Anatomic Classification of Focal Epilepsies



Anatomic Classification of Focal Epilepsies


Peter D. Williamson

Jerome Engel Jr.



Introduction

The International Classification of Epilepsies, Epileptic Syndromes, and Related Seizure Disorders2 recognizes five symptomatic localization-related epilepsies, four of which are anatomically defined: Temporal lobe epilepsies, frontal lobe epilepsies, parietal lobe epilepsies, and occipital lobe epilepsies. Only the fifth, chronic progressive epilepsia partialis continua of childhood, is not classified by location. If we accept the definition of an epileptic syndrome as a complex of signs and symptoms, with only one feature being the occurrence of electroclinically characteristic epileptic seizures.1 then we can make a cogent argument for the existence of a syndrome of mesial temporal lobe epilepsy (MTLE)47,55,208,218,222 (see Chapter 247). Otherwise, the classification of epilepsies by anatomic lobe appears to refer predominantly to conditions with characteristic seizure types as the only unifying feature. Recent reports of the International League Against Epilepsy (ILAE) argue against the view that these conditions are syndromes rather than seizures.51 However, as there is as yet no new classification of syndromes, anatomically defined epilepsies with variably specific seizure types will be retained in this chapter. Other than MTLE, there may be a few examples with constellations of findings that do constitute true syndromes. These possible exceptions will be discussed later.

Much effort has gone into creating an anatomic classification of epileptic seizures, based largely on data obtained from epilepsy surgery centers. The anatomic classification has been greatly assisted by the introduction of high-resolution magnetic resonance imaging (MRI) more than 20 years ago, enabling the detection of small, circumscribed, potentially epileptogenic lesions in many patients with symptomatic localization-related epilepsy. Nevertheless, the validity, or for that matter, clinical value of devising an anatomic classification of epileptic disorders remains the topic of considerable debate. This chapter will review the subject, but emphasis, of necessity, will be placed on seizures rather than syndromes.


Historical Perspectives

John Hughlings Jackson is credited with being the first neurologist to recognize the correlation between ictal behavior and the location of structural abnormalities in the brain.190 His work derived largely from observations of ictal signs and symptoms, followed by identification of lesions at postmortem examination, in an era of pioneering investigations of localization of function within the brain.48 This work led directly to the application of surgical treatment for refractory partial epilepsy,82 which, in turn, provided opportunities for invasive investigations of the human brain, such as direct stimulation to delineate anatomic substrates of specific ictal manifestations.151 With the advent of the electroencephalogram (EEG), the location of focal interictal spikes was used as evidence for the site of origin of habitual seizures, and surgical treatment for “temporal lobe epilepsy” was introduced on the basis of EEG evidence alone.11 Shortly thereafter, the pioneering efforts of Bancaud et al. related intracranial seizure recording to clinical seizure characteristics.12

The development of sophisticated electrophysiologic techniques for recording spontaneous seizures with intracranial electrodes and correlating this information precisely with videotaped ictal behavior has made anatomic classification of seizures more, rather than less, controversial. Because of variability of propagation patterns of ictal discharge, the area of brain giving rise to clinical signs and symptoms can be at considerable distance from the actual site of seizure onset. Consequently, virtually no signs or symptoms can be considered pathognomonic of the anatomic localization of the primary epileptogenic region.4,121,124 Although the timing or sequencing of seizure manifestations may be as important as their presence or absence, this feature fails to localize precisely because of the potential for seizure origin in clinically “silent” brain regions.

Improvements in structural imaging also have not definitively clarified anatomic classification. High-resolution MRI can identify many structural cerebral lesions in patients with localization-related epilepsy,107 and they are usually indicative of the epileptogenic region.25 Many patients with partial seizures, however, have normal MRI evaluations; others have lesions that are unrelated to their epilepsy. Furthermore, it has become increasingly apparent from intracranial recordings of patients under evaluation for surgical treatment that a discrete, well-circumscribed epileptogenic region often does not exist in the human brain; instead there are diffuse or multiple areas of functional abnormality capable of generating interictal, and even ictal, discharges, even in patients for whom a specific localized structural lesion has been identified.46,214,220 For these reasons, anatomically classifying a seizure disorder merely on the basis of the location of an observed structural abnormality is not always justifiable. Nevertheless, a continued effort to categorize epileptic seizures, and even epileptic disorders, according to a presumed anatomic substrate can be considered a reasonable exercise of some value, particularly for devising presurgical diagnostic strategies, as long as the limitations of such a classification are recognized and surgical interventions are not undertaken on the basis of ictal manifestations or MRI lesions alone. Combining clinical seizure characteristics, structural neuroimaging findings, and EEG results, however, has improved our understanding of partial seizures and helped to better define an anatomic classification.

Functional imaging using positron emission tomography (PET) has been employed at a few medical centers for many years and is currently becoming available at many more institutions. While early results in patients with temporal lobe seizures were and still are promising,28,117 PET utility in patients with seizure origin outside of the temporal lobe
has been less consistent.128 Newer radioligands and improved technologies will almost certainly improve results in extratemporal epilepsy with normal MRI77,117 (see Chapter 80).

Ictal single positron emission computed tomography (SPECT) is being used widely to help identify regions of seizures origin.75,92,101,117,130,143,182,201 Early injection is essential, particularly with brief seizures that propagate rapidly.113,201 The subtraction product of interictal SPECT from ictal SPECT coregistered with MRI improves the resolution of this relatively difficult-to-obtain procedure.141,142,143 Ictal SPECT has become an integral part of the presurgical evaluation in some centers. This is particularly relevant when the MRI is normal.

Magnetoencephalography (MEG)45 superimposed on MRI and EEG-triggered functional MRI104 have great promise for localizing sources of ictal, as well as interictal, epileptiform activity. In the former, dipole localization is easier than with EEG because differences in conductivity between brain, cerebrospinal fluid (CSF), bone, and skin do not need to be taken into account. In the latter, simultaneous EEG and MRI recording permit MRI acquisition during epileptiform discharges to be subtracted from MRI acquisitions in the absence of epileptiform discharges, so that the difference in blood oxygenation can be used to localize tissue activated during the EEG events of interest. Magnetic resonance spectrometry (MRS) is also being used to identify metabolic changes associated with interictal disturbances, and this has been particularly useful for confirming hippocampal epileptogenic lesions.109,186


Definitions

With the exception of MTLE, the designation of an epileptic condition by anatomic location denotes the occurrence of a specific type of seizure or seizures, rather than a complex of signs and symptoms that would ordinarily constitute a syndrome. In the purest sense, a temporal lobe epilepsy, for example, would be an epileptic condition in which seizures originate somewhere in the temporal lobe, as indicated directly by invasive electrophysiologic monitoring, or inferred indirectly from clinical seizure characteristics, scalp EEG monitoring, perhaps the presence of a discrete lesion on structural neuroimaging, functional disturbances revealed by neurologic examination, neuropsychological testing and functional neuroimaging, and disappearance of habitual events following a temporal lobe resection of some type. There remains some controversy, however, with this terminology. In much of the published literature on temporal lobe epilepsy, some patients are given this diagnosis when extratemporal ictal onsets preferentially project to mesial temporal structures that, in turn, are responsible for mediating the characteristic ictal manifestations. In many cases, such patients are diagnosed as having temporal lobe epilepsy because the extratemporal origin of their seizures is unrecognized, but in others the actual site of the primary epileptogenic region is ignored and classification is based on ictal characteristics.208 If a temporal lobe seizure is one that is caused by seizures involving temporal lobe structures, then this could result from ictal initiation within the temporal lobe or an extratemporal “silent area” that preferentially projects to the temporal lobe. Thus, if temporal lobe epilepsy is characterized by the existence of temporal lobe seizures, then the broader definition is acceptable. Rather than dwell on the semantic difficulties associated with anatomic classification, this example can be used to argue the folly of attempting to construct an anatomic classification of epileptic disorders at all. On the other hand, if a specific temporal lobe epilepsy syndrome can be defined, then seizures originating elsewhere can, in part, be identified by virtue of not fulfilling all the characteristics of this syndrome.

Attempts to define seizures or epilepsies according to anatomic lobes of the brain are also confounded by the fact that none of the classical four lobes represents functionally homogenous or unique regions. For instance, parts of the frontal lobe, such as the precentral gyrus, are capable of generating specific ictal signs and symptoms, whereas others, such as the orbital frontal cortex, belong, in part, to the limbic system and can give rise to ictal signs and symptoms by virtue of propagation to other frontal regions or mesial temporal structures, or both (see section on frontal lobe seizures). The International Classification of the Epilepsies divides frontal lobe epilepsy into seven anatomic subtypes and temporal lobe epilepsy into two anatomic subtypes, based on the belief that these areas give rise to different characteristic ictal manifestations.2 However, there are also functional areas of the brain that give rise to relatively stereotyped clinical seizure characteristics, despite the fact that they encompass more than one anatomic lobe. Thus, seizures arising from the perirolandic area can generate indistinguishable signs and symptoms from motor or sensory cortex, and there may be characteristic features associated with ictal discharges in the temporal-parietal-occipital junction, as well as regions of frontal parietal operculum and insula.

Nevertheless, as our understanding of the various specific and anatomically originating seizures improves, there may emerge clearer evidence on how these clinical characteristics can help localize regions of seizure origin, both by virtue of their uniqueness and their masquerading qualities. This may be particularly true of some, but not all, seizures of frontal lobe origin.


Epidemiology

Among the localization-related epilepsies, the relative incidence and prevalence of temporal lobe, frontal lobe, parietal lobe, and occipital lobe epilepsy have not been adequately determined, due largely to the recognized inaccuracies of identifying anatomic substrates without sophisticated diagnostic evaluation. Data from epilepsy surgery centers indicate that temporal lobe epilepsy is by far the most common of the four when only medically refractory patients are considered. In a review of surgical procedures performed worldwide for medically refractory epilepsy, approximately 70% have involved resections of temporal lobe abnormalities, whereas <20% involved resection of extratemporal cortex.49,162,163,212 Of the latter group, frontal lobe resections are more common than occipital resections, and parietal lobe resections are least commonly performed. It is important to note, however, that these figures reflect the ease of identification of temporal lobe, frontal lobe, occipital lobe, and parietal lobe epileptogenic regions; the degree of intractability and disability of seizures associated with these areas; and the desirability of their surgical resection, as much as they reflect the incidence of epileptogenic abnormalities in these cerebral areas.

Mesial temporal lobe epilepsy may well be the most common epileptic syndrome, partly because of the peculiar epileptogenicity of hippocampal sclerosis (see Chapter 247). Studies of regional differences in epileptogenicity have shown that temporal lobe EEG spike foci are the most likely, and central EEG spike foci the least likely, to be associated with clinical seizures in children,98 although these data are biased by the inclusion of patients with the benign centrotemporal spike EEG trait.26 A different pattern of epileptogenicity is revealed by studies of posttraumatic epilepsy, which indicate that injury to the perirolandic region is most likely to give rise to seizures87 and that frontal and temporal lesions are more epileptogenic than those in parietal and occipital areas.31,91



Etiology and Basic Mechanisms

There are no specific causes or pathophysiologic substrates of the various anatomically defined epileptic disorders, with the exception of hippocampal sclerosis, which underlies most cases of MTLE (discussed in Chapters 13, 41, and 247). Any insult, lesion, or primary genetic abnormality capable of producing a localized epileptogenic region can account for the localization-related epileptic disorders classified by cerebral lobe. There are, for example, variants of the idiopathic benign childhood epilepsy with centrotemporal spikes that appear to originate in other cortical areas.122


Clinical Presentation

The symptomatic localization-related epilepsies discussed in detail below include only those for which there is reasonable documentation. Much of the information is derived from recently reported series using currently available technologies with localization often confirmed by surgery.


Clinical Characteristics of Seizures Related to Anatomic Region of Origin


Temporal Lobe Epilepsy

Wieser206 extensively evaluated a limited number of carefully studied patients and concluded that there were five different types of “psychomotor” seizures. Four of these were of temporal lobe origin. While this concept failed to achieve wide acceptance, much of what he concluded remains valid, and it did direct attention toward the possibility of different temporal lobe epilepsy syndromes.


Mesial Temporal Lobe Seizures

Currently, temporal lobe epilepsy is divided into two types: Those with seizures originating in the medial temporal structures (MTLE), and those with seizures beginning elsewhere in the temporal lobe (e.g., neocortical temporal lobe epilepsy or lateral temporal lobe epilepsy). MTLE refers to a specific subset of patients with seizures originating in the medial temporal lobe structures (mesial temporal lobe seizures [MTLSs]); MTLE with hippocampal sclerosis (MTLE-HS) comprises the great majority of patients with MTLE. Although MTLE-HS is characterized by many features that are not shared by conditions with MTLS caused by other lesions, the seizures themselves are indistinguishable. For purposes of this chapter on anatomic classification, therefore, MTLE, with and without HS, need not be further discussed here, as the phenomenology is described extensively in Chapter 247. MTLE, however, serves as a prototypical benchmark for localization-related epilepsies. As noted previously, one of the first steps used to identify other anatomic symptomatic focal epilepsies is to examine them in comparison to MTLE to determine how far they stray from the typical syndrome.


Neocortical Temporal Lobe Epilepsy

Since the lateral temporal cortex only comprises a portion of the temporal neocortex, the preferred terms here are neocortical temporal lobe epilepsy (NTLE) and neocortical temporal lobe seizures (NTLSs).

Several reviews of temporal lobe epilepsy do not even attempt to differentiate medial from neocortical temporal lobe seizures.102,133,208 A review specifically addressing the issue of neocortical temporal lobe epilepsy concludes that currently there is very little information about this type of epilepsy and that distinguishing seizure characteristics do not exist.203 Patients with neocortical temporal lobe seizure origin are rare, particularly if examples with obvious epileptogenic lesions are excluded. Although exact statistics are not available, a reasonable estimate would be that <10% of patients with temporal lobe epilepsy, who do not have obvious circumscribed lesions, have seizure origin in the temporal neocortex. Patients with seizures due to temporal neocortical lesions are also uncommon. In one study spanning 14 years from a large epilepsy center, only ten examples of neocortical temporal lesional epilepsy were identified.173

One study comparing mesial versus neocortical temporal lobe seizures found very little difference in terms of risk factors, demographics, and scalp EEG patterns.27 The only significant difference was increased lateralized memory impairment during intracarotid Amytal testing in patients with MTLSs. There was a trend toward early risk factors (below age 2 years) in the patients with MTLSs. Clinical seizure characteristics were not examined. An important contribution of this study is that patients with lateral temporal lobe seizure origin (none of whom had detectable lesions) had surgical outcomes as good as those associated with mesial temporal lobe seizure origin. This is in contrast to other studies reporting less favorable postsurgical results in patients with nonlesional lateral temporal lobe epilepsy.76,184

Extensive reciprocal connections between lateral temporal neocortex and medial limbic structures might explain why clinical features of seizures could be similar and possibly indistinguishable from both regions.22,69,207 Even the auras may require participation of both areas.22,69

Nevertheless, despite some generally negative results obtained in attempting to define lateral temporal lobe seizures, there are some features that may help identify them. Auditory, vertiginous, and complex visual hallucinations have been equated with lateral temporal origin.12,212 No one in a large “pure culture” of patients with MTLSs had this type of aura.55 One study comparing patients with hippocampal sclerosis and temporal lobe tumors, however, did report auditory, visual, and vertiginous auras in both patient groups.173 Three studies have noted that the motor manifestations observed in MTLSs are less common in lateral temporal seizures.52,67,173 There are some scalp EEG and neuroimaging findings that might help distinguish the two types of temporal lobe seizures.43,44,76,202

During the past 10 years, additional converging evidence has been provided that there are some characteristics that can help differentiate seizures associated with MTLE from those originating in the temporal neocortex.29,53,68,119,126,148 Patients with NTLSs have no history of febrile convulsions and have experiential auras (auditory and other sensory/experiential illusions and hallucinations, but not fear), early contralateral dystonic posturing in the absence of oral alimentary automatisms, early loss of contact and shorter seizure duration, and greater propensity to generalize compared to patients with MTLSs. Patients with MTLSs often have a history of febrile seizures (usually complicated), younger onset, evidence of hippocampal sclerosis on MRI, auras of visceral sensations, fear or olfactory hallucinations (rare), dreamy state, early oral alimentary automatisms, delayed contralateral dystonic posturing, and less tendency to generalize compared to patients with NTLS. All authors stress the importance of early findings in differentiating the two conditions. Many of the clinical seizure characteristics described in autosomal dominant lateral temporal epilepsy are the same as described here in neocortical temporal lobe seizures.132



Frontal Lobe Epilepsy

Patients with seizures originating in the frontal lobes are the second most common to have localized surgical resections.93,99,161,162,163,223 Frontal lobe seizures are not rare but, because of lack of familiarity with some of the clinical manifestations, they are often misdiagnosed.96,205,211,214,217,219,220 Before 1985, scattered reports described different types of frontal lobe seizures,4,63,65,125,151,193 but during the past 20 years, there has been heightened interest in frontal lobe seizures as reflected in the publications from two symposia specifically concerning this topic.34,90 A current PubMed search of frontal lobe seizures from 2000 to 2006 yields 907 citations.

Whereas MTLSs are all similar with differences representing variations on a theme, the same is not true for frontal lobe seizures. A number of entirely different seizure types are associated with frontal lobe origin. Initially, because of protean clinical seizure manifestations, frontal lobe seizures seemed to defy classification into subtypes,213 but recently some order has begun to shape this formerly chaotic situation. Frontal lobe seizures can now be grouped into six broad categories: Focal clonic motor seizures, asymmetric tonic seizures, frontal lobe hyperkinetic (psychomotor, hypermotor, complex partial) seizures, frontal lobe absence seizures, frontal opercular seizures, and frontal lobe seizures that closely resemble typical MTLSs, recognizing that there can be some admixture among the different types.32,33,92,99,205,211,212


Focal Clonic Motor Seizures

When focal clonic motor seizures occur in isolation, they are not associated with impaired consciousness and reflect ictal activity in the primary motor area. Focal clonic motor seizures are often part of other frontal and extrafrontal secondary activations of the primary motor cortex. They are discussed in detail in Chapter 44.


Asymmetric Tonic Seizures

Supplementary motor area (SMA) seizures are the classic type of asymmetric tonic seizures. SMA seizures were first described over 50 years ago.5,150,151 Renewed interest in this relatively uncommon seizure type is reflected in recent publications describing both clinical seizure manifestations and cortical stimulation studies.13,14,38,56,58,72,92,113,135,159,164,212,214,220 Despite the often dramatic and identifiable clinical seizure manifestations, the frequent absence of EEG abnormalities can still lead to the erroneous diagnosis of nonepileptic psychogenic attacks in patients with SMA seizures.58,96

SMA seizures can have subjective symptoms that include bilateral, contralateral, and ipsilateral somatosensory sensations of tightness, pressure, numbness, or tingling. This is fairly common, causing some investigators to prefer the term “supplementary sensorimotor area.”38 A nonspecific general feeling of constriction or tightness can immediately precede visible motor activity. Because the majority of patients do not report specific somatosensory symptoms, the preferred term remains SMA seizures.92 The objective manifestations of SMA seizures usually begin with the sudden, often explosive, assumption of a fixed posture, classically with the arm contralateral to the side of seizure origin abducted at the shoulder, externally rotated, and flexed at the elbow with the head and eyes deviated as if looking at the up-raised hand.4 The leg on the side contralateral to seizure origin can be held in rigid extension or can be flexed at the hip. Forced vocalization can occur, usually repeated vowel sounds rather than formed words. More often, there is speech arrest. Seizures are brief, lasting seconds, and only rarely last longer than 1 minute. Toward the end of the seizure, there can be clonic motor twitches of the hand or face. These seizures can secondarily generalize into tonic–clonic seizures. Even when generalized convulsive seizures occur, the postictal clinical and EEG suppression is often surprisingly short when compared to other convulsive generalized seizures.220 If they do not secondarily generalize, they stop as suddenly as they start. They can occur in clusters of many seizures per day. Nocturnal preponderance is common, but there are many exceptions. Patients with SMA seizures are often conscious during seizures. There seems to be no fundamental difference in terms of lateralization or clinical seizure characteristics that determines whether or not consciousness is impaired during SMA seizures. Responsiveness, however, often is not tested during these typically brief seizures. When given test phrases or words during these seizures with intense bilateral motor activity, one is often surprised to discover that patients clearly recall the test phrase during the immediate postictal period.

The classic motor manifestations of SMA seizures are probably the exception rather than the rule. Although to some extent representing variations on a theme, the motor manifestations of SMA seizures among different patients can differ widely, though they are usually stereotyped for a given patient.14,38,58,92,99,135,145,164,220 Motor manifestations in relatively well-documented examples of SMA seizures can include symmetric tonic or dystonic posturing of both upper and lower extremities; unilateral rigid straight tonic upper extremity posturing; adducted, flexed posturing of the upper extremity with the fist clenched; flailing, thrashing movements of the ipsilateral arm; kicking and stepping activity of the lower extremities; tonic/dystonic posturing involving the contralateral lower extremity; and athetoid dystonic movements of the contralateral hand, arm, leg, or both.214,217,220 SMA seizures can rarely present with automatisms without tonic motor activity.92,211

Startle epilepsy with asymmetric tonic seizures of presumed SMA origin has been the topic of several reports.36,92,144,177 SMA origin was documented in some, but not all, of the reported patients. Similarly, reflex seizures due to more casual somatosensory stimuli (i.e., movement or rubbing) have been attributed to SMA seizure origin,95,154 but SMA origin was not documented in any of these patients. In some of these patients seizures probably began in the parietal cortex.95

Both positive (tonic) and negative (atonic) motor activity have been attributed to SMA seizure activity.95,131 Gelastic seizures in the absence of mirth have been described as a part of SMA seizures.19,33 Finally, many of the features of SMA seizures seen in adults are also seen in children.13

Despite a wealth of reports on SMA seizures, some confusion persists, largely because asymmetric tonic seizures can be evoked practically anywhere in the neocortex.3,83,215,216 In many of the reports of SMA seizures, origin of seizures is not well documented. In a careful study of four patients with asymmetric tonic seizures, various MRI lesions, and good surgical results, seizure origin was documented in the SMA in only one patient.86 Two of these patients had seizure origin in the lateral frontal regions and the other patient had seizures beginning in the medial parietal cortex. All propagated to the SMA. These authors emphasized that asymmetric “SMA-type” seizures do not equate with SMA origin. These same authors, in a later report, examined clinical progression patterns in ten patients with presumed SMA seizures. SMA origin was based on clinical seizure characteristics and scalp EEG findings. Only three patients had intracranial studies, while two patients had parietal lesions on MRI. In an earlier report, they described two patients with parietal lesions and asymmetric tonic seizures.83 Another study of SMA seizures with tonic limb posturing found the posturing to have poor localizing and lateralizing results in 14 patients studied with intracranial electrodes. Eleven patients had surgery and only three became free of seizures.3 The lack of specificity of asymmetric tonic seizures has been well documented in earlier reports.64,214,215,216,220 Although it is
often assumed that these tonic motor manifestations represent spread to the SMA,3,83,84 this has seldom been proven215 and sometimes has been specifically denied.64

Considering the sometimes confusing and conflicting data associated with seizures of SMA origin, can they be diagnosed with any degree of accuracy? The scalp EEG in patients with SMA seizures is notoriously inaccurate and may be misleading or completely uninformative.15,33,54,179,184,200 A recent paper describing ictal and interictal magnetoencephalography in patients with medial/frontal lobe epilepsy reported good results but provided no evidence to verify the findings.179 Ictal SPECT was used to study eight patients with presumed seizure origin in the SMA.113 The importance of early (5 seconds or less) isotope injection was emphasized. Excellent results were realized in all eight patients. The ictal SPECT results in part explained the clinical seizure characteristics. Outcome was excellent in the five patients who had surgery. In a recent paper examining frontal lobe seizures in general, seven had seizure origins located in the SMA.92 While not specifically stated, ictal SPECT was used to direct intracranial electrode placements in many of the patients included in this study. Of the seven patients with SMA seizure onset who had surgery, five are seizure free and one has had brief nocturnal seizures. The one patient with a poor result had a major postoperative complication that resulted in death 2 years later. No seizures occurred after surgery, but the patient was in a persistent vegetative state. Although normal MRI is said to predict less favorable outcomes,50 the six patients from the Dartmouth series with good outcomes all had normal MRIs.92 A small series of patients with presumed SMA seizure origin studied with invasive recording all experienced good results following surgery.14

Only gold members can continue reading. Log In or Register to continue

Stay updated, free articles. Join our Telegram channel

Aug 1, 2016 | Posted by in NEUROLOGY | Comments Off on Anatomic Classification of Focal Epilepsies

Full access? Get Clinical Tree

Get Clinical Tree app for offline access