Cerebral and Cerebellar Volume Reduction in Children with Intractable Epilepsy

Transcription

1 Epilepsia, 41( I I): , 2000 Lippincott Williams & Wilkins, Inc., Baltimore 0 International League Against Epilepsy Clinical Research Cerebral and Cerebellar Volume Reduction in Children with Intractable Epilepsy John A. Lawson,?Simon Vogrin, $Andrew F. Bleasel,?Mark J. Cook, and Ann M. E. Bye *Department of Paediatric Neurology, Sydney Children s Hospital & School of Paediatuics, University of New South Wales, Randwick, New South Wales; fst. Vincent s Hospital, Fitzroy, Victoria; and #New Children s Ho.rpita1, Westmead, New South Wales, Australia Summary: Purpme: Adult epilepsy studies have demonstrated cerebral and cerebellar volume reduction beyond the epileptogenic zone, correlating this with an inferior surgical outcome. We determined whether brain volumes were reduced in childhood epilepsy and the significance of this. Methods: Cerebral, cerebellar, and hippocampal volumes were measured by quantitative magnetic resonance imaging on 112 children (ages 4-18) with epilepsy syndrome determined by video-eeg telemetry. Eighty-seven had partial epilepsy and 25 had generalized epilepsy or indeterminate syndrome. Normative volumes were obtained from 44 child controls from the community. Results: A significant reduction in cerebral (12.6%) and cerebellar (7.9%) volume was present in the epilepsy group compared with controls. Analysis of subgroups revealed that cerebral volume was significantly decreased in frontal lobe and nonlocalized partial epilepsies. The mean hippocampal ratio of 0.73 for mesial temporal lobe epilepsy was significantly less than for all other syndromes and controls. There was no difference in the rate of hippocampal volume reduction between syndromes. There was a significant correlation between IQ and cerebral and cerebellar volume, but not duration or age of onset of epilepsy. Conclusions: Cerebral and cerebellar volume reduction is common in intractable epilepsy syndromes of childhood. These cross-sectional data suggest that brain volume reduction is present at epilepsy onset and is not a result of intractable seizures. Hippocampal asymmetry is more sensitive than volume reduction as a marker for mesial temporal lobe epilepsy, but neither measure is specific. Key Words: Cerebral-Cerebellar- Epilepsy-Reduction-Magnetic resonance imaging. Recent magnetic resonance imaging (MRI) studies have provided new insights into the question of loss of cerebral and cerebellar volume in patients with epilepsy. Quantitative MRI studies in adults with epilepsy have found loss of cerebral volume beyond the epileptogenic zone (1-3). Sisodiya et al. (1) demonstrated extrahippocampal volume loss in mesial temporal lobe epilepsy (MTLE) and correlated this with an inferior surgical outcome. Cerebellar atrophy is common in intractable epilepsy; the cause is still debated (4), but it has been correlated with a poorer surgical prognosis (5) and a lower threshold to drug side effects (6). Pediatric MRI studies have not addressed these issues. We aimed to determine whether cerebral and cerebellar volume reduction was evident in intractable childhood epilepsy syndromes, with potential significance for pathophysiology and treatment. Accepted June I, Address correspondence and reprint requests to Ann M. E. Bye, Department of Paediatric Neurology, Sydney Children s Hospital, Randwick, 203 I New South Wales, Australia. abye@unsw. edu.au Hippocampal volume (HV) loss has been thoroughly examined in the adult epilepsy literature. Indices of hippocampal asymmetry (HA) are the quantitative MRI method most commonly used to lateralize the epileptogenic focus in MTLE. Absolute HV can, in addition, be used to assess bilateral disease to guide surgical decisions (7). Our group previously reported (8,9) on the efficacy of quantitative MRI in childhood epilepsy, with high sensitivity and specificity of HA for MTLE. We required normal volume data from childhood controls to assess the utility of absolute HV measurement in children with epilepsy. METHODS Patients Patients were enrolled from the two tertiary referral pediatric epilepsy centers in the state of New South Wales, Australia, between June 1995 and June Entry criteria were children age 4-8 yr who underwent an MRI volume acquisition sequence and whose predominant seizure type was recorded on video-eeg te- 1456

2 REDUCED BRAIN VOLUME IN CHILDHOOD EPILEPSY 1457 lemetry monitoring. The study population were predominantly children with intractable epilepsy. The study had approval from the institutional ethics committees of the South-Eastern Sydney Area Health Service and New Children's Hospital, Westmead. One hundred twelve children were enrolled in the study (61 boys, 51 girls; median age 11 yr [range 4-18]). Controls The control population consisted of 44 healthy children (23 boys, 21 girls) from the community. The parents gave informed consent before the MRI. No sedation or anesthetic was used, and the child was required to lie still for <8 min. Median age was 10.3 yr (range ). A clinical evaluation with neurologic examination was performed on each child. Exclusion criteria for selection of controls were history of significant perinatal insult, previous neurologic insult (e.g., head injury), febrile convulsion, intellectual disability, or abnormal results of a neurologic examination. Classification of epilepsy syndrome Clinical classification of the epilepsy syndrome and seizure type according to the International League Against Epilepsy (10,ll) was determined independently by three investigators (A.M.E.B., J.A.L., and A.F.B.) unaware of MRI data. The classification was based on four criteria: clinical data, ictal semiology, interictal EEG epileptogenic activity, and ictal EEG abnormality. Clinical data, a medical history, and a neurologic examination were obtained in all patients when they were admitted for telemetry. Ictal semiology was determined and the predominant seizure type was classified (1 1) from continuous video-eeg recording (duration 1-5 d). Medications were reduced as necessary to induce the patient's typical events. The EEG was recorded with the international system, supplemented with sphenoidal electrodes or closely spaced electrodes when necessary to include other regions of potential epileptogenic activity. Interictal EEG was recorded in sleep and wake cycles and during routine activation procedures (hyperventilation and photic stimulation). The interictal EEG was reviewed for evidence of generalized or partial epileptiform discharges and classified on the basis of distribution maps of maximum amplitude (12). Localization of the ictal EEG was based on the region of onset of epileptiform activity (13). Epilepsy syndromes were first classified into partial or generalized. The acronym and number of cases is included in brackets. Generalized epilepsies were subclassified into symptomatic or idiopathic (GEN-SYMPT-12 or GEN-IDIO- 10). Partial epilepsies were subclassified into frontal (FLE-25) or mesial temporal (MTLE-11). Patients with temporal lobe epilepsy that could not be localized to mesial structures were classified as having other temporal lobe epilepsy (TLE-other-8). Parietal and occipital lobe epilepsies were combined as extratemporal lobe epilepsy, other (ETLE-other- 10). A localizationrelated epilepsy that could not be localized to a single region as above was classified as partial (PARTIAL-33). A syndrome that was not identified as either partial or generalized was called indeterminate (INDETERM-3). Quantitative MRI analysis Subjects underwent MRI studies using a Siemens Magnetom Vision MRI 1.5-Tesla (SCH) or a Phillips I.5-Tesla (NCH). Control data were obtained on the NCH scanner only. The MRI parameters for the coronal gradient echo volume acquisition sequence were slightly different. (NCH: TR = 17 ms, TE = 4 ms, flip angle 20, field of view 250 cm, matrix 256 x 256, slice thickness 1.5 mm; SCH: TR = 11.4 ms, TE = 4.4 ms, flip angle 8", field of view 250 cm, matrix 160 x 256, slice thickness 1.5 mm). Quantitative methods used were similar to those in published studies (14). Images were archived to magneto-optical disc and transferred to a UNIX workstation running ANALYZE 7.55 (Mayo Foundation). Raw 16-bit data were scaled to 8 bits before further analysis. A semiautomated three-dimensional (3D) morphometric procedure was used to convert the eight-bit gray-scaled image volume into a binary image through intensity thresholding, based on an initial interactive visual definition of cortical boundaries to enable the extraction of cerebrum, cerebellum, and brain stem from the surrounding extra-axial structures. The definition of this boundary relied on visual inspection. Image algebra tools were used to erode tissue surfaces while retaining homogeneity of connected cerebral and cerebellar forms. A conditional dilation tool restored the eroded layers of the cerebral and cerebellar structures. The number of dilations was determined as one greater than the number of surface erosions. The extracted binary representation of the brain was multiplied by the eight-bit gray-scale volumetric sequence to generate a segmented 3D gray-scaled brain image volume. The segmented 3D volume was then broken up into cerebral hemispheres, cerebellum, and brain stem. This was done by first moving through the volume coronally, one slice at a time, and disarticulating the cerebral hemispheres from the brain stem at the level of the superior colliculus, and from the cerebellum at the tentorium cerebelli. Right and left cerebral hemispheres were disarticulated by passing a vertical line from the falx, again using the straightened coronal views. The cerebellum was manually edited, using autoconnecting neighboring voxels, from the cerebrum along the tentorium cerebelli on coronal sections, and then turned and viewed in the axial plane to enable disarticulation from the brain stem through the isthmus of the cerebellar peduncles, using the lateral recess of the fourth ventricle as the medial boundary. Cerebellar hemispheres were defined in axial Epilepsia, Vol. 41, NU, 11, 2000

3 I458 J. A. LA WSON ET AL. sections by manual editing, splitting the structure through the roof of the fourth ventricle to the vermis. Absolute volumes were measured using seed-point autoconnection of neighboring voxels. Ventricles were excluded from the measurements. HVs were measured using an established protocol (15). All measurements were made by one observer (S.V.) who was unaware of all clinical information, including group status (controls vs. patients, epilepsy classification, gender). To determine the validity of the volumetric measurements, 10 randomly selected children had their total intracranial, cerebral, and cerebellar volumes remeasured by the technician who performed the original studies (intrarater reliability). Further, a neurologist familiar with the technique remeasured the same 10 children s total intracranial volumes (interrater reliability). To confirm that the above anatomic landmarks were easily reproducible, two investigators measured another 10 children s cerebral and cerebellar hemisphere volumes on brains already extracted from the extra-axial structures. Because two different scanners were used, using slightly different volume sequences, phantom studies were performed on both to ensure good agreement in estimates between devices. The phantom was an object constructed from Perspex and filled with doped water. Visual analysis of MRI was undertaken to determine the presence of structural lesions by a neuroradiologist. Risk factors Risk factors were collected from clinical history on admission for telemetry and medical records. Putative risk factors included febrile convulsions, intracranial infection, family history of seizures, intellectual disability, age of onset, and duration of epilepsy. Intellectual disability was determined from psychometric tests where available or educational placement. Forty-seven patients had formal psychometric testing with various instruments used because of age-appropriateness and retrospective collection of these data. To overcome the variability in instruments used, IQ scores were converted to z scores (that is, the standard deviation [SD] from the age-standardized mean of each test). The tests used (and age range) were Wechsler Intelligence Scale for Children 111 (6-16 yr), Wechsler Preschool and Primary Scale of Intelligence-Revised (3-7.5 yr), Differential Abilities Scale (3-16 yr), Stanford-Binet Intelligence Scales (22 yr), and Griffiths Mental Development Scales (0-8 yr). Data analysis Pearson s correlation coefficient was used to obtain intra- and interrater reliability coefficients. Differences in phantom volumes between the two MRI machines were analysed with Student s t test. Brain volumes of controls and patients were analyzed by linear regression using age and sex as covariates. Abnormal HV and HA was defined as being <2 SDs below the control mean. Regression analysis of controls using age and sex as independent variables allowed calculation of predicted cerebral and cerebellar volume. In patients, the residual of actual minus predicted volume was calculated. Significant reduction in cerebral and cerebellar volume was defined by a residual >2 SDs below the predicted mean. Significant asymmetry was defined by a smaller to larger volume ratio <2 SDs below the control mean. To determine the significance of the difference between mean volumes in epilepsy syndromes, analysis of covariance was applied with age, sex, and intellectual disability as covariates. The p value was adjusted for multiple comparisons using the Bonferroni transformation. The significance of putative risk factors was examined by linear regression analyses. IQ data, when available, were converted to z scores, and the relationship to volume data were analyzed by linear regression. RESULTS MRI phantom and reliability data The phantom measured 2.2% larger (t = 5.6, p < 0.001) in the NCH MRI. Because of this small but significant difference, volume data were analyzed controlling for scanner location, but the results remained unchanged. Estimates of volumes of cerebrum and cerebellum were not different between the two machines when age-matched samples from the epilepsy group were compared (p = 0.61 and p = 0.09, respectively). The interrater reliability estimates were 0.79 (p = 0.006) for total intracranial volume, 0.87 (p = 0.001) for cerebral volume, and (p = 0.013) for cerebellar volume. The intrarater reliability estimates were 0.76 (p = 0.01) for total intracranial volume. From brains already extracted from extra-axial structures, intrarater reliability estimates were 1.OO for cerebral volume and 0.99 for cerebellar volume. Control population Analysis of cerebral volume in two controls was excluded because of artifact generated by teeth braces, and in four controls HVs were excluded as a result of carotid pulsation artifact. The mean volumes and asymmetries, comparing control subjects with patients, are summarized in Table 1. Cerebral volume was affected by gender, with girls 11% smaller than boys (t = 4.8, p < 0.001). Cerebral volume increased by an average of 1.3% per year (t = 4.1, p < 0.001). The cerebral hemispheres were highly symmetric, with a mean ratio of 0.98 (SD ). Cerebellar volume was affected by age, with a 1.6% increase on average in size per year (t = 4.7, p < 0.001). The cerebellum was 3% smaller in girls (p = 0.07) after controlling for age. Cerebellar hemispheres were highly symmetric, with a mean ratio of 0.97 (SD 0.02). Epilepsiu, Vol. 41, No. 11, 2000

4 REDUCED BRAIN VOLUME IN CHILDHOOD EPILEPSY I459 TABLE 1. Brain volume data for control subjects and patients Measure (Clll ) Control subjects Patients Cerebral volume 1,236 (I, 197-1,276) Cerebral ratio 0.98 ( ) Cerebellar volume 145 ( ) Cerebellar ratio 0.97 ( ) Hippocampal volume 26 (25-28) Hippocampal ratio 0.96 ( ) 95% confience interval given in parentheses. 1,ox 1 (1,056-1,104) 0.96 ( ) 133 ( ) 0.97 ( ) 28 (27-29) 0.89 ( ) HV was not significantly related to either age or gender. Mean HV was 2,809 mm3 (SD 389), with abnormality defined as <2,031 mm3. HV was significantly related to total intracranial volume (R = 0.47, p = 0.003). The hippocampi were relatively symmetric, with no right/left predominance; the mean ratio was (SD 0.03), with abnormality defined as < Patient population Cerebral volumes and hemisphere asymmetry Controlling for age and gender effects, patients estimated mean cerebral volume was 12.6% less than that of control subjects (F = 43.9, p < 0.001). Excluding children with intellectual disability, the difference was 9. l % less than control subjects (F = 28.5, p < 0.001). On subgroup analysis, the FRONTAL and PARTIAL groups were significantly smaller (F = 4.1, p = 0.009) than control subjects. Cerebral volume reduction was present in 53/112 (47%) of patients. Cerebral hemisphere asymmetry was present in 20/112 (18%). Sixteen of these had a lesion, ipsilateral to the smaller hemisphere in 14 and contralateral in 2. Cerebellar volumes and hemisphere asymmetry Controlling for age and gender, patients estimated mean cerebellar volume was 7.9% less than that of con- trol subjects (F = 12.3, p = 0.001). Excluding children with intellectual disability, the difference was 3.4% less than control subjects (F = 3.65, p = 0.06). No individual epilepsy syndrome had mean cerebellar volume significantly smaller than control subjects. Significant cerebellar volume reduction occurred in (36%). Cerebellar hemisphere asymmetry was present in 8/112 (8%). Four of these patients had partial epilepsy and the smaller hemisphere was contralateral to electroclinical lateralization. Hippocampal volumes and asymmetry Two patients were excluded from analysis because of artifact through the temporal lobes. There was no significant overall difference between patients and control subjects for mean HV (F = 0.85, p = 0.36). Individual epilepsy syndromes were not significantly different from each other or control subjects. Absolute HV reduction was present in 17/110 (15%). There was no significant difference in the rate of unilateral HV reduction for each syndrome. HA occurred in 33/110 (30%) of patients. The rate of HA and the mean hippocampal ratio are listed in Table 2. The mean hippocampal ratio of 0.72 for MTLE was significantly less than for all other syndromes and control subjects (F = 6.6, p = 0.001). Risk factors The median age of onset of epilepsy was 4 yr (range 0-14) and the median duration was 5 yr (range 0-16). Febrile convulsions occurred in 24 (21 %). Lesions were present on MRI in 33 (29%). There was no significant difference in cerebral volume when a lesion was present (F = 0.51, p = 0.48). Intellectual disability was present in 41 patients (37%). Patients with intellectual disability had smaller cerebral (F = 22.5, p < 0.001) and cerebellar (F = 23.4, p < 0.001) volumes. In the subgroup of 47 patients who underwent formal psychometric testing, the average IQ was 1.69 SDs below the age-standardized mean of the tests used. In this subgroup, there was a significant linear relationship of IQ with cerebral (R = 0.43, p = 0.002) and cerebellar (R = 0.51, p < 0.001) volumes. Febrile convulsions, age of onset, duration of TABLE 2. Comparison of epilepsy syndromes Cerebral vol. Epilepsy syndrome (cm )* Cerebellar vol. HVR rate HA rate HA ratio MTLE 1,122 FLE 1,077 (p = 0.004) TLE-other I,I 15 ETLE 1,116 PARTIAL 1,057 (p < 0.001) GEN-SYMPT 1,134 GEN-ID10 1, t / / /33 11/ /10 2/ MTLE, mesial temporal lobe epilepsy; FLE, frontal lobe epilepsy; TLE, temporal lobe epilepsy; ETLE, extratemporal lobe epilepsy; HVR, hippocampal volume reduction; HA, hippocampal asymmetry. * Estimated mean after controlling for age, gender, and intellectual disability. t p = Epikpsia, Vol. 41, No. 11, 2000

5 I A. LA WSON ET AL. epilepsy, intracranial infection, and family history were not associated with cerebral, cerebellar, or HV reduction. DISCUSSION Control population Our control population exhibited maturational changes in brain volume consistent with other MRI studies and large autopsy series. From autopsy series, total brain weight is greater in male subjects than in female subjects, with a mean difference of 9.8% (16). Most brain growth occurs in the first 3 yr of life, with an additional 20% growth from 3-18 yr. The study by Dekaban and Sadowsky (1 6) did not separate intracranial structures, but the mean weight in g was remarkably similar to the mean intracranial volume of the controls in cc (e.g., yrold boys, 1,440 g/l,463 mm3), strongly validating our data. Many studies of maturational changes in brain volume on MRI have been published, but few have sufficient numbers or encompass a broad age range. MRI studies by Reiss et al. (17) and Giedd et al. (18) are exceptions because of large sample size and inclusion of normal children defined prospectively. Giedd et al. examined 104 healthy children aged 5-18 yr using fine contiguousslice MRI. They reported a gender and age effect on cerebral and cerebellar volumes. There was no effect of age or sex on HV (19). Our control population showed similar trends, but the absolute volumes reported are different. This may be due to a difference in the scanners used in the two studies. In our study, although two different machines were used at different centers, comparisons were made to ensure this did not introduce bias into the volume estimates, Phantom studies demonstrated a small difference between the two MRI scanners, but comparison of patient groups showed no significant variation; this small effect resulted in no difference in the final results. The reliability and reproducibility of the volumetric methodology were high, with inter- and intraobserver reliability similar to published data (20). Cerebral volume reduction This study found that children with intractable epilepsy have mean cerebral volumes 12.6% smaller than control subjects. Cerebral volume reduction has not been reported previously in childhood epilepsy, and its significance is uncertain. After controlling for the effect of intellectual disability, patients with frontal lobe epilepsy and nonlocalized partial epilepsy remained significantly smaller than control subjects. Frontal lobe epilepsy was our most common localized syndrome but had the lowest yield of lesions on MRI, as previously reported (9). In the FLE group, we plan to measure whether frontal cortical and lobar volume reduction is present. It would be of great value if the localizing and lateralizing value of this measure in FLE was found to be similar to HV reduction in MTLE. Reduction in extratemporal volume has been reported in adult studies, but only in MTLE. Cook s group (21) studied 51 adults with pathologically proven hippocampal sclerosis and found a 7% decrease in cerebral volume compared with 100 control subjects. In contrast, DeCarli et al. (22) found no decrease in cerebral volumes of 21 adults with MTLE. Sisodiya et al. (1 ) reported widespread extrahippocampal volume deficits in MTLE, suggesting that hippocampal sclerosis may be a marker of more widespread cerebral dysgenesis. These extrahippocampal abnormalities correlated with a poor outcome from temporal lobe surgery. Thus, the finding of brain volume reduction in partial epilepsies has implications for etiology and prognosis. There is a well-recognized association of intellectual disability with the extremes of both small and large brains, but recent MRI studies have been conflicting (17,23) regarding the association of cerebral volume and IQ. Reiss et al. (17), in 69 normal children, reported that a larger volume of gray matter predicted higher IQ scores. In children with intractable epilepsy, we have found an association between IQ and cerebral and cerebellar volume. An earlier age of onset and a longer duration of epilepsy were not related to cerebral volume reduction, identical to Cook s findings in adults (21). Lack of a relationship with epilepsy duration supports the hypothesis that these patients had a reduction in brain volume at the onset of epilepsy rather than volume loss secondary to seizures or their treatment. This hypothesis is limited by the retrospective cross-sectional nature of the data and requires elaboration, ideally by a longitudinal series examining the effect of epilepsy and its treatment on brain growth. Cerebellar volume reduction Patients cerebellar volumes were significantly smaller than those of control subjects, but after exclusion of children with intellectual disability, the reduction was not significant. More than one third of our sample had cerebellar volume reduction, a finding not specific to epilepsy syndrome. Cerebellar volume reduction is a common finding in adults with chronic epilepsy, present in 45% of Specht et al. s series (5) of 78 adults with TLE. Specht et al. found that a smaller cerebellum correlated with a poorer outcome after surgery. The etiology of cerebellar volume reduction remains controversial, with debate on the role of phenytoin and duration of epilepsy (4,24). Botez et al. (25), in a computed tomography study of 134 adults with epilepsy, found that a small cerebellum correlated with a longer duration of epilepsy and treatment with phenytoin. Specht et al. (5) reported that age, generalized seizures, and intellectual disability were the main contributors to cerebellar atrophy in their group, but they did not examine medication effects. In our series, similarly, age and Epilepsia, Vol. 41, No. 11, 2000

6 REDUCED BRAIN VOLUME IN CHILDHOOD EPILEPSY 1461 intellectual disability were significant predictors of cerebellar volume reduction. These factors clearly need to be considered in future studies of the etiology of cerebellar atrophy. Polypharmacy and frequent seizures are a feature of our population, so an expanded group including less severe epilepsies and analysis of individual medications is being collected to address this etiologic question. Hippocampal volume reduction Measurement of absolute HV has been advocated to detect the presence of bilateral disease, with implications for surgical (7) and neuropsychological (26) outcome. Quigg et al. (27) found that 18% of 40 patients with MTLE had evidence of bilateral hippocampal atrophy, but there was no correlation with surgical outcome. We previously reported (9) on the utility of ratios of HA, with a high sensitivity for MTLE in children. In this study, the use of absolute HV was actually less helpful than hippocampal ratios. Unilateral HV reduction was present in only four of the MTLE group, whereas asymmetry was present in nine. There were no patients with MTLE who had evidence of bilateral disease, in contrast to adult series (7,27). There was no significant difference in mean HV between epilepsy syndromes and control subjects, in contrast to the clearly significant decrease in mean hippocampal ratio confined to MTLE. In addition, although HA was a sensitive marker of MTLE, it is clearly not specific, occurring in nearly a quarter of extratemporal epilepsies. This is in marked contrast to adult MRl studies (15,28), which have noted the rarity of HA in extratemporal lobe epilepsy. The differing efficacy of quantitative MRI is probably related to the population studied: it is highly specific when applied to an adult surgical sample but much less so in a broader childhood population. Progressive hippocampal damage from recurrent seizures over many years is a possible explanation why adults with MTLE exhibit more significant volume reduction (29-32). Our data showed no relationship between the duration of epilepsy and HV, but conclusions are limited by the heterogeneous sample and the crosssectional nature of the analysis. Summary This is the first study to present evidence of significant reduction in cerebral and cerebellar volumes in children with intractable epilepsy. We demonstrated that even with partial epilepsy syndromes, there is evidence of widespread abnormality. Children with FLE, our most common localized syndrome, had a significant reduction in cerebral volume. Whether this signifies more specific isolated frontal lobe volume loss is being investigated. Volume reduction was independent of age of onset and duration of epilepsy. These cross-sectional data support the hypothesis that the brains of children with intractable epilepsy are abnormal at the onset of the seizure disorder, but this theory requires confirmation in a prospective longitudinal series. HA is more sensitive than volume reduction as a marker for MTLE. Neither measure is specific to MTLE, prohibiting reliance on MRI data alone to localize the epileptogenic zone. We are collecting data on a broader childhood epilepsy population to evaluate whether brain volume reduction is confined to children with a severe epilepsy phenotype. These findings also need to be explored in a pediatric surgical population to determine whether evidence of a diffuse abnormality of the brain-that is, brain volume reduction-is a predictor of surgical outcome. Acknowledgment: This study was supported by the NH&MRC of Australia, Sydney Children s Hospital Foundation, Brain Foundation, and the Ramaciotti Foundation. We thank the MRI departments of both children s hospitals, and Dr. John Pereira for review of MRIs. REFERENCES 1. Sisodiya SM, Moran N, Free SL, et al. Correlation of widespread preoperative magnetic resonance imaging changes with unsuccessful surgery for hippocampal sclerosis. Ann Neurol 1997;41: Marsh L, Morrell MJ, Shear PK, et al. Cortical and hippocampal volume deficits in temporal lobe epilepsy. Epilepsia I997;38: Briellmann RS, Jackson GD, Kalnins R, Berkovic SF. Hemicranial volume deficits in patients with temporal lobe epilepsy with and without hippocampal sclerosis. Epilepsia 1998;39: Ney GC, Lantos G, Barr WB, Schaul N. Cerebellar atrophy in patients with long-term phenytoin exposure and epilepsy Arch Neurol 1994;51: Specht U, May T, Schulz R, et al. Cerebellar atrophy and prognosis after temporal lobe resection. J Neurol Neurosurg Psychiatry 1997; Specht U, May TW, Rohde M, et al. Cerebellar atrophy decreases the threshold of carbamazepine toxicity in patients with chronic focal epilepsy. Arch Neurol 1997;54: King D, Spencer SS, McCarthy G, Luby M, Spencer DD. Bilateral hippocampal atrophy in medial temporal lobe epilepsy. Epilepsia 1995;36: Lawson JA, Cook MJ, Bleasel AF, Nayanar V, Morris KF, Bye AM. Quantitative MRI in outpatient childhood epilepsy. Epilepsia 1997;38: Lawson JA, Nguyen W, Bleasel AF, et al. ILAE-defined epilepsy syndromes in children: correlation with quantitative MRI. Epilepsia 1998;39:134S Commission on Classification and Terminology of the Internatioiial League Against Epilepsy. Proposal for revised classification of epilepsies and epileptic syndromes. Epilepsia 1989;30: Commission on Classification and Terminology of the International League Against Epilepsy. Proposal for clinical and electrographic classification of epileptic seizures. Epilepsia 1981;22: Morris HH, Luders H. Electrodes. In: Gotman J, Ives JR, Gloor P, eds. Long-Term Monitoring in Epilepsy. New York: Elsevier, 1985: Lee BI, Lee JD, Kim JY, et al. Single photon emission computed tomography-eeg relations in temporal lobe epilepsy. Neurology 1997;49: Hohne KH, Hanson WH. Interactive 3D segmentation of MRI and CT volumes using moifhologica1 operations. J Comput Assist Tomogr 1992;16:28S-94. 1s. Cook MJ, Fish DR, Shorvon SD, Straughan K, Stevens JM. Hip- Epilepsiu, Vnl. 41, No. 11, 2000

7 1462 J. A. LAWSON ET AL. pocampal volumetric and morphometric studies in frontal and temporal lobe epilepsy. Brain 1992;115: Dekaban AS, Sadowsky D. Changes in brain weights during the span of human life: relation of brain weights to body heights and body weights. Ann Neurol 1978;4: Reiss AL, Abrams MT, Singer HS, Ross JL, Denckla MB. Brain development, gender and IQ in children. A volumetric imaging study. Brain 1996;119: Giedd JN, Snell JW, Lange N, et al. Quantitative magnetic resonance imaging of human brain development: ages Cerebral Cortex 1996;6: Giedd JN, Vaituzis AC, Hamburger SD, et al. Quantitative MRI of the temporal lobe, amygdala, and hippocampus in normal human development: ages 4-18 years. J Comp Neurol 1996;366: Bartzokis G, Mintz J, Marx P, et al. Reliability of in-vivo volume measures of hippocampus and other brain structures using MRI. Magnetic Resonance Imaging 1993;11: Murrie V, Cook MJ, Kilpatrick C, Wilkinson L, Desmond P, Cox I. Cerebral and cerebellar volumes in patients with hippocampal sclerosis (abstract). Epilepsia 1996:37: DeCarli C, Hatta J, Fazilat S, Gaillard WD, Theodore WH. Extratemporal atrophy in patients with complex partial seizures of left temporal origin. Ann Neurol 1998;43: Tranio MJ, Loftus WC, Stukel TA, Green RL, Weaver JB, Gazzaniga MS. Brain size, head size, and intelligence quotient in monozygotic twins. Neurology 1998;50: Luef G, Burtscher J, Kremser C, et al. Magnetic resonance volumetry of the cerebellum in epileptic patients after phenytoin overdosages. Eur Neurol 1996;36: Botez MI, Attig E, Vezina JL. Cerebellar atrophy in epileptic patients. Can J Neurol Sci 1998;15: Baxendale S. Amnesia in temporal lobectomy patients: historical perspective and review. Seizure 1998;7: Quigg M, Bertram EH, Jackson T, Laws E. Volumetric magnetic resonance imaging evidence of bilateral hippocampal atrophy in mesial temporal lobe epilepsy. Epilepsia 1997;38: Gilmore RL, Childress MD, Leonard C, et al. Hippocampal volumetrics differentiate patients with temporal lobe epilepsy and extratemporal lobe epilepsy. Arch Neurol 1995;52: Theodore WH, Bhatia S, Hatta J, et al. Hippocampal atrophy, epilepsy duration, and febrile seizures in patients with partial seizures. Neurology 1999;52: Van Paesschen W, Duncan JS, Stevens JM, Connelly A. Longitudinal quantitative hippocampal magnetic resonance imaging study of adults with newly diagnosed partial seizures: one-year follow-up results. Epilepsia 1998;39: Kalviainen R, Salmenpera T, Partanen K, Vainio P, Riekkinen P, Pitkanen A. Recurrent seizures may cause hippocampal damage in temporal lobe epilepsy. Neurology 1998:50: Tasch E, Cendes F, Li LM, Duheau F, Andermann F, Arnold DL. Neuroimaging evidence of progressive neuronal loss and dysfunction in temporal lobe epilepsy. Ann Neurol 1999;45: Epilepsia, Vol. 41, No. 11, 2000