Pathology of temporal lobe epilepsy : An analysis of 100 consecutive surgical specimens from patients with medically refractory epilepsy.
The neuropathological features of temporal lobe epilepsy were studied utilising 100 consecutive surgical specimens from patients with medically refractory complex partial seizures. A wide spectrum of neuropathological changes was recorded in 98 specimens. Fifty-eight specimens showed features of Ammon's horn sclerosis. Diffuse accumulation of corpora amylacea were demonstrated in the resected temporal lobes from 54 patients. Six patients had neoplastic lesions of temporal lobe. One unique case of dysembryoplastic neuroepithelial tumour showed a melanotic component within the tumour. The neuropathological features were regarded as nonspecific in 31% of cases. Our results indicate that a majority of patients with medically intractable epilepsy of temporal lobe origin reveal significant neuropathological features. Careful documentation of the neuropathological features and its correlation with radiological, electrophysiological and pre- and post-surgical clinical features will help in predicting the seizure outcome after temporal lobectomy for medically refractory epilepsy.
Temporal lobectomy with amygdalo-hippocampectomy is being performed with increasing frequency in patients with medically intractable complex partial seizures of temporal lobe origin.,, In the past, neuropathologist's contribution to the study of epilepsy was essentially confined to evaluation of the lesions in the autopsy material. With the increasing number of surgical specimens available from patients with temporal lobe epilepsy, the role of neuropathologists in the overall management of epilepsy has assumed a greater importance in the recent past.
Our objective in this study was to evaluate the spectrum of neuropathological findings ascertained from a sizeable number of patients with temporal lobe epilepsy. A literature survey from the Indian subcontinent revealed no published reports which highlighted the neuropathological features in patients with medically refractory temporal lobe epilepsy. This communication, thus, will offer useful information about the neuropathological substrate of medically refractory temporal lobe epilepsy in our country and how it compares with the data from elsewhere.
From March 1995 to March 1998, over 130 temporal lobectomies for medically refractory epilepsy have been undertaken at the Sree Chitra Tirunal Institute for Medical Sciences and Technology, Trivandrum. The surgical specimens from the first 100 patients formed the material for this study. We excluded from this analysis patients who underwent temporal lobectomy for reasons other than medically refractory epilepsy.
All patients with hippocampal lesions observed on magnetic resonance imaging (MRI), underwent a tailored anterior temporal lobectomy with microsurgical excision of amygdala and `enbloc' dissection of hippocampus. In 31 cases, enbloc dissection was not feasible due to technical reasons during surgery and in such instances, the hippocampus was removed `piecemeal'. The excised temporal lobectomy specimen along with mesial temporal structures was fixed in 10% neutral buffered formalin. Following optimal fixation, the specimen was sliced in coronal plane and the hippocampus removed enbloc was processed in toto for routine histopathological studies. Specimens received in piecemeal were also processed in toto and the entire specimen was embedded in paraffin. Five micron thick paraffin sections were stained with haematoxylin and eosin, cresyl violet, Luxol fast blue (LFB), PAS and Bodian stain. Ten blocks of hippocampus from age-matched controls were collected at autopsy, from patients who died from non-neurological diseases. This was undertaken to correlate the specific neuropathological features in the hippocampus from patients with temporal lobe epilepsy and with that from age-matched controls. We defined Ammon's horn sclerosis as the loss of neuronal cell population of more than 30% in the CA1 sector of cornu ammonis (CA) when compared to the CA1 sector in the control hippocampus.
The neuropathological features of 100 surgical specimens from patients with medically refractory temporal lobe epilepsy are summarized in [Table I]. The 100 patients comprised 52 males and 48 females. Their age ranged from 3 to 48 years (mean 26 years). The mean duration of epilepsy prior to surgery was 16 years (range 2 to 42 years).
Ammon's horn sclerosis : Fifty-eight specimens showed features of Ammon's horn sclerosis. The loss of neuronal cell population in the CA1 sector [Figure 1] was accompanied by reactive glial fibrillary proliferation which appeared as a meshwork of finely fibrillated glial matrix. The reduction in the neuronal cell population was also evident in the granular cell layer of the dentate gyrus which was better delineated in the cresyl violet preparation [Figure 2]. Besides these findings, there was definite disarray in the arborisation of axonal fibres. The network of axonal fibres between the neurons was absent.
Nonspecific changes : Nonspecific findings were noted in 31% of cases. These changes were essentially seen in patients in whom hippocampus was removed in piecemeal. In the paraffin sections, microscopic anatomical integrity of hippocampus could not be defined with any degree of certainty. Focal gliosis, loss of neuronal cells, chromatolysis and pyknotic changes in the neurons, and focal demyelination were seen.
Corpora amylacea : A massive deposition of corpora amylacea was a notable feature in the PAS stain in 33 out of 58 specimens with Ammon's horn sclerosis and in 21 out of 31 cases showing non-specific changes. In the specimens with Ammon's horn sclerosis, corpora amylacea were preferentially accumulated in the pyramidal cell layer (CA1 and CA2 sectors). In the control specimens, corpora amylacea were more often located in the subpial and subependymal regions and were fewer in number.
Neoplastic lesions : Six patients had various neoplastic lesions in the temporal lobe. A histopathological diagnosis of dysembryoplastic neuroepithelial tumour was made in one case based on the characteristic glioneural elements. This tumour at some microscopic foci showed melanotic differentiation [Figure 4]. Two of our cases showed histopathological features of ganglioglioma and they could easily be distinguished from astrocytoma by the presence of neoplastic neuronal cells along with the fibrillary astrocytic component. One of our cases exhibited classical features of oligodendroglioma.
In all chronic seizure disorders of temporal lobe origin, the neuropathological evaluation of surgical specimens should address the following four issues: 1) describe the composition of the specimen, 2) delineate the presence or absence of a focal lesion, 3) characterize the presence or absence of Ammon's horn sclerosis, and 4) ensure that an adequate removal of neoplastic lesions have been accomplished.
Neuropathological evaluation in patients with temporal lobe epilepsy to a great extent depends upon structural integrity of excised specimens of hippocampus. In our data, neuropathologic features like Ammon's horn sclerosis could be delineated well in the `enbloc' specimens rather than in the piecemeal specimens of the hippocampus. The microscopic anatomical integrity of hippocampus in the piecemeal specimens could not be defined and the changes like neuronal loss, gliosis could not be quantitated. Usage of ultrasonic aspirator during surgery can also alter the anatomic integrity of the specimen.
Ammon's horn sclerosis : 58% of the specimens in the present study had features of Ammon's horn sclerosis. Many of our patients with piecemeal specimens had forme fruste findings of Ammon's horn sclerosis, which we have categorized as nonspecific because of insufficient histopathological data. Hence the actual proportion of specimens with Ammon's horn sclerosis is likely to be much more than 58%. In three published reports of more than one hundred cases of refractory temporal lobe epilepsy, the frequency of Ammon's horn sclerosis has ranged from 20.6 to 71.8%.,, Sommer was the first to describe the pathological changes in the Ammon's horn of patients with chronic epilepsy, predominantly characterised by the loss of pyramidal cells in the CA1 sector (often called the Sommer sector). Margerison and Corsellis, in an autopsy study of temporal lobe epilepsy, observed two major types of hippocampal sclerosis : `classical Ammon's horn sclerosis' (with neuronal loss and gliosis in the CA1 sector and the dentate gyrus) and `end folium sclerosis' involving the end folium or CA3 and CA4 sectors. Falconer and Taylor observed that with increasing severity of the pathology, other parts of mesial temporal lobe also demonstrated a varying degree of cell loss and coined the term `mesial temporal sclerosis' to describe the full spectrum of pathological changes involving the mesial temporal structures. Most of the neuropathological features of mesial temporal sclerosis, however, are concentrated in the Ammon's horn.
We relied on the changes involving Ammon's horn to delineate the hippocampal pathology in patients with chronic temporal lobe epilepsy. In 31% of the specimens from our patients in whom the hippocampus was removed in piecemeal, the microscopic anatomy and pathology of the complete mesial temporal structures could not be assessed with any degree of certainity. Therefore, we prefer to use the term Ammon's horn sclerosis to designate the histopathological entity of mesial temporal sclerosis in this article.
Although it is well known that mesial temporal scleros is the most frequent pathological marker of temporal lobe epilepsy, there is a great deal of controversy regarding whether this lesion is the primary cause of temporal lobe epilepsy, the result of repetitive seizure activity or hypoxic injury, or both. Several recent reviews have dealt with this subject in detail,,, and is beyond the scope of this article.
Corpora amylacea : In our data 54 out of 100 specimens with temporal lobe epilepsy contained a significant number of corpora amylacea in the paraffin sections of the hippocampus which were scattered diffusely in the CAI sector and in the dentate gyrus. Corpora amylacea are well known to occur in a number of neurodegenerative diseases and also during the course of normal ageing and have been regarded as a nonspecific finding. The mean age of our patient group was 26 years (range 3 to 48 years). Furthermore, the distribution of corpora amylacea observed in the specimens with Ammon's horn sclerosis in this study is quite distinctive and can be easily differentiated from those generally regarded as age related. In the paraffin sections of hippocampus from the controls, the number of corpora amylacea were significantly less and they were limited to the subpial and subependymal regions. None of the control hippocampus showed a significant number of corpora amylacea in the pyramidal cell layer CA1 and CA2 sectors as observed in patients with Ammon's horn sclerosis. From a diagnostic standpoint, the presence of large number of corpora amylacea in the pyramidal layer, endfolium, and the dentate gyrus provides a readily identifiable marker for Ammon's horn sclerosis. Their occurrence is more helpful in cases where assessment of neuronal loss and gliosis becomes difficult due to technical problems such as tissue fragmentation and piecemeal removal of the hippocampus. The corpora amylacea appear more prominently in the PAS stain. Since routine haematoxylin and eosin stained slides greatly underestimate their number, we suggest that LFB-PAS combined stain should routinely be undertaken during processing of temporal lobectomy specimens. LFB - PAS stain demonstrates corpora amylacea better because of the bright red colour it imparts and also because of a better appreciation of the pyramidal cell layer in the sections.
The presence of corpora amylacea in temporal lobe epilepsy has been described only recently. Loiseau et al14 described numerous polyglucosan bodies in the subcortical white matter of a female subject with intractable partial seizures. In the remainder of their 65 temporal lobectomy specimens, only additional 4 showed polyglucosan bodies. In contrast, Mackenzie15 found excess number of corpora amylacea in 15 out of 40 temporal lobectomy specimens. Chung et al16 observed increased number of corpora amylacea in 58% of temporal lobectomy specimens.
Neoplastic and other structural lesions : An intracranial mass lesion may be surgically verified in 10-20% of patients with chronic medically refractory partial epilepsy., In patients with mass lesions, neoplastic lesions outnumber non-neoplastic lesions. In a report of 216 consecutive surgical specimens from patients with chronic medically refractory temporal lobe epilepsy, there were 75 (34.7%) tumours and 51 (23.4%) non-neoplastic focal lesions such as heterotopias, vascular malformations and non-specific gliotic lesions. Slow growing, low-grade and well differentiated gliomas are the most epileptogenic lesions.,, Among 39 patients with chronic intractable epilepsy as the only symptom of a primary brain tumour, the pathology was ganglioglioma in 15, low-grade astrocytoma in 13, oligodendroglioma in 4, dysembryoplastic neuroepithelial tumour in 2 and other glial neoplasms in 5 cases. Thus, the histopathological profile of neoplasms associated with chronic partial epilepsy is quite different from the usual pattern of primary brain tumours. In our data, one patient had features of dysembryoplastic neuroepithelial tumour and this tumour showed characteristic melanotic differentiation. In a series of 39 cases of dysembryoplastic neuroepithelial tumours reported by Daumas-Duport et al melanotic differentiation was not recorded.
Relation between postoperative seizure outcome and histopathology : There has been considerable interest in recent years about the relation between the outcome of epilepsy surgery and structural lesion in the temporal lobe. In an actuarial analysis of postoperative seizure outcome in 135 patients, Berkovic et al observed a 2 year seizure-free state in 80% of patients with foreign tissue lesion, 62% of those with hippocampal sclerosis, and 36% of those with no lesion. Histopathologically demonstrated lesion was a strong predictor of a favourable seizure outcome among 184 patients who underwent anterior temporal lobectomy at the Mayo Clinic, for medically refractory epilepsy. Thus, the histopathological evaluation in patients with temporal lobe epilepsy provides prognostically relevant information. In addition, the neuropathological examination documents the extent and adequacy of the surgical resection. The seizure remission is better in those with a complete tumour removal when compared with those with postoperative residual lesions.,
In the present study, 80 out of 100 patients have completed a one year postoperative follow-up. 46 (57%) are completely free of seizures. Sixty percent of patients with MTS, 67% of patients with neoplasms and 50% of patients with nonspecific lesions had a seizure-free outcome. A detailed analysis of their long-term outcome and the predictors of a favourable outcome will be communicated later.