Temporal Lobe Epilepsy (TLE) and Neuroimaging
Temporal Lobe Epilepsy (TLE) comprises 30% of all epilepsies and is the most common cause of focal seizures in both adults and children, accounting for 60% of all cases of focal epilepsy evaluated in specialized centers. Nearly 30% of these patients will develop drug resistance and, of these, 30% will have negative MRIs with the routine epilepsy protocol. Detection of epileptogenic lesions is crucial in both the initial diagnosis and the presurgical assessment, and Clinical Neuroradiology plays a fundamental role in managing these patients.
The most common cause of TLE is mesial temporal sclerosis (MTS), a syndrome which displays signs of hippocampal sclerosis (HS) on MRI, accompanied by a characteristic electroclinical profile. Alternative causes of TLE include other focal lesions located in the temporal lobes, some of them undetectable with current technology (cryptogenic TLEs); there are also familiar forms associated with various genetic mutations.
Patients with refractory temporal seizures are candidates for surgical treatment. The detection of a structural lesion on MRI is related to poorer pharmacological control but better surgical results. However, when MRI is negative, other more expensive and invasive investigations must be considered. Therefore, studying these cases always requires a specific protocol and, frequently, a personalized diagnostic strategy, so the appropriate use of the different radiological techniques is essential. Structural MRI is the principal radiological technique in both diagnostic and presurgical settings, although functional imaging is required when the MRI is inconclusive. Findings from imaging should always be interpreted considering the EEG data, and patients with refractory seizures should be managed by multidisciplinary teams in specialized units.
KeywordsTemporal lobe epilepsy Hippocampal sclerosis
List of Abbreviations
Focal cortical dysplasia
Initial precipitating injury
Lateral temporal lobe epilepsy
Malformation of cortical development
Mesial temporal lobe epilepsy
Mesial temporal sclerosis
Subtraction ictal SPECT co-registered to MRI
Temporal intermittent rhythmic delta activity
Temporal lobe epilepsy
Temporal Lobe Epilepsy
Entity: Definition and Types
Temporal lobe epilepsy (TLE) encompasses a heterogeneous group of diseases that have the anatomical location of the epileptogenic focus in common, and similar symptoms characterized by auras and complex partial seizures with automatisms.
Possible causes of TLE
Mesial temporal sclerosis (MTS)
Other structural lesions
Malformation of cortical developmental (MCD)
Scar (trauma, infarct)
The International League Against Epilepsy (ILAE) recognizes two types of TLE, with different clinical and pathophysiological characteristics: Mesial temporal lobe epilepsy (MTLE) and lateral temporal lobe epilepsy (LTLE). The most frequent is MTLE (the main cause of focal seizures in adults), in which the seizures could originate in the temporal medial structures, such as the hippocampus, the amygdala, and the parahippocampal gyrus. Most cases correspond to mesial temporal sclerosis (MTS) which includes hippocampal sclerosis (HS), although there is frequent extension to other limbic structures, especially the amygdala. Alternative causes of MTLE include tumors localized in the temporal mesial structures, malformation of cortical development (MCD), or other rare entities such autoimmune encephalitis or temporal encephalocele. In some cases, HS can coexist with another epileptogenic lesion (dual lesion).
LTLE, which originates in the neocortex of the temporal lobe, is less common and the causes are almost always secondary to a structural lesion, mainly tumors, malformations of cortical development (MCD) which occasionally are difficult to visualize in the temporal lobes, vascular malformations, and gliotic lesions related to previous trauma and brain infarcts. These entities have been discussed in the Neocortical Epilepsy chapter.
Patients with TLE where MRI scans reveal no significant alterations, known as “non-lesional,” constitute a heterogeneous group, which includes genetic cases as well as those with lesions that are not visible with current technology, most of them focal cortical dysplasias (FCD). Management of these patients may be different from those with classical MTS; therefore, accurate classification is essential requiring a combination of electroclinical and imaging data.
Clinical Scenario and Indication for Imaging
The natural history, semiology of the seizures, and EEG findings are characteristic and can often be differentiated from other types of focal epilepsy.
MTLE usually starts in the first decade of life, with occasional seizures that are usually controlled. From the second decade onwards, most patients will develop drug-resistant epilepsy, with partial seizures characterized by their autonomic and psychological components, with frequent epigastric and déjà vu auras with preserved consciousness, or complex partial seizures, with impairment of awareness or responsiveness followed by oral and manual automatisms. Generalization of the seizures is rare. As the disease evolves, a gradual loss of memory function can be seen, depending on the disease’s relationship with the dominant hemisphere.
Routine EEG (20–30 min.) is usually normal, although a longer recording may detect temporal epileptiform activity in most of the cases. Slow temporal activity, when it appears in the form of delta wave trains (temporal intermittent rhythmic delta activity (TIRDA)), may also have a localizing value. The ictal EEG shows the ipsilateral onset of the focus in 70% of cases, although it may appear contralateral to the affected side in 13%. Prolonged slow activity can be seen in the postcritical phase and be helpful for localization.
In LTLE auditory auras, ictal aphasia, and other neocortical auras will be more common. In comparison with medial temporal seizures, neocortical seizures may progress with early contralateral dystonia and/or later oral/manual automatisms. Ictal patterns on EEG may be similar to those of patients with MTLE (O’Brien et al. 1996). Neocortical origin must be suspected when the ictal pattern consists of a rhythmic peak followed by rapid activity and evolves into theta activity, as this pattern is more common in FCDs. Similarly, late ictal patterns (i.e., non-lateralized theta activity which develops late in a temporal region) are more frequent in patients with lateral temporal epilepsy.
If there are doubts about lateralization, or if a neocortical origin is suspected, depth electrodes should be implanted, allowing depiction of periodical hypersynchronous discharges during the onset of the seizure (including during the aura) evolving into fast low-voltage activity which rapidly spreads to other regions.
The response to AEDs is usually poor and is clearly inferior to what is seen in other focal epilepsies, usually not exceeding 30%. In these cases, presurgical assessment should not be delayed, as early resection has better results and improves psychosocial status (Bernhardt et al. 2009). Long-term control of seizures usually reaches 60%, being crucial the response in the first 2 years, because up to 90% of patients who have responded within this period will have no seizures in the following 10 years (Elsharkawy et al. 2009).
Imaging Technique and Recommended Protocol
MRI has been one of the most significant breakthroughs in the management of epilepsy patients, allowing to identify the epileptogenic substrate more effectively. MRI provides relevant information at all stages of the management of these patients. MRI is one of the tools used for syndromic classification in the initial assessment, and diagnosis of MTS depends on positive signs in the MRI. In the case of refractory seizures, MRI is decisive when selecting surgical candidates and contributes to treatment planning.
The most important role of MRI is to detect structural lesions, which is particularly important in epilepsy, which frequently involves subtle lesions which are difficult to detect and characterise; this fact has become evident as technical advances have been made in both hardware and software. For example, the sensitivity of MRI in MTLE was less than 50% in the 1980s, but the application of high-resolution sequences with rapid techniques (T2-FSE, 3DT1), and then FLAIR (fluid-attenuated inversion recovery) sequences, has led to a significant increase in sensitivity (Taillibert et al. 1999) which had already exceeded 90% by the 1990s. Related to the magnetic fields, the difference in results between 0.5 T and 1.5 T magnets is also significant, and studying epilepsy patients at field strength below 1.5 T is no longer acceptable.
Structural MR Protocol Valid for Epileptic Patients
Routine brain MRI protocols are inappropriate for the study of epileptic patients, as many lesions will be unnoticed. Therefore, it is crucial to use a dedicated epilepsy protocol and an experienced neuroradiologist to obtain the highest accuracy in the radiological diagnosis. (Von Oertzen et al. 2002).
2D-T2 sequences have the advantage of attaining higher in-plane resolution than 3D-T2 sequences allowing better assessment of the internal structure of the hippocampus. It is recommended to add T2* or magnetic susceptibility sequences to detect minor calcification or hemorrhage.
MR Protocol for temporal lobe epilepsy
3D-T1 gradient Echo (MPRAGE, SPGR, FFE) with isotropic voxel (1 mm)
Multiplanar reformation, volumetry
3D-FLAIR with isotropic voxel (1 mm)
Axial T2FSE: 2–3 mm slice thickness
Cor T2FSE: <3 mm thickness
High resolution (512 matrix if possible)
Axial EPI-T2GE <3 mm
3D-SWI: More sensitive than T2*
Coronal T1-IR: Better grey-white matter contrast
DWI: Suspected acute lesion
Post-contrast T1: Suspected tumor, inflammation
3D-T2: Suspected encephalocele
This protocol would be sufficed to detect the vast majority of the cases with MTS and many MCDs, although questionable cases should be assessed individually in an epilepsy unit, with the possibility of expanding the study with multi-array coils, 3 T, or by using advanced techniques such as volumetry, diffusion, perfusion, relaxometry or spectroscopy.
Quantification: Volumetry and Relaxometry
When signs of atrophy or signal alteration are not obvious, quantification of hippocampal volume and calculation of the T2 relaxation time may be useful (preferably on 3 T) (Coan et al. 2014; Mueller et al. 2007). Atrophy could be demonstrated in 13% of patients with MTLE and negative MRI, and changes in T2 in 19%.
Volumetry requires high-quality 3DT1 sequences, but actual volumes will depend on method and operator experience; as a result, normative values should be established for each centre. We must also account for changes due to age, as well as the normal asymmetry of the hippocampi (a ratio of 0.96 with left side usually being smaller). A difference of more than 10% (the standard deviation in healthy controls is 5–9%) is regarded as significant, and it is also recognized that visual analysis by an expert can detect changes from 15% upwards, which explains the low percentage of sensitivity gain in volumetry compared to visual analysis. However, volumetry is very reliable for longitudinal studies(Bernhardt et al. 2013).
Calculation of T2 relaxation time is more robust than volumetry, as it is less dependent on acquisition and of measurement method and has less physiological variability. The normal value of T2 relaxation is considered 100 ms with an SD of 2–3 ms. while in HS, the T2 relaxation is increased to around 120 ms. Some studies relate the loss of volume to the number of seizures and to drug resistance, while the changes in T2 have been linked more to prolonged seizures, regardless of their frequency.
Advanced MR Techniques
Nuclear Medicine Techniques
PET ictal SPECT, and SISCOM are indicated in TLE when (3 T) MRI scans are negative, questionable or there is a disagreement between clinical and imaging finding, improving the detection of the epileptogenic focus by showing hypometabolism and hypoperfusion in the temporal lobe affected. They can also circumvent the use of deep electrodes when include very typical electroclinical data, contribute to planning depth electrodes if necessary and also helping in determine the extension of the surgical resection.
Mesial Temporal Sclerosis (MTS)
The ILAE considers MTS in the syndromic classification of epilepsy as a clinical-radiological entity in which an electroclinical pattern characteristic of MTLE coexists with radiological signs of ipsilateral HS on the structural MRI (Scheffer et al. 2017). Although the term MTS indicates that other limbic structures are often involved, the necessary and suffice condition for presurgical diagnosis of MTS is that HS on MRI should be demonstrated.
MTS is the cause of the 70% of surgical procedures for TLE, although the real incidence of MTS is not known due to the inherent bias of the surgical series, as there are few cases of hippocampus resection with non-lesional MRI. The age of onset is usually between 4 and 16 years old, with no gender predilection, although cases may occur outside this age range. It is often associated with a childhood history of febrile seizures (especially complicated ones) or with an initial precipitating insult (IPI): trauma, infection, hypoxic-ischemic encephalopathy, and status epilepticus. A family history of epileptic seizures, especially febrile seizures may be present. Most familial cases, with autosomal dominant inheritance but incomplete penetrance, show signs of HS on MRI; EEG pattern is similar as to what is found in sporadic MTS, although in the autosomal dominant inherence, hippocampal atrophy without signal alteration and bilateral hippocampi atrophy are more commonly found.
Anatomy and Histopathology
The hippocampus, in the anteroposterior orientation, can be divided into head, body and tail, and various fields can be histologically identified in the coronal plane: inferomedial CA1 (adjacent to the subiculum), followed by CA2 (located above), and CA3 and CA4 (centrally located, also referred to as the endfolium). The DG is also located centrally, surrounding CA4. Pyramidal neurons and stellate cells predominate in the hippocampus, but the main cells of the DG are granular cells (mossy fibers).
HS ILAE type 1: Considered the classical form of HS and most common (60–80% of operated cases). Damage is generalized, although CA1 is most severely affected, with more than 80% neuronal loss. This type is more clearly associated with a history of initial precipitating injury before the 5 years of age, as well as an earlier onset, and usually has a better postoperative prognosis.
HS ILAE type 2: Less common form (10%) characterized by a significant involvement of CA1 (similar to the classical form) which suffers severe neuronal loss and prominent astrogliosis, with other fields less affected.
HS ILAE type 3: Rarest form (around 5%), although it is frequently associated with dual lesions and has been associated with limbic encephalitis. In this type, CA4 and DG are mainly affected, and because of that, it is also called endfolium sclerosis.
- Gliosis only-no HS: Up to 20% of surgical cases of MTS and autopsies fail to demonstrate significant neuronal loss, although reactive astrogliosis consistent with the epileptogenic focus can be observed. The ILAE does not classify these cases as HS, although astrogliosis itself may play an epileptogenic role.
A relationship between the number of seizures and the degree of hippocampal atrophy has been observed in patients with MTS. The different degrees of damage within the hippocampal fields suggest a selective vulnerability, possibly mediated by an increase in intracellular calcium, due to the glutamatergic excitatory predominance over the GABAergic inhibition caused by the initial precipitating insult (Morimoto et al. 2004). The cells in each field will have different responses to the increased calcium, resulting in more or less cell death. The subiculum is not usually affected in HS but probably plays an important role in spreading and amplifying the excitatory activity of the hippocampus. Cell death also activates the microglia, which will increase the astrogliosis. Cell death and epileptiform activity can both cause the release of neurotrophic factors which stimulate the mossy fibers within the DG, creating aberrant and recurrent circuits which facilitate the epileptogenesis, and contributing to create a vicious circle which eventually leads to the development of HS (Bernardino et al. 2005).
Other Causes of Temporal Lobe Epilepsy
There are other causes that can course with TLE, some of them, such tumors and MCD, are revised in other chapters of the book, (Cross-reference with chapters “Long-Term Epilepsy-Associated Tumours” and “Neocortical Epilepsy”) but there are other entities such cavernomas and temporal encephaloceles that can be easily missed if we do not scrutinize all the temporal mesial structures.
Image Interpretation and Structured Reporting
Interpretation checklist in MTS
Main findings: Hippocampal sclerosis
Reduced size: Body, head (loss of digitations)
Signal changes: HypoT1/HyperT2-FLAIR
Internal structure: Blurring of layers and internal architecture
Enlarged temporal horn
Decreased size of
Signal changes: Temporal pole WM
Cortical ribbon blurring (FCD?)
Hippocampal atrophy is usually more pronounced in the anterior part of the hippocampus and can be seen as flattening of digitations in the hippocampal head.
Treatment and Management
Surgical management in MTLE
MRI: Structural lesion
Typical HS: Amygdalo-hippocampectomy (AHE)
Dual lesion: Temporal lobectomy
Encephalocele: Focal resection
MRI negative: Perform 3 T MRI
Typical HS: Amygdalo-hippocampectomy (AHE)
Negative/atypical: PET/SISCOM + deep electrodes
Resection based on multidisciplinar decision
If the structural study is inconclusive, functional imaging should be used, with either PET or SISCOM. In many cases, when functional imaging shows a clear, unilateral lesion, consistent with electroclinical MTS data, surgery may be advisable without the need of other examinations.
In non-lesional cases, the surgical decision will be based on lateralizing the focus by functional imaging and invasive depth electrodes. Therefore, the potential for controlling refractory seizures with surgical treatment it highly depends on obtaining and interpreting the various imaging techniques and electroclinical data. On the one hand, when the hippocampus is normal, there is a greater risk of memory loss, especially in the dominant hemisphere. This means that hemispheric dominance must be determined, and functional MRI is a good alternative to the Wada test. (Benke et al. 2006). A resection which respects the hippocampus should be considered, especially in the dominant hemisphere if high resolution MRI is clearly negative. Stereo EEG is particularly useful in these cases, as it allows electrodes to be placed in the hippocampus to verify whether the seizures begin there.
Cases of TLE with negative MRI usually have poorer surgical results, and the pathological anatomy findings are often far from specific. Nowadays, there is a trend towards considering cases of TLE with negative MRI as being different from MTLE syndrome(Mueller et al. 2007), not just because of the different prognosis but because they frequently are associated with neocortical alterations (Hammen et al. 2003; Cohen-Gadol et al. 2006). Therefore, it would be worthwhile establishing with certainty that the MRI does not show any signs of hippocampal pathology in order to classify the syndrome and the therapeutic approach. It is also important to consider that dual pathology can occur in up to 15% of adult cases and up to 67% of pediatric cases, being the FCD generally of type I, the most common dual pathology associated to HS (Bocti et al. 2003). This points out again the importance to examine patients on a 3 T scan, especially in children, if the clinical signs are not typical or if the 1.5 T MRI raises doubts of dual pathology.
When the electroclinical data do not clearly reveal a temporal origin, studies with invasive electrodes should be added, as the origin can be extratemporal (pseudotemporal epilepsy) in some cases; or extend beyond the temporal lobe (temporal plus epilepsy). This should be suspected if the functional disruption seen on PET or SISCOM goes beyond the temporal lobe, but it must be confirmed by means of deep electrodes, as the metabolic or perfusion changes may be secondary to the aberrant functional state produced by the epileptogenic focus, which would function as an abnormal network, as is happens in the epileptic encephalopathy. Although cases with typical manifestations of MTE and negative MR findings can have surgical results similar to cases with positive MRI, atypical cases and temporal plus epilepsy have a worse response. Good prognostic factors are the presence of epileptiform activity confined to the temporal lobe with consistent PET-FDG and/or SISCOM, and a short duration of the epilepsy.
Sample Report 1
A 40-years-old female without no familiar history of epilepsy, was evaluated in the epilepsy unit. She had past medial record of febrile seizures at 9 months of age and onset of partial complex seizures at 15 years of age. Seizures begin with a “strange sensation” and/or déjà vu, with posterior disconnection and oral automatisms, sometimes with unintelligible language. Frequency of seizures is variable, five in the last month but may have few in a day. Working in a call center, describe some memory complaints. Treated with different AEDs, now is taking Lacosamide, Lamotrigine, and Perampanel without control of seizures.
Neurophysiological examination shows better performance of verbal than nonverbal tasks, suggesting disfunction in nondominant hemisphere. 5-days Video-EEG shows ictal EEG with theta activity in anterior temporal lobe. Interictal EEG shows right temporal spikes.
Purpose of Study
To identify a structural lesion in right temporal lobe, mainly HS. Whereas electroclinical data are highly suspicious of right MTE, if MRI findings are positive, surgery will be proposed, with good prognostic expectations.
No significant abnormalities were seen in neocortex.
Gyral pattern was normal.
There was marked decreased size of the right hippocampus with loss of head digitations.
Right hippocampal signal is hyperintense on T2 and FLAIR (more pronounced).
In coronal T2, blurring of hippocampal layers.
There were some hyperintensity on T2/FLAIR in the right amygdala.
No abnormalities in fornix and mammillary body.
Size and signal appearances of temporal pole are normal.
There were no abnormal findings in left temporal lobe.
MRI findings are characteristic for MTS, and concordant with electroclinical data.
As patient was right handed, and all data suggest a nondominant hemisphere lesion, amygdalo-hippocampectomy was performed without any other study. Neuropathological report: Diffuse cell loss, although CA1 is most severely affected (ILAE type I HS). Patient is an Engel’s I, 2 years after surgery.
Sample Report 2
Right handed 56-year-old male with refractory seizures visited for presurgical evaluation. Patient had a previous 1.5 T MR with nonspecific white matter and basal ganglia lesions, likely related to microvascular disease. With no other previous medical record, patient has a history of 20 years of partial complex seizures, controlled with AEDs until 2 years ago. Seizures begin with epigastric aura, with subsequent disconnection, and amnesia after recovery. There were some oral automatisms. Frequency of seizures is about three every month but is increasing in intensity with more recovery time, and in the last 2 years lack of clinical control despite multiple AEDs.
Neuropsychological evaluation exhibits data of nondominant hemisphere disfunction, with decrease of nonverbal memory. Interictal EEG show bilateral temporal spikes with 55% right predominance. There is also a persistent frontotemporal, bilateral spike. Video-EEG with reduction of AEDs displayed a tonic-clonic seizure, preceded by masticatory movements and seconds after, cephalic version to the left. Simultaneous EEG show focal theta activity in anterior temporal region during 40 seconds with secondary generalization.
Purpose of MRI Study
Electroclinical data suggest a temporal origin, probably right but lateralization is not clear. If MRI shows definitive findings of MTE, surgery may be indicated. FDG-PET may be also considered.
No significant abnormalities were seen in neocortex.
Gyral pattern was normal.
Both hippocampi have normal size and internal architecture is preserved.
Both hippocampi return normal signal in T2 and FLAIR.
The hippocampal layers are well seen on FLAIR and T2.
Normal size and signal of both amygdala.
No abnormalities in fornix and mammillary body.
Size and signal appearances of temporal pole are normal.
There are some confluent foci of increased T2/FLAIR signal within the white matter and basal ganglia, likely related to small vessel disease.
The impression of the report was negative for temporal epileptogenic lesions. Non-specific white matter and basal ganglia hyperintensities.
Non-lesional temporal epilepsy, with probable right origin, but some findings point also to the frontal lobe. There are not enough data to proceed to surgery. Therefore, nuclear medicine neuroimage studies are indicated together with deep electrodes evaluation, to localize the epileptogenic zone.
The results of FDG-PET and deep electrodes confirm that EZ is in right temporal lobe, probably in the anterior part.
Small encephaloceles are a frequent cause of “non-lesional” temporal lobe epilepsy. Even with a dedicated epilepsy protocol it may be difficult to identify small lesion like in this case. 3D-T2 sequences may be very useful in difficult cases, showing small cystic component, which may be overlooked on T1, an FLAIR. In addition, high resolution CT can show small erosions in the inner diploe related to the encephalocele. The point in this case is to consider temporal encephalocele as an entity that can cause epilepsy and the anterior temporal pole must be evaluated in detail above all, if previous MR where considered negative.
- Engel J Jr, Wiebe S, et al. Practice parameter: temporal lobe and localized neocortical resections for epilepsy: report of the quality standards Subcommittee of the American Academy of neurology, in association with the American Epilepsy Society and the American Association of Neurological Surgeons. Neurology. 2003;60(4):538–47.PubMedGoogle Scholar
- Von Oertzen J, Urbach H, et al. Standard magnetic resonance imaging is inadequate for patients with refractory focal epilepsy. J Neurol Neurosurg Psychiatry. 2002;73(6):643–7.Google Scholar
- Cendes F. Neuroimaging in investigation of patients with epilepsy. Continuum (Minneap Minn). 2013;19(3 Epilepsy):623–42.Google Scholar