ARTICLE
Auteur(s) : Lara E
Jehi1, Hans Ö Lüders2, Richard Naugle1, Paul Ruggieri1, Harold Morris3, Nancy Foldvary1, Elaine Wyllie1, Prakash Kotagal1, Bill Bingaman1, Dudley Dinner1, Richard Prayson4, Beate Diehl1, Andreas Alexopoulos1, Jocelyn Bautista1, Robyn Busch1
1Epilepsy Center, Cleveland Clinic Foundation
2Epilepsy Center, University Hospitals, Cleveland
3Department of Neurology, University of Vermont College
of Medicine, Burlington, VA
4Department of Pathology, Cleveland Clinic Foundation,
Ohio, USA
Article reçu le 5 Octobre 2007, accepté le 22 Novembre 2007
Presentation of the “March 2008 Cleveland case”
A 26-year-old, right handed woman was seen in the outpatient
epilepsy clinic for the evaluation of recurrent seizures. Her first
spell was a generalized tonic-clonic (GTC) convulsion that occurred
while awake, at 10 years of age. Subsequently, she remained
seizure-free and off antiepileptic drugs, until 17 years of age,
when she had her second convulsion. Phenytoin was then initiated.
The GTC resolved, however, a new type of event began. These would
start with an unusual “sweet smell” or a feeling of “butterflies in
her stomach” followed by loss of awareness, during which observers
described her as “staring”, “chewing”, and “talking out of
context”. These seizures lasted 1-3 minutes, and occurred twice per
week, despite treatment with phenytoin and levetiracetam, and prior
treatment with carbamazepine and oxcarbazepine. The patient had no
prior history of CNS infections, head trauma, or febrile seizures.
Her past medical and surgical history and her physical examination
were unremarkable. She had an uncle with a history of seizures. She
had no mental health problems, but was not driving or working for
fear of recurrent seizures.
Neuropsychological testing revealed that she was generally
functioning in the low average to average range with some mild
difficulties in intellectual functioning and confrontation naming.
Her brain MRI is shown in (figure 1).
Radiological diagnosis
Dr Ruggieri: “This MRI shows an isolated, non-enhancing lesion in
the dorsal aspect of the right temporal lobe at the junction of the
middle and superior temporal gyri. This lesion is heterogeneous in
intensity on FLAIR and T-2, and has a cystic component in the
underlying white matter without significant mass effect.
Hyperintensity seen on unenhanced T-1 suggests a component of
dystrophic calcification. The wedge shape of the lesion suggests a
malformation of cortical development (MCD), yet its striking
heterogeneity suggests a neoplasm.
The underlying mesial temporal structures appear normal and
symmetric in signal intensity characteristics, size and
morphology.”
Further investigation
Dr Lüders: “Do you feel that any further investigations are needed,
or should the patient be directly referred to surgery for removal
of the lesion? ”
Dr Morris: “She needs a video-EEG evaluation for several
reasons. First, we need to confirm that the patient’s episodes are
indeed seizures, by recording the corresponding electrical
abnormalities. Second, we need to decide what tissue to remove:
obviously the neocortical lesion has to be resected because it
appears to be a tumor, a ganglioglioma or more likely a
dysembryoplastic neuroepithelial tumor (DNET), but her seizure
semiology, specifically her olfactory aura suggests additional
involvement of her mesial temporal structures raising concerns for
dual pathology. A normal hippocampus on MRI does not exclude it
being pathologically abnormal or epileptogenic. If her hippocampus
is involved, removing the MRI lesion will not make her
seizure-free. ”
A noninvasive video-EEG evaluation was performed, using the
10-20 international system, with additional sphenoidal electrodes.
Interictal sharp waves were seen in the right mesial temporal
region, maximum at the SP2 electrode (frequency: 1 every 5-10 sec,
90% of her overall spikes). In addition, isolated sharp waves were
seen in the left mesial temporal region, maximum at the SP1
electrode (10%). EEG source analysis of the interictal spiking
corroborated a mesial temporal origin. One clinical seizure lasting
90 seconds was recorded. The patient pushed the seizure button to
alert the nursing staff, but lost awareness before she could
describe her exact aura. This was associated with manual and oral
automatisms and followed by postictal confusion. Ictal EEG started
with a rhythmic delta activity in the right mesial temporal region,
which spread within three seconds to the left temporal region, and
was obscured seven seconds later by muscle artifact (figure 2).
An intracarotid Amobarbital test was consistent with a left
hemispheric dominance for speech, and bilateral memory
representation (she had a 67% retention score and remained
nonverbal for 59 sec following left internal carotid injection,
while she never lost her speech and had a 75% retention score
following right internal carotid injection).
Management
Invasive evaluation
Dr Foldvary: “She has right temporal lobe epilepsy (TLE). The
question is whether it is neocortical or mesial. The answer to this
question determines whether a lesionectomy alone is an adequate
treatment option, or whether the mesial temporal structures should
also be removed. Her MRI lesion is posterior temporal but her
interictal spiking is right anterior temporal. She also has left
temporal interictal abnormalities and an ictal pattern that appears
early in the left mesial temporal region. Her epilepsy could then
be originating from the surface or depths of the lesion with spread
to involve mesial temporal structures bilaterally. A neocortical
temporal lobe lesion can often be associated with ipsilateral
mesial temporal spikes and we may not even see any posterior
temporal spikes. I suppose if the duration of the epilepsy is long
enough, as in this case, you could see some degree of bilaterality
as with mesial TLE. However, the semiology of an olfactory and
abdominal aura suggests that this could all be of mesial origin.
I would proceed with an invasive evaluation to study the
relationship between her neocortical lesion and her mesial temporal
structures. This could be designed in a number of ways: one way
would be to just use subdural grids to cover as much as possible of
the right temporal lobe, the lesion and getting close to the mesial
structures. A better way would be to combine those grids with
depths inserted in her right mesial temporal structures. Depending
on how strongly you feel we need to investigate the left temporal
abnormalities, you can go one step further and also implant depths
and grids over the left temporal lobe. The reality is though that
we are not going to operate over the left side, so we could
probably forego that.”
Dr Wyllie: “I’d like to hear from our neuropsychologist about
the risks of removing the hippocampus in this non-dominant temporal
lobe?”
Dr Naugle: “She was able to achieve average scores in measures
of her visuospatial abilities, which would suggest a risk of
decline in those measures postoperatively. The issue though is that
this correlation between preoperative abilities and postoperative
changes is less strong in individuals undergoing a non-dominant
temporal resection.
Even if she shows a decline in her testing scores, it is
unlikely that she will complain of any memory change in her daily
life.”
Dr Wyllie: “Given the potential risk of neuropsychological
changes associated with removing the hippocampus, an invasive
evaluation would be additionally helpful as it may allow us to
spare her mesial structures if they prove to be non-involved.”
Dr Kotagal: “Even an invasive evaluation using a combination of
grids and depths has its limitations though. One problem is that we
don’t know how to modify our surgical plans based on results of
such evaluations in patients with dual pathology. We don’t know if
spread occurs to the mesial temporal region within for example 5
sec versus 10 sec; whether that means we have to remove the
hippocampus or not.”
Dr Dinner: “Even ictal patterns seemingly starting in the
hippocampus could actually be originating from non-monitored
regions within the depth of the lesion.”
Dr Bingaman: “We also need to discuss the risk of having
subdural electrodes and depths implanted. There is 20-30% chance of
having some sort of a problem, including stroke and hemorrhage
which could be devastating. It is also costly: $25 000 for a
non-invasive video-EEG. Here with two surgeries, the cost may be up
to $100 000. Although we don’t always think about cost, it should
be factored in.”
Lesionectomy
Dr Morris: “Given the limitations of an invasive evaluation, and
the 70% chance of seizure-freedom quoted in most papers following
lesionectomies, we could make an argument for just doing a complete
resection of the lesion at this point and hoping for the best. I
previously looked at a series of failed DNETs that required
reoperation, and in all of them, on reoperation, there was either
residual tumor or dysplastic changes (Morris et al. 1993). If
seizures recur in this patient after a lesionectomy, I would think
they’re more likely to originate from residual tissue around the
lesion, than from her mesial temporal structures, but we would then
need to evaluate both with an invasive evaluation.”
Dr Bingaman: “An important question is what is this lesion? Is
it neoplasm, or is it dysplasia? And if it is neoplasm, is there an
associated dysplasia that extends perhaps throughout the temporal
lobe, including the mesial structures. If dysplasia is involved,
the chance of seizure-freedom would probably be less than the 70%
quoted. Another issue is that lesionectomies can be tailored
intraoperatively according to how the temporal lobe feels on
palpation. A dysplastic tissue “feels” different. The problem here
is that with a lesion that is so far posterior, we cannot feel the
amygdala intraoperatively and tailor the lesionectomy accordingly.
Intraoperative EcoG would not be very helpful because the mesial
structures are not something you can sample easily in the operating
room.”
Right temporal lobectomy
Dr Wyllie: “Another possible approach, given that this is her
non-dominant temporal lobe, might be to just do a large right
anterior temporal lobectomy, removing both the lesion and the
hippocampus. The problem with this approach in this case is that
the lesion is located too far posteriorly. In cases where the
neocortical lesion is closer to the hippocampus, the decision to
remove both together is easier.”
Dr Bingaman: “A temporal lobectomy could be presented to the
patient as a method of removing the EEG abnormality, in addition to
the MRI lesion. But the question is then: are you removing normal
brain that can be spared? To answer that, you can do some form of a
subdural and depth evaluation. This would be the most scientific
option, to try to define the relationship of the lesion to the EEG.
In view of how posterior the lesion is and how remote it is from
the mesial structures, and how anterior and mesial the ictal EEG
looks, I would probably favor an invasive evaluation.”
Actual treatment: stereotcatically guided lesionectomy.
Pathological diagnosis
Dr Prayson: “The pathology of this lesion is interesting. It does
not fit into any of the current classification systems used for
either dysplasia or neoplasia. On the whole, I think it is best
classified as a tumor (figure 3).
It was multinodular, with sections containing roundish cells
with perinuclear clearing and capillary vascular patterns,
occasional larger neuronal or ganglion type cells without much
atypia, features characteristic of a DNET rather than a
ganglioglioma which tends to be more of a uninodular mass. Other
sections showed very cellular, well circumscribed nodules, with a
mixture of both atypical glial and neuronal cells in one region.
Similar foci can be encountered in a low grade glioma or mixed
astrocytoma, or in a ganglioglioma. Sections from adjacent gyri
ranged from reasonably normal cortex to areas of very abnormal,
very cellular, disturbed architecture, typical of Type Ia cortical
dysplasia.
I believe the lesion represents a tumor with a component of
cortical dysplasia adjacent to it. We did a K67 immunostain as a
marker of cell proliferation, there was little staining
(< 1%), typical of DNETs or gangliogliomas. This particular
tumor is multinodular with areas that look more like DNET and
others more like a ganglioglioma. It appears to be a hybrid. Such
cases have previously been documented in the literature.
I can not make comments on whether the margins of the resection
were free from any cellular changes because we don’t usually
marginate those lesionectomies.”
Dr Ruggieri: “You can see a very focal resection in the dorsal
aspect of that superior temporal gyrus. There is residual
hyperintensity on FLAIR along the dorsal margins of the resection
and along the mesial aspect of the lesion. It is, however,
difficult to discern how much of this hyperintensity relates to
surgical manipulation versus residual abnormal tissue. My guess is
the mesial hyperintensity at least, is a portion of the underlying
abnormal tissue seen preoperatively projecting towards the
ventricle.”
Postoperative course: typical seizures recurred five weeks after
surgery, with an unchanged frequency, duration and semiology.
Routine EEG showed right temporal spikes (Max FT10).
With neuropsychological testing, all scores were essentially
unchanged when compared to preoperative values except for less
efficient reading ability and problem solving, an unexpected
finding given that she was previously tested using the same
measures. This may have occurred because she was quite fatigued
during the testing session.
Further management
Dr Bingaman: “There is residual dysplasia in the deep white matter
underlying the lesion. Resection there puts her at risk for visual
field loss. EEG still shows the same findings. The discussion of
whether to do a large temporal lobectomy versus invasive evaluation
still applies. Typically, after a first failed surgery we do
invasives, so I think that’s a reasonable course of action.”
Dr Morris: “Based on my prior series, I am sure there is
residual dysplasia. The problem is in defining the borders of the
dysplasia: she has already had surgery there, so the margins will
not feel normal to the surgeon, and they already removed what
looked grossly abnormal on the MRI, so it will be harder to
determine margins for a new lesionectomy. You don’t have an option
to define those margins except by doing an invasive evaluation,
acknowledging that interictal spikes may just reflect neuronal
irritation from the grids and do not necessarily have to come
out.”
Dr Wyllie: “I am more interested in the postoperative MRI. The
hole is much smaller than I anticipated based on the preoperative
evaluation. It seems there are issues related to visual field loss.
Were there other elements that contributed to this rather small
resection?”
Dr Bingaman: “The mass effect present in the preoperative study
is resolved, so it contributes to the resection looking smaller
than it actually is. There is also the fear of injuring visual
radiations in the deep white matter. The parietal extent of the
lesion is around the sylvian fissure which is a highly vascularized
structure, another potentially complicating factor.”
Dr Bingaman: “How long do you wait in between evaluations?”
Dr Morris: “A period measured in weeks or months, but not years.
I don’t think there is an exact time.”
Dr Bingaman: “According to Sperling’s data, there was increased
mortality risk in patients who failed a first surgery, especially
when they their typical seizures recurred and at high frequency. So
it might be useful not to wait too long.”
Dr Lüders: “Usually we wait six months. If seizures recur like
this, the chance of controlling them with medication is virtually
zero.”
Actual management
There was a six-month lag between the initial surgery and an
invasive evaluation that consisted of subdural grids only (figure 5). No depths
were used.
Interictal spiking was predominantly anterior temporal (figure 6).
A typical seizure was recorded. This started with an olfactory
aura, followed by unresponsiveness, manual and oral automatisms,
followed by postictal confusion.
The EEG onset was characterized by a rhythmic fast activity that
evolved in amplitude over the lateral temporal contact SA17,
followed by spread within 13 seconds to the anterior part of the
lateral temporal A plate and the entire, subtemporal C plate. The
frontoparietal and posterior temporoparietal plates remained
electrically quiet (figure 7). The electrode
of ictal onset (SA17) was shown to be 3-4cm anterior to the margin
of the lesionectomy by MRI reconstruction.
Proposed treatment
Extended lesionectomy
Dr Kotagal: “I am concerned because the ictal rhythm is starting
3-4cm anterior to the lesion in an area that appears normal on
imaging. Is it then really starting from the depths of the lesion
and just spreading to the apparent ictal onset? In that case, one
might consider just performing a bigger resection of the tumor and
surrounding gliosis while sparing the rest of the temporal lobe.
This might also reduce the risk of a quadrantanopsia associated
with a bigger resection.”
Dr Bingaman: “If we remove all the white matter underneath the
lesion and go deep to the ventricle, there is a risk of some visual
field loss. Nowadays we would perform diffusion tensor imaging
(DTI) to map the optic radiations, co-register it and follow it
intraoperatively, although this might not necessarily help us avoid
it.”
Temporal lobectomy
Dr Morris: “We have to go with the results of the studies we
ordered. We knew the limitations of grids, but we identified an
active epileptogenic focus, so we can’t ignore it. If we were going
to ignore it, why put the electrodes in? Given what we have, I
think she needs a large anterior temporal lobectomy, and then it
will be interesting to see if, pathologically, she had or did not
have hippocampal sclerosis. I would do a corticectomy around the
lesion, but avoid all the deep white matter because I still think
you will find dysplasia and or tumor back there. I think she will
probably have some deficits on neuropsychological measures upon
retesting (chances about 20%), but she would probably not notice
this.”
Dr Bingaman: “The problem is that we still don’t know whether
the anterior temporal patterns we are seeing on the surface are
originating from the depths of the lesion or from the mesial
temporal structures. Nowadays, we would have put depth electrodes
in the hippocampus and temporal pole to better assess those
structures. But, given this information, I would also recommend a
temporal lobectomy, going posteriorly to the previous lesionectomy
margin together with some cleaning of the deep white matter.”
Dr Diehl: “I agree with temporal lobectomy since all current
evidence points to epileptogenicity anterior to the lesionectomy.
Electrodes posterior to the lesion were quiet, suggesting that that
area does not need to be investigated further and that anterior
resection is probably enough to treat the epilepsy. This would not
compromise functions that Dr Bingaman is worried about if a new
resection was to extend more superiorly in the parietal lobe.”
Drs Alexopoulos, Bautista, and Lachwaani: “We agree on a
temporal lobectomy. We can’t give an exact chance of
seizure-freedom, but we think it should be above 50% - probably
around 60-70% - because we would have removed both imaging and
electrical abnormalities recorded, and the mesial structures.”
Dr Bingaman: “In general, the chances of a successful outcome
following a first failed surgery and dysplasia are less than 50%.
The nice thing about this case is that everything is anterior to
the lesion, and it is non-eloquent brain so we can make a bigger
hole.”
Actual treatment: anterior temporal lobectomy and a separate
temporoparietal resection around the previous lesionectomy.
Dr Ruggieri: “We see a more extensive resection along the margins
of the lesion itself superficially, and deeper in the white matter
to include resection of the previous FLAIR hyperintensity that
extended to the ependymal surface. Areas of hyperintensity on this
study are probably related to postoperative changes. In addition,
we see an anterior temporal lobectomy with volume studies showing
complete resection of the amygdale, head and body of the
hippocampus with some residual tissue from the body of the
hippocampus mesially and posteriorly.”
Dr Jeha: “Superimposing the preoperative and postoperative MRI
reconstruction, we see the more extensive lesionectomy and the
amount of brain tissue removed with the ATL, including SA17, the
electrode of ictal onset.”
Dr Wyllie: “What was the thought process behind leaving this
island of tissue behind in between lesionectomy and ATL?”
Dr Bingaman: “Probably to avoid a field cut, a risk that is
usually unacceptable to patients, especially if they’re driving or
want to drive.”
Dr Prayson: “Sections taken from the area around the previous
lesionectomy show a cellular tissue populated mostly with
CD68-(macrophage marker) positive cells, consistent with the
reactive gliosis expected in an operative bed. Sections from the
anterior temporal lobe show some chronic inflammation on the
surface, likely related to the invasive monitoring, but no evidence
of dysplasia and no evidence of tumor. The hippocampus looked
normal. The section of amygdale had one aggregate of small neurons
which we call hamartia, which probably represents a localized
microscopic area of dysplasia of unkown significance.”
Outcome
The patient has been seizure-free for five years now
postoperatively, without any new neuropsychological deficits.
Final outcome from the audience
Dr Dinner: “It is good that she had a seizure-free outcome, but it
would have been tantalizing to see if the same result could have
been achieved with a more extensive lesionectomy alone. I come back
to Dr Kotagal’s comments about inserting electrodes in the lesion
to investigate the pathophysiology of the seizures in relation to
the lesion.”
Dr Kotagal: “This case illustrates the idea that if you have a
lesion such as a DNET or a ganglioglioma, even in the non-dominant
hemisphere, invasive evaluation might be helpful in defining the
epileptogenic zone in relation to the lesion to produce a better
outcome.”
Dr Lüders: “I am not sure how helpful an invasive evaluation
using grids alone would be. If the epileptogenic zone is within the
depths of the lesion or in the tissue surrounding it, as in this
case, ictal patterns recorded with surface grids may be somewhere
else while actually representing a spread pattern. It could be
different with depth electrodes.”
Dr Bingaman: “I am still confused as to what cured this
patient’s epilepsy. Was it the more aggressive resection of the
lesion? In that case, as Dr Wyllie suggested in the beginning, a
“bigger hole” might have resolved the issue to start with. Was it
resection of the EEG abnormality? If so, could we have just
biopsied the lesion and watched it, as we frequently do with those
developmental lesions. One message is that these dysplasia cases
are all unique with their metabolic and anatomic variations, and
should be approached as such. But perhaps if we adopt a
standardized approach for evaluation (such as the use of depths and
grids) we will eventually learn enough about their pathophysiology
to answer those questions.”
Authors’ conclusions
This case illustrates the frequent association seen between
developmental tumors and cortical dysplasia (Prayson et al. 1993;
Takahashi et al. 2005).
It is our impression that this patient’s epileptogenic zone was
in the tissue surrounding the lesion, likely in the deeper parts of
the tumor, with the ictal temporal patterns recorded from the
subdural plates representing spread rather than true ictal onset.
This is supported by the normal histology of the resected temporal
lobe and the absence of hippocampal sclerosis. We suggest that
similar patients with developmental tumors should undergo extensive
lesionectomies to avoid reoperation.
The invasive coverage in this case did not include depth
electrodes within the lesion. This was therefore inadequate to
answer the question of whether seizures started independently in
the anterior temporal lobe or spread there from deeper parts of the
tumor. We suggest that invasive evaluations of temporal lobe
developmental tumors should include subdural electrodes and depth
electrodes inserted in the depths of the lesion and in the
hippocampus to evaluate appropriately the electrophysiological
correlation between these structures.
Although multiple studies of patients with dual pathology have
shown better outcomes with temporal lobectomy compared to
lesionectomy (Li et al. 1999; Chan et al. 2006), an adequately
designed invasive evaluation might allow sparing of the mesial
temporal structures when the only “pathology” seen on imaging is
the tumor itself.
References
Chan et al 2006 Chan CH, Bittar RG, Davis GA,
et al. Long-term seizure outcome following surgery for
dysembryoplastic neuroepithelial tumor. J Neurosurg 2006; 104:
62-9.
Li et al 1999 Li LM, Cendes F, Andermann F,
et al. Surgical outcome in patients with epilepsy and dual
pathology. Brain 1999; 122: 799-805.
Morris et al 1993 Morris HH, Estes ML, Gilmore R,
et al. Chronic intractable epilepsy as the only symptom of
primary brain tumor. Epilepsia 1993; 34: 1038-43.
Prayson et al 1993 Prayson RA, Estes ML,
Morris HH. Coexistence of neoplasia and cortical dysplasia in
patients presenting with seizures. Epilepsia 1993; 34: 609-15.
Takahashi et al 2005 Takahashi A, Hong SC,
Seo DW, et al. Frequent association of cortical dysplasia
in dysembryoplastic neuroepithelial tumor treated by epilepsy
surgery. Surg Neurol 2005; 64: 419-27.
|
Version imprimable
Version PDF
