Dysplasia

Patient selection.

All cases from 1992 to 2000 of isolated temporal lobe resection for refractory epilepsy were collected from the files of the Pediatric Regional Epilepsy Program at the Children’s Hospital of Philadelphia. Patients with multilobar resections or temporal lobe resections with extratemporal subpial transections were excluded. Thirty-three of 34 patients identified with refractory TLE were eligible for inclusion in the study. One case was excluded because the medical records did not clearly state whether there was a history of FC. An arterial–venous malformation was identified in the resected specimen from this patient.

Clinical data.

All charts were reviewed for a history and description of FC, family history of epilepsy, early risk factors for epilepsy such as prior infection, head trauma, and birth injury, neurodevelopmental disabilities, MRI studies, and postoperative seizure outcome data.

Seizure semiology varied widely but was consistent with complex partial epilepsy. All patients underwent video-EEG monitoring to capture and localize ictal onset. In addition, intracranial video-EEG using subdural electrode grids was performed in 13 patients. MRI studies, including T1- and T2-weighted images in sagittal, transverse, and coronal planes, were performed in all patients using a 1.5 T magnet.

All patients had cognitive assessments by their attending neurologist. Formal neuropsychological testing was performed in 21 of 33 patients. Patients were labeled as “mentally retarded” if their full-scale IQ (FSIQ) was ≤70 or “developmentally delayed” if their developmental quotient was <50% for age. Thirty-one of 33 patients could be classified using these criteria. Surgical outcome data with a minimum follow-up length of 6 months are reported here (n = 30). Patients were considered seizure-free if they had only an isolated postoperative seizure or a seizure with medication withdrawal.

Surgical data.

A cortical incision was made from the temporal tip to the posterior margin of the desired resection, 5 to 6 cm in a language-dominant or 6 to 7 cm in a nondominant hemisphere resection or tailored using data from intracranial EEG and functional cortical mapping. The lateral temporal cortex was removed en bloc and underwent extensive histologic examination. With sharp and blunt dissection, a broad, wedge-shaped section of hippocampus extending throughout the exposed length was then removed for examination.

Pathologic data.

Two neuropathologists personally evaluated all of the cases except for Patient 1, for whom the slides and tissue could not be relocated and the original pathology report was used. In all cases, hematoxylin/eosin-stained sections were examined. Selected sections were incubated with RMDO20 (hypophosphorylated neurofilament M, H) monoclonal antibody (provided by V. Lee, University of Pennsylvania, Philadelphia, PA) at a 1:10 dilution of tissue culture supernatants and developed with a biotinylated anti-mouse streptavidin peroxidase and 3,3′-diaminobenzadine as the chromogen. The slides were then counterstained with Mayer’s hematoxylin.

Although there is no consensus on nomenclature or the pathologic features necessary for the diagnosis of CD, the presence of features commonly seen in cortical dysplasia (CD; cortical laminar disorganization, abnormal and maloriented neurons, and balloon cells) was used to assign a diagnosis of CD in this study16,17⇓ (figure). CD was diagnosed if these features were identified in any portion of the resected lateral temporal cortex. More subjective features of CD, such as increased number of neurons in the white matter and increased number of neuronal-appearing cells in layer I of neocortex, were used only in Patient 29 owing to a marked increase in the neuronal cellularity of layers I and II.

• Download figure
• Open in new tab
• Download powerpoint

Figure. Features of cortical dysplasia in temporal cortex. (a) Balloon cells. Hematoxylin/eosin; ×40. (b) Abnormally oriented and dystrophic processes of cytologically atypical neuron stained for neurofilament protein. RMDO20 antibody; ×20. (c) Cortical laminar disorganization: lack of the expected layered organization of neurons grouped together by size and orientation typical of six-layered neocortex. Randomly oriented neuronal processes that no longer have parallel apical dendrites extend toward the pial surface. RMDO20 antibody; ×10.

Statistical analysis.

“Jump In” software (SAS Institute, Cary, NC) was used to calculate Fisher’s exact two-tailed test and standard deviations. A p value of <0.05 was considered significant.

The study was carried out with the approval of the Institutional Review Board at the Children’s Hospital of Philadelphia.

Risk factors.

Clinical details of the febrile seizures are found in table E-1 (additional material can be found on the Neurology Web site; go to www.neurology.org). There were a total of 33 patients, of whom 15 had a prior FC (FC+), 9 had at least one FC of ≥20 minutes, and 9 had more than one FC. Six had brief generalized convulsions consistent with simple FC.18 Clinical details for children without FC (FC−) are listed in table E-2 (go to www.neurology.org). Demographic comparison of children (FC+ and FC−) is shown in the table. The groups did not differ with respect to sex distribution, side of surgical resection, mean age at onset of unprovoked seizures, and mean age at surgery. Mental retardation was more common in the FC− group (8/17; 42%) vs FC+ (1/14; 7%) (p = 0.018), and there was a trend toward a higher mean FSIQ in the FC+ group (FC+, 89.9 ± 15.5 SD; FC−, 75.5 ± 14.6 SD).

Risk factors for epilepsy other than FC are also listed in tables E-1 and E-2. A history of Haemophilus influenzae meningitis was present in Patients 6 and 16, and a history consistent with viral meningitis was present in Patient 3. There were five patients with a family history of FC (four in a first-degree relative and one in both maternal and paternal second-degree relatives), all in the FC+ group. Six patients, three FC+ and three FC−, had a family history of epilepsy in a first- or second-degree relative.

Surgical outcome.

In patients with at least 6 months’ follow-up, 17 of 30 (57%) were seizure-free. The seizure freedom in the FC+ cohort was slightly higher (9/13; 69%) than in the FC− group (8/17; 43%) (p = 0.28). The proportion of seizure-free patients was lower in mentally retarded patients: 1/7 (14%) vs 14/19 (74%) without mental retardation (p = 0.02). A seizure-free outcome was reported in 4 of 5 (80%) tumor patients compared with 11 of 20 (55%) with CD (p = 0.26).

Pathology.

Using our criteria to identify CD, we found similar proportions of CD in the FC+ groups: 73% (11/15) vs 56% (10/18) in the FC− group (p = 0.47) (see the figure). CD was also commonly associated with other risk factors. Two of three children with a history of a CNS infection had CD. All five children with a family history of FC and five of six with a family history of epilepsy had CD present. There were five patients with low-grade tumors (four in FC− and one in the FC+ group) and no patients with pathology characteristic of tuberous sclerosis. Neuropathologic examination of the hippocampus was available in 19 patients (9 FC− and 10 FC+). There was evidence of neuronal loss in the hippocampus in all but one of the patients (18/19; 95%). The degree of cell loss varied from severe neuronal loss to mild loss of pyramidal cells with some gliosis. Of the 18 patients with hippocampal neuron loss, 17 patients had additional pathology such as CD (n = 13), gliosis (n = 3), and a ganglioglioma (n = 1).

Preoperative MRI did not identify patients with CD. Only three patients (Patients 11, 22, and 26) had a report of suspected CD on their preoperative MRI reports.