Neuroimaging is one of the most important advances made in the past decade in the management of

seizure disorders. Magnetic resonance imaging (MRI) has increased substantially the ability to detect
causes of seizure disorders, to plan medical or surgical therapy, and to prognosticate the outcome of
disorders and therapy. However, MRI must be performed with techniques that will maximize the
detection of potentially epileptogenic lesions, especially in candidates for epilepsy surgery. Functional
imaging has an established role in evaluating patients for epilepsy surgery. It is relied on when results
from standard diagnostic methods, such as clinical information, electroencephalography, and MRI, are
insufficient to localize the seizure focus. Also, functional imaging is a reportedly reliable alternative to
invasive methods for identifying language, memory, and sensorimotor areas of the cerebral cortex.
Despite the availability of multimodality imaging, the epileptogenic zone is not determined solely by a
single imaging modality. Evidence and experience have shown that concordance of results from clinical,
electrophysiologic, and neuroimaging studies is needed to identify the epileptogenic zone accurately.
With modern techniques in image processing, multimodality imaging can integrate the location of
abnormal electroencephalographic, structural, and functional imaging foci on a "map" of the patient's
brain. Computer image-guided surgery allows surgically exact implantation of intracranial electrodes and
resection of abnormal structural or functional imaging foci. These techniques decrease the risk of
morbidity associated with epilepsy surgery and enhance the probability of postsurgical seizure control.
Surgical management of epilepsy associated with temporal lobe tumors
Suresh Sankhla, GM Khan
 Department of Neurosurgery, Saifee Hospital, Dr. B. Nanavati Hospital, Mumbai, Maharashtra, India

Correspondence Address:
Suresh Sankhla
A-503, Chaitanya Towers, Appasaheb Marathe Road, Prabhadevi, Mumbai - 400 025, Maharashtra 
India

Source of Support: None, Conflict of Interest: None

DOI: 10.4103/1817-1745.40601

 
    Abstract  

Object: Despite advancement in neuroimaging and improvement in the knowledge of tumor behavior,

the optimal surgical treatment of patients with temporal-lobe tumors remains unclear. Controversies
still exist regarding the type of tumor surgery, i.e., lesionectomy alone or in combination with resection
of mesial temporal structures and epileptogenic cortex adjacent to the tumor and the extent of tumor
removal. In this retrospective study, the authors have analyzed seizure outcome in a series of children
with epileptogenic temporal lobe tumors who had been treated only by lesionectomy. Materials and
Methods: Sixteen patients including 7 boys and 9 girls, with a mean age of 9.9 years (range: 5-17
years) underwent lesionectomy as the only surgical treatment for temporal lobe tumor related
epilepsy. The interval between onset of seizures and surgery ranged from 3 months to 7.5 year (mean
5.9 year). Ten patients (62.5%) presented with complex partial seizures and the remaining 6 (37.5%)
had simple partial seizures with secondary generalization, with seizure frequency varying from several
per day to only a few per month. All tumors were located within the temporal lobes: 8 temporomesial,
5 temporolateral and 3 temporobasal. Cortical dysplasia adjacent to tumor, hippocampal sclerosis or
tumor infiltration was not demonstrated in any patient. Results: Gross total resection of the tumor was
achieved in 11 (69%) and subtotal resection in 5 (31%) patients. The histopathological diagnosis was
ganglioglioma in 6 (37.5%) patients, dysembryoplastic neuroepithelial tumor in 5 (31.3%), low-grade
astrocytoma in 2 (12.5%), juvenile pilocytic astrocytoma in 2 (12.5%) and pleomorphic
xanthoastrocytoma in 1 (6.3%). At a mean follow-up of 5.2 years (range, 1.4-9.3 years), 11 (69%)
patients were seizure-free (Engel class I) and 4 (25%) others had a significant improvement in the
seizure frequency (Engel class II). Conclusions: The results of this study suggest that good long-
term seizure control can be achieved with lesionectomy alone in majority of children suffering from
medically resistant epilepsy associated with temporal lobe tumors.

Keywords: Dysembryoplastic neuroepithelial tumor, epilepsy, ganglioglioma, juvenile pilocytic

astrocytoma, lesionectomy, low-grade astrocytoma, pleomorphic xanthoastrocytoma, temporal lobe
tumor

How to cite this article:

Sankhla S, Khan GM. Surgical management of epilepsy associated with temporal lobe tumors. J Pediatr
Neurosci 2008;3:121-5

How to cite this URL:

Sankhla S, Khan GM. Surgical management of epilepsy associated with temporal lobe tumors. J Pediatr
Neurosci [serial online] 2008 [cited 2010 Jun 22];3:121-5. Available
from: http://www.pediatricneurosciences.com/text.asp?2008/3/1/121/40601

    Introduction  

Low-grade temporal lobe tumors are now increasingly recognized as the cause of seizures in children
and young adults. [1],[2],[3],[4],[5],[6],[7],[8] Although our understanding of the clinicopathological behavior of
these tumors has improved in the past few years, a clear consensus regarding the surgical
management of these tumors is lacking. Controversies still exist whether lesionectomy alone is
sufficient to control epilepsy [9],[10],[11],[12] or if epileptogenic areas adjacent to the tumor and/or mesial
temporal structures should also be resected to improve the seizure outcome.  [1],[6],[13],[14],[15],[16] In this
report, the seizure outcome has been retrospectively analyzed in 16 children who had received
lesionectomy as the only surgical treatment of their temporal lobe tumors.

    Materials and Methods  

The medical records were reviewed to obtain clinical data of the children who underwent surgery for
temporal lobe tumors at our institution between January 1998 and December 2004. The selection
criteria included the following: (1) seizure as the main presenting feature, (2) temporal lobe mass
lesion demonstrated on neuroimaging, (3) resection of the lesion, (4) histopathological confirmation of
a neoplasm, (5) a minimum 16-month follow-up period. All patients underwent comprehensive pre-
and postsurgical workups, including a thorough seizure history, neurological examinations,
neuroimaging studies (magnetic resonance imaging with seizure protocol) and routine
electroencephalogram (EEG) recordings. Preoperative and postoperative neuroimaging studies were
reviewed by the author with two senior radiologists to verify the location of the lesion and its
relationship to the adjacent cortex and structures of the limbic system, as well as to determine the
extent of tumor removal. Follow-up examination results were obtained from patient charts.
Postoperatively, all patients continued to receive anticonvulsant medications for a variable period
before attempting gradual drug withdrawal. Medication was continued in those children who did not
have satisfactory seizure control after surgery. Postoperative seizure outcome was categorized using
the Engel classification. [17] Seizure frequency at the time of the last follow-up was compared with that
prior to surgery.

    Results  

Patient characteristics and surgical outcomes are shown in [Table - 1]. There were 16 children, 7 male
and 9 female, with a mean age of 9.9 years (range: 5-17 years). In almost all the cases, the onset of
seizures was during infancy or childhood (range: 0.5-13 years) and the interval between the onset of
seizures and surgery ranged from 3 months to 7.5 years (mean: 5.9 years). Ten patients (62.5%)
suffered complex partial seizures, while the remaining 6 patients (37.5%) had presented with simple
partial seizures and secondary generalization. Seizure frequency varied from several episodes per
day to only a few per month. Neurologic examination was unremarkable in all patients. According to
the neuroimaging studies, all tumors were located within the temporal lobes, 10 on the right side and
6 on the left. Eight tumors were classified as temporomesial, 5 as temporolateral and 3 as
temporobasal. Cortical dysplasia adjacent to tumor, hippocampal sclerosis or tumor infiltration was not
demonstrated in any patient. Tumor resection was gross total in 11 (69%) [Figure - 1] and subtotal in 5
(31%) patients [Figure - 2]. Because of the close proximity to the eloquent areas, smaller tumor pieces
were left behind in 5 patients. The postoperative complications were observed in 4 patients (25%),
including temporary hemiparesis in 2 cases, superior quadrant field defect in 3, transient psychosis in
1 and wound infection in 1. The histopathological diagnosis was ganglioglioma (GG) in 6 (37.5%)
patients, dysembryoplastic neuroepithelial tumor (DNET) in 5 (31.3%), low-grade astrocytoma (LGA)
in 2 (12.5%), juvenile pilocytic astrocytoma (JPA) in 2 (12.5%) and pleomorphic xanthoastrocytoma
(PXA) in 1 (6.3%). The mean follow-up period was 5.2 years (range, 1.4-9.3 years). Postoperatively,
good seizure outcome was observed in 94% of the cases. Eleven (69%) patients were seizure free
(Engel class I) and 4 (25%) others had a significant improvement in seizure frequency (Engel class II).
One patient with Engel class III outcome who had subtotal resection of the low-grade astrocytoma
underwent reoperation for the residual tumor and continued on anticonvulsant treatment for 24
months until last follow-up. In 9 (56%) patients, antiepilepsy medication was gradually reduced and
stopped after a minimum 2-year seizure-free period. During the last follow-up, 4 patients with
recurrent seizures and 2 without seizures were receiving antiepilepsy treatment with a plan to stop all
medications in the near future.

    Discussion  

Intrinsic low-grade temporal lobe tumors in children most commonly present with epilepsy that is often
resistant to medical treatment. It has been estimated that approximately 38-76% of all cases with
supratentorial tumors considered for epilepsy surgery have tumors in the temporal lobes.  [18],[19],[20] The
primary goal of treatment in these patients is to achieve complete control of seizures. Clinical data
from several studies indicate that the surgical treatment is effective and provides good long-term
seizure outcome. [9],[10],[11],[12] Tumor removal is not only associated with effective seizure control, it can
also improve psychological and intellectual development of the patient, reduce the incidence of
secondary epileptogenesis, ensure an accurate diagnosis and prevent the malignant transformation of
otherwise benign tumors. [1],[7],[10],[21],[22],[23] 

However, the optimal surgical treatment for children suffering from epilepsy associated with temporal-
lobe tumors remains controversial. Some authors consider complete tumor resection alone as an
adequate treatment, [9],[11],[12] whereas others recommend the additional removal of epileptogenic
surrounding brain tissue to provide a good seizure outcome.  [1],[6],[13],[14],[15],[16],[19],[24],[25] The literature is also
unclear whether the extent of tumor resection has any direct impact on the long-term seizure
outcomes. While some investigators [19],[26],[27] report significant seizure control following radical tumor
removal, others [28] demonstrate similar results with incomplete tumor excision. From a surgical point of
view, another controversial issue is the resection of mesial temporal structures in addition to the tumor
removal. Whether the additional vital brain tissue resection improves the outcome significantly or not
continues to be a matter of debate. It is thus obvious that a proper understanding of the
pathophysiological mechanisms leading to tumoral epilepsy is essential, and without this, it may be
difficult to determine the optimal surgical strategy.

The exact mechanism of epilepsy in the patients with temporal-lobe tumors is unclear. The proposed
hypotheses include direct pressure and irritation in the cortical tissue, gliosis, disrupted
vascularization in the surrounding cortex, morphological neuronal alterations in cellular levels,
changes in the level of neurotransmitters, denervation hypersensitivity and immunological activities.  [18],
[27],[29]
 Although the peritumoral brain tissue is generally considered as the region where epileptic
discharges are generated, [30],[31] there is now clear evidence that certain tumors such as GGs and
DNETs have specific epileptogenic activities because of their neuronal and glial components. The
neurochemical profiles of these tumors seem to demonstrate some similarities to neocortical neurons
in the expression of various enzymes and receptors. [1],[3],[32],[33],[34] These profiles support the possibility
that these tumor may have intrinsic epileptogenic properties and that the neuronal component of the
tumor itself may contribute to epileptic activity. [21] The origin of some epileptiform activity in
gangliogliomas has been demonstrated by few other authors also.  [35],[36] Beaumont et al .,[30] Zaatreh et
al . [37] and others [21],[38] demonstrated relatively low incidence of Ammon's Horn sclerosis in the
hippocampus of patients who had undergone en bloc resections of epileptogenic temporal tumors.
These studies indicate that the temporal lobe tumors may be the primary source of the epileptic
disorder in a large number of cases.

It therefore seems logical to consider lesionectomy as the primary treatment in patients who present
with epilepsy secondary to temporal lobe tumors. There are several published reports in the literature
supporting good long-term seizure outcomes following tumor resection alone. In the study by
Clusman et al ., [22] all the children who underwent only temporal lesionectomies became seizure free
during follow-up. In a series of 29 children who were treated surgically for a temporal lobe tumor-
related seizure disorder, Cataltepe and coworkers[39] concluded that resection of tumor with or without
amygdalohippocampectomy provides a high rate of seizure-free outcome. Giulioni and
associates [40] analyzed 15 children suffering from epilepsy, who were associated with glioneuronal
tumors treated by simple lesionectomy, and observed that 86.6% of patients achieved excellent
(Engel Class I) postoperative seizure control. Our experience with lesionectomy is similar to others.
Out of 16 children who underwent lesionectomy alone for temporal lobe tumor-associated epilepsy, 15
had an outcome that can be considered as good, Engel Class I in 11 and Class II in 4. Nine patients
were totally free of medication and four of them continued gradual withdrawal of antiepileptic drugs at
the last follow-up.

Another controversy in the management of tumoral epilepsy is involved with the extent of tumor
resection. The complete excision of the tumor is still considered as the main factor affecting seizure
outcomes. [6],[8],[10] Blume et al . [26] reported seizure-free outcomes in 88% of their cases after gross-total
resection, whereas only 25% of patients were seizure-free after a subtotal resection. Several other
authors have reported similar results [19],[27],[39],[40] Eleven (69%) patients in our study had total tumor
removal; 8 had Engel Class I and 3 had Engel Class II seizure outcomes. Interestingly, subtotal
resection is not always associated with poor seizure control. Some studies have indicated favorable
outcomes even in patients who had undergone incomplete tumor removal.  [3],[6],[13],[28] Kirkpatrick et
al . [28] reported seizure-free outcomes in 81% of patients with incomplete tumor resection. Three
(60%) out of 5 cases with subtotal tumor resection in our series had seizure-free outcomes. It has
been postulated that the favorable outcome observed in cases with incomplete tumor resection may
perhaps be due to the reduced volume of critical epileptogenic mass and interruption of the critical
seizure propagation pathways. [37] Some authors however doubt the role of subtotal tumor removal and
argue that even if the seizure control is achieved in some cases, it is often temporary and increasing
epileptogenic mass secondary to tumor regrowth may cause recurrence of seizures in long term. For
this reason we believe that gross-total resection, where possible, should be the goal of the surgery in
these patients. Complete tumor excision is also of paramount importance for preventing malignant
transformation of the residual mass. [2],[23] Although three of our patients with incomplete tumor removal
had a good seizure outcome, we could not draw a definite conclusion from this observation because
the size of our patient population is small and statistically insignificant.

The role of resection of the cortex surrounding the epileptogenic tumor and/or the mesial temporal
structures, in addition to tumor excision is debatable. Although it has been estimated that
approximately 40% of glioneuronal tumors are associated with dysplasia, no clear association has yet
been established between epilepsy and dysplasia of the cortex adjacent to the tumor.  [1],[3],[6],[13],[22],
[32]
 Jooma et al . [25] in 1995 reviewed their experience in 30 patients with seizures who received surgery
in the form of resection of the temporal mass lesions. The seizure-free outcome was significantly
lower in these patients as compared to those who underwent electrophysiologically guided resection
of the lesion and epileptogenic focus. In contrast, Khajavi et al . [27] and Zaatreh [37] found that the
seizure-free outcomes correlated with the extent of the tumor resection but not with the additional
excision of the surrounding epileptogenic zone. They concluded that radical resection of the tumor
either removes the primary epileptogenic drive completely or reduces the epileptogenic tissue beyond
a critical mass required to generate clinical seizures.

There are convincing data regarding surgical strategy if the temporal lobe tumor is associated with
hippocampal sclerosis, in which case a resection of the mesial temporal structures is recommended
along with lesionectomy. Hippocampectomy as the treatment of tumor-associated epilepsy in the
absence of hippocampal atrophy is debatable. The proponents of the amygdalohippocampectomy
argue that the mesial temporal structures are capable of generating independent seizures and that
their epileptogenicity is because of the abnormal synaptic reorganization of the hippocampi induced
by seizures secondary to temporal tumors. [37],[41],[42] Mathern et al . [43] found that in patients with
temporal lobe lesions, ictal EEG onset started in the mesial temporal contacts in 94% of the patients.
However, the continuity of epileptogenic capability after the removal of the lesion is almost unknown.
Furthermore, there is no clear evidence that the hippocampi in patients with tumor-related epilepsy
are morphologically abnormal. [27] Therefore, it is now generally believed that the mesial structures may
be resected only when the tumor is involving the mesial structures and should be preserved in all
other cases. [39],[41],[42] Giulioni and coworkers [40] reported surgical results in 15 children with tumoral
epilepsy and concluded that lesionectomy alone may be sufficient to provide good long-term seizure
control and additional resection of the surrounding brain tissue or mesial temporal structures may not
be necessary in majority of cases. Our results further support the findings of these authors. In none of
the patients in our series, there was any evidence of hippocampal abnormality on MRI or cortical
dysplasia on histopathological examination.

Although the duration of seizure has been reported to be adversely related to the seizure outcome by
some authors, [6],[13] we have not observed any definite association between the seizure control and
duration, type or frequency of seizure and age of the patient. Because of the small number of patients
in this study, it was also not possible to establish any meaningful relationship between the type of
tumor and seizure outcome. Surgical complications were essentially minor, transient and treatable. No
recurrence or regrowth of the residual tumor was documented in our study.

    Conclusions  

The epileptogenic temporal lobe tumors are usually slow-growing and histopathologically benign or
low-grade lesions. Gross-total resection of the tumor is associated with a good seizure-free outcome
and should, therefore, be the goal of surgery. Although the results of incomplete tumor excision are
not completely unfavorable in this study, a larger patient population may be required to draw more
meaningful conclusions. Our findings further indicate that lesionectomy alone can result in good
seizure outcome in majority of patients and additional surgical procedures such as amygdalo-
hippocampectomy and the resection of the adjacent epileptogenic cortex should be reserved for the
selected patients who have dual pathology such as the mesial temporal sclerosis or cortical dysplasis
in association with temporal lobe tumor.
Surgical Management of Epilepsy

By Adam N. Mamelak, MD and Ajay Ananda, MD

Introduction

Epilepsy is the most common disorder of the brain and the second leading
cause of neuropsychiatric disability worldwide. The prevalence of epilepsy in
developed countries is estimated to be 0.3-4 percent of the population, making
epilepsy as common a disorder as Type I diabetes. This prevalence is probably
much higher in less developed countries. Due to the social ostracism and
negative stigma often associated with seizures, many people with epilepsy
either hide their disorder or become removed from the mainstream of society by
choice resulting in poor socio-economic status, high levels of unemployment
and less satisfying interpersonal relationships than many other discords, which
lack these stigmata.

Well over half of all cases of epilepsy are characterized by partial seizures,
generally indicating a discrete (and potentially removable) seizure focus or zone
of ictal onset within the brain. Unfortunately, approximately 30 percent of
patients with epilepsy will continue to have seizures despite aggressive
treatment with multiple antiepileptic drugs (AEDs) and other interventions. The
untoward cognitive side effects of AEDs coupled with the long-term effects of
recurrent seizures result in shorter life spans, reduced quality of life and
increased co-morbidities for the majority of patients. For these patients, surgery
to excise the seizure focus or diminish the frequency and/or severity of seizures
is often the best option available.

Selection of Surgical Candidates

As a general rule, patients without clear-cut lesions such as brain tumors or

malformations are not considered for seizure surgery unless they have failed a
good trial of AED rational polytherapy. This typically implies that at least two
(often many more) AEDs and/or a combination of such drugs have failed to
control the seizures. Depending on the seizure type and the goals of the
surgery, patient for whom surgery can be considered fall into two broad
categories: intent to cure or palliation. Of note, age is only a relative
contraindication for epilepsy surgery. Data suggests that the younger a patient
with refractory seizures is rendered seizure-free, the more likely they will
complete their schooling, become viable members of the work force and have
healthy social interaction. Similarly, there is no absolute age limit above which
seizure surgery cannot be performed, even though the risk of surgical
complications is clearly higher for patients above age 45 and the impact of
rendering such patients seizure-free is probably diminished from a socio-
economic perspective.

"Intent to cure" procedures include operations such as temporal lobectomy,

amygdalo-hippocampectomy, hemispherectomy, or neocortical excision of a
seizure focus. These operations are performed via craniotomies and are usually
limited to those patients with localization-specific epilepsy (LSE). LSE implies
that all, or at least the majority of the seizures arise from a well-defined region
in the brain, for whom there is a concordant amount of data from a variety of
non-invasive and invasive tests that indicate that the seizure focus has been
reliably identified. A typical example of this would be a patient with medically
refractory complex partial seizures arising from the mesial temporal lobe
structures (e.g.amygdala, hippocampus). Such a patient might be an ideal
candidate for an anterior temporal lobectomy or selective removal of the
amygdala and hippocampus. For patients in this category surgical resection has
been demonstrated to carry a 60-85 percent chance of rendering the patient
seizure-free, compared to a less than 8 percent chance with medications. The
chances of a patient becoming seizure-free from an extratemporal resection are
less, but still in the range of 60-70 percent, far higher than with any
combination of AEDs.

In some cases, a very large volume of a patient's brain may be involved in

seizure generation. For these patients who often have a catastrophic form of
epilepsy, removal of large portions of the diseased hemisphere may be needed.
Dramatic operations such as hemispherectomy (removal of half the brain) and
functional hemispherectomy, while large, have an excellent track record for
controlling seizures in these patients. Typical patients undergoing
hemispherectomy include children with Rasmussen's encephalitis,
hemimegaloencephaly or those suffering from seizures as a byproduct of a
hemispheric stroke. More recently, hemispherotomy (otomy rather than ectomy)
has been reported and popularized by several epilepsy surgeons. In these
surgeries all of the major white matter pathways connecting the lobes of the
hemisphere to each other and the contralateral hemisphere are divided, but the
bulk of the brain tissue is left in place. These procedures have gained popularity
to avoid the many complications associated with massive surgical resections
such as infection, hydrocephalus, hemosiderosis and brain shift. These
procedures require an excellent understanding of the brain functional anatomy
and have not yet been proven to provide superior results compared with more
standard procedures. Nonetheless, these creative approaches are quite
appealing and with the development of sophisticated white matter fiber tracing
for intraoperative use with surgical navigation system, it may gain further
acceptance as a less destructive option for these patients.

Palliative Procedures

For patients in whom a discrete seizure focus does not exist or for whom
surgical removal of the zone of seizure origin is felt to be too risky, there are
some palliative surgical options. These surgeries do not aim to eliminate all
seizures but to reduce the frequency and severity of the most disabling ones,
and hopefully permit reduction in the amount of AEDs.

For patients with secondary generalized epilepsy due to congenital or acquired

brain injury (e.g. Lennox-Gastaut syndrome, tuberous sclerosis) surgical
excision of a discrete seizure focus is rarely warranted. However, these patients
often suffer frequent life-threatening generalized seizures or drop attacks in,
which they injure themselves or are hospitalized. For these patients, corpus
callosotomy or surgical division of the corpus callosum can be used with great
success. Corpus callosotomy typically results in a greater than 90 percent
reduction in drop attack seizures, and a 60-70 percent reduction in generalized
seizures. While not curative, these procedures can have a dramatic impact on
the quality of life for both patient and caregiver. The utilization of intra-
operative MRI or intra-operative surgical navigation to ensure division of at least
the front 2/3rd of the corpus callosum is a recent adjunct that has improved
outcomes for the patients.

An alternative for many patients is the use of the Vagus Nerve Stimulator (VNS).
VNS is an FDA-approved device consisting of an electrode coil that is wrapped
around the left Vagus nerve in the neck and a pulse generator implanted in the
chest wall. Electrical stimulations of the nerve are transmitted to the brain stem
via afferent fibers in the Vagus nerve and induce a neuromodulatory affect on
the brain that can reduce seizures. VNS implant is a much smaller procedure
than all of the other surgical procedures discussed. It typically takes 45
minutes, and is performed on an outpatient basis. The device is well-tolerated
and very easy to program. A large database of over 8,000 patient implanted in
the US indicates that seizure reduction is observed in almost 2/3rds of patients,
with a greater than 50 percent reduction seen in over half the patients. The
frequency and intensity of seizures is also reduced in many patients and
improvements in mood and alertness are also frequently reported, especially if
AEDs are reduced. Unfortunately, very few patients become seizure-free
following VNS implant, and 25 percent will experience no change in their
seizures. Finally, patients who are potential candidates for more definitive
excision surgery are still far more likely to benefit from resection than VNS and
should be evaluated for candidacy before considering VNS as a palliative option.
Nonetheless, the VNS is an attractive surgical alternative to AEDs or more
invasive epilepsy surgery in appropriately selected candidates.

Seizure Localization with Intracranial Electrodes

For many patients with LSE the exact location of the seizure focus cannot be
demonstrated convincingly by non-invasive tests. For these patients, video-EEG
monitoring utilizing intracranial electrodes is often needed to confirm the
seizure focus, better define the extent of a resection or potentially indicate that
surgical resection might not be a safe option. Often, these invasive electrodes
are used to perform cortical stimulation mapping, a technique wherein electrical
current is delivered to the individual electrodes to map out higher cortical
functions such as language, motor, sensory and memory specific regions. By
combining information regarding the zone of seizure origin with a detailed
patient-specific map of the cortical surface anatomy, very precise and tailored
removal of epileptic tissue can be removed. Both surface electrode arrays such
as subdural grids and electrodes that penetrate into the brain parenchyma such
as depth electrodes can be utilized. Depth electrodes are particularly useful in
confirming that seizures arise from the medial temporal lobe region, to more
accurately lateralize the seizure site and to identify patients with multiple
independent zones of seizure onset. In contrast, subdural grid electrodes are
ideal for providing a detailed map of a complex zone of seizure origin over a
broad cortical region such as the frontal or parietal lobes.

Alternate Non-invasive Methods of Seizure Localization

While surgical excision of an epileptic focus and the techniques for invasive
electrode monitoring are quite sophisticated and involved, they are not really
very new ideas. Pioneers such as Wilder Penfield originally defined similar
methods in the 1940's and with some technical improvements have been a core
tool of epilepsy surgery since the 1980's. The development of more
sophisticated methods for seizure localization that avoid or minimize the use of
invasive intracranial recordings has been the focus of heavy research ever since
the potential surgical complications of invasive monitoring such as infection or
brain injury became apparent. Non-invasive methods have clear advantages for
patient safety, cost, and convenience. The use of imaging methods such as
18FDG PET and Magnetic Resonance Spectroscopy (MRS) can be a useful adjunct
to identifying a seizure focus in a subset of patients. 18FDG is measure of brain
glucose utilization and areas of brain scarring may have diminished glucose
utilization, providing an indirect but concordant test for localizing a seizure
focus. Similarly MRS can detect brain areas with diminished number of neurons,
again suggesting a region of brain injury. Ictal SPECT has been used in a similar
fashion to detect regions of increased metabolism during a seizure.
Unfortunately, these methods do not directly measure brain electrical activity
and as such are concordant tools.

Recently, there has been interest in the use of Magnetoencephalography (MEG)

to help localize a seizure focus. MEG detects the magnetic dipole equivalents of
electrical current in the brain. Because of the unique properties of magnetic
dipoles, major issues affecting EEG dipole modeling such as volume conduction
of electrical currents in the brain are avoided. However, MEG is very expensive
and requires significant technical expertise to perform and interpret the studies
accurately. Better methods of dipole mapping are also needed to correctly
identify the sources of activity in many cases. Nonetheless, several investigators
have demonstrated MEG to be a useful tool for directing placement of
intracranial electrodes, modifying surgical planning and performing non-
invasive brain mapping. Ongoing efforts in this arena will likely improve our
ability to correctly localize more patients, resulting in more surgical "cures".

Advising Patients

Surgery continues to represent the most viable opportunity to cure epilepsy for
most patients with localization-specific, medically refractory epilepsy. This is a
bold statement when one considers the fact that in the past decade over eight
new AEDs have been introduced in the U.S. While a useful adjunct, these drugs
have had a relatively minor impact on the control of seizures in this patient
population. Unfortunately, the majorities of these patients are never referred to
a comprehensive epilepsy program and suffer the life-long effects of seizures.
The careful work-up and evaluation of a patient with epilepsy is one of the most
detailed and investigative work-ups in all of medicine. For this reason,
successful surgery is critically dependent not just on the surgeon, but the entire
epilepsy team comprised of neurologists, neuropsychologists, EEG technicians,
nurses, and others. Close interaction between the surgeon and neurologist and
meticulous review of all data collected remain the core feature of successful
epilepsy surgery. With increased physician awareness of the significant benefits
and relatively low risks of epilepsy surgery there is a greater chance that
patients will be referred to centers that can more definitively treat their seizures.

Multimodal Seizure Localization and Resection

(Figure 1) A 33 year old woman with a life-long history of medically refractory
complex partial seizures had pre-surgical work-up indicating a probable right
mesial temporal lobe onset to her seizure, but without clear localization by
scalp EEG. A-C) MEG revealed multiple interictal spikes in the right posterior
temporal lobe. D) Depth electrodes were placed, with extra electrodes placed in
the posterior temporal lobe at the region of MEG spikes. Continuous video-EEG
demonstrated onset in this region, with rapid spread to the mesial temporal
lobe. E) A subdural grid electrode array was placed over this area and further
localized the seizure onset zone (red) to the posterior inferior temporal lobe
with rapid spread (2 seconds) to the mesial temporal lobe. A tailored removal of
the temporal lobe to include the area of seizure onset and the mesial temporal
lobe was carried out. The patient has now been seizure-free for over three
years, is currently driving and has discontinued all AEDs.