Seizure outcome of intractable partial epilepsy in children

Seizure Outcome of Intractable Partial Epilepsy in Children Lan S. Chen, MD, Nina Wang, BA, and Meei-Ing Lin, RN The intractable partial epilepsy outcome information is important in determining not only when epilepsy surgery evaluation should begin but also in deciding who would benefit and what is the likelihood of any benefit from surgery. Medical records of 50 children diagnosed with nontumor-related partial seizures, confirmed by video-electroencephalography (video-EEG), had at least one seizure per week and were followed for at least 2 consecutive years after video-EEGs were reviewed. There were 30 patients who continued with antiepileptic drug treatment after video-EEG. The seizure outcome analysis revealed a significant improvement of seizure control in the first year of follow-up but no difference between the first year and the following 3 years. Only 30% had excellent longterm outcome (seizure free or less than one seizure per 6 months). The presence of focal lesions on neuroimaging was the only risk factor of poor outcome. The other 20 patients underwent epilepsy surgery after video-EEG; 60% attained excellent outcome despite the fact that 90% had focal neuroimaging abnormality. Children whose partial epilepsy remained intractable after 1 year of antiepileptic drug treatment should be evaluated for candidacy of epilepsy surgery, particularly those who have focal lesions on neuroimaging. © 2002 by Elsevier Science Inc. All rights reserved. Chen LS, Wang N, Lin M-I. Seizure outcome of intractable partial epilepsy in children. Pediatr Neurol 2002;26: 282-287.

Introduction Epilepsy surgery is a well-established treatment modality for children who have partial seizure disorder refractory to antiepileptic drug treatment. Long-term outcome information in children with intractable partial seizure disorder is important because it provides guidelines con-

From the Division of Neurology; Children’s Hospital Los Angeles; Keck School of Medicine; University of Southern California; Los Angeles, California.

282

PEDIATRIC NEUROLOGY

Vol. 26 No. 4

cerning when epilepsy surgery evaluation should begin and what would be the likely benefit. The ideal way to obtain these data is to prospectively randomize children who are candidates for epilepsy surgery to either stay on antiepileptic drug treatment or proceed with surgical treatment. However, it is ethically impossible to perform such a controlled comparison. We conducted a retrospective study focusing on children who had extreme intractable epilepsy and whose diagnosis of partial seizure disorder was confirmed by simultaneous video and electroencephalography (videoEEG) evaluation. We limited this study to a period of 5 years when epilepsy surgery was fully operational in our institute and our neurologists could easily refer patients for surgical evaluation. Methods This study was comprised of patients who underwent video-EEG monitoring from June 1990 through December 1995. The candidates’ medical records were reviewed. The inclusion criteria included the following: (1) video-EEG diagnosis of partial seizure disorder; (2) a minimum of one seizure per week (Class IV) despite at least one antiepileptic medication at the time of the video-EEG study; and (3) at least 2 consecutive years of follow-up at Childrens Hospital Los Angeles after video-EEG. The diagnosis of partial seizures was made if ictal electroencephalogram demonstrated focal paroxysmal discharges or video-recorded ictal behaviors were consistent with simple partial seizures or complex partial seizure as classified by the International League Against Epilepsy [1]. According to guidelines of the 10-20 International System, a scalp electroencephalogram was recorded using a 16-channel bipolar montage. Seizure outcome was classified by the seizure frequency as follows: Class I: seizure free or less than one seizure per 6 months; Class II: less than one seizure per month but at least one seizure per 6 months; Class III: less than one seizure per week but at least one seizure per month; and Class IV: at least one seizure per week. Neurologists at Childrens Hospital Los Angeles followed these patients once every 2-4 months. According to their medical records, all patients in this study had at least two clear documented seizure frequencies every year. For each patient, each annual seizure outcome was determined by at least two neurologic follow-ups that documented the same seizure frequency class. The

Communications should be addressed to: Dr. Chen; Division of Neurology, #82, Children’s Hospital Los Angeles; 4650 Sunset Blvd.; Los Angeles, CA 90027. Received May 22, 2001; accepted October 12, 2001.

© 2002 by Elsevier Science Inc. All rights reserved. PII S0887-8994(01)00404-0 ● 0887-8994/02/$—see front matter

seizure frequency was estimated according to the reports provided by caretakers who recorded seizures on calendars. Patients who were gradually withdrawn from antiepileptic drugs after being seizure free for more than 2 years were assumed to be seizure free if they did not return to our neurology clinic. Patients who had reached Class I seizure control were analyzed as “loss for follow-up” if they did not return to our clinic before antiepileptic drugs were gradually withdrawn. Neuropsychologic tests were performed before and after surgery on every patient who underwent epilepsy surgery to assess cognitive function. However, only a few patients who stayed on medical treatment had neuropsychologic evaluation. All patients were at least 3 years of age on this retrospective review. Therefore classification of developmental level was based on simple clinical measurements, including ambulating, basic verbal communication skills, and school attendance. Children who were neither ambulatory nor verbal were categorized as severely delayed. Children who were ambulatory but not verbal were classified as moderately delayed. Children who were both ambulatory and verbal but were not able to attend regular education were considered mildly delayed. Normal development was applied to those who were in a regular educational program. We performed one-sided paired t tests to compare the differences in seizure outcome class in either the medical or surgical group or between these two groups and the number of antiepileptic drugs taken over a 4-year period of follow-up (Figs 1A and B, and 2). Chi-square tests were used to compare the difference between the medical and surgical groups (Table 1) and to evaluate various factors related to the outcome (Tables 2 and 3). In most comparisons the resulting table of interest had cells with small (less than five) counts, which can affect the accuracy of P-value. To compute a more reliable test of significance, we simulated 10,000 random tables for each comparison table in such a way that the marginal totals were fixed. The percentage of the simulated tables with a chi-square statistic greater than the chi-square of the true table was calculated. This percentage represents an adjusted P-value, Pa⬘. The significance of the chi-square tests were defined as Pa⬘ less than 0.05.

Results General Clinical Data A total of 65 consecutive patients met our criteria. Five of these 65 patients were excluded because they also had primary generalized seizures or pseudoseizures documented on video-EEG. Thirty patients underwent epilepsy surgery within 6 months after video-EEG studies (surgical group), and 30 patients remained on medical treatment (medical group). In the surgical group, eight patients were not included in the outcome data analysis because they were diagnosed with brain tumors, and their uncontrolled seizures were not the main factor leading to the decision of surgery. Two other infants (less than 1 year of age), who were diagnosed with Sturge-Weber syndrome or hemimeganencephaly and who underwent surgery shortly after their seizures appeared resistant to antiepileptic drugs, were also excluded. Table 1 lists the age of seizure onset, the duration of seizure disorder at the time of video-EEG studies, developmental level, clinical diagnosis, interictal electroencephalogram, and neuroimaging data of the medical vs the surgical groups. The age of seizure onset was 4.1 ⫾ 0.7 years in the medical group and 3.2 ⫾ 0.6 years in the surgical group. The significant differences between the medical and surgical groups were the duration of seizures

Figure 1. Seizure outcome over 4 years of follow-up after videoelectroencephalogram (video-EEG) in the medical group (A) or after epilepsy surgery in the surgical group (B) is depicted. The abscissa is the percentage of patients and the ordinate is the year of follow-up after either video-EEG or surgery. The numbers in the parentheses are the total number of patients for the year.

and the neuroimaging findings. At the time of video-EEG studies, patients in the surgical group had a longer duration of seizure than ones in the medical group (6.0 ⫾ 1.0 years vs 3.9 ⫾ 0.8 years). All except one patient had computerized tomography and/or magnetic resonance imaging. A single focal lesion on these routine neuroimagings was present in significantly more patients in the surgical group than in the medical group, 13 (65%) patients vs six (21%) patients (P ⬍ 0.005), respectively. If F18-deoxyglucose proton emission tomography or Tc-99m single-photon

Figure 2. The number of antiepileptic drugs used at the time of video-electroencephalogram and over 4-year follow-up is displayed. The abscissa is the percentage of patients and the ordinate is the year of follow-up. The numbers in the parentheses are the total number of patients for the year.

Chen et al: Seizure Outcome in Intractable Partial Epilepsy 283

Table 1.

Clinical data of medical v. surgical group

Age of seizure onset ⬍ 2 yr of age Duration of seizure ⬎ 2 yr of age Development delay* Mild delay Moderate delay Severe delay Clinical diagnosis Encephalomalacia/cerebral infarct Infection Tuberous sclerosis Congenital malformations Trauma Unknown Abnormal interictal EEG Epileptiform abnormality Focal slowing General slowing A single focal abnormality on neuroimagings†

Medical

Surgical

Paⴕ

17/30

9/20

Ns

17/30

18/20

0.03

21/30 15 4 2

10/20 4 5 1

Ns

3 (10%)

6 (30%)

Ns

5 (17%) 3 (10%) 2 (7%) 0 (0%) 17 (57%) 22/30 11 2 9 6/29‡

3 (15%) 0 (0%) 1 (5%) 1 (5%) 9 (45%) 19/20 11 1 7 13/20

Ns Ns Ns Ns Ns Ns

0.005

* Mild delay ⫽ ambulatory and verbal; not able to attend regular education; Moderate delay ⫽ ambulatory but not verbal; Severe delay ⫽ neither ambulatory nor verbal. † MRI and/or CT. ‡ One patient did not have neuroimaging study. Abbreviations: CT ⫽ Computed tomography EEG ⫽ Electroencephalogram MRI ⫽ Magnetic resonance imaging Ns ⫽ Not significant Pa⬘ ⫽ P-value adjusted as described in text

emission computed tomography (obtained only for epilepsy surgery evaluation) were included, 90% of patients in the surgical group had a single focal abnormality on neuroimaging. There was no difference in clinical diagnosis, which was made by clinical evaluation and neuroimaging studies between these two groups (Table 1). None of the patients in this study had progressive encephalopathy from metabolic or degenerative disorders. Therefore the presence of a single focal lesion on computed tomography or magnetic resonance imaging and a longer seizure history were the most likely bias factors in selecting patients for epilepsy surgery evaluation. Seizure Outcome in the Medical Group Twenty-nine patients in the medical group had multiple seizures daily when they had video-EEG evaluation. One patient did not have daily seizures but had multiple seizures a week. There were seven patients who were lost on the fourth year of follow-up after video-EEG. The seizure outcome analysis revealed a significant improvement of seizure control in the first year after video-EEG (Fig 1A). The seizure outcome appeared to

284

PEDIATRIC NEUROLOGY

Vol. 26 No. 4

reach a plateau in the second year of follow-up. However, there was no statistical difference of the seizure outcome between the first year and the following years of follow-up. The number of patients who stayed in Class IV decreased from 30 (100%) to 18 (60%) in the first year after video-EEG, 14 (47%) in the second year, 14 (47%) in the third year, and nine (40%) in the fourth year. Parallel to the decreased number of children with Class IV seizure outcome, the number of children who achieved Class I outcome increased from zero to five (17%) in the first year of follow-up, nine (30%) in the second year, eight (27%) in the third year, and seven (30%) in the fourth year. Children who stayed in Class IV the second year after video-EEG had significantly higher incidences of focal lesions demonstrated on neuroimaging than children who were able to achieve Class I outcome (Table 2). Other factors, such as the age of seizure onset, seizure duration, developmental level, or the presence of interictal focal abnormality, did not differ between these two groups. The number of patients who were on polypharmotherapy significantly increased from 12 (40%) to 19 (63%) in the first year of follow-up (Fig 2). In subsequent years of follow-up, the number of patients with polypharmotherapy did not change. Seizure Outcome in the Surgical Group In this group, 17 (85%) patients had epileptogenic foci at extratemporal locations, and three (15%) had foci in the temporal lobe. Epileptogenic zones were defined by ictal onset of seizure discharges that were mapped extraoperatively using subdural electrodes in 19 patients. The epileptogenic focus was mapped intraoperatively by interictal recording in one patient. All 20 patients in this group had multiple daily seizures at the time of epilepsy surgery. Excellent seizure outcome (Class I) was achieved in seven (35%) children within the first year of follow-up. As demonstrated in Figure 1B, the

Table 2. group

Factors that determined seizure outcome in the medical

Age of seizure onset ⬍ 2 yr of age Duration of seizure ⬎ 2 yr of age Developmental delay Focal interictal EEG abnormality Focal neuroimaging abnormality*

Class I (n ⴝ 9)

Class IV (n ⴝ 14)

Paⴕ

5 (56%) 5 (56%) 6 (67%) 3 (33%) 1 (13%)†

10 (71%) 9 (64%) 11 (79%) 4 (29%) 8 (57%)

Ns Ns Ns Ns 0.04

* Including six cases of a single abnormality and six cases of multifocal abnormalities (tuberous sclerosis). † One patient did not have neuroimaging. Abbreviations: EEG ⫽ Electroencephalogram Ns ⫽ Not significant Pa⬘ ⫽ P-value adjusted as described in text.

Table 3.

Pathology and surgical outcome

Histopathology

n

Class I

Class II ⴙ III

Class IV

Neuronal necrosis Congenital anomally* Gliosis or normal

10 7 3

8 4 0

1 1 1

1 2 2

* Includes focal cortical dysplasia, tuberous sclerosis, and cavernous hemangioma.

number of patients who improved to Class I were 10 (50%), nine (45%), and 12 (60%), respectively, during the subsequent years of follow-up. One patient’s seizure control worsened from Class I the second year to Class IV the third year after felbamate was discontinued. Then the patient improved back to Class I the fourth year. Most of these patients were within Class II or Class III the first year. Compared with the outcome of the first year, the improvement of seizure control was significant in years 2 and 4 of follow-up. The number of patients who did not benefit from surgery was nine (45%) in the first year of follow-up, five (25%) in the second year, six (30%) in the third, and five (25%) in the fourth year. There were no differences in the age of seizure onset, seizure duration, developmental level, presence of focal epileptiform discharges on interictal electroencephalogram, or presence of focal abnormality on computed tomography or magnetic resonance imaging scan between children who reached Class I outcome and those whose seizures remained intractable in this group. If patients with Class II and Class III outcome were combined as one group, the destructive cortical lesions with extensive neuronal loss appear to have a more favorable outcome than developmental cortical lesions (tuberous sclerosis, cortical dysplasia, cavernous hemangioma), or no abnormality or mild nonspecific increase of astrocytes. However, the distribution of outcome class was not statistically different among pathology groups (Table 3). Comparison of Seizure Outcome Between Medical and Surgical Groups Statistically there was no significant difference in the distribution of seizure outcome classes between the medical and surgical groups for each of the years 1 through 4 (P ⬎ 0.05). The seizure class decreased between years 1 and 2, 1 and 4, and 2 and 4 in the surgical group; however, there was no significant decrease of seizure class after year 1 in the medical group.

control was significantly improved in the first year after video-EEG, (2) children with intractable partial seizure, in general, did not have a favorable long-term seizure outcome; approximately 50% of them remained intractable, and only 30% attained Class I seizure outcome; (3) the presence of focal lesions on neuroimaging was a risk for poor outcome with antiepileptic drug treatment; and (4) almost all children (90%) who underwent epilepsy surgery had this risk of poor prognosis, although 60% of them achieved Class I seizure outcome. Patients and Outcome Assessment This study was not prospectively randomized for seizure outcome. The vast majority of patients in this study were originally referred by general pediatricians in the Los Angeles County area for seizure diagnosis and management instead of epilepsy surgery evaluation. Except for three patients who were out-of-state referrals, neurologists at Childrens Hospital Los Angeles had treated all patients in the surgical group for a number of years before epilepsy surgery was considered. Therefore patients in this study were representative of the patients with intractable partial seizures usually observed in a pediatric neurology practice. Significantly more children in the surgical group had seizure duration longer than 2 years (90% vs 60%) and a single focal lesion on routine neuroimaging (65% vs 20%) than children who stayed in the medical group. These differences are a common and well-accepted bias in selecting patients for epilepsy surgery evaluation. The comparison of the seizure outcome between these two groups is, therefore, clinically reasonable because a controlled, randomized trial of comparing antiepileptic drug treatment vs surgical intervention would be ethically difficult. The main concern for a retrospective review of seizure outcome is the accuracy of seizure frequency estimation. The advantages of the outcome classification system used in this study were twofold. First, the substantial difference of seizure frequency in each outcome class can eliminate counting minor changes. Second, seizure frequency was classified based on simple clinical assessment. For example, an improvement from Class II to Class I indicated that patients could be seizure free up to 6 months, and patients could be seizure free for up to 1 month when seizure control improved from Class III to Class II. This classification system emphasized the seizure-free period more than the total number of seizures, thus, implying an improvement in the quality of life. Seizure Outcome of Medical Treatment

Discussion This retrospective study focused on the outcome of seizure control in children with extreme intractable partial seizure disorder which was confirmed by video-EEG study. The main observations were as follows: (1) seizure

Published outcome studies on partial epilepsy in children (excluding benign focal epilepsy of childhood) did not focus on the patients with extreme intractable partial seizures. The diagnosis of a partial seizure was not made by video-EEG [2-4] in most studies. Bye and Foo [5]

Chen et al: Seizure Outcome in Intractable Partial Epilepsy 285

retrospectively studied 17 young children (less than 10 years of age) with complex partial seizure disorder who had video and telemetry evaluations. The seizure outcome was poor in 12 patients who stayed on antiepileptic drug treatment, seven (58%) of whom remained intractable (more than two seizures per month). In another study using clinical and strict electroencephalogram criteria, Kotagal et al. [6] reported that seven of 12 patients remained at the same poor seizure control (more than one seizure per week) after 5 years of antiepileptic drug treatment. A long-term outcome study on children who had at least 2 years of poorly controlled seizures (more than one seizure per month), diagnosed by the clinical observation and presence of epileptiform discharges on routine scalp electroencephalogram, was reported by Huttenlocher and Hapke [7]. Among those 59 patients who had partial seizure disorder, only 36% of them achieved excellent seizure control (less than one seizure per year) after 5-20 years of follow-up. Our data also indicated an unfavorable seizure outcome in children with medically refractory partial seizure disorder that began in early childhood (less than 5 years of age). Almost one half of these children remained intractable (one or more seizures per week) on long-term polypharmotherapy. Although the percentage of patients who achieved Class I seizure outcome increased after 1 year and appeared to plateau in 2 years, there was no statistical difference between year 1 and year 2 followups. Therefore if partial seizures remain intractable after 1 year of antiepileptic drug therapy, it is unlikely that further significant improvement of seizure control will be attained by adding more antiepileptic drugs. However, a number of antiepileptic drugs and the new treatment modality of vagus nerve stimulation were approved for epilepsy treatment after 1995. Whether or not these newly available treatments will improve long-term outcome in these children remains to be evaluated. In this study, seizure outcome was significantly improved in 1 year after video-EEG. The improvement of seizure control in the first year of follow-up could be a result of an adjustment of an antiepileptic drug regimen after a seizure type was confirmed. This possibility was supported by our previous data that an alteration of antiepileptic drug treatment was made in 45% of patients after video-EEG studies and our current data demonstrating that the number of patients receiving polypharmotherapy was significantly increased in the first year after video-EEG [8]. The patients in this study were primarily referred by general pediatricians instead of neurologists. It is possible that some patients did not have adequate antiepileptic drug management before video-EEG study; therefore, excellent seizure control could be achieved once sufficient treatment was initiated. Seizure Outcome of Surgical Treatment The majority of our patients (85%) in the surgical group had focal cortical resectioning in the extratemporal lobes.

286

PEDIATRIC NEUROLOGY

Vol. 26 No. 4

The surgical outcome of our patients is similar to a few published reports on children with nontumor-related intractable partial epilepsy from extratemporal lobes. Fish et al. [9] reported a series of 45 children who had epilepsy surgery involving the frontal lobe caused by various abnormalities other than brain tumor, 13 (29%) of whom attained a good prognosis of no more than two seizures per year. A seizure-free outcome was achieved in nine of 23 children (39%) who underwent resectioning for nontumorrelated epileptogenic foci from extratemporal location [10]. In Gilliam’s study [11], 10 of 33 consecutive children had focal resectioning in extratemporal locations for intractable epilepsy resulting from cortical dysplasia (nine patients) or gliosis (one patient). Five of these patients (50%) were seizure free for 7 months to 6 years. The same outcome was reported by Wyllie et al. [12], in which 11 of 22 patients (50%) were seizure free after epilepsy surgery for cortical dysplasia in extratemporal locations. Clinical factors associated with a favorable or unfavorable seizure outcome in children who have had epileptogenic foci resections at extratemporal locations are unclear. Various predictors of the surgical outcome for temporal lobe epilepsy have been reported, including age of seizure onset, age of surgery, developmental level, severity of seizure, duration of epilepsy before surgery, and focal abnormality on neuroimaging or electroencephalogram. However, none of these factors predicted the outcome of children who had surgery for temporal lobe epilepsy [13,14]. Lesions with developmental abnormalities or extensive neuronal loss (e.g., hippocampal sclerosis) are associated with better surgical prognosis for temporal lobe epilepsy than normal histology [15-18]. However, a relationship between the pathology and the surgical outcome was not evident in a study of pediatric patients [14]. Our data, which was based on a small patient population, also did not support this relationship. Clinical Implications The study, which was based on a small number of patients, did not demonstrate a statistically different outcome on follow-up comparison in years 1-4 between medical and surgical treatments in children who had intractable partial seizure disorder. The presence of focal lesions on neuroimaging was the only significant risk factor for predicting a poor outcome in children who received antiepileptic drug treatment for intractable partial epilepsy. Most children who underwent epilepsy surgery in this study had this risk factor. Therefore these children more than likely represent those who would remain in Class IV seizure outcome with antiepileptic drug treatment. However, nearly 60% of these children were able to achieve Class I seizure control after epilepsy surgery. Our data implied that children with intractable partial seizures should be evaluated for epilepsy surgery candidacy if their seizures remain intractable after 1 year of adequate anti-

epileptic drug treatment, particularly those who have focal imaging abnormalities that are likely associated with extensive neuronal loss. References [1] International League Against Epilepsy Proposal for revised classification of epilepsies and epileptic syndromes. Commission on Classification and Terminology of the International League Against Epilepsy Epilepsia 1989;30:389-99. [2] Glaser GH. Limbic epilepsy in childhood. J Nerv Ment Dis 1967;144:391-7. [3] Todt H. The late prognosis of epilepsy in childhood: Results of a prospective follow-up study. Epilepsia 1984;25:137-44. [4] Deonna T, Ziegler AL, Despland PA, van Melle G. Partial epilepsy in neurologically normal children: Clinical syndromes and prognosis. Epilepsia 1986;27:241-7. [5] Bye AM, Foo S. Complex partial seizures in young children. Epilepsia 1994;35:482-8. [6] Kotagal P, Rothner AD, Erenberg G, Cruse RP, Wyllie E. Complex partial seizures of childhood onset. A five-year follow-up study. Arch Neurol 1987;44:1177-80. [7] Huttenlocher PR, Hapke RJ. A follow-up study of intractable seizures in childhood. Ann Neurol 1990;28:699-705. [8] Chen LS, Mitchell WG, Horton EJ, Snead CO III. Clinical utility of video-EEG monitoring. Pediatr Neurol 1995;12:220-4. [9] Fish DR, Smith SJ, Quesney LF, Andermann F, Rasmussen T. Surgical treatment of children with medically intractable frontal or

temporal lobe epilepsy: Results and highlights of 40 years’ experience. Epilepsia 1993;34:244-7. [10] Guldvog B, Loyning Y, Hauglie-Hanssen E, Flood S, Bjornaes H. Surgical treatment for partial epilepsy among Norwegian children and adolescents. Epilepsia 1994;35:554-65. [11] Gilliam F, Wyllie E, Kashden J, et al. Epilepsy surgery outcome: Comprehensive assessment in children. Neurology 1997;48: 1368-74. [12] Wyllie E, Comair YG, Kotagal P, Bulacio J, Bingaman W, Ruggieri P. Seizure outcome after epilepsy surgery in children and adolescents [see comments]. Ann Neurol 1998;44:740-8. [13] Goldstein R, Harvey AS, Duchowny M, et al. Preoperative clinical, EEG, and imaging findings do not predict seizure outcome following temporal lobectomy in childhood. J Child Neurol 1996;11:44550. [14] Paolicchi JM, Jayakar P, Dean P, et al. Predictors of outcome in pediatric epilepsy surgery. Neurology 2000;54:642-7. [15] Davidson S, Falconer MA. Outcome of surgery in 40 children with temporal-lobe epilepsy. Lancet 1975;1:1260-3. [16] Turner DA, Wyler AR. Temporal lobectomy for epilepsy: Mesial temporal herniation as an operative and prognostic finding. Epilepsia 1981;22:623-9. [17] Kuzniecky R, Ho SS, Martin R, et al. Temporal lobe developmental malformations and hippocampal sclerosis: Epilepsy surgical outcome. Neurology 1999;52:479-84. [18] Berkovic SF, McIntosh AM, Kalnins RM, et al. Preoperative MRI predicts outcome of temporal lobectomy: An actuarial analysis. Neurology 1995;45:1358-63.

Chen et al: Seizure Outcome in Intractable Partial Epilepsy 287