A pediatrician’s guide to epilepsy surgery

Abstract

Surgical intervention for epilepsy emerged in the second half of the 20th century as an important option for pediatric patients with medically refractory epilepsy. Both the number of patients undergoing epilepsy surgery and the available surgical procedures for epilepsy have expanded in the last 3 decades, and now range from surgical resection to neuromodulatory device placement^1,2 Studies showing that many patients who would be excellent candidates for surgery are still not being offered appropriate interventions have prompted an interest in ensuring that all providers who see patients with epilepsy are aware of the options for epilepsy surgery to facilitate earlier referrals when medications have not been effective³ In this article, we will introduce the pediatrician to the process involved in determining epilepsy surgery candidacy and to surgical outcomes, with the goal of empowering pediatric providers to refer their medically refractory epilepsy patients to a pediatric epilepsy center.

Identifying patients for consideration for epilepsy surgery

Today, there are many surgical options even for patients who would not have been considered traditional resective surgical candidates. Pediatricians should recognize that any patient with epilepsy who continues having seizures despite adequate trials of two medications should be referred early on to a level 4 epilepsy center who can evaluate the patient for candidacy for epilepsy surgery (Fig 1), regardless of seizure type.^4–6 The National Association of Epilepsy Centers offers an online search function for epilepsy centers in all regions of the United States (www.naec-epilepsy.org) and can assist the pediatrician in finding a nearby level 4 epilepsy center.

Fig. 1. Flow diagram of epilepsy surgery workup and treatment options:

Any patient with epilepsy who continues to have seizures despite adequate trials of 2 or more medications (orange oval) is referred for evaluation at an epilepsy center. Some of these patients may already have known multifocal or generalized epilepsy (red oval), others may require further evaluation to determine their epilepsy type and/or localization. During the workup, there are investigations that target localization of seizure onset(s) (green box), the function of specific brain regions (blue box), or which can have dual function (intersection of blue and green boxes). After these studies, patients should be categorized more clearly into multifocal or generalized (red oval) or focal with 1–2 foci (yellow oval). All patients could be considered for nonspecific treatment strategies (left grey box), while patients with focal epilepsy may be considered for resection/ablation versus focal neurostimulation, depending on the functional status of the region(s) of seizure onset, and specific cases may qualify for hemispherotomy (right grey box). LITT=laser intersitial thermal therapy; RFTC=radio frequency thermal coagulation; HIFU=high intensity focused ultrasound.

Workup for epilepsy surgery

In general, the studies performed during evaluation for epilepsy surgery serve two purposes, either to elucidate the localization(s) of seizure onset or to identify the function of brain regions. Different tests may be needed depending on the seizures an individual has.⁷ In addition to a dedicated EEG to evaluate seizure type(s), there are several other tests that can evaluate seizure onset or brain function that your patient may receive during the workup:

• Phase 1 electroencephalography (EEG) study: This test is designed to capture the seizures of interest on scalp EEG. Most patients will require this type of study during evaluation. Ideally several seizures are captured to characterize consistent electrical and behavioral signatures. Some patients may therefore need to be prepared to stay for up to several weeks to capture enough data to move forward with a surgical evaluation.

• Magnetic resonance imaging (MRI): Most patients will also need to have an MRI to look at the structure of the brain and determine whether there are lesions. Additional navigational studies with contrast and a computed tomography (CT) scan may be required prior to any surgery as well. Sedation is often needed to get these detailed images.

• Functional MRI (fMRI): During this type of MRI, the patient must be awake and cooperative. They are asked to do specific tasks during imaging to identify regions where changes in blood flow localize motor and language functions.

• Magnetoencephalography (MEG): This test uses magnetic fields to both localize epileptiform activity and, when the patient can remain still and cooperative during the study, to also identify functional brain regions such as motor, sensory, and language.

• Transcranial Magnetic Stimulation (TMS): This noninvasive stimulation technique uses generated magnetic fields as another way to map language and motor function.

• Positron emission tomography (PET): This study uses an injected tracer to track the brain’s glucose metabolism. Areas of seizure onset often have lower metabolism between seizures and can be identified on PET to assist with localization.

• Ictal single photon emission computed tomography (SPECT): This test may sometimes be completed during the phase 1 EEG to help localize seizures and consists of a tracer being injected at the time of a seizure followed by imaging after the seizure to identify areas of the brain that experienced increased blood flow at the time of the injection.

• Genetic evaluation: Understanding whether there are any genetic underpinnings to a person’s epilepsy, whether focal or generalized, can have implications on the counseling that a patient and their family receive about likelihood of response to surgery and recurrence risk. Most patients are recommended to receive a genetic evaluation. A genetic diagnosis does not necessarily preclude successful epilepsy surgery.

• Neuropsychiatric evaluation: This type of evaluation, particularly important for resective surgical patients, consists of a battery of cognitive tests that can be performed over one or several days evaluating baseline and postoperative functioning in different domains to identify brain areas that may be particularly strong or weak for an individual patient and is outlined in more detail in Scotti-Degnan et al. in this issue.^8,9

• Phase 2 intracranial EEG: This study uses surgically placed intracranial electrodes to localize seizure onset and is limited to patients thought to have focal seizures. Electrode placement is designed through multidisciplinary discussion of results of the above studies. Patients will either have a stereo EEG (sEEG) where electrodes are placed in the brain tissue through burr holes in the skull or have electrocorticography (ECoG) using grids or strips of electrodes placed over the brain surface after removal of a portion of the skull.^10–12 More important than during a phase 1 EEG study, patients must be prepared to stay in the hospital until sufficient seizure data have been collected.

• Direct cortical stimulation: Using the electrodes placed during Phase 2 EEG monitoring, current can be passed to brain regions of interest to either map functions such as language and motor control, or to elicit seizures for clarification of the seizure onset zone.

• Wada test: This is a rarely used test in pediatrics at this time, but in some cases where language or other critical functional regions are not identified with other techniques, an anesthetic can be injected into one hemisphere of the brain via angiography to disable one hemisphere and evaluate the function of interest in that hemisphere.

Surgical options

It is helpful for the pediatrician working with patients with epilepsy to know what surgical options exist for their patients, and thus what type of intervention a referral to a level 4 epilepsy center might offer them. Surgical options range from resection or disconnection procedures to implantation of neuromodulatory devices and can be categorized as either targeted therapies for patients with focal epilepsy or as nonspecific surgical interventions for patients with multifocal or generalized epilepsy (summary in Fig 1). A more detailed summary of novel surgical techniques is found in Baumgartner et al. in this journal issue.¹³

Specific surgical options are reserved for patients with focal epilepsy from a defined and limited number of locations:

• Focal resective surgery: If the seizure focus can be identified and is in a non-eloquent region where removal would not yield intolerable deficits, the focus is removed through open surgery.

• Ablation techniques: For focal epilepsy with onset in small or difficult to access regions of the brain, various less invasive techniques are becoming available to ablate the region(s) of concern, including:

  1. Laser interstitial thermal therapy (LITT): This technique involves a laser (introduced stereotactically through burr holes rather than by open surgery) to heat and ablate a single or several small seizure foci. LITT can even be used for other types of surgery (such as corpus callosotomy, see below).^14–17

  2. Radiofrequency thermal coagulation (RFTC): This technique for inducing a local thermal lesion is useful for small foci and can be performed through already-implanted intracranial sEEG leads outside of the OR. Seizure freedom is rare, but can help guide subsequent resective surgery of the involved region.¹⁸

  3. High intensity focused ultrasound (HIFU): This is a true minimally invasive technique which does not require entry to the calvarium. HIFU is still being explored on a case by case basis, but shows promise for delicate deep structures such as the hypothalamus.^17,18

• Hemispherotomy: This technique of disconnecting an entire hemisphere of the brain is reserved for patients with large hemispheric lesions. While deficits are likely, the rate of seizure freedom in patients found to be good candidates for this procedure are high. Multiple approaches exist for performing this surgery, and it can now also be achieved in some cases through a burr hole or endoscopically, making the procedure less invasive.^13,18

• Responsive neurostimulation (RNS): When patients have either two separate seizure foci or a seizure focus in an eloquent area of brain that would yield unacceptable deficits if resected, an RNS device can be implanted. The device is programmed to detect seizures and deliver stimulation directly to the site of seizure onset to disrupt the seizure and gradually reduce seizure frequency over time. While currently only FDA approved in adults, a significant number of pediatric patients have now been followed for several years and outcomes are comparable to adults.^19–23 The RNS has recently also been adapted to be used as nonspecific stimulation (see below). More information on the RNS is available in Han et al. and Baumgartner et al. in this journal issue.^17,24

• Chronic subthreshold cortical stimulation (CSCS): Like the RNS, leads are placed in a location suspected to represent seizure onset but, rather than activating only when seizures are detected, provides intermittent regular stimulation at the site of seizure onset.^2,25

Nonspecific surgical treatment options can also be used in patients with generalized or multifocal epilepsy:

• Corpus callosotomy (CC): This surgery is a palliative option primarily considered for patients whose generalized seizure types include atonic or tonic seizures that cause drop attacks and may result in injuries. This technique can now also be achieved using LITT (see above) or endoscopically in some cases, making the procedure less invasive.

• Vagal nerve stimulation (VNS): The VNS is one of the older neurostimulation devices. Placed around the vagus nerve in the neck with a battery pack under the skin of the chest, like a pacemaker, the VNS does not involve brain surgery and provides intermittent regular stimulation which over time reduces seizure frequency in patients with any epilepsy type. The VNS also allows for a magnet-activated on-demand stimulation option for aborting ongoing seizures, which can be useful for some patients.

• Deep brain stimulation (DBS): Originally used to treat movement disorders such as Parkinson’s, the DBS has two implanted leads connected to a battery pack that is placed over the chest wall and for epilepsy has been placed in various thalamic nuclei as well as in other brain locations to provide intermittent regular stimulation which over time reduces seizure frequency. More examples can be found in Baumgartner et al. in this issue.¹⁷

• Responsive thalamic neurostimulation (thalamic RNS): Like the RNS technique described above, thalamic neurostimulation with the RNS device first detects seizures and treats only when they are detected, however, by placing one or both leads in the thalamus, a brain region with broad connections, the RNS can be converted to a more nonspecific option for patients with multifocal or generalized seizures. Initial data in pediatric patients using this technique has been promising.^26–28

Post-surgical expectations

The immediate post-surgical course following epilepsy surgery varies by the specific surgical intervention. In the case of device placement (RNS, DBS, VNS), post-operative recovery and discharge to home is often rapid (within days). Devices are typically left off for an initial period of healing and then stimulation is activated weeks later in the outpatient setting by a trained medical provider, with frequent return visits over subsequent months for device setting adjustments. On the other hand, resective, some ablative procedures, and disconnection surgeries can require a few days of post-operative observation in an intensive care unit and can lead to new neurological deficits that require acute inpatient rehabilitation and/or outpatient therapy.

Surgeries in which tissue has been damaged as part of the procedure (ablations, resections, disconnections) may be associated with new neurological deficits, which are typically worst in the first 3 months and then continue to improve over the next 12 months.²⁹ Specific deficits depend on the underlying function of the area being resected, as well as the degree of pre-existing injury in that area. For example, in hemispherotomy cases where there is an existing underlying deficit due to the unilateral injury (such as perinatal stroke), there is often minimal new deficit following surgery. While some regions such as the hand motor area often yield the most apparent persistent motor deficits when affected, certain brain regions are amenable to plasticity and recovery of function, such as the supplementary motor area, or to functional recovery by compensatory mechanisms such as with walking which is assisted by subcortical circuitry.³⁰

Corpus callosotomies are associated with a disconnection syndrome that peaks in the first 3 months with alien limb syndrome, apraxia (difficulty with performing complex gestures or tasks), tactile/visual anomia (written or spoken word finding difficulty), agraphia (inability to write despite intact verbal language), neglect, and/or dyslexia. In the long term, recovery is typically excellent, although there can be subtle cognitive differences.³¹

Neuromodulation devices have an overall favorable adverse outcome profile beyond the low-rates of immediate surgery-related complications (such as hemorrhage, infection) and the risk of damage to the leads. Some specific concerns with each of the devices exist. The VNS device can cause vagal-nerve related voice changes, dysphagia, coughing, neck pain, and can also worsen central and obstructive sleep apnea.^32,33 DBS with ANT placement has been associated with paresthesias, depression, and memory impairment.^34,35 One unique disadvantage with RNS is its conditional compatibility with brain MRI.^36–38 Rare sterile inflammatory reactions have been seen.³⁹ Additional information about complications seen with implantable devices such as RNS can be found in this issue in Han et al.²⁴

Outcome data

The primary desired outcome of all surgeries is seizure reduction. In a randomized control trial comparing pediatric patients with medically refractory epilepsy who were randomized to undergoing epilepsy surgery (resection, hemispherotomy, or corpus callosotomy) versus adjusting medications alone, 77% of the patients who underwent surgery achieved seizure freedom at 1 year as opposed to 7% of those who had medical treatment only.²⁹ Importantly, gains in seizure control appear to be sustained into adulthood – in a retrospective questionnaire over a decade from childhood epilepsy surgery, 87% were seizure free 1 year from surgery and 63% were seizure free 10 years from surgery; only a minority who still had seizures reported worsening in their seizures since the surgery.⁴⁰

In pediatric patients the rate of long term seizure freedom is highest when focal resective surgery is felt to be appropriate (56% or 76% of patients for extratemporal or temporal onset seizures, respectively)^41,42 And despite the deficits often expected with hemispherotomy, seizure freedom rates in the specific patients for whom this specific surgical option is recommended are between 70 and 93%.³⁰ Corpus callosotomy provides a worthwhile seizure reduction in 59–88% of cases (with anterior only vs complete callosotomy, respectively)³¹ Ablation outcomes data are still being collected. LITT seems to have a lower rate of seizure freedom than resective surgery of the comparable brain region^14–16 Seizure freedom from RFTC is rare, closer to 23% at 1 year after this technique, though a response to this technique can portend a good response to subsequent resection.⁴³

Neuromodulatory devices have overall lower rates of seizure reduction and seizure freedom than resective/disconnection surgeries but can provide an important adjuvant to medication in patients who are not candidates for removal of the seizure onset region(s). In the long term, half of patients with VNS report a 50% reduction in seizures, and 5% report seizure freedom.^44–47 The original clinical trials of DBS targeting the anterior nucleus of the thalamus in adult patients reported a 69% seizure reduction after 5 years, and 16% with at least one year of seizure freedom.^48–50 Newer studies of DBS targeting the centromedian nucleus of the thalamus for patients who have Lennox-Gastaut syndrome show significant but more modest improvements^51,52 In a pediatric cohort, 85% of patients had a > 30% reduction in seizures with use of DBS.^2,53 For the RNS, the longest outcome data suggest that 73% of patients can expect > 50% seizure reduction at 9 years, with many reporting at least 1 full year of seizure freedom in the time since device implantation; the newer pediatric studies have demonstrated a > 50% reduction in seizures in 65% of patients over the first year.^19,54,55 Also see Han et al. in this issue for more information about RNS outcomes in pediatric patients.²⁴

It is important to remember that not every patient is a good candidate for any one of the devices, thus direct comparison of the devices will remain a challenge due to the necessary selection bias of the patients undergoing each procedure. A recent meta-analysis in adults with devices showed a 14% decrease in seizures with VNS and DBS in first 3 months, compared with 32% initial decrease for patients with RNS; with each device there is gradual improvement over time, approximately 35% seizure freedom at 5 years with VNS versus 65–66% with either RNS or DBS, and 75% percent seizure reduction at 7–9 years with either RNS or DBS.^38,56 Amongst patients at a single institution, overall response to different neuromodulatory devices was found to be comparable, with cortical stimulation options being slightly more effective overall than subcortical stimulation options.⁵⁷ Importantly, long-term follow-up data in adults suggest that for patients with any of the three devices, risk of sudden unexplained death in epilepsy patients (SUDEP) falls from 6 in 1000 to 2 in 1000,⁵⁸ underscoring the life-saving potential of these devices.

In addition to the benefits of seizure freedom, epilepsy surgery has beneficial effects on neuropsychological outcome. Epilepsy surgery does not worsen underlying cognitive impairment,⁵⁹ and patients who have better seizure control by surgical or medical means have improved cognitive outcomes compared to peers with worse seizure control.⁶⁰ Some families report clinically meaningful improvements that are not measurable on standardized testing, and in the case of early-onset structural epilepsy, early surgery in the toddler years is linked with improvement developmental quotients.⁶¹ A review of neuropsychological issues in epilepsy surgery patients can be found in this issue in Scotti-Degnan et al.⁹

Conclusions

Epilepsy surgery has an important role in achieving seizure control in patients with medically refractory epilepsy. Newer surgical techniques and the increasing capabilities of neuromodulatory devices have expanded the population of patients that can benefit from surgery to include all refractory epilepsy patients regardless of seizure type and age. When patients have continued seizures despite adequate trials of two medications, primary care providers should be aware of the need for referral to a level 4 epilepsy center to further consider surgical options.

Vitae

Ania Dabrowski is a pediatric epilepsy fellow and physician scientist at Children’s Hospital of Philadelphia. She previously trained at Johns Hopkins Hospital and obtained her MD/PhD degree at University of Michigan, Ann Arbor. Her clinical and research interests include epilepsy genetics and neural circuit development. Caren Armstrong is a pediatric epileptologist and physician scientist at Children’s Hospital of Philadelphia. She trained previously at Johns Hopkins Hospital and obtained her MD/PhD degree at the University of California, Irvine. Her clinical work focuses on the genetic evaluation of children with epilepsy, and her research is in understanding the micro- and macro-circuitry of the brain that leads to seizures in pediatric patients undergoing epilepsy surgical evaluation.