The author: Professor Yasser Metwally
INTRODUCTION
December 15, 2008 — Histopathologic causes of temporal lobe epilepsy include
|
-
Mesial temporal sclerosis
Mesial temporal sclerosis consists of cell loss and astrogliosis in the mesial temporal cortex, the hippocampal formation, amygdala, parahippocampal gyrus, and entorhinal cortex. These changes have been best described in the hippocampus, partly due to its severe involvement and the lamellar pattern of hippocampal organization that lends itself to histopathologic study. Two forms of hippocampal cell loss have been identified in mesial temporal sclerosis. Classic sclerosis, also known as Ammon’s horn sclerosis, is the more frequent form. This consists of marked loss of the pyramidal cells in CA1, CA3, and the dentate hilus, with sparing of pyramidal cells in the CA2 sector. The second form is denoted as end folium sclerosis, which consists primarily of cell loss and astroglial proliferation in the end folium with relative sparing of the other sectors. Autopsy studies have demonstrated that mesial temporal sclerosis is present bilaterally, in up to 80% of cases. However, it is usually asymmetric in that one side is more severely involved than the other; the more severely involved of the two hippocampi typically denotes the site of origin of a patient’s seizures.
 |
Figure 1. Amygdala; 2, head hippocampus; 3,subiculum; 4, temporal horn; 5, collateral sulcus; 6, parahippocampal gyrus;7 fornix; 8, mammillary bodies; 9, 3rd ventricle, 10 occipitotemporal gyrus; 11, inferior temporal gyrus;12 entorhinal gyrus (Click to enlarge figure)
The MR image counterparts to these two pathologic abnormalities are atrophy and signal change. Hippocampal atrophy is best identified on Tl-weighted images obtained coronally or, ideally, perpendicular to the principal axis of the hippocampal formation, which is variably canted downward from posterior to anterior. The identification of atrophy by MR imaging corresponds to cell loss identified in histologic specimens.
 |
Figure 2. Atrophy of the left hippocampus with dilated left temporal horn (Click to enlarge figure)
The other principal finding is a signal intensity change consistent with increased tissue-free water resulting in decreased signal intensity on Tl-weighted images and an increased signal intensity on Tl- weighted images. It is logical to assume that the abnormality in signal intensity is a function of astroglial proliferation. Several other findings on MR images have been helpful in identifying the temporal lobe predominantly involved in mesial temporal sclerosis; these include (1) loss of normal internal architecture of the involved hippocampus; (2) unilateral atrophy of the mammillary body; (3) unilateral atrophy of the columns of the fornix; (4) unilateral atrophy of the amygdala; and (5) unilateral atrophy of white matter bundle in the parahippocampal gyrus.
 |
Figure 3. Hippocampal hyperintensity (Click to enlarge figure)
 |
Figure 4. Normal hippocampus [left image], hippocampal atrophy with dilated temporal horn [middle image] and hippocampal hyperintensity [right image]i (Click to enlarge figure)
Although all of these findings may occur in cases of mesial temporal sclerosis, the author believes that the small, bright hippocampus is the most reliable . Both of these findings are usually present in an individual patient. In some patients with mesial temporal sclerosis, however, the hippocampus may appear to be normal-sized but of increased signal, or atrophic without an obvious signal abnormality. Studies of the accuracy of visual perception have demonstrated that predominantly unilateral mesial temporal sclerosis can be identified with an accuracy of about 90% by knowledgeable readers. Although the unilateral dilatation of the temporal horn has been suggested as a useful marker of mesial temporal sclerosis, the author regards this finding as unreliable. Whereas this exists in cases of severe hippocampal atrophy because of mesial temporal sclerosis, it is not a reliable indicator of mesial temporal sclerosis because it also occurs as a common normal variant.
As mentioned, mesial temporal sclerosis is found bilaterally in approximately 80% of autopsy cases. However, the goal of imaging and the assumption underlying treatment by temporal lobectomy would indicate that in most cases of mesial temporal sclerosis, only one of the temporal lobes actually produces seizures. This apparent discrepancy is best explained by regarding the entire spectrum of mesial temporal sclerosis in four categories: (1) category 1, one hippocampus is entirely normal and the other is abnormal; (2) category II, one hippocampus is slightly abnormal and the other severely abnormal; (3) category III, both hippocampi are abnormal to an equal degree; and (4) category IV, both hippocampi are normal. Clinical experience to date indicates that in category II patients, the more severely involved hippocampus typically represents the site of seizure onset. It is the distinction between categories I/II and III/IV that appears to be the most critical, both in terms of surgical outcome and imaging identification of the substrate of epilepsy.
Patients in categories I and 11 both respond well following removal of the abnormal temporal lobe. Furthermore, visual identification of the more abnormal hippocampus is fairly straightforward when there is a significant side-to-side discrepancy in volume and signal intensity. On the basis of imaging criteria, however, it is impossible to identify the more involved hippocampus in categories III and IV because, by definition, the two hippocampi are either symmetrically abnormal or symmetrically normal. These two categories appear to have a similarly mediocre response to temporal lobectomy: fewer than 50% of patients are seizure free postoperatively.
-
Other causes of temporal lobe epilepsy
Approximately 25% to 30% of patients with complex partial seizures originating from the temporal lobes have neoplasms, vascular malformations, cortical dysplasias, or posttraumatic lesions . Neoplasms can be differentiated from medial temporal sclerosis because they are usually associated with mass effect, and when they are not hippocampal in origin, they displace the hippocampus. In adults, neoplasms tend to be low-grade astrocytomas and in children neoplasms tend to be oligodendrogliomas.
Cavernous angiomas are the most common vascular malformation associated with temporal lobe seizures. This lesion is thought to occur more frequently in adults. Cavernous angiomas are characterized by mixed signal intensity: The increase in TI signal is related to recent hemorrhages, the leakage of blood products into the adjacent gliotic tissue, and slowly flowing blood within abnormal vessels; the decrease in T2 signal is the result of intracellular hemosiderin caused by old hemorrhages. These lesions frequently have a periphery of low signal intensity. Although most cavernous angiomas do not enhance with contrast material, a minority of them may enhance. Abnormal vessels are not visible either on MR angiography or conventional cerebral angiography. Arteriovenous malformations have enlarged arteries and veins, which are invariably visible on MR scans. The appearance of the nidus of the malformation is influenced by the presence of previous hemorrhagic products and gliotic changes. Venous angiomas are characterized by the presence of many small flowing veins that converge into a large draining vein, which usually enhances with contrast material. There is an association between venous and cavernous angiomas. Venous angiomas are almost always asymptomatic.
 |
Figure 5. Left hippocampal cavernous angioma,the high signal intensity focus in the T1 image [left] relates to the increased methemoglobin content while the T2 hypointensity [right image] is probably due to deoxyhaemoglobin or haemosiderin (Click to enlarge figure)
References
-
Metwally, MYM: Textbook of neuroimaging, A CD-ROM publication, (Metwally, MYM editor) WEB-CD agency for electronic publication, version 10.1a January 2009 [Click to have a look at the home page]
