The estimated prevalence of epilepsy in North America is 5 to 10 cases per 1000 people, and approximately one-third of those with epilepsy fail to achieve adequate seizure control with antiepileptic drugs or are unable to tolerate the side effects required to do so. A large body of evidence indicates that failure to adequately control seizures impairs cognition, decreases overall quality of life, and increases mortality.

Medical intervention is the first step in treating epilepsy. Despite the use of various antiepileptic drugs, when medical therapy fails to achieve seizure freedom, comprehensive surgical evaluation is appropriate. The most common and well-characterized focal epilepsy syndrome is medial (mesial) temporal lobe epilepsy (MTLE). For patients with intractable MTLE, an anterior temporal lobectomy and amygdalohippocampectomy (ATL) offers the possibility of seizure freedom with reduced reliance on antiepileptic medications.

The ATL has been refined over the past century from its earliest descriptions by Falconer and Morris into a safe and effective procedure that has been validated with class I evidence. This procedure has consistently provided seizure-free rates approaching 70% or better in MTLE patients. This chapter outlines the modalities used for diagnosis and evaluation of temporal lobe epilepsy (TLE), my surgical approach to MTLE, and outcomes after surgery.

From an electrical and clinical perspective, TLE can be divided into two different subtypes: MTLE and neocortical temporal lobe epilepsy (NTLE). The distinction between these subtypes is made based on differences with respect to electrophysiology, neuropsychological profile, pathophysiology, and response to surgery.

The most common cause for TLE is medial (mesial) temporal sclerosis (MTS). Pathologically, this is characterized by segmental loss of pyramidal cells, dispersion of granule cells, and reactive gliosis. Other causes of TLE include tumors, infections, vascular malformations, cortical dysplasia, and trauma. The differential diagnosis for NTLE is similar to MTLE, with the exception of MTS.

Among patients with epilepsy caused by an extratemporal epileptogenic focus or lesion, approximately 15% have associated MTS; these patients harbor what is termed dual pathology. It is not clear which lesion (or sometimes both) serves as the true epileptogenic focus in these patients; resection of both foci generally yields the highest likelihood of attaining seizure freedom as long as preoperative evaluation demonstrates localization of the seizure focus to one or both of the two sites. The kindling phenomenon has been implicated in causation of MTS in this patient population. Despite resection of the extratemporal focus, MTS can continue to act as an autonomous source of seizure generation.

The goal of epilepsy surgery is removal of the epileptogenic zone, which, if completely resected, potentially results in seizure freedom. The preoperative evaluation of TLE seeks to identify this zone and determine the safety of resective surgery for the seizure focus. Herein I will review the various diagnostic modalities involved with the diagnosis of the epileptogenic zone in TLE.

The preoperative evaluation includes a detailed history and physical exam. Seizure semiology, past medical history, family history, remote history of febrile seizures/trauma/meningoencephalitis and details regarding attempted antiepileptic medications should be discussed. Often it is helpful if a family member or friend is present who has witnessed typical or habitual seizure episodes. A complete neurologic examination, in conjunction with the history, may help estimate the location of the seizure focus.

The presentation of patients with TLE often follows a characteristic electroclinical syndrome that includes an abdominal sensation (aura) followed by hand/mouth automatisms and possible ictal contralateral dystonic posturing of the upper extremity. Postictal speech dysfunction indicates lateralization of seizure onset to the dominant hemisphere. Unfortunately, patients with MTLE and NTLE have similar clinical presentations and cannot be distinguished from one another based on clinical features alone. As expected, NTLE is associated with more frequent generalized seizures.

Scalp electroencephalogram (EEG) is an essential component of the initial evaluation and is usually performed on an outpatient basis for convenience. A 30-minute awake/sleep-deprived analysis may help localize the epileptogenic focus by correlation of the ictal and interictal discharges with clinical history and imaging findings.

Many patients with MTLE harbor unilateral anterior temporal interictal spikes on surface EEG. Additionally, some investigators report unilateral temporal rhythmic theta activity less than 30 seconds after electrical seizure onset to be associated with ipsilateral MTLE. The sensitivity of detecting an EEG abnormality can be increased by performing the test within 48 hours of a known seizure.

A critical component in determination of surgical candidacy requires inpatient admission to the epilepsy monitoring unit for continuous scalp EEG and video monitoring. Correlation of the clinical seizure manifestations with interictal and ictal discharges helps localize the epileptogenic focus. Provocative measures, such as medication reduction, sleep deprivation, and hyperventilation, may be used to induce epileptiform activity. If scalp EEG fails to lateralize the epilepsy focus, invasive EEG monitoring using intracranial electrodes (an intracranial study) may be necessary to provide further localization.

Scalp EEG fails to lateralize the seizure focus in up to one-third of patients, and even when noninvasive tests are lateralizing, up to 10% are falsely localizing. Therefore, invasive EEG monitoring (an intracranial study) should be undertaken when there is discordance among the different preoperative tests, suspicion of multifocal epilepsy, or MRI-negative TLE that requires differentiating NTLE from MTLE. This intracranial study will delineate the borders of the epileptogenic focus and its proximity to the eloquent cortices.

The characteristic electrical activity from the recording hippocampal depth electrodes for MTLE includes periodic spiking activity from the hippocampus, followed by episodes of high-voltage rhythms lasting up to 1 minute. As the adjacent entorhinal cortex and temporal neocortex are recruited, a synchronous and regular 5 to 9 Hz ictal rhythm is often seen in the subtemporal and anterior scalp EEG electrodes.

Another common EEG pattern for MTLE includes low-voltage, high-frequency discharge without preictal spiking. This activity is recorded by the hippocampal depth electrodes and medial-basal subdural electrodes in contact with the entorhinal cortex. In NTLE, the most common ictal onset is characterized by low voltage and high-frequency discharge that is associated with loss of background activity.

After scalp/invasive video EEG monitoring, patients who are found to suffer from psychogenic nonepileptic seizures, multifocal epilepsy, a generalized seizure disorder, or a nonlocalizing seizure focus are deemed not appropriate candidates for resective surgery. Almost half of the patients who undergo scalp EEG and video monitoring harbor a well-defined epilepsy focus or require an intracranial study to determine their surgical candidacy.

Magnetic resonance (MR) imaging is the primary diagnostic imaging modality of choice for TLE. The MR protocol should scan the whole head, thin-sectioned high-resolution T1- and T2-weighted images, in addition to gradient echo T2 sequence to investigate the presence of blood.

If a mass lesion or tumor is found, gadolinium should be administered to better establish the character of the lesion and its anatomic location. If MTLE is suspected, the required sequences include coronal T1-weighted imaging, fluid-attenuated inversion recovery (FLAIR) coronal sequences, and T2-weighted coronal sequences through the hippocampus.

Typical findings of mesial temporal sclerosis are hippocampal atrophy and increased FLAIR and T2 signal hyperintensity in the medial temporal lobe structures. Hippocampal atrophy on preoperative MR imaging is associated with improved seizure control following temporal lobectomy. MR imaging, in combination with the other diagnostic modalities discussed in this chapter, affords a more accurate localization of the epileptogenic focus.

Neuropsycological evaluation is important for patients with TLE because it can identify preoperative functional deficits and predict postoperative neuropsychological outcomes. Planned surgical removal of the temporal lobe and hippocampus necessitates the evaluation of language and memory, respectively. Patients with dominant lobe TLE display subtle verbal memory deficits and word-finding difficulty. In contrast, patients with nondominant TLE typically display visuospatial memory dysfunction.

The most common postoperative deficit following TLE surgery is memory decline. Verbal memory deficits are more likely to occur after surgery on the left temporal lobe, compared with visuospatial deficits following surgery on the right side. Patients undergoing a left temporal lobectomy who have average or better memory and language function are at higher risk for developing postoperative deficits. However, no such relationship has been found in patients who undergo right temporal lobectomy. A preoperative discussion with the patients at risk for postoperative memory or language dysfunction must be conducted before surgical treatment is offered.

Before the advent of functional MRI (fMRI), the Wada test was used to assess language and memory function of the two cerebral hemispheres independently. It is now used when fMRI is inconclusive, not available, or not achievable. Amobarbital or another anesthetic agent, such as methohexital, propofol, or etomidate, is injected into the internal carotid artery, temporarily suppressing function on that side while language and memory tests are performed. Baseline memory function is typically tested the day before the actual test.

Before injection of the anesthetic agent, the patient is asked to raise both arms and count aloud. Hemiplegia signifies adequate anesthesia. Language and memory are assessed while hemiplegia persists, evaluating the side harboring the suspected epilepsy focus first. Next, the contralateral hemisphere is tested 30 minutes later.

Language representation is considered to be unilateral (right or left) if global aphasia occurs after anesthetic injection into one hemisphere, and bilateral (with or without predominance) if some degree of preservation of language is seen after injection in both hemispheres.

For memory evaluation, patients are asked to correctly identify items shown during the hemiparesis phase. An overall memory score is assigned based on the ability of the contralateral side to support memory during anesthetic injection of the side ipsilateral to the suspected epilepsy focus. Memory scores of 50% to 67% have been deemed passing scores. The goal of the memory portion of the Wada test is to establish risk for postoperative memory decline. The less memory support the contralateral hemisphere owns, the greater the risk for postoperative memory deficit.

Despite the high accuracy of the Wada test in determining language and memory dominance, this test is still associated with false positives and negatives. Additionally, Wada test results can be affected by drug dose, unblinding of test assessors, and patient cooperation. Furthermore, the Wada test is associated with risks such as seizures, contrast allergy, catheter site hematoma, vascular dissection, stroke, and infection. The risk of arterial dissection or stroke is estimated at 1%.

Positron emission tomography (PET) utilizes radioactive isotopes linked to metabolically active molecules to analyze functionality in various regions of the body based on metabolic activity. When investigating TLE, PET seeks to localize the regions of interictal hypometabolism in the temporal lobe.

EEG recording during PET is important to ensure that hypometabolism in one hemisphere is not secondary to an active seizure on the contralateral side resulting in hypermetabolism.

Fluorodeoxyglucose (FDG) is the most commonly used isotope in PET. FDG-PET has a high specificity for MTLE. Mesial temporal sclerosis is associated with hypometabolism within the hippocampus, amygdala, entorhinal cortex, and temporal pole.

It has also been shown that hypometabolic regions identified using FDG-PET correlate well with the predicted laterality when compared with the results of depth electrodes. The sensitivity of PET is increased when the metabolic activity from both temporal lobes is sampled to examine hypometabolism in one temporal lobe as compared with the other.

PET does not provide any additional information if MTS is found on MRI. Therefore, PET is typically reserved for cases in which MRI and other preoperative tests fail to localize the seizure focus.

Cerebral blood flow to the brain is increased in regions of the brain undergoing seizure activity due to increased metabolic demand. Single photon emission computed tomography (SPECT) measures local cerebral perfusion using technetium-99m hexamethyl propelene amine oxime or technetium-99m bicisate. These compounds are taken up by neurons within seconds of injection and remain in the cells for several hours. Therefore, injecting radiotracers immediately following a seizure may help localize the area of ictal onset.

The sensitivity of SPECT is increased if interictal studies are used to compare the relative change in cerebral perfusion during seizure activity. Additionally, the difference in blood flow between interictal and ictal SPECT can be correlated and coregistered with high-resolution MRI to identify the epilepsy focus. SPECT is typically reserved for cases in which MRI and/or EEG are nonlocalizing.

Functional MRI (fMRI) examines neuronal activity via alterations of the MR signal due to changes in blood oxygenation levels. It is mainly used for localization of the eloquent cortex, such as motor and language cortices. It can also be coupled with EEG analysis to identify the zone of ictal onset. It has a spatial resolution of a few millimeters, and can be used as a noninvasive alternative to the Wada test for language lateralization and localization of the cortical speech areas.

As a newer imaging modality, only a handful of studies have correlated fMRI and surgical outcome. Increased signal activation during language and memory tasks ipsilateral to the ictal focus has been associated with greater deficits after surgical resection, a correlation that may be a stronger predictor than neuropsychological testing.

Patients with medically intractable epilepsy of medial temporal lobe origin, as defined by failure of two or more antiepileptic medication trials, are candidates for anteromedial temporal resection and amygdalohippocampectomy.

Patients may undergo resective surgery without invasive intracranial monitoring if they have: 1) ictal onset lateralized to one hemisphere on EEG, 2) unilateral MTS on MRI and/or evidence of unilateral mesial disease on PET or SPECT, 3) neuropsycological and Wada/fMRI tests to confirm memory and speech dysfunction consistent with the side of ictal onset, and 4) Wada confirmation that the contralateral side supports memory function. Patients who do not meet these criteria may need an intracranial study using electrodes and further testing with the aforementioned modalities before a surgical resection is performed.

The need for an intracranial study is dictated based on a suspicion raised by EEG or imaging modalities that epileptogenesis is not limited to the medial temporal lobe. The absence of hippocampal atrophy and presence of a neocortical dysplastic structural lesion on imaging studies, as well as documentation of neocortical ictal onset by basic invasive electrode recordings, indicate the need for a more in-depth intracranial study with adequate coverage of candidate cortices by subdural grid electrodes for ictal localization and language mapping. The use of depth electrodes within the medial structures is routine.

Right-handed individuals do not require a Wada study for right temporal lobe resections due the high likelihood of the left temporal lobe housing verbal memory.

In the above sections, I discussed the details of preoperative comprehensive epilepsy surgery evaluation for TLE. The use of intraoperative image guidance using MR imaging is advised for the novice surgeon. This tool guides the entrance into the temporal horn via the middle temporal gyrus.

The perioperative events and anesthetic medications lead to a drop in the serum levels of anticonvulsant medications, exposing the patient to the risks of postoperative seizures. Supratherapeutic levels of anticonvulsant medications are therefore advised during the perioperative period because seizures complicate the patient’s recovery from surgery.

I do not use intraoperative electrocorticography to guide the extent of neocortical or hippocampal resection. If the extent of seizure focus is in question, intracranial monitoring is more effective to guide the surgeon.

The anatomy of the medial temporal lobe is certainly the most complex part of the intraparenchymal anatomy.

The details of technique for a standard ATL involve a very defined set of steps which prevent operator’s disorientation that can lead to injury to the thalamus and the contents of the basal cisterns.

Further details about the performance of the exposure and craniotomy are summarized in the Pterional Craniotomy chapter.

Bridging veins from the Sylvian fissure to the sphenoparietal sinus are identified at the anterior temporal tip. Retraction of the temporal lobe can cause bleeding from these veins unless they are cauterized during this portion of the procedure.

One should recognize that the ventricle is more posterior than expected due to the size of the amygdala that occupies the anteromedial temporal lobe. If the operator is disoriented during subcortical dissection, he or she may continue medial white matter dissection and completely miss the ventricle and inadvertently enter the brainstem, thalamus, or structures in the center of the hemisphere.

The Meyer’s loop courses along the roof and lateral aspects of the temporal horn. The operator can avoid a superior temporal quadrantanopsia by maintaining the plane of dissection within the medial fusiform gyrus.

The extent of hippocampal resection is associated with improved postoperative seizure outcome. Hippocampectomy to the level of the quadrigeminal plate is advised. Retained mesial structures is the most common cause of recurrent seizures.

If there is evidence of MTS associated with an adjacent temporal epileptogenic lesion, dual pathology is suspected. I remove the lesion and the nondominant hippocampus to maximize the chance of postoperative seizure freedom. However, if the lesion does not directly affect the dominant hippocampus, the dominant hippocampus should be spared to minimize the risk of postoperative memory dysfunction.

Supratherapeutic levels of anticonvulsant medications are advised during the perioperative period. Seizures during the immediate postoperative period are not uncommon and do not indicate a poor long-term seizure outcome from surgery. The patient and his or her family should be reassured.

Anticonvulsant drugs are continued for approximately 1 to 2 years postoperatively, even in the absence of seizures during this period. If the patient remains seizure free during this time, cessation of these medications may then be considered.

• Early identification of the temporal horn orients the operator anatomically and prevents misguided operative trajectories.
• Limiting the extent of occipitotemporal fasciculus dissection along the lateral hippocampus minimizes the breadth of expected visual field defects caused by injury to the Meyer’s loop.
• Transient neuropathies of cranial nerves III and IV can be avoided by meticulous protection of the arachnoid planes and by avoiding bipolar cauterization near the tentorial edge.
• Compromise of the anterior choroidal artery can be avoided by minimizing bipolar cauterization and manipulation of the choroid plexus. Cottonoid patties cover the choroid plexus to avoid its direct manipulation which leads to bleeding.
• The posterior cerebral artery can remain safe during hippocampus mobilization by carefully cauterizing and transecting the perforating vessels entering the hippocampal sulcus once they are well within the fissure.

Contributor: Stephen Mendenhall, MD

DOI: https://doi.org/10.18791/nsatlas.v7.ch02.1