Address correspondence and reprint requests to Dr. F. Andermann at Departments of Neurology and Pediatrics, McGill University, Montreal Neurological Hospital & Institute, 3801 University St., Montreal, Quebec, Canada, H3A 2B4. E-mail: firstname.lastname@example.org
Summary: Purpose: We evaluated the results of cortical resection of epileptogenic tissue for treatment of intractable porencephaly-related epilepsy.
Methods: We examined clinical features, electrophysiological data, surgical findings, and seizure outcomes after cortical resection in eight patients with intractable epilepsy related to porencephalic cysts.
Results: All eight patients had hemiparesis. Five retained motor function in the hemiparetic extremities; six retained visual fields. All had partial seizures, six with secondary generalization. Seven patients had simple and three had complex partial seizures (CPSs); two also had drop attacks. Four patients had multiple seizure types. Long-term scalp video-EEG (LVEEG) localized interictal epileptic abnormalities that anatomically corresponded to the cyst location in three patients. LVEEG recorded ictal-onset zones in five; these anatomically corresponded to the cyst location in three of the five. EEG recorded generalized seizures in two patients, hemispheric in one, and multifocal in two. Intraoperative electrocorticography (ECoG) revealed interictal epileptic areas extending beyond the margins of the cyst in seven patients. We resected ECoG-localized interictal epileptic areas completely in five patients and partially in two. Cortical resection was based on seizure semiology and LVEEG in one patient whose ECoG showed no epileptiform discharges. After a minimum follow-up of 1 year, six patients had excellent seizure outcome (Engel class I), and two had a >90% seizure reduction (Engel class III) without complications.
Conclusions: Cortical resection guided by ECoG allows preservation of motor function and visual field and provides an effective surgical procedure for treatment of intractable epilepsy secondary to porencephaly.
Porencephalic cysts are cavities in the brain resulting from prenatal or perinatal vascular occlusion. Commonly congenital hemiparesis, intellectual impairment, and medically intractable epilepsy are present (1–6). Physicians often discourage surgical intervention because accurate localization of seizure foci in patients with intractable porencephaly-related epilepsy is difficult (7). When surgery is recommended, neurologic status determines the choice of the surgical procedure. Hemispherectomy is the classic treatment option for patients with porencephaly-related epilepsy and maximal or near-maximal congenital hemiparesis and hemianopsia (8–14), whereas partial hemispherectomy or cortical resection with excision of multiple lobes is used in patients who have relatively mild hemiparesis with useful finger movements or preserved visual fields. Lesionectomies, focal epileptogenic corticectomies, or lobectomies have occasionally been performed in patients with focal extratemporal or temporal lobe epilepsy and porencephalic cysts (15–21).
Recently some patients have benefited from temporal lobectomy for coexisting hippocampal sclerosis with extratemporal porencephaly when clinical and EEG findings are concordant with temporal lobe onset of seizures (22,23). However, the majority of patients have still undergone functional hemispherectomy (22).
No systematic studies have focused on focal cortical resection as a surgical option, as an alternative to hemispherectomy, in patients with porencephaly. In this study, we evaluate the benefit of cortical resection for seizure control in eight patients. We examine the clinical histories, electrophysiologic data, epileptogenic area localization corroborated by intraoperative electrocorticography (ECoG), surgical findings, and seizure outcomes.
PATIENTS AND METHODS
We studied eight patients who had undergone cortical resections for intractable epilepsy secondary to a porencephalic cyst between 1986 and 2002 at The Montreal Neurological Institute and Hospital. Six were male subjects. We included only patients whose conditions met the clinical diagnosis of porencephalic cyst: a circumscribed brain defect in any cerebral artery territory, revealed on computerized tomographs (CTs) or magnetic resonance images (MRIs) (3–5). Patients may have had a previous history of prenatal or perinatal ischemic brain insults but no other etiologic factors. We excluded patients with cystic lesions that occurred later in life (e.g., after trauma, tumor resection, circulatory disturbances, or infection) and cystic developmental malformations (e.g., schizencephaly or arachnoid cyst) that were visible on CT (24) or MRI (3,25,26). We classified seizures according to the International League Against Epilepsy guidelines (27,28).
We reviewed the medical records of all patients with respect to their past histories; neurologic findings, especially motor disabilities; seizure descriptions and frequencies; neuropsychological assessments; neuroimaging studies; scalp and intracranial EEG findings; and surgical findings. Motor disabilities were evaluated with specific attention to functional aspects. When patients had no voluntary finger movement or foot tapping, we considered them to have maximal hemiparesis. Usually in such patients, their hands barely moved or were spastic, their fists clenched, elbows were fixed, and they limped. When patients had voluntary finger movements, we considered them to have mild hemiparesis. All patients were referred to an ophthalmologist for visual field examination by Goldman perimetry.
All patients had long-term scalp video-EEG (LVEEG) monitoring with electrodes placed according to the International 10-20 system. Sphenoidal electrodes were used when we suspected seizure onset from the mesial temporal region. Preexcision ECoGs with 16 carbon-ball electrodes were recorded to cover the all the borders of the cysts, under general or local anesthesia, supplemented by neuroleptanalgesia with droperidol and fentanyl. One patient underwent intracranial EEG monitoring with depth and epidural electrodes before cortical resection. Only patient 2 was studied with positron emission tomography (PET) for motor-function mapping before the second operation.
We based the area of cortical resection on intraoperative ECoG, LVEEG data, seizure semiology, and anatomic location of the cystic lesion.
We evaluated seizure outcome by using Engel's classification (29) and postoperative complications after a minimum 1-year follow-up.
Table 1 summarizes the clinical profiles of the patients. Five had perinatal ischemic insults, one was premature, and the remaining two had normal deliveries. One had status epilepticus in childhood. One patient (patient 7) had prior epilepsy surgery consisting of partial removal of gliotic cortical tissue surrounding the cyst.
Table 1. Clinical profiles and seizure semiology
Age at surgery/sex
Age at seizure onset
y, years; M, male; F, female; m, months; L, left; R, right; CPS, complex partial seizure; SPS, simple partial seizure; SPSG, SPS with secondary generalization.
SPSG (L sensorimotor)
2-3 up to 30/day
SPSG (R focal motor)
CPS with R tonic arm extension, CPS with oroalimentary automatism
SPS (L sensorimotor), CPS with R leg automatism, Drop attacks
SPSG (L sensorimotor), Drop attacks
SPSG (R focal motor), CPS with epigastric aura
SPSG (R sensorimotor)
R homonymous hemianopsia
SPSG (L focal motor)
Age at seizure onset ranged from birth to 9 years (mean, 3.8 years), whereas age at surgery ranged from 10 to 25 years (mean, 16.5 years). Seizure duration was from 6 to 25 years (mean, 12.9 years).
All patients had partial seizures confirmed by LVEEG monitoring. Seven had sensorimotor or focal motor simple partial seizures (SPSs) restricted to the hemiparetic side, and six of these also had secondarily generalized seizures. Three patients had CPSs associated with epigastric aura, tonic arm elevation, and oroalimentary or ambulatory automatisms. Drop attacks occurred in two patients. Four had multiple seizure types: one with SPSs, CPSs, and drop attacks; one with SPSs, CPSs, but no drop attacks; one with CPSs with right tonic arm extension and CPSs with oroalimentary automatism; and one with SPSs and drop attacks. Seven patients had one or more seizures per day before surgery.
Neurologic examination revealed maximal hemiparesis without useful finger movements in three patients and mild hemiparesis with some voluntary finger movements in five.
Two patients had visual loss: patient 7 had homonymous hemianopsia, and patient 8 had concentric field contraction associated with congenital amblyopia.
Neuropsychological examinations revealed mental retardation in one patient, low-average intelligence in six, and average intelligence in one.
Magnetic resonance imaging/computed tomography
Six patients underwent MRI examinations; and two had CT scans (Table 2). Porencephalic cysts were present in the frontocentroparietotemporal region in three patients, in the frontal lobe in three, in the centroparietal region in one, and in the parietooccipital region in one. Cysts in six patients (patients 1, 2, 3, 4, 6, and 7) were in the territory of the middle cerebral artery (Fig. 1). In two patients (5 and 8), the cysts were in the anterior cerebral territory. Three patients (2, 6, and 7) had significant atrophy of the affected hemisphere.
Table 2. Locations of porencephalic cysts on MRI/CT, LVEEG results, intraoperative ECoG findings, and areas of cortical resection
Location of porencephalic cyst on MRI/CT
Epileptic area on LVEEG (Interictal area/ictal onset area)
Generalized discharges with occasional predominant L mid F-T/Bilat ant head regions
L sup F
L sup F, SMA
L C-P/Bilat C
L C sup P
L P, lower 3rd precentral, Cyst removal
R ant-mes F
R Hemispheric discharges with predominant mid T-C-P
R F not including precentral, Cyst removal
L mid-lat F
L F, ant T
No epileptiform discharge
L ant F sparing the operculum, ant T with amygdalectomy
L mid-in f C-P
L mid-inf C, P
R parasagital F
Generalized discharges with occasional predominant R F, mid-post T/R F
R F, Cyst removal
LVEEG and intracranial EEG
Results of LVEEG are described in Table 2. In three patients 1, 4, and 7), LVEEG localized interictal epileptiform discharges that corresponded anatomically to the location of cystic lesions on MRI/CT. Two patients (2 and 6) had independent epileptiform discharges from multiple regions in the affected hemisphere. Generalized epileptiform discharges occurred in two patients (patients 3 and 8), and diffuse hemispheric discharges in one (patient 5).
LVEEG localized the ictal-onset zone in five patients, and this corresponded anatomically to the cystic lesions in three (patients 1, 7, and 8). Seizure onset was bilateral central in patient 4, although interictal epileptiform discharges localized to the left central region corresponding to the cyst. In patient 3, the ictal-onset area involved bilateral anterior head regions, and because she had two types of CPSs involving tonic arm elevation and oroalimentary automatism, she also underwent preoperative intracranial EEG monitoring with stereotactically implanted epidural and depth electrodes (30). The stereo-EEG revealed that most of the seizures originated from the left supplementary motor area (SMA), whereas a few attacks came from the precentral gyrus adjacent to the cyst.
In five patients (patients 1, 2, 4, 6, and 7), craniotomies were large, covering the areas of cystic lesions and the regions of localized or multifocal interictal and/or ictal discharges. In the three patients (3, 5, and 8) who had generalized or hemispheric interictal epileptiform discharges, craniotomies were based on predominant interictal and ictal-onset areas, and the results of stereo-EEG monitoring.
When cortical surfaces were exposed during surgery, thin transparent or translucent membranes covered the cysts of seven patients. Gliotic or atrophic cortical tissue surrounded the cysts of five patients: abnormal cortical tissue overlaid the thin membranes of the cysts in two patients (2 and 5), represented the posterior margin of the cyst in patient 8, and the superior margin of the cysts of three patients (1, 4, and 5). In patient 6, the cyst was entirely covered by cerebral cortex.
All patients underwent intraoperative ECoG before cortical resection (Table 2). This localized the interictal epileptic area in seven of the eight patients, whereas one (patient 6) had no epileptiform discharges on ECoG.
In the seven patients with ECoG localizations, epileptic areas extended contiguously beyond the cyst margins. In the five patients (1, 2, 4, 5, and 8) who had abnormal cortical tissue over or in the margin of the cysts, interictal areas on ECoG discharges also extended beyond the cyst margins. Patient 7, who at age 14 years had undergone prior surgery for partial removal of gliotic tissue surrounding the cyst, had an ECoG that showed a contiguous extensive interictal epileptic area anterior to the cyst. The interictal ECoG area of patient 3 was superior to the cyst, concordant with the preoperative stereo-EEG monitoring
Table 2 summarizes the areas of cortical resection. We completely resected the areas that showed interictal epileptic discharges on ECoG in five patients (1, 2, 4, 7, and 8), and visibly abnormal cortical tissues in five patients (1, 2, 4, 5, and 8). We performed subpial cortical resection by using cavitron ultrasonic aspiration (CUSA) of the gray matter and attempted to spare the white matter. We removed the cysts in three patients (4, 5, and 8) because the cortical excision areas included most of the cyst wall (Fig. 2).
Patient 2, who had had multiple areas of interictal epileptiform discharge on LVEEG, had two operations. In the first, we resected the left frontal cortex anterior to the cyst and the gliotic tissue over the cyst, and performed multiple subpial transection of the central area because of extensive left frontocentral epileptic discharge on ECoG. Although the frequency of seizures decreased remarkably for 5 years, seizures eventually increased. He then underwent a second operation in which we resected the central area after bilateral representation of hand and leg functions was confirmed by PET.
When ECoG showed interictal epileptic discharges or stereo-EEG monitoring indicated an ictal-onset zone in eloquent cortex, we left functioning cortex intact (patients 3 and 5). In patient 3, stereo-EEG recorded most of the seizures from the left SMA and only a few attacks from the precentral gyrus. She had a cortical resection of the superior frontal gyrus including the SMA, sparing the precentral gyrus because of residual voluntary finger movement. In patient 5, who had a right mesial frontal porencephalic cyst, the intraoperative ECoG showed epileptic discharges in an extensive frontal area involving the operculum and precentral gyrus. Because of only mild left hemiparesis, we performed a right frontal lobectomy and spared the precentral gyrus.
In patient 6, whose intraoperative ECoG showed no epileptiform discharges, we performed an anterior temporal resection with amygdalectomy and frontal corticectomy, based on the patient's seizure type (CPSs with epigastric aura) and left frontal and temporal interictal discharges on LVEEG.
Specimens from six patients were available for histologic examination. The diagnosis in all was porencephalic cyst (3,4,31,32). The cortical tissues in four patients were slightly or moderately disorganized with focal neuronal loss and gliosis and showed polymicrogyria in two.
Seizure outcome and postoperative complications
Follow-up periods ranged from 1 to 18 years (mean, 8.3 years; Table 3). All patients benefited from surgery. Six (1, 2, 3, 4, 6, and 8) had excellent seizure outcome (Engel class I). Four of the six had complete cessation of attacks. Patient 1, who had been seizure free for 7 years after surgery, experienced very rare generalized tonic–clonic seizures that differed from the preoperative attacks. Patient 8 had focal motor seizures once a year without the secondary generalization she had had before. The focal motor seizures were not disabling for the patient. Only two (5 and 7) of eight patients had suboptimal result (Engel class III); however, even they had a >90% seizure reduction, which improved their quality of life.
Table 3. Follow-up after cortical resection
GTC = generalized tonic clonic seizures; SPS = simple partial seizure; SPSG = SPS with secondary generalization; yrs = years.
Seizure free, very rare GTC for the last 2 yrs
Transient deterioration in hemiparesis
1/month, drop attacks
1/month, SPSG (focal motor)
1/yr, SPS (focal motor)
Patient 2, who underwent resection of the central area at second surgery, showed transient deterioration of motor function in the hemiparetic limbs but improved within a week. Postoperative hematoma occurred in one patient, and one experienced aseptic meningitis; neither complication necessitated additional treatment.
Cortical resection for porencephaly-related partial epilepsy
In this group of patients, intraoperative ECoG localized interictal epileptic discharges in the pathological cortex that extended contiguously beyond the margin of the cyst. Cortical resections guided by ECoG achieved satisfactory seizure outcomes.
Cortical resections using intraoperative ECoG have been mentioned in some studies of extratemporal lobe epilepsy in patients with porencephaly (18–21). These reports suggested that cortical resection was efficacious for seizure control.
Porencephalic cyst formation is related to pre- and perinatal cerebral artery occlusion (1–6). Infarcts may cause gliosis adjacent to the cyst or polymicrogyria if these occur in early intrauterine life (3,4,31,32). The ischemic impact is not confined to the tissues surrounding the cyst but may involve wider or even more remote areas beyond the occluded arterial territory (7,33). The pathogenic mechanism of porencephaly in the pre- and perinatal period may determine the epileptogenic involvement of the cerebral cortex (7,34,35), resulting in an extensive epileptogenic zone, compared with the small, limited, and perilesional epileptogenic localization seen in acquired lesions such as brain tumor (21). The epileptogenic zone may extend from the margin of the cyst and include the visible lesions such as gliosis. We recommend not only lesionectomy but also additional cortical resection by using ECoG.
Another surgical procedure that has been used for the treatment of porencephalic cysts is “uncapping,” or removal of the outer wall of the cyst (36–39). Surgery that combines uncapping and fenestration of the cyst to the lateral ventricle (39) and resection of atrophic and gliotic cerebral cortex surrounding the subcortical cyst (40) also has been performed for heterogeneous seizure types in infants with porencephaly. The results of fenestration or uncapping are, however, not comparable to those obtained by resection of epileptogenic areas.
Some patients have benefited from temporal lobectomy for coexisting extratemporal porencephaly with hippocampal atrophy on MRI when seizure semiology and EEG findings were concordant with temporal lobe seizure onset (22,23). Among 24 patients with porencephaly, six had an epileptogenic focus in the temporal region with coexisting hippocampal atrophy (23). Five of the six patients were seizure free after temporal lobectomy (23). Among a larger group of 41 patients with intractable epilepsy and vascular congenital hemiparesis, five of 25 surgically treated patients had temporal resection with excellent outcome and only residual auras in three of the five (22). Hippocampal damage has been considered to result from kindling, secondary epileptogenesis, or a sequel of the perinatal vascular insult (7,22). Temporal lobectomy has been currently considered to be a surgical option for patients with porencephaly, although the majority have still undergone functional hemispherectomy (22). In this study, no patient had coexisting hippocampal atrophy as dual pathology, and only patient 6 had a temporal lobe cortical resection without hippocampectomy.
Preservation of useful motor functions in adults with porencephaly
Neurologic status determines the choice of surgical procedure for the treatment of adolescent and adult patients with intractable attacks secondary to porencephaly. Five of eight patients in this study had residual useful motor function in a hemiparetic extremity. Only one had homonymous hemianopsia related to the porencephalic cyst. Selective cortical resection in patients with intractable porencephaly-related epilepsy and minimal hemiparesis without hemianopia can preserve useful motor and visual function.
Patients with congenital hemiparesis have been considered good candidates for hemispherectomy (8–11). Hemispherectomy has been an appropriate surgical procedure for patients with porencephaly secondary to prenatal or perinatal middle cerebral artery occlusion (11,12). This procedure has a high yield for seizure-free outcome (22). However, the criteria for this operation have limited its use to patients with maximal or near-maximal hemiparesis and hemianopsia (8,13,14) whose motor disability involves loss of fine finger movement and foot tapping. Patients with porencephaly have a continuum of spastic weakness from mild hemiparesis to maximal hemiplegia or monoparesis (6). In adolescents and adults with mild hemiparesis and no hemianopia, hemispherectomy is not an ideal surgical option for seizure control. In children, the porencephaly-related epilepsy eventually interferes with the function of the nonaffected hemisphere and with the possible development of compensatory functional mechanisms (40). The resection of regional epileptogenic cortex with preservation of useful function provides a valuable surgical option.
Partial seizures in porencephaly-related epilepsy
All eight patients had partial seizures. Although four had multiple seizure types and five had multiple epileptic, diffuse, or generalized scalp EEG abnormalities, the intraoperative ECoG localized the interictal epileptic area in seven of the eight.
SPSs (40,41), CPSs (40,42), infantile hemispasms (36,39,40), Lennox–Gastaut syndrome (37,39), and generalized seizures (38,39,42) have all been reported in association with porencephalic cysts. Naef (6) described the various electroclinical profiles of seizures including focal motor, generalized, adversive, and psychomotor attacks in 32 patients with porencephaly. Only half the patients had focal activity, most with multiple spike foci, corresponding to the location of the cysts, but in the others, diffuse or even contralateral epileptic discharges were seen on scalp EEGs (6).
Generalized and diffuse epileptic EEG abnormalities are common in porencephaly-related epilepsy (8). Bilaterally synchronous spike-and-wave discharges, when not accompanied by independent epileptic discharges from the “good” hemisphere, may represent secondary bilateral synchrony originating from the porencephalic hemisphere because the more damaged hemisphere cannot generate potentials of as high a voltage. Intraoperative ECoG in our study identified the regional epileptic discharges from pathological cortex extending from the margins of the cysts. The cessation of generalized seizures, including drop attacks, suggests that the epileptogenic zone must have been localized in the porencephaly-related surrounding cortex. Intraoperative ECoG, however, has some disadvantages because the recording is short, mostly interictal, under anesthesia, and does not cover the whole brain. Therefore scalp video-EEG, semiology, and other noninvasive information regarding the extent of the epileptogenic zone must be reviewed before the intraoperative ECoG.
West syndrome (WS), Lennox–Gastaut syndrome, generalized seizures, and, more rarely, partial seizures have been described in infants with porencephaly (39). Scalp EEGs showed hypsarrhythmia, generalized slow spike and waves, and hemispheric or focal epileptiform discharges. The surgical outcome of those infants was seizure freedom in 15 of 18 patients younger than 2 years and residual generalized and partial seizures in 11 of 19 patients older than 2 years (39). The older children with additional partial seizures had less-favorable seizure outcomes after surgery.
Our adolescent and adult patients had partial seizures, multiple seizure types, and a longer seizure history compared with previous reports of infants and had a good response to resection of epileptogenic areas.
Cortical resection is an effective surgical procedure for intractable epilepsy secondary to porencephaly. It can spare preoperatively preserved motor function and visual fields. The epileptic cortex extends contiguously from the margin of the cyst. Intraoperative ECoG can localize the epileptic cortex, even when multiple seizure types and diffuse scalp EEG abnormalities are present. We recommend not only lesionectomy but also additional cortical resection with ECoG in patients with porencephaly-related intractable partial epilepsy.
Acknowledgment: We thank Mrs. Carol L. Squires for editorial assistance.