15
CHAPTER
Epilepsy Surgery Evaluation
Tung T. Tran
Consideration of epilepsy surgery is important in the treatment of patients with persistent epileptic seizures despite optimal medication management. About half of all patients with epilepsy will have successful seizure control with their first appropriately chosen and well-tolerated antiepileptic drug (AED). Another 10% to 15% will find success with a second appropriate AED. Unfortunately, after two good AED trials, less than 5% will achieve success with the addition of more medications (1). Those patients whose seizures are not well controlled despite multiple trials of AEDs, traditionally referred to as having intractable epilepsy, are also described as having drug-resistant epilepsy (DRE). The International League Against Epilepsy’s (ILAE) definition of DRE is “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” (2). Risk factors for DRE include poor response to the first AED and signs of higher epilepsy burden, including frequent seizures, a long duration of epilepsy, and a history of status epilepticus.
There are multiple reasons why a patient treated with AEDs continues to have seizures. One of them is the aforementioned DRE. The others are inappropriate or poorly tolerated AED treatment and misdiagnosis. In any of these cases, if seizures are not controlled, the patient should be referred to an epilepsy specialist. Diagnostic tests, such as an epilepsy monitoring unit (EMU) admission, may have great impact on a patient’s quality of life. Before a patient has any sort of surgical intervention, their seizures must be thoroughly evaluated.
Partial seizures are generally more difficult to control with medications when compared to idiopathic generalized epilepsy. Fortunately, partial seizures are sometimes amenable to surgical resections, while generalized seizures are not. The majority of this chapter describes the use of surgical resection for partial seizures. However, there are surgical treatment options for generalized seizures, including corpus callosotomies and vagus nerve stimulators, to be discussed in Chapters 30 and 31. Callosotomies, in particular, can be very useful for patients who fall and injure themselves. These patients often require helmets because of the severity of their falls. A callosotomy can reduce the injuries and falls, if not necessarily the seizures. Therefore, an epilepsy center evaluation should be considered for all patients with any persistent debilitating seizures.
REFERRAL FOR EPILEPSY SURGERY
All patients with debilitating partial seizures and DRE should be referred for epilepsy surgery work-up because DRE implies life-long impairments. In these patients, epilepsy surgery may offer a possible cure. The goal of any patient with epilepsy is “no seizures, no side effects.”
Benefits
Patients with DRE may have a multitude of complications. Persistent seizures mean increased risk of injuries from falls. Quality of life is often affected because driving, employment, social isolation, and stigma may be associated with intractable seizures. The continued use of AEDs may cause cognitive and mood impairments, as well as other long-term side effects associated with chronic medication use, such as impaired bone health. DRE patients are often on multiple medications, which can lead to toxicity effects. Furthermore, persistent seizures are related to increased mortality, whether due to status epilepticus or sudden unexpected death. These topics are discussed in chapters located in Part IV of this book and are the emphasis and motivation behind controlling seizures. This is especially important because many patients with DRE can potentially be cured.
While surgery is often considered by patients and practitioners as a drastic measure, studies show that, for select DRE patients, the long-term benefits and tolerability of surgery are significantly better than prolonged unsuccessful treatment of DRE with AEDs. About two out of every three patients who undergo temporal lobe resections become seizure-free, with an additional percentage having significant decrease in their seizure frequency. Many of these patients, after surgery, can drive, work, and live lives relatively free from the persistent fear that a seizure can occur at any time. Furthermore, some are weaned off of all seizure medications. These patients avoid the long-term side effects of chronic AED use. Also, from an economic standpoint, the up-front cost of surgery is much cheaper than the cost of years of medications, emergency room visits, and loss of productive time.
Additional advantages of epilepsy surgery will be discussed in a later section on positive outcomes, including the persistent improvements in quality of life. The number of patients with drug-resistant temporal lobe epilepsy one needs to treat with epilepsy surgery before demonstrating an advantage over conservative treatment is only two (3). This demonstrates the high effectiveness of epilepsy surgery and why it is considered the standard of care for patients with DRE.
Whom to Refer
The American Academy of Neurology (AAN) Clinical Practice Guidelines states that any patient with disabling complex partial seizures who has failed appropriate AED treatment be considered for epilepsy surgery referral (4). In fact, anyone who continues to have partial seizures that affect their quality of life, despite their practitioner’s best efforts, should be offered an evaluation for possible surgery. This may include those patients with only simple partial seizures, as frequent auras can sometimes have an impact on a person’s life. Similarly, even one major seizure, however rare, can cause significant injury if occurring at an inopportune time. Thus, there is no minimal criterion for seizure intensity or frequency with regard to epilepsy surgery, as long as their epilepsy continues to have a negative impact on a patient’s life.
Furthermore, referrals to epilepsy centers are not restricted by seizure type. While surgical resection is restricted to patients with partial seizures, there are other advanced techniques, discussed in Part III of this book, which may be very helpful. Also, sometimes secondarily generalized partial seizures may mimic primary generalized seizures. Prolonged video electroencephalography (vEEG) monitoring may help differentiate between the two. If a practitioner is unable to satisfactorily control seizures, that is enough reason for an epilepsy center referral.
The Canadian Appropriateness Study of Epilepsy Surgery (CASES) group has placed a free, simplified questionnaire online at www.epilepsycases.com to evaluate the appropriateness of epilepsy surgery referral. It is based on seizure characteristics such as type, severity, frequency, and first onset, along with treatment success, side effects, and prior diagnostic results.
For the purpose of encouraging patients to be seen at an epilepsy center, it may be helpful to know that certain characteristics correlate with patients having better outcomes with surgery. One is the identification of a focal brain lesion. Concordant imaging and EEG suggest a higher chance of seizure freedom after surgery. That said, sometimes lesions are not identified initially. With additional work-up, however, surgery can often be successfully performed when a lesion was not initially seen (5). Therefore, the absence of a focal brain lesion should not prevent one from being referred to an epilepsy center.
Other factors influence success of surgery. Along with imaging abnormalities, EEG abnormalities correlate to more positive outcomes. This suggests that better localization of epileptogenic regions lead to better results. On the other hand, not surprisingly, more severe preoperative seizures relate to worse outcomes. Patients with secondarily generalized seizures have relatively reduced remission rates than those without. Patients with lower IQ and comorbid psychiatric disease are also thought to have higher seizure recurrence rates after surgery. Another important predictor of surgery success is duration since the time of onset, with longer durations having worse outcomes. However, no one poor outcome risk factor necessarily precludes a patient from getting an epilepsy surgery evaluation.
When to Refer
If someone with DRE is a good surgical candidate, they should be referred for epilepsy surgery as soon as possible. In any case, after two appropriate AED treatment trials, there is little evidence that postponing surgery is helpful. Some patients with DRE may go months without a seizure, but the majority of them are very likely to have another seizure at some point (6). Each seizure puts the patient at increased risk of brain and possibly bodily injury. Furthermore, multiple studies have suggested that the outcome of epilepsy surgery is dependent on the time since seizure onset; the less time to epilepsy surgery, generally the better prognosis after surgery. Even for patients with disabling mesial temporal lobe epilepsy for not more than 2 years, a recent study showed that surgical therapy is better than medical therapy in terms of both seizure freedom and quality of life (7). Based on all these factors, once a patient is thought to have DRE, discussion should be initiated with the patient regarding an epilepsy center referral.
Obstacles to Referral
Despite a consensus that epilepsy surgery should be considered for patients with DRE, a large percentage of eligible patients are not referred to an epilepsy surgery center, and few are referred in a timely manner. Even after the release of clinical guidelines from the AAN in 2003, the average time from seizure-onset to surgical evaluation has remained about 18 years (8). There is some evidence that epilepsy referrals occur less in minorities and those without private insurance (9). In addition, the number of epilepsy surgeries has not dramatically increased.
Where to Refer
The National Association of Epilepsy Centers (NAEC) defines guidelines and standards of care for epilepsy centers. Level 3 epilepsy centers provide basic noninvasive epilepsy care, sufficient for the Phase I monitoring described in the next section. NAEC Level 4 centers offer a more complete range of epilepsy treatment options, including Phase II monitoring and epilepsy surgery.
SURGICAL EVALUATION PROCESS
The goal of surgical intervention is to remove all epileptogenic brain tissue, thereby hopefully eliminating all seizures. The purpose of the presurgical workup is to both identify the epileptogenic region and assess the risk of removing it. Identifying the seizure origin involves multiple electrophysiological and neuroimaging methods described later and in other chapters. Assessing the risk of surgical resection involves measuring the function of relevant regions of the brain.
Ideally, the epileptogenic region is a stable, single focus, which can be removed without any deficit in the patient’s function. A fluctuating lesion, by contrast, may suggest a more systemic etiology that could recur despite resection. Similarly, multiple sources of seizures, particularly if coming from both hemispheres independently, would also imply a poor surgical outcome. If a patient does not have a resectable epileptogenic region, other advanced therapies must be considered.
Phase I Monitoring
Surgical work-up is commonly divided into phases. Phase I is noninvasive and involves electrophysiological monitoring, neuroimaging, and functional testing. Sometimes, phase I testing is conclusive and sufficient, and the patient can proceed directly to surgical resection.
The following subsections summarize some of the techniques used to noninvasively evaluate for surgery. More details of each can be found in their respective chapters in Part II of this book.
Electroencephalography and Neuroimaging
Epilepsy resection surgery workup includes prolonged inpatient vEEG monitoring in an EMU. The goal of monitoring is to capture as many seizures as needed in order to localize their site of origin.
Multiple seizures should be recorded, because a patient may have multifocal epilepsy, which is less amendable to surgical resection. Capturing only one or two seizures may miss different seizure foci. Also, seizures that occur close to each other in time may be the result of one persistent seizure event, and thus they should not be considered independent of each other. In other words, seizures should be separated by several hours.
Ideally, for confident ictal localization, at least four to five independent seizures with the same EEG pattern should be recorded. If the EEG pattern is different, there may be more than one site of seizure onset. Similarly, if the patient describes multiple seizure types, each type should be recorded, with the hope that they all arise from only one epileptogenic region.
Sometimes insufficient ictal recordings are captured. If this is the case, ancillary data are used. Interictal discharges are often suggestive of an epileptogenic zone. For example, frequent spike-and-slow-wave discharges over right anterior temporal head regions, without discharges anywhere else, are suggestive of right temporal seizures. In addition, the semiology of seizures often reflects its origin, along with neuroimaging results, such as MRI, PET, SPECT, and functional MRI (fMRI). Of particular use, unilateral mesial temporal sclerosis when identified on MRI is particularly suggestive of a good surgical outcome.
Concordant imaging, semiology, and interictal discharges greatly increase the confidence of seizure localization when combined with ictal EEG. On the other hand, if results are not concordant, further discussion and possibly additional testing should be done before proceeding to surgery.
Neuropsychological Testing
Neuropsychological testing involves a battery of cognitive tests used to quantify psychological function associated with neuro-anatomical structures and pathways. Preoperatively, these provide two benefits with regard to epilepsy surgery evaluation. First, they can demonstrate that a particular brain region is impaired, which may suggest brain abnormality and possible seizure origin, much like imaging studies. Secondly, neuropsychological testing evaluates brain function, and thus it could possibly predict the cognitive deficits that would occur from removal of the examined brain regions. Cognitive skills testing include memory, language, executive function, and visual-spatial perception. Impaired language skills, for example, might suggest a dominant hemisphere dysfunction. Presurgical testing is important for comparison to postsurgical testing, particularly in regard to measuring outcome. Neuropsychological testing is often an all-day process, as the battery of tests can be time consuming.
Wada Test
The Wada test involves temporarily impairing one hemisphere of the brain in order to assess language and memory dependency on that side. This is done, typically, by injection of amobarbital by a neuroradiologist into the internal carotid artery via the femoral artery, one side at a time. This essentially puts one half of the brain “to sleep.” Once this happens, the patient is usually paralyzed on the contralateral side, sometimes with a visual field deficit. A series of brief tests are performed at this time in order to assess how memory and language are affected. The effects of amobarbital usually wear off quickly, so after about 30 minutes testing on the other side can be performed. EEG is often recorded during the procedure to document effects on brain wave activity. EEG and motor strength are often used to measure effectiveness of cortical anesthesia.
If injection into the left carotid artery greatly disrupts language and memory, but sequential injection into the right carotid artery produces minimal abnormality, then the patient likely depends on the left hemisphere for language and memory. They are likely left-hemisphere dominant, and memory is poorly supported by the right hippocampus, but well supported by the left. Resection of the right may not produce any significant cognitive deficit.
Wada test is generally well tolerated. Complications may arise from application of the medication, but serious adverse effects, such as stroke, bleeding, or infection, are rare, occurring in less than 1% of patients.
Magnetoencephalography and Other Tests
Unlike EEGs, which measure electrical activity on the scalp, magnetoencephalography (MEG) measures magnetic fields around the scalp. These magnetic fields are generated by electrical activity in the brain but are less distorted by structures like the skull and muscles. Therefore, magnetic fields are good at detecting electrophysiological activity deep in the brain such as in cortical sulci. MEG electrophysiological testing is often complimentary to EEG testing.
Unfortunately, MEG machines are much more unwieldy and much less available than EEGs. They are large machines in specially isolated rooms. MEG, unlike EEG, MRI, and neuropsychological testing, is not considered standard of care in epilepsy surgery evaluation. However, epilepsy centers may choose to use them if the additional information they provide can affect treatment.
Along with MEG, there are several new diagnostic techniques being developed for the evaluation of patients for epilepsy surgery. Some involve improved sensitivities of current tests, such as increased resolution of MRI. Others look at different modalities, such as the connectivity of white-matter tracts. Still other techniques incorporate multiple modalities to improve outcome. Given the ever-changing nature of seizure diagnosis and treatment, it may be warranted to reevaluate a patient previously thought not to be eligible for surgical treatment every few years.
Phase II Monitoring
Sometimes noninvasive Phase I monitoring is not conclusive, and the patient may or may not be a good surgical candidate. In these cases, the patient will need to undergo intracranial monitoring before a resection can be recommended. This Phase II monitoring involves placement of electrodes directly on brain tissue in the operating room (OR). This is done after a plan is discussed between the epileptologist and neurosurgeon regarding where the electrodes should be placed. This is important because intracranial electrodes cannot cover the entire brain. Therefore, seizures will only be detected where the electrodes are placed, and seizures cannot be found where no one is looking. In other words, there is a selection bias. It is the responsibility of the care team to appropriately narrow the regions of interest. Sometimes this requires multiple stages of intracranial monitoring.
Intracranial Electrodes
There are two standard intracranial electrode types: subdural and depth electrodes. Subdural electrodes are thin, flat disks, usually arranged in linear strips or grids, placed on the surface of the brain under the dura. They are usually positioned on the cortex where epileptogenic activity is suspected to occur. The goal is to cover all epileptogenic cortex as well as neighboring eloquent tissue. Ideally, by the end of Phase II monitoring, the epilepsy team is able draw the boundaries of both tissue types and plans a resection to include all tissue of seizure origin, without removing important brain tissue. Common placements of subdural electrodes include the lateral surfaces of frontal, temporal, and parietal lobes, inferior surface of the temporal lobe, and sometimes the medial inter-hemispheric surface.
Electrodes may also be embedded in thin flexible wires that pierce the brain. These are referred to as depth electrodes, because they are typically used to measure EEG activity deep in the brain. A common example is a depth electrode directed toward the hippocampus, a typical source for seizures. Depth electrodes are placed using stereotactic surgery, guided by coordinates determined by MRI and X-ray. They are inserted through small burr-holes, which involves less trauma to the skull and surrounding tissue than a craniotomy, which is needed to place grid electrodes. Compared to subdural electrodes they are better tolerated by the patient, and complications are generally lower. Placement of depth electrodes bilaterally can be used to lateralize the hemisphere of seizure onset. However, compared to subdural electrodes, they usually cover less cortex and involve penetrating the brain. They are used to identify deep borders of epileptogenic cortex. Sometimes, a combination of subdural and depth electrodes is used.
Disadvantages of Intracranial Monitoring
Regardless of the type of electrodes used, surgical intervention is required to implant the electrodes and thus has the associated risks of general anesthesia, bleeding, and infections. The number of electrodes placed within the skull is limited by the volume the electrodes occupy and the amount of brain exposure that is required. In general, more electrodes require more exposure and a higher risk of complications.
Advantages of Intracranial Monitoring
The primary advantage of intracranial monitoring is increased proximity to epileptogenic tissue, and thus increased resolution of seizure-origin localization. That resolution depends on placement of the intracranial electrodes. Intracranial monitoring allows this to be customized based on prior monitoring results. It is generally limited by anatomical consideration.
Another advantage of having electrodes adjacent to the cortex is better detection of high-frequency (> 80 Hz) oscillations. There is some evidence that removal of all regions demonstrating these higher frequencies, which may be beyond the usually determined epileptogenic zone, correlates with better surgical outcome. Along with some of the advanced techniques mentioned previously, these new techniques bring hope that epilepsy surgery will continue to improve. However, more work needs be done to prove a correlation between better technology and better outcome.
SURGICAL RESECTION
Once it is determined that a patient is a good surgical candidate, the treating health care team, including epileptologist and neurosurgeon, should discuss the options available with the patient. This includes a review of all prior work-up and their implications, resection plan, expected outcomes, and possible complications. The patient should demonstrate understanding of the plan and its possible consequences. If Phase 2 monitoring is required, this discussion should take place before intracranial electrodes are removed because often resection occurs in conjunction with removal of electrodes.
Once a plan for surgery is in place and the patient is informed and comfortable with it, surgery should be scheduled as soon as the patient is ready. In some institutions, cortical mapping may occur in the operating room before resection, in order to confirm that important eloquent brain tissue is spared. However, if adequate preoperative testing was done, resection is performed without additional mapping.
Complications
Epilepsy surgery puts patients at risk of the typical neurosurgical complications, including stroke, hemorrhage, and infections. However, the most common concerns about removing brain tissue are the associated cognitive deficits. Obviously, the presurgical evaluations discussed earlier provide some guidance on potential risks. Even so, despite Wada testing, neuropsychological testing, and language mapping, up to 40% of patients, after dominant lobe resections, experience some difficulty with language (10). Thus, dominant anterior temporal resections are usually more conservative, with the posterior margin of resection not being as far back as nondominant anterior temporal lobe resections.
Another functional deficit that may arise from temporal lobe resections is a superior quadrant visual field defect. Although this defect occurs in about half of such cases, the extent of this deficit is variable and often does not affect daily functions. Other less common complications include nerve palsies and hemiparesis. Fatality is extremely rare.
Despite the associated risks of an epilepsy surgery resection, the positive outcomes of surgery generally compensate for its potential complications, such that overall quality of life improves afterward.
Positive Outcomes
The commonly used Engel epilepsy surgery outcome scale divides outcomes into four categories. Classes I and II mean seizure free and rare disabling seizures, respectively. These are the best outcomes. Class III indicates worthwhile improvement, while class IV suggests no worthwhile improvement. Another system for classifying seizure outcomes after surgery was devised at Duke University. The Duke system grades patients from Class 1 to Class 3. Class 1 includes patients who are seizure free or have only auras. Class 2 patients have 10 or fewer seizures per year, and Class 3 patients have more than 10 seizures per year. The Engel system is used more often.
There has been one completed randomized control trial and multiple other studies showing that the percentage of patients seizure free, ie, Engel class I, after temporal lobe surgery is about two in three (3). The success of epilepsy surgery often persists for the long term. Most patients, about 75%, who are seizure free after 2 years remain seizure-free after 15 years, and this holds especially true for patients who are seizure free for 5 years (11). While about half of patients who have undergone an anterior temporal lobectomy are seizure free, an additional 30% achieve intermittent seizure control. Only 20% never achieve any measure of seizure control.
While the majority of epilepsy surgeries are standard temporal lobectomies and most outcome studies are based upon these surgeries, extratemporal epilepsy is often also successfully treated with neocortical resection. While patients undergoing a temporal lobectomy have about a 67% chance of seizure-free outcome, patients undergoing extratemporal surgery enjoy a less than 50% chance of seizure freedom.
As important as seizure control is on a patient’s well-being, their quality of life after surgery should also be considered. This is helped by the combination of AED reduction, or even elimination, with reduced seizures and their effects. These effects include improved independence, driving capability, employment opportunity, and overall social and lifestyle options. There is evidence that even mental health status can improve after surgery. Furthermore, the mortality risk after surgery is generally lowered and the vast majority of patients who undergo surgery say that they would repeat the process (12).
Characteristics of Success/Failure
As discussed earlier, success with regard to seizure control depends in part on the location of the epileptogenic zone. Temporal lobe resections are more successful than frontal lobe resections. This partially correlates with the difficulty of completely defining the extent of some epileptogenic zones, as better identification ensures resection of all epileptogenic tissue. Temporal lobe epilepsy surgeries are often successful because the epileptogenic zone is often isolated to medial temporal structures.
Preoperative factors that influence surgical outcome were discussed previously. In general, a better baseline with regard to both seizure control and cognitive and mental health, along with an identified MRI abnormality, predict better postoperative quality of life.
Postoperative evaluations also may help predict surgery success. Not surprisingly, a lack of deficits and seizures postoperatively predicts better long-term outcome. More specifically, recurrent seizures in the first month to year after surgery predict worse long-term outcome (11). There is some debate whether postoperative EEGs showing interictal epileptiform discharges also portend a worse result.
POSTSURGICAL CARE
Withdrawal of AED
Epilepsy surgery is supposed to improve seizure control. However, particularly for the more severe cases, expectations of a cure should be tempered. The hope is to at least reduce seizure burden and also possibly reduce treatment burden by using less AEDs. If there are no seizures, then there is even the possibility of totally eliminating all AEDs.
Immediately after a resection, there is the risk of seizures due to the brain inflammation and trauma associated with surgery. AED are usually continued for at least 1 year after surgery. If a patient is seizure free during that period, then AEDs may be decreased and eventually withdrawn permanently. Ideally, epilepsy surgery reduces seizure risk enough that AED are no longer needed, but that is certainly not always achieved. There are risks to withdrawing AEDs completely, and these must be discussed with the patient.
While there is a correlation between how early AED are reduced and the likelihood of seizure recurrence, there is no clear evidence that early AED reduction predicts long-term seizure freedom in temporal lobe resections (13). In other words, reducing AED treatment earlier only reveals persistent postoperative epilepsy earlier, but it does not seem to influence long-term prognosis. This is probably particularly true in cases where seizure freedom is more likely, but is less true when the probability of seizure freedom is less. However, if medication withdrawal is being considered, there may be evidence that doing this early is not harmful.
Repeat Surgery
Sometimes surgery fails because a patient’s epilepsy is actually not amendable to surgery, despite the presurgical estimation of success. Even so, sometimes, the potential benefits and hope of success justified the attempt. This should always be made in conjunction with the patient’s full understanding and agreement.
However, occasionally, surgery fails because of technical limitations. For example, there may be complications that arise during surgery. The neurosurgeon may determine that a more extensive resection is not worth the risk. Resections may be incomplete because diagnostic testing suggests that a wider margin would cause functional impairments. In these cases, the patient may be a repeat surgery candidate. As medical technology improves, if any patient continues to have seizures, they should continue to follow-up or revisit an epilepsy center at least every few years.
