Key words
epilepsy surgery - pharmacoresistance - development - focal cortical dysplasia - glioneuronal
tumor
Introduction
Epidemiology
Epilepsy is one of most common neurological disorders in children. The incidence of
unprovoked seizures is 44–57 per 100 000 population; the incidence of epilepsy is
33–61 per 100 000 population. The incidence of epilepsy is considerably higher in
the first year of life and remains high in childhood and adolescence to decrease between
the third and sixth decade of life. Beyond the first year of life, the incidence of
epilepsy remains constant at 46 per 100000 population to decrease markedly after the
tenth year of life [1]
[2].
Pharmacoresistance
In a prospective pediatric cohort study with a follow-up of over 10 years, 23% of
children and adolescents diagnosed with epilepsy were drug-resistant at any given
time [3]. Although this study indicated a higher proportion of children in remission than
previously reported [4], this remission with at least one seizure-free year was achieved on average after
5 or more years, and only after several anticonvulsant trials, whereas two-thirds
of these children experienced later recurrences. In this study, only 4% of patients
who did not respond to the second anticonvulsant remained seizure-free over 4 years.
Likewise, early remission with anticonvulsants followed by subsequent enduring remission
at last follow-up (smooth sailing) in a third of the children and adolescents should
be taken with caution, since follow-up was limited to a few years.
It is thus presumed that 23% of children with epilepsy meet the criteria for pharmacoresistance
[3], defined as the failure of 2 appropriate treatment trials with well-tolerated and
suitably selected anticonvulsants used alone or in combination [5]. Pharmacoresistance is usually established relatively early in the course of epilepsy
[6]
[7], and is often associated with cognitive deficits, psychiatric comorbidities, reduced
quality of life and increased SUDEP risk (Sudden Unexpected Death in Epilepsy). Neonatal
seizures, high seizure frequency at epilepsy onset, developmental delay and structural
lesions correlate with poor chances of remission [8]. In particular, a structural etiology has been identified as the best predictor
of pharmacoresistance [9]. Because of this, the initiation of presurgical workup at an early stage is indicated
for children with structural epilepsies after the failure of 2 anticonvulsants[10].
Epilepsy surgery
In recent years, epilepsy surgery has been established as a particularly effective
treatment option for children and adolescents with drug-resistant structural epilepsies
[11]. Epilepsy surgery should be considered at the latest when pharmacoresistance has
been ascertained and presurgical evaluation should be promptly initiated (Cross et
al., 2006). The primary goal of epilepsy surgery is the resection or disconnection
of the epileptogenic zone, defined as the cortical area that is necessary and sufficient
for initiating seizures [12]. In epilepsies associated with a structural lesion, the epileptogenic zone may correspond
to the extent of the lesion, but also extend beyond the lesion. In addition to curative
surgery, palliative procedures have been developed to avert secondary generalization.
Palliative surgery is indicated in cases of diffuse or multifocal seizure origin or
if the epileptogenic zone overlaps with eloquent cortical areas [13]
[14].
Epilepsy onset in early childhood is associated with considerable developmental deficits
as well as behavioral issues [15]. Furthermore, longer epilepsy duration is associated with unfavorable developmental
outcomes in affected children [16]. In recent years, numerous studies have emphasized the need for early intervention
in young children with severe early-onset epilepsy in order to prevent these comorbidities
[17]
[18]
[19]
[20]
[21]. Thanks to the latest technical advances, particularly in neuroimaging, neurosurgery
and neuroanesthesia, epilepsy surgery has been established as a safe and effective
therapeutic option, including the first years of life [22]
[23]
[24]
[25]. Despite the so far encouraging results, multicentric studies with longer observation
intervals are needed to confirm the positive effects of early epilepsy surgery on
children’s cognitive development, especially considering the diversity of etiologies
associated with a refractory course.
Duration from pharmacoresistance to referral for surgery
Despite the increased acceptance of epilepsy surgery in the last decades, there is
still a substantial delay between the establishment of pharmacoresistance and the
referral for presurgical evaluation in an epilepsy center with pediatric expertise
[11].The international pediatric epilepsy surgery survey from 2004 showed that only one-third
of children underwent surgery within the first 2 years after epilepsy manifestation,
whereas two-thirds of these children had presented with seizures in the first 2 years
of life [26].In a cohort study in the United States, the incidence of early-onset drug-resistant
focal epilepsy was 11.3 per 100 000 per year, while the incidence of epilepsy surgery
in the same cohort was 1.3 per 100 000 per year. Only 45 % of children with drug-resistant
epilepsy were referred to specialized epilepsy centers for further evaluation [9].
Indications for Epilepsy Surgery
Indications for Epilepsy Surgery
Although there are generally accepted principles of epilepsy surgery, the precise
goals of this intervention are individually defined [10]
[27]. Resective surgery is generally indicated when seizure semiology, EEG and imaging
findings allow for a precise delineation of the epileptogenic zone that can be removed
without unacceptable neurological and neuropsychological sequelae. The broad range
of childhood epileptic syndromes and underlying etiologies as well as the wide variability
of clinical and electroencephalographic seizure patterns in the respective age groups
define the spectrum of pediatric epilepsy surgery.
In a survey by the ILAE (International League Against Epilepsy) conducted in 2004
[26], focal cortical dysplasias (FCD) (42%) and glioneuronal tumors (19%) were the most
common etiologies among children and adolescents undergoing surgery in 20 epilepsy
centers worldwide. Furthermore, FCD was the most frequent etiology among young children
with extratemporal epilepsy. In children and adolescents undergoing temporal lobe
resections, FCDs and glioneuronal tumors, i. e. dual pathology, were observed significantly
more frequently than isolated hippocampal sclerosis. In addition, this survey provided
crucial information regarding the frequency of more rare epilepsy syndromes as well
as related etiologies in the pediatric population undergoing epilepsy surgery. Epilepsy
surgery was performed in less than 5% cases in the following rare syndromes: Landau-Kleffner
syndrome (<1 %), ESES (Electrical Status Epilepticus During Slow Sleep) (1%), Rasmussen
encephalitis (3%), Ohtahara syndrome (<1 %), West syndrome (4%), Sturge-Weber syndrome
(3%), and hypothalamic hamartoma (4%).
[Fig. 1]
[2] show the most frequent etiologies of drug-resistant structural epilepsies in infancy
and childhood.
Fig. 1 The most frequent etiologies of drug-resistant structural epilepsies in childhood.
a Right frontal focal cortical dysplasia, b Multiple cortical tubers in tuberous sclerosis, c Right hemimegalencephaly, d Left hippocampal sclerosis following meningo-encephalitis in infancy, e Right temporal glioneuronal tumor, f Extensive left frontal cortical scarring following a perinatal cerebrovascular incident.
Fig. 2 The most frequent etiologies of pharmacorefractive structural epilepsies in childhood.
a Right frontal cortical dysplasia, b Right parietal DNET, c Right hemispheric Rasmussen encephalitis, d Left hippocampal sclerosis, status after meningeal encephalitis in infancy, e Left parieto-occipital Sturge-Weber syndrome, f Extended left frontal substance defect after perinatal cerebral hemorrhage.
Focal cortical dysplasia
FCDs are the most common cortical malformations in epilepsy surgery and the most frequent
etiology in pediatric epilepsy surgery cohorts [26]. FCDs may present as focal, multi-focal or even hemispheric malformations. The FCD-related
epileptogenic zone is often more extensive than the MRI-visible lesion [28]. The most common localizations are frontal and temporal, although FCDs can affect
all lobes of the brain [29]. The completeness of resection of FCD-related epilepsies determines the postoperative
seizure outcome [30]. Despite considerable advances in imaging with development of epilepsy-specific
protocols, the accurate demarcation of FCDs remains challenging, especially in the
first years of life [17]. Prior to the onset of myelination in the first 6 months of life, there is a crucial
time window during which FCDs can be better imaged than in the following 1.5 to 2
years [31]. For this reason, epilepsy-specific imaging should be initiated early in young infants
at the first signs of a severe course. In the case of refractory course and electro-clinical
findings suggestive of focal epilepsy, the MRI should be repeated every 6 months,
if initially unremarkable, particularly after the age of 24–30 months[32]. FCDs that were not initially detected can become visible with advanced myelination
[33].
FCDs constitute a particularly complex group of malformations that vary substantially
with respect to anatomo-pathological localization and cellular presentation. In 2011,
a new ILAE classification was established to address this diversity of localization
and extent of the malformation [34]. FCD Type I is associated with a greater extent across multiple brain lobes as well
as with early onset of epilepsy [34]
[35]. FCD Type II is most frequently found in the frontal and parietal lobes, and can
manifest as a barely visible bottom-of-sulcus dysplasia or as an extensive lesion
comprising several gyri. Similarities have been recently reported at the cellular
level between FCD Type II , hemimegalencephaly and cortical tubers. The new ILAE Type
III includes FCDs associated with other lesions, such as hippocampal sclerosis, tumors,
vascular , post-traumatic or ischemic lesions.
Despite major diagnostic advances in recent years, FCD-associ- ated epilepsy remains
difficult to treat. Optimal outcomes can be achieved by using various diagnostic modalities:
62% of patients are seizure-free following surgery [28].
Tumors
Glioneuronal brain tumors (dysembryoplastic neuro-epithelial tumor: DNT, ganglioglioma)
are predominantly localized in the temporal lobe and are more frequently associated
with refractory epilepsy compared to all other pediatric brain tumors [36]. Other less common low-grade gliomas of children and young adults such as supratentorial
pilocytic astrocytomas, pleomorphic xanthoastrocytomas, and angiocentric gliomas are
also frequently associated with with drug-resistant epilepsy [37]. In addition, FCDs are often associated with these benign pediatric tumors, with
a variable prevalence in different studies [36]
[38]. The completeness of tumor resection determines postoperative seizure freedom [36]
[39], as long as there is no associated FCD. Epilepsy-specific imaging is needed to rule
out any adjacent malformations. In case of discordant seizure semiology or EEG findings,
an extraoperative invasive EEG recording or intraoperative electrocorticography (ECoG)
should be performed to determine the extent of the resection. Further predictors of
postoperative seizure freedom are age at surgery, duration of epilepsy and severity
of seizures [40].
Early surgical intervention also plays a significant role in this age group. Longer
epilepsy duration correlated with lower cognitive functioning at surgery in a recent
pediatric study [36]. This is of great importance, since presurgical functioning determines postsurgical
functioning. Cognitive stabilisation or even improvement are expected benefits of
successful epilepsy surgery in this context.
Tuberous sclerosis complex
90 percent of children with tuberous sclerosis present with epilepsy; two-thirds develop
pharmacoresistance [41]. Children with tuberous sclerosis and refractory epilepsy are potential candidates
for epilepsy surgery. The presence of multiple tubers, multiple seizure types and
multifocal epileptiform discharges in scalp EEG represent special challenges in the
presurgical evaluation of these patients. As a result, these patients are often excluded
from presurgical evaluation, or from surgical intervention. However, epilepsy surgery
can achieve seizure freedom in a subgroup of these complex patients, or at least provide
some improvement of their condition [42]
[43]. Overall, the results of epilepsy surgery in tuberous sclerosis patients are comparable
to those in extratemporal FCD-associated epilepsies. A meta-analysis has demonstrated
that 57 % of children are seizure-free after surgery, and an additional 18 % experience
a significant improvement in seizure frequency [44]. In these cases, 1–2 of the multiple tubers were identified as leading tubers and
resected. In a recent multi-center study, the extent of resection beyond the tuber
correlated with postsurgical seizure freedom [45]. This supports the notion that the epileptogenic zone is not limited to the tuber,
but also encompasses adjacent cortical areas. Diagnostic methods that can be particularly
useful in the presurgical evaluation of patients with tuberous sclerosis include interictal
alpha-11C methyl-L-tryptophan (AMT) PET (Positron Emission Tomography) [46]
[47]; ictal SPECT (single photon emission computed tomography) with SISCOM (Subtraction
Ictal SPECT co-registered to MRI) [48]
[49]; source analysis methods [50]; functional (f)MRI [51], as well as EEG-fMRI (simultaneous EEG and fMRI).
Polymicrogyria
Polymicrogyria (PMG) is one of the most common cortical malformations, and can be
associated with additional malformations such as FCD, hemimegalencephaly, dysgenesis
of the corpus callosum, cerebellar hypoplasia, periventricular heterotopia or schizencephaly
[52]
[53]
[54]
[55]
[56]. The underlying etiology is heterogeneous and comprises genetic as well as non-genetic
fac-tors [57]. PMGs can occur in the vicinity of ischemic lesions, in conjunction with congenital
infections (particularly cytomegalovirus infections), metabolic diseases (particularly
peroxisomal disorders) as well as gene mutations and copy number variations. PMGs
offer a variable clinical presentation depending on the etiology, syndrom-specific
characteristics, extent and localization as well as the presence of additional cortical
malformations [52]
[53]
[58]
[59]. The age of epilepsy onset varies between the neonatal period and late adulthood,
with most patients exhibiting hemiplegia, microcephaly, global developmental delay,
and particularly epilepsy, with most patients exhibiting hemiplegia, microcephaly,
global developmental delay, and especially epilepsy.
Previous studies regarding seizure freedom after surgery for refractory PMG-associated
epilepsy are sparse and their results are somewhat contradictory. This is most likely
due to the widely variable epilepsy syndromes, diagnostic methods and surgical procedures
considered in each study [52]
[53]
[55]
[58]. Postoperative seizure freedom in these studies ranges from 25 to 78 %. It remains
unclear whether the genetic etiology of certain PMG types [60] can affect postoperative seizure freedom. On the other hand, some PMG types are
apparently associated with a favorable long-term outcome despite the occurrence of
pharmacoresistance at some point in the course of the disease [61]. In such cases, epilepsy surgery should be considered with caution. Major challenges
in surgery for PMG-associated epilepsy are related to (1) the extent of bilateral,
multilobar or multifocal malformation, and (2) the presence of epileptogenic as well
as electrophysiologically normal cortex within the malformation, with the epileptogenic
zone often extending beyond the malformation [53]
[62]. Despite this inherent complexity, the outcome of epilepsy surgery is encouraging.
Surgical evaluation requires the use of different methods, including invasive EEG
recordings [53]
[54]
[63]. Two recent studies of invasive EEG recordings (SEEG: Stereo-electroence-phalography)
in PMG-associated epilepsy underlined the complexity of interrelations between the
epileptogenic zone and the PMG [53]
[62]. The epileptogenic zone can encompass the entire PMG or only a part of the PMG as
well as distant cortical regions. In some cases, the epileptogenic zone may be limited
to remote cortical regions. For this reason, seizure freedom can be achieved with
a focal resection even in patients with a multilobar or multifocal PMG, if only part
of the PMG is shown to be epileptogenic [64].
Hemispheric syndromes
Hemispheric syndromes, with refractory epilepsy as a consequence of extensive hemispheric
damage are classified as: (1) acquired (e. g. perinatal cerebrovascular incident,
hemiconvulsion hemiplegia syndrome, traumatic or infectious causes); (2) congenital
(e. g. FCD, PMG, hemimegalencephaly); and (3) progressive (e. g. Rasmussen encephalitis
and Sturge-Weber syndrome [65]
[66]. Hemispheric syndromes usually show hemispheric MRI abnormalities, seizure onset
in the damaged hemisphere and a contralateral permanent or progressive neurological
deficit (motor, language). The integrity of the contralateral hemisphere in hemispherotomy
candidates should be verified prior to surgery, since this determines both postsurgical
seizure freedom and stabilized or even improved cognitive development [67]. In cases of left-hemispheric pathology, it is important to verify language transfer
to the contralateral hemisphere by fMRI or Wada test [65].
MRI-negative epilepsy
MRI is one of the most important diagnostic tools in the presurgical evaluation [10]
[32]. An MRI-visible structural lesion determines the epilepsy classification and sets
the indication for presurgical evaluation and, eventually, for epilepsy surgery [9]
[68].The identification of a structural lesion depends on both the available imaging
methods (e. g. high-field MRI, postprocessing) and the experience of the neuroradiologist.
The proportion of MR-negative patients with refractory epilepsy that undergo presurgical
evaluation varies between 16% and 32% [69]
[70]
[71]
[72]
[73]
[74]. A survey by the ILAE Pediatric Epilepsy Surgery Survey Taskforce [26] demonstrated a clear lesion in 77% of patients, a subtle lesion in 6%, and no lesion
in 17%. A meta-analysis revealed a significantly higher proportion of MRI-negative
cases among children compared to adults (31 vs. 21%) [73]. MRI-negative epilepsy is related to poorer chances for a presurgical evaluation
and resulting resection [69]
[74]
[75]. A comprehensive non-invasive presurgical evaluation [76]
[77]
[78]
[79]
[80] is needed in order to form a valid hypothesis regarding the epileptogenic zone.
In the majority of MRI-negative focal epilepsies, invasive EEG recordings are required
in order to demarcate the epileptogenic zone [70]
[81]. In subdural EEG recordings, a multifocal or diffuse seizure onset is found 3 times
more often in MRI-negative patients compared to MRI-positive cases [82]. Although the likelihood of postoperative seizure freedom is generally reduced among
MRI-negative patients [73], it still amounts to 40–50 %, thus clearly supporting surgical intervention in relation
to conservative treatment with continuation of anticonvulsants [83]. Furthermore, some studies of surgery for FCD-associated drug-resistant epilepsy
report comparable rates of postsurgical seizure freedom for negative and MR-positive
patients [70]
[71]
[84].
Presurgical Evaluation
Scalp EEG recordings
Based on the guidelines of the ILAE Subcommission for Pediatric Epilepsy Surgery [10], interictal scalp EEG recordings including spontaneous sleep phases are a key component
of presurgical evaluation in childhood and adolescence. In addition, simultaneous
video-EEG recordings of epileptic seizures are strongly recommended. Particularly
in the case of infants and toddlers, sequential EEG recordings are often necessary
to establish or rule out a progression. EEG abnormalities in young children with focal,
multifocal or bilateral epilepsies are frequently extensive [17], and thus less useful regarding the localization of the epileptogenic zone. Moreover,
semiology of focal seizures n the first years of life, is less frequently lateralizing
and hardly ever localizing; the level of consciousness is hard to assess [85]. Focal epileptogenic lesions in the first year of life can manifest with infantile
spasms in the context of West syndrome [17]
[18]
[21].
Invasive EEG recordings
Invasive recordings are performed to obtain a precise demarcation of the epileptogenic
zone in MRI-negative epilepsy as well as in lesion-related epilepsy or for the functional
mapping of adjacent and overlapping eloquent cortical areas [86]. Two essentially different methods of invasive evaluation are available: recordings
using subdural strip or grid electrodes [36]
[87]
[88] and recordings using intracerebral depth electrodes [36]
[53]. Placement of grid electrodes requires an open craniotomy, whereas both strip electrodes
and depth electrodes can be inserted through burrholes. A combination of subdural
and depth electrodes is possible [89]. A recent study reports clinically-relevant complications in 9% of recordings using
subdural electrodes and in 6% of recordings with depth electrodes. Overall, postoperative
complications were reported for 48% of subdural recordings and 25% of recordings with
depth electrodes [90]. Use of subdural electrodes correlated with an increased risk of postoperative hemorrhage
and extraaxial collection. In addition, patients with prior craniotomies demonstrated
an increased risk of postoperative hemorrhage. Neurological deficits associated with
the invasive recordings were found in<5% of patients; less than 1 % were permanent
[90].
Selection of the appropriate method for invasive evaluation depends on individual
objectives and limitations. Recordings using depth electrodes enable the assessment
of deep structures such as the temporo-mesial areas, the insular cortex or subcortical
heterotopias. In addition, depth electrodes are used according to the principles of
stereo-encephalography (SEEG) in order to obtain a 3-dimensional illustration of the
seizure onset and early spread [91]. On the other hand, recordings with depth electrodes are less suitable for the evaluation
of extensive, multilobar malformations or for functional mapping of eloquent cortex.
In these cases, subdural recordings are preferable. In the first 3 years of life,
recordings with depth electrodes are technically impracticable, since the electrodes
cannot be sufficiently fixed to the points of entry due to the particularly thin skull
bones [92]
[93].
Intraoperative electrocorticography (ECoG) [94], is limited to relatively brief recordings in the operating room, thus providing
almost exclusively interictal data. Awake surgery, as performed on adults in some
epilepsy centers, is unfeasible in pediatric epilepsy surgery due to the lack of cooperation.
Moreover, intraoperative recordings are limited to the exposed cortical surface during
surgery, whereas deep structures such as the temporal-mesial areas or the insular
cortex remain unaccessible.
Imaging
Due to immature myelination in the first 2 years of life, interpretation of imaging
is fraught with difficulty, particularly with respect to the diagnosis of FCDs. Repeated
MRI examinations are necessary for this age group in order to identify possible cortical
malformations during postnatal brain development.
In addition, functional imaging, such as interictal positron emission tomography (PET),
interictal or ictal single-photon emission computed tomography (SPECT), magnetoencephalography
(MEG) or functional MRI can provide useful data regarding the localization of the
epileptogenic zone. The Wada test can contribute significantly to language lateralization
[95].
Finally, source analysis deriving from scalp or invasive EEG recordings as well as
MEG recordings can be applied, combined with MRI-based head models for optimal visualization
[87]
[88]
[96]
[97].
Neuropsychology
Detailed neuropsychological testing can supplement localization of the epileptogenic
zone by identifying deficits while objectifying the functional significance of the
affected area, so that the risk of postoperative neuropsychological deficits can be
estimated and reduced. This is particularly important, since children with drug-resistant
epilepsies who can benefit from epilepsy surgery frequently have accompanying developmental
disorders and behavioral problems [15]
[17]
[65]
[98].
Concept of cortical zones
The epileptogenic zone is the cortical region that is critical for the generation
of epileptic seizures [12]. The aim of epilepsy surgery is the complete resection of the epileptogenic zone,
sparing the eloquent cortex areas. The complete resection or disconnection of the
epileptogenic zone is the necessary and sufficient prerequisite for the achievement
of seizure freedom that should be retained even after reduction and discontinuation
of anticonvulsive medication. The epileptogenic zone is thus a theoretical concept.
The irritative zone is the cortical region generating interictal epileptiform potentials
that can be recorded by scalp and intracranial EEG, MEG and EEG-fMRI. The seizure
onset zone is the cortical region where epileptic seizures originate. It is determined
using scalp and intracranial EEG or ictal SPECT. The symptomatogenic zone is the cortical
region that generates the initial ictal symptoms and which can be localized based
on an analysis of seizure semiology. An epileptogenic lesion is delineated using high-resolution
MRI and may contribute directly to seizure onset due to its intrinsic epileptogenicity,
as in the case of FCD, or be secondarily involved by propagation from the surrounding
cortical regions. The functional deficit zone is the cortical region that demonstrates
functional abnormalities in the interictal period, as indicated by the findings of
the neurological and neuropsychological examination.
Resection is generally indicated when (1) seizure semiology, long-term video EEG and
imaging findings allow for a clear delineation of the epileptogenic zone, and (2)
resection of this area is possible without unacceptable neurological and neuropsychological
sequelae. The broad range of childhood epileptic syndromes and related etiologies
as well as the wide variability in seizure semiology and electroencephalographic seizure
patterns encountered in different age groups account for the main distinctive features
of pediatric epilepsy surgery.
Developments in pediatric epilepsy surgery
Bilateral EEG and MRI findings
Recent developments in the classification of epilepsy syndromes as well as in surgical
techniques have expanded the range of epilepsy surgery indications and available procedures
[99]. It has been shown in recent years that postsurgical seizure freedom is attainable
even in patients with bilateral extensive and diffuse EEG abnormalities and extensive
multifocal or multilobar lesions. These observations apply to cases of (1) early-onset
catastrophic epilepsy associated with a focal lesion that frequently present with
generalized seizure semiology and electroencephalographic abnormalities [17]
[36]
[87], (2) tuberous sclerosis with multiple tubers, with the resection of the leading
tuber resulting in -at least temporary- seizure control [17], and (3) polymicrogyria, with a partial resection guided by invasive EEG recordings
resulting in seizure freedom [53].
Extent of resection
In a single-center study involving 580 cases [100], the outcomes of pediatric epilepsy surgery in Los Angeles in 1986–1997 were compared
to those of 1998–2008. In the later time period, the number of curative as well as
palliative procedures per year almost doubled, reflecting the dynamic development
in pediatric epilepsy surgery. More focal resections were performed in contrast to
the dwindling rates of multilobectomies, more patients with tuberous sclerosis were
operated, and only rarely did histopathology provide evidence for non-specific gliosis.
Moreover, the number of invasive EEGs declined, the outcomes improved and the complication
rate decreased. The improvement of seizure outcomes has been ascribed to recent advances
in non-invasive presurgical evaluation [87]
[88], to the improved selection of surgical candidates [65], and particularly to the objective of a complete resection of the epileptogenic
lesion [39]. The authors commented that, in the past years, incomplete resection was performed
to avoid neurological deficits. For example, temporo-parieto-occipital resections
have been performed in children and adolescents with hemispheric epileptogenic lesions
and incomplete hemiparesis. In the long-term follow-up, the majority of patients with
incomplete resections of the epileptogenic region suffered recurrent seizures and
underwent reoperation relatively soon after the first surgery. In recent years complete
resections have gained ground, especially in young children at risk of epileptic encephalopathy,
despite the involvement of sensory or motor cortex in some cases [100].
Surgical Interventions
According to the ILAE [26] survey of epilepsy centers with a pediatric focus in the USA, Europe and Australia,
multilobar and hemispheric resections or disconnections dominate in childhood, while
temporal resections account for only 25% of surgeries. This is in stark contrast to
adult cohorts with predominantly temporal lobe surgeries and only few extratemporal
and hemispheric resections. The extent of the required resection decreases with the
age of the surgical candidates.
Focal cortical resection
The epileptogenic zone and, consequently, the extent of the resection is defined individually
depending on the findings of the presurgical evaluation. The complete removal of the
epileptogenic zone is the prerequisite for postoperative seizure freedom. In FCD,
this corresponds to the complete resection of the epileptogenic lesion. The completeness
of FCD resection determines postsurgical seizure freedom [30]. Seizure outcomes following temporal lobe epilepsy surgery are more favorable than
the outcomes of extratemporal surgery. Children with MRI-negative epilepsy reportedly
have lower chances of seizure freedom. However, the proportion of MRI-negative cases
has significantly decreased in recent years, due to advances in imaging technologies
[11]. In view of the particularly favorable results in children with clearly delineated
epileptogenic lesions, e. g. glioneuronal brain tumors, lesionectomy should be undertaken
relatively early in the course of epilepsy [36].
Temporal lobe epilepsy surgery
Temporal resections account for a quarter of surgical interventions for epilepsy among
children and adolescents [26]. Younger age at surgery correlates with higher rates of seizure freedom and more
favorable cognitive development [101]. A recent pediatric epilepsy surgery study with follow-up to adulthood reported
85 % seizure-freedom as well as a significant increase of IQ [102]. The improvement in IQ was demonstrated 6 years or longer after epilepsy surgery
and was associated with the discontinuation of anticonvulsants.
Glioneuronal brain tumors (40%) and FCDs (30%) represent the most frequent etiologies
in temporal lobe epilepsy surgery, followed by hippocampal sclerosis (22%) [26]. In contrast to adult cohorts, isolated hippocampal sclerosis appears in only 7
% of children undergoing surgery. In pediatric cohorts hippocampal sclerosis is diagnosed
mainly in adolescents and hardly ever in the first years of life [17]. Dual pathology, i. e. coexistence of hippocampal sclerosis with an additional epileptogenic
lesion, is found more frequently in children than adults. In most cases, hippocampal
sclerosis is associated with FCDs [28]. Other, less frequent etiologies in temporal lobe epilepsy surgery include polymicrogyria,
phacomatosis, atrophic or ischemic lesions [36]
[53].
An important issue in temporal lobe epilepsy surgery is the indication of an additional
resection of mesial temporal structures in the presence of a clearly delineated epileptogenic
lesion, such as a glioneuronal tumor [36]. This issue can often be clarified only with invasive EEG recordings using depth
electrodes [53]. This is crucial, since the resection of mesial temporal structures without clear
indications of pathology can produce severe cognitive deficits. The results of temporal
lobe epilepsy surgery in cases of dual pathology are comparable to those of selective
amygdalohippocampectomy, provided that the complete resection of the epileptogenic
zone is feasible. The incomplete resection of the epileptogenic zone due to the risk
of functional deficits strongly correlates with seizure recurrence in long-term follow-up
[39].
Extratemporal epilepsy surgery
Extratemporal resections account for 20% of pediatric epilepsy surgery [26]. Despite advances in presurgical evaluation with improved selection of suitable
candidates, extratemporal epilepsy surgery is less successful than temporal lobe resections
[100].However, outcomes of extratemporal epilepsy surgery in children and adolescents
have considerably improved in recent years and are superior to those in adult cohorts
[11]. Cortical malformations represent the leading etiology [28].
Frontal lobe epilepsy surgery:
Frontal lobe resections account for 36% of intralobar epilepsy surgery in children
and adolescents [26]. Postoperatively, up to 66 % of patients remain seizure-free, although outcomes
can differ according to etiology, localization as well as duration of follow-up. An
epileptogenic lesion clearly demarcated in MRI that can be completely resected, due
to lack of overlap with eloquent cortical areas, is linked to a particularly favorable
prognosis. However, electroencephalographic findings in structural frontal lobe epi-lepsy
are often of little help and can lead to the mislateralization and mislocalization
of the epileptogenic zone, or even to misclassification of the epileptic syndrome
as genetic epilepsy [87].
In a recent study of 158 patients who underwent frontal lobe epilepsy surgery [103] both younger age at surgery (<18 years) and shorter epilepsy duration (<5 years)
correlated with higher rates of seizure freedom. In this study, children presented
significantly higher rates of seizure freedom compared to adults, especially after
early intervention, although FCD was the prevailing etiology in this subgroup. These
results can be attributed to the protective effect of early surgery in preventing
secondary epileptogenesis. Furthermore, more radical surgery is possible in the first
years of life, since neurological deficits can be compensated by increased functional
plasticity. This underpins the necessity of early presurgical evaluation and surgical
intervention in suitable pediatric candidates.
Posterior cortex epilepsy surgery:
Resections in the parietal and/or occipital lobe, sometimes including the posterior
part of the temporal lobe, account for 10% of intralobar epilepsy surgery in children
and adolescents [26]. Presurgical evaluation and epilepsy surgery in these brain regions present a significant
challenge, due to the rapid connections to other cortical areas and to the functionality
of the cortex. Several studies have highlighted the difficulty of delimiting the epileptogenic
zone within the parieto-occipital and posterior temporal regions, owing to the often
non-specific seizure semiology which, due to rapid propagation, can be mislocalized
within the anterior brain regions. In childhood, this challenge is all the greater
since electroencephalographic abnormalities are frequently diffuse or generalized
and offer little evidence for the localization of the epileptogenic zone.
Postsurgical seizure freedom ranges between 43% and 86%, depending on the particular
cohort and follow-up time, although few studies have been conducted in exclusively
pediatric cohorts. In the largest pediatric study of posterior cortex epilepsy surgery
to date, including 62 children and adolescents [104], 86% of patients remained seizure-free in the long-term follow-up of 2–16 years.
Postoperative seizure control strongly correlated with a well-delineated, surgically
accessible epileptogenic zone. In addition, later epilepsy onset was associated with
higher rates of seizure freedom. This may be attributed to the susceptibility of the
immature brain to the establishment of extensive epileptic networks, well beyond the
epileptogenic lesion. A recent study [105] demonstrated the efficacy of posterior cortex epilepsy surgery: 60% of the patients
remained seizure-free in the long-term follow-up of 1.5–18 years, 30% were off drugs.
Longer epilepsy duration at surgery was identified as the sole independent predictor
of seizure recurrence. These results support the early consideration of surgical intervention
in children and adolescents with drug-resistant epilepsy originating in the parietal
and/or occipital lobes.
Hemispherotomy
Hemispherotomy, i. e. the functional disconnection of a cerebral hemisphere, currently
accounts for 20–40% of resections in pediatric epilepsy surgery [26]. This procedure is considered in children with hemispheric damage due to a congenital
(e. g. FCD, PMG), acquired (e. g. perinatal infarct) or progressive (e. g. Rasmussen
encephalitis) etiology, drug-resistant epilepsy and neurological deficits such as
hemiparesis and hemianopsia.
The vast majority of children and adolescents undergoing hemispherotomy remain seizure-free
postoperatively [11]. The etiology was identified as the main predictor of seizure freedom [65]. Congenital lesions, especially hemimegalencephaly, are associated with less favorable
long-term seizure outcomes. Electroencephalographic or MR-tomographic abnormalities
contralateral to the affected hemisphere are generally considered as negative predictors
of postsurgical seizure freedom. However, in a recent study it was shown that even
contralateral EEG and MRI abnormalities are compatible with postoperative seizure
freedom in selected patients [65].
The postsurgical cognitive development of affected children is determined by seizure
control, epilepsy duration and presurgical cognitive development. Furthermore, there
are indications that postsurgical cognitive functions correlate with the integrity
of the contralateral hemisphere.
A recent study [65] identified a subgroup of patients with ac-quired etiology, older age at surgery
and excellent seizure outcomes as “double winners”. These results should encourage
pediatric neurologists to set aside their reservations regarding possible functional
loss and to consider hemispherotomy for older, less affected children or adolescents,
especially since these patients might particularly benefit from this surgery.
Special Features of Pediatric Epilepsy Surgery
Special Features of Pediatric Epilepsy Surgery
Epilepsy surgery in the first years of life
Epilepsy surgery is an established treatment method for selected pediatric candidates
of all age groups with drug-resistant structural epilepsy. However, diagnosis in the
first few years of life is fraught with limitations with respect to the MR-tomographic
demarcation of epileptogenic lesions. A further challenge in the presurgical evaluation
in early childhood are the often diffuse or bilateral interictal epileptiform discharges
and electroencephalographic seizure patterns that may arise from focal brain lesions
[39]
[65]. Despite the vast progress in neurosurgery, anaesthesia and intensive care, increased
risks are assumed due to the limited blood volume and immature physiology of the developing
brain and the extensive procedures often required to achieve seizure freedom in this
age group. Nevertheless, there is a trend in offering surgery as soon as intractability
is ascertained, including the first years of life
This view is based on the high incidence of epilepsy in the first year of life with
one-third of affected children developing pharmacoresistance [6] and on the severity of epilepsy syndromes in this age group [26]. The possibility of epilepsy surgery should be explored at an early stage, considering
the cognitive and behavioural impairment associated with early epilepsy onset, longer
epilepsy duration, high seizure frequency, continuous epileptiform discharges and
polytherapy. Epilepsy surgery during the first years of life benefits from the plasticity
of the immature brain. This approach is supported by studies that demonstrate (1)
the superiority of surgical treatment compared to pharmacotherapy [19],and (2) the prospects of seizure control and developmental benefits resulting from
early intervention [18].
Pediatric epilepsy surgery aims toward seizure control and reduction or discontinuation
of anticonvulsants. In view of the devastating effect of seizures and anticonvulsants
on the developing brain, a successful surgical intervention is expected to provide
long-term benefits regarding cognitive development. After epilepsy surgery, two thirds
of children remain seizure-free at long-term follow-up and present a stable [17] or even improved [18] cognitive development. Parents frequently report dramatic improvements in quality
of life and an amelioration of social adaptation deficits that go along with seizure
control. Morbidity and mortality are low, although extensive resections are often
performed. Very young age is not a contraindication for epilepsy surgery.
Reoperations
The high rate of reoperations is another particular aspect of pediatric epilepsy surgery
. This may be partly attributed to the inherent challenges in identifying and delineating
FCD, the most prevalent etiology, in MRI, thus increasing the risk of an incomplete
first resection. Ten percent of epilepsy surgery in children and adolescents involve
reoperations [11]. This is especially relevant in cases of catastrophic epilepsy in the first years
of life. After a first epilepsy surgery, 30–40% of patients continue to suffer seizures;
however, only 6–21% of these children and adolescents undergo reoperation [17]
[39].
A recent study of reoperations in children and adolescents reported seizure freedom
in 61 % of patients, whereas altogether 83% (Engel I and II: [106]) significantly
benefited from reoperation [106]) significantly benefited from reoperation [39]. These outcomes are considerably more favorable compared to those previously reported
for adult cohorts with only 20-40 % seizure freedom after reoperation. This is most
likely due to the disparities in the underlying pathology and its localization in
different age groups. These particularly favorable outcomes of reoperation in childhood
derive from the improved MR-tomographic detection of epileptogenic lesions as well
as from the latest technological developments. These recent advances have significantly
contributed to the accuracy of presurgical evaluation and surgical treatment and expanded
the spectrum of potential candidates [28].
Incomplete resections due to fear of functional deficit are often encountered in cases
of extensive multilobar or even hemispheric dysplasia, if the epileptogenic zone overlaps
with eloquent cortical regions [17]
[65]. The decision for multilobar or hemispheric surgery is further hampered by the insufficient
MR-tomographic delimitation of FCD in infants and toddlers, especially in the absence
of a neurological deficit. Reoperation in the form of multilobectomy or hemispherotomy
offers excellent chances for seizure freedom [39].
Children and adolescents with recurrent seizures after an initial surgery for epilepsy
can substantially benefit from reoperation, especially following the incomplete resection
of extensive cortical dysplasia. The possibility of reoperation should be promptly
considered following seizure recurrence, which in most cases manifests within the
first 3 months after epilepsy surgery. Reoperation, when indicated, should be performed
at the earliest possible time so that the child can benefit from the higher functional
plasticity in early life to compensate for possible neurological deficits.
Discontinuation of anticonvulsants
The goals of pediatric epilepsy surgery include seizure freedom, discontinuation of
anticonvulsants and improvement of cognitive development [10]. In the first years of life, anti-convulsants can negatively influence brain development
by triggering neuronal apoptosis and preventing neurogenesis, synaptogenesis, cell
proliferation and migration as well as synaptic plasticity [107]
[108]. In addition, several anticonvulsants have well-known cognitive side effects, particularly
affecting attention, vigilance and psychomotor speed [109]
[110]. The cumulative effects of polypharmacotherapy during crucial phases of cognitive
development can have a critical influence in affected children and adolescents. The
prospect of developmental improvement can encourage anticonvulsant discontinuation
in seizure-free patients. A retrospective multicenter study demonstrated that early
discontinuation of antiepileptic drugs has no influence on long-term seizure freedom
[111]. In some cases, early discontinuation of anticonvulsants can reveal a less favorable
outcome. Thus, children and adolescents requiring anticonvulsants in the long-term
can be identified, whereas others can be spared unnecessary continuation of drug treatment.
A further retrospective multicenter study [112] showed that the initiation of antiepileptic drug withdrawal, the number of drugs
reduced, and the complete anticonvulsant withdrawal were associated with improved
postoperative cognitive functions.
Summary
Epilepsy surgery is an effective treatment option for children and adolescents with
drug-resistant structural epilepsy. The vast majority of patients benefit substantially
from surgical treatment. Very young age, severe developmental delay and psychiatric
comorbidities are no contraindications for epilepsy surgery. Likewise, a normal routine
MRI should not discourage from pursuing presurgical diagnosis in cases of drug-resistant
epilepsy with focal seizures. The extent of the required resection decreases with
age. Younger children often require more extensive resections. Postsurgical neurological
deficits are generally well-compensated thanks to the increased functional plasticity
of the brain. On the other hand, in adolescents with a presurgical neurological deficit,
epilepsy surgery can offer excellent chances for seizure freedom and favorable neurocognitive
and psychosocial development. In addition to the development of non-invasive methods
for presurgical evaluation, it is of utmost importance to shorten the interval between
determination of pharmacoresistance, presurgical evaluation, and epilepsy surgery
in suitable candidates, in order to improve postsurgical cognitive outcomes. Multicentric
studies with longer observation intervals are needed to identify predictors of seizure
freedom and improved cognitive development in affected children. This would facilitate
the selection of surgical candidates and improve presurgical counseling of patients
and their families. Discontinuation of anticonvulsants after successful epilepsy surgery
can lead to cognitive improvement without compromising long-term seizure freedom.
