Identification of Potential Epilepsy Surgery Candidates

In the past 10 years, epilepsy surgery has increasingly been recognized as a viable treatment for patients with medically refractory seizures. At the time of the first Palm Desert Conference, in 1987, only 26 epilepsy surgery centers existed in the United States, having performed approximately 500 therapeutic surgical procedures in 1985. By the time of the second Palm Desert Conference, in 1992, 67 epilepsy surgery programs were operational in the United States, having performed approximately 1500 epilepsy surgeries in 1990. ^[1]

At present, over 100 epilepsy centers exist in the United States. Unfortunately, it often takes 20 years before patients are referred for an epilepsy surgery evaluation. This delay is likely due to physician’s perception of epilepsy surgery as a “last resort” procedure.

Most epilepsy centers define intractability as failure of at least 2 or 3 first-line antiepileptic medications. ^[2] However, many physicians define medical intractability differently. They will often try numerous antiepileptic drugs (AEDs) trials before referral for a presurgical evaluation is even considered, even though the prospect of becoming seizure free is only 5%.

The introduction of newer AEDs with better tolerability and fewer drug-drug interactions has made a significant impact on the treatment of epilepsy, in that efficacy can be optimized with less dose-limiting adverse effects. However, a significant proportion of patients still have intractable epilepsy.

More than 250,000 people in the United States have intractable partial epilepsy, and a conservative estimate is that at least 50% of these patients are potential surgical candidates. These patients are at an increased risk for sudden death and personal injury. They are less likely to be employed than patients with well-controlled seizures.

In addition, the epileptogenic process itself produces interictal dysfunction with adverse consequences on cognition and mood, which can be irreversible, particularly in children. After failure of 2 first-line AEDs, the chance for seizure freedom with additional therapeutic regimens can be as low as 5-10%.

In 2009, Berg and colleagues identified a prospectively identified community-based cohort of children from 1993-1997 and found the frequency of magnetic resonance imaging (MRI) lesions potentially relevant to nonidiopathic epilepsy was 16% and that 8% underwent an epilepsy-related surgical procedure. ^[3]

On the basis of these data, Berg et al estimated that 127 of 1,000,000 new cases of epilepsy per year would be pharmacoresistent, 52 of 1,000,000 patients with childhood-onset epilepsy would undergo epilepsy surgery evaluations, and 27 of 1,000,000 patients would have an epilepsy-related surgical procedure. As of 2006, there were 74 million children younger than 18 years in the United States.

A study by Engel et al found that early surgical treatment may be beneficial. For patients who had mesial temporal lobe epilepsy and disabling seizures for no more than 2 consecutive years following adequate trials of 2 brand-name AEDs, resective surgery plus AED treatment resulted in a lower probability of seizures for at least 2 years posttreatment, as well as improved health-related quality of life, than continued AED treatment alone. ^[4]

Certain readily definable, surgically remediable syndromes are easily identified, have an excellent prognosis for seizure control, and have minimal surgical morbidity. Postoperatively, many patients with these syndromes have improved quality of life (QOL), improved cognitive, psychosocial, and occupational outcome. In children with catastrophic epilepsy due to diffuse hemispheric syndromes, early surgical intervention stops seizures and reverses the cognitive decline so that these patients can develop normally, eventually leading relatively normal lives.

For more information, see Epilepsy Surgery and Epilepsy and Seizures.

In the United States, approximately 3 million people (0.5-0.9% of the population) have been diagnosed with epilepsy. Half of these patients have partial seizures, and in 30% seizures are not controlled adequately with antiepileptic drugs (AEDs). A conservative estimate is that one half of patients with medically intractable epilepsy are potential candidates for epilepsy surgery. Another 10-15% of patients with epilepsy have severe symptomatic or cryptogenic generalized seizure disorders that do not respond to AEDs.

In the United States alone, there are 200,000 potential surgical candidates. According to earlier estimates, at least an additional 5,000 potential surgical candidates are added to this pool each year.

Unfortunately, many primary care providers and even some neurologists believe that epilepsy surgery is a last-resort treatment. Many patients who could become seizure free with surgery instead undergo treatment with multiple medications over many years, suffering the adverse effects of recurrent seizures, the short- and long-term side effects of AEDs, and the psychosocial and occupational consequences of recurrent seizures.

Medical intractability is defined currently by many investigators as seizures that are not controlled after an adequate trial with 2 first-line AEDs. Some advocate at least 3 regimens, including 1 trial of 2-drug therapy. If 3 trials of monotherapy with first-line drugs are not successful, the chance that the patient will respond to a fourth drug as monotherapy or polytherapy is only 5%.

Determining intractability also requires an understanding of how the seizures affect the patients’ quality of life (QOL) in terms of their psychological, interpersonal, and occupational functions. For example, even as few as 2-3 seizures a year may be disabling to an individual whose occupation requires transportation with a motor vehicle.

A consensus proposal by the Task Force of the International League Against Epilepsy (ILAE) Commission created an operational definition of drug-resistant epilepsy, noting that this definition may change as more empirical evidence becomes available. The ILAE defines drug-resistant epilepsy as “a failure of adequate trials of two tolerated and appropriately chosen and used AED schedules (whether as monotherapies or in combination) to achieve sustained seizure freedom.”

Seizure freedom is defined as freedom from seizures for a minimum of 3 times the longest preintervention interseizure interval (determined from seizures occurring within the preceding 12 months) or 12 months, whichever is longer. ^[5]

Many clinical trials of adjunctive therapy for AEDs have a high placebo responder rate, therefore confounding how much of the response is due to the AED itself. In this regard, Beyenburg and colleagues performed a systematic meta-analysis of the evidence to determine the placebo-corrected net efficacy of adjunctive therapy with modern AEDs; they found that for both adults and children, only 6% became seizure free as compared with placebo and only 21% had a less than 50% reduction in seizures after placebo correction. ^[6]

Mattson et al reported similar results from 2 Veterans Administration (VA) cooperative trials in patients untreated or inadequately treated with AEDs, finding that 60-70% of patients with generalized tonic-clonic seizures alone attained seizure freedom for at least 12 months. ^[7] In patients with complex partial seizures, on the other hand, the prognosis was less favorable, with only 23-26% attaining seizure freedom for at least 12 months.

A study by Kwan and Brodie revealed that only 47% of patients with untreated epilepsy became seizure free during treatment with the first AED, and 14% became seizure free with a second or third AED. ^[8]

These studies underscore the fact that patients with partial epilepsy who do not become seizure free after trials with 2 first-line AEDs are less likely to achieve seizure freedom with additional AED trials. In this regard, early identification of these patients with surgically remediable epilepsy is crucial, since the rate of seizure-free surgical outcome ranges from 70-80% in well-selected cases. ^[9]

Furthermore, evidence from Berg and colleagues indicate that the prototype surgically remediable syndrome, mesial temporal lobe epilepsy (TLE), can have prolonged periods of remission before becoming intractable. ^[10] For example, the Yale multicenter epilepsy surgery study found that, in the group with TLE, the average time to develop intractable epilepsy, defined as failure of 2 AEDs, was 9 years.

Often, medial TLE begins in childhood but does not become intractable until adolescence or early adulthood indicating that medial TLE can have a period of prolonged remission before becoming intractable.

A study of childhood onset epilepsy by Sillanpaa and colleagues showed similar findings: 19% percent of children with epilepsy had intractable epilepsy without remission from the onset of their epilepsy, whereas 14% of children had an early or late remission (of 5 or more years) with subsequent intractability. ^[11]

In view of the finding that some cases of intractable epilepsy were well controlled early in their course, both physicians and patients alike may falsely believe that prolonged remission will once again occur, thus leading to prolonged AED trials even though this is not the case.

Whereas in some patients, a delay occurs from onset to intractability, the time from medical intractability to surgery is also prolonged. In patients with childhood-onset and adolescent-onset epilepsy, the average time to referral is 15-16 years. In those with adult onset (age >20 y), the time to surgery is 4.5-7 years. In the UCLA series, the time from seizure onset to surgery has increased over the years from 12.5 years to 19 years. Another study reported similar findings, where the average duration of epilepsy before referral was 18 years.

Furthermore, 39% of patients were self referred, and 36% of this group had been advised by their primary neurologist not to consider surgery. Therefore, both the early remission(s) of intractable epilepsy, which may lead to unnecessarily prolonged AED trials in hopes of once again attaining prolonged remission, and the misperceptions of the risks of epilepsy surgery by physicians and patients lead to prolonged delays before surgical referral.

With the approval of several newer AEDs and vagus nerve stimulation in recent years, more treatment options are available. However, prolonged trials of AEDs in patients with well-defined, surgically remediable epilepsy (eg, hippocampal sclerosis or a well-defined structural lesion) offer diminishing chance for seizure freedom (5-10%) and delay surgical treatment that can reduce substantially or perhaps eliminate seizures. ^[12]

A study by Semah and colleagues showed that the pathologic substrate of epilepsy indicated a poor long-term prognosis for remission. ^[13] For example, patients with hippocampal sclerosis or cortical dysplasia had a poor prognosis, with the worse prognosis in patients with dual pathology defined as a structural lesion with concomitant hippocampal sclerosis.

Other studies have shown similar findings. Therefore, recognition that patients with localization-related epilepsy with hippocampal sclerosis or a well-defined lesion have a universally poor prognosis with medical treatment but a good prognosis with surgical treatment is of the utmost importance. These patients need to be identified early in life before the psychosocial consequences of prolonged disability prevent useful rehabilitation, even if the patient eventually undergoes epilepsy surgery, and becomes seizure free.

Patients with uncontrolled epilepsy often have low self-esteem, impaired social relationships, and reduced occupational function. Vickrey and colleagues compared quality of life (QOL) in patients with epilepsy to that of patients with hypertension, diabetes, heart disease, or depression and found that patients who were seizure free after surgery scored better on QOL measures than the other patient groups. ^[14] Patients who continue to have seizures after epilepsy surgery, on the other hand, scored worse in terms of emotional well-being and overall QOL.

A classic study by Sillanpaa and colleagues showed that adolescents with intractable epilepsy are less likely to finish high school, find employment, and get married. ^[15] Evidence in the literature supports the idea that patients with more frequent seizures have worse QOL than those with fewer seizures. QOL decreases as seizure frequency increases to 10-12 seizure per year. However, once seizure frequency exceeds 1 or seizure per month, the degree of impairment of QOL is essentially the same whether the seizure frequency is 2 or 20 per month.

Mood disorders also occur with greater frequency in patients with epilepsy than in other patients with other medical conditions with a similar degree of disability. Approximately 20-30% of patients with epilepsy have a comorbid mood disorder. In patients with temporal lobe epilepsy (TLE), approximately 50% have concomitant mood disorders consisting of depression and anxiety. Recent studies have shown that patients with intractable epilepsy and comorbid depression who have similar seizure frequencies score worse on QOL measures than patients without depression.

Patients with intractable epilepsy also have loss of autonomy, in that they are unable to drive and are often shielded from responsibility by family members out of fear of injury. In fact, Gilliam et al reported that patients with intractable epilepsy report driving, independence, and employment as their most common concerns. ^[16]

Occupational status is also impaired in patients with intractable epilepsy. Sperling and colleagues reported that 71% of patients who were seizure free were employed, as compared to only 44% of patients with persistent seizures. ^[17] Jacoby and colleagues also found that reduced employment levels were related to worsening seizure control. ^[18]

A European study involving 5000 patients in 15 countries showed that more than 50% of patients reported feeling a stigma associated with epilepsy. Public misperception also adversely affects the ability of patients with epilepsy to obtain employment.

Intractable epilepsy is also associated with reproductive endocrine disorders such as polycystic ovary syndrome, hypogonadotropic hypogonadism, and anovulatory cycles. Part of this increased risk is associated with certain antiepileptic drugs (AEDs), such as valproate.

Patients with epilepsy can also have reduced libido and reduced sexual arousal; this may be related in part to the older enzyme-inducing AEDs (EIAEDs). ^[19] Recent evidence indicated that these EIAEDs can produce sexual dysfunction in men by increasing the metabolism of testosterone and increasing the production of sex hormone–binding globulin, thereby reducing free testosterone levels. ^{[20, 21]}

The risk for injury and mortality in patients with intractable epilepsy is greater than that in the general population. The overall mortality rate in persons with epilepsy is 2-3 times greater than that in the general population. This increased mortality risk is related to the etiology of the epilepsy, the degree of seizure control, and the extent of neurologic handicap. The most common causes of mortality are sudden unexplained death and accidents.

Overall, persons with epilepsy have a 24-fold greater risk of sudden unexplained death in epilepsy (SUDEP) than the general population. Patients with medically intractable epilepsy, however, have an even greater risk, approaching 50-100 times the risk observed in the general population. In fact, SUDEP accounts for 7-17% of deaths in patients with epilepsy and more than 50% in patients with intractable epilepsy.

In patients aged 15 years and older with uncontrolled epilepsy, the risk of dying is approximately 0.5% per year, and this risk accumulates over their lifetime. Patients whose seizures stop after surgery have mortality rates as low as those in the general population. Therefore, early surgical intervention in appropriate candidates can potentially reduce mortality.

These patients are also at increased risk for injury. One community-based study found that 35% of patients had sustained 1 or more injuries, including head injury (24%), burn or scald (16%), dental injury (10%), and fracture (6%). The risk for injury is related to higher seizure frequency, seizure severity, and seizure type (generalized tonic-clonic and tonic).

Recurrent seizures themselves, the epileptogenic process, or reactive inhibitory mechanisms may contribute to the progressive nature of epilepsy. Evidence for the progressive nature of epilepsy comes from phenomena such as kindling and secondary epileptogenesis in experimental animals. Whether these phenomena occur in humans is unknown. However, patients with refractory epilepsy clearly have impaired psychosocial and occupational functions that are less likely to improve the longer the seizures remain uncontrolled with surgery.

Epilepsy usually is viewed as an “ictal” disorder, and the interictal effects of chronic epilepsy were largely ignored until comparatively recently. Patients with epilepsy also have interictal dysfunction, which includes material-specific memory deficits, mood disorders, neuropsychological dysfunction, and metabolic abnormalities in brain regions outside the epileptogenic region. For example, Krauss et al demonstrated that interictal spikes on depth electrode recording interfered with working memory in patients with TLE. ^[22]

This interictal dysfunction also significantly contributes to the morbidity of chronic epilepsy and may be potentially reversible. In this regard, Cendes and colleagues performed a study of patients with TLE, demonstrating that the interictal metabolic abnormalities on magnetic resonance spectroscopy observed in the contralateral temporal lobe before surgery partially returned to normal 3 months after epilepsy surgery. ^[23]

Longitudinal studies of patients with chronic TLE have shown progressive hippocampal atrophy over time on volumetric magnetic resonance imaging (MRI) studies, and this atrophy correlates with the duration of epilepsy and the number of generalized tonic-clonic seizures. Functional imaging data from positron emission tomography (PET) show a relationship between epilepsy duration and the degree of hypometabolism in the temporal lobe.

The progressive nature of epilepsy is best exemplified by children with catastrophic epilepsy due to diffuse hemispheric disturbances. These patients develop normally until refractory infantile spasms develop. These seizures occur several times each day and are associated with developmental decline with loss of motor and language functions. The epileptogenic discharges from the abnormal hemisphere interfere with the normal development of the contralateral nonepileptogenic hemisphere.

The developmental decline becomes irreversible when the seizures persist beyond the critical period for development of certain skills, such as language. Invariably, these patients develop intellectual disability and require chronic institutional care. Early identification of children who are candidates for functional hemispherectomy or multilobar resection is crucial; cessation of seizures results in developmental improvement sufficient to allow these patients to develop normally and, eventually, to lead relatively normal lives.

Intractable epilepsy has an enormous impact on various components of an individual’s life, including social, psychiatric, occupational, cognitive, and sexual functioning. It also increases their risk for injury and mortality. Intractable epilepsy may also be associated with progressive cognitive and structural/functional brain changes that can be reversible if identified and treated early.

Epilepsy typically affects individuals before or during the most productive years of their life. Furthermore, the annual estimated cost of epilepsy in the US is $12.5 billion, with most of the cost attributable to patients with intractable epilepsy. For these reasons, it is imperative that patients with intractable epilepsy be referred for an epilepsy surgery evaluation with the goal of abolishing or reducing seizures, eliminating AED side effects, and restoring psychosocial function and QOL.

In the authors’ experience, some patients report only simple partial seizures. However, upon further questioning of the patient and family, the occurrence of amnestic episodes of unresponsiveness or loss of time for which the patient is unable to account becomes apparent, suggesting unrecognized complex partial seizures.

Some patients, particularly those who live or sleep alone, may have episodes of trauma to the lateral aspect of their tongue or may have urine incontinence upon awakening, indicating unrecognized secondarily generalized tonic-clonic seizures.

Patients are usually unaware of seizures, unless environmental cues are present. In a study of patients admitted to an epilepsy-monitoring unit, Blum et al found that only 26% of patients always identified their seizures. ^[24] Patients with right-sided temporal lobe seizures were aware of 50% of their seizures, and patients with left-sided temporal lobe seizures were aware of only 10% of their seizures.

Thus, obtaining a careful history from both the patient and the family members is essential. Determining how the seizures affect the patient’s quality of life (QOL) is important. Family members can provide information about possible unrecognized seizures. With regard to assessing the impact of seizures on QOL, a useful question to ask the patient is “How would your life change if you no longer had seizures?”

Ask the patient about auras; these have great value in localization. The importance of historical information has been recognized since Jackson used ictal characteristics to localize a patient’s seizure. In a study by Palmini and Gloor, auras were found to be highly localizing. ^[25] For example, an elementary visual aura suggests onset in the occipital lobe, a complex visual hallucination involving faces suggests onset near the fusiform or the inferior temporal gyrus, and an auditory aura is often associated with seizure onset in or near the superior temporal gyrus.

Whereas auras may represent spread from adjacent epileptogenic silent cortical regions to nonepileptogenic symptomatic regions, in many cases the aura can direct imaging with high-resolution magnetic resonance imaging (MRI) to identify a subtle cortical malformation that would otherwise be overlooked (see Neuroimaging of Epilepsy).

Similarly, postictal deficits may provide lateralizing as well as localizing information. For example, postictal hemiparesis suggests onset near primary motor cortex, postictal language deficit suggests dominant hemisphere lateralization, and postictal visual field defect indicates onset in the occipital lobe.

Temporal lobe epilepsy

The syndrome of medial temporal lobe epilepsy (TLE), which is associated with hippocampal sclerosis, is an example of a surgically remediable syndrome. Approximately 40-67% of these patients have a history of a complicated febrile convulsion (a febrile seizure lasting >30 min). These patients typically present with seizures in late childhood, at which time seizures are well controlled with antiepileptic drugs (AEDs). As the child enters adolescence and early adulthood, the seizures recur and become refractory to multiple medication trials.

Many patients with TLE experience auras, which are simple partial seizures that precede most or all of their complex partial seizures and which often occur in isolation. These auras consist of autonomic phenomena, such as epigastric rising/nausea; olfactory auras, such as a strange taste or odor; or psychic auras, such as fear, déjà vu/jamais vu, depersonalization, or derealization. Secondarily generalized tonic-clonic seizures, when present, are infrequent and easily controlled with AEDs.

These patients are hypothesized to have isolated auras and few generalized seizures because the hippocampus has few connections with the contralateral hippocampus, making seizure propagation slow. Their complex partial seizure consists of behavioral arrest; wide-eyed stare with pupillary dilation; and oral or alimentary automatisms, such as repetitive chewing and lip smacking. Because the hippocampus is involved bilaterally during the complex partial seizure, patients are amnestic for events that occur during the seizure.

Certain clinical features of temporal lobe seizures have lateralizing value. For example, ictal speech usually is associated with nondominant temporal lobe seizure onset. Dystonic limb posturing, when present, is contralateral to the side of temporal lobe seizure onset. This dystonic posturing usually involves flexion of the arm at the elbow with internal or external rotation of the forearm, flexion at the wrist, and extension of the fingers. Some patients may have ipsilateral head turning and contralateral dystonic posturing.

Postictal nose wiping with one hand is defined as wiping of the nose twice in the postictal period. Geyer and colleagues demonstrated that the nose-wiping hand is usually ipsilateral to the temporal lobe of onset. ^[26] Postictal dysnomia lasting for greater than 2 minutes suggests onset in the dominant temporal lobe. Other lateralizing signs in TLE include postictal thirst, peri-ictal urinary urge, and ictal spitting, which lateralize to the nondominant hemisphere.

Early clonic or tonic activity or an aura suggesting extratemporal onset (eg, unilateral somatosensory aura, visual aura) places the diagnosis of medial TLE in doubt, in that extratemporal foci may propagate to the medial temporal lobe and produce seizure semiology that is indistinguishable from that of medial temporal lobe onset seizures.

Common extratemporal to medial temporal propagation pathways, as described by Ajmone-Marsan and Ralston, ^[27] include (1) the posterior cingulate gyrus to the medial temporal lobe through the cingulum, (2) the orbitofrontal cortex through the uncinate fasciculus, (3) the parietal lobe, particularly the inferior parietal lobule, through the middle longitudinal fasciculus, and (4) the occipital lobe through the inferior longitudinal fasciculus.

The interictal electroencephalogram (EEG) typically shows anterior to midtemporal epileptiform discharges, which are usually unilateral but may be bilaterally independent in approximately 20-33% of cases. However, bilaterally independent, medial temporal interictal spikes do not necessarily indicate the presence of bilateral epileptogenic regions capable of generating spontaneous seizures. In fact, many patients with bilaterally independent temporal lobe interictal epileptiform abnormalities have their habitual seizure onsets in a single temporal lobe.

Patients with bilaterally independent temporal lobe seizure onsets and lateralizing data from magnetic resonance imaging (MRI) or positron emission tomography (PET) and Wada testing can also be good surgical candidates. Infrequently, a patient with TLE has rare or infrequent interictal epileptiform discharges.

High-resolution MRI using a T1-weighted, spoiled-gradient recall sequence with contiguous slices perpendicular to the long axis of the temporal lobe is most sensitive in detecting unilateral hippocampal atrophy in approximately 85% of patients. Hippocampal atrophy on MRI correlates with the presence of hippocampal sclerosis.

Other MRI findings include increased signal in the hippocampus on conventional spin echo T2-weighted imaging. Clifford Jack and the Mayo Clinic group reported that MRI with fluid-attenuated inversion recovery (FLAIR) shows an even greater sensitivity for detecting signal changes within the abnormal sclerotic hippocampus than conventional spin echo T2-weighted imaging does. ^{[28, 29]}

More recent data from Bernasconi and colleagues indicate that some patients may have entorhinal cortex atrophy in the absence of hippocampal atrophy. ^[30] Other imaging findings include anterior temporal tip atrophy, thinning of the anterior temporal white matter, particularly the temporal stem, or smaller left temporal lobe volume.

Extratemporal lobe epilepsy

Extratemporal lobe epilepsy also may be treated effectively with epilepsy surgery, particularly when a clearly defined lesion is present on high-resolution MRI. In fact, surgical outcome improves from 20% seizure free in patients without a lesion to 70% seizure free in patients with a lesion.

Extratemporal seizures may have variable seizure semiologies that represent seizure propagation patterns rather than the region of onset; however, these regions usually have selected pathways of propagation that may help to narrowly define the potential epileptogenic region.

Frontal lobe epilepsy

Frontal lobe seizures are typically brief, lasting seconds, with minimal postictal state. They usually occur in clusters, and they occur at night or predominantly in the morning. Unlike the semiology of temporal lobe seizures, that of frontal lobe seizures varies depending on the region of onset. For example, frontal lobe seizures propagate so rapidly that the clinical semiology may represent seizure propagation rather than the region of onset.

The semiology of frontal lobe seizure varies according to where the seizure originates: basal frontal (orbital frontal cortex, gyrus rectus), medial frontal (cingulate cortex, supplementary sensorimotor cortex), dorsolateral frontal (lateral frontal lobe extending posteriorly to the precentral sulcus and includes premotor cortex) or primary motor cortex. However, some clinical features of frontal lobe seizures may suggest a particular region of onset.

For example, seizures characterized by agitation with prominent motor activity and thrashing (ie, hypermotor seizure) suggest orbitofrontal onset. Typically, orbitofrontal seizures are clinically silent when the discharge remains there, and the prominent hypermotor activity becomes apparent after propagation to other parts of the frontal lobe, such as the medial frontal lobes. For example, frontal lobe seizures characterized by staring for several seconds or bilateral tonic facial grimacing suggest medial frontal lobe onset (ie, anterior cingulate cortex).

Seizures arising from the supplementary motor region in the medial frontal lobe anterior to the motor strip (area 6) are characterized by asymmetric tonic posturing of bilateral limbs (usually with a figure 4 posture with 1 arm tonically flexed and 1 arm tonically extended), monotonous vocalization, and variable preservation of consciousness; seizure onset is usually contralateral to the tonically extended limb.

Forced head, eye, and body version suggests onset in the dorsolateral frontal lobe. With dorsolateral frontal lobe seizures, forced head and body version is usually contralateral to the hemisphere of onset. However, a cautionary note is that posterior parietal and occipital lobe seizures can also have head and eye turning that may be versive or cursive (nonforced). One study showed that nonforced head turning at ictal onset without a tonic component or hemi-facial clonic twitching was usually ipsilateral to the hemisphere of onset.

Gyratory seizures are defined as a rotation around the body axis during the seizure at least 180 degrees. ^[31] Gyratory seizures that begin with forced head version are usually contralateral, whereas gyratory seizures without forced head turning are usually ipsilateral to the hemisphere of onset.

Seizures arising from central regions consist of focal clonic, tonic, or tonic-clonic activity of the face, arm, or leg. In the authors’ experience, seizures arising from the pericentral region include tonic extension of the contralateral limb(s) upper, lower and tonic contralateral facial contraction, or both. Seizures arising from the frontal operculum consist of unilateral facial clonic twitching and profuse salivation, immediately followed by tonic posturing of all limbs.

With most frontal lobe epilepsies, except orbitofrontal epilepsy, secondarily generalized tonic-clonic seizures are common because the frontal lobe neocortex has dense connections to other frontal and extrafrontal regions. In addition, the frontal lobe has dense callosal connections to the contralateral frontal lobe, allowing rapid contralateral propagation and subsequent secondary generalization. Auras are uncommon; when present, they are usually nondescript, except with seizure arising in primary motor cortex, where a primary somatosensory aura can be present.

Parietal lobe epilepsy

Parietal lobe seizures may be associated with a lateralized somatosensory phenomenon in the face or limb(s), vertigo, or a sense of motion. ^[32] Anterior parietal lobe seizures usually mimic frontal lobe seizures because of spread to frontal lobe regions. Posterior parietal lobe seizures often spread to the temporal lobe, producing temporal lobe semiology indistinguishable from that of seizures with primary origin in the temporal lobe.

Seizures arising from the dominant supramarginal gyrus can present with an aura of a “presence” as though someone is in the room with the person experiencing the seizure. Seizures arising from the nondominant supramarginal gyrus can present with autoscopy, described as a feeling of being outside and above one’s body.

Occipital lobe epilepsy

Occipital lobe seizures may be associated with a stereotyped visual aura consisting of unformed elementary visual phenomena, such as the sparkles of light that some patients describe as “television static.” Unlike a migraine aura, which lasts several (usually >5) minutes, an epileptic visual aura is usually brief, lasting several seconds or as long as 1-2 minutes.

In addition, an epileptic visual aura usually does not migrate across the visual field; however, the visual image may rotate in place. It may consist of colorful shapes that are present in the central visual field. Migraine aura also may consist of colorful shapes, but usually these are in the periphery and not in the central visual field.

Postictal blindness, when present, is a highly localizing finding that suggests occipital lobe onset. Although forced bilateral eye blinking during a seizure is nonspecific with regard to localization, it suggests an occipital lobe seizure origin if it occurs at seizure onset.

When an occipital lobe seizure without loss of awareness (aura) evolves into a focal seizure with loss of awareness, it may be indistinguishable from a temporal lobe seizure or a frontal lobe seizure because of spread of the seizure discharge through the inferior longitudinal fasciculus (ie, temporal propagation) or the superior longitudinal fasciculus (ie, frontal propagation).

Eye version in occipital lobe epilepsy can be either ipsilateral or contralateral to the epileptogenic region, and is likely due to propagation of the region for visually guided saccades in the inferior parietal lobule.

The pathologic substrate of extratemporal lobe seizures includes low-grade gliomas: developmental tumors such as gangliogliomas and dysembryoplastic neuroepithelial tumors, arteriovenous malformations (AVMs), cavernous malformations, encephalomalacia, and malformations of cortical development.

Schizencephaly is defined as a cleft that is lined with gray matter. This cleft extends from the pial surface to the ventricle, usually is located in the central region, and often is associated with contralateral hemiparesis. It presents in 2 forms: a closed lip (type 1) and an open lip (type 2).

If patients with schizencephaly have seizures with temporal lobe semiology, they may be candidates for an anterior temporal lobectomy. For example, if ictal electroencephalographic (EEG) onset localizes to one temporal lobe, and magnetic resonance imaging (MRI) shows unilateral hippocampal atrophy or fluorodeoxyglucose positron emission tomography (FDG-PET) shows unilateral temporal hypometabolism, a unilateral anterior temporal lobectomy offers seizure freedom in 70% of cases.

Porencephaly is defined as a cyst that is contiguous to the lateral ventricle and most commonly is associated with a perinatal ischemic insult. Ho et al reported that patients with unilateral porencephaly, like those with schizencephaly, also might be excellent surgical candidates when they have seizures with temporal lobe semiology and noninvasive data that localize seizures to a single temporal lobe. ^[33]

Patients with intractable epilepsy and vascular congenital hemiparesis can be surgical candidates for functional hemispherectomy if the seizures lateralize to the affected hemisphere or for anteromedial temporal resection if hippocampal atrophy and lateralized EEG findings are present. Similarly, patients with severe cerebral hemiatrophy and unilateral hippocampal sclerosis can also be good candidates for anteromedial temporal resections.

Earlier studies showed that patients with periventricular nodular heterotopias with seizure localization to the temporal lobe did poorly after temporal resection. Subsequent studies showed that resection of the nodular heterotopias and sometimes adjacent neocortex can result in a favorable surgical outcome in some patients.

These patients usually require invasive EEG monitoring with depth electrodes with or without concomitant subdural grid placement for localization of their epilepsy. In these cases, preoperative studies with magnetic source imaging and subtraction ictal single-photon emission computed tomography (SPECT) are of importance in planning intracranial electrode placement.

Several studies from Miami Children’s, the Mayo Clinic, the Montreal Neurologic Institute, and the Cleveland Clinic have shown that patients with tuberous sclerosis and localization-related epilepsy can also be good surgical candidates even when multiple cortical tubers are present and/or multifocal interictal epileptiform abnormalities are seen. Other conditions such as Sturge-Weber syndrome are also amenable to surgical resection.

With a thorough preoperative evaluation, the epileptogenic region and functional regions can be delineated such that a surgical resection can be offered with little or no deficit.

Children with intractable infantile spasms and diffuse hemispheric abnormalities, such as hemimegalencephaly, Sturge-Weber syndrome, porencephalic cyst, Rasmussen encephalitis, or perinatal unilateral cerebral infarct, can be candidates for functional hemispherectomy or multilobar resection if the patient has a useless hand. These patients need to be identified early, particularly those with the dominant hemisphere affected, since language can shift to the opposite hemisphere if surgery is performed while the patient is younger than 6 years.

Some patients with infantile spasms and an electroencephalogram (EEG) showing diffuse abnormalities (eg, hypsarrhythmia) may have a focal abnormality causing the seizures. Chugani and colleagues at UCLA demonstrated that fluorodeoxyglucose positron emission tomography (FDG-PET) is sensitive in identifying a focal region of hypometabolism in infants with medically intractable infantile spasms. ^[34]

Sankar and colleagues reported that magnetic resonance imaging (MRI) performed in individuals older than 2 years, when myelination has occurred, identifies focal regions of cortical dysgenesis. ^[35]

Patients who appear to have multiple seizure types arising from different brain regions are excluded from seizure surgery, unless 1 of the seizure types present is clearly the most frequent and disabling. Clinicians must be cautious because some patients have a clinical history that indicates multiple seizure types while video-electroencephalographic (VEEG) long-term monitoring of the patient reveals a single habitual seizure type.

Patients with bilateral hippocampal atrophy on MRI with poor memory and bilaterally independent seizure onsets usually have poor outcome. Those patients with bilateral hippocampal atrophy or bilaterally symmetric temporal hypometabolism on fluorodeoxyglucose positron emission tomography (FDG-PET) and unilateral ictal onsets usually are excluded from epilepsy surgery, unless the memory testing on Wada testing shows poor memory function ipsilateral to ictal onset and intact memory function on the contralateral side, which is rare.

Patients with an idiopathic (ie, genetic) epilepsy, such as benign rolandic epilepsy or benign childhood epilepsy with occipital paroxysms, are not surgical candidates since their seizures remit by adolescence.

Although patients with low intelligence quotient (IQ) scores are more likely to have more than 1 seizure localization and a less favorable outcome with surgery, they should not be excluded from preoperative evaluation. Many patients with a low IQ score have partial epilepsy that may allow a good seizure-free outcome with surgery.

In addition, patients with disabling tonic/atonic seizures (eg, drop attacks) and low IQ scores may benefit from corpus callosotomy as a palliative procedure; a trial of lamotrigine, felbamate, rufinamide; or both after failing vagal nerve stimulation.

Adults with diffuse (multilobar) polymicrogyria or subcortical band heterotopia are poor surgical candidates even when localization-related epilepsy is found.