neurology, neurosurgery, surgical treatment epilepsy · zentner, hufnagel, wolf, ostertun, behrens,...

3rournal ofNeurology, Neurosurgery, and Psychiatry 1995;58:666-673

Surgical treatment of temporal lobe epilepsy:clinical, radiological, and histopathologicalfindings in 178 patients

Josef Zentner, Andreas Hufnagel, Helmut K Wolf, Burkard Ostertun, Elga Behrens,Manuel G Campos, Laszlo Solymosi, Christian E Elger, Otmar D Wiestler,Johannes Schramm

AbstractThe surgical treatment ofpharmacoresis-tant temporal lobe epilepsy is increasingrapidly. The correlation of preoperativeMRI, histopathological findings, andpostoperative seizure control is reportedfor 178 patients with chronic medicallyintractable temporal lobe epilepsy whowere operated on between November 1987and January 1993. Histopathologicallythere were distinct structural abnormali-ties in 97-2% of the surgical specimens.Signal abnormalities on MRI were pre-sent in 98-7% of patients with neoplasticlesions (n = 79), 76'6% of patients withnon-neoplastic focal lesions (n = 55), and69-2% of patients with Ammon's hornsclerosis (n = 39). Overall, structuralabnormalities were detected by MRI in82-7% of all patients. The mean postoper-ative follow up period was three years.Some 92% of the patients benefited fromsurgery: 103 patients (61.7%) were seizurefree, 26 (15.5%) had no more than twoseizures a year, and 24 (14-4%) showed areduction of seizure frequency of at least75%. Fourteen patients (8.4%) had a<75% reduction of seizure frequency. Thepercentage of patients who were com-pletely free of seizures after operation was68-5% for patients with neoplastic lesions,66-7% for Ammon's horn sclerosis, and54-0% for patients with non-neoplasticfocal lesions. By contrast, none of thepatients in whom histopathological find-ings were normal became seizure freepostoperatively. The data show that thepresence of focal lesions or Ammon'shorn sclerosis as determined byhistopathological examination is associ-ated with improved postoperative seizurecontrol compared with patients withoutspecific pathological findings. Brain MRIwas very sensitive in detecting neoplasms;however, its sensitivity and specificitywere limited with respect to non-neoplas-tic focal lesions and Ammon's hornsclerosis. Improvement of imagingtechniques may provide a more precisedefinition of structural lesions in thesecases and facilitate limited surgical resec-tions ofthe epileptogenic area rather thanstandardised anatomical resections.

(7 Neurol Neurosurg Psychiatry 1995;58:666-673)

Keywords: epilepsy; temporal lobe; magnetic reso-nance imaging; histopathology

Temporal lobe epilepsy is the most commonform of epilepsy, with a prevalence of around0 1% in the general population.' Despite opti-mal pharmacotherapy, about 30% of thepatients do not become seizure free.2Recently, there has been increased interest inthe relation between structural lesions andepileptic seizures. Abnormalities that are oftenfound in association with intractable seizuresinclude low grade tumours,3-'3 non-tumorousdisorders,"4-19 and Ammon's horn sclerosis.2024Moreover, the presence of a "dual pathol-ogy"-that is, a focal lesion in associationwith Ammon's horn sclerosis-has beenfound in many cases.25-27 Magnetic resonanceimaging has been found to be valuable forboth identification and localisation of variouslesions.28-38 Most recent reports on temporallobe epilepsy have been limited- to small num-bers of patients and predominantly considerspecific aspects. Here we provide a compre-hensive report on 178 patients with a meanfollow up period of three years. The relationbetween clinical, radiological, and histopatho-logical findings and postoperative seizure con-trol is presented in detail.

Patients and methodsBetween November 1977 and January 1993227 patients underwent surgical treatment fortemporal lobe epilepsy at the Department ofNeurosurgery, University of Bonn. Thepatients had a well documented chronic andmedically intractable epilepsy lasting for aminimum of two years. For all patients ade-quate periods of treat-ment with at- least twofirst line anticonvulsant agents such as carba-mazepine, phenytoin, phenobarbitone, or val-proic acid were required before referral forpresurgical evaluation. Histopathological eval-uation of surgical specimens showed uncer-tain findings in 49 cases. These specimens didnot show focal lesions; however, they were notappropriate to allow definite conclusions as tothe presence or absence of Ammon's hornsclerosis. Thus these 49 patients were omit-ted, leaving 178 patients in this series.

All patients underwent continuous videoEEG monitoring by cable telemetry with scalpand sphenoidal electrodes. If the site of theseizure origin remained undetermined, bilat-eral subdural strip electrodes and-morerecently-additional stereotactic depth elec-trodes were used to identify the epileptogenicarea. Extraoperative electrocorticographicstudies were done in 74 of the 178 patients(41 8%). Of these, 61 patients had subdural

Department ofNeurosurgeryJ ZentnerE BehrensM G CamposJ SchrammDepartment ofEpileptologyA HufnagelC E ElgerDepartment ofNeuroradiologyB OstertunL SolymosiInstitute ofNeuropathology,University ofBonn,GermanyH K Wolf0 D Wiestler

Correspondence to:Dr Josef Zentner,Department ofNeurosurgery, University ofBonn, Sigmund-Freud-Strale 25, 53105 Bonn,Germany.Received 15 June 1994and in revised form10 November 1994Accepted 23 January 1995

666 on D

ecember 12, 2020 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.58.6.666 on 1 June 1995. D

ownloaded from

http://jnnp.bmj.com/

Surgical treatment of temporal lobe epilepsy: clinical, radiological, and histopathologicalfindings in 178 patients

electrodes only, and 13 had both subduraland depth electrodes. The site and extent ofthe resection were based mainly on the ictaland interictal EEG pattern and-if present-on the site of a localised structural lesion asdetermined by MRI. Other information wasused to confirm the location of the epilepto-genic area and to assess the risks and prognosisof resecting this. The information included adetailed evaluation of seizure history andsemiology, neuropsychological studies, psy-chiatric evaluation, and intracarotid amylo-barbitone testing in most cases. Studies withsingle proton emission computed tomography(SPECT) have been performed in a subgroupof patients. A detailed analysis will be pub-lished elsewhere.39

In all, 168 MRI examinations were avail-able for review. Of these, about 50% wereperformed at the University of Bonn MedicalCentre. The rest were done at various otherinstitutions. At our institution, MRI was car-ried out with a 1-5 Tesla unit (PhilipsGyroscan S15) with at least sagittal TIweighted images (TR/TE/slice thickness =500-600 ms/15-25 ms/8-10 mm) and axialproton density weighted and T2 weightedimages (TR/TE/slice-thickness = 2000-2500ms/20-30 and 80-120 ms/6-8 mm). Usually,spin echo scans were performed. T2 weightedgradient echo scans were done only rarely. If atumour was suspected, additional coronal andaxial Ti weighted images with and withoutgadolinium-DTPA were acquired with similarparameters. If there was prior evidence for atemporal focal lesion, axial images were oftenacquired with a modified angulation parallelto the long axis of the temporal lobe-that is,about 250 off the orbitomeatal line.

Examinations with MRI from other institu-tions included a wide range of technical varia-tions. Employing systems of variousmanufacturers and field strengths, slices weregenerally 1-2 mm thicker than those obtainedat our institution, whereas acquisition of oneTI weighted and one T2 weighted scan intwo different planes was standard. Some-

Figure 1 Schematic presentation of different resective procedures usedfor this series:standard anterior temporal lobectomy with and without hippocampectomy (A); "key-hole"resection with hippocampectomy (B); extended lesionectomy with and withouthippocampectomy (C); and selective amygdalohippocampectomy (D).

times, no proton density weighted images hadbeen performed. The modified "temporal"orientation for axial images was not used inthese examinations. The differences in imag-ing techniques reflect the fact that there is nostandardised protocol for the examination ofpatients with temporal lobe epilepsy.

All preoperative MRI studies were retro-spectively evaluated by a neuroradiologistwith no other information available. Findingson MRI were classified as follows: tumour,non-tumorous lesion, signal abnormalities of"uncertain" significance, Ammon's hornsclerosis, and no detectable abnormality. Theterm "uncertain" was used with respect to theinterpretation of imaging studies if the find-ings were not conclusive due to insufficientimage quality, questionable partial volumeeffects, or the lack of appropriate sequences orsectional planes. The diagnosis of Ammon'shorn sclerosis was based on the following cri-teria: increased signal on T2 weighted imagesor temporomesial atrophy with an enlargedtemporal horn of the lateral ventricle.Volumetric measurements of hippocampalsize were not performed.The following surgical procedures were

performed: anterior temporal lobectomy(standard or "keyhole") with hippocampec-tomy (n = 144), anterior temporal lobectomywithout hippocampectomy (n = 8), extendedlesionectomy with hippocampectomy (n = 9),extended lesionectomy without hippocam-pectomy (n = 15), and selective amygdalo-hippocampectomy (n = 2) (fig 1). Thirteenpatients in whom anterior temporal lobec-tomy with hippocampectomy was performedhad been operated on previously at otherinstitutions for tumours; however, theirseizure outcome had not been improved.

All operations were performed under gen-eral anaesthesia without intraoperative elec-trocorticography. In every case, completeremoval of a radiologically identified lesionwas intended. A standard anterior temporallobectomy consisted of a 4-5 cm (non-domi-nant hemisphere) and 4 cm (dominant hemi-sphere) resection of the superior and middletemporal gyri and a 5*5 cm (non-dominanthemisphere) and 5 cm (dominant hemisphere)resection of the inferior temporal gyrus. Thiswas followed by removal of the amygdala, theparahippocampal uncus, and total resection ofthe hippocampus and parahippocampal gyrus.If there was adequate memory performance ofthe non-operated temporal lobe as evaluatedby neuropsychometric and intracarotidal amy-lobarbitone testing, the anterior 3 to 3-5 cm ofthe parahippocampal gyrus and hippocampuswere resected.

In cases with insufficient contralateralmemory performance, resection of the hip-pocampus was limited to 1-2 cm, or thehippocampus was preserved completely.Extended lesionectomy was performed by ananterolateral or posterolateral approachdepending on the location of the lesion. Insome patients, these procedures were com-bined with amygdalohippocampectomy.Selective amygdalohippocampectomy was

667 on D



j.com/

J Neurol N


ownloaded from


Zentner, Hufnagel, Wolf, Ostertun, Behrens, Campos, Solymosi, Elger, Wiesder, Schramm

performed by the trans-sylvian route asdescribed by Wieser and Yasargil.40

All glass slides of operative specimens werereviewed by two neuropathologists. Sectionsstained with haematoxylin and eosin wereavailable from all cases. In most specimensthere were also slides with Nissl stains andcombined haematoxylin-eosin-luxol-fast bluestains available. In selected cases additionalspecial stains (elastic van Gieson, reticulin,Bodian) were used. In all tumour cases someor all of the following immunohistochemicalreactions were carried out: glial fibrillary acidprotein, synaptophysin, neurofilament pro-tein, neuron specific enolase, and S 100 pro-tein. Tumours were classified according to therevised World Health Organisation classifica-tion for tumours of the nervous system.40Histological findings were divided into fourgroups: neoplasm, non-neoplastic focal lesion,Ammon's horn sclerosis, and absence of spe-cific histopathological alterations.

Follow up information, available from 167patients, was based on regular outpatient visitsto hospital at three to 12 month intervals.Postoperative observation time ranged from12 to 72 months (mean three years). Withrespect to the postoperative seizure state,patients were assigned to four different out-come classes as described by Engel et al30:(a) seizure free or only auras since surgery,(b) rare seizures (4 2 per year), (c) reductionof seizure frequency >75%, and (d) un-changed (<75% reduction of seizure fre-quency). Student's t test was used forstatistical comparison of continuous variablesif the distribution was normal, and Wilcoxon'stest if it was not. Discrete variables were com-pared by cross tabulation.

ResultsCLINICAL FINDINGSThere were 82 male and 96 female patientswith ages ranging from 3 to 64 years and amean age of 27-7 years. All patients had com-plex partial seizures, 78 of them with sec-ondary seizure generalisation.

Patients' ages at onset of seizures rangedbetween 1 and 45 (mean 15-9) years.

Table 1 Main histopathological diagnoses in 178specimens from patients with temporal lobe epilepsy

PatientsHistopathological diagnosts No (%)

Neoplasm 79 (44 4)Non-neoplastic lesion

Glioneuronal hamartia 19 (10-7)Glioneuronal hamartoma 5 (2 8)Vascular malformation 14 (7 9)Old necrosis 3 (1-7)Abundant neurons in white matter 3 (1-7)Chronic encephalitis 2 (1-1)Other 9 (5 0)

Ammon's hom sclerosis 39 (21-9)Normal 5 (2 8)Total 178 (100-0)

In some specimens more than one pathological abnormalitywas present. Only 72 specimens were sufficiently preserved topermit an evaluation as to the presence of Ammon's hornsclerosis. The incidence of Ammon's horn sclerosis in thissubgroup was 72-2%.

Duration of epilepsy varied between two and52 years with a mean of 12-2 years. Patients'mean age at surgery was 26-2 years with neo-plastic lesions, 26-5 years with non-neoplasticfocal lesions, and 29-2 years with Ammon'shorn sclerosis. Mean age at onset of seizureswas 14-2 years in patients with neoplasticlesions, 14-7 years in patients with non-neo-plastic focal lesions, and 17-9 years in patientswith Ammon's horn sclerosis. As determinedby Student's t test the patients' ages at thetime of surgery and the ages at the first onset ofseizures were significantly lower in patientswith focal lesions than in those without focallesions (P < 0X01). There was considerable over-lap between the two groups. No statisticallysignificant differences were present for dura-tion of seizures between any of these groups.

HISTOPATHOLOGICAL FINDINGSTable 1 gives the results of histopathologicalexamination of the specimens obtained atoperation. In 106 patients (56-7%), a definitehistopathological statement as to the presenceor absence of Ammon's horn sclerosis wasimpossible. In some cases, segmental neu-ronal loss and gliosis were obscured by infil-trating tumour or inflammation. More often,however, the portion of the hippocampal for-mation that was actually submitted for patho-logical evaluation consisted only of smalltissue fragments that did not permit a properorientation during embedding and a reliableidentification of the different segments ofAmmon's horn. In many cases, the lack of anappropriate specimen was due to the extensiveuse of ultrasonic aspiration. Most specimenscontained minor and non-specific changessuch as focal gliosis of the end folium or otherparts of the hippocampal formation, subpialgliosis, or diffuse white matter gliosis irrespec-tive of the presence or absence of a specificfocal lesion or classic Ammon's horn sclerosis.As representative specimens of the hippocam-pus were not available in many cases, it wasimpossible to give a specific figure for thenumber of entirely normal samples for thewhole series.The most frequent tumour was a gangli-

oglioma, and all but two tumours were oflow histopathological grade (table 2). Dysem-bryoplastic neuroepithelial tumours were pre-sent in eight patients. This entity has recentlybeen included in the 1993 World HealthOrganisation (WHO) classification of CNS

Table 2 Histopathologcal diagnoses in 79 tumours

PatientsDiagnosis No (%)Ganglioglioma (WHO grade I) 29 (36-7)Ganglioglioma (WHO grade II) 3 (3 8)Anaplastic ganglioglioma (WHO grade III) 1 (1-3)Pilocytic astrocytoma (WHO grade I) 18 (22-7)Astrocytoma (WHO grade II) 6 (7-6)Anaplastic astrocytoma (WHO Grade III) 1 (1-3)Oligodendroglioma (WHO grade II) 10 (12-6)Oligoastrocytoma (WHO grade II) 1 (1-3)Pleomorphic xanthoastrocytoma 1 (1-3)Epidermoid 1 (1-3)Dysembryoplastic neuroepithelial tumour 8 (10-1)Total 79 (100-0)

668 on D



j.com/

J Neurol N


ownloaded from



Table 3 MRI diagnoses in 77 histopathologically verified neoplastic lesions

MRI diagnosis

Tumour Non-tumorous "Uncertain" Normal TotalHistopathological diagnosis No lesion No No No No

Ganglioglioma (WHO grade I) 21 1 5 1 28Ganglioglioma (WHO gradeII) 3 - - - 3Ganglioglioma (WHO grade III) 1 - - - 1Astrocytoma (WHO grade I) 16 1 - - 17Astrocytoma (WHO gradeII) 5 - 1 - 6Astrocytoma (WHO grade II) 1 - - - 1Oligodendroglioma (WHO grade II) 10 - - - 10Oligoastrocytoma (WHO grade II) 1 - - - 1Pleomorphic xanthoastrocytoma 1 - - - 1Dysembryoplastic neuroepithelial tumour 7 - 1 - 8Epidermoid 1 - - - 1Total 67 2 7 1 77

tumours41 and its detailed histopathologyremains to be determined. In the presentseries, two tumours that had been previouslydesignated as gangliogliomas were reclassifiedas dysembryoplastic neuroepithelial tumours.Among non-neoplastic focal lesions, glioneu-ronal hamartias and cavemomas were mostoften found (table 1). Ammon's horn sclerosiswas the only abnormality in 39 cases (17.2%).Correlation of focal lesions and Ammon'shorn sclerosis showed that Ammon's hornsclerosis was significantly more frequent inpatients without focal lesions. There was asmall group, however (25 0%), of patientswith Ammon's horn sclerosis and focal lesions("dual pathology"). Only five patients hadneither Ammon's horn sclerosis nor focallesions. More detailed descriptions on thehistopathological findings have recently beenpublished elsewhere.1942

MRI FINDINGSBrain MRI showed structural abnormalities in139 of 168 cases (82.7%). There were signalabnormalities on MRI in 76 of the 77histopathologically verified tumours (98&7%),and in 67 cases (87.0%) the MRI diagnosis

Table 4 MRI diagnoses in 47 histopathologically verified non-neoplastic focal lesions

MRI diagnosis

Non-tumorousHistopathological lesion Tumour "Uncertain" Normal Totaldiagnosis No No No No No

Hamartia - 5 4 10 19Hamartoma - 3 2 - 5Cavernoma 11 - - - 11AV malformation 2 - 1 - 3Necrosis - 1 1 1 3Heterotopia 3 - - - 3Abscess wall - 1 - - 1Encephalitis - 1 - - 1Gliosis - 1 - - 1Total 16 12 8 1 1 47

Table S MRI diagnoses in 44 patients with and without definite Ammon's horn sclerosis(AHS)

MRI diagnosis

Histopathological AHS Tumour "Uncertain" Normal Totaldiagnosis No No No No No

AHS 21 2 4 12 39No AHS - - - 5 5Total 21 2 4 17 44

AHS = Ammon's horn sclerosis.

was that of a tumour. In seven tumour casesMRI showed signal abnormalities of uncertainrelevance, and in one case no signal abnor-mality was found. In six of these eight cases,histopathological examination showed a cir-cumscribed ganglioglioma (table 3).Of 47 histopathologically verified non-neo-

plastic focal lesions, MRI was abnormal in 36cases (76 6%). Cavernomas and heterotopiaswere correctly assessed in every case, whereashamartias and hamartomas were often missedor interpreted as neoplasms. In 11 cases(23.4%), MRI was normal. In 10 of thesepatients, histopathological examinationsshowed microscopic glioneuronal hamartias(table 4).Of the 39 patients with histopathologically

verified Ammon's horn sclerosis and no focallesion, MRI was abnormal in 27 cases(69-2%). In 21 of them (53-8%), Ammon'shorn sclerosis was correctly diagnosed,whereas in two cases a tumour was suspecteddue to high signal intensity. All five patientswith inconspicuous histopathological findingshad normal MRI (table 5). Figures 2-5 arerepresentative MRI findings from ourpatients.

FOLLOW UPThere was no operative or postoperative mor-tality. The postoperative course was compli-cated in 18 of 178 patients (10.1%). Inparticular, six patients had deep wound infec-tion and two had meningitis. Neurologicalimpairment was encountered in sevenpatients: aphasia (one), third nerve paresis(two), hemianopsia (one), and hemiparesis(three). Deep vein thrombosis was found intwo cases and peripheral nerve deficits due toinadequate positioning on the operating tablein one case. Almost all of these complicationsresolved completely without further sequelae.Permanent morbidity occurred in threepatients (1-7%) and included hemiparesis intwo cases, and hemianopsia in one case.

Table 6 shows the outcome of patients withrespect to seizure control. In total, 91 6% ofthe patients were seizure free or had a worth-while improvement after surgery, whereas8-4% had no appreciable reduction in seizurefrequency or were unchanged. We did notencounter any significant differences inseizure outcome with respect to the side ofoperation (right v left). The most favourable

669

on Decem

ber 12, 2020 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.58.6.666 on 1 June 1995. Dow

nloaded from


Zentner, Hufnagel, Wolf, Ostertun, Behrens, Campos, Solymosi, Elger, Wiestler, Schramm

Figure 2 Ammon's hornsclerosis: MRI with T2weighted temporallyoriented image (left) andTl weighted coronal image(inversion recovery)(right) shows increasedsignal (T2 weighted) andright hippocampal atrophy(Tl weighted) (arrows).

results were in patients with neoplastic lesionsand Ammon's horn sclerosis. With thesehistopathological diagnoses, 68-5% and66-7% of patients were seizure free. The ratioof seizure free patients with non-neoplasticfocal lesions was somewhat lower (54 0%).These differences were statistically significant(P < 0-05). None of the five patients in whomthere were no specific histopathological find-ings became seizure free postoperatively (table7). The differences in seizure outcomebetween patients without specific histopatho-logical alterations and those with the diagno-sis of a focal lesion or Ammon's horn sclerosiswere highly significant (P < 0-001).

DiscussionThe surgical management of epilepsy involvesthe identification of the epileptogenic cortex

Figure 3 Cavernoma.There is a righttemporomesial lesion withcentral hyperintensity andperipheral signal voidtypicalfor cavernoma onaxial Tl weighted MRI.Note enlarged tip ofinferior horn.

and its excision as completely as possible.3643Despite complex diagnostic tools the defini-tion of the epileptogenic area remains hypo-thetical. It is generally accepted, however, thatstructural abnormalities play a significant partin epileptogenesis and that altered brain hasto be removed completely to achieve seizurecontrol. Brain MRI has proved to be highlysensitive both for detection and localisation ofstructural lesions. 1528323844-47 This study of178 patients treated surgically for temporallobe epilepsy was designed to evaluate thesensitivity and specificity of MRI for identify-ing structural abnormalities and to define therelevance of histopathological findings withrespect to seizure control.

In our series, histopathological examinationshowed definite structural abnormalities in173 of 178 specimens (97.2%). Forty ninepatients who did not show a focal lesion andwhom histopathological findings did not allowdefinite conclusions as to the presence orabsence of Ammon's horn sclerosis wereexcluded from the total series of patientsoperated on during the respective period.Neoplastic lesions occurred most often(44 4%), and all but two tumours werehistopathologically benign. Among non-neo-plastic focal lesions (309%), hamartias andcavernomas were predominant. The relativelylow rate of definite Ammon's horn sclerosis(21-9%) reflects the fact that many specimenssubmitted for pathological evaluation con-sisted only of small tissue fragments that didnot allow a reliable identification of the differ-ent segments of Ammon's horn. In manycases the lack of an appropriate specimen wasdue to the extensive use of ultrasonic aspira-tion. Regarding only specimens in which thehippocampal formation was well preserved(n = 72), 72-2% showed Ammon's horn scle-rosis. This number includes coincidence ofAmmon's horn sclerosis with focal lesions as adual pathology (25-0% of focal lesions). ThusAmmon's horn sclerosis represents the most

670 on D



j.com/

J Neurol N


ownloaded from



Figure 4 Glioneuronalhamartia. MRI with fourcoronal T2 weighted slicesshows a pardy cystic lesionwith pronounced corticaland subcorticalhyperintensity in the rightmesial temporal lobe.

important structural abnormnality in temporallobe epilepsy, whereas low grade gliomas andnon-neoplastic focal lesions are found with alower frequency. 12 23 24 26 47 48 The fact thatalmost all lesions found in our patients werehistopathologically benign corresponds well

with the usually long history of seizures,which was similar in patients with and withoutlesions, although the age at time of surgeryand the age at the first onset of seizures wassignificantly lower in patients with focallesions than in those without focal lesions.

Figure 5 Ganglioglioma(WHO grade I). Protondensity (left) and T2weighted (right) axialMRI illustrate a minimallyspace occupying lesion ofthe left uncus withpronounced hyperintensityin both images.

671 on D



j.com/

J Neurol N


ownloaded from


Zentner, Hufnagel, Wolf, Ostertun, Behrens, Campos, Solymosi, Elger, Wiestler, Schramm

Table 6 Seizure outcome in 167 patients divided intofour classes related to the operation side

Patients

Outcome Right Left Totalclass No (%) No (%) No (%)I 59 (59 6) 44 (64-7) 103 (61-7)II 16 (16-2) 10 (14-7) 26 (15-5)III 15 (15-1) 9 (13-2) 24 (14-4)IV 9(9-1) 5(7.4) 14(8-4)Total 99 (100 0) 68 (100-0) 167 (100 0)

The high anatomical resolution and thelack of artifacts from overlying bone havebeen stressed as advantages of MRI over CT;MRI is more sensitive in detecting smallstructural lesions than CT with the exceptionof small areas of calcification for which CT ismore sensitive.'346 49-52 In view of these find-ings, MRI is used in most institutions includ-ing our own as the initial imaging procedure.Abnormalities are found on MRI in 40% to70% of patients with temporal lobe epilepsy,28 48and 20% to 25% of these abnormalities aremissed by CT.46 50-52 In our series, MRIshowed structural abnormalities in 82-7%.Sensitivity and specificity ofMRI were highestin tumours (98-7% and 87'0%). In non-tumorous lesions, MRI was abnormal in 36 of47 cases (76-6%); however, only 16 of theselesions (34 0%) were recognised to be non-tumorous. It is remarkable that none of thehamartias and hamartomas were recognisedas lesions by MRI. This may be in part due tothe small size and variable distribution patternof these lesions. In some patients, however,the glioneuronal malformations were of con-

siderable size and were therefore classified as

probable neoplasms based on MRI criteria.Special attention has been paid to the sensitiv-ity of MRI for detection of Ammon's hornsclerosis. By contrast with Sperling et al,38 whodid not find MRI abnormalities in patientswith Ammon's horn sclerosis, others haveshown abnormal signals in 50% to 80%.21624Even severity of pathological changes seemsto correlate with the relative intensity of theabnormal signals.'6202' In our series, MRI was

abnormal in 69-2% and Ammon's horn scle-rosis was correctly diagnosed in 53-8%. Theseabnormalities were best seen on T2 weightedand proton density weighted images in thecoronal plane. Abnormal high intensity signalin two of our cases gave rise to the diagnosisof a tumour. In 12 cases, MRI was inconspic-

Table 7 Seizure outcome related to histopathological diagnosesHistopathological diagnosis

Non-neoplasic Ammon's hornOutcome Neoplasm lesion sclerosis Normal Totalclass No (%/.) No (%/.) No (%) No (%o) No (%Io)I 50 (68-5) 27 (54 0) 26 (667) - (-) 103 (61-7)II 13 (17-8) 7 (14-0) 6 (15-4) - (-) 26 (15-5)III 9 (12-3) 8 (16-0) 5 (12-8) 2 (40 0) 24 (14-4)IV 1 (1-4) 8 (16-0) 2 (5-1) 3 (60 0) 14 (8-4)Total 73 (100 0) 50 (100-0) 39 (1000) 5 (100 0) 167 (100-0)

uous; however, in all of these cases the qualityof the images was unsatisfactory. Volumetricor signalometric measurements of the hip-pocampus may represent a valuable diagnos-tic tool for demonstration and quantificationof hippocampal sclerosis.345354 These mea-surements have not been performed at ourinstitution so far.With respect to seizure outcome, about

92% of our patients benefited from surgery.These results correspond favourably withthose reported from other centres.30 "5 Wedid not encounter significant differences inseizure outcome comparing the side ofsurgery (right v left). Patients with Ammon'shorn sclerosis or focal lesions had a signifi-cantly better outcome than patients withoutspecific histopathological findings. It seems tobe remarkable that none of the patients inwhom specific structural changes were absentbecame seizure free postoperatively. Thisindicates that histopathological evaluationprovides prognostically relevant information.

Controversy exists with respect to the surgi-cal approach in temporal lobe epilepsy andintraoperative techniques. At some centres theindication for surgery is based on the locationof interictal events as determined by scalpEEG, and the resection is carried out underlocal anaesthesia and tailored to the results ofintraoperative electrocorticography (ECoG)and functional mapping.58-60 Other centresbase the selection of surgical candidates onextraoperative recordings of seizure onsetthrough depth electrodes. This is followed byan anatomically standardised resection thatmay be either a standard anterior temporallobectomy with resection of both mesial andlateral structures6' or a more limited resectionsuch as a selective amygdalohippocampectomy.40In the present series, we have combined fea-tures of these two opposite approaches. Thedefinition of the epileptogenic area was basedon extraoperative evaluation of the seizurefocus. In earlier non-lesional cases, resectionof lateral structures was a standardisedremoval of the anterior 4 to 5 cm of the tem-poral lobe. More recently, we have limitedand individualised the lateral resectionsaccording to the patient's EEG and functionallocalisation results. In our experience, com-bined subdural and hippocampal depth elec-trodes have proved to be valuable todistinguish the foci of mesial and lateralseizures. Obviously, no single approach toresective surgery is suitable for all patients.Although in most cases the location of a struc-tural lesion roughly indicates the epilepto-genic area the relation is not absolute and theextent of resection should be tailored toextraoperative electrocorticographic findingsto include the cortical zones of seizure originand maximal interictal epileptiform activity.

In conclusion, detailed histopathologicalevaluation of temporal lobe epilepsies empha-sises the role of structural abnormalities forepileptogenesis and allows important prognos-tic conclusions. Magnetic resonance imaging isreliable in identifying tumorous lesions. Imagingtechniques need further improvement,

672 on D



j.com/

J Neurol N


ownloaded from



however, for detection of non-tumorous focallesions and Ammon's horn sclerosis. Furtherimprovement of morphological and functionalimaging techniques and refined electrophysio-logical examinations can be expected to allow amore precise preoperative definition of theepileptogenic area. This will further facilitatethe use of tailored resections instead of classicstandardised operations.We thank DK Boker and W Entzian for their contribution tothis study.

1 Keranen T, Reikkinen P. Severe epilepsy: diagnostic andepidemiological aspects. Acta Neurol Scand 1988;78(suppl) :7-14.

2 Hauser WA. The natural history of temporal lobe epilepsy.In: Luders HO, ed. Epilepsy surgery. New York: RavenPress, 1992:133-41.

3 Bergen D, Bleck T, Ramsey R, et al. Magnetic resonanceimaging as a sensitive and specific predictor of neo-plasms removed for intractable epilepsy. Epilepsia 1989;30:318-21.

4 Boon PA, Williamson PD, Fried I, et al. Intracranial,intraaxial, space-occupying lesions in patients withintractable partial seizures: an anatomo-clinical, neu-ropsychological, and surgical correlation. Epilepsia199 1;32:467-76.

5 Cascino DG. Epilepsy and brain tumors: implications fortreatment. Epilepsia 1990;31(suppl):37-44.

6 Cavanagh JB. On certain small tumours encountered inthe temporal lobe. Brain 1958;81:389-405.

7 Daumas-Duport C, Scheithauer BW, Chodkiewiez JP, etal. Dysembrioblastic neuroepithelial tumor: a surgicallycurable tumor of young patients with intractable partialseizures. Neurosurgery 1980;23:545-56.

8 Falconer MA, Serafetinides EA, Corsellis JAN. Etiologyand pathogenesis of temporal lobe epilepsy. Arch Neurol1964;10:233-48.

9 Ketz E: Brain tumours and epilepsy. In: Vinken andBruyn, eds. Handbook of clinical neurology. Vol 16,Amsterdam: Elsevier 1974:254-69.

10 Rasmussen T. Surgery of epilepsy associated with braintumors. In: Purpura DP, Penry JK, Walter RD, eds.Advances in neurology. Vol 8. New York: Raven Press,1975:227-39.

11 Rich KM, Goldring S, Gado M. Computed tomography inchronic seizure disorder caused by glioma. Arch Neurol1985;42:26-7.

12 Spencer DD, Spencer SS, Mattson RH, Williamson PD.Intracerebral masses in patients with intractableepilepsy. Neurology 1984;34:432-6.

13 Tampieri D, Moumdjian R, Melanson D, et al.Intracerebral gangliogliomas in patients with partialcomplex seizures. CT and MR Imaging findings. AJNRAmJ7Neuroradiol 1991;12:749-55.

14 Armstrong DD. The neuropathology of temporal lobeepilepsy. JfNeuropathol Exp Neurol 1993;52:433-43.

15 Dowd CF, Dillon WP, Barbaro NM, et al. Magnetic reso-nance of intractable complex partial seizures: pathologicand electro-encephalographic correlation. Epilepsia1991;32:454-9.

16 Kuzniecky R, de la Sayete V, Ethier R, et al. Magnetic res-onance imaging in temporal lobe epilepsy: pathologicalcorrelations. Ann Neurol 1987;22:321-47.

17 Smith A, Weinstein M, Quencer R, et al. Association ofheterotopic grey matter with seizures: MR imaging.Radiology 1988;168:195-8.

18 Vinters HV, Armstrong DL, Babb TL, et al. The neu-ropathology of human symptomatic epilepsy. In: EngelJ, ed. Surgical treatment of the epilepsies. 2nd ed. NewYork: Raven Press, 1993:593-608.

19 Wolf HK, Campos MG, Zentner J, et al. Surgical pathol-ogy of temporal lobe epilepsy. Experience with 216cases. J Neuropathol Exp Neurol 1993;52:499-506.

20 Berkovic SF, Andermann F, Oliver A, et al. Hippocampalsclerosis in temporal lobe epilepsy demonstrated bymagnetic resonance imaging. Ann Neurol 199 1;29:175-82.

21 Bronen RA, Cheung G, Charles JT, et al. Imaging findingsin hippocampal sclerosis: correlation with pathology.AJNR Am 7Neuroradiol 1991;12:933-40.

22 Cascino DG, Jack CR, Parisi JE, et al. Magnetic resonanceimaging-based volume studies in temporal lobe epilepsy:pathological correlations. Ann Neurol 1991;30:31-6.

23 Falconer MA. Mesial temporal (Ammon's horn) sclerosisas a common cause of epilepsy: etiology, treatment, and

_prevention. Lancet 1974;2:767-70.24 Gates JR, Cruz-Rodriguez R. Mesial temporal sclerosis:

pathogenesis, diagnosis, and treatment. Epilepsia 1990;318(suppl):55-66.

25 Dam AM. Epilepsy and neuron loss in the hippocampus.Epilepsia 1980;21:617-29.

26 Drake J, Hoffmann HJ, Kobayashi J, et al. Surgical man-agement of children with temporal lobe epilepsy andmass lesions. Neurosurgery 1987;21:792-7.

27 Fried I, Kim JH, Spencer DD. Hippocampal pathology inpatients with intractable seizures and temporal lobemasses. JfNeurosurg 1992;76:735-40.

28 Brooks BS, King DW, Gammal TE, et al. MR Imaging inpatients with intractable complex partial epilepticseizures. AJNRAm J Neuroradiol 1990;1 1:93-9.

29 Elster AD, Mirza W. MR imaging in chronic partialepilepsy: role of contrast enhancement. AJNR Am JfNeuroradiol 1991;12:165-70.

30 Engel J, Van Ness P, Rasmussen TB, et al. Outcome withrespect to epileptic seizures. In: Engel J, ed. Surgicaltreatment of the epilepsies. 2nd ed. New York: RavenPress, 1993:609-21.

31 Feindel W, Robitaille Y, Tampieri D, et al. Electro-encephalographic, magnetic resonance imaging andpathology in patients surgically treated for temporal lobeepilepsy. Can _J Neurol Sci 1991;18:577-9.

32 Jabbari B, Gunderson CH, Wippold F, et al. Magnetic res-onance imaging in partial complex epilepsy. Arch Neurol1986;43:869-72.

33 Kilpatrick CJ, Tress BM, O'Donnell Ch, et al. Magneticresonance imaging and late-onset epilepsy. Epilepsia1991;32:358-64.

34 Lencz T, McCarthy G, Bronen RA, et al. Quantitativemagnetic resonance imaging in temporal lobe epilepsy:relationship to neuropathology and neuropsychologicalfunction. Ann Neurol 1992;31:629-37.

35 Lesser RP, Modic MT, Weinstein MA, et al. Magnetic res-onance imaging (1:5 Tesla) in patients with intractablefocal seizures. Arch Neurol 1986;43:367-71.

36 McLachlan RS, Nicholson RL, Black S, et al. Nuclearmagnetic resonance imaging, a new approach to theinvestigation of refractory temporal lobe epilepsy.Epilepsia 1985;26:555-62.

37 Sharbrough F, Jack C, Cascino G. Presurgical quantitativemagnetic resonance based on hippocampal volume mea-surements in patients undergoing temporal lobectomiefor medically intractable seizures. Epilepsia 1989;30:675.

38 Sperling MR, Wilson G, Engel J, et al. Magnetic resonanceimaging in intractable partial epilepsy: correlativestudies. Ann Neurol 1986;20:57-62.

39 Grunwald F, Menzel C, Pavics L, et al. Ictal and interictalbrain SPECT imaging in epilepsy using technetium-99m-ECD. J Nucl Med 1994 (in press).

40 Wieser HG, Yasargil MG. Selective amygdalhippocampec-tomy as a surgical treatment of mesiobasal limbicepilepsy. SurgNeurol 1982;16:445-456.

41 Kleihues P, Burger PC, Scheithauer BW. Histologicaltyping of tumors of the central nervous system. Berlin:Springer, 1993.

42 Wolf HK, Wiestler OD. Surgical pathology of chronicepileptic seizure disorders. Brain Pathol 1993;3:371-80.

43 Luders H, Engel J, Munari C. General principles. In:Engel J, ed. Surgical treatment of the epilepsies. 2nd ed.New York: Raven Press 1993:137-53.

44 Heinz ER, Crain BJ, Radtke RA, et al. MR imaging inpatients with temporal lobe seizures: correlation ofresults with pathologic findings. AJNR Am J Neuroradiol1990;11:827-32.

45 Ryvlin P, Garcia-Larrea L, Philippon B, et al. High signalintensity on T2-weighted MRI correlates with hypo-perfusion in temporal lobe epilepsy. Epilepsia 1992;33:28-35.

46 Theodore WH, Dorwart R, Holmes M, et al. Neuro-imaging in refractory partial seizures: comparison ofPET, CT, and MRI. Neurology 1986;36:750-9.

47 Triulzi F, Franceschi M, Fazio F, Del Maschio A. Nonrefractory temporal lobe epilepsy: 1,5-T MR imaging.Radiology 1988;166:181-5.

48 Turner DA, Wyler AT. Temporal lobectomy for epilepsy:mesial temporal herniation as an operative and prognos-tic finding. Epilepsia 1981;22:623-9.

49 Holland BA, Kucharcyzk W, Brant-Zawadzki M, et al. MRimaging calcified intracranial lesions. Radiology 1985;157:353-6.

50 Latack JT, Abou-Khalil B, Siegel GJ, et al. Patients withpartial seizures: evaluation by MR, CT, and PET imag-ing. Radiology 1986;159:159-63.

51 Ormson MJ, Kispert DB, Scharbrough FW, et al. Crypticstructural lesions in refractory partial epilepsy: MRimaging and CT studies. Radiology 1986;160:215-9.

52 Schorner W, Meencke HJ, Felix R. Temporal-lobeepilepsy: comparation of CT and MR imaging. AJ3R AmJ7Neuroradiol 1987;149:1231-9.

53 Cendes F, Andermann F, Gloor P, et al. Atrophy of mesialstructures in patients with temporal lobe epilepsy: causeor consequence of repeated seizures? Ann Neurol 1993;34:795-801.

54 Jack CR, Sherbrough FW, Twomey CK, et al. Temporallobe seizures: lateralization with MR volume measure-ments of the hippocampal formation. Radiology 1990;175:423-9.

55 Duncan JS, Sager HJ. Seizure characteristics, pathologyand outcome after temporal lobectomy. Neurology 1987;37:405-9.

56 Elwes RDC, Dunn G, Binnie CD, Polhey CE. Outcomefollowing resective surgery for temporal lobe epilepsy:a prospective follow-up study of 102 consecutive cases.J Neurol Neurosurg Psychiatry 1991;54:949-52.

57 Olivier A: Relevance of removal of limbic structures insurgery for temporal lobe epilepsy. Can J Neurol Sci1991;18:628-35.

58 Bronen RA. Temporal lobectomy. In: Spencer SS,Spencer DD, eds. Surgery for epilepsy. Boston:Blackwell Scientific, 1991:103-49.

59 Ojemann GA. Surgical therapy for medically interactableepilepsy. J7Neurosurg 1987;66:489-99.

60 Penfield W, Jasper HH. Epilepsy and the functional anatomyof the human brain. Boston: Little-Brown, 1954.

61 Falconer M. Anterior temporal lobectomy for epilepsy. In:Logue V, ed. Neurosurgery, 2nd ed. Operative surgery. Vol14. London: Butterworths, 1971:142-9.

673 on D



j.com/

J Neurol N


ownloaded from


neurology, neurosurgery, surgical treatment epilepsy · zentner, hufnagel, wolf, ostertun, behrens,...

Documents