General Epilepsy Surgery Outcomes

The earliest epilepsy surgeries performed in the late 1800s and early 1900s were targeted toward neocortical seizure foci.¹² Because of the limited utility of imaging and neurophysiologic modalities available to clinicians at that time, localization of the seizure focus was based primarily on anatomic features such as external signs of head trauma and underlying cortical encephalomalacia. Over time, improved imaging and electrophysiologic evaluation allowed delineation of medial temporal lobe epilepsy syndrome,^13,14 which is characterized by hippocampal sclerosis.^15,16 The work of several surgeons in developing and analyzing surgical approaches to this syndrome^17–21 has led to standardized approaches to epilepsy surgery for these patients. With these standardized approaches and distinct imaging characteristics of hippocampal sclerosis, many epilepsy surgery centers have proposed protocols that allow medial temporal lobe epilepsy surgery to be performed with a defined evaluation.²² In addition, epilepsy surgery for medial temporal lobe epilepsy with hippocampal sclerosis is associated with a high seizure-free outcome rate on the order of 53% to 84%, a rate that has not been matched by extratemporal neocortical resections.^23–26 Probably for these and other reasons discussed later, modern series evaluating epilepsy surgery show that extratemporal lobe epilepsy surgery is much less common than temporal lobe epilepsy surgery.

The most recent multicenter source of information regarding current practice patterns of major epilepsy centers comes from a seven-center prospective observational study of resective epilepsy surgery. In this study, which included 355 patients at 1-year follow-up and 339 patients at 2-year follow-up, only 12% of the epilepsy surgery patients underwent extratemporal surgery.^27,28 In a review of 708 epilepsy surgeries at a major epilepsy surgery center in Germany, 429 of which were therapeutic, 35% of the therapeutic surgeries were extratemporal and consisted of frontal resection in 14%, parietal resection in 2%, occipital resection in 3%, multilobar resection in less than 1%, callosotomy in 8%, and hemispherectomy in 8%.²⁹ Based on these and other studies, it is clear that temporal lobe epilepsy surgery, especially medial temporal lobectomy, is much more common than extratemporal lobe epilepsy surgery. There are several probable reasons for this discrepancy, including a higher incidence of medial temporal lobe epilepsy that is intractable, better seizure-free rates with medial temporal lobe resection than with extratemporal resection, and frequently, more diffuse or obscure extratemporal seizure foci necessitating a more tailored approach for extratemporal resection and requiring additional evaluation for determination of the exact seizure location, often with surgically implanted electrodes.

Outcomes of extratemporal lobe epilepsy surgery are based mostly on retrospective case series reviews. Only one study has provided class I evidence for epilepsy surgery; however, this study looked only at medial temporal lobe resections but demonstrated a clear benefit of early surgery over maximal medical therapy for medically intractable epilepsy.²⁵ The multicenter epilepsy surgery study just mentioned was a prospective case series that grouped neocortical temporal with general extratemporal cases.^27,28 This study found that patients undergoing neocortical resection had a lower seizure-free rate than did patients undergoing medial temporal lobe resection at 1-year follow-up (56% for neocortical versus 77% for medial temporal resection) and 2-year follow-up (50% for neocortical versus 68% for medial temporal resection). These differences were statistically significant only at the 1-year follow-up. Interestingly, the seizure relapse rate for patients who were initially seizure free after surgery was lower with neocortical resection than with medial temporal lobe resection in both the 1-year (4% for neocortical versus 24% for medial temporal resection) and 2-year studies (19% for neocortical versus 25% for medial temporal resection), again only statistically significant at 1 year. A recent review of the adult and pediatric epilepsy surgery literature reported 1-year or greater freedom from seizures in 53% to 84% of patients undergoing medial temporal lobe resection, in 66% to 100% of patients with dual pathology, including medial temporal sclerosis and temporal neocortical involvement, and in 36% to 76% of patients undergoing neocortical resection.²³

The American Academy of Neurology in association with the American Epilepsy Society and the American Association of Neurological Surgeons published a position paper that reviewed all the outcome studies for epilepsy surgery before the paper’s publication in 2003.³⁰ The Quality Standards Subcommittee identified 33 studies reporting seizure-free outcomes after epilepsy surgery, but just 1 class I study was included (the one discussed earlier) and 32 class IV studies. Only 8 of these studies described outcomes after neocortical resection, including both temporal and extratemporal locations, and all were class IV studies. This group was able to conclude that the benefits of medial temporal lobe resection surgery for medically intractable disabling complex partial seizures are greater than the benefits with continued maximal medical therapy and that the risks related to surgery are at least comparable to the risks associated with antiepileptic drugs. However, they were not able to draw similar conclusions from the neocortical resection studies because of the low number of studies, lack of class I evidence, and great variability among the different neocortical epilepsies and their surgeries based on the lobe involved. This position paper simply states that further studies are needed to determine the benefits of surgery for treating neocortical epilepsies.

Many studies have supported the overall benefit of successful epilepsy surgery in several areas of patients’ daily lives. These studies have not typically differentiated among patients according to the site of the resection, but it is likely that these quality-of-life studies are applicable to all patients who have successfully undergone epilepsy surgery, including surgery on the extratemporal lobe. The Multicenter Study of Epilepsy Surgery looked at some quality-of-life outcomes in their study patients. They found that quality of life was improved by 3 months after surgery regardless of seizure outcome but that scores at 1 and 2 years were statistically significantly lower in patients who were not seizure free than in those who were seizure free.²⁷ Others have found improved scores on quality-of-life measures after epilepsy surgery.³¹ A study of pediatric epilepsy surgery also found that quality of life was improved to a greater degree in children rendered seizure free after surgery than in those who continued to have seizures.³² Both anxiety and depression decreased significantly by 3 months after surgery, more in the seizure-free patients than in those who continued to have seizures, and scores remained improved at 1 and 2 years’ follow-up.³³ Full-scale IQ has also shown improvements in long-term follow-up after temporal, parietotemporal, and frontal epilepsy surgery.³⁴ At 2 years after surgery, 75.5% of patients said that they would definitely undergo surgery again, 79.1% thought that they had a very strong or strong overall positive impact from the surgery, but in only 7% was employment status improved.³⁵ Other studies have shown no improvement in employment status for epilepsy surgery patients in comparison to epilepsy patients without surgical treatment despite a greater decline in the former group in both monthly seizure frequency and antiepileptic medication intake.³¹ In a review of epilepsy surgery outcome papers, two studies evaluated multiple outcome factors in both temporal and extratemporal epilepsy surgery. Both studies found improvements in seizure outcome and decreased antiepileptic drug use, but only one of these studies found an improvement in social outcomes, and the one study that evaluated quality of life found no difference between surgical patients and controls.²⁴ This paper reported that in studies in which temporal lobe and extratemporal lobe resections were analyzed together, an average of 20% of patients were able to discontinue their antiepileptic drugs and an average of 41% were able to achieve monotherapy. Adverse outcomes associated with epilepsy surgery are relatively infrequent, with a series of 429 therapeutic epilepsy surgeries reporting no mortality, transient morbidity in 3%, and permanent morbidity in 2.3%.²⁹ Taken together, these results indicate a positive impact of epilepsy surgery on patients’ lives that outweighs the risks.

Lobar Distribution of Extratemporal Lobe Epilepsy

The distribution of lobes involved in various reported surgical series of extratemporal epilepsy may be skewed by the relative risk associated with ablative surgery and other factors that affect the decision for surgery, which differ for each lobe and may not represent the natural distribution of seizures in the general population of all patients with epilepsy. The surgical series are particularly instructive, however, in that they provide outcome data for cases in which relatively good localization has been achieved. The confidence in localization increases with invasive subdural or depth electrode monitoring, stimulation-induced replication of the seizure, and verification of seizure control postoperatively. Any or all of these criteria can be applied to the classification of surgical epilepsy patients but not nonsurgical patients.

Several case series have looked at epilepsy surgery in general or extratemporal epilepsy surgery in particular. These studies give an idea of the relative incidence of surgically remediable epilepsy in each of the lobes and the frequency of types of extratemporal epilepsy surgery. In adult epilepsy surgery series, extratemporal surgery represents 13% to 37% of operations.^36–38 Of the extratemporal resections, 60% to 84% were frontal, 4% to 20% were parietal, 3% to 20% were occipital, and 0% to 46% were multilobar.^36,38 Series combining adult and pediatric patients have reported extratemporal surgery in 12% to 44%.^{27–29,36,38–41} Of these, 33% to 64% were frontal, 7% to 14% were parietal, 2% to 23% were occipital, 0% to 37% were multilobar, 0% to 34% involved corpus callosotomy, and 0% to 22% involved hemispherectomy.^{29,36,38–44} The percentage of extratemporal epilepsy surgery cases in pediatric series varies from 32% to 83%.^{36,38,45–52} The pediatric extratemporal epilepsy surgeries included frontal resection in 18% to 73%, parietal in 0% to 18%, occipital in 0% to 9%, multilobar in 0% to 72%, corpus callosotomy in 0% to 50%, and hemispherectomy in 0% to 52%.^{36,38,45–52} Extratemporal resections, including corpus callosotomy and hemispherectomy, are more common in the pediatric population.

Parietal lobe epilepsy and occipital lobe epilepsy are rare in surgical series, a finding reported by several authors.^53,54 This very low incidence may be due to the relative resistance of these regions of the brain to the development of seizures, difficulty identifying these seizures because of the ambiguity of symptoms referable to seizures from these areas (especially in parietal lobe epilepsy), and the potential risk for permanent postoperative neurological deficit. The data also illustrate one of the characteristics of extratemporal lobe epilepsy that make it less amenable than temporal lobe epilepsy to surgical therapy—a higher incidence of widespread pathology involving more than one lobe.

Frontal Lobe Epilepsy

The frontal lobe is the largest lobe of the brain, and it encompasses several distinct anatomic-functional units, including the primary motor region; supplementary motor areas; language areas in the dominant frontal operculum; the frontal eye fields; part of the cingulate gyrus; a component of the limbic system; the orbitofrontal and ventromedial regions, which play a major role in the regulation of emotions; and the dorsolateral frontal region, which has major cognitive function, especially in executive functions and working memory. The regions responsible for clearly observable motor functions, the primary and supplementary motor areas, were the earliest to be classified anatomically^55,56 and the earliest targets for surgical treatment of epilepsy.^57–60 The clinical syndromes of primary and secondary motor cortex seizures are fairly well agreed on. Characterization of primary motor cortex seizures has remained essentially unaltered since the investigations of Penfield and Jasper¹²: focal clonic jerks without loss of consciousness if generalization does not occur. Supplementary motor cortex seizure morphology has been described by Ajmone-Marsan and Ralston⁶¹ and later investigators and has been characterized by more complex motor semiology, including combined movements of the extremities and head version.

Seizures in other regions of the frontal lobe have shown significant variability, which has resulted in difficulty characterizing classic frontal lobe epilepsy syndromes.⁶² Based on characterization of seizures in patients via depth electrode recording, Bancaud and Talairach⁶³ proposed that frontal lobe epilepsies be classified into (1) inferior frontal gyrus seizures in either the dominant or nondominant hemisphere with speech arrest, tonic or tonic-clonic contractions at the ipsilateral angle of the mouth, swallowing, salivation, gustatory hallucinations, vegetative signs, respiratory deficits, and possibly simple motor manifestations; (2) medial intermediate frontal seizures originating in the mesial frontal lobe, anterior to the supplementary motor cortex, superior to the cingulate gyrus, and posterior to the polar region, with frontal-type absence or complex motor seizures; (3) dorsolateral intermediate frontal seizures with contralateral deviation of the eyes followed by aversion of the head; (4) anterior cingulate gyrus seizures with intense fright, expressions of fear, and aggressive verbalizations and acts; (5) frontopolar seizures with dissociation from the environment, fixed eyes, immobility, flexion and turning of the head, falling, and tonic-clonic generalization; (6) orbitofrontal seizures with either olfactory illusions and hallucinations or vegetative symptoms, including cardiovascular, respiratory, or digestive system involvement; and (7) operculoinsular suprasylvian seizures with a variety of symptoms, including somatomotor involvement of the face and upper and lower limbs, disorders of verbal expression, contralateral oculocephalic deviation, dissociation from the environment, and postictal speech deficits. These syndromes are not universally accepted, with some authors expressing skepticism concerning the feasibility of anatomic localization by seizure semiology.⁶⁴

Localization of a single resectable focus in patients with frontal lobe epilepsy is typically difficult because of the propensity for extensive epileptogenic zones with multiple pathways that allow rapid ictal spread within the frontal lobe and to other lobes and the contralateral side.⁶⁵ Even a determination of lateralization can be difficult.⁶⁶ In the absence of a distinct focal structural lesion, localization of frontal lobe epilepsy foci almost always requires intracranial ictal recording with either subdural electrodes⁶⁵ or depth electrodes.⁶³ The findings from these evaluations, combined with the variety of anatomic imaging modalities available, are the basis on which surgical resection plans are made.

Because of the difficulty in localization of the seizure focus, it is not surprising that reports of a reduction in seizure episodes after surgery for frontal lobe epilepsy are fewer than for temporal lobe seizures. An additional factor may be increased restraint in performing cortical resection because of the relative involvement of critical functions in the frontal lobe.^67,68 Generous resections of the frontal lobe for epilepsy have been described, including total frontal lobectomy on the nondominant side and resection of the lateral sensorimotor cortex to nearly 3 cm above the sylvian fissure.^41,69,70 Any transient deficit from loss of a frontal eye field or the supplementary motor cortex is likely to resolve eventually. Special care is taken, however, to preserve speech areas in the frontal operculum of the dominant hemisphere, a region in which incurred deficits are likely to remain permanent. Care is exercised in resecting primary motor cortex, especially in the hand region or in the dominant hemisphere face region, because fine motor movement could be permanently affected.^12,57,71 Functional MRI (fMRI) can be helpful in delineating these motor areas preoperatively.⁷² A recently developed MRI technique, diffusion-tensor imaging (DTI), enables delineation of the subcortical motor pathways.⁷³

Frontal lobe epilepsy surgery is somewhat successful in achieving seizure-free outcomes. A systematic review of epilepsy surgery studies showed a seizure-free rate of 27% with frontal lobe surgery in all age groups.²⁴ In clinical series that looked at adult and pediatric patients together, the range of seizure-free outcomes for frontal lobe resections with an average of 1.8 to 16 years’ follow-up was 23% to 64%.^{36,40,42,43,74–78} For adults only, at an average follow-up of 1 to 5 years the seizure-free rate after frontal lobe epilepsy surgery was 25% to 50%.^36,40,79 In the pediatric frontal lobe epilepsy surgery population, after an average of 2 to 3.6 years the seizure-free rate varied widely from 9% to 75%.^{36,40,49,51,52} There is some variability in postoperative seizure control in this population, and it is not as good as with temporal lobe surgery, but with seizure-free rates possibly being 50% or better, these surgeries should still be considered a possible option. Lower seizure-free rates in series with longer average follow-up periods may indicate long-term relapse of seizures after frontal lobe epilepsy surgery (Fig. 61-1) (Case Study 61-1).

FIGURE 61-1 Seizure-free outcome based on average follow-up time for a number of reported frontal lobe epilepsy surgery case series. Lower seizure-free rates are seen in series with longer follow-up time.

Case Study 61-1

A 29-year-old woman underwent surgical evaluation for complex partial seizures dating back to 8 years of age, when Sturge-Weber syndrome was diagnosed. Her seizures were characterized by unresponsiveness, a blank stare, eyelid flutter, and rubbing her fingers together or rubbing her arms. These seizures would typically last 10 to 40 seconds; she had 15 to 25 seizures per month with occasional generalization. Her seizures had proved resistant to multiple pharmacologic agents. Her interictal EEG showed generalized frontotemporal slowing greater on the right than on the left, frontally predominant polyspike and wave discharges on the right, and isolated right temporal spike discharges. During EEG video telemetry, the patient’s seizures were characterized by wide-eyed staring, lifting of both arms above her head, rubbing her face, and moving the pillow and covers. EEG recorded during these episodes suggested frontal lobe seizures with the possibility of right frontal onset, but because of rapid propagation to the left frontal lobe, the possibility of bilateral or left frontal onset could not be ruled out. MRI and an angiogram during intracarotid sodium amobarbital testing showed a right frontal venous angioma (Fig. 61-E1A). The intracarotid sodium amobarbital test indicated that the left hemisphere was dominant for language. MEG results were consistent with widespread right anterior frontal and orbitofrontal activity, but with some activity in the left frontal region (Fig. 61-E1B).

FIGURE 61-E1 Case Study 61-1. A, Preoperative magnetic resonance imaging (MRI) shows a right frontal venous angioma in the region consistent with electroencephalographic localization of the seizure focus. B, A magnetoencephalogram superimposed on MRI shows epileptiform spikes localized to the frontopolar region, consistent with the area of the venous angioma seen on MRI in A. C, Postoperative MRI shows resection of the frontopolar region, including the venous angioma seen in A.

Based on the MEG findings, subdural grid and strip electrodes were surgically placed over the right frontal region and extended over the mesial and lateral frontal and the orbitofrontal regions. Ictal recordings from these electrodes showed onset in the anterior right frontal lobe. Electrocorticography at the time of resective surgery showed spiking activity at the anterior, mesial, and lateral extents of the craniotomy in the right frontal lobe. The tissue surrounding the venous angioma was resected en bloc and consisted mostly of the middle frontal gyrus with some additional cortical tissue at the frontal pole (Fig. 61-E1C). The resection included the anterior mesial frontal region. Somatosensory evoked potential mapping indicated a margin of safety between the posterior border of the resection and the rolandic cortex. Postresection electrocorticography showed no residual spiking activity, so the resection was carried no further. The pathology reading for the resected tissue was cortical dysplasia. Postoperatively, the patient was seizure free and remained so at 5-year follow-up.

This case of frontal polar involvement shows the usefulness of integrating several different modalities of evaluation, including semiology, EEG, MRI, MEG, subdural electrodes, and electrocorticography, when there is ambiguity concerning the seizure focus. Even if no single test is definitive, if all tests suggest the same region as the seizure focus, resective surgery is likely to be successful. This case also shows that a lesion seen on MRI cannot stand alone in the planning of resection for epilepsy because anatomic findings, such as the lesion identified preoperatively as venous angioma, may not be responsible for generation of the seizures; in this case, a large area of cortical dysplasia was resected after the MEG-guided grid study delineated the origin of the seizures.

Parietal Lobe Epilepsy

Parietal lobe epilepsy is relatively rare in series of epilepsy surgeries, perhaps because of an innate resistance to seizures in the region; the difficulty of localizing these seizures, which may be accompanied by symptoms referable to other lobes; and the reluctance of surgeons to resect tissue in this area. Seizures originating in the parietal area have not received as much investigation as frontal lobe or temporal lobe seizure syndromes. Parietal lobe seizures are typically characterized by somatosensory auras,^80,81 pain, paresthesias, vertigo, head and eye deviation, complex visual hallucinations, sensations of body movements, and actual complex movements of the extremities.^82–84 Except for the primary sensory cortex, there is a paucity of reported clinical and electrophysiologic correlates of seizure activity.⁸² Combined with a tendency for seizures to spread quickly to other regions of the brain,⁸¹ this paucity makes characterization and localization of the seizure focus based on semiology and scalp EEG difficult. Similar to frontal lobe epilepsy, demonstration of focal lesions on MRI or other imaging modalities is paramount in determining the appropriate surgical resection site.

Tailored resections of the parietal lobe do not usually include the primary somatosensory cortex because of the importance of this cortex in the production of skilled movement, as described by Penfield and Erickson.⁸⁵ Similar to the primary motor cortex, it has been suggested that the primary sensory cortex can be resected nearly 3 cm above the sylvian fissure in the nondominant and dominant hemispheres without significant deficit as long as the tongue, thumb, and lip areas are identified and preserved.⁷⁰ Any resection of the parietal lobe in the dominant hemisphere should remain above the intraparietal sulcus to prevent damage to the receptive language center.⁷⁰ In general, resections in the parietal dominant hemisphere should take into consideration the high probability of language deficits in the vicinity of the sylvian fissure and more superiorly in regions such as the angular gyrus. As a general rule, resections in the dominant parietal lobe should be carried out only after electrical stimulation mapping. Preoperative fMRI can be useful in providing a general idea of language representation in the region but cannot be relied on with respect to final surgical decision making. With significant resection of the parietal lobe, a contralateral inferior quadrantanopia should be expected.⁷⁰ The nondominant parietal lobe mediates important visuospatial functions. Large resections in this lobe cannot be undertaken without severe impairment in spatial cognition. We have performed focal resections in the nondominant parietal lobe when invasive monitoring showed a fairly circumscribed epileptogenic region (Case Study 61-2) but have resorted to multiple subpial transections when a large nondominant parietal territory was involved (see Case Study 61-4).

Case Study 61-2

An 18-year-old, right-handed man with seizures since the age of 6 was admitted for presurgical evaluation. His seizures were characterized by an aura of impending doom. Initially, his seizures consisted of 20 seconds of staring, but later the semiology changed to staring and dystonic posturing in a sitting position with outstretched arms that lasted 20 seconds, occasionally followed by left face and arm twitching with drooling. Postictally after the longer seizures, he exhibited left-hand weakness and dysarthric speech. On average, he had more than one seizure per day. The patient was prescribed several different antiepileptic medications and had a vagal nerve stimulator placed without alleviating the seizures. Video EEG monitoring showed rare right temporal spike and wave complexes with seizure activity originating in the right temporal region. MRI was unremarkable, but PET showed right inferior parietal hypometabolism (Fig. 61-E2A). MEG showed dipoles in the right posterior temporal and inferior parietal regions (Fig. 61-E2B). These findings suggested possible temporal and parietal lobe involvement; accordingly, subdural electrodes were surgically implanted to cover the frontotemporoparietal region, with strips covering the basal temporal lobe, anterior frontal lobe, posterior parietal region, and superior frontoparietal junction. These subdural electrodes localized the region of seizure onset to the right temporoparietal cortex, including the supramarginal gyrus and posterior superior temporal gyrus. The region of ictal onset corresponded well to the region of interictal abnormalities, but was smaller. For resective surgery, intraoperative evoked potentials and electrical stimulation mapping were used to define the rolandic cortex. Intraoperative electrocorticography confirmed localization of the epileptogenic region to the temporoparietal junction and was similar to the interictal localization obtained by long-term recordings with the subdural grid. Resection included the area surrounding the posterior edge of the sylvian fissure on the parietal and temporal sides. Subpial transections were performed at the anterior border of the resection cavity, where interictal abnormalities were observed close to the somatosensory cortex. Postresection electrocorticography showed slowing but no residual spikes. Pathologic evaluation of the resected tissue revealed no histologic abnormalities. The patient has remained seizure free for more than 1 year, although longer follow-up is required to assess the success of surgery.

FIGURE 61-E2 Case Study 61-2. A, Positron emission tomography (PET) shows a region of hypometabolism in the right lateral parietal lobe. B, A magnetoencephalogram superimposed on magnetic resonance imaging shows spiking activity focused in the right parietal lobe, consistent with the findings of PET in A.

This case shows the combined use of various modalities to plan surgery. Scalp EEG, MEG, and SPECT were useful in directing the placement of invasive electrodes. The resection included the entire ictal onset zone but not the entire zone of interictal abnormalities, which was larger. Multiple subpial transection was used to address the remaining region of interictal abnormalities, close to the rolandic cortex, and to avoid large resection in the nondominant parietal lobe.

Seizure control outcomes after parietal resection are reportedly slightly better than outcomes after frontal lobe resection. A systematic review of epilepsy surgery studies showed a seizure-free rate of 46% with parietal lobe surgery in all age groups.²⁴ Combined adult and pediatric series with an average follow-up of 2 to 20 years show seizure-free rates of 46% to 100% after parietal lobe resection.^{36,42,43,53,81,86,87} Adult series of parietal lobe epilepsy surgery with an average follow-up of 2 to 2.5 years report 90% to 100% seizure-free rates.^36,53 Pediatric series of parietal lobe epilepsy surgery with an average 2 to 3.6 years’ follow-up have reported 33% to 100% seizure-free rates.^36,49 Lower seizure-free rates in series with longer average follow-up periods after parietal lobe epilepsy surgery, as with frontal lobe surgery, may indicate long-term relapse over time (Fig. 61-2).

FIGURE 61-2 Seizure-free outcome based on average follow-up time for a number of parietal lobe epilepsy surgery series. Lower seizure-free rates are seen in the series with longer follow-up.

Occipital Lobe Epilepsy

Occipital lobe epilepsy is also rare. Clinically, occipital lobe epilepsy is characterized most often by visual auras.^54,88,89 These visual auras are elementary in nature and are described as lights, spots, or simple shapes that can be flashing or moving. Formed visual hallucinations are more suggestive of temporal lobe epilepsy, either primary or from spread of an occipital lobe seizure. Another clinical feature of occipital lobe epilepsy is episodic blindness,^88–91 which can involve half the visual field or the entire visual field. Other signs observed with occipital seizures include blinking and tonic or clonic eye deviation.^88,89 Occipital lobe seizures can spread rapidly to the temporal lobe, thus making localization of their onset difficult and occasionally leading to relatively ineffective temporal lobectomy.⁸⁹

Resective surgery for occipital lobe epilepsy is almost certainly going to lead to some degree of visual deficit,⁷⁰ and this procedure must be considered carefully by the patient and the surgical team. Impairment of sight could be more disabling than the seizures. It would be unfortunate for a patient to undergo surgical intervention for seizure control in anticipation of regaining a driver’s license simply to have the license denied because of a dense hemianopia. Many of these patients have preoperative visual field deficits, especially if there is a mass underlying the seizure disorder, and there is a risk in not controlling the seizures because permanent blindness has been described after recurrent, uncontrolled occipital seizures.⁹² Given the frequent spread of occipital lobe seizures to the temporal lobe, resections are sometimes designed to include some portion of the temporal lobe.^44,88,89

Postoperative seizure control outcomes for occipital lobe epilepsy patients have varied significantly among series. A systematic review of epilepsy surgery studies showed a seizure-free rate of 46% with occipital lobe surgery in patients of all ages.²⁴ In several series of adult and pediatric patients combined at an average follow-up of 2 to 17 years, seizure-free rates were 20% to 88%.^{36,42,43,81,87,88,93} Adult series alone showed a seizure-free rate of 46% to 100% after an average 2- to 6-year follow-up.^36,54 Pediatric occipital lobe epilepsy patients had a seizure-free rate of 0% with an average follow-up of 3.6 years.⁴⁹ Similar to frontal lobe resective surgery for epilepsy, seizure control outcomes for occipital lobe epilepsy surgery may suffer because of a more conservative approach to the extent of cortical resection. Interestingly, occipital lobe epilepsy surgery series do not show lower seizure-free rates in series with longer average follow-up periods (Fig. 61-3) (Case Study 61-3).

FIGURE 61-3 Seizure-free outcome based on average follow-up time for a number of occipital lobe epilepsy surgery case series.

Case Study 61-3

A 38-year-old man who has had seizures since the age of 13 initially had generalized tonic-clonic seizures that would occur at night, but 5 years before surgery, he began having seizures during the day in which he would become disoriented, experience vertigo and partial loss of awareness, and turn his head and eyes to the right. The generalized tonic-clonic seizures began with lifting of his right arm and lasted for 0.5 to 2 minutes. He typically had 5 to 10 focal seizures per day and 1 to 2 generalized seizures per month. On video EEG monitoring, the seizures appeared to begin in the left occipital region but almost instantly spread to the left parasagittal area. On MEG, he had dipoles not only in the mesial part of the left occipital lobe but also in bilateral central parietal areas (Fig. 61-E3A). MRI and SPECT were unremarkable, and PET showed no regions of focal hypometabolism. Neuropsychological testing indicated involvement of the nondominant hemisphere. Intracranial depth electrodes were stereotactically placed in the right occipital, right parietal, right parahippocampal gyrus, right supplementary motor, left occipital, left posterior parietal, left parietal, left parahippocampal gyrus, and left supplementary motor areas (Fig 61-E3B). Evaluation of seizures with these electrodes indicated a left frontal onset of the seizures. The depth electrodes were removed, and at a later time, frontal subdural grid and strip electrodes were placed through bilateral frontal craniotomies with extensive sampling of the left mesial frontal region. A left mesial frontal focus was delineated, and this region was resected with electrocorticographic guidance and somatosensory evoked potential and stimulation mapping of the motor cortex. The pathology was consistent with a cavernous hemangioma, a lesion that was not seen on any of the preoperative imaging studies. The patient is currently nearly seizure free 3 years after surgery.

FIGURE 61-E3 Case Study 61-3. A, A magnetoencephalogram superimposed on magnetic resonance imaging (MRI) shows bilateral parietal spikes, more on the left than the right. B, MRI shows placement of depth electrodes (MRI artifact exaggerates the size of the electrodes, which have a diameter of 1.3 mm).

This case shows the complexity involved in localizing an extratemporal seizure focus and illustrates how easily noninvasive tests can point to an incorrect seizure focus. In such cases, depth electrodes are used to identify a broad region of onset. When the region has been identified, extensive sampling can be achieved by focused subdural electrode arrays, which are used to identify the discrete epileptogenic region and map the region by electrical stimulation. In this case, depth electrodes in the parietal and occipital lobes and frontal subdural electrodes were essential in localizing the seizure focus to the frontal lobe rather than to the posterior target as suggested by noninvasive monitoring.

Multilobar Resection

Extratemporal lobe epilepsy may involve more than one lobe, particularly in pediatric patients,^50,94 and in such cases multilobar resection involving two or three lobes may be necessary. The more common multilobar resection patterns are frontal-temporal,⁴⁴ frontal-parietal, and temporal-occipital.⁵⁰ These resections are designed around preserving regions of important function, such as motor-sensory cortex and language areas. The range of seizure-free rates in combined adult and pediatric patients undergoing multilobar resection was 15% to 60% with an average of 2 to 11 years’ follow-up.^36,38,69,94 Adult series of multilobar resection showed a 25% to 50% seizure-free rate after 1 to 5 years.^38,79 Pediatric multilobar resection patients with an average of 2 to 5 years’ follow-up had a seizure-free rate of 22% to 100%.^36,38,46,51 The higher rates of success in children may be due to more aggressive removal of tissue than in adults because the greater plasticity in an undeveloped brain allows better functional recovery (Fig. 61-4).

FIGURE 61-4 Seizure-free outcome based on average follow-up time for a number of multilobar epilepsy surgery case series.

Hemispherectomy/Hemispherotomy

When the epileptogenic region is extensive enough to involve the cortex of one cerebral hemisphere entirely or almost entirely, removal or disconnection of the whole hemisphere may be in order, especially for pediatric patients.^95–97 The original procedure to accomplish this was anatomic hemispherectomy, in which the entire hemisphere neocortex was resected.⁹⁸ Functional hemispherectomy and hemispherotomy procedures were developed in an attempt to leave some of the brain tissue intact for structural purposes to alleviate the issues of cerebral shift, hemosiderosis, and large subdural hygromas. These procedures were designed to remove only the amount of tissue necessary to completely disconnect the hemisphere’s neocortex from other parts of the brain. Several variants of the procedure have been described.^99–103 Hemispherectomy and hemispherotomy are more commonly reserved for pediatric patients, who have the best opportunity for regaining neurological function over time. In the series by Eriksson and coworkers,³⁶ in which pediatric and adult epilepsy surgeries were compared, eight hemispherectomies were performed in the pediatric group and none in the adult group. The procedure has been used in adult patients with hemispheric atrophy and fixed unilateral motor deficit. Given the inevitable minimum neurological deficit of decreased fine motor control and the higher risk for surgical complications than with other epilepsy surgeries (16% in one series²⁹), these procedures should be considered only for severe epilepsy syndromes with full comprehension of the risks by the family and be performed only by a surgeon who is experienced in this specific type of surgery. Outcomes after hemispherectomy have been reported only in pediatric epilepsy series and range from 40% to 83% seizure-free rates after an average 1- to 5-year follow-up.^{36,48,51,104–107} Outcomes are somewhat dependent on the underlying pathology¹⁰⁵ and the method used for disconnection of the hemisphere.¹⁰⁸ Patients with hemispheric cortical dysplasia and Rasmussen’s encephalitis had the best postoperative seizure-free rates, 81% at 1 year and 60% to 62% at 5 years, followed by patients with infarct or ischemia, who had seizure-free rates of 75% at 1 year and 71% at 5 years, and hemimegalencephaly patients, who had the worst seizure-free rates, 57% at 1 year and 33% at 5 years.¹⁰⁵ Based on the technique used for hemispherectomy, the highest reoperation rate for recurrent seizures occurred after functional hemispherectomy, the longest hospital stay and highest rate of shunt requirement were associated with anatomic hemispherectomy, and the least blood loss and lowest complication rate were associated with the lateral hemispherotomy technique.¹⁰⁸

Other Surgical Techniques

Direct resection is not the only surgical procedure available to treat intractable extratemporal lobe epilepsy. In general, the other surgical methods can be classified as disconnection procedures and stimulation procedures.

Disconnection Procedures

Multiple Subpial Transection

Multiple subpial transection (MST) involves inserting a specially designed instrument through the pia at one side of a gyrus and transecting the cortical ribbon of that gyrus subpially to the other side of the gyrus.^109,110 Subpial transections are traditionally performed approximately every 5 to 10 mm along the length of each gyrus within a region of epileptogenicity. These transections are thought to divide the fibers connecting adjacent regions of the cortex while leaving the projection fibers in and out of the region intact. MST has shown some efficacy in eliminating or reducing seizures without compromising function of the cortical region. Although one series of pediatric epilepsy patients reported that all 20 had transient hemiparesis after MST of the primary motor cortex,¹¹¹ permanent motor deficit after MST of the primary sensory-motor cortex is not typically reported.^111,112 A very small series of MST in the language cortex suggests that MST in the posterior language cortex leads to language dysfunction that will improve over time but that MST in the frontoparietal language areas does not lead to language deficits.¹¹³ In contrast, a meta-analysis of MST showed a 22% incidence of new neurological deficits after MST,¹¹⁴ so the possibility of neurological deficits should be considered carefully and discussed with the patient and family. Subpial transections are typically used in regions of cortex with critical functions (Case Study 61-4) and can be performed in combination with tissue resection techniques.^109,110

Case Study 61-4

A 36-year-old man had his first seizure at the age of 32. The seizures were typically characterized by visual hallucinations of the number 103 followed by loss of awareness and cheering and clapping behavior and then postictal confusion and sleepiness. He was having approximately two to three seizures per week, only a third of which began with a visual aura. The seizures had proved resistant to pharmacotherapy with four different medications in various combinations. Interictal scalp EEG showed right parietal spike and slow wave activity, and video EEG monitoring captured three complex partial seizures that localized to the right central parietal region and one rapidly secondarily generalized seizure in which his head turned to the left before generalization. MRI findings were normal, but PET showed mild right temporal hypometabolism. MEG showed clusters of spike activity over the posterior temporal and parietal regions of the right hemisphere (Fig. 61-E4). The patient had subdural grid electrodes surgically implanted through a right temporal/parietal/occipital craniotomy. Activity recorded from these subdural electrodes showed diffuse onset of the seizures from the right parietal and temporal regions. After extraoperative monitoring, the patient returned to the operating room for surgical treatment of the epilepsy. Stimulation motor mapping and electrocorticography were performed. A biopsy specimen taken from a region well behind the rolandic cortex showed no abnormality on pathologic evaluation. Multiple subpial transection (MST) guided by electrocorticography and the information obtained from the subdural grids before the operation was carried out during surgery. MST was used because the seizure focus overlapped regions of rolandic cortex and extensive regions of the nondominant parietal lobe.

FIGURE 61-E4 Case Study 61-4. A magnetoencephalogram superimposed on magnetic resonance imaging shows spikes primarily localized to the right lateral parietal and posterior temporal lobes but with some spikes in the left parietal lobe. LD1, LD2, and LD5 denote dipoles related to the somatosensory representation of the left-hand digits.

Electrocorticographic findings were much improved after MST, but residual spikes remained, so further subpial transections were performed. The final electrocorticographic findings were improved further. The patient’s postoperative course was unremarkable, and in a 4-year follow-up period, he remained seizure free. This case shows the potential efficacy of subpial transections in alleviating seizures located in cortex that is not amenable to wide surgical resection.

The most comprehensive analysis of outcomes after MST is a meta-analysis of six epilepsy surgery programs with a total of 211 patients.¹¹⁴ This study classified patients with excellent outcome as those with a greater than 95% reduction in seizures. In patients who underwent MST in addition to resection, 87% with generalized seizures, 68% with complex partial seizures, and 68% with simple partial seizures had excellent outcomes. In patients undergoing MST alone, 71% of generalized, 62% of complex partial, and 63% of simple partial epilepsy patients had excellent outcomes. However, only 16% of patients became seizure free after MST in another systematic review of adult and pediatric epilepsy surgery.²⁴ Case series of outcomes after MST, with or without cortical resection, vary in their reported success rates, with 0% to 50% being seizure free, 30% to 62% having a significant reduction in seizures,^{107,110,111,115–120} and one report of relapse of seizures between 2 and 5 years after surgery in 18.5% of patients.¹²⁰

Corpus Callosum Section (Corpus Callosotomy)

In corpus callosotomy, a portion of the corpus callosum, usually the anterior two thirds, is divided in an attempt to eliminate most connections from one cerebral hemisphere to the other, in this manner curtailing sudden generalization by preventing spread of seizures from one side of the brain to the other.¹²¹ Corpus callosotomy is indicated in patients with drop attacks and generalized epilepsy in which the seizures spread throughout the brain so quickly that they cause complete flaccid paralysis and falling, which poses considerable risk for serious injury. These patients are severely affected by the resulting repeated head trauma. Corpus callosotomy rarely eliminates seizures (6% to 19% of patients)^122–125 but is primarily designed to change the character of the seizures to eliminate drop attacks. A systematic review of the epilepsy surgery literature showed that only 35% of corpus callosotomy patients became free of their most disabling seizures.²⁴ However, 33% to 92% of patients had significant improvement in their seizures after corpus callosotomy,^{107,121–139} depending on the extent of the callosotomy (partial versus total section)^138–140 and type of seizures.¹²²

Corpus callosotomy can lead to disconnection syndromes that may be disabling.^121,141 The reported incidence of disabling disconnection syndromes varies from series to series, with one reporting 4 of 20 having a disconnection syndrome,¹³⁴ one reporting only transient disconnection syndrome in 57% of patients,¹³⁵ one reporting no severe neuropsychological deficits in 15 patients,¹³³ and one reporting neuropsychological benefits in all 25 patients in the series after corpus callosotomy.¹⁴² Because of concern about surgical morbidity and disconnection syndromes, many epilepsy surgery programs consider vagal nerve stimulator implantation the first-line surgical therapy for severe generalized seizures before considering corpus callosum section.¹²⁶