The localization-related epilepsies are epileptic syndromes in which seizures arise in a geographically restricted area within a part of one or alternate hemisphere(s) (partial or focal seizures).³⁰ Lesional localization-related epilepsies are the most common group of medically intractable epilepsies seen in adults and are also an important cause of intractable seizures in children.^58,162 Magnetic resonance imaging (MRI) is the most sensitive and specific modality for imaging lesions in patients with focal epilepsies and has greatly improved our understanding of the nature and frequency of these lesions.^19,31,72,85

Approximately 30% of patients who undergo surgical treatment for intractable epilepsy have a foreign tissue lesion detected on pathologic examination.⁶ There is a strong correlation between the site of the lesion and the site of the epileptogenic zone.^5,14,25,119 The identification of an epileptogenic lesion on MRI has been cited as being as predictive of a poor response to antiepileptic drugs.⁴³ Furthermore, such patients are more likely to become seizure free postoperatively than those in whom no structural abnormality is found.^94,127,144 In a randomized, controlled trial for surgery versus medical treatment for medically refractory mesiotemporal lobe epilepsy, 58% of those treated surgically in addition to drugs were seizure free at 12 months compared with 8% in the only conservatively treated group.¹⁵⁹

Focal structural lesions, particularly brain tumors, have been recognized as a cause of seizures since ancient times.¹⁴⁸ Hughlings Jackson⁷³ wrote extensively of the relationship between partial seizures and underlying focal brain pathology. He stressed that seizures could be the first and only manifestation of the tumor, that ictal behavior may predict the cerebral localization of the lesion, and that the severity and type of seizures were not predictive of the nature of the underlying pathology. Horsley in 1886⁶⁶ reported three patients who had been cured of seizures by surgical excision of an underlying focal structural lesion. In the earlier part of this century, it was particularly the work of Penfield and colleagues^87,121 at the Montreal Neurological Institute (MNI) and of Falconer and Serafetinides⁴⁸ in London that advanced our understanding of localized cerebral lesions and epilepsy and of the surgical treatment of these conditions. It has been only since the advent of modern neuroimaging, however, that the true importance of local structural lesions as a common, surgically treatable cause of both temporal and extratemporal focal epilepsy has been appreciated.⁶

For the purposes of this chapter, “specific lesions” are defined as discrete local (or regional) structural pathologies that are associated with chronic focal epilepsy. These local lesions most commonly occur in an otherwise structurally normal brain. More diffuse cerebral pathologies that may be associated with partial seizures but do not typically present as discrete mass lesions (e.g., diffuse neuronal migration disorders, Rasmussen encephalitis) are discussed elsewhere (Chapters 259 and 243). Mesial temporal sclerosis is discussed fully in Chapter 247.

A large surgical series of patients with focal structural lesions comes from the MNI; Table 1 summarizes the results of this series and other selected epilepsy surgery series along with reported pathology. Surgical series, however, are subject to significant biases because patients with a known mass lesion are more likely to be referred to an epilepsy center and then proceed to surgery.¹⁰¹ There is also a bias regarding the sites of the lesions reported in these series, with patients having temporal lobe seizures more likely to undergo epilepsy surgery. Furthermore, most of the patients in these series were collected before the advent of modern neuroimaging, when only large masses could be detected preoperatively. It is important to note that a number of pathologies associated with focal epilepsy, notably focal cortical dysplasias (FCDs) and dysembryoplastic neuroepithelial tumors (DNETs), were not recognized until the advent of high-quality magnetic resonance imaging.

Studies using high-resolution MRI have the potential to reduce some of the biases of these surgical series, and they may give a more accurate representation of the incidence and sites of occurrence of neocortical lesions in focal epilepsy. Because these series also come from large epilepsy referral centers, however, they are subject to some of the same referral biases. Ta-ble 2 summarizes four large, high-resolution MRI series. In comparison with the pathologic series, it is noteworthy that the proportion of extratemporal lesions is higher, as is the incidence of certain types of lesions, particularly FCDs and DNETs.

Epidemiologic population-based studies of epilepsy have the potential to reduce the biases seen in both the surgical and MRI series.^57,71 Hauser and Kurland,⁵⁸ in one of the largest such studies (Rochester, Minnesota, from 1935 through 1967), found a focal mass lesion in only 27 (5.2%) of 516 cases given a diagnosis of epilepsy during this period. Of these patients, 21 had brain tumors (18 primary and 3 metastatic), 4 had vascular malformations, and 1 patient had tuberous sclerosis. This study was also performed before the advent of modern neuroimaging, however, and therefore it is likely to have underestimated the true incidence of lesions because most are indolent and produce
no focal clinical impairment or electroencephalographic (EEG) slowing.¹⁵⁰

The incidence of cerebral tumors in children undergoing epilepsy surgery may be higher than in adults, with estimates as high as 46%.^13,101 Studies using MRI also suggest that before the age of 12 years, lesional epilepsy, particularly gangliomas and disorders of cortical development (DCDs), may be more common and mesial temporal sclerosis (MTS) less common.^85,119,122

The pathophysiologic mechanisms by which intracranial mass lesions cause chronic seizure activity is poorly understood, but a number of theories have been advocated. One proposed mechanism is “denervation hypersensitivity,” which results from the partial isolation of a part of the neocortex through tumor growth or brain scarring, thus creating enhanced excitatory status and epileptogenic potential.⁴¹ The degree of mass effect of the lesion, however, is unrelated to the incidence of epilepsy.^8,95 There is some evidence that there may be a familial predisposition to epilepsy developing with mass lesions, but much of the data on this are conflicting.^14,109

Low-grade tumors, which predominate in series of chronic lesional epilepsy, are rarely associated with pathologic evidence of hemorrhage, necrosis, inflammation, or ischemia, nor are they usually associated with significant mass effect.¹⁰¹ Cerebral tumors may induce changes in the surrounding neocortex that affect the balance of neurotransmitter levels, synaptic receptors (especially for N-methyl-D-aspartate [NMDA] or γ-aminobutyric acid [GABA]), or ion channels (e.g., increased leakage of axonal calcium or chloride channels).²¹ In support of this theory, Bateman et al.⁸ demonstrated an increased concentration of glutamine (the precursor to the excitatory neurotransmitter glutamate) in the gliomas of patients with epilepsy compared with gliomas from patients without seizures. Glutamine has been shown to be released and taken up by glioma cells.^106,156 Decreases in concentrations of the inhibitory
neurotransmitter GABA have also been demonstrated in gliomas of patients with seizures⁸ and in the surrounding non–tumor-infiltrated neocortex.¹⁰ Expanding tumors may also interfere with vascularization of the surrounding cerebral cortex, creating a region of relative cerebral ischemia having an increased epileptogenic potential.¹²⁶

In vascular malformations, pathologic studies have shown the presence of neuronal loss, gliosis, demyelination, and hemosiderin deposition in the surrounding cerebral cortex, which may act as a focus for epileptogenesis.^97,145 It has been suggested that the increased epileptogenic potential associated with these lesions is caused at least in part by the effects of repeated subclinical hemorrhage and resultant hemosiderin deposition.^21,145,169 Dodick et al.,³⁸ however, showed that hemosiderin deposition could not be the sole mechanism of epileptogenesis in patients with vascular malformations. Alternatively, ischemia in the brain surrounding an arteriovenous malformation (AVM) caused by arteriovenous shunting of blood may result in an area of epileptogenic encephalomalacia.¹⁶⁷

It has been appreciated for more than a century that the clinical history and ictal behavior may give a clue to the site of the underlying epileptogenic lesion.⁷³ Since the advent of MRI, however, it has become clear that a large overlap exists in the ictal symptomatology produced by lesions at different cortical locations. Lesions at any site may result in simple partial, complex partial, or secondarily generalized seizures. Complex partial seizures are often thought to indicate temporal lobe seizures, but in a study of high-resolution MRI in 129 consecutive patients with video-EEG-proven complex partial seizures, discrete neocortical lesions were detected in 58 (45%), of which 22 (37.9%) were extratemporal (15 frontal, 4 frontoparietal, 1 parietal, and 2 occipital).¹⁰⁸ Boon et al.,¹⁴ in 51 patients with lesions, found that although all patients with temporal lesions had complex partial seizures, 74% of patients with extratemporal lesions also had complex partial seizures. This study also found that although visual auras may give a clue to the presence of an occipital lesion, the nature of the aura was not otherwise useful in predicting the location of the lesion.

The clinical features, including seizure type, age of patient at onset, and duration of epilepsy, response to antiepileptic drugs, and findings of neurologic examination, are not useful in predicting the nature of the underlying lesion.¹²⁶ Patients tend to have a long history of seizures before surgery, and even patients with tumors do not commonly have an increasing frequency of seizures.^87,101,140 Careful neurologic examination in patients with low-grade tumors may occasionally detect focal signs, such as visual field loss, unilateral facial weakness, or progressive sensory loss or hemiparesis, but findings are normal in the vast majority of patients.^21,101

The identification of a lesion in a patient with intractable epilepsy is not sufficient grounds to proceed directly to surgical excision, because the zone of seizure onset may occasionally be at a site remote from the lesion.^5,14,20,85
Furthermore, patients with potentially epileptogenic lesions have been found after evaluation to have idiopathic generalized epilepsy or nonepileptic seizures. It is therefore important that all patients with intractable epilepsy have a comprehensive presurgical evaluation to ensure that the identified lesion is the source of the seizures. A great deal of caution, however, should be exercised before determining that the MRI lesion is not the source of the seizures; both extracranial and intracranial EEG can be misleading in lesional epilepsy,^5,23 and if the apparent epileptogenic zone is excised without the lesion, the results are likely to be poor in these patients.⁵⁰ Of course the reverse also holds true. The second important aspect of the presurgical evaluation is precise definition of the location and extent of the lesion, so that an operative strategy can be planned allowing maximal potential for a seizure-free outcome while minimizing the chance of a disabling postsurgical neurologic deficit.