions. 14,15 Other findings that were documented when noted included the presence of acute and/or chronic leptomeningeal inflammation, meningeal fibrosis, and evidence of recent infarct and/or contusion. Results
Clinical Information The study included 20 males and 17 females. The patients ranged in age from 3 months to 37 years (mean,
9.6 years) at the time of surgery. Five patients were over the age of 21 years at the time of initial surgery. One patient was known to have epidermal nevus syndrome.
Histopathologic Findings The histopathologic findings in the 37 patients are presented in Table 1, Cortical dysplasia was observed in 14 resections. The single most common pattern of dysplasia observed was that of architectural disorganization consisting of either abnormalities of cortical lamination or a haphazard arrangement of neuronal cells within the cortex (Fig 1). This finding was seen in 13 of 14 cases. In 11 cases, increased numbers of molecular layer neurons were identified. Cytomegalic neuronal cells were focally noted within the cortex in 11 cases (Fig 2). Evidence of gyral fusion was seen in four cases. Discrete areas of gray matter situated within deep white matter was also noted in four resections. Pachygyric cortex was identified in three resections. Balloon cells characterized by abundant eosinophilic cytoplasm and round to slightly oval nuclei with prominent nucleoli were identified focally in three cases (Fig 3). Occasional balloon cells were multinucleated. Discrete areas of heterotopic white matter were observed in the cortex in two cases. Discrete clustering of atypical neuronal and glial cells was observed situated in the cortex in one case. Foci of glial-neuronal tissue situated within the pia arachnoid region were seen in four cases (Fig 4). Heterotopic neurons situated in deep white matter were identified in 17 cases. In all but one patient, more than one of the above-mentioned patterns of dysplasia were noted to coexist. Histologic evidence of remote ischemic injury or infarct was observed in six cases involving the middle cerebral artery distribution. These areas were characterized by focal cystic degeneration of the parenchyma Table 1, Summary of Histopathologic Findings in 37 Functional Hemispherectomies N (%) Cortical dysplasia Infarct/remote ischemia Sturge-Weber disease Rasmussen's encephalitis Hippocampal neuronal Ioss/gliosis* Hippocampal sclerosis* No significant pathologic findings1Chronic encephalitis, not otherwise specified
14 (37.8) 6 (16.2) 6 (16.2) 3 (8.1) 3 (8.1) 2 (5.4) 2 (5,4) 1 (2.7)
*Most of these patients had coexistent cortical dysplasia or radiographic evidence of remote infarct. tBoth of these patients had radiographic evidence of remote infarct, which was the presumed cause of seizures.
Histopathology of Functional Hemispherectomy
Figure 1. Loss of the normal laminar architecture of the cortex with gliosis involving cortical layers 1 to 3 and prominent subpial gliosis. (Hematoxylin-eosin stain; medium power.)
Figure 2. Cytomegalic cortical neurons in a cortical dysplasia involving layer 5. (Hematoxylin-eosin stain; high power.)
207
associated with pericystic gliosis (Fig 5). In four of these cases, adjacent cortical dysplasia-like changes were also observed; these were presumed in some cases to be secondary to the ischemic/infarctive event. In three of these four cases, focal architectural disruption with loss of the normal lamination pattern and a haphazard arrangement of neuronal cells was observed. In two cases, a slight increase in the number of molecular layer neurons was noted. In at least one of these cases, the infarct was known to be associated with previous surgery that involved excision of an arteriovenous malformation. In six resections, changes consistent with those described in Sturge-Weber disease were noted. These changes involved a proliferation of predominantly venous-type vessels in the leptomeninges (Fig 6). Increased numbers of vessels were also noted within the cortex, associated with extensive perivascular and parenchymal mineralization of both cortex and less notably of underlying white matter. Increased gliosis was also noted within the cortex. In three resections, clinicopathologic features consistent with Rasmussen's encephalitis were identified. The histopathologic findings in these cases were somewhat similar and involved the presence ofperivascular chronic inflammation in both the leptomeninges and focally surrounding parenchymal vessels associated with a diffuse microglial cell proliferation, focal formation of more discrete microglial nodules, and a mild nonspecific reactive astrocytosis (Fig 7). Specific viral inclusions were not identified in any of these cases. One additional resection demonstrated similar histopathologic findings but lacked the typical clinical and radiographic features
208
Prayson et al
Figure 3. Scattered balloon cells with abundant eosinophilic cytoplasm in a dysplasia. (Hematoxylin-eosin stain; high power.)
of Rasmussen's encephalitis. Viral inclusions were also not identified in this case by light microscopy. Histopathologic findings ofhippocampal sclerosis were definitely identified in two cases. In one of these cases, a concomitant increased number of molecular layer neurons (cortical dysplasia) was observed. In the other case, a coexistent remote infarct was identified. In three additional hippocampal resections, focal areas of neuronal loss and gliosis were identified. Due to the fragmentation and orientation of these particular specimens, a definite diagnosis of hippocampal sclerosis was precluded, although the possibility that these lesions represented that process could not be excluded. In one of these three cases, coexistent cortical dysplasia character-
Figure 4. Glial-neuronal tissue located in the pia-arachnoid region. (Hematoxylin-eosin stain; medium power.)
ized by gray matter heterotopia and a focally disrupted cortical architecture pattern were identified. The other two patients had evidence of remote infarct on imaging studies. No significant histopathologic findings were identified in two limited resection specimens. Both of these patients had evidence of remote left middle cerebral artery infarcts radiographically. Nonspecific subpial gliosis was observed in 31 of 37 resections. Focal areas of perivascular atrophy, most notably invoMng white matter vessels, was seen in 13 cases. Excluding the four cases of encephalitis, focal areas of chronic leptomeningeal inflammation were seen in eight cases and meningeal fibrosis in 11. Most of these patients had a prior history
Histopathology of Functional Hemispherectomy
209
Figure 5. An area of cystic degeneration associated with a remote infarct. (Hematoxylin-eosin stain; medium power.) of recent invasive seizure monitoring. Focal leptomeningal acute inflammation was observed in one resection. Discussion
Figure 6. Leptomeningeal vascular proliferation in SturgeWeber disease. (Hematoxylin-eosin stain; low power.)
The single most common histopathologic finding in this series is that of cortical dysplasia, which was observed in slightly more than one third of the resections. This was not surprising given the nature of the lesion, which can at times be quite extensive to the extreme of hemimegalencephaly. Although comparable series specifically documenting the histopathologic features in functional hemispherectomy cases are uncommon, a fairly comparable incidence of dysplasia was observed in a series of children who underwent hemispherectomy or multilobar cortical resection as treatment for chronic epilepsy at UCLA Medical Center. 16 Cortical dysplasia also was one of the most common etiologies among children and adolescents who had epilepsy surgery at the Cleveland Clinic Foundation.': Farrell et a116 reported evidence of malformations or hamartomatous lesions in 15 of 38 resections (39.5%). In a series of 58 hemispherectomy cases] a 26 of which were functional hemispherectomies, cortical dysplasia, likewise, was the single most common pathologic finding in 33 of 58 patients (57%). The most common patterns of dysplasia observed in this series included architectural laminar or neuronal orientation abnormalities, increased numbers of molecular neurons, and evidence of neuronal cytomegaly. In Mischel et al's 9 evaluation of 77 cortical dysplasia cases derived from a variety of surgeries, the single most common pattern of dysplasia was cortical laminar disorganization, which was observed in
210
Prayson et al
Figure 7. Rasmussen's encephalitis marked by perivascular chronic inflammation and reactive astrocytosis. (Hematoxylin-eosin stain; medium power.)
all cases. According to their proposed grading system, Mischel et aP suggested that neuronal cytomegaly most probably represented early developmental disruptions (occurring during the first trimester of development), whereas cortical laminar disorganization and increased numbers of molecular layer neurons can arise at any time during development, but the most severe forms probably arise earlier during development. It therefore appears that the more severe and extensive cases of cortical dysplasia, which may require a functional hemispherectomy procedure, are more likely to arise from migration defects that occur earlier during development. It appears that the extent of dysplasia and its specific location may be at least, if not more, significant than the histologic type observed with regard to seizure severity. In the vast majority of cases, multiple patterns of coexistent cortical dysplasia were encountered in the same resection. The etiologies of cortical dysplasia still remain uncertain and they likely arise from a variety of conditions that disrupt the process of neuronal migration and differentiation,a,~9 Defects involving any one of a number of factors that modulate neurogenesis, including involvement of the radial glia, programmed cell death, nerve growth factors, mechanisms controlling cell proliferation, neuronal developmental genes, hormonal factors, and cell adhesion molecules, may potentially give rise to abnormalities that affect neuronal migration and result in cortical dysplasia. 19 Superimposed on these factors are environmental conditions, such as ischemic or infectious states, that may potentially alter neurogenesis. Of particular interest in this regard are the cases in which there was clear histologic evidence of a remote,
likely perinatal, ischemic event associated with adjacent architectural disorganization. Whether one chooses to use the term cortical dysplasia for such abnormalities may be a matter of semantics. Cortical dysplasia is generally thought of as a maldevelopmental-associated lesion. It is quite likely that at least a subset of lesions we currently recognize as dysplasia are the result of ischemic events that occurred in utero during the development of the brain but were not severe enough to cause a cavitary lesion. The presence of pathologic patterns virtually identical to those currently recognized as cortical dysplasia adjacent to areas of infarct raise interesting questions regarding the nature of dysplasia and its precise definition. Patients with Sturge-Weber disease likewise have been previously reported in most hemispherectomy series. It was a relatively common etiology in the Cleveland Clinic pediatric epilepsy series. 17 Farrell et a111 noted that three patients had pathologic findings associated with Sturge-Weber disease, u Two of 58 hemispherectomy specimens reported by Peacock et aP a showed changes consistent with Sturge-Weber syndrome. Because of the generally more extensive distribution of pathology in Sturge-Weber syndrome, many of these patients tend to benefit from a hemispherectomy procedure and have generally excellent seizure control following surge~. 2° Another commonly encountered pathologic finding in hemispherectomy series is that of Rasmussen's-type encephalitis. The extensive involvement of the process (usually confined to one hemisphere) and the progressive nature of the disorder both lend themselves to this surgical approach. Numerous studies have examined a
Histopathology of Functional Hemispherectomy
variety of viral agents in an attempt to pinpoint a culprit in terms of the etiology of Rasmussen's encephalitis. Although occasional reports have demonstrated evidence of cytomegalovirus, Epstein-Barr virus, or herpes simplex virus in patients with Rasmussen's encephalitis, this has been a variable finding, n,~1-23 A fairly recent study has found evidence ofautoantibodies to glutamate receptor (GluR3) in these patients, suggesting that an autoimmune process may be in part responsible for the pathologic finding. 24 Certainly, the possibility that a viral infection may trigger this autoimmune mechanism cannot be excluded. Alternately, the somewhat nonspedfic histologic findings in patients with Rasmussen's encephalitis may be due to a variety of etiologies that happen to share a similar pathology. A recent report of a group of patients who developed seizures during adolescence or adulthood with identical histopathologic findings to those found in patients with Rasmussen's encephalitis demonstrated more variable clinical characteristics. 25 The precise relationship of these types of patients with the classic Rasmussen's syndrome is not well understood and indicates we still have much to learn. Hippocampal sclerosis is well known as the most common pathologic finding found in the majority of patients with temporal lobe epilepsy. 14,26 Generally, however, one would not expect the patient to require a more extensive functional hemispherectomy procedure for hippocampal sclerosis. Of particular interest is the fact that in both patients with hippocampal sclerosis, a significant coexistent lesion was present. In one patient, coexistent cortical dysplasia was identified and in the other patient, evidence of remote infarct was noted on imaging studies. The exact relationship between dysplasia and hippocampal sclerosis is unknown. 15,27,28 Whether these lesions merely coexist or represent two different phenotypic manifestations of a common event that occurred in utero or whether one lesion may beget the other is not certain. In two cases, no signifcant pathologic findings were noted by light microscopy; however, both patients had radiographic evidence of infarct. The etiology of seizures in these cases is presumably related to these lesions, which were not excised at the time of surgery.
Acknowledgment The authors thank Denise Egleton for her help in the preparation of this manuscript.
211
References 1. Tinuper P, Andermann F, Villemure J-G, et ah Functional hemispherectomy for treatment of epilepsy associated with hemiplegia: Rationale, indications, results, and comparison with callosotomy. Ann Neurol I988;24:27-34 2. Dandy W: Removal of right cerebral hemisphere for certain tmnors with hemiplegia: Preliminary report.JAMA 1928;90:823-825 3. Oppenheimer DR, Griffith HB: Persistent intracranial bleeding as a complication of hemispherectomy.J Neurol Neurosurg Psychiatry 1966;29:229-240 4. Rasmussen T: Postoperative superficial hemosiderosis of the brain, its diagnosis, treatment and prevention. Trans Am Neurol Assoc 1973;98:133-137 5. Rasmussen T: Hemispherectomy for seizures revisited.J Can Sci Neurol 1983;10:71-78 6. Rasmussen T, Villemure JG: Cerebral hemispherectomy for seizures and hemiplegia. Cleve ClinJ Med 1989;56:562-568 7. Villemure JG: Hemispherectomy: Techniques and complications, in Wyllie E (ed). The Treatment of Epilepsy-: Principles and Practice. Philadelphia, PA, Lea & Febiger, 1993, pp 1116-1119 8. Prayson RA, Estes ML: Cortical dysplasia: A histopathologic study of 52 cases of partial lobectomy in patients with epiiepsy. Hum Pathol 1995;26:493-500 9. Mischel PS, Nguyen LP, Vinters HV: Cerebral cortical dysplasia associated with pediatric epilepsy. Review of neuropathologic features and proposal for a grading system. J Neuropathol Exp Neurol 1995;54:137-153 10. Rasmussen T, OlszewskiJ, Lloyd-Smith D: Focal seizures due to chronic localized encephalitis. Neurology 1958;8:435-445 11. Farrell MA, Droogan O, Secor DL, et ah Chronic encephalitis associated with epilepsy: Immunohistochemical and ultrastructural studies. Acta Neuropathol 1995;89:313-321 12. DiTrapani G, DiRocco C~ Abbamondi AL, et ah Light microscopic and uhrastructural studies of Sturge-Weber disease. Childs Brain 1982;9:23-36 13. Prayson RA, GrewaI 1D, MdvlahonJT, et ai: Leukocyte adhesion molecules and x-ray dispersive spectroscopy in Sturge-Weber disease.J Pediatr Neurol 1996;15:332-336 14. Babb T, Brown W: Pathological findings in epilepsy, in EngelJJ (ed). Surgical Treatment of the Epilepsies. New York, NY, Raven, 1987, pp 511-540 15. Prayson RA, ReithJD, Najm IM: Mesial temporal sclerosis: A clinicopathologic study of 27 patients, including 5 with coexistent cortical dysplasia. Arch Pathol Lab Med 1996;120:532-536 I6. Farrell MA, DeRosa MJ, Curran JG, et ah Neuropathologic findings in cortical resections (including hemispherectomies) performed for the treatment of intractable childhood epilepsy. Acta Neuropathol 1992;83:246-259 17. Wyllie E, Comair YG, Kotagal P, et al: Seizure outcome after epilepsy surgery in children and adolescents. Ann Neurol 1998;44:740748 18. Peacock WJ, Wehby-Grant MC, Shields WD, et ah Hemispherectomy for intractable seizures in children: A report of 58 cases. Childs Nerv Syst 1996;12:376-384 19. Rorke LB: The roIe of disordered genetic control of neurogenesis in the pathogenesis of migration disorders. J Nenropathol Exp Neurol 1994;33:105-i 17 20. Hoffman HJ, Hendrick EB, Dennis M, et al: Hemispherectomy for Sturge-Weber syndrome. Childs Brain 1979;5:233-248
212
Prayson et al
21. Jay V, Becker LE, Otsubo H, et ah Chronic encephalitis and epilepsy (Rasmussen's encephalitis): Detection of cytomegalovirus and herpes simplex virus I by the polymerase chain reaction and in situ hybridization. Neurology 1995;45:108-I 17 22, Power C, Poland SD, Blume WT, et ah Cytomegalovirus and Rasmussen's encephalitis. Lancet 1990;336:1282-1284 23. Atkins MR, Terrell W, Hulette CM: Rasmussen's syndrome: A study of pot ential viral etiology. Clin Neuropathol 1995;14:7-12 24. Rogers SW, Andrews PI, Gahring LG, et al: Autoantibodies to glutamate receptor GluR3 in Rasmussen's encephalitis. Science 1994; 265:648-651 25. Hart YM, Andermann F, Fish DR, et al: Chronic encephalitis
and epilepsy in adults and adolescents: A variant of Rasmussen's syndrome? Neurology 1997;48:418-424 26. Wiiliamson PD, FrenchJA, Thadani VM, et al: Characteristics of medial temporal lobe epilepsy:II. Interictal and ictal scalp electroencephalography, neuropsychological testing, neuroimaging, surgical results, and pathology. Ann Neurol 1993;34:781-787 27. Rush E, Morrell MJ: Cortical dysplasia with mesiotemporal sclerosis: Evidence of kindling in humans. Epilepsia 1993;34:15A (suppl 6) (abstr) 28. Raymond AA, Fish DR, Stevens JM, et al: Association of hippocampal sclerosis with cortical dysgenesis in patients with epilepsy. Neurology 1994;44:1841-1845