Neuroimaging_of_Epilepsy.6 (1)
Neuroimaging_of_Epilepsy.6 (1)
Neuroimaging of Epilepsy
C O N T I N UU M A UD I O By Samuel Lapalme-Remis, MDCM, MA, FRCPC;
I NT E R V I E W A V AI L A B L E Dang K. Nguyen, MD, PhD, FRCPC
ONLINE
Downloaded from http://journals.lww.com/continuum by i+cC+52ngAdUtedgsTdfNOxhCCW4G4FOZD2fZ8QmmlT+aE4SHYOyhfBeNRul5dxwn6Zo6W0bDxT5atFk5uMUborvp0zVJYRPr6d8Hsq0mYcMLwYZhV4NTIFVRAalaljE4RP+otqkOehaIsIqDVVzVbs0jTKveTIVGaV7hr+9Oc4= on 04/12/2022
ABSTRACT
PURPOSE OF REVIEW: This article provides an overview of imaging modalities,
important imaging pathologies, and the role each imaging modality can
play in the diagnosis, evaluation, and treatment of epilepsy, including
epilepsy surgery.
INTRODUCTION
E
Address correspondence to
Dr Nguyen, 1000 Saint-Denis,
pilepsy is a paroxysmal disease of the brain that presents diagnostic and
Montreal, Quebec, H2V 2L9, treatment challenges for neurologists. From the beginning of the
Canada, [email protected]. disease course, brain imaging assists neurologists in reaching the
RELATIONSHIP DISCLOSURE: correct diagnosis, selecting initial treatments, and estimating the
Drs Lapalme-Remis and Nguyen prognosis for treatment response. For patients whose disease later
report no disclosures.
demonstrates resistance to pharmacologic treatments, neuroimaging becomes
UNLABELED USE OF even more important when considering epilepsy surgery because multiple
PRODUCTS/INVESTIGATIONAL imaging modalities are often necessary to maximize the likelihood of localizing
USE DISCLOSURE:
Drs Lapalme-Remis and Nguyen
the epileptogenic onset zone, thus optimizing seizure freedom and minimizing
report no disclosures. the risk of neurologic impairment.
Traditionally, epilepsy was diagnosed after a patient had at least two
© 2022 American Academy unprovoked seizures occurring greater than 24 hours apart. A definition
of Neurology. proposed by the International League Against Epilepsy (ILAE) in 2014 allows for
CONTINUUMJOURNAL.COM 307
CASE 3-1 A 29-year-old right-handed woman was seen in clinic for an episode of
loss of consciousness that had occurred 3 weeks previously during an
overnight transcontinental flight. She was on a business trip, and no
family or friends traveled with her. She had no memory of the loss of
consciousness, recalling only that she regained consciousness to find
herself lying on the floor of the aisle surrounded by flight personnel. She
was confused for a few minutes but returned to normal quickly. She had
no tongue laceration or incontinence.
The patient had not received any information about the manifestations
of the loss of consciousness from flight personnel or other witnesses and
did not seek medical care when she landed at her destination. After
returning home, she reported the episode to her family physician, who
referred her to an urgent neurology clinic. On questioning, she reported a
similar episode approximately 12 months previously, also while traveling
internationally with disturbance in her sleep schedule. Neurologic
examination was normal.
Although the episodes were
suspicious for bilateral
tonic-clonic seizures, the
patient was hesitant to
accept a diagnosis of epilepsy
or begin treatment, given the
lack of corroborating data.
An EEG obtained the same
day of the clinic visit was
normal. An EEG with sleep
deprivation and MRI were
requested within 1 week.
MRI was performed first and
revealed the presence of
periventricular nodular
heterotopia (FIGURE 3-1).
This MRI finding supported
a diagnosis of epilepsy,
and the patient FIGURE 3-1
began treatment with an MRI from the patient in CASE 3-1 identified the
antiseizure medication presence of a large periventricular nodular
without recurrence. The heterotopia (arrow) adjacent to the right lateral
ventricular atrium, seen here as a solid mass
sleep-deprived EEG was isointense to the cortex on a coronal
canceled. T2-weighted slice.
● The Harmonized
STRUCTURAL DIAGNOSTIC IMAGING IN EPILEPSY Neuroimaging of Epilepsy
All patients with seizures or epilepsy should receive head imaging early in their Structural Sequences
disease course as part of their basic diagnostic and prognostic workup. (HARNESS-MRI) protocol is
recommended for all
patients with seizures. It
MRI consists of three mandatory
MRI is the mainstay of epilepsy imaging; it is highly sensitive to many types of sequences and two optional
intracranial pathology and provides a level of structural detail unattainable by sequences, optimized for
3-Tesla (T) scanners but
other forms of imaging. The ILAE Neuroimaging Task Force reviewed the role of
compatible with 1.5T
MRI in epilepsy diagnosis and proposed methods to improve its yield.8 The Task scanners.
Force recommended that, resources allowing, MRI be performed for all patients
presenting with a first seizure or newly diagnosed epilepsy. Neurologists should
not rely on a report of a normal examination; if the MRI was done at another
center where the MRI protocol was suboptimal or the radiologist was not
experienced with neuroimaging of epilepsy, a repeat examination can provide
a higher yield (CASE 3-2).
To ensure that appropriate sequences are obtained and to standardize imaging
practices across different centers, the ILAE Neuroimaging Task Force
recommends using the Harmonized Neuroimaging of Epilepsy Structural
Sequences (HARNESS-MRI) protocol for all patients with seizures. This
protocol, which is optimized for 3-Tesla (T) scanners but remains compatible
with a 1.5T scanner, consists of three mandatory sequences and two optional
sequences. The sequences are summarized in TABLE 3-1. Taken together, these
sequences permit evaluation of the brain’s anatomy and morphology, pathologic
lesions, and the internal structures of the hippocampus. When clinically
indicated, the optional postgadolinium T1-weighted sequence can evaluate
tumors, vascular malformation, or infectious processes, and T2 susceptibility-
weighted imaging (SWI) is sensitive to blood products and calcifications.
CT
When MRI was first used in the early 1980s for patients with epilepsy, studies
immediately demonstrated the superiority of the new technique over CT
scanning, both in sensitivity and specificity, of identifying epileptogenic lesions.9
The importance of CT imaging in epilepsy has since been greatly reduced.
Nevertheless, being faster, cheaper, and more rapidly available, CT is often used
CONTINUUMJOURNAL.COM 309
FIGURE 3-2
Three-Tesla MRI from the patient in CASE 3-2 shows increased T2 signal of the left hippocampus
(arrows) on axial (A) and coronal (B) fluid-attenuated inversion recovery (FLAIR) sequence and
loss of hippocampal internal architecture on a coronal T2-weighted image (arrow) (C).
CONTINUUMJOURNAL.COM 311
then can the neurologist make an educated hypothesis as to the role of the lesion
in causing the patient’s symptoms, and that hypothesis must be subject to
reevaluation as more information becomes available over time.
For example, an arachnoid cyst is usually an incidental finding in the
evaluation of a patient with suspected seizures and can be dismissed as the
epileptogenic lesion if its location is not highly compatible with the patient’s
symptomatology and EEG findings. However, highly epileptogenic lesions such
as mesial temporal sclerosis or cavernous malformations require a far lower
threshold, and a putative diagnosis can be made in the clinic if the history and
seizure symptomatology are compatible with the lesion.
This section reviews different lesion types that are known to cause epilepsy,
focusing on representative pathologies of different etiologic categories, as well as
others that have received recent attention. A detailed description of all potential
epileptogenic lesions seen on imaging is beyond the scope of this review; more
comprehensive lists can be found in the tables. For the purpose of this review,
lesions have been classified into six categories: malformations of cortical
development, vascular lesions, tumors, gliosis and lesions caused by physical
deformation of the brain, autoimmune conditions, and infectious causes.
Sequence
Sequence name type Focus Clinical scenario
Magnetization-prepared rapid T1-weighted Optimal evaluation of brain anatomy Mandatory for all
gradient echo (MPRAGE)/three- and morphology patients
dimensional spoiled gradient echo/
three-dimensional turbo field echo
High in-plane resolution two- T2-weighted Examination of hippocampal internal Mandatory for all
dimensional coronal (turbo spin echo) structure (images are acquired patients
perpendicular to the long axis of the
hippocampus by using high resolution)
Susceptibility-weighted imaging (SWI) T2*-weighted Assessment of venous blood, Optional; when tumor,
hemorrhage, iron deposits, and vascular malformation,
calcifications or infectious process is
suspected
a
Data from Bernasconi A, Epilepsia.8
CONTINUUMJOURNAL.COM 313
Focal cortical dysplasias are classified into three subgroups, each of which has
multiple types.12 Briefly, focal cortical dysplasia type I is characterized on
pathologic inspection by alterations in the cytoarchitecture of neurons and can
affect large or small regions of the cortex. Focal cortical dysplasia type I is often
subtle and often invisible or missed on MRI imaging. When visualized on MRI, it
may appear as regions of hypoplasia or thin cortex, blurring of the gray-white
matter interface, or abnormally shaped or deep sulci.
Focal cortical dysplasia type II is characterized by disruption of the cortical
lamination, accompanied by morphologically abnormal cell types. Focal
cortical dysplasia type IIa contains only dysmorphic neurons, whereas focal
cortical dysplasia type IIb also contains balloon cells. Focal cortical dysplasias
type II are generally easier to detect on MRI than focal cortical dysplasias type I.
Like focal cortical dysplasia type I, focal cortical dysplasia type II may be
characterized by blurring of the gray-white junction or abnormal gyral or sulcal
patterns. However, it may also present additional features, including areas of
thickened cortex, increased T2/fluid-attenuated inversion recovery (FLAIR)
signal in the adjacent subcortical white matter, or the well-described
transmantle sign, seen in cortical dysplasia type IIb, which is a long region of
T2/FLAIR hyperintense signal tapering between the affected region of cortex
and the ventricular wall (FIGURE 3-4). Some focal cortical dysplasias type II are
located at the bottom of sulci, making them a challenge to identify, but when
found, bottom-of-sulcus dysplasia is often highly amenable to successful
surgical treatment.13
Focal cortical dysplasia type III refers to dysplasia that is found in the same
region of the brain as another lesion such as hippocampal sclerosis or a vascular
malformation. The focal cortical dysplasia itself may have imaging characteristics
of focal cortical dysplasias type I or type II or may not be visible at all (CASE 3-3).
An important feature of focal cortical dysplasias to be considered in
pediatric epilepsy is that they can be masked by the maturation of myelination
◆ Dysgyria
◆ Focal cortical dysplasia
◆ Hemimegalencephaly
◆ Lissencephaly
◆ Megalencephaly
◆ Microcephaly
◆ Periventricular nodular heterotopia
◆ Polymicrogyria
◆ Schizencephaly
◆ Subcortical band heterotopia
◆ Sturge-Weber syndrome (vascular phakomatosis)
◆ Tuberous sclerosis (neurocutaneous syndrome)
Vascular Lesions
Many different abnormalities of brain vasculature may cause epilepsy, and many
mechanisms of epileptogenesis can occur across different lesion types. For
CONTINUUMJOURNAL.COM 315
CASE 3-3 A 27-year-old man was assessed at an epilepsy center for drug-resistant
focal epilepsy. He experienced bilateral tonic-clonic seizures as a child but
had been successfully weaned off antiseizure medications at age 14. His
seizures recurred at age 22. These were characterized by an onset of
palinopsia and oscillopsia, followed by a rising sensation of heat, nausea,
hypersalivation, alteration of consciousness, and confusion with preserved
language, rarely followed by evolution to a bilateral tonic-clonic seizure.
Brain MRI showed a region of encephalomalacia and hemosiderin in the right
occipital lobe, suggestive of a chronic ruptured vascular lesion (FIGURE 3-5).
Recorded seizures first showed EEG onset in the right posterior temporal
region, then spreading to the anterior temporal region.
Seizure symptomatology and imaging suggested that the primary seizure
focus was in the right parieto-occipital region, with propagation to the
ipsilateral mesial temporal structures. A right occipital lesionectomy was
performed. Surgical pathology showed that the region of encephalomalacia
was caused by an obliterated primitive vascular malformation. In the
adjacent cortex, neuronal lamination was anomalous, consistent with focal
cortical dysplasia type III.
COMMENT Focal cortical dysplasia type III arises in the presence of an adjacent lesion
during development and may explain the epileptogenesis of a static
encephalopathy, in this case a long-ruptured vascular malformation. In this
patient, the focal cortical dysplasia could not be radiologically
distinguished from the patient’s known lesion.
Tumors
All varieties of brain tumors can cause seizures, and seizures are often their
presenting symptom, leading to diagnosis through neuroimaging. Malignant
tumors such as high-grade gliomas and metastases from systemic cancers may
cause seizures, but their treatment is primarily directed at preserving the life and
neurologic function of the patient rather than controlling seizures. Some intra-axial
tumors such as gangliogliomas are benign and almost never expand, but they are
epileptogenic and usually operated for the sole purpose of controlling epilepsy.
Extra-axial tumors such as meningiomas cause epilepsy less frequently, but they
may become epileptogenic as they expand and distort the adjacent cortex, often
requiring surgery by the time they cause symptoms. Low-grade gliomas, discussed
in the following section, hold a middle ground between highly malignant and
benign tumor types. Although they may expand and threaten the patient’s life and
neurologic function, for years or decades the primary difficulty they present may
FIGURE 3-5
MRI from the patient in CASE 3-3 shows a region of encephalomalacia and hemosiderin
deposits (A, B, arrows) in the right occipital region on axial T1 (A) and coronal fluid-
attenuated inversion recovery (FLAIR) image (B). Although pathology of the adjacent
cortex (C) revealed focal cortical dysplasia type IIIc with cortical dyslamination, neuronal
loss most dramatic in cortical layers 2 and 3 (area within the box), and reactive gliosis (area
within the circle), no focal cortical dysplasia was visible on MRI imaging.
GFAP = glial fibrillary acidic protein; NEUN = neuron-specific nuclear protein.
Pathology images courtesy of France Berthelet, MD, and Romain Cayrol, MD, Centre hospitalier
de l'Université de Montréal.
CONTINUUMJOURNAL.COM 317
FIGURE 3-6
Images showing the evolution of a focal cortical dysplasia through infancy to adolescence. A,
MRI showing the prominence of a right frontal focal cortical dysplasia (arrow) when imaged
by a dual echo T2-weighted sequence in a patient at the age of 8 months. B, When the patient
was 2 years old, the same focal cortical dysplasia was undetectable on fluid-attenuated
inversion recovery (FLAIR) because of ongoing changes in myelination. C, At the age of
14 years, the focal cortical dysplasia (arrow) became visible again, but it remained subtle
compared with the initial MRI.
Images courtesy of David Dufresne, MD, Université de Sherbrooke.
FIGURE 3-7
MRI showing multiple periventricular nodular heterotopias in one patient. A, Axial T1-weighted
image shows multiple periventricular nodular heterotopias (arrows) lining the occipital horns
of the lateral ventricles. B, Coronal T2-weighted image shows an additional nodule (arrow)
adjacent to the right hippocampus. The nodules represent masses of neurons that have failed
to migrate normally and are therefore isointense to cortex on all sequences.
Benign or Slowly Progressive Intracranial Tumors Strongly Associated With TABLE 3-3
Epilepsy That Can Be Seen on Imaging
CONTINUUMJOURNAL.COM 319
KEY POINTS
● Cerebral cavernous
malformations may have a
classic “popcorn”
appearance on MRI. T2*
sequences such as gradient
recalled echo (GRE) or
susceptibility-weighted
imaging (SWI) show areas of
hypointensity and the
epileptogenic hemosiderin
rim that surrounds the
structure.
● Multinodular and
tumors (DNETs) are among the most common tumors causing focal epilepsy.
vacuolating neuronal tumors These benign tumors, thought to be congenital, are usually static over decades or
(MVNTs) are recently grow very slowly. They are, however, highly epileptogenic and are a frequent
described epileptogenic target of epilepsy surgery. On imaging, most tumors are located in the temporal
lesions with an MRI
appearance of multiple
or frontal lobes, usually within the cortex. On CT, they appear as a wedge-shaped
discrete ovoid intra-axial hypodense mass and may contain calcifications or cause remodeling of adjacent
nodules found at the bony structures. MRI shows the classic “bubbly” appearance, with multiple
junction of superficial lobulated regions of hyperintensity on T2-weighted sequences. The calcifications
subcortical white matter
seen on CT also appear as a blooming artifact on T2* sequences. DNETs do not
and a deep cortical ribbon,
often surrounding a sulcus. avidly enhance or cause edema (FIGURE 3-10).
FIGURE 3-11
MRI of a patient with left posterior temporal multinodular and vacuolating neuronal
tumors (MVNTs) causing focal seizures with impaired awareness. Coronal (A) and axial (B)
T2-weighted images show multiple discrete small nodules (arrows) at the junction of
superficial subcortical white matter and the cortex surrounding the bottom of a sulcus.
Axial fluid-attenuated inversion recovery (FLAIR) image (C) shows the nodules (arrow)
more clearly as T2 hyperintensities. The lesion is barely visible on axial postcontrast
T1-weighted image (D) as scattered hypointensities (arrow) that were nonenhancing.
CONTINUUMJOURNAL.COM 321
FIGURE 3-12
Coronal MRI shows right hippocampal hyperintensity and atrophy on fluid-attenuated
inversion recovery (FLAIR) image (A, arrow) and loss of internal structure on T2-weighted
image (B, arrow).
COMMENT The diagnosis of epilepsy with mesial temporal sclerosis can be made in
the clinic based on history, EEG, and MRI. Patients with mesial temporal
sclerosis often benefit from surgical treatment of their epilepsy and should
be referred to a comprehensive epilepsy center.
CONTINUUMJOURNAL.COM 323
CONTINUUMJOURNAL.COM 325
NEUROCYSTICERCOSIS.
Neurocysticercosis is a leading
cause of focal epilepsy in many
FIGURE 3-15
low- and middle-income regions
Axial postcontrast T1-weighted MRI of autoimmune and can cause epilepsy
glial fibrillary acidic protein (GFAP) astrocytopathy. throughout the several-year life
Patterns of enhancement include radial cycle of its causative parasite,
periventricular (A), leptomeningeal and punctate
Taenia solium, or long after its
(B), serpiginous (C), and periependymal (D).
Reprinted with permission from Kunchok A, et al, Curr Opin death. The parasite is ingested
Neurol.24 © 2019 The Authors. through fecal-oral contamination
and can release larvae into the
human bloodstream, and from
there, some can reach the brain. The larvae then proceed through four stages of
development, each with unique imaging characteristics (TABLE 3-5).25 When the
parasite dies, it progresses to the nodular calcified stage, remaining within the
brain for the rest of the patient’s life as a fibro-calcified nodule.
On CT imaging, the chronic lesions appear as single or multiple small
nodular hyperdense masses, without perilesional edema or enhancement.
These lesions appear hypointense on T1-weighted and T2-weighted MRI
sequences. T2* sequences reveal a strongly hypointense blooming artifact
(FIGURE 3-16).
◆ Cerebral malaria
◆ Cryptococcosis
◆ Herpes simplex virus encephalitis
◆ Neurocysticercosis
◆ Toxoplasmosis
◆ Paragonimiasis
◆ Subacute sclerosing panencephalitis
◆ Syphilis
◆ Tuberculosis
Stage 1: vesicular (living parasite) A scolex may be seen a within cyst of Cyst appears as a hypodense
approximately 1-2 cm; when visible, the lesion isodense to CSF
scolex may enhance and is best seen on
fluid-attenuated inversion recovery (FLAIR)
images; minimal or absent enhancement of
the cyst wall; cyst fluid isointense to CSF
Stage 2: colloidal vesicular (early Scolex is no longer visualized; a fluid-fluid Cyst appears as a hypodense
degeneration and host inflammatory level may be seen within the cyst, lesion with ring enhancement
response) suggesting internal debris; cyst fluid with contrast
hyperintense on T1-weighted sequences;
gadolinium ring enhancement surrounds the
cyst; T2-weighted sequences may show
perilesional edema; absence of diffusion-
weighted imaging (DWI) restricted diffusion
Stage 3: granular nodular (further Morphology evolved to a smaller nodular Contrast images may show
degeneration) lesion; gadolinium enhancement of lesion enhancement within the lesion
with possible small ring enhancement; or small ring enhancement
T2-weighted sequences may show mild
perilesional edema
Stage 4: nodular calcified (end stage as a Hypodense nodule on T1- and T2-weighted Hyperdense nodule;
nonviable calcified granulomatous lesion sequences; usually no edema or nonenhancing
without host inflammatory response) enhancement
CONTINUUMJOURNAL.COM 327
FIGURE 3-16
Imaging from a patient with anterior temporal neurocysticercosis causing left temporal onset
seizures. Two other lesions were seen in other regions of the brain (not shown). On axial
T1-weighted (A) and fluid-attenuated inversion recovery (FLAIR) (B) images, the lesion
(arrows) was seen as a small nodular hypointensity, with prominent hypointense “blooming”
artifact (arrow) on axial susceptibility-weighted imaging (SWI) (C). MRI could not distinguish
whether the lesion represented calcification or hemosiderin deposit, so CT was performed
(D), confirming highly dense calcifications (arrow) and the diagnosis of chronic
neurocysticercosis lesions.
seizure focus must be established with maximal confidence. On the other hand,
the function of any region of cortex under consideration for removal must be
established to avoid a permanent disabling functional deficit from the surgery.
The presurgical evaluation is therefore highly individualized, and imaging data
are essential to focus both identification and surgical planning for preservation of
neurologic function.
Focus Identification
Identifying the seizure focus for resection is a simple endeavor in some
patients but highly complex in others. A general principle of focus
identification is to collect different types of data, which may include the
details of seizure symptomatology, neuropsychological profile, interictal EEG,
video-EEG recordings of seizures, structural imaging, and, when necessary,
different methods of functional imaging or intracranial EEG data, each of
which may suggest a particular seizure focus. When these different methods
of evaluation are concordant, the seizure focus can be predicted with higher
confidence, and the likelihood of surgical success is increased. When they are
inconclusive or discordant, surgical planning becomes more challenging.
In the evaluation of a patient with focal epilepsy, the presence or absence of an
epileptogenic lesion that can be visualized on structural imaging is usually the
most important factor predicting the eventual success of surgery. For many
lesion types such as mesial temporal sclerosis, DNETs, or cavernous
malformations, rates of seizure freedom after surgery can exceed 70%.27 Simply
put, a lesion that can be seen can more easily be identified and removed. This
concept is so basic that focal epilepsy was traditionally divided as “lesional” or
“nonlesional,” referring to whether patients have a visible lesion on MRI. Without
a visible lesion, intracranial EEG monitoring is often necessary before focus
resection. The lower rates of postoperative seizure freedom attest to the difficulty
of correctly identifying the seizure focus in patients with so-called
“nonlesional” epilepsies.28
CONTINUUMJOURNAL.COM 329
FIGURE 3-17
Imaging from a patient with nocturnal seizures. Axial 3-Tesla MRI shows an area of focal
atrophy in the left orbitofrontal region (purple cursor in all images) in axial (top row),
sagittal (middle row), and coronal (bottom row) images (note that, contrary to the usual
orientation, the patient’s left side is shown on the left in the axial and coronal images).
The sulcogyral pattern is abnormal and includes many small gyri. On T1-weighted (first
column) and fluid-attenuated inversion recovery (FLAIR) (second column) images, a blurring
of the gray-white matter junction can be seen. The FLAIR signal is also increased. These
anomalies were confirmed by image processing (third and fourth columns) with texture
analysis showing hyperintensity and decrease of gradient, with a large darkened area.
This anomaly is suggestive of focal cortical dysplasia type IIa.
Image courtesy of Andrea Bernasconi, MD, Montreal Neurological Hospital.
POSITRON EMISSION TOMOGRAPHY. The epileptic focus and the region surrounding it
may have reduced metabolism during the interictal period. This region with
interictal hypometabolism consumes differentially less fludeoxyglucose (FDG)
than regions of healthy tissue, a phenomenon that can be detected and expressed
with FDG–positron emission tomography (PET) imaging. A glucose analogue
radiotracer (18F-FDG) is injected into the patient. Brain cells, which rely primarily
on glucose for energy, avidly take in the tracer, where it remains trapped
unmetabolized. The patient is scanned no less than 30 minutes later, and the PET
images display the regions of the brain that contain relatively more or less of the
radiotracer. Areas that light up less than expected, either through visual
inspection in comparison with the contralateral side or by using quantitative
algorithms, can be considered hypometabolic. PET images can also be exported to
existing MRI images of the same patient, adding anatomic detail to the regions of
altered metabolism.
In temporal lobe epilepsy, unilateral temporal hypometabolism is correlated
with the side of seizure focus, allowing lateralization in such cases. Even in
MRI-negative epilepsy, patients with temporal hypometabolism on FDG-PET
have more favorable postsurgical outcomes. It should be noted, however, that the
region of hypometabolism often extends well beyond the true epileptic focus,
such that PET imaging cannot be used to guide delineation during surgical
planning (CASE 3-5). Recent research suggests that for patients with mesial
temporal epilepsy, a more widespread region of hypometabolism portends a
lesser likelihood of postsurgical seizure freedom, especially when extending into
extratemporal or contralateral regions, with the best outcomes found when the
hypometabolism is restricted to the anteromesial region.31 Although FDG-PET
CONTINUUMJOURNAL.COM 331
can also be used to detect extratemporal epileptic foci, the reported sensitivity in
such patients is much lower.
Compared with many other functional imaging techniques, the logistic
barriers to obtaining PET imaging are significantly fewer because this technique
can be obtained routinely on an outpatient basis. As a result, it can sometimes be
used as an adjunctive technique in the diagnosis and classification of epilepsy,
even in patients who are not presently candidates for epilepsy surgery.
COMMENT The FDG-PET shows a region of hypometabolism that extends well beyond
the epileptogenic lesion. In such cases, the region of hypometabolism
should not be used to delineate the surgical margins, as this phenomenon
is well described. The larger region of hypometabolism does suggest a
lesser probability of postsurgical freedom from seizures, even when the
presumptive seizure focus is fully resected, but should not be a
contraindication to surgery.
FIGURE 3-19
Ictal-interictal subtracted single-photon emission computed tomography (SPECT)
coregistered to axial MRI showing activation centered on the right mesial temporal
structures during seizure onset, suggesting a right temporal seizure focus.
CONTINUUMJOURNAL.COM 333
FIGURE 3-20
EEG–functional MRI (fMRI) showing activation clusters superimposed on axial (A), sagittal (B),
and coronal (C) T1-weighted fMRI correlated with interictal spikes and slow waves diffusely
maximal over the right temporal and parietal head regions on EEG. These discharges were
correlated with maximal activation of the right supramarginal and angular gyri.
FIGURE 3-21
Magnetic source imaging showing the location of magnetoencephalography dipoles
coregistered to MRI, superimposed on coronal (A), sagittal (B), and axial (C) T1-weighted
images. The images show that the biggest clusters of magnetic dipoles are located in the
left opercular and anterior insular regions. Note that, contrary to the usual orientation, the
patient’s left side is shown on the left in the coronal and axial images.
Surgical Planning
Once the seizure focus has been identified, it must be removed or destroyed
surgically to achieve seizure freedom. Patients are often surprised to learn that a
piece of the brain can be taken out without causing serious neurologic injury. But
given the specialization of different regions of the brain, it is essential to ensure
that the region to be removed is not responsible for some critical function such as
speech or motor activity. Knowledge of brain anatomy is essential but not
sufficient to avoid neurologic impairment, given the variability in functional
topography among patients, especially in regions affected by a lesion that may
have been present since birth and influenced brain development from a young
age. fMRI and diffusion tensor imaging (DTI) are imaging methods that can
predict the location of critical brain tissue.
CONTINUUMJOURNAL.COM 335
the brain are more active when a patient performs a specific standardized task
using the function under review compared with a rest or control condition.
Although fMRI can be used to evaluate the localization of sensory or motor
function, it is most frequently used in epilepsy surgery planning to lateralize
and localize language areas. Different language functions (eg, naming,
comprehension) can be tested individually by using protocols and mapped onto
the patient’s brain anatomy. This information can be used to estimate the risk
of language impairment after surgical resection, especially when considering a
left temporal lobectomy, or to help define surgical margins. Being less invasive,
this technique has partially replaced the Wada test for language lateralization
(FIGURE 3-22). When a surgery is considered high risk for loss of function,
however, intracranial cortical stimulation mapping may be necessary
before resection.
FIGURE 3-22
Functional MRI (fMRI) used to evaluate lateralization of language function in a right-handed
patient with left mesial temporal sclerosis and drug-resistant epilepsy before lobectomy.
A, Statistical representation of changes in blood oxygen level–dependent (BOLD) activation
on a normalized brain during a language completion task to evaluate expressive speech;
darker areas show greater activation. The patient read an incomplete written sentence and
spoke out the missing words. This task showed bilateral activation in the frontal regions,
near the Broca area, left greater than right. B, Areas of higher BOLD activation (yellow) are
superimposed onto left sagittal, coronal, and axial T1-weighted MRI images of the patient’s
brain. Bilateral occipital activation was also seen due to the visual nature of the task. Note
that, contrary to the usual orientation, the patient’s left side is shown on the left in the axial
and coronal images. C, Areas of higher BOLD activation (see the yellow-red scale) are
superimposed onto a three-dimensional reconstruction of the patient’s brain MRI, shown
from different perspectives. Other language tasks (not shown) also mostly showed bilateral
activation, suggesting bilateral language function. A left temporal amygdalohippocampectomy
and anterior temporal lobectomy was performed without impact on the patient’s language
function.
CONCLUSION
It is the neurologist’s task to understand each patient’s epilepsy well enough to better
explain and treat their disease. In this endeavor, the importance of neuroimaging
cannot be underestimated, especially for more challenging cases. For neurologists
who treat patients with epilepsy, understanding the advantages and limitations of
each modality is essential and will lead directly to better care and a higher likelihood of
seizure freedom in their patients. As imaging technology improves, the judicious
selection and interpretation of imaging modalities will become all the more critical.
REFERENCES
1 Fisher RS, Acevedo C, Arzimanoglou A, et al. ILAE 4 Rosenow F, Alonso-Vanegas MA, Baumgartner C,
official report: a practical clinical definition of et al. Cavernoma-related epilepsy: review and
epilepsy. Epilepsia 2014;55(4):475-482. recommendations for management—report of
doi:10.1111/epi.12550 the Surgical Task Force of the ILAE Commission
on Therapeutic Strategies. Epilepsia 2013;54(12):
2 Krumholz A, Wiebe S, Gronseth GS, et al.
2025-2035. doi:10.1111/epi.12402
Evidence-based guideline: management of an
unprovoked first seizure in adults: report of the 5 Feyissa AM, Hasan TF, Meschia JF. Stroke-related
Guideline Development Subcommittee of the epilepsy. Eur J Neurol 2019;26(1):18-e3.
American Academy of Neurology and the doi:10.1111/ene.13813
American Epilepsy Society. Neurology 2015;
6 Consoli D, Bosco D, Postorino P, et al.
84(16):1705-1713. doi:10.1212/WNL.
Levetiracetam versus carbamazepine in patients
0000000000001487
with late poststroke seizures: a multicenter
3 Scheffer IE, Berkovic S, Capovilla G, et al. ILAE prospective randomized open-label study (EpIC
classification of the epilepsies: position paper of Project). Cerebrovasc Dis 2012;34(4):282-289.
the ILAE Commission for Classification and doi:10.1159/000342669
Terminology. Epilepsia 2017;58(4):512-521.
7 Lapalme-Remis S, Cascino GD. Imaging for adults
doi:10.1111/epi.13709
with seizures and epilepsy. Continuum (Minneap
Minn) 2016;22(5, Neuroimaging):1451-1479.
doi:10.1212/CON.0000000000000370
CONTINUUMJOURNAL.COM 337
8 Bernasconi A, Cendes F, Theodore WH, et al. 20 Campbell ZM, Hyer JM, Lauzon S, et al. Detection and
Recommendations for the use of structural characteristics of temporal encephaloceles in patients
magnetic resonance imaging in the care of with refractory epilepsy. AJNR Am J Neuroradiol 2018;
patients with epilepsy: a consensus report from 39(8):1468-1472. doi:10.3174/ajnr.A5704
the International League Against Epilepsy
21 Gillinder L, Britton J. Autoimmune-associated
Neuroimaging Task Force. Epilepsia 2019;60(6):
seizures. Continuum (Minneap Minn) 2022;
1054-1068. doi:10.1111/epi.15612
28(2, Epilepsy):363-398.
9 McLachlan RS, Nicholson RL, Black S, et al.
22 Irislimane M, Guilbert F, Leroux J, Carmant L,
Nuclear magnetic resonance imaging, a new
Nguyen D. Serial MR imaging of adult-onset
approach to the investigation of refractory
Rasmussen's encephalitis. Can J Neurol Sci 2011;
temporal lobe epilepsy. Epilepsia 1985;26(6):
38(1):141-142. doi:10.1017/S0317167100120815
555-562. doi:10.1111/j.1528-1157.1985.tb05691.x
23 Bien CG, Urbach H, Deckert M, et al. Diagnosis
10 Severino M, Geraldo AF, Utz N, et al. Definitions
and staging of Rasmussen's encephalitis by serial
and classification of malformations of cortical
MRI and histopathology. Neurology 2002;58(2):
development: practical guidelines. Brain 2020;
250-257. doi:10.1212/wnl.58.2.250
143(10):2874-2894. doi:10.1093/brain/awaa174
24 Kunchok A, Zekeridou A, McKeon A. Autoimmune
11 Barkovich AJ, Guerrini R, Kuzniecky RI, et al. A
glial fibrillary acidic protein astrocytopathy. Curr
developmental and genetic classification for
Opin Neurol 2019;32(3):452-458. doi:10.1097/
malformations of cortical development: update
WCO.0000000000000676
2012. Brain 2012;135(pt 5):1348-1369.
doi:10.1093/brain/aws019 25 Lerner A, Shiroishi MS, Zee C-S, et al. Imaging of
neurocysticercosis. Neuroimaging Clin N Am
12 Blümcke I, Thom M, Aronica E, et al. The
2012;22(4):659-676. doi:10.1016/j.nic.2012.05.004
clinicopathologic spectrum of focal cortical
dysplasias: a consensus classification proposed 26 Culler GW, Jobst BC. Surgical Treatments for
by an ad hoc Task Force of the ILAE Diagnostic epilepsy. Continuum (Minneap Minn) 2022;
Methods Commission. Epilepsia 2011;52(1): 28(2, Epilepsy):536-558.
158-174. doi:10.1111/j.1528-1167.2010.02777.x
27 Englot DJ, Chang EF. Rates and predictors of
13 Harvey AS, Mandelstam SA, Maixner WJ, et al. seizure freedom in resective epilepsy surgery: an
The surgically remediable syndrome of epilepsy update. Neurosurg Rev 2014;37(3):389-404;
associated with bottom-of-sulcus dysplasia. discussion 404-405. doi:10.1007/s10143-014-0527-9
Neurology 2015;84(20):2021-2028.
28 Noe K, Sulc V, Wong-Kisiel L, et al. Long-term
doi:10.1212/WNL.0000000000001591
outcomes after nonlesional extratemporal lobe
14 Eltze CM, Chong WK, Bhate S, et al. Taylor-type epilepsy surgery. JAMA Neurol 2013;70(8):
focal cortical dysplasia in infants: some MRI 1003-1008. doi:10.1001/jamaneurol.2013.209
lesions almost disappear with maturation of
29 Opheim G, van der Kolk A, Markenroth Bloch K,
myelination. Epilepsia 2005;46(12):1988-1992.
et al. 7T Epilepsy Task Force consensus
doi:10.1111/j.1528-1167.2005.00339.x
recommendations on the use of 7T MRI in clinical
15 Khoo HM, Gotman J, Hall JA, Dubeau F. Treatment of practice. Neurology 2021;96(7):327-341.
epilepsy associated with periventricular nodular doi:10.1212/WNL.0000000000011413
heterotopia. Curr Neurol Neurosci Rep 2020;20(12):
30 Lapalme-Remis, Witte RJ, Wong-Kisiel LC.
59. doi:10.1007/s11910-020-01082-y
Anteroinferior temporal encephalocele: a
16 Huse JT, Edgar M, Halliday J, et al. Multinodular surgically treatable cause of pharmacoresistant
and vacuolating neuronal tumors of the epilepsy. Pediatr Neurol 2017;77:89-90.
cerebrum: 10 cases of a distinctive seizure- doi:10.1016/j.pediatrneurol.2017.07.003
associated lesion. Brain Pathol 2013;23(5):
31 Chassoux F, Artiges E, Semah F, et al. 18F-FDG-PET
515-524. doi:10.1111/bpa.12035
patterns of surgical success and failure in mesial
17 Thom M, Liu J, Bongaarts A, et al. Multinodular temporal lobe epilepsy. Neurology 2017;88(11):
and vacuolating neuronal tumors in epilepsy: 1045-1053. doi:10.1212/WNL.0000000000003714
dysplasia or neoplasia? Brain Pathol 2018;28(2):
32 O'Brien TJ, So EL, Mullan BP, et al. Subtraction
155-171. doi:10.1111/bpa.12555
peri-ictal SPECT is predictive of extratemporal
18 Nunes RH, Hsu CC, da Rocha AJ, et al. epilepsy surgery outcome. Neurology 2000;55(11):
Multinodular and vacuolating neuronal tumor of 1668-1677. doi:10.1212/wnl.55.11.1668
the cerebrum: a new “leave me alone” lesion with
33 Kowalczyk MA, Omidvarnia A, Abbott DF, et al.
a characteristic imaging pattern. AJNR Am J
Clinical benefit of presurgical EEG-fMRI in
Neuroradiol 2017;38(10):1899-1904.
difficult-to-localize focal epilepsy: a single-
doi:10.3174/ajnr.A5281
institution retrospective review. Epilepsia 2020;
19 Saavalainen, Jutila L, Mervaala E, et al. Temporal 61(1):49-60. doi:10.1111/epi.16399
anteroinferior encephalocele: an
34 Mohamed IS, Toffa DH, Robert M, et al. Utility of
underrecognized etiology of temporal lobe
magnetic source imaging in nonlesional focal
epilepsy? Neurology 2015;85(17):1467-1474.
epilepsy: a prospective study. Neurosurg Focus
doi:10.1212/WNL.0000000000002062
2020;48(4):E16. doi:10.3171/2020.1.FOCUS19877